Economic Re 6 rm in Subt-Saharan Africa____ Report No. 10434 edited by Ajay Cbhiibber Stanley Fischer A World Bank Symposi-um I Economic Reform in Sub-Saharan Africa A World Bank Symposium Economic Reform in Sub-Saharan Africa edited by Ajay Chhibber Stanley Fischer The World Bank Washington, D.C. 01991 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 Manufactured in the United States of America First printing December 1991 The findings, interpretations, and conclusions expressed in this study are entirely those of the authors and should not be attributed in any manner to the world Bank, to its affiliated organizations, or to members of its Board of Executive Directors or the countries they represent. Because of the informality of this series and to make the publication available with the least possible delay, the manuscript has not been edited as fully as would be the case with a more formal document, and the World Bank accepts no responsibility for errors. The material in this publication is copyrighted. Requests for permission to reproduce portions of it should be sent to the Office of the Publisher at the address shown in the copyright notice above. The World Bank encourages dissemination of its work and will normally give permission promptly and, when the reproduction is for noncommercial purposes, without asking a fee. Permission to photocopy portions for classroom use is not required, although notification of such use having been made will be appreciated. The complete backlist of publications from the World Bank is shown in the annual Index of Publications, which contains an alphabetical title list and indexes of subjects, authors, and countries and regions. The latest edition is available free of charge from the Distribution Unit, Office of the Publisher, at the address in the copyright notice or from Publications, World Bank, 66, avenue d'Iena, 75116 Paris, France. Library of Congress Cataloging-in-Publication Data Economic reform in sub-Saharan Africa / edited by Ajay Chhibber and Stanley Fischer. p. cm. Includes bibliographical references. ISBN 0-8213-2063-7 1. Africa, Sub-Saharan-Economic policy. 2. Africa, Sub-Saharan- Economic conditions-1960- . I. Chhibber, Ajay, 1954- II. Fischer, Stanley. HC800.E277 1992 338.967-dc20 92-908 CIP Preface The decade of the 1980's witnessed the adoption of conference was to bring together a group of econo- programs of economic reform in most countries of mists from African and non-African academic and Sub-Saharan Africa. Given the enormity of the Sub- research institutions, as well as international organi- Saharan economic crisis, some reforms-particularly zations and governments, to discuss economic macroeconomic reforms designed to reduce large reform in a professional forum. In addition to the budget and balance of payments deficits-were both authors, a range of distinguished African academics unavoidable and well understood; others-which and policy makers participated actively in the con- had the purpose of improving efficiency and restor- ference. ing growth-were crucial but less clearly The conference was organized by the World Bank understood. with substantial assistance from the Africa Econom- While the analysis of economic reform draws on a ics Research Consortium (AERC) in Nairobi. Finan- common set of economic principles, it must be recog- cial Assistance was provided by DANIDA-Den- nized that at the time that African economic reform mark's Development Agency and the Norwegian programs were being implemented, relatively little Foreign Ministry. The conference could simply not academic research had been conducted on how to have been put together without the assistance of the apply these general principles in specific African AERC, particularly Jeffrey Fine and Benno Ndulu. reform programs. As the reform process has pro- The support of Peter Eigen, the World Bank's Resi- gressed, it has become increasingly clear how vital it dent Representative in Nairobi was also critical to the is for both African economists and outside re- success of the entire effort. At the World Bank, the searchers interested in African development to pur- conference was very ably managed by Carol Sadler, sue further the economic analysis of African reform with help from Lesley Davis and Sook Bertelsmeier. programs. The manuscript was edited by Sandra Grain, with Research on the economic problems confronting desk-toppingbyMetadeCoquermontandLauraLee Africa is important, because the results of that re- Wilson. The figures and graphs were prepared by search will help policymakers in each country make Elizabeth Dvorscak and Catherine Cocak. Special better policy choices. But the importance of the re- thanks are due to Nga Lopez for managing the pro- search process, in which arguments have to be devel- duction of the manuscript. The preparatory work for oped logically and discussed professionally, extends the conference was managed by a steering committee beyond its technocratic impact; the informed public composed of John Holsen, Johannes Linn and discussion of policy that becomes possible as re- Stephen O'Brien, under the overall guidance of Stan- search results are disseminated improves the public's ley Fischer and Edward Jaycox. understanding of the choices confronting society, We hope that this conference, besides stimulating and thereby improves the policy process. further research on African economic policy issues, This volume presents papers on the economics of will also lead to a biennial conference on African reform in Sub-Saharan Africa presented at a confer- economic issues. The first in this series is scheduled ence in Nairobi in June 1990. The main aim of the for June 1992 in Lome, Togo. v Contents 1. Introduction 1 Ajay Chhibber and Stanley Fischer I. Exchange Rate Policy 2. Exchange Rate Policies and the Social Consequences of Adjustment in Africa 12 Patrick Guillaumont and Sylviane Guillaumont Jeanneney 3. Membership in the CFA Zone: Odyssean Journey or Trjan Horse? 25 Shantayanan Devarajan and Jaime de Melo 4. Comments, Ravi Kanbur and Stephen A. O'contnell 34 II. Parallel Markets 5. Exchange Reforn, Parallel Markets and Inflation in Africa: The Case of Ghana 39 Ajay Chhibber and Nemat Shafik 6. The Macroeconomics of the Unofficial Foreign Exchange Market in Tanzania 50 Daniel Kaufmann and Stephen A. O'Connell 7. Niger and the Naira: Some Monetary Consequences of Cross-Border Trade with Nigeria 66 Jean-Paul Azam 8. Comments, Vikram Nehru and Paul Collier 76 III. Fiscal Deficit and Expenditure Policy 9. Issues in Public Expenditure Policy in Africa: Evidence from Tanzania's Experience 80 Laurean Rutayisire vii 10. Poverty Conscious Restructuring of Public Expenditures 90 Marco Ferroni and Ravi Kanbur 11. Manufacturer's Responses to Infrastructure Deficiencies in Nigeria: Private Altematives and Policy Options 106 Kyu Sik Lee and Alex Anas 12. Comments, Nii Kwaku Sowa and Yao Kouadio 122 IV. Financial Sector Policy 13. The Informal Financial Sector and Markets in Africa: An Empirical Study 125 Ernest Aryeetey and Mukzvanason Hyuha 14. Mobilizing Domestic Resources for African Development and Diversification: Structural Impediments to Financial Intermediation 137 Machiko Nissanke 15. Monetary Cooperation in the CFA Zone 148 Patrick Honohan 16. Comments, F.M. Mwega and Diery Seck 161 V. Trade Policy 17. Commodity Exports and Real Income in Africa 170 Arvind Panagariya and Maurice Schiff 18. The Response of Firms to Relative Price Changes in C6te d'Ivoire: The Implications for Export Subsidies and Devaluations 182 John L. Newman, Victor Lavy, Raul Salomon and Philippe de Vreyer 19. Do Africans Pay More for Imports? Yes 198 Alexander Yeats 20. Comments, AdemolaAriyoandC.D.Jebuni 211 VL Regional Integration 21. Integration Efforts in Sub-Saharan Africa: Failures, Results and Prospects--A Suggested Strategy for Achieving Efficient Integration 217 Ali Mansoor and Andras Inotai 22. The Record and Prospects of the Preferential Trade Area for Eastern and Southern African States 233 Nguyuru H.l. Lipumba and Louis Kasekende viii 23. Enhancing Trade Flows Within the ECOWAS Sub-Region: An Appraisal and Some Recommendations 245 Ademola Ariyo and Mufutau 1. Raheem 24. Comments, Atchi Atsain and Sylviane Guillaumont Jeanneney 259 VII. Human Capital and Entrepreneurship 25. Human Resources, Technology and Industrial Development in Sub-Saharan Africa 263 Sanjaya Lall 26. Entrepreneurship and Growth in Sub-Saharan Africa: Evidence and Policy Implications 277 Ademola Oyejide 27. Comments, S.M. Wangwe and William F. Steel 285 VIII. Growth-Oriented Adjustment 28. Growth and Adjustment in Sub-Saharan Africa 287 BennoJ. Ndulu 29. The Liberalization of Price Controls: Theory and An Application to Tanzania 303 Paul Collier and Jan Willem Gunning 30. The Prospects for an Outward Looking Industrialization Strategy Under Adjustment in Sub-Saharan Africa 316 S. Olofin 31. Comments, Hafez Ghanem and C. Obidegwu 327 ix Contributors Alex Anas Stanley Fischer Department of Economics and Engineering Development Economics University of Illinois, USA The World Bank Ademola Ariyo Hafez Ghanem University of Ibadan World Bank Office Ibadan, Nigeria Abidjan, Cote d'Ivoire Ernest Aryeetey Patrick Guillaumont Department of Economics Centre d'Etudes et de Recherches sur le University of Ghana Developpement International Accra, Ghana Clermont-Ferrand, France Atchi Atsain Jan Willem Gunning Department of Economics Free University University of Abidjan Amsterdam, The Netherlands Abidjan, C6te d'Ivoire Patrick Honohan Jean Paul Azam Country Economics Department Centre d'Etudes et de Recherches sur le The World Bank Developpement International Clermont-Ferrand, France Mukwanason Hyuha Department of Economics Ajay Chhibber University of Dar-es-Salaam Development Economics Dar-es-Salaam, Tanzania The World Bank Andras Inotai Paul Collier Country Economics Department Institute for the Study of African Economics The World Bank Oxford University, England Sylvaine Guillaumont Jeanneney Shantayanan Devarajan Centre d'Etudes et de Recherches sur le Kennedy School of Government Developpement International Harvard University Clermont-Ferrand, France Marco Ferroni Charles D. Jebuni Social Dimensions of Adjustment Unit Department of Economics The World Bank University of Ghana, Accra, Ghana xi Ravi Kanbur Benno J. Ndulu World Bank Economic Review and Observer African Economics Research Consortium The World Bank Nairobi, Kenya Louis Kasekende Vikram Nekru Research Department Western Africa Department Bank of Uganda The World Bank Kampala, Uganda Machiko Nissanke Daniel Kaufmann International Development Center South Africa Department University of Oxford The World Bank Oxford, England Yao Kouadio Chukwumna Obidegwu CIRES Strategic Planning and Review Department Universite Nationale de Cote d'Ivoire The World Bank Abidjan, Cote d'Ivoire Stephen A. O'Connell Sanjaya Lali Department of Economics Institute of Economics and Statistics University of Dar-es-Salaam University of Oxford Dar-es-Salaam, Tanzania Oxford, England Samuel Olofin Victor Lavy Department of Economics Population and Human Resources Department University of Ibadan The World Bank Ibadan, Nigeria Kyu Sik Lee Ademola Oyejide Infrastructure and Urban Development Department of Economics Department University of Ibadan The World Bank Ibadan, Nigeria Nguyuru H.I. Lipumba Arvind Panagariya Department of Economics Country Economics Department University of Dar-es-Salaam The World Bank Dar-es-Salaam, Tanzania Mufutau I. Raheem Ali Mansoor Department of Economics Africa Technical Department University of Ibadan The World Bank Ibadan, Nigeria Jaime de Melo Laurean Rutayisire Country Economics Department Department of Economics The World Bank University of Dar-es-Salaam Dar-es-Salaam, Tanzania Francis Mwega Department of Economics Raul Salomon University of Nairobi Population and Human Resources Department Nairobi, Kenya The World Bank John L Newman Maurice Schiff Population and Human Resources Department Country Economics Department The World Bank The World Bank xli Nemat Shafik Philippe de Vreyer International Economics Department Population and Human Resources Department The World Bank The World Bank Diery Seck S.M.Wangwe University of Windsor Department of Economics Windscr, Ontario, Canada University of Dar-es-Salaam, Tanzania Nil K Sowa Alexander Yeats Department of Economics International Economics Department University of Ghana, Accra, Ghana The World Bank WilUiam F. Steel Industry and Energy Department The World Bank xiii 1 Introduction Ajay Chhibber and Stanley Fischer The Roman Road is the greatest monument ever raised to economic reform-Exchange Rate Policy, Parallel human liberty by a noble and generous people. It is built Markets, Fiscal Deficits and Expenditure Policy, Fi- broad, straight, andfirm. It joins city with city and nat- nancial Sector Policy, Trade Policy, Regional ion with nation. It is tens of thousands of miles long, and Integration, Human Capital and Entrepreneurship, always thronged with grateful travellers. And while the and Growth Oriented Adjustment. Great Pyramid, a few hundred feet high and wide, azves In this introductory paper, we first briefly describe siglht-seers to silence-though it is only the rifled tomb of the papers within each section, and then at the end an ignoble corpse and a monument of oppression and draw lessons from the contributions in the volume. misery ... Exchange Rate Policy Tiberius Claudius, Emperor of the Romans 10 BC - 54 AD The first section in this volume begins with a dis- cussion of the fundamental question of exchange rate I immediately empowered her to grant a large number of policy of fixed versus floating (or adjusting) ex- monopolies ... Six months later the removal of competition change rates. The adoption of adjustment programs in the monopoly trades, which included necessaries as well in a number of African economies, with devaluation as luxuries, had sent prices up to a most ridiculous height- as a center-piece, has led to an exceptionally vigorous the merchants were recoveringfrom the consumers what debate on this subject. they had paid in bribes to Messalina-and the City became The two papers in the first section deal with the more restless than at any time since thefamine winter ... I issue of the CFA Franc Zone. In the first, Patrick and fixed the prices, for the ensuing twelve months, of the Sylviane Guillaumont question the common IMF/- commodities affected ... However, as soon as any com- World Bank advice to use nominal devaluation to plaints reached me that a certain class of goods was not bring about a realignment of the real exchange rate. reaching the City in sufficient quantities I added another The authors do not deny the need to be competitive. firm to those already sharing the monopoly. Rather they argue that the route of nominal devalu- ation to achieving a reduction in the overvaluation of Tiberius Claudius, Emperor of The Romans the real exchange rate has excessively heavy social 10 BC - 54 AD costs. These costs arise from the very high inflation which has typically accompanied devaluations in African economies. They argue that the better way to achieve real exchange rate depreciation is through The papers in the volume are divided into eight expenditure reduction, which lowers the price of broad and interrelated topics that matter in African non-tradeables relative to tradeables. They argue 1 Introduction that the reduction in expenditure can, through effi- The first paper by Daniel Kaufmann and Stephen cient targeting, be controlled in a way that protects A. O'Connell focuses on three issues related to the the real income of the poor. parallel exchange rate in Tanzania. First, it examines While there is no question that internal price ad- the transmission of the parallel exchange rate into justment can in principle achieve exactly the same domestic prices, focusing on the effect of widespread real results as a devaluation, the CFA Franc Zone has shortages in rural areas in the early 1980s, and the not yet achieved the required domestic wage and structural change that occurred with the introduction price reductions. In the second paper on the CFA of the own-funds import window in 1984. Second, it Zone, Devarajan and de Melo show that the failure specifies and estimates a reduced-form equation for to adjust in the 1980's meant low or negative growth the parallel market premium that allows for both rates of both exports and GDP in the CFA zone coun- trade and portfolio effects and that takes into account tries. These countries have benefitted from price sta- the structural shift that occurred in 1984. The paper bility, but at the cost of severe balance of payments shows that as in Ghana (in the next paper by Chhib- problems (see table 1-1). ber and Shafik), both the trade and portfolio effects The experience of countries which have under- were important in determining the parallel market taken exchange rate devaluations has also been premium in Tanzania. However domestic prices did mixed. An important factorcontributing towards the not fully reflect the parallel exchange rate - rationing lack of effective devaluations has been insufficient- was effective in some markets, leading to a 'money ly strong accompanying fiscal and monetary cont- overhang" in 1979-83. The final section of the paper rol. As a result, in many countries an exchange rate- addresses the question of exchange rate unification. price spiral hasbeen set-off, as is evident in table 1-1. It points out that the own-funds scheme, while wel- This evidence has allowed policy-makers in CFA fare-improving, is a step away from unification of the Zone countries to argue that in practice devaluat- official and the parallel rates. ions have not generally worked in other African The second paper on parallel markets by Ajay countries, and that fixed exchange rates at least bring Chhibber and Nemat Shafik focuses on Ghana, the benefits of greater stability not only in prices but which has carried out one of the most thorough also in output. structural adjustment programs in Africa. Ghana's An important issue which was not addressed in increasinglyhighinflationratehasbeenattributedto this section-both in the papers and the discussion major devaluations of the official exchange rate. Us- that followed-is how precisely the fixed exchange ing a macroeconomic model estimated with Ghana- rate contributes to low inflation. At least two factors ian data, Chhibber and Shafik dispute this conclu- are present. First, on the supply side, the fixed ex- sion. Their results show that over the period 1965-88 change rate keeps the prices of imported inputs con- there is no direct relationship between the official stant. Second, the need to maintain convertibility exchange rate and inflation; prices had already ad- constrains monetary policy, typicallykeepingmone- justed to the exchange rate prevailing in parallel tary growth low. A further factor that is extremely markets. Their results also show that official devalu- important in considering the CFA experience is to ation had a positive effect on Ghana's budget. Reve- note that most other developing countries that have nue improvements came from three channels: the attempted to operate with fixed rates have not suc- higher grant aid disbursed at a more depreciated ceeded in maintaining capital account convertibility. exchange rate, a reduction in the subsidies that had Rather, they have typically resorted to capital con- accrued to importers through an overvalued ex- trols, which in turn led to the development of parallel change rate, and an increase in export taxes as cocoa markets. These are analyzed in the next section. farmers increasingly marketed their output through official channels. Official devaluations therefore did Parallel Markets not produce higher budget deficits, demand pres- sures did not spill onto the parallel market, and the The focus of the three papers on parallel markets exchange premium narrowed considerably. is on how to achieve unification of the official and Chhibber and Shafik's results suggest that the key parallel exchange markets. The prolonged overvalu- to the success of Ghana's adjustment program was ation of the official exchange rate in many countries the adequate level of foreign financing, combined has led to the widespread emergence of a parallel with a coherent set of fiscal policies. They argue that exchange system. Further, research shows that the although Ghana's inflation had structural causes in parallel exchange rate was the crucial determinant of the past, the acceleration in recent years is primarily tradeable goods prices in many African countries. a monetary phenomenon. It also reflects weakness in 2 Introduction Table 1-1. Inflation in Sub-Saharan Africa 1975-89 Recorded 1980-89 1975-89 Highest Industrial Countries 5.0 6.4 Asia 7.9 7.3 Sub-Saharan Africa 17.2 16.7 More than 20 percent Ghana 43.7 52.2 122.9 Sierra Leone 63.7 43.9 178.7 Somalia* 41.1 32.9 91.2 Sudan* 33.1 25.1 64.7 Tanzania* 30.5 24.1 35.3 Uganda 104.6 - 238.1 Zaire 58.8 62.1 101.0 Zambia* 30.8 245 55.6 Between 10-20 percent Botswana 10.5 11.0 16.4 Burundi 7.6 10.1 36.6 Gambia 17.9 15.1 56.7 Kenya 10.4 11.3 20.4 Lesotho 13.6 13.9 18.0 Madagascar* 17.2 14.3 31.8 Malawi 16.6 _ 33.9 Mauritius 7.6 10.8 42.0 Nigeria 20.5 18.9 40.9 Swaziland* 13.9 13.7 20.8 Zimbabwe 13.5 12.3 23.1 Under 10 percent Burkina Faso 4.1 6.4 30.0 Cameroon* 9.4 9.8 17.2 Central African Republic*" 3.9 - 14.6 Congo* 8.0 8.5 17.4 C6te d'lvoire* 5.3 9.5 27.4 Djibouti 4.2 - 18.1 Ethiopia 4.3 8.3 28.5 Gabon* 7.7 9.8 28.4 Liberia* 4.3 6.1 19.5 Niger 2.4 6.7 23.5 Rwanda 4.4 6.8 31.1 Senegal 6.5 6.6 31.7 Togo 4.0 6.3 22.5 Note: *For these countries averages are for 1980-88 and 1975-88. For these countries averages are for 1981-89. Source World Bank data the financial system that must be tackled to sustain currency of Niger (the CFA franc) is convertible. Al- reform. though the price of the Naira can be regarded as In the third paper, Jean-Paul Azam shows that in exogenous for Niger, as it does not adjust to clear the analyzing parallel markets it matters whether the foreign exchange market, it is not constant, and its country's currency is convertible ornot. Azam shows fluctuations have significant macroeconomic effects. how Niger can acquire convertible foreign exchange Azam shows that it has pass-through effects on the by its parallel trade with Nigeria, despite the latter CPI in Niger, and that Niger can in effect be regarded having an inconvertible currency. This is because the as a small price taker. The parallel exchange rate for 3 Itltroduction the Naira is shown to have a significant effect on real expenditure in Africa during the adjustment decade, cash balances in Niger, this is the monetary counter- especially in the social sectors. The authors ask the part of the balance of payments surplus that Niger pertinent question-how should governments pri- has had in its cross-border trade with Nigeria. oritize the use of scarce resources for poverty reduc- All three papers on parallel exchange markets tion? Their answer is that the decision should be show a significant impact of portfolio shifts on the made on the basis of both the direct effects of basic determination of parallel exchange rates'thereby needs expenditures, and the indirect effects of other erasing a general perception that parallel markets are income-enhancing policies and expenditures. It will driven only by controls on the prices and movement also depend on the weights attached to various com- of goods. These results show that the parllel capital ponents of the standard of living, and information on market shifts resources across borders in response to the fraction of social expenditures reaching the poor. changing rates of return. They lead to the important Using household survey data for Cote d'lvoire, Kan- conclusion that unification of exchange markets re- bur and Ferroni show how the government can set quires appropriate fiscal and monetary policies, levels, based on poverty considerations, for the con- which should take into account the de-facto imper- sumption of rice, housing, and education. fect control on movements of capital. A further im- The third paper in this section, by Kyu Sik Lee and portant message, emerging from the papers on AlexAnas,examinestheimplicationsofpublicsector Ghana and Tanzania, is that exchange rate unifica- inefficiencies for the response of the private sector in tion-official towards parallel-has significant fiscal Nigeria. It asks questions such as: How do firms benefits which can be anti-inflationary. The assump- respond to the constraints caused by deficient infra- tion here is that the Government is a net supplier of structure? What alternatives do firmns have, and foreign exchange to the rest of the economy. what do they cost? Is the private provision of serv- ices a viable alternative to their public provision? Fiscal Policy and Public Expenditure The paper also examines supply side questions, such as the causes of the failure to deliver adequate serv- One of the most important and controversial ques- ices. To what extent are such failures caused bya lack tions in designing a reform program is the extent of of capacity, or by poor operations and maintenance? the needed reduction in the fiscal deficit. Often the How do inappropriate pricingand user charges con- disagreement is not over the desirable size of the fisc- tribute to the problem? What are the options, in al deficit, but instead on the speed of adjustment and terms of investment, technology, institutions, regula- the composition of budget cuts-over how quickly tions, and financing, for remedying these failures? and in what areas (capital vs. current, social sectorvs. Based on empirical observations, Lee and Anas military, primary education vs. tertiary) expendi- suggest policy options for improving the provision tures should be cut. Of course, the current state of ex- of infrastructure services in Nigeria. These include penditure incidence methodology does not allow ve- regulatory changes to encourage the fuller use of ry precise answers to these questions. However, it is existing capacity (for example, allowing the sale of more often than not possible to suggest the needed excess private electrical power); participation of the direction of change. Policy distortions are usually so private sector in the supply of infrastructure-related egregious that the direction of reform and its rough services; and pricing policies that take account of the orders of magnitude are easily identifiable. The three presence of congestion, system failures, and vari- papers included in this section demonstrate this on ations (by firm size and location) in the private pro- three different fiscal and expenditure questions. vision of services. The first, by Laurean Rutayisire, focusses on the macroeconomic and debt implications of the fiscal Financial Sector deficit in Tanzania, and analyzes the sustainable level of the fiscal deficit. Rutayisire first examines the fac- Structural adjustment programs adopted in many tors underlying the increase in government spend- Sub-Saharan African countries have placed high pri- ing. He then discusses these factors in the context of ority on the liberalization of financial markets. Given various theories of the motivation behind govern- the shallowness of the financial markets in many of ment spending, and concludes that unless the gov- these economies, policies relyingheavilyon the mar- ernment's income increases, Tanzania's present level ket mechanism alone can, according to Machiko Nis- of govemment expenditure is unsustainable. sanke, generate unexpected side effects. For example, The second paper, by Marco Ferroni and Ravi Kan- interest rates are likely to be excessively high if deter- bur, starts with a brief look at patterns of public mined by market forces in very thin financial mar- 4 Introduction kets. Painstaking efforts at institution building are ment to exchange rate and price stability that mem- needed to achieve market depth, so that market bership represents-and that would be hard to sus- forces can operate beneficially. The simple switch of tain by unilaterally pegging their own currency. policy from financial repression to financial liberali- Membership offers other potential advantages. zation can by itself do little to revitalize such econo- Within a monetary union, capital might flow more mies. freely to where it is most needed. If the distribution Nissanke shows that because of the absence of of union benefits is reasonable, this could benefit all effective mechanisms for domestic resource mobili- members-even those who because of low capital zation in these countries, not only has the system productivity became net lenders within the union. been segmented into the infornal and formal sectors Moreover, the operation of monetary policy and but no interactions have taken place among formal the prudential supervision of the banking system institutions. The economies are faced with little or no might be more effective if the resources of several capacity to generate long-term credit or loan provi- small countries were pooled in a strong and inde- sions for self-sustained diversification efforts. In pendent Central Bank. In the third paper in this spite of the high liquidity in the overall banking section, Patrick Honohan asks whether these advan- system, the potential productive sectors in the econ- tages have been realized in the CFA zone? Honohan omy are starved of adequate institutional credit. argues that the experience has notbeen encouraging. In the second paper in this section, Ernest Aryeetey Despite the fixed exchange rate and an elaborate set and Mukwanason Hyuha argue that analysis of the of rules for avoiding over-expansion of credit, the impact of monetary and other economic policies in CFA zone has almost foundered in widespread bank many Africaneconomiesusuallyunderestimatesthe insolvency. The zone's institutional set-up seems role of the informal financial sector, and sometimes equitable, but in practice the burden of paying for leads to the pretence among policy makers that this losses appears to have fallen disproportionately on activity is inconsequential. Even where it is acknow- the poorer countries-whereas most of the non-per- ledged that a substantial informal financial sector formingcreditshavebeen made in some of thezone's exists, policy makers usually do not take into account more prosperous countries. its linkages with the fonnal financial sector. They argue that while the sector thrives and affects growth Trade Policy processes in several ways, it is usually unaccounted for in both economic analysis and economic statis- It has been argued frequently that while an indi- tics. vidual primary exporter may benefit from increasing Current interest in the role the informal financial its exports, an expansion of primary product exports sector can play in the economic transformation of by several developing countries is likely to lead to a many African countries derives from the dearth of decline in their terms of trade, export revenues and financial resources. Recent studies carried out in real incomes. The paper by Arvind Panagariya and Ghana and Tanzania and reported by Ernest Ary- Maurice Schiff presents a systematic analysis of this eetev and Mukwanason Hyuha (as well as other issue with respect to a commodity-cocoa-in which studies) have helped throw more light on the internal many African countries have a large share of world workings of the informal markets, as well as on the exports. It examines how real incomes and export relationships between them and the formal financial revenues compare under the existing and altemative markets. The authors argue on the basis of the evi- (Nash) taxes. It also examines the impact of export dence that the informnal financial sector in the region expansion resulting from increased efficiency on real is relatively large and often plays a crucial role in the income, export revenues and tax revenues under savings-investment process. Thus, for optimal finan- alternative tax regimes, and compares the effects of cial resource allocation, informal financial markets export expansion by African Countries with that by needb to be given due consideration in policy design. non-African countries. The shallowness of financial markets in many The paper shows that there are no easy answers to countries in SSA is obviously a function of macro- the choices facing primary commodity producers. economic and financial sector policies. However, it is An economist can devise a set of welfare maximizing also to a considerable extent due to the small size of Nash taxes, but the poor record of adherence to inter- these economies. One way to increase the size of the national arrangements on production quotas, export financial market is to join a monetary union. In con- quotas or prices means these taxes are unlikely to be sidering the merits of joining a monetary union, instituted or maintained. The entry of Asian cocoa small countries naturally value the credible commit- producers, notably Malaysia, makes coordination 5 Introdution more complicated. The entry of Asian producers also including the heavy fiscal costs, suggest that this shows that coordination is unlikely to work in the complicated administrative attempt to mimic a de- long run, since there are always potential new en- valuation will not work in the long-run. It should be trants. Thus withholdingproduction, or usingexport pointed out that East Asian countries have used ex- taxes to control exports, is likely to be self-defeating, port subsidies and devalued exchange rates as com- since new entrants will take over market share. The plementary rather than substitute measures: they best long-run option may be to pass international typically kept undervalued exchange rates even price changes on to the farmers and assist them in when they were using export subsidies. diversification. Numerous studies have examined the influence of The response of firms in Africa to changes in rela- market structure on performance in domestic mar- tive prices and their ability to expand exports as this kets of industrial countries. These investigations activity becomes more profitable has been central in show that prices and profits are higher, and resources the debate overthe use of export subsidies. This issue less efficiently allocated, in markets where aggres- was very relevant in the 1980s for many of the mem- sive competition is absent. Using techniques similar ber countries of the West African Monetary Union, to those in the earlier studies, Alexander J. Yeats who kepta fixed exchange ratewith theFrenchFranc examines the relative prices paid for iron and steel and decided not to devalue their currencies. Search- products by selected African and other developing ing for alternative means to alter the relative price of and developed countries. The findings parallel those exportables, and with the recommendation of the of the industrial country market studies. Typically, World Bank, some of these countries opted to mimic prices are higher in international markets that are a devaluation by increasing both import tariffs and more concentrated (less competitive), or that rely on export subsidies. Opponents of the policy claim that fewer trade contacts. is doomed to fail because supply response in Africa Yeats' analysis shows very high excess price mar- is very sluggish, export elasticities are low, and be- gins on iron and steel imports from France to 20 cause the heavy flow of resources through the gov- former French colonies in Africa. Over the longer emnment budget suggests that the policy is time in- term (1962-1987), the African countries paid an aver- consistent, undermining its credibility and discour- age premium of 20 to 30 percent over other import- aging firms from responding to it. ers. The losses from these excess prices came to about The second paper on trade policy, by Victor Lavy, $2 billion by 1987. This study also finds that the John Newman and Phillipe De Vreyer, addresses overpricing extends to other (non-French) African these issues and provides relevant empirical evi- countries. Former colonies of Belgium, Portugal and dence from C6te d'lvoire, the first country to adopt the United Kingdom still pay premiums of 20 to 30 the tariff cum subsidy program in 1986. Their empiri- percent on imports from the former colonial power. cal work is based on data from the six years that The three papers all point to substantial potential preceded the implementation of the program (1980- benefits from freer trade and greater competition in 1985). The paper focuses on modeling the response Sub-Saharan Africa. On the export side the attempt of firms to exogenous changes in export prices. The to manage exports of primary commodities through new elements introduced are the endogeneity of the commodity boards has led to a very substantial loss domestic price, and the explicit modeling and inte- of market shares. On the import side too, the evi- gration of domestic demand so that it is estimated dence indicates that managed trading relationships, jointly with the producer behavioral equations. The typically between parastatals and former colonial estimated model is then used to simulate the effects countries, have cost African countries dearly. A more of the subsidy program, compare it to a simulated open trading structure wouldbringsubstantial bene- devaluation, and finally, in light of the two simula- fits. tions, to evaluate the 1986-1989 export subsidy pro- gram of the Cote d'lvoire. Regional Integration Lavy, Newman and De Vreyer show that the com- bination of an export subsidy with an import tariff, The sixth section of the book deals with the impor- which mimics a devaluation, would counter some of tant and much discussed issue of regional integra- the short-run negative effects of an overvalued ex- tion.InthefirstpaperAliMansoorandAndrasInotai change rate. However, they point out that the scheme argue that attempts at economic integration among introduced in Cote d'lvoire in 1986 did not result in developing countries (EIDC) have generally failed. a uniform increase in export prices and has also They attribute these failures to (1) the costs of not involved heavy fiscal costs since then. Their results, allowing free trade for products produced in more 6 Introduction than one member country, (2) the emphasis on re- tional phase in July 1984. The major focus of PTA gional import substitution behind high barriers; and activities during this period was the preferential re- (3) the attempts at industrial planning to ensure an duction of tariffs, buyer-seller meetings to encourage equitable distribution of benefits among participat- intra-PTA trade, with the aim of increasing intra-PTA ing countries. trade and economizing on the use of foreign ex- The costs of these inefficiencies could not be sus- change. PTA has also emphasized the development tained after commodity prices collapsed and debt of transport and communication links among mem- burdens grew in the early 1980s. Since then most ber states. EIDC arrangements have either stagnated or suf- The record does not show an increase in the ratio fered setbacks. However, most developing countries of intra-PTA trade to total trade. Lipumbaand Kasek- maintain their membership in at least one EIDC in- ende argue that tariffs are not the effective barriers to itiative, and interest in the subject has been renewed trade. Rather, non-tariff barriers, particularly foreign by greater integration among industrialized coun- exchange allocation schemes and import licensing, tries, such as the US-Canada Free Trade Agreement are the main hindrance. Theauthors argue furtherth- and the EC 1992 single market agreement. at as member states liberalize their overall trade and Recent regional integration initiatives among de- payment regimes, it willbe important to reduce tariff veloping countries reflect the new consensus on the preferences in the promotion of intra-PTA trade. A importance of an outwardorientation. Theyalso reflect planning approach to a common market for the PTA a more pragmatic approach where flexibility of trad- region is not compatible witha market oriented econ- ing arrangements is more important than consensus omy, in the sense that the multinational allocation of among a large (often disparate) membership. This industries in unlikely to be feasible in a diverse re- reorientation of EIDC may be particularly important gion. Closer economic integration and expanded for SSA where economic fragmentation limits many trade will require factor mobility and currency con- of the opportunities for horizonal and vertical inte- vertibilityamong member states, as well as improve- gration that are available in Latin American and ment in transport and communication links and in- Asian countries. Further, EIDC offers the means for creased availability of trade and market information. the entrepreneurship talent that has developed in the On the other side of the continent, the formation more outward oriented economics to be made avail- and eventual commencement of operations of able in other member countries. ECOWAS in 1975 was seen as one of the surest means In Sub-Saharan Africa, the aim of the Lagos Plan of acceleating development in West Africa. The pa- of Action of 1980 for an eventual common market, per by Ademola Ariyo and Mufutau I. Raheem looks which was restated at the recent summit of the Or- at several aspects of ECOWAS. First, the consolida- ganization of African Unity, sets out an ambitious tion period for the operation of ECOWAS was ex- objective which will yield economic benefits pro- pected to last 15 years. Hence, this year marks an vided the common market has an outward orien- important milestone for evaluating its performnance. tation with substantially lower trade barriers than Second, most of the studies so far reported were those in place today at the national level. A comm- undertaken before the introduction of stabilization on market could yield efficiency gains as a result of programsby some members of the ECOWAS sub-re- greater internal (regional) competition as well as gion. Hence, the possible incremental impact of this through greater factor mobility. Inotai and Mansoor development as reflected in the structure and volume argue that to ensure that MFN trade liberalization of intra-ECOWAS trade is yet to be documented. proceeds in parallel with iptra-regional liberalizat- Third, little attention had been paid to the trade ion of labor and capital flows, a common external matrix amongst members of the community, a de- tariff should be avoided until tariffs are generally tailed analysis of which should enhance our under- low. The first step should be to eliminate non-tariff standing of current and prospective trade flows barriers on the basis of mutual concessions among within the community. Finally, attention had earlier participating countries, with a phased unilateral ex- been focussed on distribution, and less on produc- tension on an MFN basis to avoid maintaining pr- tion, of tradeable goods within the region. eferences that result in inefficient trade and invest- Ariyo and Raheem's findings suggest that there ment flows. has been a small increase in the volume of trade The second paper, by Nguyuru Lipumba and within the region, perhaps arising from the benefits LouisKasekende, analyses the rationaleand achieve- of stabilization prgrams which encourage intra-re- ments of the Preferential Trade Agreement of Eastern gional economic cooperation. The paper reveals that and Southern Africa (PTA) since it entered its opera- the structure of trade flows has broken the language 7 Introduction and geo-politicalbarriersespeciallyalongtheAnglo- can industrialization. The analysis of industrial de- phone-Francophone dimension. The paper's find- velopment requires a broader approach that takes ings also suggest that trade within the region is lim- into account not just macro-economic conditions and ited, mainlybecause so much of the required imports the incentive framework, but also the ability to re- are not produced within the region. This finding spond to incentives. This ability is determined by a formed thebasis for Adiyoand Raheem's suggestion mixture of entrepreneurship, industrial skills, tech- that the success of ECOWAS requires agreed produc- nological effort and institutional development. tion arrangements to increase the availabilityof some Lall suggests that a very large part of the explana- goods hitherto not available within the sub-region. tion of the African industrial experience lies in the This view is quite the opposite of that reached by small base of industrial capabilities and institutional Kasekende and Lipumba. support with which most countries started, and the Despite political efforts towards the goal of re- inadequate growth over time in these relative to the gional integration, the evidence in the three papers industrial facilities set up"in other words, that indus- on the topic in this volume is that official policies trial development was initially too ambitious. Data have, if anything, thwarted regional integration. The on education levels and technological effort are used good news is that despite these contra-integration to illustrate the gap between African countries and policies, unofficial trade has expanded, bringing the newly-industrializing countries of East Asia and about much greater de-facto integration than is sug- elsewhere. The policy conclusions drawn differ sig- gested by official statistics. This unofficial trade has nificantly from those commonly advanced, which contributed to reducing the economic costs of poor aim to provide "quick-fix' solutions to African indus- trade and pricing policies. trialization. Changes in price and trade policies are In future work the issue of regional integration not enough to foster industrialization-although must also be viewed from the angle of its implica- they are necessary to avoid mis-allocation of re- tions for the degree of openness of the overall trade sources and investment in wrong industries. Lall regime"the central question of trade diversion versus argues instead for longer-ternm educational and insti- trade creation. Will greater regional integration (a la tutional changes to enhance the quality of entrepre- the EEC) come at the expense of less openness to- neurship and the ability to respond to price signals. wards the rest of the world? Is that becoming neces- The second paper in this section, by T. Ademola sary in a world of managed trade-blocs? These are Oyejide, shifts focus from the question of human issues on which furtherresearchbyeconomists inter- capital development to the somewhat less tangible ested in Africa is clearly needed. but equally relevant issue of entrepreneurship. Against the background of poor economic perfor- Human Capital, Entrepreneurship and mance during the 1970's and 1980's, many structur- Industrialization al adjustment programmes are placing increasing emphasis on private initiative and market forces. The seventh section of this volume deals with the Advocates of the new policies assume that this new interrelated questions of human capital, industriali- orientation will improve the flexibility with which zation, and entrepreneurship. In the first paper, San- African economies can respond to rapid and unex- jaya Lall examines the problems of industrialization pected changes in the international economic envi- in Africa, and asks why industrial development re- ronment. The new orientation also implies an ex- mains at fairly low levels of technological sophistica- panded role for the private sector in the development tion, integration with the domestic economy, and process. competitiveness in world markets. Lall traces the However, there is increasing concem that the pri- poor performance of African industry to several vate sector's supply response appears to be inade- causes. External shocks and poor macroeconomic quate. This concern suggests the need for studies management in many countries have reduced the aimed at acquiring a deeper understanding of the supply of foreign exchange as well as domestic de- determinants of African entrepreneurship and its ca- mand for industry, causing severe dislocations. pability for response, as well as analysis of the factors Highlyinterventionist, inward-looking policies have which constrain the development of entrepreneur- created and sustained considerable inefficiencies, ship and the private sector in Africa. while holding back the emergence of industrial ex- Oyejide's paper makes a contribution to this broad porters. research agenda. Its primary purpose is to survey the Exogenous shocks and interventionist strategies available literature with the aim of identifying major do not, however, fully explain the problems of Afri- elements of the dynamics of private enterprise devel- 8 Introduction opment in Africa. The main focus of the paper is an The second paper on adjustment, by David Bevan, analysis of the characteristics of African entrepre- Paul Collier, and Jan Willem Gunning emphasizes neurship, followed by an examination of its pattern the need to re-examine the recommendations of ad- of upgrading graduation and growth, leading to a justment programmes in the agricultural sector. Gen- discussion of the constraints impeding enterprise eralized price controls on consumer goods, when development and growth. Oyejide shows that very enforced, give rise to shortages. Such policies have few firms in the informal and micro-enterprise sector been fairly common in Africa. This paper develops a have grown to become modern firms of any size. He theory which argues that shortages of consumer also shows that special targeted assistance programs goods have serious consequences for the rural econ- for these enterprises have not worked. The results of omy. First, the unavailability of consumer goods Oyejide's paper suggest that the emphasis in assis- changes the marketed supply response to price tance programs and interventions should not be on changes (a point that is well-known in the context of firm size, but on the constraints a particular industry the former socialist economies): higher crop prices faces. The technology in each industry thrown-up an can actually reduce crop sales. Second, depending optimal firm size which could be a number of small upon how the shortages are distributed, they may firms or one large firm which competes internation- induce a build-up in money balances which give rise ally. Size of firm is not the relevant issue. to an overhang problem in the transition back to market clearing. Third, the rural economy is liable to Growth and Adjustment implode into subsistence. Farmers reduce incomes because they are unable to make purchases, but in The last section of the book contains three papers, aggregate these income reductions lower the supply which examine the impact on growth of the adjust- of consumer goods in subsequent periods. ment strategy that has been followed in Sub-Saharan The first part of the Bevan, Collier and Gunning Africa. Benno J. Ndulu focuses on key trade-offs paper sets out the theory, usinga variant of a diagram between growth and macroeconomic adjustment in due to Malinvaud. The second part applies the the- the medium term consistent with longer term sus- ory to the economic reforms which have taken place tainable growth. The special characteristics of the in Tanzania since 1984. This paper provides a con- growth process in Sub-Saharan Africa are high- vincing-and surprising-rationale for the impor- lighted. Key among these is the fact that reasonable tant role for financing of consumer goods (as op- levels of investment during the last two decades were posed to capital and intermediate goods) imports as not accompanied by commensurate growth rates. part of the financing package that underpins adjust- Ndulu distinguishes capacity growth from actual ment programs in Sub-Saharan Africa. growth, and examines the factors influencing the In the last paper of the volume, Samuel Olofin wedge between them. Ndulu argues that this wedge examines the prospects for an outward looldng in- can be explained to a largeextentbyimport compres- dustrialization strategy in Sub-Saharan Africa. This sion. Two factors are responsible for the import com- is the Export Oriented Industrialization strategy pression: (a) the poor export performance; and (b) (EOI), or what is sometimes referred to as the G-4 increased debt burden. The paper discusses how for- model, following the remarkable success of its fore- eign aid can be used to relieve import compression. most practitioners, the so-called East Asian 'gang of Ndulu presents a simple growth model incorpo- four': Hong Kong, Souith Korea, Singapore and Tai- rating the above special features. The model high- wan. Most countries in Sub-Saharan Africa have be- lights the trade-offs between capacity expansion and gun to adopt a semblance of the EOI strategy as part capacity utilization in an import-compressed econ- of World Bank/IMF assisted Structural Adjustment arny. It explains the coexistence of high investment Programs (SAPs). Very few studies have been under- and low actual growth and discusses requirements taken to analyze the implications of SAP for indus- for breaking out of this scenario and implied policy tralization in Sub-Saharan Africa. choices. By modelling the financing of the fiscal gap, Given the successes of the Asian NICs, and the fail- it also highlights the trade-off between growth and ures of the inward looking industrialization strategy price stability in situations of heavy reliance on the (ISI) which many African countries have pursued, it inflation tax to finance growth. The role of foreign is natural to ask whether African countries should resources in filling financing gaps is also discussed pursue an Asian style EOI industrialization strategy. and fungibility in the use of such resources to support Olofin addresses the extent to which the EOI or G-4 measures for improved efficiency in the future is modelmaybeexpectedtoworkinindividualAfrican emphasized. countries. The paper points out that serious supply 9 Introduction constraints exists in most African economies, which expense of well-functioning informal financial mar- will not allow them to easily duplicate the East Asian kets. experience. It is therefore not enough to simply chan- Fifth, formal efforts at regional integration in Af- ge relative prices and expect immediate and sizeable rica have so far failed. However, regional trade export-led growth. Rather, active measures will have (mostly illegal) is much larger than is recorded in to be taken to develop supply response in Africa. official statistics. This suggests that current trade and investment policies hinder rather than foster regional Concluding Remnarks integration. African governments would help re- gional integration by encouraging more free trade What lessons can be drawn from this particular (even unilaterally) and by focussing on improving collection of papers and the comments that follow? the infrastructure for moving commodities across the No doubt each reader will draw her or his own con- various sub-regions such as Western Africa, among clusions, but we believe the balance of the evidence the SAADC countries and in East Africa. presented here warrants the following conclusions: Sixth, incentive policies are one important element First, a realistic exchange rate policy is crucial for of the package needed to restore growth, but they are growth. The exchange rate can sometimes be used not enough. They need to be accompanied by sub- temporarily as a nominal anchor to fight inflation. stantial investments in infrastructure - roads, ports, But fundamentally inflation control requires control telecommunication and power in order to reduce the over fiscal deficits and money growth. Using the cost of doing business. These complementarities be- exchange rate as a nominal anchor frequently ends tween incentive policies and the provision of infra- in overvaluation of the currency, which guarantees structure are well known and understood. An inter- low growth. Experience shows that most East Asian esting twist to the supply response story in the Afri- countries have kept their exchange rates extremely can context is the need for provision of consumer competitive, in order to boost exports and growth. goods (requiring extemal financing) and thereby re- African countrieswould need to match these policies ducing overall shortages to elicit the necessary sup- to remain competitive. ply response from the farm sector. Second, returns to investment are typically very Seventh and last is the issue of human capital. The low in Africa. This is due partly to poor policies but, level of human capital development has been very more so than elsewhere, it appears to be due to low-largely because African countries were subject inefficient management of public capital. One solu- to colonialism which did not encourage education. tion is to privatize industries and reduce the share of The consequences of this are widespread-they af- public investment in total investment. The other is to fect entrepreneurship, public sector management, encourage more joint ventures with foreign compa- and the adoption of new technology, as well as what nies, in both the public and private sectors, accompa- we now call govemance. In short it affects the ability nied by strongefforts at capacitybuilding to improve to grow and participate effectively in the global econ- overall managerial capabilities. omy. Its rect: ication will require substantial invest- Third, trade diversification-into processed and ments over several decades. This is not an area where industrial goods-is obviously necessary to get away quick fixes are possible. from the common problem faced by many African It is our hope as editors that this collection of countries of declining terms of trade for their major papers and the conference at which they were pre- exports. However, the way to diversify is not by sented will help encourage further research on Afri- taxing primary commodity producers as in the past. can economic issues. Each of the papers raises many This dries up investment in primary commodity ex- further questions. In addition, a number of issues ports without encouraging it in other sectors and important for African economic development have results in African countries losing market shares to not been touched on in this volume-among them, producers in Asia and Latin America. Rather the labor markets and employment, public enterprise emphasis should be on encouraging exports through reform, and the broader issue of governance. maintenance of appropriate exchange rates, and de- These critical issues, togetherwith furtherwork on regulation of the domestic economy, as well as fo- the topics that are covered in the present volume, de- cussed efforts at export promotion. serve further discussion, which can be taken up in Fourth, the role of the informal financial system in the next conference and the volume that will accom- resource mobilization and allocation needs to be rec- pany it. This volume also does not include a detailed ognized. Past financial sector policies have tended to examination of African agricultural issues on which concentrate on formal financial systems, often at the there is considerable literature as well as the question 10 Introduction of African debt which is covered in a recent World Ret&ences Bank publication (Husain and Underwood, 1991). 'The challenge for African economic policy, and for Husain, I and J. Underwood (1991): African Extemal Fi- researchers on African economies, is the restoration nance in the 1990's, World Bank, Washington D.C. of growth in the 1990's. More and better research can Mellor, J.W., C.L. Delgado, and M.J. Blackie (1987): Accel- play a crucial role in assisting policymakers and the erating Food Production in Sub-Saharan Africa, John public to understand more clearly what works and Hopkins Press, Baltimore. what does not. 11 2 Exchange Rate Policies and the Social Consequences of Adjustment in Africa Patrick Guillaumont and Sylviane GuillaumontJeanneney During the 1980s the majority of African countries duction must change in favor of tradable goods. This implemented structural adjustment programs desig- can be achieved in two ways and these can occur ned to ensure lasting balance of payments equilib- simultaneously. rium and the resumption of growth. The adjustment The first is through a reduction in the real exchange policies were based on a wide variety of economic rate, which is the ratio of the price of nontradable policy instruments in the context of diverse exchange goods to that of tradable. The real exchange rate rate systems. The social consequences of structural changes in line with the ratio between the price indi- adjustment policies are significantly influenced by ces in the country and abroad multiplied by the index the choices made in setting the exchange rate and of the nominal effective exchange rate. In this paper, regulating exchange transactions. We try to examine the International Monetary Fund principles of calcu- here the social consequences linked to the exchange lations of the effective exchange rate are used to rate policy implemented during the adjustment. express the cost of local currency in foreign exchange. Thepaperis divided intofourmainparts. Theaims Another way in which structural adjustment can and social implications of structural adjustment are be achieved is to increase productivity. When pro- summarized. The diversity of African exchange rate ductivity improves in the production of tradable systems is reduced to three main types and an at- goods, their profitability is increased directly. When tempt is made to show their implications for real it improves in thenontradablegoods sector, theprof- exchange rate decline and inflation. The social con- itability of the production of tradable goods is in- sequences of the different modalities of real exchange creased if the prices of nontradable goods decline. rate decline are drawn and the social consequences This is likely if the price elasticity of the demand for of exchange rate policy are analyzed through its ef- nontradable goods is low. fects on productivity in structural adjustment. Structural adjustment by definition has a socially beneficial aspect, since it is intended to restore the The Aims and Social Implications of Structural economic growth compromised by the external defi- Adjustment cit. However the two ways to achieve structural ad- justment outlined above have different social impli- In the case of a major disequilibrium in thebalance cations. Reducing the real exchange rate involves of payments currentaccount, it is generallynecessary strong control over demand and induces a transitory to bringabout an initial reduction in overall demand. reduction in national income and the standard of This stabilization reduces economic activity and in- living. Those whose incomes are linked to price evitably has a high social cost. To reduce the current changes in tradables, such as small farmers and en- deficit without sacrificing economic growth-to trepreneurs, should benefit at the expense of those achieve structural adjustment-the structure of pro- whose incomes depend on the price of nontradables, 12 Exchange Rate Policies and the Social Consequences of Adjustment in Africa such as local businesses and transport companies, rate, a parallel market developed between the na- and those with fixed incomes (wage earners) irre- tional currency (the naira) and the currencies of the spective of their sector of activity. Increasing produc- neighboring countries (West and Central African tivity enables the incomes of small farmers and com- CFA francs, which are convertible). The smooth func- panies in the tradable goods sector to increase with- tioning of this market is apparent from the fact that, out necessarily reducing the incomes of wage earners over the entire length of Nigeria's long frontiers, the and the earnings of companies in the home goods rate for the naira in CFA francs tends to be constant. sector. Increasing productivity requires more time Nigeria implemented its adjustment policy in two and sometimes leads to a reduction in employment, phases. The first phase was begun in 1983, without but it involves a smaller reduction in the remunera- assistance from international institutions. It con- tion of labor compared to reducing the real exchange sisted of a drastic contraction in public expenditure rate. and tightening of the quantitative restrictions on im- ports, without any change in the official exchange Three Exchange Rate Policies Aimed at Adjustment rate. Inflation remained high and the overvaluation of the naira became much more marked. The ex- Various exchange rate policies were implemented change rate on the parallel market, on the other hand, in Africa during the 1980s.1 These different situations showed a pronounced depreciation which more than can, for the simplicity of exposition, be classified offset the difference due to inflation with the neigh- under three major exchange rate policies: devalu- boring countries. ation in the presence of a significant parallel foreign During the second phase, beginning in 1986, Nige- exchange market, the standard case of devaluation, ria was assisted by international institutions. The and adjustment without devaluation. This classifica- new policy was characterized by a measure of liber- tion of major exchange rate policies is a simplification alization of foreign trade and prices, recourse to ex- compared with the diversity of actual exchange rate ternal capital, and flexibility in the official exchange systems and policies. Table 2-1 shows the different rate. Since September 1986 a system of foreign ex- currencies or units to which African currencies are change auctions has been operating, which at first pegged and those which have flexibility. In the fol- coexisted with an administered exchange rate. Since lowing, each of the above three cases is illustrated by 1987 the auction-set rate has been applied to all offi- the experiences of African countries. cial transactions. The parallel market for the naira has not disappeared, but the difference between the offi- Devaluation in the Presence of a Significant Parallel cial rate, set by the auctions, and the parallel market Foreign Exchange Market rate is now much smaller than it was in the previous phase. The majority of African countries have practiced The depreciation of the official rate was antici- strict exchange control, even applied to current trans- pated on the parallel market. For an appreciable part actions. This has resulted in a parallel foreign ex- of the economy the actual reduction in the exchange change market which is more or less tolerated, if not rate was more progressive than might have appeared encouraged, by the national authorities. The parallel from the movement of the official rate. As a conse- market covers an appreciable proportion of the for- quence the rate of inflation was not so high (see table eign exchange transactions. Access to official foreign 2-2). The ratio of the variation in the real effective exchange is limited and depreciation of the official exchange rate to that of the nominal effective ex- exchange rate appears as a policy aimed at unifying change rate, a ratio that Edwards calls "effectiveness the two exchange markets. Devaluation in the pres- of devaluation"-whatcanbecalled here the"appar- ence of a significant parallel foreign exchange market ent" effectiveness-was 93 percent. Thus the core of is the experience in Ghana, Guinea, Nigeria, Tanza- the reforms in Nigeria consisted not only of greater nia, and Uganda. In some cases the parallel market is flexibility in the official exchange rate, but also in a authorized; in others it is simply tolerated. trend toward unification of the market exchange rate. The situation in which the parallel exchange mar- ket is more or less tolerated can be illustrated by Nigeria. All along both sides of its far-flung borders, Standard Case of Devaluation there are concentrations of population belonging to the same ethnic groups which have traditionally This case is the one most frequently referred to in maintained intense commercial relations. Due to lim- adjustment theory. In this case the country conducts ited availability of foreign exchange at the official a policy of flexibility in its official nominal exchange 13 Exchange Rate Policies and the Social Consequences ofAdjustment in Africa Table 2-1. Exchange rate arrangements of African counties, 1988 Exchange rate pegged to a single currency a basket of currencies Country USS FF Rand SDR Other Flexible Benin x Botswana x Burkina Faso x Burundi x Cameroon x Cape Verde x Central African Republic x Chad x Comoros x Congo x C6te d'lvoire x Djibouti x Ethiopia x Equatorial Guinea x Gabon x Gambia x Ghana x Guinea x Guinea-Bissau x Kenya x Lesotho x Liberia x Madagascar x Malawi x Mali x Mauritius x Mauritania x Mozambique x Niger x Nigeria x Rwanda x Sao Tome and Principe x Senegal x Seychelles x Sierra Leone x Somalia x Sudan x Swaziland x Tanzania x Togo x Uganda x Zaire x Zambia x Zimbabwe x Source: IMF, Annual Report, 1988 14 Exchange Rate Policies and the Sodal Conequences of Adjustment in Afiica Table 2-2. Variation in the nominal and real effective exchange rates and inflation in Sub-SaharanAfrican countries that have implemented structural adjustment policies Ration of Variation in variation in Inflation rate Real effective nominal Variation in real effective during exchange effective real effective exchange rate adjustment period rateat start of exchange rate exchange rate to variation in (percentage) adjustment during during nominal period (index, adjustment adjustment effective Average Country Period 1970=100) period period exchange rate Total Annual Burundi 1983-89 142 -52 -44 0.85 35 6.2 Central African Republic 1980-87 110 -10 14 - 77 10.3 Congo 1981-87 96 -00.8 -003 0.38 2 22 C6te d'lvoire 1980-87 125 -12 -17 1.42 48 4.4 Gabon 1981-87 122 00.1 -2.7 - -0.9 -0.9 Gambia 1983-87 99 -69 -27 0.39 180 29.3 Ghana 1982-87 1,011 -99 -92 0.93 498 43 Guinea 1985-88 - -96 n.a. n.a. - - Guinea-Bissau 1983-87 74 -93 -88 0.71 581 61 Kenya 1980-88 91 -58 -33 0.59 121 10.4 Liberia 1980-87 83 9.2 0.5 0.05 28 3.5 1984-87 113 -25 -26 1.04 8 2.6 Madagascar 1981-87 120 -74 -58 0.78 151 17 Malawi 1979-87 86 40 -14 0.35 214 15.4 1985-87 101 -37 -26 0.70 43 19.4 Mali 1982-87 86 -5.5 -15 2.73 23 4.3 Mauritius 1978-88 94 -40 -26 0.65 180 10.8 Niger 1981-87 122 -5.5 -29 5.27 5.5 0.9 Nigeria 1984-88 318 -88 -82 0.93 41 14.1 1985-88 284 -87 -80 0.93 38 17.1 Senegal 1975-87 127 -10 -14 1.4 148 8.6 1979-87 110 -8 0 0 94 7.8 Sierra Leone 1982-87 104 -97 -34 0.35 2,395 90.3 1984-87 158 -95 -56 0.59 791 107.3 Somalia 1981-87 247 -94 -67 0.71 671 40.5 Tanzania 1982-87 172 -88 -63 0.72 297 31.7 1985-87 205 -81 -69 0.85 115 31.2 Togo 1977-87 114 -16 -25 1.56 74 5.7 Zaire 1978-87 248 -99.8 -77 0.77 4,594 53.4 Zambia 1982-87 104 -89 -67 0.75 328 33.7 Zimbabwe 1982-87 89 -49 -25 0.52 106 15.6 Notes: The countries included in the table are those which had one or several agreements with the IMF between 1980 and 1987. The nominal effective exchange rates are a geometric mean of the official bilateral exchange rate indexes (annual average of the cost of the na- tional currency expressed in foreign exchange) weighted by the official structure of imports (excluding oil) over the period from 1980 to 1986. Only convertible currencies are included. The real effective exchange rates are equal to the product of the nominal effective ex- change rates and the ratios of the annual consumer price indexes in the country considered and elsewhere.(geometric mean of the con- sumer price indexes caluculated with the same weighting as the nominal effective exchange rates). In case of Mali the GDP deflator was used because of the lack of a consumer price indez for the whole period of adjustmenL The adjustment periods have been defined by the commencement of the period of nominal depreciation of the currency for the countries that have devalued, and by the start of the de- mand restriction policy for the Franc Area countries and Liberia (which have not devalued). As the adjustment policy does not generally start at the beginning of the calendar year, and since the exchange rates are annual indexes, the variation percentages calculated provide orders of magnitude for the movements and not a percise meaurement. The ratio of variation in real effective exchange rate to variation in nominal effective exchange rate corresponds to what Edwards (1988) calls the effectiveness of the devaluation. 15 Exchange Rate Policies and the Social Consequences of Adjustment in Africa rate, and this rate is actuallyapplied in thebulkof all ment policies without devaluation. The official fixed foreign transactions. The parallel market is relatively rate is used for all transactions. Countries in this negligible, either because the official rate makes it category are Cote d'Ivoire, Mali, Niger, Senegal, possible for the supply of and demand for foreign Togo, the Central African Republic, Congo, and Ga- exchange to be kept in balance, or because clandes- bon. These countries are members of the CFA Franc tine operations are hindered by geographic factors or Zone. 2 vigorous repression. This is not the most common The fourteen African countries in the Franc Zone experience in Africa; however, it can be illustrated by have no national currencies. Except for Comoros, the exchange rate policies followed in Kenya, they are members of two monetary unions. This Malawi, Mauritius, Zimbabwe, and Madagascar. means that devaluation would require a joint deci- Since 1982 Madagascar has conducted a policy of sion by all the members of each union. Each currency dramatic depreciation of the nominal exchange rate. of the two distinct monetary unions is called CFA Under the liberalized import regime, when import- franc and is pegged to the French franc. The member ers applied for foreign exchange they paid a set fee countries have made monetary agreements with regardless of the degree to which their requests were France enabling them to maintain the convertibility met. As this fee system was not applied to priority of their currency. imports or to those financed by international assis- The counterpart of this guarantee of convertibility tance, it led to a sort of dual official exchange rate. It given to the CFA franc consists of the commitment of was replacedbytheliberalized import system,which member countries to conform to certain rules of is applied to most imports and does not include a fee. monetary policy, including a limitation on advances All requests for foreign exchange are met at the offi- by central banks to nafional treasuries. Franc Zone cial rate, which is adapted according to the Central countries are subject to certain institutional con- Bank's exchange reserves position. straints leading to monetary stability. The real exchange rate did not begin to decrease Despite the absence of devaluation of CFA francs, until 1984. Edwards' effectiveness of devaluation ra- the nominal effective exchange rate of Franc Zone tio was 78 percent. The decline in the real exchange countries decreased between 1980 and 1985 as a re- rate and a large inflow of international aid enabled sult of changes in the value of the French franc in the progressive removal of foreign exchange restric- terms of other currencies (see table 2-2). However, tions. All applications for goods imports are now during the 1980s the CFA franc appreciated strongly fully met. The strong depreciation of the nominal on the parallel market in relation to the naira and the effective exchange rate, which resulted in the decline cedi. When changes in the value of the naira and the in the real exchange rate, simultaneously induced a cedi are taken into account, the result is a smaller rapid increase in domestic prices. nominal depreciation of the CFA franc (Guillaumont In other countries, such as Kenya, Mauritius, and and Guillaumont 1989). Zimbabwe, the official exchange rate was the actual During the adjustment period monetary and exchange rate for the main part of transactions. These budgetary policies were restrictive. The increase in countries devalued an initial real exchange rate domestic prices was kept below that in the rest of the which had not increased, compared to that of 1970 world, and prices decreased in some years (see table (see table 2-2). They did so because of a need for 2-2).3 In Cote d'Ivoire, Niger, Senegal, and Togo the structural adjustment due to a terms of trade decline real exchange rate was reduced by more than the and to an increase in debt service. They devalued in nominal exchange rate. The ratio of effectiveness of several steps and succeeded in lowering the real the depreciation of exchange was greater than one. exchange rate. However devaluation caused a rather high inflation in which the price level more than The Social Implications of Real Exchange Rate doubled in the three countries during the adjustment Decline With and Without Devaluation period. This inflation resulted both from the rise of tradable goods induced by devaluation and from the Devaluation is an apparently simple way of reduc- monetary and fiscal policy which accompanied the ing the real exchange rate, which by definition de- devaluation. rives from the nominal exchange rate and from the ratio of domestic to foreign prices. However, devalu- Adjustment Without Devaluation ations must be accompanied by monetary and budg- etary policy designed to maintain overall demand During the 1980s other African countries, whose and to avoid an increase in nominal remunerations. currencies are convertible, have implemented adjust- Without devaluation, a decline in the real exchange 16 E&change Rate Policies and the Social Consequences of Adjustment in Africa rate can be brought about only by keeping inflation countries than in other African countries, it was simi- lower at home than abroad through control of de- lar in adjusting countries in both groups. Second, the mand. As world inflation is generally low, adjust- growth performance in the CFA Franc Zone is higher ment without devaluation is only possible in coun- than in other African countries, for the whole sample tries which did not experience a too high inflation and for the adjusting countries. These comparisons beforehand and in which the real exchange rate ap- do not seem to support the idea that the adjustment preciated only moderately. without devaluation in Franc Zone countries has This section looks at the social consequences of real implied a more restrictive policy than in the adjust- exchange rate decline when it is obtained with and ment with devaluation in other African countries. without devaluation. Transitional income, relative price, and inflation effects of real exchange rate de- The Relative Price Effects cline are assessed. Structural adjustment policy is based on a decrease The Transitional Income Effect in the real exchange rate, which alters the relative price for home goods in relation to tradable goods. Demand contraction is intended to induce real This often causes a shift in the ratio of urban to rural currency depreciation. Without devaluation, de- income, to the advantage of the latter. Since rural mand must contract in order to slow inflation down income is typically lower, this effect could reduce to a point where it is below world inflation. With urban bias. However this depends on the manner in devaluation, and if there are not inflationary antici- which prices and incomes are detennined in rural pations, the contraction of demand is designed only and urban areas. Here we look at how exchange rate to curb the inflationary effects of the improvement in policy affects rural and urban incomes. the trade balance resulting from the devaluation. A given decline in the real exchange rate would be Exchange rate policy and rural incomes. The adjust- less costly with devaluation than without. However, ment policy can have a preponderant influence on devaluation reduces the real exchange rate onlyif the farmer income throughagricultural prices. However, increase in tradable goods prices is not offset by the many of the rural poor do not produce tradable increase in domestic prices and wages. Othervise, a goods because they do not have enough land. They restrictive policy would be necessary to reduce the cannot benefit from higher prices and are gradually real exchange rate, as in the case where there is no becoming the laborers of the landowning farmers. devaluation (Guillaumont Jeanneney 1988). When a Moreover, there is a division of labor between the country is obliged to devalue, it is generally still in genders, with men responsible for tradable goods, the grip of inflation. If this inflation is not cur4bed by and women concentrating on home goods (Azam, restrictive monetary and budgetary policy, it be- Chambas, Guillaumont, and Guillaumont 1989, p. comes permanent and contributes to a new real ap- 19-20). preciation of the currency. It then becomes more The nominal level of agricultural prices, p, is ex- likely that economic agents will anticipate the effect pressed in the following (Guillaumont 1986, Bonjean of these devaluations on prices, making a restrictive and Marodon 1988): policy all the more necessary. , A simple test of the restrictiveness of the policy in (i) p African countries adjusting with and without de- valuation is given in table 2-3. GDP growth rates in which p'w is the international price expressed in during 1980 to 1987 are compared for Franc Zone foreign exchange, r is the exchange rate, a is the countries and other African countries. Within each public levy ratio, and m is the unit marketing and category, data is given for adjusting countries. In processing costs. The real price of the good sold, pr, order to control for other factors of growth (country or its unit purchasing power, depends on the nomi- size, climatic vulnerability, terms of trade, initial nal price and on the domestic prices of the goods debt, etc.) a cross-sectional model of the exogenous consumed by the fanner. The price index for domes- (non-policy) determinants of growth has been esti- tic goods consumed by the farmer is pc . Then, mated. Each country residual has been interpreted as an indicator of its growth performnance relative to the (2-2) pr= E = A A (1 - a) - A norm. pc pc-r Pc There are two main results. First, although the Let us call p'AC the foreign consumer price index observed growth was higher in CFA Franc Zone and r the product of the initial exchange rate r, and 17 Exchange Rate Policies and the Social Consequences of Adjustment in Africa the exchange rate index p. Then r - ro * p and expressed in local currency, provided that there is a decrease in the real exchange rate. A second require- A, r ment is that the increase in export prices be passed (2-3) Pr Aw * pc. r} (1 - a) - A on to the producer and not seized by the state, a p wc Pc stabilization board, or a marketing agent (the in- crease is not offset by an increase in a or m/pc). When The real exchange rate, RER, is by definition equal to there is no devaluation, the increase in the producer A A pc/Pw, therefore price involves a decline in either the public levy ratio or the unit marketing and processing costs. The im- p m pact of an increase in the nominal producer price is (2-4) pr, r w * E1 (1 - a) - -, . stronger, since there is no inflation induced by de- P wc rP * RER pc valuation. In rural areas the social consequences of adjust- Thus, the real producer price depends by defini- ment policies are not limited to the direct effects of tion on: the real intemational price of the product, prices on farmer income. Major indirect conse- p'w/'wc which could be called the international terms quences result from the manner in which the farmers of trade for the product; the real exchange rate, RER; react to the new prices. Higher prices and income the public levy ratio, a; and the real marketing and normally tend to encourage farmers to increase pro- processing costs, m/pc. duction; they often also provide the means to do so, A first requirement for a devaluation to increase as they facilitate access to inputs and equipment, the real producer price is that the rise in domestic particularly through greater borrowing power. This prices does not offset that in the intemational price favorable impact cannot occur if the infrastructure is Table 2-3. ObservedGDPgrowthrateandgrowthperfonnance, 1980-87 Growth perfornance Type of avergage Observed growth Case 1 Case 2 Africa South of Sahara a 1.8 0.0 -0.1 (25) b 1.5 -0.4 -0.4 c 1.7 0.4 0.1 Franc Zone African countries a 2.4 1.0 0.7 (9) b 1.9 1.0 0.5 c 2.7 1.3 1.0 Of which with adjustment program a 1.3 0.3 -0.1 (6) b 1.3 0.6 0.2 c 1.8 0.3 0.2 OtherAfrican countries a 1.5 -0.6 -0.5 (16) b 1.4 -0.8 -0.8 c 1.6 -0.9 0.0 Of which with adjustment program a 1.2 -0.7 -0.7 (12) b 1.4 -0.8 -0.8 c 1.5 -0.9 -0.5 Note: a) simple average b) weighted average c) median Growth performance is defined as the difference between observed growth and growth estimated as a function of 'environmental' fac- tors on a sample of 58 developing countries. For growth performance case (1) these factors are the following: population size in 1980; cli- matic vulnerability as measured by the instability in agricultural value added during 1970 to 1987 (weighted by the ratio of agricultural value added to GDP); variation in the terms of trade (during 1980 to 1987, relative to 1975 to 1979), weighted by the exports to GDP ra- tio; and debt structure (private debt to GDP in 1980). Growth performance case (2) includes all of the above factors and a dummy vari- able for major political disturbances. The six Franc Zone countries with adjustment programs are Central African Republic, C6te d'Ivoire, Mali, Niger, Senegal, and Togo. The twelve other African countries with adjustment programs are Ghana, Kenya, Madagascar, Malawi, Mauritania, Uganda, Sierra Leone, Sudan, Tanzania, Zaire, Zambia, and Zimbabwe. Source: Gross data from World Bank tables 1988-89. 18 Exchange Rate Policies and the Social Consequences of Adjustment in Africa run down or the marketing and credit channels are official agricultural prices-to sell their products not organized, especially if the farners cannot sell once again through the official channels, which gives their output and/or cannot buy anything in ex- them an advantage since clandestine marketing in- change for their increased monetary income. Such a evitably involves risks and costs. A return to the shortage of goods due to lack of foreign exchange official channels did in fact occur in Guinea and does not occur in the adjusting countries benefitting Ghana (for Ghana, see Heller et al 1988). from theconvertibilityof theircurrency, as in theCFA The pattern of change in the real producer prices Franc Zone. in the countries of the Franc Zone, where nominal In terms of country experiences, a special case prices are usually administered, is an interesting test occurs in the case of devaluation in the presence of a of the possible impact of an adjustment policy with- significant parallel foreign exchange market, when out devaluation. In most cases during the 1980s, the the farmers can sell their harvests in neighboring official producer prices for export crops were first countries outside the official marketing channels. increased and then maintained, whereas the trend in They are then paid in the currency of those countries, international prices was becoming unfavorable. As a which they exchange on the parallel market at a rate disinflation policy was simultaneously being imple- which, especially if that currency is convertible, is mented, real producer prices changed relatively lit- much higher than the official rate. Under such cir- tle; real per capita income in the country frequently cumstances a devaluation of the official exchange declined (see table 2-4). This was the trend until 1985 rate does not have a major direct effect on their because of the fall in the French franc exchange rate income, but can nevertheless encourage the farmers against the dollar. It continued throughout the period -if the depreciation of the currency is reflected in the because of a decline in the government levy ratio at Table S. Movement of real official producer prices in the African Franc Zone countries that have implemented adjustment prices 1980 1981 1982 1983 1984 1985 1986 1987 1988 1981 1982 1983 1984 1985 1986 1987 1988 1989 Central African Republic Coffee 100 96 114 78 79 73 77 - - Cotton 100 88 92 92 95 97 - 82 - C6te d'Ivoire Cocoa 100 92 86 94 97 101 95 90 83 Coffee 100 92 86 94 98 101 85 90 83 Cotton 100 87 86 101 111 110 103 98 90 Palm Nuts 100 138 128 121 116 145 136 129 119 Niger Groundnuts (shelled 100 92 97 100 101 124 135 149 Cotton 100 105 157 145 134 145 150 137 - Cowpeas 100 118 138 149 153 167 129 138 Millet 100 114 130 120 138 112 - - - Paddyr 100 104 113 116 107 115 91 100 - Senegal Groundnuts (unshelled) 100 132 112 100 90 102 96 100 80 Cotton 100 107 94 64 75 95 90 93 95 Millet 100 118 101 99 98 100 95 98 - Paddy 100 117 100 104 104 118 111 116 118 Togo CocoaL 100 85 80 86 97 109 114 144 - Coffee 100 90 89 100 113 133 140 140 Cotton 100 90 82 87 107 127 122 122 Note: The nominal prices have been deflated by the African consumer price index (annual index relating to the start of the season). Sources: BCEAO, Bulletin d'information et statistiques, and Ministry of Cooperation and Development - Caisse Centrale de Coop6ra- tion Economique, Bulletin de conjoncture, various numbers. 19 Exchange Rate PolUcies and the Social Consequences of Adjustnent in Africa the same time that the countries were attempting to Table 2-5. Declining real wages in adjusting cut theirbudget deficits. Nevertheless, in some coun- countries tries, after international ptices for some commodities (coffee, cocoa and groundnuts) fell again, maintain- Chaunge in minimum ing producer prices has had to be dramatically recon- adgustment sidered, owing to the government's financial straits Country Adjustmcnt period (percentage) (for example in C6te d'Ivoire). C6te d'lvoire 1980-1987 -25.6 Niger 1981-1987 -5.5 Exchange Rate Policy and Urban Incomes Senegal 1979-1987 -11.6 Urban remuneration is especially affected by the Togo 1977-1987 -27.0 adjustment policy when it decreases the real ex- Source. BCEAO, Notes d'information et statistiques (1985-88). change rate. Urban production of nontradable goods and services (financial services, transportation, to adaptbecause of its greater flexibility in determin- trade, public works and construction, etc.) is rela- ing the real remuneration of workers and the number tively high. Moreover, in general a larger share of of peopleitemploys.Theinformalsectorparticipates remunerations in urban than in the rural areas con- in an informal structural adjustment. One of the most sists of fixed income (business and government sala- visible aspects of the urban consequences of adjust- ries, rents, etc.). In the event of devaluation, their real ment policies is precisely growth in the informal value is reduced by the increase in prices. That effect sector. This phenomenon can be heightened when is attenuated when there is a parallel foreign ex- the real exchange rate is decreased without devalu- change market where the currency is depreciated ation, as in the Franc Zone countries (for the case of before the official devaluation. Niger, see CERDI 1988a). In those countries, a drop Without devaluation, a decrease in income results in nominal wages (or unit remunerations) is more from the restrictive policy aimed at blocking or even easily achieved in the informal sector than in the diminishing the nominal level of remunerations. modern sector, under the pressure of foreign compe- This is achieved through monetary policy for busi- tition or the effect of a decline in domestic demand ness salaries, budget policy for government salaries, (as far as home goods are concerned). and regulation of minimum salaries. The price of tradable goods increases in tandem with world infla- The Inflation Effects tion and sometimes even more rapidly when those prices-after having been controlled-are deregu- Adjustment policies in Africa have led to inflation lated. Thus, a marked drop in real wages was seen in rates that differ widely from one country to another, the African countries that implemented adjustment depending on how devaluation is used. There is vari- policies (see table 2-5, for a more detailed analysis of ation ranging from nearly stable prices to rates of inf- the situation in C6te d'Ivoire, Levy and Newman lation in the vicinity of 100 percent a year (table 2-2). 1989). One effect of inflation is to increase the burden Another important determinant of the urban imposed by progressive income taxes, and to reduce standard of living during adjustment is the level of that imposed by lump taxation. This is true insofar employment. It is often reduced by the stabilization as tax brackets and lumps are not indexed. Wage policy that is generally a prerequisite of structural earners in developing countries often pay a gradu- adjustment. The decrease in demand primarily affect ated tax, even at comparatively low income levels, urban activities because of the high proportion of while some rich merchants are lump taxed. In this home goods production. Structural adjustment in- case, the inflation that accompanies an adjustment volves a shift of jobs from nontradable goods produc- policy involving devaluation might well add to the tion towards production of tradable goods. distortions between wage earners in the modern sec- Analysis of the urban social consequences of ex- tor, whose tax burden would increase, and individu- change rate policy must take into account the dual- als active in theinformal sector, whowouldfind their ism within the urban manufacturing and tertiary tax obligations reduced. sector. In addition to the modern sector there is an Another general consequence of inflation is its informal sector, which remains largely outside offi- wealth redistribution effect. The increase in prices cial regulation and taxation. The modern sector has drastically reduces the value of monetary and quasi- in general been more affected by adjustment policies monetary balances, while real estate assets maintain than has the informal sector. The latter is better able their value. The tax charge payable on cash balances 20 Erchange Rate Policies and the Social Consequences of Adjustment in Africa works to the advantage of those who benefit from As shown a quarter of century ago by Balassa money creation (the state and private borrowers), (1964) and more recentlyby Kravis et al (1975), coun- which is why it is described as an inflation tax. It is a tries with higher relative income per capita have a type of taxation that appears to weigh proportionally higher real exchange rate. In the long run a relatively more on poor groups, since a higher percentage of rapid rate of growth involves an increase of the real their wealth, if not most of it, is in monetary form. It exchange rate. In this perspective, the strong decline is also possible that the inflation tax hits small farm- of the real exchange rate in some African countries, ers harder than urban workers because of the usually such as Ghana and Madagascar, although considered seasonal nature of inflows of farm income and the as a success of adjustment policies, is a symptom of resulting need to hold cash balances. However, this the failure of long run development. situation may be mitigated when farmers have the It appears that the relation between the real ex- opportunity to set up savings in the form of goods change rate and average labor productivity runs in (grain, livestock, etc.). Urban groups would be less two directions. In the short and medium term, the affected by inflation than are their rural counterparts real exchange rate decline may induce a decline in because they are more likely to invest their money labor productivity. In the long tenn, a decline in the profitably and have more monetary debt. Certain relative level of labor productivity implies a decline people only in urban centers are in a position to profit in the real exchange rate. from devaluations by investing in foreign exchange. Inflation may cause urban real estate prices to rise, Fixed Parity as an Incentive to Improve Efficiency which would result in reduced urban property taxes during the lag in reassessing the tax base. From a social standpoint, decreasing the real ex- change rate is reflected by a decrease in the remu- T-he Social Consequences of the Exchange Rate neratiorr for labor and a change in the employment Policy Through its Effects on Productivity Growth structure. In the case of devaluation, companies can easily protect their profits by adjusting their prices; After stabilization and after getting the prices they therefore have less incentive to enhance their right, productivity increase can be seen as the next management efficiency. Conversely, the constraint of stage of adjustment. When there is a large nominal stable prices is an incentive to reduce costs. depreciation of the currency, adjustment is sought The Franc Zone countries lack the means of raising predominantly through a decrease in the real ex- prices in the agroindustrial export subsectors. They change rate. In countries where price distortions re- have made significant productivitygainswhich have quire only a moderate decline in the real exchange contributed to adjustment. Although international rate, adjustment programs rely more on productivity prices for cotton (in Chad, Central African Republic, increases. Mali, and Togo) have been low, producer prices have In the following, the relation between exchange not been reduced much. This has been possible due rate policy and productivity is explored. We discuss to improvements in the productivity of extension, the negative effects of reduced real wages, fixed par- processing, transportation and marketing. ity as an incentive to improve efficiency, exchange Increasing productivity does not entail a decrease rate policy and private investment, and exchange in remunerations, but generally an increase in the rate policy and public social expenditures. intensity of labor and a modification in the employ- ment structure. Efforts to increase productivity may Real Exchange Rate and Labor Productivity prompt some businesses to make layoffs that are not immediately offset by new jobs stemming from the Reduction of the real exchange rate has been increase in production. The change in the employ- viewed in the foregoing as an objective. But in many ment structure may lead-at least temporarily-to African countries the large reduction in real wages an increase in urban unemployment. Therefore pro- has impacted negatively on the motivation of wage ductivity improvement has at least a temporary so- earners. These effects have not been offset by im- cial cost, which is less drastic than that resulting from proved social protection. In low-income African a decrease in the real exchange rate. countries, wage earners' ability to work has been The social cost of increased productivity is directly adversely affected by poor nutrition and health.4 felt by the businesses whose productivity increases, There are similar negative productivity effects in the which explains their resistance to it. An open ques- nontradable goods production sector. tion is whether in Africa reducing employment is 21 Exchange Rate Policies and the Social Consequences of Adjustment in Africa socially more or less costly and politically more or GDP, p is an index of the general price level, and ps is less feasible than lowering wages. an index of the price of social goods. This equation Moreover, when adjustment is based more on an shows that the level of social expenditure depends increase in productivity than a decrease in the real on four factors: GDP per capita, the ratio of govern- exchange rate, the informal sector has a more impor- ment expenditure to GDP, the structure of tant role. This is because of its greater flexibility in government expenditure, and relative prices (see labor remuneration and the number of people em- CERDI 1988b). All four factors may be affected by ployed. This would explain why during the adjust- exchange rate policies. ment period Franc Zone countries seem to have ex- The ratio of government expenditure to GDP is perienced an especially rapid expansion of their in- determinedbytherateof taxationandbythesustain- formal sector. able budget deficit. A drop in the exchange rate, whether sought through devaluation or not, implies Exchange Rate Policy and Private Investment strict control of the budget deficit. This is why the level of government expenditure is limited by the The social consequences of structural adjustment level of government revenue. depend on the country's ability to sustain growth. Exchange rate policy can exert a significant impact Lasting growth requires a sufficient level of invest- on the tax ratio. In the African countries, taxation on ment especially in the modem sector. Therefore as far tradable goods, and particularly on extemal trade, as the exchange rate policy influences investment, it appears to be heavier than taxation on home goods has a long term social impact. and incomes. A drop in the real exchange rate has a The influence of exchange rate policy on invest- favorable effect on the tax/GDP ratio. ment depends on the relationship between inflation This pattem can be seen in the African countries and savings. It depends also on the effects of the real that have devalued. When the drop in the real ex- exchange rate level and instability in the inducement change rate has been sharp, public revenue gener- to invest. Investment maybe increasedbygetting the ated by foreign trade has increased more quickly prices right and by an appropriate real exchange rate. than the general level of prices. Increases in public However investment is likely to decrease due to revenue drawn from trade have tended to be greater exchange rate instability, such as that resulting from in countries with a stabilization board system and a floating exchange rate or discontinuous devalu- where the official producer prices have not reflected ations. Moreover, strict exchange control makes for the increases in the local currency export prices. The low investment and inefficiency because of its ad- expected favorable effect of devaluation on agricul- ministrative complications and the risks of break- tural income and output has then been moderated downs in supplies associated with it. and sometimes even reversed. Hence, there is a trade-off between generating more revenue in order Exchange Rate Policy and Public Social Expenditures to finance government expenditure and increasing primary agricultural income. Even if devaluation The exchange rate policy exerts a range of effects allows for greater fiscal pressure, this effect can be on the real level of social spending per capita. The compensated by an output decline caused by pro- level of public social expenditures has a direct social gressive taxation and variable income imposition. impact. It also effects workers' productivity, which is A change in the rate of government expenditure sensitive to the workers' health and education D/Y has repercussions on the level of social expen- status.5 diture. This depends on the structure of government The factors which in the course of adjustment act expenditure DID and the relative price of social on the per capita level of social expenditure, expr- goods ps/p, which are both influenced by the ex- essed in constant prices, can be examined by means change rate policy. of the following equation: Because of debt service obligations expressed in foreign currencies, exchange rate policy acts directly Ds Ds D Y P on the structure of government expenditure. If there (2-5) p is a devaluation, the relative share of public expendi- P . ps D Y P p ps ture earmarked for debt service increases automat- ically. The proportion available for other spending, where Ds is govemment social expenditure, D is particularly social expenditures, shrinks. In the total government expenditure, P is population, Y is course of the adjustment the authorities often have 22 Exchange Rate Policies and the Social Consequences of Adjustment in Africa to cut back more on capital than on recurrent expen- exchange rate is probably weaker when devaluation ditures and more on operating expenses than on is used and there is some money illusion. However, labor costs. this advantage of devaluation and the extent of In countries facing foreign exchange shortages, money illusion are often overstated. The relative reducing expenditures is also a matter of import price effect of a given decline of the real exchange rate content. Under a given nominal government (non- does not seem to differ greatly whether the country debt) expenditure structure, exchange rate policy adjusts with or without devaluation. Real exchange may have an effect on the real level of social spend- rate decline has the favorable social impact of reduc- ing. A decline in the real exchange rate, by altering ing urban bias. A main social difference between the relative prices, changes the distribution of govern- two ways of decreasing the real exchange rate results ment expenditure. The volume of social expenditure from the effects of inflation. is affected according to its content of tradable goods. Productivity improvement is the other major way Education expenditure has a relatively low tradable to achieve structural adjustment. Adjustment with- goods content, unlike health expenditure, which in- out devaluation leads to greater emphasis on this cludes a quite high proportion of tradable goods and alternative. Decreasing the real exchange rate in- services (drugs, equipment). For this reason, and also volves decreasing labor remunerafion, which may because it involves a smaller labor component than have a negative impact on labor productivity. Main- education expenditure, real health expenditure is taining a fixed parity is an inducement to efficiency. likely to be particularly affected during adjustment. Hence, once major distortions are eliminated, there This seems to have happened in various African may be a trade-off between the real exchange rate countries including those which have not devalued, decline and productivity increase. as far as their real exchange rate has declined. The effects of the exchange rate policy on invest- ment in material and human capital are complex. Concluding Remarks However, inflation and instability of relafive prices are often associated with successive devaluations This study has aimed to compare the social conse- and floating exchange rates. These may have a nega- quences of different exchange rate policies based on tive impact on investment. Monetary illusion may the adjustment experience of African countries. help to correct distorted relative prices; however, it The African exchange rate policies have been clas- neither increases productivity nor favors the poor. sified in three categories. Some countries have dra- In the short run productivity increase involves matically devalued or adopted floating rates after greater labor intensity and flexibility. In the long run experiencing a very strong overvaluation of the offi- it is sociallybeneficial. In the African countries where cial rate associated with a large parallel market for the real exchange rate has reached a sustainable level, foreign exchange. Their policy was targeted to unify productivity increase constitutes the priority for the foreign exchange market, since depreciation of structural adjustment. the currency had already been achieved for a large part of the economy. In those countries, devaluation Notes was unavoidable. Some other countries have experienced moderate 1. A presentation on the different exchange policies of or no real appreciation of their currency. They chose sixteen Sub-Saharan African countries that have devalued to devalue their currency in order to facilitate their willbe foundinMinistere delaCooperationetduDevelop- balance of payments adjustment. Other countries, pement - Caisse Centrale de Cooperation Economique because they belong to monetary unions, adjusted (1988, vol. II and III), and in Jacquemont and Assidon without devaluation. Countries in these two catego- (1988). ries have achieved a decline in their real exchange 2. Other countries, however, have conducted their eco- rate, which has been on average a little more pro- nomic policy without devaluing and without a lasting nounced in the devaluing countries. But in these recourse to the international lending institutions for adjust- countries inflation was much higher than in the non- ment aid (Eithiopia, Rwanda, and some other Franc Area devaluing adjusting countries. countries). Real exchange rate decline, which is the main way 3. The adjustment experiences of Congo and Gabon to achieve structural adjustment, has important so- (see table 2-2) are too recent to be illustative. cial consequences, partly depending on the nominal 4. The negative effect of low wages on prductivity has exchange rate policy. The transitional income effect been especially noted in the public sector (Lindauer et al of the control of demand needed to decrease the real 1988 and Klitgaard 1989). 23 Exchange Rate Polices and the Social Consequences of Adjustmnent in Africa 5. Despite the dificulty of accurately distinguishing oux, A. M. Geourjon, R. Marodon, T. Montalieu, P. Plane, public spending of a social nature from other categories of December, 520 p. government expenditure, the convention is to regard edu- Guillaumont, P. 1986. Facteurs determinant le prix reel cation and health costs and welfare payments as the most paye au producteur, CERDI, June. characteristic items of social expenditure. A good many Guillaumont, P. and S. Guillaumont. 1989. La politique d'a- infrastructure facilities (rural roads, urban developments, justement au Niger," to be published. etc.) obviously have a direct influence on the standard of Guillaumont Jeanneney, S. 1988. "Devaluer en Afrique?," living of poor groups. Observations et diagnostics economiques, Revue de 6. In cases where exchange control is stringent and a l'OFCE, No. 25, October. parrallel exchange market exists, a considerable propor- Heller, P. S. et al. 1988. The Implications of Fund-Supported tion of trade eludes official channels and taxation. The Adjustment Programs for Poverty: Experiences in Selected major effect of a new exchange policy (devaluation and Countries, IMF Occasional Paper No. 58, Washington, lighter controls) which helps to bring some part of import D.C. and export traffic back into official channels is to increase IMF. 1988. Annual Report. government revenue. IMF. Various years. International Financial Statistics. IMF. Various years. Direction of Trade Statistics. References Jaquemot, P. and E. Assidon, in collaboration with A. H. Akanni. 1988. Politiques de change et adjustement en Afri- Azarn, J. P., G. Chambas, P. Guillaumont, and S. Guillau- que. L'experience des 16 pays d'Afrique subsaharienne et de mont. 1989. Impact of Macroeconomic Policies on the Rural L'Ocean Indien, Ministry of Cooperation and Develop- Poor, UNDP, New York. ment, Etudes et documents, distr. La Documentation Balassa, B. 1964. "The Purchasing Power Parity Doctrine: Francaise. A Reappraisal," Journal of Political Economy, Dec. p. 584- Klitgaard, R. 1989. "Incentive Myopia,' World Development, 596. Vol. 17, No. 14, April, p. 447-460. BCEAO. Various years. Notes d'information et statistiques. Kravis, I. B. and alii. 1975. A System of International Compari- Bonjean, C. and R. Marodon. 1988. Contraintes et efficacitede sons of Gross Product and Purchasing Power, The Johns la politique des prix agricoles. Examples de la C6te d'Ivoire, Hopkins University Press, 1975. du Kenya, deMadagascar, du Niger, duRwanda et du Senegal, Levy, V. and J. L. Newman. 1989. 'Wage Rigidity: Micro CERDI, University of Clermont I. and Macro Evidence on Labor Market Adjustment in the CERDI. 1988a. La politique d'ajustement au Niger (1982-87), Modern Sector," The World Bank Economic Review, Vol. 3, report produced at the request of the Government of the No. 1, January, p. 97-128. Republic of Niger, under the direction of Sylviane Guil- Lindauer, D. L., 0. A. Meesook, P. Snebsaeng. 1988. "Gov- laumont (with J. P. Azam, T. Montalieu, J. Mathonnat, G. emient Wage Policy in Africa: Some Findings and Chambas, A. M. Geourjon, P. Plane, C. Bonjean, C. De- Policy Implications " The World Bank Research Observer, jou), November, 650 p. Vol. 3, No. 1, January, p. 1-26. CERDI. 1988b. Les consequences sociales des politiques d'ad- Ministry of Cooperation and Development, Caisse Cen- justement. Elemnents d'analyse, portant plus particulierement trale de Cooperation Economique. 1988. Politique de sur le cas de la COte d'Ivoire et du Senegal. Study made for change et adjustement en Afrique. Une etude sur 16 pays the Ministry of Research and Technology, under the d'Afrique sub-saharienne hors zone monetaire (Coordinator: direction of P. Guillaumont, with F. Andre, M. Bergoug- P. Jacquemot) (3 vol.). noux, C. Bonjean, G. Chambas, M. Demeocq, C. Foumni- World Bank. 1989. World Bank Tables 1988-89. 24 3 Membership in the CFA Zone: Odyssean Journey or Trojan Horse? Shantayanan Devarajan and Jaime de Melo For most of the thirteen African members of the CFA formance of its members in the 1980s. We ask Franc Zone, the 1980s have been a decade of slow or whether on average CFA countries fared worse than negative growth in per capita GDP, worsening bal- a group of comparator countries. Since there is no ance of payments, debt crises, financial crises, single clear-cut group of comparators, we look at declining competitiveness and-most distressing of three: other Sub-Saharan African countries, other all-an apparent lack of adjustment to the changed low- and middle-income countries and other pri- external environment they inherited from the 1970s. maryand fuel exporters. Recognizingthata compari- Of the few recently-documented "success stories' of son of averages neglects differences within a group adjustment in Africa, none is a member of the CFA of countries, we perform some statistical estimations. Zone "Norld Bank 1989). Assuming that year-to-year GDP growth rates for all This disappointing performance is curious in light countries are drawn from a random distribution, we of the cautious optimism about CFA Zone member- investigate whether there is evidence that the distri- ship voiced earlier in the decade by, among others, bution of CFA countries' growth rates is significantly Guillaumont and Guillaumont (1984), and Devara- different from that of the comparators. We take a jan and de Melo (1987a). Their optimism stemm-ed closer look at the adjustment experience of CFA from the notion that participation in the CFA Zone countries vis-a-vis their comparators. Did CFA coun- would foster growth and reduce the need for adjust- tries adjust less, controlling for the size of the external ment. Guaranteed convertibility of the CFA franc and shock, than other countries? Did they adjust differ- the fixed exchange rate with the French franc would ently? One argument is that, given that they cannot lead to a stable investment climate for domestic and devalue their nominal exchange rate, CFA countries foreign investors, thereby stimulating economic cannot levy an inflafion tax to finance a fiscal deficit. growth. As for adjustment, the rules of the CFA Zone We ask whether this led to lower inflation and higher led to monetary and fiscal discipline. By avoiding current account deficits. Another argument is that some of the excesses of their African neighbors, CFA the fixed exchange rate makes expenditure-switch- Zone members' need to adjust would be less-even ing more difficult, so that CFA countries rely more on though they lacked an important instrument of ad- expenditure-reduction as a means of adjustment. We justment, namely, currency devaluation. Further- test this hypothesis. more, as Devarajan and de Melo (1987b) pointed out, CFA countries had enough instruments with which Growth and Adjustment in the CFA Zone: An to depreciate the real exchange rate which was, after Overview all, the relevant signal for structural adjustment. The purpose of this paper is to reassess the benefits In this section we compare the average perform- and costs of the CFA Zone in light of the poor per- ance of CFA members with that of three groups of 25 Membership in the CFA Zone: Odyssean journey or Trojan Horse? Table 3-1. A comparison between the 1970s and comparator countries. Rather than undertake a de- 1980s tailed analysis, we look for broad patterns that will suggest the statisfical evaluations of subsequent sec- Count rygroup 1973-81 1982-89 tions. This approach is based on simple, unweighted Average Annual Real GDP Growth Rate averages of countries' performance indicators over (pencentage) different periods. The approach does not recognize CQA (11) 3.7 2.6 that during the last two decades countries have been Other: subjected to extemal and internal shocks that have SSA (20) 2.7 2.0 varied in timing and magnitude across countries. Low-income (41) 4.4 2.9 Our method of aggregation by country groupings Primary (52 4.6 3.9 implies that the shocks were uniform within each group. Later we will allow for some diversity in Real Total Investment/Real GDP comparisons based on an error components model. (percentage) We compare CFA members with three groups of CFA 24.3 18.9 comparators. The most important comparator group Other: is other Sub-Saharan African countries. These coun- SSA 20.3 17.8 tries have most in common with CFA Zone members Low-income 21.6 19.8 in terms of economic structure, history and culture. Primary 22.5 19.4 Furthermore, being their neighbors, they provide CFA Zone members with the best perspective on life Debt/GDP (debt service/exports in parenthesis) outside the CFA Zone. We also compare CFA coun- (percentage) tries with other low- and middle-income countries CFA 30.6 (7.7) 62.5 (19.2) and other primary and fuel exporters. Except for Other. Gabon, all CFA countries had a per capita income in SSA 28.6 (9.7) 70.5 (20.9) 1980 below $1200. We use this figure as the cut-off for Low-income 26.0 (13.0) 58.4 (22.3) low- and middle-income countries. Higher income Primary 24.9 (15.1) 56.4 (25.2) countries, such as those in Latin America and East Asia, tend to have very different economic structures Average annual inflation rate and human capital endowments; hence their exclu- (percentage) sion from the comparator set. In the same vein, there CFA 12.0 4.3 is some evidence that countries producing and ex- Other: porting primary and fuel products have different SSA 24.3 29.7 adjustment histories from those which emphasize Low-income 18.4 33.3 manufactured goods (Faini and de Melo forthcom- Primary 24.4 44.9 ing). Since every CFA country is either a primary or fuel exporter, we also compare them with other pri- Real exchange rate mary and fuel exporters. (index, 1980=100) Table 3-1 displays the averages for seven indica- CFA 107.0 108.0 tors for CFA countries and their comparators during Other: 1973 to 1989. We look at two periods, corresponding SSA 115.0 147.0 (roughly) to the pre- and post-adjustment periods for Low-income 103.0 121.0 most countries. The results broadly confirm our ear- Primary 103.0 119.0 lier results (Devarajan and de Melo 1987a) for the sample period from 1960 to 1982. There we showed Average annual export growth rate that CFA countries' average GDP growth rate was (peentage) slightly higher than that of other Sub-Saharan Afri- CFA 6.8 1.5 can countries, but lower than that of other develop- Other: ing countries. Furthermore, we found that the rela- SSA 1.9 2.6 tive performance of CFA countries improved after Low-income 4.9 5.0 1973. We attributed the differences to the benefits Primary 4.8 7.6 from the stability of a fixed exchange rate regime, Note: Unweighted averages. Number of countries inespeally in the turbulent post-1973 era of floating sis. exchange rates. Source: World Bank data 26 Membership in the CFA Zone: Odyssean Journey or Trojan Horse? The pattern we discerned using data up to 1982 while thefixed exchange rate mayhave exerted some appears to have persisted through the late 1980s. monetary discipline in the CFA Zone, it did not de- Table 3-1 indicates that CFA countries enjoyed an crease CFA countries' reliance on extemal finance. average annual real GDP growth rate of over half a The debt service-to-exports and real exchange rate percentage point higher than their African neigh- indicators should be examined together with the bors. Their performance relative to other low- and debt-to-GDP ratio. A real exchange rate depreciation middle-income and primary and fuel exporting should cause a decrease in the debt service ratio if countries continued to be inferior. While the basic there is an export supply response -- although it will trends established up to 1982 appear to be sustained, also cause an increase in the debt-to-GDP ratio (see, the gap between CFA countries and other African for example, Rodrik 1989). While CFA countries' av- countries seems to be narrowing. The difference in erage debt-to-GDP ratio over the 1982 to 1989 period GDP growth rates was a full percentage point in the was lower than that of their African counterparts, the 1970s. average debt service ratios of the two groups were In the 1970s CFA Zone members had a higher almost the same. In other words, the debt servicing investment ratio than any of the comparators (and needs and, implicitly, the creditworthiness of these four percentage points higher than other African two groups of countriesare roughlycomparable. Yet, countries). The gap in investment-to-GDP ratios be- while the real exchange rate of the other African tween CFA countries and their African counterparts countries depreciated almost 30 percent from the narrowed during the 1980s. The decline in invest- 1970s to the 1980s, that of the CFA Zone stayed the ment and the narrowing of the gap between CFA same. This may provide a clue to the debt service countries and others is a two-edged sword. Lower puzzle mentioned earlier. By depredating their real investment ratios could be associated with an in- exchange rates, the other African countries were able crease in the efficiency of investment, as various to increase exports so that their average debt service- white elephants are abandoned. However, if there is to-exports ratio was comparable to that of CFA coun- no improvement in the marginal efficiency of invest- tries, however their average debt-to-GDP ratio was ment, the shortfall would signal a further slowdown much higher than that of the CFA countries. By not in GDP growth in the future. As we will see, the latter depreciating their real exchange rate, the CFA coun- possibility appears more likely when we look at the tries have probably not generated a comparable ex- period from 1982 to 1989 more closely. port supply response. Their debt service ratio is the The debt-to-GDP and inflation indicators high- same as that of other African countries, although light particular aspects of the CFA Zone rather their debt-to-GDP ratio is lower. sharply. Having a fixed and rigid nominal exchange Some of the speculation in the above paragraph is rate with the French Franc-changes in parity re- vindicated by a comparison of export growth in CFA quire the unanimous consent of CFA Zone mem- countries vis-a-vis their comparators. CFA Zone bers-effectively limits the seignorage tax that CFA members experienced faster growth in their exports countries can levy. Consequently, the CFA countries than all of their comparators during 1973 to 1981; have experienced dramatically lower inflation rates during 1982 to 1989, theiraverage export growth rate than their African counterparts and other low-and was slower than those of the comparators, perhaps middle-income and primary and fuel exporting becauseof thelackof real exchangerate depreciation. countries. While average inflation rates increased in This comparison between the 1970s and 1980s other parts of the developing world, as a conse- masks the evolution of adjustment throughout the quence of the deterioration in the external environ- 1980s. When we look at the two subperiods, 1982 to ment which began around 1982, they decreased in 1985 and 1986 to 1989, the deterioration in the CFA the CFA Zone where the average rate in the 1980s was Zone's position becomes clearer. The results appear under 5 percent. in table 3-2. Several striking patterns emerge. First, The CFA countries have had, until recently, an while the CFA countries' average GDP growth was unlimited line of credit from the French treasury, higher than that of other African countries in the first known as the compte d'operations. This enabled them subperiod, it was actually lower in the second. When to have an average debt-to-GDP ratio in the 1970s the growth rate in the rest of Africa accelerated after that was higher than that of all of the comparator 1985, that in the CFA Zone declined. Second, the groups. As it did for other low- and middle-income investment rate in the CFA Zone fell sharply during countries, this ratio doubled for the CFA Zone in the the 1980s, to the point where it was (marginally) 1980s. However, the debt-to-GDP ratio increased less lower than that of all of the comparators (in the 1973 rapidlythan in the Sub-Saharan Africangroup. Thus, to 1981 period it was the highest). Third, while the 27 Membership in the CFA Zone: Odyssean Joumey or Trojan Horse? real exchange rate in other African countries depre- Table 3-2. A closer look at the 1980s ciated sharply in the second subperiod, it appreci- - ated in the CFA Zone, partially as a reflection of the Counirygroup 298245 1986-9 depreciation of the dollar vis-a-vis the French franc. Average Annual Real GDP Growth Rate One interpretation of this change in relative posi- (pe-tage) tions is that, by undertaking adjustment programs CFA (11) 3.5 1.8 that emphasized a real exchange rate depreciation, Other: the other African countries were able to benefit from SSA (20) 1.0 3.0 the improvement in world commodity prices and Low-income (41) 2.4 3.4 demand after 1985. In particular, they were able to Primary (52) 4.8 2.9 enjoy export growth which translated into faster GDP growth. Another interpretation would empha- Real Total Investment/Real GDP size that the competitiveness of CFA countries was (pentage) undermined by the continued depreciation of their CFA 21.3 16.6 neighbors' currencies. There is also some evidence Other: that after 1982 the line of credit from the compte SSA 18.4 17.1 d'operations was no longer completely open. CFA Low-income 20.7 18.8 Zone members had to cut back their current account Primary 20.6 18.2 deficits. Given the lack of real exchange depreciation, this must have come through expenditure-reduction. Debt/GDP (debt service/exports in parenthesis) In particular, they reduced investment sharply. This (p-ennhge) led to a reduction in GDP growth rates which may CQA 58.0 (16.1) 67.1(215) continue into the future. Other: SSA 57.1(17.1) 83.5 (24.9) Are CFA Zone Growth Rates Different from Those Low-income 49.3 (19.5) 67.6 (25.0) of the Comparators? Primary 47.1(22.4) 65.9 (27.9) As mentioned earlier, a comparison of averages Average annual inflation rate assumes implicitly that all countries within a group (p 8entage) are uniform. We now relax that assumption. Spec- C0A 8.6 1.0 ifically, we assume that the GDP growth rate for each Other: country in each year is drawn from a random dis- SSA 26.2 35.7 tribution. We then ask whether the distribution from Low-income 19.5 50.4 which CFA countries' growth rates are drawn is sig- Primary 28.9 64.6 nificantly different from that of the comparators. This is a strong assumption since it assumes away the role Real exchange rate of a host of other factors which influence growth. (index, 1980=100) A common method for answering the question CFA 115.0 100.0 asked by the title to this section is by pooling the Other: cross-section and time-series data and using least SSA 124.0 177.0 squares regression with dummy variables. To cont- Low-income 109.0 136.0 rol for country-specific differences, this method re- Primary 106.0 136.0 quires a dummy variable for each country, which severely restricts the number of degrees of freedom. Average annual export growth rate Instead, we use a modified approach, known as the (Pntage) "error-components framework" which assumes that CFA 3.0 0.1 the intercept term in the regression of the logarithm Other: of GDP on the time trend (and a dummy variable) is SSA 0.1 5.0 also a random variable. This random variable is as- Low-income 1.2 8.8 sumed to pick up the influence of omitted variables Primary 7.6 7.7 in determining growth. The error-components m deth requinres the s of a gra d leas Note: Unweighted averages. Number of countries in paenthe. method requires the use of a generalized least sis. squares estimator to get efficient estimates (see Ful- Source: World Bank data lerand Battese 1974), but results in greater degrees of 28 Membership in the CFA Zone: Odyssean Journey or Trojan Horse? freedom. As this was the model estimated in our Table 3-3. Estimated growth ratesfrom error- previous work on the CFA Zone (Devarajan and de components model (standard errors in parenthesis) Melo 1987a), it has theadvantage of providingabasis Countrygroup for comparison. 1973-1 1982-89 Table 3-3 presents the estimated coeffidents for CFA 3.9 2.1 and ' in the following model: (0.33) (0.43) Other: Yit =a iDit +a'OD'lt +b DitT + b'DitT + uit SSA 2.5 2 (0.25) (0.33) where Yit is the logarithm of GDP of country i in year Low-income 4.2 3.0 t, Dit is the dummy variable for CFA members, D'it is (0o19) (0.23) the dummy variable for comparator group members, Tis the time trend, and Ujt is the composed error term. Primary 4.5 2.8 The estimates of b and b' represent the growth rates (0.17) (020) for CFA and comparator countries, respectively. Note: All the coefficients are significantly different from zero The regression results confirm the pattern sug- at the 95% confidence level. gested by the comparisons of averages above. While they enjoyed significantly higher growth rates than countries, CFA Zone members fellbehind the pack in their African counterparts in the 1970s, CFA coun- the 1980s. tries fell behind in the 1980s. The estimated growth rate for CFA countries is lower than that for other A Control Group Approach to Adjustment Sub-Saharan African countries for the 1982 to 1989 period as a whole. The comparison of averages above The statistical approach in the previous section. showed this to be true only for the latter half of this does not control for factors that are likely to affect period, i.e., for 1986 to 1989. Furthermore, the gap performance. In particular, developing countries between CFA countries and other low- and middle- were differentially affected by external shocks dur- income countries and other primary and fuel export- ing the 1970s and 1980s. Could the differences in ers appears to have widened. growth ratesbetween the CFA Zone and its compara- Table 3-3 shows that the growth rates of CFA coun- tors be attributable to these left-out factors? In ttis tries and their comparators were all significantly section we attempt a partial answer to that queston different from zero, and that the growth rate of CFA by relying on a statistical approach often used in Une countries fell more in the second period than did evaluation of adjustment programs, namely, the those of the others. It does not, however, indicate modified control-group approach (see, for example, whether the growth rate of CFA countries was sig- Faini et al forthcoming). This method amounts to nificantly different from those of its comparators. To looking for a fixed effect (in this case, belonging to determine this, we need to test whether the differ- the CFA Zone) in explaining performance after con- ences in the coefficients are significantly different trolling for changes in the external environment fac- from zero. Table 3-4 reports the results of such a test, ing each country. As before, our comparisons are for where 'significantly different from zero' is defined at the two periods, 1973 to 1981 and 1982 to 1989. We the 95 percent confidence level. now specify that the change in performance between The results corroborate the fact that the CFA Zone's the two periods-where performance is measured by perfornance has declined in relative termns during the average value of selected indicators during each the 1980s. The CFA members' growth rate was signi- period-is a function of autonomous policy changes ficantly higher than that of Sub-Saharan Africa dur- ing the 1970s. During the 1980s it was lower, but not Table 3-4. Growth comparisons significantly so. If we think of economic growth as a - running race, CFA countries in the 1970s were clearly 1973-81 1982-89 ahead of the pack of all African countries. By the SSA + NS- 1980s theywere indistinguishable from the restof the Low-income NS- pack. Similarly, while CFA members' growth rate Primary - NS- was not significantly different from that of other low- Note. + (-) indicates that growth in the CFA Zone was and middle-income countries during 1973 to 1981, it significantly higher (lower) than in the comparator group, NS+ became significantly lower during 1982 to 1989. That (NS-) indicates the growth rate in the CFA Zone was higher is, in the race with other low- and middle-income (lower) but not significantly so. 29 Membership in the CFA Zone: Odyssean Jounmey or Trojan Horse? after controlling for changes in the extemal environ- periods, had lower inflation and less current account ment between the two periods. The set of perform- improvement in the 1980s relative to the 1970s. The ance indicators j for country i is denoted by yy. We coefficients on the dummy variables for the GDP postulate that changes in the value of each perform- growth and investment equations are also lower, but ance indicator depend on the vector of autonomous not significantly so. The overall impression from this policy changes, x;, on changes in the extemal envi- set of regression coefficients is that, while average ronment, SHi; and on membership in the CFA Zone: GDP growth and investment-to-GDP ratios may have been lower in the 1980s in the CFA Zone, the yyj =a q + x' a i + SHia 2j +aqd +e ij more significant differences appear in the average inflation rates and current account improvement. As where a prime denotes a transpose and d is a dummy mentioned earlier, the rules of CFA Zone member- variable that takes the value 1 for countries that ship prevent CFA governments from levying an in- belong to the CFA Zone and 0 otherwise. flation tax, so that the fixed effect on inflation for CFA Since autonomous policy changes are difficult to countries would be lower. The current account coef- observe, we postulate that they are a function of the ficient is much more troubling. It implies that in all difference between the realized value of a perform- of the comparator groups, CFA countries are con- ance indicator and some target value. Hence, we use spicuous by their inability to reduce their current lagged values of the performance indicators as prox- account deficits, even taking into account the fact ies for the changes in policy (see Faini et al forthcom- that the shocks they face (i.e., the need to adjust) may ing, for details). The statistical results reported below have been different. show the coefficients for a model in which each This observation is consistent with our earlier re- observation is an average over the first or second sult about the lack of real exchange rate depreciation period. in the CFA Zone. While the restof Sub-Saharan Africa The shock variable is a weighted average of depreciated its real exchange rateby an averageof 28 changes in the world real interest rate (R) and the percent during the 1980s, the CFA Zone's real ex- export price index (PX) and import price index (PM) change rate appreciated. Such a result would not for each country. The weights are the ratios to GDP have been a problem if the CFA countries did not (Y) of debt (D), exports (X) and imports (M), respec- need to depreciate their real exchange rate. However, tively. By "changes' we mean the difference in the the coefficients in table 3-5 show that they needed to averagevalueof thesevariablesbetween 1973 to 1981 do so. Contmlling for differences in the shocks they and 1982 to 1989. In symbols, faced, the amount of current account improvement in the CFA Zone was systematically lower than that SH = (R2 - RI) (D/Y)i - (PX2 - PXi)(X/Y)i in the comparator groups. + (PM2 - PM1) (M/Y)I The coefficients of the lagged variables are usually significant with the expected sign. The higher the where the subscripts refer to averages over the first value of the own-lagged variable, the lesser the and second periods, and country subscripts have change in performance between the two periods. For been omitted. example, other things equal, the higher is average The estimate of the shock variable turns out to be growth in the period 1973 to 1981, the less is the lower for CFA countries than for any of the compara- increase in average growth during 1982 to 1989 com- tor groups. Indeed, the shock faced by other Sub-Sa- pared with 1973 to 1981. While the extemal shock haran African countries was over twice that facing variable has the expected sign for the inflation and CFA countries. Hence, some of the lack of adjustment current account equations, it enters positively with in the CFA Zone may be explained by a reduced need the change in growth and investment, which is un- for adjustment. The question is whether, given the expected. Although, in most instances, these coeffi- reduced need for adjustment, the observed adjust- cients are not statistically significant, the low ex- ment was still too little. planatory power of the proxy measure for extemal The results are presented in table 3-5. The CFA shocks suggests some measurement inaccuracies. dummy variable has a negative coefficient for all Anotherpotential problem with theresults in table comparisons. The two coefficients that are consis- 3-5 is that we may have misspecified the model. In tently significant for the CFA Zone are those in the particular, by controlling for autonomous policy inflation and current account equations. The inter- changes separately from CFA Zone membership, we pretation is that CFA countries, after controlling for- maybe neglecting those aspects of Zone membership differences in the shocks they faced between the two that are associated with policy. An alternative formu- 30 Membership in the CFA Zone: Odyssean Jouney or Trojan Horse? Table 3-5. Pefortnance in the CFA Zone Dependent variable Y-1 I/Y.1 INF.i CA.i CTA SH Comparson with other Sub-Saharan countries Y -.889* .094 .021 .057 -.009 .049 (-4.73)* (1-47) (.80) (.54) (-.82) (1.01) I/Y 1.032* -.656* .010 .061 -.022 .235 (2.37) (-3.72) (.09) (.27) (-.89) (1.70) INF -2.151* -.012 .212 -1.397* -.138* .135 (-2.47) (-.05) (.67) (-4.24) (-3.01) (.63) CA .282* .052 .021 -.661* -.032* -.106* (1.73) (.85) (.95) (-5.84) (-3.24) (-2.18) Comparison with other low-income countries y -.433* .035 .039 -.093 -.021 .014 (-2.63) (.58) (1.27) (-.95) (-1.58) (.37) I/Y .457* -.36* -.004 .185 -.01 -.078 (2.07) (-3.74) (-.06) (1.54) (-.77) (-1.28) INF -1.70* - 276* .163 -.827* -.134* .081 (-4.4) (1.75) (.65) (-4.64) (-3.85) (1.17) CA .111 -.025 .015 -.765* -.024* -0.56 (.76) (-.48) (.62) (-6.87) (-2.48) (-1.40) Compaison with primary exporting countries y -.63* .069 .025 .01 -.015 .02 (-5.1) (1.60) (1.58) (.16) (-1.30) (.66) I/Y .55* -.391* -.027* -.091 -.005 .116 (2.77) (-5.41) (-1.85) (-.63) (-.36) (2.00) INF -.75* -.003 -.123 -.537* -1.06* .046 (-1.72) (-.002) (-.59) (-2.74) (-2.58) (.59) CA -.033 .032 -.001 -.781* -.026* -.068 (-.23) (.58) (-.06) (-7.01) (-2.40) (1.51) Notes, The constant term is omitted from the results. Asterisks denote those coefficients which are significant at least at the 90 percent level. Y=GDP growth; I/Y=investment/GDP; CA=current account/GDP; INF=inflation rate. The subscript (-1) denotes lagged values. Results are corrected for heteroskedasticity by weighing each observation by the inverse of its estimated standard error. Extreme influential observations are excluded. lation would be to leave out the independent vari- CFA Zone members have enough instruments with ables for autonomous policy changes or their proxy, which to adjust their economies (Devarajan and de the lagged dependent variables. Unfortunately, this Melo 1987b). In practice, they have been reluctant to specification yielded significant coefficients only for use these other instruments or, when they have cho- the CFA variable in the inflation equation. The effect sen to use them, the results have been disppointing. of CFA membership on current account improve- The advantage of a nominal devaluafion is that it ment was not statistically significant. Thus, isolating permits expenditure-switching to accompany the autonomous policy changes sharpens our estimates necessary expenditure-reduction of an adjustment of the fixed effects due to CFA Zone membership, program. In this case, the amount of expenditure possibly because there was no systematic relation- reduction required would be less (Corden 1988). It ship in the use of these policies among CFA Zone follows, therefore, that CFA members would have members. had to rely more on expenditure-reduction as op- Despite these qualifications, there is some evi- posed to expenditure-switching in reducing their dence that adjustment in the CFA Zone has been currentaccountdeficits.Wenowtest thishypothesis. insufficient in the 1980s, arguably the era when al- Using the data from the two periods, 1973 to 1981 most everybody else in Africa was undertaking ma- and 1982 to 1989, we estimate an equation which jor adjustment programs. We have suggested one links the target of adjustment-the resource balance reason for the CFA Zone's lack of adjustment, expressed as a rafio to GDP-with the investment namely, the inability of CFA Zone members to effect rate and the real exchange rate. The investment rate a nominal devaluation. But the nominal exchange is used here as an instrument for expenditure chang- rate isbutoneinstrumentof adjustment. In principle, ing policies. Of course, there are other such instru- 31 Membership in the CFA Zone: Odyssean Jommey or Trojan Iforse? Table 3-6. Changes in resource balance, investment and the real exchange rate betweem 1978/79 and 1987/88 Average predicted change in resource balancdGDP Investment component Real exchange rate component CFA .046- .071 -.033 C)ther lo w-income .050- .053 .232 Note: Column I is equal to the sum of colums 2 and 3 plus the intercept tern. rnents, includinggovernment consumption expendi- come countries. Second, the real exchange rate com- ture. However, it is easier to cut investment first, ponent has the wrong sign for CFA countries. Instead especially in a slow-growing economy. Therefore, it of contributing to the reduction in the current ac- is worth examining the extent to which improve- count deficit, the real exchange rate may have ments in the resource balance reflected declines in worked against it in the CFA Zone, though the lack i'esntment. of statistical significance of the coefficient on the real The regression linking the resource balance with exchange rate variable calls for caution in interpret- investment and the real exchange rate also contains ing this result. Finally, and most importantly, the a dummy variable for each country, to capture coun- relative contributions of investment reduction and tryspecific effects, and a time trend. The results of real exchange rate depreciation are very different the regTession yield a negafive-and highly signif- between the two groups of countries. For other low- cant-coefficient on the investment variable for both income countries, it is about one-quarter whereas for CFA and non-CFA countries. By contrast, the coeffi- CFA members it is over two (and with the wrong dent on the real exchange rate variable is small and sign). indistinguishable from zero. This is consistent with other studies that have attempted to link the real Concdusions exchange rate to the resource balance or the trade balance. For example, Pritchett (1990) found only a The purpose of this paper has been to assess weak relationship between the merchandise trade whether the particular institutional arrangement of balance and the real exchange rate, even when con- the CFA Zone has aided or hurt its members. It has trollingforterms of trade movements. His reasoning. been argued that the convertible currency with a which may also apply here, was that while exports fixed exchange rate results in monetary and fiscal ma3y respond to changes in the real exchange rate, discipline which, in turn, benefits CFA Zone mem- imports were determined by foreign exchange avail- bers. Just as Ulysses tied himself to the mast, CFA ability, i.e., exports, and hence may move perversely. governments abdicated the right to levy an inflation Undaunted by the insignificance of the real ex- tax so that they would neverbe tempted to do so. The change rate variable, we use the estimated regression evidence of the relative performance of CFA countri- coefficients to compute the predicted resource bal- es'economies vis-a-vis theircomparators shows that ance in 1978-79 and 1987-88. For both CFA and non- this argument was persuasive until the early 1980s. CFA countries, the resource balance improved be- After 1981, changes in the world environment and tween the first and second two-year periods. This is persistent current account deficits meant that CFA a reflection of the cutback in foreign lending in the countries needed to adjust their economies along 1980s and the simultaneous need for these countries with most other developing countries. Our statistical to make increasingly higher debt service payments. results show that they did not adjust as much as they We can then ask how much of the improvement in needed. Furthermore, their growth performance was the resource balance between 1978-79 and 1987-88 disappointing. Under every estimate, CFA Zone was due to investment reduction, and how much to members' GDP growth rates fellbehind those of their real exchange rate depreciation. Were the relative counterparts, including the other African states. Fi- proportions different between CFA and non-CFA nally, the burden of adjustment appears to have countries? The results are reported in table 3-6. fallen disproportionately on expenditure reduction First, although the improvement in the resource in general, and investment reduction in particular- balance-to-GDP ratio was roughly comparable for an ominous sign for future growth. the two groups of countries, the investment compo- Of course a change in external circumstances does nents were quite different. CFA countries relied more not necessarily mean that the original commitment heavily on cutting investment than did other low-in- to a fixed exchange rate was unwise. It is possible that 32 Membership in tihe CFA Zone: Odyssean Journey or Trojan Horse? CFA Zone members took all these contingencies into of the CFA Zones,w World Development, Vol. 15, No. 4, pp. account in making the original decision to join the 483-96. CFA Zone and hence forego the opportunity to de- Devarajan, S. and J. de Melo, (1987b) "Adjustment with a value their nominal exchange rate in the future. In Fixed Exchange Rate: Cameroon, Cote d'Ivoire and Sene- this case, there are no policy implications from the gal,' World Bank Economic Review, Vol. 2, No. 2. recent deterioration in the CFA Zone's economic per- Faini, R., J. de Melo, A. Senhadji-Semlali and J. Stanton, formance. The members may have drawn a bad (forthcoming) 'Growth-Oriented Adjustment Programs: hand, but not one which renders their original deci- A Statistical Analysis', World Development sion sub-optimal. Faini, R. and J. de Melo, (forthcoming) 'Adjustment, In- An altemative interpretation is that the current vestment and the Real Exchange Rate in Developing circumstances facing the CFA Zone lie outside the set Countries," Economic Policy. of events which were considered when the original Fuller, W. and G. Battese, (1974) 'Estimation of Linear decision to join the CFA Zone was made. In particu- Models with Crossed Error Structure," Journal of Econo- lar, the adverse terms of trade shocks of the 1970s and metrics, Vol. 2, pp. 67-78. 1980s, and the attendant need to shift resources from Guillaumont, P. and S. (1984) Zone Franc et Developpement nontradables to tradables, may not have been ex- Africain, Paris: Economica. pected in the 1960s. Such an argument is compelling Guillaumont, P., S. Guillaumont and P. Plane, (1988) "Par- because the exchange rate is both an instrument for ticipating in African Monetary Unions: An Alternative transforming resources from nontradables to trad- Evaluation,' World Development, Vol. 16, No. 5, pp. 569- ables as well as an inflation creating (or controlling) 76. tool. It could be that the founders of the CFA Zone Guillaumont, P. and S. Guillaumont, (1988b) eds., Strategis calculated the inflation controlling benefits of a fixed de Developpement Comparees: Zone Franc et Hors Zone exchange rate without anticipating the costs in terns Franc, Paris: Economica. of the countries' inability to adjust to unfavorable Honohan, P. (1990) 'Monetary Cooperation in the CFA external circumstances. Thus, the very institutional Zone," Policy, Research and External Affairs Working arrangement which enabled these countries to enjoy Paper No. WPS 389, World Bank. faster and more stable growth in the 1970s is prevent- O'Connell, S. (1989) 'Uniform Trade Taxes, Devaluation ing them from adjusting to the external and internal and the Real Exchange Rate: A Theoretical Analysis,w shocks of the 1980s. In short, what began as an Odys- Policy Planning and Research Working Paper No. WPS sean journey may have turned into a Trojan horse. 185, World Bank. Pritchett, L., (1990) 'The Merchandise Trade Balance and References the Real Exchange Rate in LDC's,' Country Economics Department, World Bank Corden, W. M., (1988) 'Macroeconomic Adjustment in De- Rodrik, D., (1989) 'The Welfare Economics of Debt Service,' veloping Countries,' Research Department, Interna- John F. Kennedy School of Government, Harvard Uni- tional Monetary Fund. versity. Devarajan, S. and J. de Melo, (1987a) 'Evaluating Participa- World Bank, (1989) Sub-Saharan Africa From Crisis to Sus- tion in African Monetary Unions: A Statistical Analysis tainable Growth, Washington, D.C. 33 4 Comments on Exchange Rate Policy Ravi Kanbur and Stephen O'Connell Ravi Kanbur surveys so much but from anthropological investiga- tions in the African context, is the gender division of First I would like to bring up two criticisms of recent labor within the household. In particular, females treatments of the exchange rate and real income dis- seem to have more control over the food crop plot tribution and poverty. One of the criticisms has been and males seem to have more control over the cash that the notion of thinking of households as having crop plot. In other words, there is a division within been involved in only tradable activities or only non- the household between the control and use of income tradable activities is clearly empirically not true. from tradable goods versus nontradable goods. If in Through diversification each household can get its this case food is a nontradable good, like root crops income from both tradable and nontradable activi- for example, then we are back to the old story in ties. Taking this to the extreme, suppose there is a which some individuals derive their incomes pri- nationally representative household, which is in fact marily from tradable goods and other individuals exactly representative of the national composition of derive their incomes primarily from nontradable tradable and nontradableactivities. Then, fora given goods. As the share of tradable goods income is total expenditure, altering the composition of na- increased relative to nontradable goods income, in tional income should not affect the standard of living fact intrahousehold inequality will be increased. I of this typical representative household. So to the think this is a feature that has not been taken into extent that households are diversified, to the extent account in the literature so far. This has not been that the household's portfolio of sources of earnings directly analyzed because the data is not available; is exactly the same as the nation's portfolio of sources however, there is plenty of indirect evidence, firstly of earnings, expenditure-switching should have no that these types of social structures exist, and sec- effect at all on poverty. So I think this is one thing that ondly that the actual division of income between would go against the simple way of looking at it, different members of a household affects the con- which is that all households get their incomes either sumption patterns of that household. Paul Collier from tradables or from nontradables. and a number of other people have looked at indirect Another criticism which actually goes in the other evidence which supports this. direction is related to the very important area of In terms of the social consequences of adjustment intrahousehold inequality. The analyses that have following through from the impact of real exchange been done so far are focused entirely on looking at rate changes, I essentially see three stages to the the distribution of income at the household level. So argument. In the first stage, we assumed that there we take the sources of income of the household as a was a complete division between tradables and non- whole and we take that as being somewhat equally tradables. On the basis of that assumption, most of distributed within the household. But in fact, one of us came to the conclusion that real exchange rate the points that is coming out, not from household changes would not be detrimental to poverty. In the 34 Comments on Ewxhange Rate Policy second stage, people said that households tend to les, and if you have alreadyabdicated its usebeca use have diversified sources of income. The third stage is you do not want to use it to change the relative price that, yes, it is true that households may be diversi- between present goods and future goods, then of co- fied, but within the household there are individuals urse you cannot use it to change the first relative price who focus on one rather than the other source of either. There isa certain difficulty in the way the pap- income and there is real conflict between these indi- er presents this, which is the notion of Ulysses tying viduals in the household so actually the impact of himself to the mast. Circumstances change so it is structural adjustment on inequality is understated. now worthwhile to untie Ulysses from the mast, but The Guillaumonts' analysis leads to these issues, as soon as you untie him, instead of rescuing the da- which are not covered in the paper. msel in distress, he is liable to go off toward the siren. The basic problem is that in abdicating the use of The Guillaumont and Guillaunont Jeanneney this instrument to be able to change one relative Paper price, you abdicate its use to change the other relative price. In conditions or circumstances where you need The first part of the Guillaumonts' paper discusses to change the other relative price, if you use it, you the relationship between the nominal exchange rate remove the precommitment not to use that instru- and the real exchange rate. The problem here is that ment to change the second relative price. two relative prices are affected by the nominal exch- Suppose now that nominal devaluation is indeed ange rate at the same time. One is the relative price tried, and that commitment has been lost. Then what between tradables and nontradables, the real ex- problem is being solved by a rational government or change rate, and the other is the relative price be- what is the problem that Devarajan and de Melo see tween present goods and future goods, the inflation rational govemments solving in this context? One rate. The nominal exchange rate affects simultane- line of argument seems to be that, in the early part of ously the real exchange rate and the inflation rate. All the period, there is a commitment to a fixed exchange of their discussion is based around this, as is a lot of rate and it seems to work. There are high growth the Homeric discussion (in the Devarajan and de rates, etc. Now the extemal circumstances change, so Melo paper). we should abandon that commitment and use this instrument to affect these relative price changes. But The Devarajan and de Melo Paper of course once you abandon it, you will have lost the earlier benefits. Devarajan and de Melo may even Essentially the point is that you have an instrume- argue that the fixed exchange rate policy commit- nt which you can either use or abdicate the use of co- ment should never have been made in the first place, mpletely. If you retain use of it, you maybe tempted but I think that is an argument that has to be made to use it to levy the inflation tax. The point made by quite separately. Devarajan and de Melo is that, by abdicating the use This discussion can be cast in terms of time consis- of the nominal exchange rate, you are abdicating the tency. The difficulty is that once the commitment to ability to levy an inflation tax. But, if circumstances the fixed exchange rate is abandoned, it is lost virtu- are such that you need to use that instrument to affect ally forever. What are the costs of that from the view the relative price between tradables and nontradab- of commitment? Stephen A. O'Connell Though differing in motivation and scope, these two The Guillaunont and Guillaumont Jeanneney stimulating papers share a concem for the relation- Paper ship between policy regime and macroeconomic performance in Sub-Saharan Africa. I will restrict my This paper begins with a useful overview of ex- comments to this area of common interest, leaving change rate arrangements in Sub-Saharan Africa. the second half of the Guillaumonts' paper, on the The authors then go on to make four main points social consequences of alternative adjustment related to the implications of the exchange rate sys- modes, to my fellow discussant. tem for structural adjustment. First, exchange rate 35 Comments on Exchange Rate Policy policy helps determine how much inflation will ac- tion of features, of the Zone accounts for the ability company real exchange rate adjustment. Second, in of these countries to maintain policies consistent many cases a devaluation will not reduce the am- with a fixed parity for so long.1 ount of expenditure reduction required for extemal The question of whether a nominal devaluation (i.e., current account) adjustment. Third, external allows external adjustment to takeplace at lowercost adjustment does not require real exchange rate depr- in terms of expenditure reduction is clearly essential eciation, since it can take place through productivity to any assessment of external adjustment strategies improvements without any change in the real exch- in a managed exchange rate system. The authors ange rate. Fourth, low inflation programs are more claim that nominal devaluations are likely to have favorable to investment, and therefore more likely to minimal transitional benefits, basing their claim on produce productivity-led external adjustment. the example of Madagascar and an appeal to lack of However broad the authors' intent, it is difficult money illusion on the part of wage earners. Both not to read this line of argument as an advertisement arguments deserve further scrutiny. In the Madagas- for maintenance of the current parity in the CFA car case, we need the counterfactual: what contrac- Zone. This is a position that the authors have vigor- tion would have been required if the exchange rate ously defended in other work and on other grounds. had not been used? Moreover, since the transitional While the presence of this sub-theme makes for lively benefits depend on the credibility of the new nominal reading (all the more so since the two sets of authors exchange rate, what lessons does the Madagascar appear to differ on both the facts and their interpre- case have in the very different institutional context of tation), it undercuts the paper somewhat as a survey the CFA Zone? With regard to the presence or ab- of the implications of exchange rate policy in Sub-Sa- sence of money illusion, thisisan empirical question. haran Africa. Thus, for example, the authors spend Elsewhere in the paper, the authors stress the effects very little time discussing the role of exchange rate of nominal effective movements in the French franc policy in stabilization, an issue of extreme relevance on the real exchange rate in CFA countries; it is outsideoftheZone;andtheylumptheZonetogether difficult to square this particular nonneutrality with other fixed exchange rate regimes in Sub-Saha- (which seems to suggest the presence of some wage- ran Africa, without discussing the mechanisms that price stickiness in the CFA countries) with the claim have allowed the Zone to accumulate such a remark- that a policy-induced nominal devaluation would able history of nominal exchange rate stability. have no transitional benefits. In discussing real exchange rate depreciation and The authors make an important contribution by inflation, the authors point to a correlation between emphasizing the potential role of productivity incr- devaluation and inflation across African countries. eases in generating external adjustment without exc- But this observation does not establish a causal role essive cost in terms of growth. One has only to contr- for nominal exchange rate policy. What role, if any, ast the expansionary adjustment pattern of some of does the exchange rate regime play in determining the Asian debtors with the contractionary adjustm- inflation and the path of the real exchange rate? What ent pattern of theirLatin American and African coun- we know from outside of the CFA Zone is that fixed terparts to appreciate the significance of this point. I exchange rates often fail to deliver stability of prices was less impressed, however, by the argument that except over very short time horizons. While this is adjustment programs that avoid nominal devaluat- most dramatic in the devaluation-cum-stabilization ion are more conducive to productivity increase than cycles of Latin America, the African record is also full others. While there isample theoretical and empirical of examples of fixed parities that were eventually work relating productivity growth to trade policy changed or abandoned altogether under the pressure and even fiscal policy (e.g., government expenditure of domestic inflation. These cases suggest that at least patterns), there is relatively little evid nce linking pr- over the medium run (e.g., the time horizon of the oductivity growth to the exchange ir.e regime, much typical adjustment program), inflation and the rate less the level of the nominal exchangerate. And given of official depreciation are the joint result of other the authors' statement that adjustment with a fixed policies, such as those determining fiscal deficits, parity is out of the question for high inflation coun- external borrowing, and the rate of domestic credit tries, it appears that we are again workingwithin the creation. From a CFA Zone perspective, it may ap- narrow confines of the CFA Zone debate, with few pear that these other policies are subordinated to the lessons for the rest of Sub-Saharan Africa. commitment to maintain the parity. But this is evi- Therelationshipbetweenexchangeratepolicyand dently not a generalizable feature of the Zone. One is productivity growth is well worth further investiga- therefore left wondering what feature, or combina- tion on both the analytical and empirical ends. As the 36 Comments on Exchange Rate Policy authors suggest in their discussion, it is probably Performance comparisons are subject to the critiq- uncertainty regarding real exchange rates and infla- ue that Zone membership is only one of many fact- tion, rather than overvaluation or inflation per se, ors affecting performance. The use of comparator gr- that discourages investment and productivity oups low-income countries, primary exporters, and growth. Recent empirical work provides some sup- Sub-Saharan Africa) addresses this and allows the port for this notion.2 authors to control to some degree for level of devel- Overall, (the first half oO this paper brings a fresh opment, economic structure, and external environ- perspective to the debate on exchange rate policy in ment. Nonetheless, it is possible that the results are Sub-Saharan Africa. Future contributions will un- driven by chance correlations between Zone memn- doubtedly develop some of the analytical arguments bership and other unmeasured factors like the terms further and strengthen their empirical content. of trade. Guillaumont, Guillaumont and Plane (1988) approached this problem by estimating a cross-sec- The Devarajan and de Melo Paper tional model of the determinants of growth, and then interpreting each country's residual as an indicator The CFA countries have long been noted for their of its performance relative to the norm. In the second choice of monetary stability and low inflation (e.g., half of their paper, Devarajan and de Melo take an Mundell 1972). Has this stability delivered benefits alternative route (the control group approach), re- in terms of increased growth? Both sets of authors gressing performance indicators on a composite eco- represented here have recently given a (qualified) nomic shocks variable,asetof macroeconomic policy yes. Devarajan and de Melo (1987) found that the indicators, and a CFA-specific dummy variable. CFA countries had grown more rapidly than the rest I found the control group approach somewhat less of Sub-Saharan Africa (though not quite as rapidlyas appealing than the rest of the paper, for two reasons. developing countries outside of Africa) over the pe- First, in order to econometrically identify the unsp- riod from 1960 to 1982. Similar methodology, includ- ecified CFA Zone effect, the approach assumes that ing more extensive attempts to control for structural the effects of Zone membership do not operate throu- differ- ences and external shocks, was applied by gh differing impacts of autonomous shocks or differ- Guillaumont, Guillaumont and Plane (1988), with ent channels of influence of autonomous policy. This similar conclusions. is a troubling starting point, since we know that pol- The present paper by Devarajan and de Melo icy regimes differ and we suspect that economic res- shakes up the debate by suggesting that Zone mem- ponses differ due to the different policy framework. bership may not be such a bargain after all. In the first Second, instead of measuring autonomous policy half of the paper, the authors extend their earlier changes directly, the authors adopt a target-adjust- work comparing macroeconomic performance in the ment model in which the lagged performance indi- Zone to perfonnance in three alternative comparator cator serves as a proxy for policy changes. However, groups. When data from 1983 to 1989 are included, since the target itself is unobservable, the fixed effect the relative growth performance of the Zone deterio- ends up picking up (among other things) differences rates noticeably, particularly after 1986. Moreover, in targetsbetweenCFAand non-CFA countries. Thus relative performance on the external front, as meas- the finding that the current account deficit fixed ef- ured by real exchange rate depreciation and export fect is higher in the CFA Zone than outside may not growth, deteriorates even more markedly. While be an indicafion of relatively poor performance in the some portion of the export recovery observed out- Zone; it may simply be evidence of a less binding side of the Zone may simply be a shift of existing external constraint. In particular, the arrangements exports onto official markets in response to declining of the Zone, including the special economic and po- direct and indirect taxation of official exports, it is litical ties to France, may give these countries better revealing that the average export growth rate in the access to international capital flows (including pro- Zone declines dramatically in the 1980s, while that of spective debt forgiveness) than otherwise, leading other developing country groups remains roughly them to choose an optimal mix of public finance that the same or improves. The authors point out that is weighted more heavily towards external finance. while inflation was extremely low in the Zone in the This would be consistent with a more appreciated 1980s, real exchange rates were essentially un- real exchange rate and a lower fixed effect in the changed over the period. Outside of the Zone, in inflation equation. contrast, real exchange rates depreciated substan- Given the difficulties associated with the control tially. In short, the Zone appears to have lost much of group approach, a more satisfactory approach might its magic in the 1980s. be to run a fixed-effects model controlling only for 37 Comments on Ezchangc Rate Policy external shocks. The authors do something similar Notes "they re-run the control group regressions with the policyvariables omitted"and find thatonlythe infla- 1. There are numerous possibilities, including (1) the tion effect remains as a clear distinction between CFA availability of short-term external finance through the op- and non-CFA countries. While the authors charac- erations account, which shelters these countries from terize this finding as unfortunate, it does support external liquidity crises; (2) the convertibility guiarantee, their basic point that the CFA sparkle has disap- which imposes a responsibility on France to monitor peared in the 1980s. Further work attempting to con- macroeconomic policy in the Zone; (3) the institutional rule trol for extemal factors, including perhaps an update requiring unanimity in deciding on any parity change; (4) of the paper by Guillaumont, Guillaumont and Plane the institutional rule limiting government borrowing from (1988), would seem highlyjustified. the central bank; and (5) the high degree of diversification It is clear that the arrangements of the CFA Zone of commodity exports, implying relatively low terms of came under unprecedented pressure in the 1980s, trade volatility for the Zone as a whole. and that many of these pressures are continuing into 2. For example, Caballero and Corbo (1989) show that thel990s.Theauthors trace therelative deteriorafion exports are negatively related to volatility of the real ex- of performance to an excessive reliance on expendi- change rate in a group of Latin American countries. ture reduction. This focuses attention on the question of the parity: can the CFA countries devalue without References endangering the very existence of the Zone? Given the severity of external shocks in the 1980s (including Caballero, R. J. and Corbo (1989), How Does Uncertainty the first Zone-wide decline in export prices and the About the Real Exchange Rate Affect Exports?, The substantial appreciation of the French franc), it seems World Bank, PPR Working Papers Series No. 221, June. at least conceivable that an adjustment tied to these Devaraian and de Melo (1987), Evaluating Participation in unusual circumstances might be viewed as consis- Aftican Monetary Unions: A Statistical Analysis of the tent with a continued commitment to a fixed parity. CFA Zones, World Development Vol. 15, No. 4: 483-496. More broadly, the issue is whether there are policy Guillaumont, Patrick, Sylviane Guillaumont, and Patrick rules for the nominal exchange rate that would pro- Plane (1988), Participating in African Monetary Unions: vide more flexibility but still be consistent with other An Altermative Evaluation, World Development Vol. 16, important features of the Zone, such as absence of No. 5:569-576. severe capital controls and continuation of the con- Mundell, Robert A. (1972), African Trade, Politics and vertibility guarantee from France. The paper by Money, in R. Tremblay, ed., Africa and Monetary Integra- Devarajan and de Melo provides essential back- tion (Montreal, Les Editions HRW). ground for any further analysis of these issues. 38 5 The Inflationary Consequences of Devaluation with Parallel Markets: the Case of Ghana Ajay Chhibber and Nemat Shafik A common concern in macroeconomic adjustment on, the large official devaluations of over 40 percent programs is the potential inflationary effects of a per annum implemented between 1983 and 1988. combination of devaluation, trade liberalization, This paper disputes the conclusion that the recent subsidy reduction and price decontrol. This issue has surge in inflation in Ghana is due to the large official become critically important in the African context devaluations. The conclusion is based on a model where inflation has been accelerating in a number of which makes it possible to quantify the effects of countries, particularly in those which are not mem- exchange rate changes in the presence of parallel bers of the CFA Franc zone. The high inflation in the markets. The results show that there is no direct non-CFA zone countries is alleged to result from the relationship between the official exchange rate and devaluations many of these countries have under- inflation since prices had already adjusted to the gone as part of the reform programs. The link parallel exchange rate. In fact, official devaluation in between devaluation and inflation has been ana- Ghana had a positive effect on the govemment lyzed in considerable detail, both theoretically and budget. Because of this improvement in the fiscal empirically. The empirical evidence suggests that devaluation is inflationary, but the degree of inflation Figre 5-1. Mark-Up Model of Inflation in Ghana depends on accompanying fiscal and monetary poli- cies. The analysis of the inflationary effects of 120 devaluation is complicated by the presence of paral- lel markets. This paper presents a theoretical framework to 95 address the question of exchange rate devaluation in the presence of parallel markets in the second section. a , -l It then uses Ghanaian data to empirically estimate a 7P0 model in the third section which is used to analyze policy trade-offs for Ghana in the fourth section. The last section draws together the lessons from the Gha- 4.5 p naian experience. Mark- Ghana has carried out one of the most thorough uP adjustment programs in Africa. Its inflation rate, wh- Urlit ich is depicted in figure 5-1, is high and has been ris- Wage ing in recent years, although it is still lower than dur- costs ing the worst crisis years. These high rates of inflat- A ion have coincided with, and have often beenblamed Source: Author's calculations. 39 The Inflationary Consequences of Devaluation with Parallel Markets: the Case of Ghana deficit, the official devaluation had favorable effects effect on unit labor costs. In the asset markets, there on the rate of monetization, on inflation and on the is a domestic asset, cedis, and a foreign asset, foreign rate of depreciation of the parallel exchange rate. exchange (F). Overall demand is a function of total financial assets which are defined as: The Model A = M + epF. The model simultaneously determines the rate of inflation and the parallel market exchange rate pre- Where M represents the money supply, ep is the mium and links them to the monetary, fiscal and real parallel exchange rate, and F is the foreign asset. The sides of the economy through several channels. A demand for foreign exchange is composed of trans- segmented goods market is hypothesized. There is actions demand and portfolio demand. The transac- an official market at which goods are available at a tions demand forforeignexchangeisanegative func- subsidisedprice (P3). Thereare tradedgoods (P1) and tion of the exchange premium ep/eo, a positive func- nontraded goods (P2) in the parallel market and the tion of the difference between prices in the open uncontrolled legal market. Traded goods prices are market and controlled prices and a positive function equal to foreign goods prices (Pf) converted at the of total financial assets, A. The portfolio demand for parallel exchange rate, plus a mark-up (s) for the unit foreign exchange is a negative function of the differ- costs of smuggling. ence between domestic and foreign interest rates and a positive function of total financial assets, A. Pi = (1 + s) (ep. Pf). ep. Fd = h(ep/eo, P/P I, A) + j [(i - i - pe), Al We assume for expository convenience that s = 0 A wherehl>0,h2 >O,h3 >O;jl 0. P1 = ep + Pf The supply of foreign exchange is a function of the exchange premium, the real exchange rate (r), and A third category of final goods transacted in the the ratio of open market prices to the controlled parallel market is nontraded goods whose produc- official price (P/P3): tion requires labor and imported inputs. Foreign ex- change for the purchase of imported inputs comes epFs = n(epleo, r, P/P3) from official sources with exchange rate eo as well as from the parallel market exchange rate ep. The follow- where ni > 0, n2 < 0, n3 > 0. ing variable mark-up model is used to explain non- A rise in P increases the transactions demand for traded goods prices (Pz): foreign exchange as the difference between the open and controlled price grows. A rise in P also increases P2 = (1 + u) [ep. Pf + (1-k) eo Pf al (wp) a 2.) expectations that the parallel exchange rate will de- preciate. This leads to an increase in the portfolio where u denotes the mark-up, wp denotes unit labor demand for foreign exchange, given domestic and costs, and k is the parallel market's share of foreign foreign interest rates. The effect on the supply of exchange used for imported inputs. The size of the foreign exchange is also positive, but the net effect is mark-up is a function of excess demand (ED) in the an increase in the demand for foreign exchange. A economy. So that the rate of change in P2 is; rise in ep is required to offset this net demand increase and restore equilibrium. The direct price effect of ep P2 =bo ED + b1k [ep + tf I + bi ( 1-k ) (eO + Pf) + b 2Wp on Fd is negative. But there isa positive wealth effect. If the substitution effect is greater than the wealth The average rate of inflation is the weighted aver- effect, the overall demand effect is negative and the age of the changes in prices of traded (Pi) and non- ee curve unambiguously slopes upwards (figure 5- traded (P2) goods and goods transacted at controlled 2). An increase in ep also lowers the expectation of prices (P3): future devaluation and thereby lowers the portfolio demand for foreign exchange. P = C1P1 + CAP2 + (1-C1-C2)P3P Figure 5-2 describes the dynamic equilibium in ep and P. In quadrant I we have an excess supply of We assume that there is a surplus in the labor goods but an excess demand for foreign exchange so market such that an increase in aggregate demand that equilibrium is restored with a decline in P and decreases the size of the labor surplus, but has no an increase in ep. In quadrant II we have an excess 40 The Inflationary Consequences of Devaluation with Parallel Markets: the Case of Ghana Table 5-1. The complete model A. Inflation ED=Excess demand 1.1. P = xi P1 +x2 1 2 + (1-xi - x2) P3 EXPT=Export tax revenue 1.2. P1=Pf + eA EXPQ=Export volume 1.3. P, =f5+ Lajw, + a2mc] F=Foreign exchange in parallel market 1.4 mc =Pf + e. FT=Transactions demand for foreign exchange in 1.5. =f(ED); fl >0 parallel market 1.6. ED=log M/p - log Mdlp FP=Portfolio demand for foreign exchange in 1.7. logMd/p= d (logy i ,pe); di > 0, d2 <0, d3 < 0 parallel market 1.8. e= keo + (1-_k) ep Fs=Supply of foreign exchange PDEBT=Foreign debt B. Parallel exchange rate GBB=Government borrowing from the Central 1.9. F=Ff + FP Bank 1.10. epFt=h(ep/eo,P/P3,A); hl < 0, h2 > 0, h 3 > 0 GDEF=Fiscal deficit 1.11. epFP7=j[i-i* - e)p]. A; ji > 0 GFB=Govemment foreign borrowing 1.12. ep.Fs=n (ep/eo,r,P/P3); ni > 0, n2 > 0, n3 > 0. GDB=Government domestic borrowing GEXP=Govemment expenditure C. Money and fiscal aspects GREV=Government revenue 1.13. M2 = m + GOEXP=Other government expenditure 1.14. R=COG + NFA + OA GOVI=Government investment (nominal) 1.15. COG=COG-1 + GBB GDEXP=Other government expenditure 1.16. GBB=GDEF - GFB - GDB gy=Growth rate of real output 1.17. GDEF=GEXP - GREV gk=Growth rate of capital stock 1.18. GREV=CUDT + EXPT + GOREV + AID gPoP=Growth rate of population 1.19. CUDT=IMPQ.Pfeo.tm IMPQ=Import volume 1.20. EXPT=EXPQ.Px.eO.tx INTF=Interest on foreign debt 1.21. AID=$AID.eo INTD=Interest on domestic debt 1.22. GEXP=WBILL+SUBS+INTD+INTF+GOVI i=Domestic interest rate on six-month deposits +GOEXP if=Foreign interest rate 1.23. WBILL=Wg.Lg Ip =Real private investment 1.24. INTF=FDEBT if.eo Ig=Real public investment 1.25. FDEBT=FDEBT-i + GFB/eo K=Capital stock 1.26. INTD=DDEBT.i Kp=Private capital stock 1.27. DDEBT=DDEBT-1 + GDB Kg=Government capital stock 1.28. PCRED=PCRED.1 + PCRED Lg=Government labor growth 129. APCRED=ATCRED -A GCRED M=Money supply (M2); includes 30-day deposits Md=Money demand D. Real side mc=Costs of intermediate imports 1.30. gy=ao + algk + a2gPoP P=CPI 1.31. K=Kp + Kg P1=CPI for traded goods 1.32. Kp=Kp l (1-d) + Ip P2=CPI for nontraded goods 1.33. Kg=Kg . (1-d) + Ig P3=CPI for controlled goods 1.34. Ip=f(PCRED) PCRED=Credit to private sector (real) Pf=lmport price index where Px=Export price index AID=Foreign aid in cedis r=Real official exchange rate $AID=Foreign aid in US dollars TCRED=Total credit (real) COG =Govt. borrowing from the Central Bank tm=Customs duty rate CUJDT=Customs duty eamings tx=Export tax rate DDEBT=Domestic debt U=Mark-up eo=Official exchange rate (cedis/US dollars) wp=Unit labor costs ep=Parallel exchange rate (cedis/US dollars) wg=Government wage level e=Weighted average exchange rate WBILL=Govemment wage bill 41 The Inflationary Consequences of Devaluation uith Parallel Markets: the Case of Ghana Figure 5-2. Comparative Statics with Devaluation comparison with the higher cost-push effect of a of the Official Exchange Rate devaluation (figure 5-2). As long as the cost push effect of the official devaluation is not larger than 12.5 PIP1, then ep, the rate of parallel market depreciation, ,, unambiguously declines, but the net effect on is an empirical question that depends on the relative mag- 10.0 , w ', ' e nitudes of different variables. This empirical ques- . , - - tion will be explored in the simulation model that follows. On the other hand if the official devaluation 7.5 a * leads to a larger budget deficit both ee and PP shift P _. - .outwards and ep and rise unambiguously. 5.0 a ' ' _Y _-P The Empirical Estimation of the Model The detailed empirical model with all the equa- 2.5 e, Itions and identities is presented in table 5-1. The e model is divided into four blocks: inflation, the par- allel exchange rate, monetary and fiscal aspects, and o.oA_ thereal side. Most of the relationships in table 5- are e self-explanatory, but those that are not are discussed Source: Author's calculations. below. supply of goods and foreign exchange and both P Block A: Inflation and ep decline. The direction of movements to equi- librium in quadrants III and IV is self-explanatory. The approach to modelling excess demand in the What happens if the official exchange rate is deval- goods market was to use Walras' law to assume that ued? There are two effects that could potentially go excess demand in the goods market is equivalent to in opposing directions. First, with an official devalu- excess supply in the money market. This implies that ation there is a direct cost-push effect which shifts PP the substitution between money and goods is far upwards to PIP'. more important than that between money and other In addition, there are budgetary effects which af- financial assets. This is a plausible assumption in a fect private asset holdings. If the budget deficit im- developingeconomysuchasGhanawitharelatively proves with an official devaluation and the rate of shallow financial system characterized by adminis- monetary creation declines, then the ee curve shifts tered interest rates and a limited number of financial inwards to e'e' due to a declining wealth effect. The assets. P1Pl curve also shifts inwards as the lower excess The excess supply of money is specified as the log demand reduces the average level of mark-up in the difference of real money supply to real money de- economy. How far PIP' shifts depends on the size of mand (equation 1-6). A standard money demand the wealth effect on demand and on the mark-up in function is hypothesized where Y is real income, i is Table 5-2. Inflation equation with wage vaniables Equation Constant eO + Pf Log (M/P-1) Log y i W w9 p-1 CHi2 DW 2.1 18.8822 0.0887 0.3939 -1.9721 0.0372 0.3447 0.4551 10.09 1.83 (1.64) (0.28) (2.34) (1.76) (1.92) (0.64) (1.85) 2.2 13.6146 0.4812 0.4420 -1.5027 0.0315 -0.1185 0.4404 12.45 1.90 (1.47) (2.04) (3.90) (1.68) (1.00) (0.26) (1.65) All equations were estimated by two stage least squares using PCGIVE. The instruments used were lagged values of inflation, output, money supply, interest rates, import prices at the official and parallel exchange rates, the parallel market premium, the wage gap, and real public sector wages. eo: Change in official exchange rate (cedis/US dollars) i: Six-month deposit rate Pr. Change in foreign price index in US dollars W: Real wage minus productivity M: Money supply (M2) Wg: Real public sector wages y: Real GDP p-1: Lagged inflation 42 The Inflationary Consequences of Devaluation with Parallel Markets: the Case of Ghana the rate of interest on deposits, and e is the expected public sectorwages hasbeen included in the inflation rate of inflation. equation. The results reported in equation 2.2 of table Changes in wages can have inflationary conse- 5-2 also confirm that public sector wages did not quences when they exceed the growth in labor pro- contribute to the inflationary process in Ghana. This ductivity, i.e. a 'wage gap' exists. In order to ass- ess is consistent with a declining trend in real public this channel for the acceleration of inflation, an index sector wages over much of the high inflation period. of nominal wages was constructed. Equation 3.1 in table 5-3 drops the wage variables Regressions of real wage growth on a constant, a as there is no evidence that wages were an important time trend, and a lagged dependent variable always factor in overall price determination in Ghana. The resulted in insignificant coefficients, implying that it resulting equations are very robust with significant was not plausible to assume that productivity fol- coefficients and diagnostic statistics. Expected infla- lowed a time trend in the Ghanaian case. Instead, it tion was specified using both perfect foresight and was necessary to analyze labor productivity more adaptive lagged inflation. The empirical results indi- directly in order to evaluate the inflationary conse- cate that lagged inflation worked better. quences of the wage gap. The proxy used for labor productivity was the growth rate of real output per Inflation and Exchange Rates capita (gypc). The following equation was estimated: The cost-push literature posits that the relation- WGPG = -0.05 CONSTANT + 1.81 (gypc). ship between the exchange rate and inflation oper- (0.84) (1.83) ates through the inelastic demand for imports and inelastic supply of exports that characterize many The residual that resulted from this regression was developing economies. Because of such rigidities, a defined as the "wage gap." The coefficient on labor devaluation implies cost increases for importers productivity is not significantly different from unity. without a concommitant rise in the income of export- Table 5-2 contains the results of econometric esti- ers, at least in the short run. mates of the reduced form inflation equation for Econometricevidenceontherelationshipbetween Ghana. All of the regressions are two stage least inflation and exchange rates is presented in table 5-3. squares estimates. The model performs well with all Equation 3.1 in table 5-3 shows the significance of the of the variables significant and appropriately signed import price index at the official exchange rate with the exception of unit wage costs. This is not (eo+Pf).Equation3.2includesonlytheparallelmarket surprising in Ghana given that a very small share of version of import prices (ep + Pf) along with the basic the labor force is in the formal sector and unionized inflation equation. Again, the coefficient is highly and that there has been tremendous downward pres- significant. These results would seem to imply that sure on wages in recent years. since both the official and parallel market rates mat- An alternative argument might be that it is public ter, the overall rate of inflation is a weighted average sector wages that rise at a faster rate than productiv- of changes in both exchange rates since the market is ity and therefore fuel inflation. Given the difficulties segmented and clearing at two different prices. in measuring public sector productivity, it has been In order to test this hypothesis, a grid search was assumed to be constant so only the growth rate in conducted using varying values of k to determine Table 5-3. Inflation equations, 1965-1988 Equation Constant eO + Pf ep + Pf Log (M/P-1) Logy i p-l CHi2 DW 3.1 14.3193 0.2909 0.4036 -1.5513 0.0390 0.3209 22.62 1.80 (2.07) (2.03) (3.80) (2.30) (2.97) (2.17) 3.2 8.8233 0.1649 0.3566 -1.0049 0.0258 0.5872 69.60 1.39 (1.97) (4.52) (6.90) (2.32) (3.25) (5.22) 3.3 9.1805 0.0691 0.1474 0.3832 -1.0533 0.0281 0.5596 42.74 1.35 (1.80) (0.39) (1.60) (5.31) (2.12) (2.86) (3.32) All equations were estimated by two stage least squares using PCGIVE. The instruments used were lagged values of inflation, output, money supply, interest rates, import prices at the official and parallel exchange rates, the parallel market premium, the wage gap, and real public sector wages. eo: Change in official exchange rate (cedis/US dollars) y: Real GDP ep: Change in parallel market exchange rate (cedis/US dollars) i: Six-month deposit rate Pr. Change in foreign price index in US dollars p-I: Lagged inflation M: Money supply (M2) 43 The Inflationary Consequences of Devaluation with Paralkl Markets: fhe Case of Ghana whether the overall fit of the inflation equation im- The supply of foreign exchange to the parallel proved with varying weights on a composite foreign market will depend on the exchange premium, the price variable. The result was that the fit deteriorated real official exchange rate, and the difference be- as the relative weight of the official exchange rate tween prices in the open market and controlled increased. This was true even when the sample pe- prices. The resulting empirical model which deter- riod for the grid search was restricted to the pre-1983 mines the parallel market exchange rate is based on period. This is an indication that consumer prices the supply and demand factors for foreign exchange reflected the parallel market price of foreign ex- in the parallel market. This model posits that the change even during the period prior to the introduc- parallel market premium depends on the real effec- tionof theauction.Thisisalso confirmedbyequation tive official exchange rate, the depreciation-adjusted 3-3 in which both the official and parallel market interest differential and the stock of real money bal- exchange rates are included in the inflation equation ances in the economy. and the parallel rate is far more significant. The infla- This model was estimated for Ghana using the tion equation was also estimated recursively and the cedi/dollar exchange rate in the parallel market on results of the one-step Chow tests indicated that the the left hand side and the real official exchange rate, parameters were stable at the 1 percent confidence the depreciation-adjusted interest rate differential level over the entire sample period. To summarize, between Ghana and the United Kingdom, and the the results on the determinants of inflation indicate real stock of cedi assets on the right hand side. The that the relevant cost-push variable is the parallel resulting estimate is reported as equation 4.1 in table market exchange rate. 5-4. The model shows that the coefficient of the relative yield variable is significant but that of real Block B. Exchange Rate Premium money balances is insignificant. The significance of the relative yield variable implies that as long as real A dynamic variant of the theoretical model in Sect- interest rates in Ghana diverge widely from world ion 2 is estimated to identify the factors which deter- rates, there will always be a parallel market pre- mine the differential between the official exchange mium. Exchange rate unification therefore implies rate and the parallel market exchange rate. The dem- eventual opening up of the capital account so that and forforeignexchangein theparallelmarket is div- there is legal arbitrage between Ghanaian cedis and ided into transactions demand and portfolio demand foreign assets. This result highlights the important (equation 1.9). The transactions demand depends on linkages between exchange rate management and the exchange premium and the degree of price cont- domestic interest rates. Although a number of other rols (equation 1.10). In determining their asset portf- factors, such as confidence and credibility, offset olio demand, agents choose between holding Ghana- movements in the premium, the interest rate consti- ian cedis and foreign exchange. The relative shares of tutes an accessible policy instrument for influencing cedis and foreign exchange in the public's asset hold- the exchange rate premium. ings will depend on the relative returns between ced- The real exchange rate and the real stock of domes- is and foreign exchange. This will depend on domes- tic assets are insignificant in equation 4.1. The equa- tic and foreign interest rates and the rate of expected tion performs much better when a dummy variable depreciation in the parallel market (equation 1.11). is included for 1978, the year in which the cedi's link Table 5-4. Exchange rate premium equations, 1965-1988 Equation Constant r Ry M2P Dum 78 PVAR CHi DW 4.1 -0.5157 -0.0004 0.9992 0.0040 2.87 2.38 (0.74) (0.17) (2.56) (0.63) 4.2 -0.0303 -0.0012 1.1486 -1.8476 8.54 2.33 (0.20) (0.85) (4.00) (4.01) 4.3 -0.4874 -0.0038 1.3173 -2.4904 1.8525 13.96 1.73 (2.81) (2.87) (5.67) (6.21) (3.56) All equations were estimated by two stage least squares using PCGIVE. The instruments used were lagged values of the exchange rate differential, the real exchange rate, relative yields, price variability, inflation, and foreign exchange as a proportion of imports. r : Real exchange rate index (cedis/US dollars) Ry: Relative yield between cedis and UK Stg. DUM 78: Dummy variable for 1978 PVAR: Index of monthly variation in prices M2P: M2/CPI 44 The Inflationary Consequences of Devaluation nith Parallel Markets: the Case of Ghana to the US dollar was severed and the official ex- have occurred, some of the underlying economic change rate was devalued. The resulting estimates relationships have been captured in the model. are reported as equation 4.2 in table 5-4. In the case of Ghana, the differential between the If there is imperfect substitutability between Gha- official and parallel market exchange rates was so naian cedis and foreign exchange, the exchange rate large that the share of transactions that occurred premium is also affected by uncertainty about future through official channels decreased steadily over exchange rates, inflation and government policies. time. The spread between the official and parallel While it is not possible to quantify much of this unc- market exchange rates is depicted in nominal terms ertainty adequately, the effects of price instability in figure 5-3. Because of this leakage to smuggling have been explored empirically. Avariable represent- and parallel market activities, the efficacy of theover- ing price volatility over time was constructed (PVAR) valued official exchange rate as an export tax dimrin- using monthly consumer price indices. In periods of ished. The Ghanaian government's strategy was one rapid price changes, there is greater pressure on the of a gradual reduction of the fiscal deficit combined authorities to devalue which results ina reduction of with a relaxation of exchange rationing and large, the exchange rate differential. The inclusion of PVAR discrete devaluations. The reduction in the fiscal also results in the real exchange rate becoming sig- deficit was necessary to reduce demand in the econ nificant. Because of the major changes in the foreign omy and thereby reduce the rate of inflation and the exchange regime, equation 4.4 was also estimated rate of depreciation of the parallel exchange rate. recursively to analyze the stabilityof the coefficients. As with the inflation equation, the one-step Chow Block C: Monetary-Fiscal effects tests indicated parameter stability at the 1 percent confidence level over the sample period. These test Monetary growth is modelled conventionallyas a results imply that, despite the policy reforms that function of the money multiplier and the growth oi Figure 5-3. Parallel and Official Nominal Exchange Rates, 1965-87 Exchange rate ($/cedi) 1.2 1.1 1 , 0.9 \ . , ~~~---------- 0.8 - 0.7 0.6 0.5 0.4 0.3 0.2- - 0.1 1965 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 1987 Year Parallel - ----Official Source: World Bank data. 45 The Inflationary Consequences of Devaluation uith Parallel Markets: the Case of Ghana reserve money, which in turn is affected by the size a share of GDP, and import prices as a ratio of the of the fiscal deficit and the manner in which it is GDP deflator. financed. The fiscal block is included in the model to incorporate the effect of price changes on various The negative relationship between agricultural elements of public expenditure and revenue. Fiscal output and subsidies shows that the govemment att- revenues are a function of the exchange rate, taxes on empted to cushion the effects of food price inflation foreign trade, and output. Government expenditures due to droughts. This cushion constitutes a greater are disaggregated between wages, subsidies, interest burden for the govemment subsidy bill when the rel- payments, public investment, and other government ative price of imports is greater than domestic prices. expenditure. These components of govemment ex- penidture depend, in tum, on variables such as the Block D: The Real Side: Investment and Output exchange rate and output. The size of the fiscal deficit is thus the endogenous outcome of govenmment pol- In order to explore the consequences of price dy- icy. The identities in the model are self-explanatory. namics for the real side of the economy, a simple For the behavioral equations the following empirical model of output and investment determination was results were observed. estimated for Ghana. In addition, an export function Othergovenmment revenues, which are essentially is defined to analyze the interaction between ex- non-trade tax receipts, were modelled as a function change rates, the balance of payments and inflation. of nominal GDP: Exports are hypothesized to be a function of the real exchange rate, agricultural output, and lagged (5-1) log (GOREV) = -6.1962 + 12793 log (GDP) exports: (6.41 ) (16.15) (5-4) log(EXPT) = -3.4206 - 0.2560 log(r)(-1) DW = 1.74; Chi2 = 5273.6; TSLS (0.94) (1.67) Instruments: lags of government expenditure and of + 1.0286 log(GDP (-1) AG) GDP. (1.35) + 0.7978 log(EXPT) (-1) Government investment is treated as an exoge- (3.38) nous policy variable. Other government expenditure is modelled as a funcfion of nominal output: DW = 2.02; R-2 = 0.58, OLSQ (5-2) log (GOEXP) = -2.5374 + 0.9516 log (GDP) The equation indicates a relatively low short run (8.2 0) (34.03) elasticity of exports of about 0.26 with respect to the real exchange rate. This is not surprising given the DW = 2.15; Chi2 = 11955.0; TSLS long gestation periods associated with some of Instruments: lags of government revenue and of Ghana's major exports-cocoa and forestry prod- GDP. ucts. Note also that the long-run elasticity of real exports to the real exchange rate in Ghana is about 1. Government expenditure on subsidies and trans- This is a very plausible number and close to that fers reflects the difference between controlled and obtained in numerous studies on export elasticities. market prices as well as operating losses of the enter- Output is hypothesized to be determined by a prises and social security payments. Econo-metri- Cobb-Douglas production function. For empirical cally, transfers and subsidies as a share of GDP estimation, it was assumed that the participation rate (SUB/GDP) have been modelled as a function of ag- was constant over the period so population growth ricultural output (GDPAG) and the import price in- could serve as a proxy for the growth of the labor dex relative to the GDP deflator (PMI/YDEF): force. Theempirical estimate of this production func- tion using TSLS was: (5-3) log(SUB/GDP) =-3.0443-0.2452 (3.72) (3.13) (5-5) log (Y/L) = 1.7052 + 0.4481 log (KIL) log(GDPAG)+0.9258 log(PMI/YDEF)(-1) (6.44) (7.47) (3.96) DW = 0.68; Chi2(B = 0) = 27313.41; TSLS DW = 1.76; Chi2 = 792.14; TSLS Instruments: lags of the capital-labor ratio, output-la- Instruments: lags of GDP in agriculture, subsidies as bor ratio, and deviation of output from trend. 46 The Inflationary Consequenwcs of Devaluation uith Parallel Markets: the Case of Ghana The capital-labor ratio is significant and appropri- interest rate and govemment capital formation to ately signed. The coefficient implies that the share of private investment determination was verified capital in output is approximately 45 percent. econometrically. The quantity of credit reflects the A numberof different models of investment deter- rationing that occurs under administered interest mination were tested on the data for Ghana. Because rates. The elasticity of private investment to real of the numerous shocks to the economy, the absence private credit is about 0.58. This implies that a ten of any real growth during much of the period, and percent reduction in real credit to the private sector the existence of considerable uncertainty, the private will result in a 5.8 percent decline in real private investment rate was fairly volatile. In addition, the investment. This is a very sizable impact and has data for private investment, which is derived as the obvious implications for the design of a program difference between aggregate investment, which is oriented towards the recovery of private investment. estimated largely on the basis of imports, and gov- However, the availability of credit is only a partial emnment investment, may not be very reliable. In explanation since banks held excess liquidity during particular, the private investment series displays a theperiod of credit rationing. The recoveryof private sharp drop in 1976 that is difficult to explain. The investment in Ghana will depend in part on alleviat- following simple equation provides the best explana- ing the financing constraint, as well as on improving tion of the determinants of private investment over the allocation of credit and addressing an array of the period: issues related to private sector confidence. (5-6) PRIVID = 9.7888 + 0.5785 (PCRED) - 212118 Did the Official Devaluations Accelerate Inflation? (3.50) (3.08) (4.93) (DLIM76) The econometric results presented in the previous section showed that a 1 percent devaluation leads to DW = 1.46; CHi2 = 97.38; TSLS about a 0.14 to 0.16 percent immediate increase in Instruments: lags of Private Investment and of Pri- inflation. The longrun effects are evenlargerbecause vate Credit. of the important role of lagged expectations. To reca- pitulate the lagged inflation variable was highly sig- This implies that investment is a function of the nificant with a coefficient of about 0.5. This implies real quantity of credit available to the private sector that if the short run effect of exchange rate devalu- and a dummy variable to capture the unusual col- ation is about 0.15 the long run effect would be twice lapse of investment in 1976. The irrelevance of the that or about 0.30. Figure 5-4a. CPI with Slower Official Devaluation Figure 5-4b. Parallel Market Exchange Rate with 1981-88 (1980 = 100) Slower Devaluation, 1982-88 CPI Cedi/U.S. dollars 2500 - 2500 2000 - 2000 Slower devaluation 1500 1500 1000 - j tual 1000- 500 5 l l00 Slower devaluation 1981 1982 1983 1984 1985 1986 1987 1988 1982 1983 1984 1985 1986 1987 1988 Year Year Source: Author's calculations. Source: Author's calculations. 47 The Inflationary Consequences of Devaluation uith Parallel Markets: the Case of Ghana These are the effects on inflation of movements in kets. The results address, among other things, a cen- the tree market exchange rate. The Ghanaian story is tral criticism of adjustment programs in Africa, i.e., somewhat more complicated because the official rate that excess-ive reliance on the exchange rate as a tool has been catching up to the parallel market rate and to bring about relative price adjustments is exces- tlis has had budgetary implications. In Ghana, the sively inflationary. In the case of Ghana, which has devaluation of the official rate led to budgetary im- carried out one of the most thorough adjustment provements in recent years since the govemment is programs in Africa including very large adjustments a net supplier of foreign exchange to the economy. in the official exchange rate, high inflation was and The government's supply of foreign exchange in- continues to be a problem. creased as a result of the devalution since a larger However, the recent inflationary surge in Ghana is prcportion of transactions occurred through official not due to the exchange reforms. The empirical res- channels. In addition, the devaluations reduced the ults for Ghana show that the official devaluations large subsidy to those importers with access to for- had nodirecteffecton consumerprice inflation.Ifan- eign exchange at the official exchange rate. ything, the official devaluations had a positive budg- In order to quantify these effects, the model in the etary effect which was anti-inflationary. Statistical U lird section was simulated with a slower official tests show that both before and after 1983 the rel- exchange rate depreciation of 25 percent in the period evant exchange rate for pricing decisions was the f: tn 194 to 198. The results of this simulation show parallel market price of foreign exchange. In fact, a that the parallel exchange rate would have depreci- slower devaluation would have led to higher depre- ated aster (figure 5-4), and the exchange premium ciation of the parallel exchange rate and higher infla- .would have increased. An additional cost of such a tion according to our model simulations. The model policy would be lower growth as government bor- also explains the evolution of the parallel market rowing to finance the higher fiscal deficit arising premium as a function of the real exchange rate from slower exchange rate adjustment would calculated at official prices, interest rate differentials sqleeze out private investment. and uncertainty. Devaluation of the official exchange One would expect lower inflation in the simula- rate leads to a narrowingof the differential. Similarly, tion as the parallel exchange rate depreciates more sl- reduction in the relative yield through either upward owiv. In fact, because of lower growth, inflation turns movement of domestic interest rates or through a out to be higher under this formulation, averaging reduction in the expectation of a devaluation also 29.3 percent per annum as against 27.8 percent per leads to a narrowing of the exchange rate premium. annum in the base case during 1983 to 1988. The net The improvement in the budgetary position of the outcome is lower growth, a larger public debt, higher government in response to official devaluation was nflation and a more overvalued official exchange a crucial element of the ability of the authorities to r'te. This result is somewhat counter-intuitive and is sustain the exchange unification. An important ele- dS le to three factors: (a) the official devaluations did ment of the credibility of the program was the heavy not have a cost-push effect as prices had already adj- inflow of concessional assistance. This external sup- usted to the parallel exchange rate; (b) the devalu- port strengthened the political hand of those advo- ation had a positive budgetary effect and therefore cating economic reform. This has significant lessons ddid not fuel inflation-as a result, a substantial real for exchange rate reforms in Africa which can be devaluation took place which narrowed the pre- destabilising where there is not adequate support for n,ium; and (c) relative yields on cedis versus foreign the adjustment program. It is useful to contrast the exchange improved because the large official devalu- Ghanaian experience with the Zambian reforms ation reduced expectations of future devaluation. In where an important factor responsible for the unrav- summary, this simulation has shown that the official ellingof the introduction of the exchange rate auction devaluations could not explain the high and persist- was the underfunding of the program. ent inflation in Ghana since the reforms and particu- laxly since 1986 when devaluations appear to have References contributed to a reduction in inflation Azam, J and T. Besley (1989), 'General Equlibrium with Conclusions Parallel Markets for Goods and Foreign Exchange: The- ory and Application to Ghana, World Developmenl, Vol. This paper has presented a theoretical framework 17, No. 12, pp. 1921-1930. for analyzing the economic consequences of ex- Balassa. B (1985), ChangeandChallengein the World Economy, change rate reform in the presence of parallel mar- MacMillan, London. 48 Thet Inflationary Consequences of Devaluation with Parallel Markets: the case of GhAana Blejer, M. and M. Khan (1984), "Government Policy and Gordon, R. (1975), "Alternative Responses of Policy to Ex- Private Investment in Developing Countries," LMF Staff ternal Supply Shocks," Brookings Papers on Economic Ac- Papers, volume 31. tivity, volume 1. Bruno, M. (1978), Exchange Rates, Import Costs and Wage- Greene, J. (1989), 'Inflation in African Countries: General Price Dynamics," Journal of Political Economy, volume 86, Issues and Effect on the Financial Sector,' IMF Working pp. 379-404. Paper. Bruno, M. and J. Sachs (1985), Economics of Worldwide Stag- Herbst, J. (1990), "Exchange Rate Reform in Ghana: Stra- flation, Cambridge, Massachusetts: Harvard University tegy and Tactics," mimeograph, Princeton Universitv. Press. Kalecki, M. (1971), Selected Essays on the Dynamics of the Cap- Chhibber A. and S. van Wijnbergen (1988), 'Public Poliicy italist Economy: 1933-1970, Cambridge University Press. and Private Investment in Turkey', World Bank PPR Khan, M. (1980), "Monetary Shocks and the Dynamics of Working Paper, Washington, D.C. Inflation," IMF Staff Papers, volume 27, number 2. Chhibber, A., J. Cottani, R. (1989), Firuzabadi and M. Wal- Khan, M. and J. Lizondo (1987), 'Devaluation, Fiscal Defi- ton, 'Inflation, Price Controls and Fiscal Adjustment in ct, and the Real Exchange Rate," The World Bank Eco- Zimbabwe," PPR Working Paper, The World Bank, nomic Review, volume 1, number 2. Washington, D.C. Killick, T. (1972), "Price Controls, Inflation and Income Chhibber, A., and N. Shafik (1990), "Exchange Reformn, Distribution: The Ghanaian Experience," Harvard Uni- Parallel markets and Inflation in Africa: The Case of versity Centerforlnternational Affairs, Economic Devel- Ghana', PRE Working Paper number 427, The World opment Report no. 223. Bank, Washington, D.C. Killick, T. (1978), Development Economics in Action: A Corbo, V. (1985), "Intemational Prices, Wages and Inflation Study of Economic Policies in Ghana, New York St. in an Open Economy: A Chilean Model," Review of Eco- Martin's Press. nomics and Statistics, 68. Leechor, C. (1991), "Ghana's Adjustment Program" in Tho- Dombusch, R. (1986), "Special Exchange Rates for Capital mas et.al. Strucutral Adjustment and the World Bank, Ox- Account Transactions," World Bank Economic Review, vol- ford University Press. ume 1, number 1. Pinto, B. (1989), "Black Market Premia, Exchange Rate Uni- Dornbusch, R. (1988), "Exchange Rate and Inflation", The fication and Inflation," The World Bank Economic Review, M.I.T. Press, Cambridge, Mass. and London, England. Volume 3, Number 3. Edwards, S. Real EchangeRates, Dewluation andAdjustment: Rocha, R. (1989), "The Black Market Premium in Algeria Exchange Rate Policy in Developing Countries, The M.I.T. mimeograph, The World Bank, Washington, D.C. Press, Cambridge, Mass and London, England. Shafik, N. (1990), "Modeling Investment Behaviorin Devel- Fardi, M. (1991), "Zambia's Adjustment Program" in Tho- oping Countries: An Application to Egypt," PRE Work- mas, Chhibber Dailami and de Melo, ed., Structural Ad- ing Paper number 452, The World Bank, Washington, justment and 7he World Bank, forthcoming, Oxford Uni- D.C. versity Press. Sundararajan, V. and S. Thakur (1980), "Public Investment, Fischer, S. (1981), "Relative Shocks, Relative Price Variabil- Crowding Out, and Growth: A Dynamic Model Applied ity, and Inflation," Brookings Papers on Economic Activity,. to India and Korea," LMF Staff Papers, volume 27. 49 6 The Macroeconomics of the Unofficial Foreign Exchange Market in Tanzania Daniel Kaufmann and Stephen A. O'Connell An active parallel foreign exchange market has ex- dow accounted for roughly 40 percent of import isted in Tanzania since the early 1970s. The licenses issued in 1988. Unoffidal estimates suggest characteristics of the market have varied over time in that the share of own-funded imports in total imports response to economic shocks and the evolving policy is even larger perhaps significantly exceeding one regime, with savingand portfoliodecisionsfeaturing half.3 Given these orders of magnitude, it is clearly importantly in some periods and illegal trade trans- critical in the current policy context to have an un- actions in others. At present, the market is very derstanding of the parallel foreign exchange market extensive, with both trade and financial transactions in Tanzania. playing important roles. The second section begins with a brief summary of This paper analyzes the macroeconomics of the major policy developments since Independence. We parallel foreign exchange market in Tanzania. It fo- then provide a detailed overview of macroeconomic cusses on the following questions: What factors led developments in Tanzania since 1967, concentrating to the emergence of an unofficial market? What fac- on the external sector and the evolution of the pre- tors determine the premium between the unofficial mium on foreign exchange in the parallel market. In and official exchange rates? What are the linkages the third section, we specify and estimate a simple between the unofficial foreign exchange market and empirical model of the parallel premium using an- the rest of the economy? The answers to these ques- nual data from 1966 to 1988. The fourth section sum- tions have important implications for macroeco- marizes the key conclusions and indicates our nomic management in Tanzania and provide essen- agenda for future work. tial background for consideration of such current policy issues as unification of the foreign exchange Economic Structure and Policy*. An Overview markets. While the exact size of the unofficial foreign ex- Figure 6-1 shows quarterly movements in the par- change market is (by its very naturel) difficult to allel premium on the U.S. dollar in Tanzania since judge, a recent policy measure legalizing one key 1970 (table 6-1 gives annual observations back to dimension of the market provides an unambiguous 1966). Data for the unofficial rate are from the World indication of current orders of magnitude. Starting in Currency Yearbook, supplemented after 1984 by a mid-1984, individuals with access to unofficial for- small survey carried out in Dar es Salaam by Mali- eign exchange were allowed to obtain import li- yamkono and Bagachwa (1990).4 The figure also censes without accounting for the source of their shows the official exchange rate against the dollar, funds.2 Based on official figures, the own-funds win- periods of discrete devaluation against the relevant 50 The Maoeconomics of the Unofficial Forigxn Exchange Market in Tanzania Figure i-I. Tanzania: Parallel Premium vs. Official Exchange Rate, 1970:3 - 1989.4 7.5 - 1974:1 1979:1 1984:1 5.0 25 2.5 i ii, - - - - - _ , , O- l l l l l l l l l l l ~~~~~~~~~~~I .I I I I 1971 72 73 74 75 76 77 78 1979 80 81 82 83 1984 85 86 87 88 89 Year Log of end of quarter parallel premiurn. - - - - -Log of end of quarter official exchange rate (TShs/$). Note: Stars (*) denote discrete devaluations of the official rate. Source: World Bank data. currency basket are identified with asterisks. Table tion; the activities of trade finance were taken over 6-2 provides a list of currency realignments over the by the National Bank of Commerce after nationaliza- sample. tion of the banking sector. The parallel premium shows substantial vari- Despite the abrupt and major institutional ationsover timeinboth trend and level. BetweenJuly changes, real GDP per capita grew at an average 1970 and March 1986, the premium increased at an annual rate of 5.2 in the period from 1967 to 1973. average rate of nearly 1 percentage point per month; With the implementation of the Second Five-Year from April 1986 to the end of the sample (the period Plan (1969 to 1974), the gross investment rate rose to of the Economic Recovery Program), it declined at a above 20 percent of GDP (from an average of 14.3 rate of over4percentagepoints permonth.5Fluctua- percent between 1964 and 1968), and Tanzania tions around trend, which are often large and persist- achieved significant improvements in the social sec- ent, occur throughout the sample in response to tors, particularly in education and health. Structural changes in the macroeconomic and regulatory envi- weaknesses were already beginning to emerge, how- ronment. ever, which would become more pronounced and affect future economic performance. Agricultural ex- 1967 To 1973 ports began to stagnate in volume terms in the late 1%0s. Domestic savings performance reached a peak After the Arusha Declaration of 1967, the Govern- of 18.1 percent of GDP in 1970 but fell to 15 percent ment rapidly consolidated its control over all major by 1973, and then dropped as low as 8 to 9 percent of aspects of the economy. In the external sector, the GDP in the crisis years 1974-75. The widening gap eight major private import-export firms were nation- between investment and domestic savings was re- alized and replaced by the State Trading Corpora- flected in the external accounts: between 1970 and 51 The Macrocconomics of the Unoffial Foreign Exchange Market in Tanznia Table 6-1. Exchange rates and the parallel premiwn, 1966-1989 Official exchange rate Unofficial exchange rate Paralld premium TShs/$ (percentage) Period (avg.) (eop) (avg.) (eop) (avg.) (eop) 1966 7.14 7.14 8.6 8.6 20.4 20.4 1967 7.14 7.14 8.7 8.8 21.8 23.2 1968 7.14 7.14 85 8.3 19.0 16.2 1969 7.14 7.14 8.7 9.1 21.8 27.4 1970 7.14 7.14 10.1 105 40.8 46.3 1971 7.14 7.14 11.6 15.0 62.2 110.0 1972 7.14 7.14 15.2 15.4 113.0 115.6 1973 7.02 6.90 14.5 13.5 106.9 94.9 1974 7.14 7.14 13.5 13.0 885 96.0 1975 7.41 8.26 20.6 25.0 176.5 202.5 1976 8.38 8.32 21.9 20.4 161.8 145.1 1977 8.27 7.96 21.5 15.1 159.1 89.1 1978 7.69 7.41 13.1 11.8 69.8 58.5 1979 8.25 8.22 12.0 13.5 45.2 64.2 1980 8.19 8.18 21.0 265 156.6 223.9 1981 8.29 8.32 27.6 24.4 232.7 192.6 1982 9.33 9.57 32.6 29.2 247.5 204.7 1983 11.26 12.46 39.6 50.0 252.8 301.4 1984 15.51 18.11 55.9 70.0 259.5 286.6 1985 17.35 16.50 100.8 150.0 487.6 809.1 1986 34.26 51.72 165.0 180.0 478.0 248.0 1987 65.62 83.72 180.0 190.0 178.3 127.0 1988 100.39 125.00 210.0 230.0 110.2 84.0 1989 144.47 190.00 254.2 300.0 76.7 57.9 Source Official exchange rates: IMF, Intemational Financial Statistics, Unofficial exchange rates: World Currency Yearbook (formerly Pick's CurrencyYearbook) for monthly rates from July, 1970 to January,1984; unofficial survey b Maliyamkono and Bagachwa at University of Dares Salaam for monthly rates from February 1984 to November 1989; our estimate for December, 1989. Unofficial rates before July, 1970 are based on occasional observations reported in Pick's Currency Yearbook. 1973, the trade deficit was already 6 or 7 percent of demand, stagnating exports, and capital flight. Tan- GDP, as compared with balanced trade in the mid zania's first (minor) balance of payments crisis oc- and late 1960s. curred in 1970-71, when international reserves fellby Under the Currency Board system, balance of pay- 25 percent between the end of 1969 and the end of ments problems had been virtually nonexistent in 1971. Relative to the risingimportbill, the decline in East Africa. The currency issue of the East African reserves was more drastic, with import coverage Currency Board was backed virtually 100 percent by dropping from 4 months to less than 2 months over sterling so that the currency stock moved one-for- the period. The Government weathered the crisis by one with the sterling reserves of the Board (see New- tightening import controls, extending exchange con- lyn 1967).6 Proponents of an independent central trols to Kenya and Uganda, and mobilizing inflows bank viewed the introduction of the Bank of Tanza- of concessionary financing (Green, Rwegasira and nia in 1966 as an opportunity to move to a less van Arkadie 1980). conservative monetary policy more in tune with the The use of direct controls for balance of payments country's ambitious development program. adjustment was consistent with the ongoing transi- Tanzania's balance of payments performed favor- tion to socialism and the Government's already-es- ably in the first three years of operation of the Bank tablished aversion to exchange rate devaluation;7 it of Tanzania, with gross reserves rising steadily from was institutionalized with the introduction of for- 1966 to 1969. Serious pressures first began to emerge eign exchange budgeting in 1970/71. A domestic in the early 1970s, in response to the rise in internal credit planning apparatus was introduced in the 52 The Macroeconomics of the Unoffiial Foreign Echange Market in Tanzania Table 6-2. Major party changes since 1970 financed by increased aid and capital inflows in 1974 Period Change and 1975, as the government froze wages and re- stricted imports other than oil and food in an attempt July 1973 revaluation 3.4% to manage the short-term situation. Intemational re- January 1974 devaluation 3.4% serves fell by nearly 70 percent in 1974, and for most October 1975 switch to SDR and devaluation 16% of 1975 covered only about three weeks of imports. January 1979 switch to undisclosed basket, and The Government's management of the 1974-75 cri- devaluation 10% sis represented a conscious decision not to sacrifice March 1982 devaluation 10% the development program in the face of adverse cir- June 1983 devaluation 20% cumstances (Green, Rwegasira and van Arkadie June 1984 devaluation 36% (1980). In practice, this meant increased reliance on June 1986 devaluation 55% and initiation of aid and capital inflows. Pressures on the external weekly crawl accounts were eased dramatically in 1976-77 with the November 1988 devaluation 23% recovery of domestic food production and the arrival December 1989 devaluation 29% of the coffee boom. By the end of 1977, reserves were at the unprecedented level of nearly 5 months of imports. Fiscal pressures were eased as well, since the same year, with the intention of implementing the Government chose to tax away most of the windfall Government's development priorities and influenc- in export proceeds. ing the overall growth of credit. The weak underlying external situation emerged With the advent of the semi-annual foreign ex- dramatically in 1978, however, when the govern- change plan, trade policy (as represented by the set ment loosened import constraints in response to the of import quotas implicit in foreign exchange alloca- boom-related inflow of foreign reserves (as it had tions) became an endogenous function of foreign done in 1973).8 As imports expanded dramatically, exchange revenues. With the exception of a brief the coffeeboom collapsed; the current account deficit period immediately following the Arusha Declara- rose to above 15 percent of GDP in 1978 and gross tion, the parallel premium in Tanzania was below 30 reserves fell by nearly $200 million over the course of percent until the 1970-71 mini-crisis. From the expe- the year. External arrears appeared for the first time rience of other countries, a premium of this magni- in 1978. tude is consistent with the operation of binding capi- The parallel premium fluctuated dramatically tal controls in a stable and otherwise relatively un- over the 1974 to 1978 period, rising to above 250 distorted macroeconomic environment. With the percent by the end of 1975, and nearly as high again emergence of the crisis, however, the premium in the first half of 1977, and then falling sharply moved above 50 percent. By the end of 1971, follow- starting in the third quarter of 1977. These move- ing the extension of exchange controls to Kenya and ments reflecta numberof macroeconomic influences, Uganda, it exceeded 100 percent. While these short- including increased savings incentives associated run movements can plausibly be attributed to inten- with the temporary coffee boom revenues (see sified desires for capital flight, the premium did not Bevan, Collier and Horsnell 1985) and the foreign aid return to its previous low level after the bulk of the inflows and import liberalization in 1978. The 1977 nationalizations had been accomplished. This cor- breakup of the East African Community was a fur- roborates ourobservation (see also Green, Rwegasira ther influence; in that year, the Tanzanian Govern- and van Arkadie (1980)) that although the 1970-71 ment closed its border with Kenya, which probably mini-crisis was successfully contained by short-run raised the cost of illegal trade between the two coun- measures, the crisis itself was an indication of the tries substantially.9 emergence of more fundamental imbalances. 1979 To 1984 1974 To 1978 This was a period of fiscal and external crisis and The 1974 to 1978 subperiod began with the coun- cumulative economic collapse. Soon after the ill- try's first serious balance of payments crisis, brought fated import liberalization of 1978, the economy was on by two years of drought and the first oil shock. hit by the second OPEC oil price increase and the The crisis exposed some of the longer-term weak- onset of war with Uganda. In contrast to the balance nesses in economic performance. Large current ac- of payments crises of 1970-71 and 1974-75, when count deficits (averaging 14 percent of GDP) were Tanzania managed to maintain consumption per 53 The Macroeconomics of the Unofficial Foreign Exchange Market in Tanzania capita through increased inflows of external capital, shortages were not averted by inflows of illegal imn- adjustment to the economic shocks of 1978-79 even- ports financed by unofficial foreign exchange and tually required a substantial cutback in both con- sold at market-clearing prices. Surveys conducted by sumption and aggregate investment. Bevan et al (1c 90) document that especially in rural The macroeconomic collapse that unfolded in the areas, goods could not in fact be obtained even first half of the 1980s has been analyzed carefully by through illegal channels. One reason is that the elabo- Bevan et al (1990), Ndulu (1987), and others. Starting rate system of cantrols on distribution meant that the in 1979, the government tightened import controls costs of avoiding detection were extremely high. A severely, while at the same time raising producer second factor is that the activity of smuggling is prices for export crops in the hope of improving transport-intensive and therefore subject to the seri- export incentives. The import compression fell par- ous deterioration of infrastructure that occurred be- ticularlystronglyon intermediates and consumption ginning in the late 1970s. Illegal activity was further goods, in line with government and donor priorities discouraged by the 1983 "economic saboteurs" cam- which remained strongly geared towards the Basic paign during which a large number of businessmen Industrial Strategy objective of increasing manufac- were jailed (Maliyamkono and Bagachwa 1990). turing capacity (and over 1978 to 1981, defense). The A concomitant of the shortages that emerged start- compression of intermediate imports drove down ing in 1979 was that the monetary expansion of that capacity utilization in the manufacturing sector; to- yeardid not immediatelypushprices up; instead, the gether with the direct compression of consumer im- velocity of money fell by 40 percent in 1979, and then ports, this produced a severe shock to the supply of stayed at the lower level until the introduction of the consumer goods. own-funds scheme and domestic price decontrol in In an exchange-control regime without domestic 1984. Although the unofficial exchange rate did not price controls or government control of internal immediately reflect the expansion of real money bal- trade, a reduction in the supply of consumer goods ances (one would expect a depreciation in the pres- would be equilibrated by a rise in the domestic rela- ence of portfolio substitution between domestic tive price of imports and import substitutes (and money and unofficial foreign exchange holdings), it probably a rise in the real consumption rate of inter- did begin to rise dramatically by the end of 1979. est, provided the shock were viewed as temporary). During the early 1980s a number of attempts at In Tanzania, where price controls were pervasive 10 policy reform (e.g., small devaluations) failed to ad- and a "confinement" policy restricted most domestic dress the keyproblems and did not always obtain the and foreign trade operations to selected parastatal necessary political support.12 The 1984/85 budget agencies, the consequence of import compression represented a turning point, and provided the first was the emergence of widespread shortages of con- indication of a major shift towards pragmatism in the sumer goods, particularly in rural areas. Moreover, Government's economic management. The ex- since the government resisted devaluing the ex- change rate was devalued by one-third, parastatal change rate, the rise in producer prices meant sub- subsidies were cut, an import liberalization program stantial losses by the exporting parastatals and a was initiated through introduction of the own-funds corresponding increase in the public sector borrow- import scheme, an export retention scheme was in- ing requirement and in inflationary pressure.11 Un- troduced allowing exporters to retain a portion of der the combined pressure of shortages and falling their proceeds to purchase imports, and restrictions real producer prices, peasants retreated into subsis- on the movement of grain were eased. Simultane- tence production (Bevan et al 1987, 1990) and, to a ously, cooperatives (which had been abolished in limited degree, increased smuggling of the export 1976) were reestablished, and took over many of the crop. functions of the parastatal crop authorities. The collapse of recorded exports in the early 1980s was dramatic: exports declined by roughly 10 per- 1985 To Present cent between 1979/80 and 1981/82, and then by a further 20 percent in 1983. Against this background This was a period of regime change and gradual of macroeconomic collapse, the parallel premium recovery. The period from 1985 to the present is one increased dramatically throughout the third subpe- of gradual economic recovery coinciding with a su- riod, with only minor and short-lived interruptions tained liberalization of economic policy. In mid-1986, in response to devaluations of the official rate. Given the Government produced a medium-term 'Eco- the key role of shortages of consumer goods in exac- nomic Recovery Program." The Economic Recovery erbating the collapse, it is important to ask why Program aimed at achieving a positive growth rate 54 The Macroeconomics of the Unofficial Forcign Exchange Market in Tanzania in per capita income, reducing the rate of inflation, portfolio factors in the determination of the parallel and restoring a sustainable balance of payments po- premium, both in the short run and over time. In sition. Its main thrust was to reduce distortions and table 6-1, we address this issue using static and dy- encourage more efficient resource allocation while namic versions of the Dombusch, et al (1983) model exercisingfiscal and monetaryrestraint. In the public for the parallel premium. The model builds on two sector, rehabilitation of the transport infrastructure basic relationships. The first is a flow equation in and support for agricultural production were identi- which the change in private holdings of unofficial fied as the most urgent priorities. The measures initi- foreign exchange, dF (the unofficial trade balance), is ated at the time of the 1986/87 budget and continued a function of the incentives for illegal trade, including thereafter include: (1) significant adjustments in the the parallel premium z, the official real exchange rate official exchange rate; (2) increases in interest rates, RER, and other variables w: resulting in positive real interest rates by 1988; (3) increases in producer prices for export crops; and (4) (6-1) dFt =f (zt, RERt, wt) a significant reduction in the number of price-con- + + trolled items. Both GDP per capita and trade volumes began to We provide a detailed rationale for an equation like rebound in 1986 after reaching their lowest points in (6-1) below. 1985. Real GDP growth averaged 4percentfrom 1986 The second equation is a portfolio equilibrium to 1989, with even higher growth evident in the condition in which the allocation of financial wealth extensive informal sector. The most visible source of between domestic assets M and unofficial foreign growth has been the agricultural sector, where over- exchange is a function of the uncovered interest par- all production increased between 4 and 5 percent in ity differential. Letting the notation tyt+1 denote the both 1987 and 1988, reflecting continued increases in expected value of yt+1 conditional on information production of foodgrains and some traditional ex- available at time t, the portfolio balance condition is port crops. The cornerstone of the Economic Recovery Pro- (6-2) Mt = g (i *t + tdInUt+i -it, xt)(Mt + utFt), g't < 0 gram has been the adjustment of the Tanzanian cur- rency. Although the devaluation of mid-1984 was where i and i are the foreign and domestic nominal substantial, it did not represent a fundamental interest rate, respectively, and x is a vector of other change in the Government's approach to exchange variables affecting portfolio behavior. The uncover- rate management. The parallel premium continued ed interest paritydifferential measures the difference its twenty-year rising trend, reaching over 700 per- in expected yields between dollar and TSh-denom- cent in early 1986, perhaps in anticipation of the inated assets (not including the expected penalty, if major devaluation and policy shift that accompanied any, associated with holding illegal foreign assets); a the 1985/86budget and agreementwith the IMF and rise in this differential lowers the desired share of World Bank. Since March, 1986, the premium has domestic bank deposits and other TSh-denominat- gradually fallen, reaching roughly 50 percent in the ed assets in the overall portfolio. Other influences on first half of 1990, a level not experienced since the relative yields, or on the relative liquidity or risk of early 1970s and briefly following the coffee boom. domestic and foreign assets, are captured by x. An While a mild premium (e.g., below 30 percent) canbe increase in penalties for violations of exchange or expected to persist reflecting capital controls, con- capital controls, for example, would simultaneous- vertibility of the exchange rate for current account ly reduce the expected yield and decrease the liq- transactions now appears to be a realistic option. uidity associated with illegal foreign exchange ho- ldings; at the same time, it might well increase the The Parallel Premium: Some Empirical Results riskiness of dollar assets. The overall effect would be to lower g for any value of the interest parity dif- The chronology presented above suggested that a ferential. variety of forces were at work in determining the Using the identity tlnzt+l = tlnUt+ . tlnEt+i where parallel premium in Tanzania. In this section, we Uand E are the unofficial and official exchange rates, present some empirical results and suggest direc- respectively, equation (6-2) yields the following dy- tions for further work on the determination of the namic equation for the parallel premium: parallel premium. One of the key questions emerging from the sec- (6-3) tdlnzt+l = h (Mt / Et, Ft, z t, xt) - (* t+dlntE t+1-i) ond section is the relative importance of trade and - + + 55 The Macroeconomics of the Unofficial Foreign Erchange Market in Tanzania Equations (6-1) and (6-3) form a second-order premium to the extent that they lower the share of dynamic system in which the parallel premium and domestic financial assets in private portfolios or de- the private stock of foreign exchange evolve together preciate the real exchange rate. In both cases, the in response to current and anticipated movements in results clearly indicate the need for complementary the real exchange rate, the domestic asset stock macroeconomic policies, since the effect of a nominal (measured in foreign exchange), the interest parity devaluation can be nullified by increases in nominal differential, and the other flow and stock determi- money or domestic prices. nants, w and x.13 With respect to the long-run effects of the portfol- It is apparent from equation (6-1) that for fixed io factors, the results are mixed. We cannot reect the values of the right-hand side variables, the model has null hypotheses that a4 and a5 are simultaneously ze- a steady state in which the parallel premium is a ro, using standard F-tests. In this sense, the resu- Its function only of the flow determinants RER and w support the prediction that portfolio factors influ- (simply set dFt = 0). On the other hand, the portfolio ence the parallel premium in the short run only, and determinants, by equation (6-3), clearly affect the that the premium is determined by flow factors alo- parallel premium in the short run. We therefore esti- ne in the long run.17 Taken separately, however, it ap- mate the following dynamic specification thatallows pears that while changes in real moneybalances have for separate short and long run effects of both groups no effect in the longrun (i.e., a5 is negligible), changes of determinants:14 in the interest parity deviation do have a cumulative effect over a two-year horizon. And when we drop (6-4) Inzt - a o + a,Inztil+ a2 d (MME)t + a3dIPDt the long-run portfolio effects from the regression + a4dRER t + as(84)t-i (columns 2 and 4), the overall performance deterio- + a6IPDEVt.l+ a7RER_1 rates noticeably. Caution is clearly appropriate in interpreting the results regarding dynamics. Table 6-1 gives the results of OLS and instrumental While the hypothesis of no serial correlation of the variables estimation of equation (6-4). The data are residuals cannot be rejected at the marginal5 percent given in table 6-2.15 We use the ex post interest parity significance level based on the Box-Pierce statistic, deviation as a proxy for the expected deviation, and both the Durbin-Watson (which is biased towards 2 apply instrumental variables to handle the implied given the lagged dependent variable) and the Box- measurement error.16 For the value of domestic as- Pierce statistic suggest that further work on the dy- sets, we use M2 (= Currency + Demand Deposits + namic specification and/or estimation with a serial Time Deposits). The real exchange rate is a trade- correlation correction maybe in order. weighted index of bilateral real exchange rates with the eight major trading partners. Extending the Analysis Given the short sample and the uncertain quality of the data, the basic results (columns 1 and 3) are The discussion above emphasized that in an ex- quite satisfactory. They give strong support to the change control regime, the parallel premium is conclusion that both trade and portfolio factors are jointly determined with the domestic price of im- at work in determining the premium on unofficial ports, since the marginal supply of imported goods foreign exchange in Tanzania. The short-run effects enters the country through smuggling and underin- of the various determinants, given by a2 - an, are all voicing. A rise in the official allocation of foreign of the expected signs: a rise in the interest parity exchange, for example, will increase the total supply deviation or an increase in the real value of domestic of imports, thus reducing the domestic price of im- financial assets leads to portfolio substitution tow- ports relative to the world prices and depreciating ards unofficial foreign exchange, raising the prem- the real exchange rate; simultaneously, it will reduce ium; a real appreciation shifts incentives away from the value of illegal imports, reducing the flow de- export smuggling and towards import smuggling, mand for foreign exchange in the parallel market and raising the premium. An appreciation of the real exc- lowering the parallel premium. hange rate raises the parallel premium in both the To incorporate these considerations, we first spec- short run and the long run, as predicted by the mod- ify flow equilibrium in the parallel foreign exchange el; moreover, since a4 and a7 > 0, an unanticipated market morecarefullyand then takecareof endogen- shock to the real exchange rate produces an 'over- eity of the real exchange rate. The flow demand for shooting' of the parallel premium in the short run. foreign exchange comes from two sources: (1) dire- The results also support the conclusion that nomi- ctly smuggled imports, M5, or imports through the nal devaluations are capable of lowering the parallel own-funds windcw, M°"n; and (2) imports brought 56 The Macroeconomics of the Unofficial Foreign Exchange Market in Tanzania in through the official window, but underinvoiced to change, the value of reported imports (and therefore avoid payment of tariff. In the underinvoicing case, the base of the import tariff) is fixed in advance, so we denote the officially reported value of imports that changes in tariffsaffectoverall profitsbutnot the and the amount by which these imports are underin- optimal degree of underinvoicing. voiced by V°ff and VU, respectively. In a strict foreign We include the tariff rate as an argument in the exchangebudgetingregime, V'ffisnotachoicevaria- underinvoicing function to capture the fact that ble, since it equals the official allocation of foreign ex- while the overall allocation of official foreign ex- change for imports. NoteVu need not be positive; im- hange may be determined in advance, individual ports maybe overinvoiced as a way of obtaining off- traders may view themselves as having some inf- icial foreign exchange for sale on the parallel market. luence over their own allocation of foreign exchange. To derive the flow demand for foreign exchange, For these traders, cost-minimizingbehavior requires consider first the case where there is no own-funds trading off the marginal benefit of an additional dol- scheme, so that the marginal supply of imports into lar of underinvoicing, which is tm, against the mar- theeconomyisthroughsmugglingandunderinvoic- ginal cost, which is z plus the marginal increase in ing. Suppose (1) that the marginal cost of smuggling expected penalties. Underinvoicing will occur on a unit of imports rises with the amount smuggled, those goods for which tm exceeds z; if tm is less than and is denominated in the good being smuggled (cf. z, the good will be overinvoiced. Similar behavior is Bhagwati and Hansen (1974)); and (2) that the cost of exhibited by individuals who illegally import either underinvoicing is denominated in foreign exchange a given quantity of goods (e.g., a single car or ma- and is an increasing function of the distance of v = chine tool) or goods withoutwell-developed markets V1A/Voff from zero (i.e., an increase in fraud in either within the country, such as specialized spare parts. direction raises marginal costs; cf. Macedo (1987)). Cost minimization yields an amount of smuggling Then letting WS and wu be variables like the govern- and underinvoicing that is an increasing funcfion of ment enforcement effort that increase the marginal the tariff rate and a decreasing function of the parallel costs of smuggling and underinvoicing, respectively, premium (and zero for tm = z). the flow demand for foreign exchange for illegal Finally, notice that the effect on FD of an increase imports takes the form.18 in officially recorded imports is uncertain, since it depends on the sign of v, i.e., on whether imports (6-5) FD = Pm Ms + Vu = p*Ms (q,z,tm,wIn) through the official window arebeingunderinvoiced + - + - or overinvoiced on average. If the average domestic + vu (qz,tm,wM)VOff price premium q isbelow the parallel premium, how- + - + -? ever, the overall incentive will be to overinvoice, so = FD ( q, z, tm,V°f, Pm, wm, wm), that v is negative; in this case a rise in Vff will de- + - + +? + - -? crease the flow demand for foreign exchange, and a rise in wmu will increase it. where q = (Pm-EP m)/EP m is the premium of the The flow supply of foreign exchange, FS, comes domestic price of imports over the world price at the from direct export smuggling in amount Xs and from official exchange rate, tm is the import tariff rate, and underinvoicing of exports officially reported to the z = (U-E)/E is the parallel premium. authorities, X". The amount of underinvoicing of For smugglers or for underinvoicers with fixed exports is given by xu = xu(t,x)[X(RER&ODA) - XS1, individual foreign exchange allocations, the optimal where X(RERx,ODA) is total export supply as a fu- illegal behavior embodied in equation (6-5) is nction of the real official exchange rate for exports straightforward: smuggle and/or underinvoice and the level of official development assistance. We those goods whose price on the domestic market is use the latter as a proxy for the demand for exports highenough to offset theparallel premium (i.e., those of housing services and tourism, two major channels goods for which q > z), and overinvoice the rest (i.e. of illegal exports in Tanzania.20 Both Xs and xu de- those for which q < z).19 pend positively on the parallel premium and on the Notice that the tariff rate does not affect the de- gap between the domestic price of exports and the mand for foreign exchange of these agents, since it world price at the official exchange rate, tx. Again does not affect their marginal incentives for illegal letting w8 and wu be variables like government en- activity once q and Vff are given. This is obvious for forcement efforts that increase the marginal cost of smugglers, who avoid tariffs altogether; for underin- smuggling and underinvoicing, respectively, we voicers who receive a fixed allocation of foreign ex- have: 57 TJhe Macroeconomics of the Unofficial Foreign Exchange Market in Tanzania (6-6) FS = P x (x$ (tx,a,wx) + x"(tx,z,wsU) {X( t,z,wu) margin; these prices are determined by overall ab- + + - + + - sorption, A, and by supply conditions. In the case of [X (RERx,ODA) - Xs (tx,z,wsv ) } nontradeds, an upward-sloping supply curve comes - + from standard general equilibrium considerations = FS (tx,z,RERxODA,Pxw,u, wx) (resources must be attracted from other sectors, in- ++ - + +- - cluding the export sector, to increase the supply of nontradeds); for the case of imports under exchange Notice that there are two sources of increases in tx, control, the marginal supply of imports comes from given the world price of exports: (1) decreases in smugglingandunderinvoicing, sothatqisafunction producer prices paid by marketingboards without a of tm and z as well as A.22 Substituting for q and P. in change in the exchange rate (e.g., through higher equation (6-9), we have overhead margins being charged by marketing para- statals, or through a policy decision to tax exports (6-10) {z = Zm (tm,tx,,MOftA,P * m,P* ,ODA,w,FS- FD) more heavily); (2) devaluations in the official ex- +? - -? + + - - + change rate that are not passed on to producers. Notice also that in Tanzania, the relevant "world Finally, we need models for A, Voff and FS-FD. In price" for key export commodities may in fact be the the Tanzanian context of foreign exchange rationing, producer price being paid in neighboring countries VIff is a policy variable. We can therefore either take (e.g., the producer price for arabica coffee in Kenya). it as exogenous or specify a "reaction function" in Taking equations (6-5) and (6-6) together gives us which the amount of foreign exchange allocated for the following version of equation (6-1) (we consoli- imports depends on other variables. Based on actual date the w's into a single vector for notational con- Tanzanian experience as outlined in the second sec- venience): tion, one possibility is to have VIff endogenously respond to the reserves position in the previous pe- (6-7) z = z (q,tm,txRERx7VODA,P*,P*xwwFS-FD) riod, and to current official exports and aid receipts: ++ -+ -? - + - + (6-11) v?ff =f (VA, (Rt_--Rt.2),Px (x-x5),ODA The real exchange rate for exports, RERx, is the + + + + ratio of the consumerprice index to theprice received by exporters (we use PPx, the producer price for where Rt-l is reserves at the end of period t-1.i In agricultural exports):21 this case, equation (6-11) becomes Da 1)1-a ~~~~~~~(6-12) (6-8) RERx = 1+Ea cc (Px/fp ) I-a Z = Z (tm,tx,Vt-l,Rt-l-Rt-2,A,Pm,Px,ODA,WwFS-FD) PPX +?-? -? + +- + (PMP,) = RERx (q, t, P,/EPtP1Z/P,), + + + + Equation (6-12), together with (6-13), provides a rich structure for incorporating flow determinants of Equation (6-7) can therefore be written the parallel premium and analyzing the linkages of the parallel foreign exchange market with the rest of (6-9) z = z( q,tmtx, Pn/EPx,Pm;, Px, ODA,V'%,w,S- FD) the economy. Intertemporal considerations, for ex- + +? + + - - -? + ample, enter through the determination of A: a com- modity boom that raises desired absorption will tend The effect of a rise in tx is uncertain a priori, since to raise the parallel premium by driving up the do- there are two opposing effects: the real exchange rate mestic relative price of imports and raising the prof- for exports rises, reducing aggregate export supply, itability of import smuggling; on the other hand, it while the share of exports that is diverted onto unof- will (i) increase the foreign exchange value of any ficial channels also rises. The net effect on illegal given volume of smuggled exports, and (ii) rapidly exports, and thus on z, is an empirical question. feed into higher allocations of foreign exchange Equation (6-9) contains four potentially endoge- through the official window, with the opposite effect nous variables: q, PpIEPx, VIff, and dF = FS-FD. Con- on the premium. The net effect on the parallel pre- sider the relative prices first. Both q and P,IEPx are mium over time is an empirical question, depending relative prices of goods that are nontraded at the largely on the degree to which the commodity boom 58 The Macroeconowics of the tlnofficial Foreign Exchange Market in Tanzania is perceived as temporary. In the case of the 1976-77 As in the case of a commodity boom, the net effect is coffee boom, which was clearly the effect of a tempo- an empirical question. rary supply shock (a frost affecting the Brazilian A major weakness of equation (6-12), in combina- crop), we would expect a very mild effect on aggre- tion with (6-3), is that it leaves out the government gate expenditure, and therefore, given the time path budget constraint. It therefore misses the endoge- of the domestic money stock and official exchange nous determination of the domestic moneystockand rate, a tendency for the parallel premium to be driven the official exchange rate. This is an important poten- down by the valuation effect on smuggling volumes tial direction for extensions. In the commodity boom and the endogenous trade liberalization. case, for example, the commodity revenues would be received by the private sector primarily in domestic Aid Inflows. With respect to an increase in Official currency, since the bulk of the export crop is mar- Development Assistance (ODA), holding desired ab- keted through official channels. In the absence of sorption constant, the effect should be to lower the sterilization, portfoliobalance considerations would premium, both by increasing the flow supply of for- then tend to push the unofficial exchange rate up in eign exchange onto the parallel market and by pro- the short run, counteracting the effect of the (actual ducing an endogenous trade liberalization. On the and anticipated) trade liberalization. other hand, increased aid should raise disposable income and therefore desired absorption, again de- Own funds. Extending the derivation of an equa- pending in part on how permanent the increase is tion like (6-12) to the case of an own-funds scheme expected to be; this would tend to raise the domestic is relatively straightforward. Since the costs of direct relative price of imports and the parallel premium. smuggling are positive, imports that were previously Table 6-3. OLS and instrumental variables estimation resultsfor equation (64) Dependent variablee Parallel premium (PPREM)1 OLS IV2 1967-1988 1968-1988 1 2 3 4 CONSTANT 191.38 -259.94 -195.95 -285.31 (-2.27) (-2.69) (-1.39) (-2.47) PPRTT.IEM 0.38 0.45 0.71 0.40 (1.49) (2.03) (1.41) (1.60) d(M2/ tE) 0.19 0.24 0.22 0.27 (2.58) (2.82) (1.55) (2.32) d(lPD,) 2.45 1.20 1.91 0.89 (2.16) (1.09) (0.76) (0.58) d(RE,R) 3.32 0.80 6.13 0.48 (2.24) (0.69) (2.03) (0.37) (M2/Et-1) -0.05 - -0.12 - (-1.40) - (-1.60) - IPDt-1 3.63 - 4.97 - (3.03) - (2.74) - REt,iR 2.46 3.08 2.56 3.17 (2.78) (3.09) (1.84) (4.26) RBA3R 0.86 0.79 0.80 0.78 (Q1l)3 -14.73 16.76 8.67 14.78 (0.20)4 (0.12) (0.56)4 (0.14)4 Note: (t-statistics are in parenthesis) a.The data are in table 6-2. b. Instruments for d(REER)t and d(IPD)t are M2.Z, REERt. ,and lPDt- 2 (along with the other right-hand side variables, which are assumed to be predetermined; note that in the case of PPREMtI, this is only valid if the disturbances are serially uncorrelated). c. Q is the Box-Pierce statistic for testing general serial correlation. For columns 3 and 4., the statistic reported is Q(10). 59 The Macroeconomics of the Unofficial Foreign Exchange Market in Tanzania own-funds scheme.24 Ignoring the cost of operating and V' (recall) is underinvoicing through the official in the (still illegal) foreign exchange market, and ass- window. The statutory tariff rate tm and the enforce- uming that expected penalties forunderinvoicingare ment variable vmu enter as determinants of te, and z an increasing function of the ratio of underinvoicing enters through the arbitrage relationship q - te = z. The to reported imports, there will be a perfectly elastic resulting reduced form can be written exactly as in supply of imports through the own-funds window (6-12), although the parameters will differ reflecting at the priceq- te = z, where teis the effective tariff paid the change in the supply function for imports fi- on own-funds imports.2 The amount of own-funds nanced by unofficial foreign exchange. imports will then be determined residually, as the One conclusion that emerges unambiguously in difference between total import demand M, and the the current model is that introduction of an own- amount of imports brought in through the official wi- funds scheme will raise the parallel premium consis- ndow: MaU = M - (fVA+VU) /Pm.We therefore have tent with any given value of the private current account surplus FD-FS, and therefore that it must (6-13) IFD = PmMown + Vu = Pm* M (A, Z, te, y) - voff raise the steady state parallel premium, ceteris pari- bus (since in the steady state FD-FS is fixed at zero). = FD (A, z, tm, wu, (V'ff/Pm), y), The reasoning is straightforward: by reducing the + - - - + costs of import smuggling, an own-funds scheme shifts out the supply of imports and drives down the where A is aggregate expenditure, y is a vector of gap between the domestic price of imports and their other variables influencingtheaggregate demand for international price at the official exchange rate (i.e., q imports (e.g., government absorption of imports), falls). Total imports are therefore higher under the Table 64. Data forparallelpremium regressions Period PPREM MZE$ IPDOP REER TAXCINV TOT AID 1966 20.40 190.54 n.a. 97.61 49 96.8 81.49 1967 23.20 215.56 1.96 99.14 50 91.9 70.29 1968 16.20 253.93 2.86 105.99 47 93.0 68.71 1969 27.40 277.39 6.26 104.91 71 112.3 117.32 1970 46.30 310.74 5.02 101.22 53 106.5 105.52 1971 110.00 364.41 3.08 97.09 54 98.8 136.10 1972 115.60 432.56 0.17 91.83 75 95.3 191.48 1973 94.93 529.42 7.13 88.03 39 118.8 170.00 1974 96.00 624.69 11.21 92.94 35 107.9 148.87 1975 202.53 671.95 16.93 99.60 35 92.8 332.93 1976 145.07 834.52 0.25 93.02 27 126.3 331.65 1977 89.07 1048.58 -5.44 92.94 23 140.3 393.36 1978 58.46 1267.25 11.72 93.11 40 115.1 404.55 1979 64.21 1679.35 6.26 87.03 34 114.9 535.91 1980 223.89 2141.34 10.37 100.00 52 100.0 546.61 1981 192.59 2486.09 25.55 131.93 52 85.2 496.25 1982 204.70 2584.86 29.04 158.73 57 88.4 405.17 1983 301.40 2338.33 43.44 171.82 52 91.1 318.69 1984 286.63 1668.01 16.39 177.94 53 96.4 310.80 1985 809.13 2387.37 103.98 206.61 50 90.5 424.92 1986 248.04 972.73 94.50 113.12 43 103.0 447.41 1987 126.95 793.62 42.36 174.52 50 89.6 704.92 1988 84.00 702.24 33.79 49.83 46 94.3 818.79 1989 57.89 n.a. n.a. n.a. n.a. n.a. n.a. Note: n.a. not available. PPREM=100*(U-E)/E is the end-of-the-year parallel premium in percentage points, with the unofficial and official exchange rates U and E taken from table 2-1. M2E$ is the end-of-the-year M2 in TShs (source: IMF, IFS), deflated by the official exchange rate. IDPOP=100*[(1 + i)(Et+i /Et)-(I + i)] is the uncovered interest parity differential, with i given by the London Eurodollar deposit rate (source: IFS) and i by the Saving deposit rate in Tanzania (source: Bank of Tanzania). REER is the ratio of the Tanzanian CPI to a trade-weighted average of WPIs of 8 major developed country partners (source: World Bank). 60 The Macroeconomics of the Unofficial Foreign Exchange Market in Tanzania own-funds scheme than previously, for a given sup- introduced, was an integrated policy package that ply of official foreign exchange. Smuggled exports included a devaluation of the official rate, decanotr must therefore rise to finance the higher imports if of some prices, and other policy actions. To the de- the private current account surplus is to remain un- gree that the package was perceived as a signal of changed. The only way this can occur is for the genuine policy reform, and thus of a prospective parallel premium to rise.26 increase in the return on domestic real assets, it would have tended to lower, rather than raise, the Implementing the Extended Model parallel premium. Finally, D83 is a dummy variable for the 1983 Table 6-3 shows the results of estimation of a ver- crackdown on "economic saboteurs"; its effect on the sion of the extended model. The parallel premium is premium is theoretically uncertain. From the illegal regressed on the portfolio determinants (imposing trade side, while a crackdown unambiguously re- the condition that these operate only in the short run) duces the volume of illegal trade, it may either uaise and on a subset of the flow detenninants in equation or lower the premium, since it simultaneously affecis (6-12). The data are in table 6-4. TAXCINV corre- both the supply and the demand for illegal foreign sponds to the inverse of tx in (6-12): it is the ratio of exchange. On theportfoliobalance side, a crackdcwn the domestic producer price of coffee to the fob ex- impairs the liquidity of foreign exchange assets and port price (converted to TShs).27 As discussed above, reduces their expected yield; both effects woukd tend the effect of export taxes on the premium is theoreti- to lower the parallel premium. cally ambiguous; a positive net effect would indicate Most of the variables have the expected sigs. Bth (i) that coffee growers are actively adjusting the share portfolio determinants enter significantly, with rnag of the crop sold through official channels in response nitudes generally close to those found in table 6-1. to taxes and the parallel premium, and (ii) that these Of the flow detenninants, only the TOT comes in st- adjustments are large enough to have macroeco- rongly, with the lagged TOT exerting a strong negat- nomic implications. ive effect on the premium (as observed, for exa.mple, TOT is the terms of trade; it enters (1) through the during and after the coffee boom). Lagged aid infio- endogenous trade liberalization that follows an im- ws also lower the premium, although theeffectis not provement in the balance of payments; (2) through a estimated precisely. The effect of lagged TOT and la- direct valuation effect on the illegal trade deficit; (3) ged aid is consistent with a substantial endogencus through resource movements in favor of exports and trade liberalization in response to balance Gf payttme- away from imports; and (4) througheffects on aggre- nts improvements; this corroborates evidertce L inn gate demand, depending on the savings response. import equations in Ndulu and Lipumba (19wo The first three of these would be expected to lower The coefficient on the coffee tax variable is consi 4a- the parallel premium; the third would raise it, to an ently negative but insignificant, implying tha; tIv extent depending on the savings response. Overall, smuggling response is more than offset by an aggreg- we expect a net negative effect. ate coffee supply response in the opposite direction. AID is net official resource transfers in dollars; a While this finding does not rule out a macecono- rise in AID should lower the premium both through mic role forcoffee smugglingin determiningthepar- direct increases in illegal export flows (expatriate allel premium (cf. Donnelly and Mshomba (19s,I housing, etc.) and through the endogenous trade who argue that up to 25 percent of the arabica coff@e liberalization effect; it should raise the premium to crop has been smuggled to Kenya in some yeans'., the degree that it raises aggregate demand. Again, we suggests that the elasticity of smuggling supplv, expect a negative effect on balance, although the low in the coffee sector, at least over the horc, Df aggregate demand effect might be rather strong year. given that changes in aid have a strong permanent The own-funds scheme appears to have raised v he component. premium, ceteris paribus, as predicted by the m oce. Own funds is a dummy variable for the period The magnitude of the increase, between 150 and 240 from 1984 to the end of the sample, during which the percentage points, is impressive, and suggests that own-funds scheme was in operation. By the argu- the low elasticity of smuggling response indicated in ments given above, we expect it to have a positive the coffee case may be a more general phenomenon. effect on the premium, given the values of the other Moreobviously, theresultsindicatethatthelowering variables. Interpretation of the estimated coefficient of the parallel premium since 1986 has been a func- will be complicated, however, by the fact that the tion of other developments in policy and extemal 1984/85 budget, in which the own-funds policy was conditions, such as (i) cumulative depreciations of 61 The Macroeconomics of the Unofficial Foreign Exchange Market in Tanzania the official exchange rate that reduced the real stock On the policy side, the key issue regarding the of domestic money, and (ii) large inflows of foreign unofficial foreign exchange market is unification. aid, and not of the own-funds scheme itself. The resource allocation gains of a single an market- Implementation of the extended model using an- determined exchange rate for commercial transac- nual data is clearly problematic given the short sam- tions are well known, and do not require elaboration ple. Nonetheless, while the results in table 6-3 should here. Holding aside the question of capital account be interpreted with caution, they suggest that there liberalization, what are the prospects for adoption of maybe significant payoff to furtherworkusing equa- an essentially market-determined official exchange tion (6-12). rate for commercial transactions? First, there is a fiscal impact of unification. Pinto Conclusions (1987) has emphasized that if the government is a net buyer of foreign exchange from the private sector, This paper has taken a macroeconomic perspec- unification will worsen the real deficit and raise the tive on the parallel foreign exchange market in Tan- inflation rate required to finance the fiscal deficit. Pr- zania. We discussed the emergence and behavior of eliminary calculations we have made suggest that a premium on foreign exchange over the period sin- Tanzania is on the favorable side of this calculation. ce 1966, and presented some preliminary empirical Given the large inflows of aid being channeled by the evidence on the determination of the premium. In Government, increases in the official exchange rate this conclusion, we briefly discuss selected empiric- provide a fiscal bonus, lowering the fiscal deficit in al and policy issues and the agenda for future work. domestic currency terms. Moreover, devaluation will Tlhe results of the third section show that both po- improve the trade tax base, both by moving activity rtfolio and trade factors are at work in determining from unofficial channels to official channels (a real the parallel premium in Tanzania. They therefore devaluation is required here), and by the valuation suggest that rises in the premium in the early 1980s effect of higher official exchange rate (since import may have been in part due to the substantial moneta- tariffs are levied on the domestic value at the official ry expansions that occurred starting in 1979. In the exchange rate). It therefore appears likely thatofficial context of the model, however, the resulting rise in exchange rate adjustments in the course of unifica- the premium should have raised the market-clearing tion will not contribute to inflation via the fiscal domestic prices of imports and import-substitutes. channel. But in this case, real moneybalances would not have Second, while unification through adoption of an risen as dramatically as they did in 1979, with veloc- across-the-board floating exchange rate has been ity falling by 40 percent (for both Ml and M2) and tried in other African countries, capital controls are remaining at the lower level until 1984 (at which time likely to be in place in Tanzania for the foreseeable it equally abruptly returned to trend). Surveys con- future. This means that a parallel rate will continue ducted by Bevan, et al (1990) and Ndulu and Hyuha to exist, and that some amount of trade activity will (1989) suggest that while official price indices did not take place at this rate. "Unification" therefore means fully reflect transactions prices during the third sub- the removal of rationing of import licenses at the period, the gaps are not large enough to explain the official rate, and the provision of a competitive offi- fall in measured velocity. One of the important issues cial outlet for export proceeds, rather than the reduc- for further investigation, therefore, is the impact of tion of the parallel premium to zero. Moreover, uni- failure of market clearing on the behavior of the fication in this sense is a complex process, involving parallel premium and its interaction with other mac- change in policy institutions and gradual adjustment roeconomic variables. on the part of market participants. In this context, the A second empirical issue has to do with handling parallel rate is likely to play an important allocative the change in regime that occurred between 1984 and and signalling function for a long time to come. 1986. For familiar reasons, a major change in policy Characterizing the nature of this role is an important regime should be expected to change the parameters part of our agenda. of behavioral equations and reduced forms. The As a final observation, one of the most interesting short sample, particularly after the regime change, aspects of the Tanzanian experience, and a potential has prevented us from investigating this issue care- lesson for other countries similarly situated, is the fully; one possible avenue for future work is to im- role of the own-funds scheme. The de facto disman- plement simple versions of the model using quar- tling of the QR-dominated trade regime through the terly data (which are available for prices, interest own-funds policy resulted in a significant inflow of rates, money stocks and exchange rates). consumer and intermediate imports, providing in- 62 The Macroeconomics of the Unofficial Foregn Exchange Market in Tanzania centives to farmers to increase production, and chan- for July, 1970 to March, 1986 and -0.041 for April, 1986 to neling scarce spare parts and transport equipment to December, 1989. industry and agriculture. The resulting supply re- 6. Starting in 1955, the Board was allowed to issue sponse was significant. This, coupled with the price currency against government securities. By the end of alignment to reflect scarcity values in virtually all 1%5, however, 82.5 percent of the currency issue was still commodities (brought about by the de facto trade backed by sterling. liberalization itself and the price decontrol), implied 7. The Tanzanian government decided not to follow the that the official devaluations initiated in 1986 did not devaluation of sterling in 1967. Arguments given then result in an acceleration of inflation, even though about the inefficacyof devaluationintheTanzaniancontext there was a marked increase in money supplygrowth became an established part of the policy canon until the during the 1987 to 1989 period. The success of the mid-1980s. own-funds scheme suggests that carefully identified 8. A second policy development in 1978 with possible policies linking the parallel market for foreign ex- implications for the balance of payments was the amend- change with the official economy can provide a sig- ment of the Bank of Tanzania Act, which abolished the nificant supply response and price alignment imme- previous limit of 25 percent of recurrent revenue on gov- diately preceding the adoption of politically contro- enment borrowing from the Central Bank (Ndulu and versial structural adjustment efforts. Hyuha 1989). It is not clear, however, that this represented a genuine loosening of fiscal constraints, since parastatals Notes had always been able to borrow from the National Bank of Commerce (itself a parastatal), and the National Bank of 1. The parallel foreign exchange market in Tanzania is Commerce from the Bank of Tanzania. an illegal or 'black' market. We use the terms 'unofficial' 9. The border was re-opened in 1984. The effect of the and "parallel interchangeably in this paper, although the border closure on the parallel premium is ambiguous, since latter term is broader and in some countries refers to a legal, it should raise the cost of both import smuggling (thus officially recognized market. reducing the demand for illegal foreign exchange) and 2. Similar schemes have been operated in Ghana and export smuggling (thus reducing the supply of foreign The Sudan in recent years. exchange). The theoretical model in the third section cap- 3. Ndulu and Hyuha (1989) give three reasons why the tures the effects of macroeconomic and regulatory changes share of licenses may underestimate the true share: (1) on the parallel premium. own-funds consignments under TShs 10,000 (approxi- 10. The National Price Commission existed from 1973 to mately USD 50) do not require licenses, (2) the utilization 1984. By the mid 1970s, it controlled some 2000 prices of rate of own-funds licenses is considerably higher than that imported and domestically produced goods. Prices were of licenses accompanied by official foreign exchange; and decontrolled starting in 1984; only 12 commodities re- (3) the incentive to underinvoice own-funds imports to mained under price control by July, 1988. avoid customs duties is much stronger than for officially 11. The fiscal developments largely accounted for the financed imports, since the cost of foreign exchange on the government's failure to meed performance criteria associ- underinvoiced portion is identical to that on the declared ated with IMF borrowings in the 1979-80. The absence of portion. IMF support meant an additional shock to external funds. 4. The World Currency Yearbook publishes monthly data 12. Ndulu (1987) gives a chronicle of official responses from July 1970 (with annual observations back to 1966). to the external crisis starting in 1979; there were three The series used in this paper is the World Currency Yearbook unsuccessful short-term government/IMF ventures in series up to January 1984, and the Maliyamkono and Ba- 1979-80, and then the unsuccessful government three-year gachwa (1990) series thereafter. While the two series move Structural Adjustment Program starting in June 1982. closely together for most of the period since 1984, an excep- 13. The system is saddle-path stable; in each period, the tion occurs in 1985, when the World Currency Yearbook parallel premium jumps to clear the asset market, and the shows a stabilization of the parallel rate in contrast to the resulting incentives for illegal trade determine the flow trend depreciation reported by Maliyamkono and addition to private foreign exchange holdings. Dornbusch Babachwa. We use the latter series based on our own et al (1986) and Rocha (1990) give a diagrammatic analysis. observations during that period, and consultation with 14. See also Rocha (1990) for an application to Algeria, other observers, who unanimously regarded the World and Fishelson (1988) for an application of the model to 19 Currency Yearbook data for 1985 as anomalous. developing countries. Fishelson uses the uncovered differ- 5. The trends are calculated by regressing the monthly ential at the unofficial rate, on the argument that parallel premium on a constant and a trend over the period movements in the unofficial rate provide a good proxy for in question. The coefficients on the trend term are 0.0091 expected movements in the official rate. Neither of these 63 The Macroeconomics of the Unofficial Foreign Exchange Market in Tanzania papers makes a distinction between the dynamic effects of tic price of exportables will equal the producer price. No- the portfolio deterninants and those of the flow determi- tice that the same is not true on the import side; there, the nants. relevant marginal price for imports is the unofficial ex- iS. The dependent variable actually used in the regres- change rate (plus the costs of smuggling), because sions is the parallel premium, defined as 100*(U-E)/E, exchange controls act like an import quota and imply that where U and E are the unofficial and official exchange rates, the marginal source of imports is illegal activity. respectively, rather than In z = In (U/E) as defined in the 21. The real exchange rate for exports is usually defined texf. as the domestic price of nontradables relative to exports. 16. Assuming market participants have rational expec- We use the CPI relative to the export price to capture the tations, the forecast error will be uncorrelated with resource puII away from exportables production more variables observed at time t or earlier. We therefore can use comprehensively. 3agged values of the right-hand side variables as instru- 22. The domestic price premium q falls towards tm, the nents for the interest parity deviation. We also instrument tariff rate, as the official foreign exchange allocation rises. for he change in the real exchange rate, on the grounds that For sufficiently high tariffs and liberal foreign exchange the irrent real exchange rate is jointly determined with allocations, the implicit quota on imports becomes non- :1he parallel premium. We treat the nominal money stock binding, and imports become a legally traded good on the and ` he official exchange rate as predetermined. margin. We then have q = tn,, and aggregate absorption is 17. This interpretation of equation (4) reflects a back- no longer a determinant of q. ward-uooking interpretation of the dynamic adjustment 23. One could also make Mff an increasing function of towards the steady state. In reality, the system formed by q and z, on the argument (suggested earlier) that official 'I' ard (3) has both a stable and an unstable root, so that foreign exchange allocations may respond to rent-seeking the rational expectations solution for z takes the form behavior. This does not affect the signs of the derivatives dtd2f = alzt - zbart], 0 < a < 1, where zbart is a function of in (10) provided that the effect of q on foreign exchange curTent and expected future levels of the flow determinants allocations is not large enough to offset the positive effect amd chlanges in the stock determinants of the parallel pre- of q on z through the incentive to smuggle. mium. Since future values of these variables must be 24. Provided, that is, that the authorities credibly com- lirected based on current information, however, the final mit not to scrutinize the source of funds. This does appear so uiton for z would include a distributed lag on current to have been the case in Tanzania. Informal sources sug- and past values of the determinants, as in equation (4). gest that surveillance on the own-funds window has been '-. The exact form of the smuggling function depends extremely loose in general. on the costs associated with the illegal activity. While our 25. Note that importsbrought in through the own-funds io7inulafion is fairly general, we have made one key sim- window are fully financed at the parallel exchange rate, plification in assuming that the costs of smuggling are regardless of the degree of underinvoicing chosen by the ultinately denominated in foreign exchange: this implies importer. The incentive to underinvoice is therefore that only the parallel premium, and not the level of the stronger for own-funds importers than for importers using official exchange rate, will directly affect smuggling incen- the official window, since in the latter case an increase in ti -es. underinvoicing raises the trader's costs by increasing the s c. Smuggling profits are equal to Rs = PmMs - U[M5 + share of the import that must be financed at the parallel C 'fvl), where C(Ms) is the smuggling cost function. The rate. The effective tariff in the own-funds case will there- first-order condition is C'(Ms) = (q-z)/(1+z), yielding a fore be below that on officially financed imports, ceteris smuggling function Ms(q,z) with the properties given. Un- paribus. In the Tanzanian case, this difference is exacer- der(or over)invoicing profits are given by Ru = batedbythegenerallaxityofsurveillanceofgoodscoming Pm:[(XA°F+VU)/Pm*J - E(1+t)Voff - U(VU+c(v)Voff) = {(q-t) - in through the own-funds window. <-zAv.* (1+z)c(v)}EWf (we have assumed that the invoic- 26. Notice the importance of our assumption that the rin cost function is homogeneous of degree one in Vu and costs of import smuggling are independent of the costs of Va 3,. Given t.ff, the first-order condition for v is similar export smuggling. If import and export smuggling were -o the smuggling case: c'(v) = (q-z)/(1+z). This yields the joint activities, then an own-funds scheme might not drive invoicing function Vu = VU(q,z)Vff given in equation (5). out all import smuggling. The assumption that smuggling 20. We use the real official exchange rate as the relative costs are private costs is also important; if the cost of price influencing export supply decisions under the as- smuggling were a loss of some portion of the good being surnption that peasants continue to voluntarily sell some smuggled, then theown-funds scheme would represent an portion of their export crop to the marketing authorities improvement in the economy's marginal terms of trade rather than selling it on the parallel market. In this case, and therefore in the productivity of a unit of illegal exports aibitrage between the two markets implies that the domes- in generating imports, In this case, while an own-funds 64 The Macroeconomics of the Unofficial ForFign Exchange Market in Tanzania scheme would raise total imports, it is an empirical ques- Correction Approach to Money Demand: The Case of the tion whether this would require an increase in the flow of Sudan', Journal of Development Economics. illegal exports, and thus in the premium (we are grateful to Donnelly, Lawrence P. and Richard Mshomba (1987), *The K. Krumm for pointing this out). Smuggling of Coffee: The Case of Kenya and Tanzania", 27. We use an average ratio for arabica and rubusta, mimeo, presented at November 1987 annual meeting of calculated by taking the ratio of payments to producers for the African Studies Association. the two types of coffee to the total fob export value of the Edwards, Sebastian (1990), Real Exchange Rates, Deval- two types. For data availability reasons, we use the ad- uation and Adjustment: Exchange Rate Policy in Devel- vance price for coffee. oping Countries (Cambridge, Massachusetts, MIT References Press). Azam, Jean-Paul, and Tim Besley (1988) 'General Equilib- Fishelson, Gideon (1988), "The Black Market for Foreign rium With Parallel Markets for Goods and Foreign Ex- Exchange: An International Comparison', Economics change: Theory and Application to Ghana", Princeton Letters 27: 67-71. University Research Program in Development Studies, Green, Reginald H., Rwegasira and Brian van Arkadie Discussion Paper #143, December. (1980): Economic Shocks and National Policy Making Bagachwa, M.S.D., N.E. Luvanga, and G.D. Mjema (1989), (The Hague, Institute of Social Studies). "Tanzania: A Study on Non-Traditional Exports (Prelimi- Greenwood, Jeremy and Kent P. Kimbrough (1987), 'For- nary Draft)," mimeo, University of Dar es Salaam, Sep- eign Exchange Controls in a Black Market Economy', tember. Journal of Development Economics 26:129-143. Bevan, David, Arne Birgsten, Paul Collier and Jan Willem Krumm, Kathie (1987), 'Medium-Term Framework for Gunning (1987), "East African Lessons on Economic Lib- Analysis of the Real Exchange Rate: Application to the eralization" Thames Essays no. 48, Trade Policy Research Philippines and Tanzania', CPD Discussion Paper No. Centre, London. 1987-9, The World Bank, May. Bevan, D., P. Collier and Jan Gunning (1990) Peasants and Macedo, Jorge Braga de (1987): "Currency Inconvertibility, Governments (Oxford, Oxford University Press). Trade Taxes and Smuggling", Journal of Development Bevan, D., P. Collier and P. Horsnell (1988), "Supply Re- Economics 27: 109-125. sponse Under Goods Market Rationing in Tanzanian Maliyamkono, T. and M.S.D. Bagachwa (1990): The Second Peasant Agriculture,' paper presented at the OECD. Economy inTanzania (Oxford, Oxford University Press). Bhagwati, JagdishN. and Bent Hansen (1973), "A Theoreti- Ndulu, Benno J. and Nguyuru H.l. Lipumba (n.d., apprax. caB Analysis of Smuggling," Quarterly Joumal of Eco- 1988): "Intemational Tradeand Economic Development nomics, May, reprinted in Bhagwati, ed., Illegal Transac- in Tanzania", mimeo, Department of Economics, Univer- tions in International Trade (Amsterdam, North sityof Dares Salaam. Holland, 1974). Ndulu, Benno and Mukwanason Hyuha (1989): "Inflation Biersteker, Thomas J. (1986), "Self-Reliance in Theory and and Economic Recovery in Tanzania: Some Empirical Practice in Tanzanian Trade Relations," in John Raven- Evidence', Uchumi 2, No.1, University of Dar es Salaam. hill, ed., Africa in Economic Crisis (New York, Columbia Newlyn, W.T. (1967): Money in an African Context (Loan- University). don, Oxford University Press). Cliffe, Lionel and John S. Saul (1972), Socialism in Tanzania: O'Connell, Stephen A. (1990): "Short and Long-run Effects An Interdisciplinary Reader, Vol. 1: Politics (Nairobi, of an Own-Funds Scheme", mimeo, Department of Eco- East African Publishing House). nomics, University of Dar es Salaam, March. Cliffe, Lionel and John S. Saul (1973), Socialism in Tanzania: Pinto, Brian (1988): "Black Markets for Foreign Exchange, An Interdisciplinary Reader, Vol. 2: Policies (Nairobi, Real Exchange Rates and Inflation: Overnight vs. Grad- East African Publishing House). ual Reform in Sub-Saharan Africaw, mimeo, World Bank Collier, Paul (1988), "Aid and Economic Performance in Rocha, Roberto De Rezende (1990): "The Black Market Tanzania," mimeo. Premium in Algeria", mimeo, World Bank. Collier, Paul and Jan Willem Gunning (1989), "The Tanza- Tanzanian Economic Trends, Vol. 1, No. 4, January, 1989 nia Recovery, 1983-1989", mimeo, November. (Economic Research Bureau, University of Dar es Sa- Domowitz, Ian and Ibrahim Elbadawi (1987), "An Error- laam). 65 7 Niger and the Naira: Some Monetary Consequences of Cross-Border Trade With Nigeria Jean-PaulAzam Nigeria is by far the largest economy of West and which are traded at this exchange rate. In many Central Africa; its macroeconomic behavior affects African countries where official imports are re- all the countries in the region. Most of its neighbors stricted, most economic agents are affected at the belong to the CFA Franc Zone and have a convertible margin by these prices, whereas official prices play currency, both internally and externally. External only an inframarginal role (Azam and Besley 1989). convertibility is guaranteed by France, which pro- By contrast, when the currency is convertible, the vides automatic balance of payments support to the parallel market for foreign exchange needs not clear, two monetary unions, the Union Mon6taire Ouest even if the official exchange rate is flexible, as any Africaine (UMOA) in West Africa and the Banque des surplus or deficit can be exchanged on the official Etats de l'Afrique Centrale (BEAC) in Central Africa. market. Here we neglect the "laundering" costs in- On the contrary, the naira, which is Nigeria's cur- curred when the money has been acquired by selling rency, is inconvertible. illegal goods. Hence, convertibility, by unifying the This fact, together with various other trade distor- foreign exchange market, dampens exchange rate tions, explainswhyaparallelmarket for the naira has fluctuations by pooling the shocks affecting each developed. Arnoult (1983), Azam (1990a), and segment of the market when the exchange rate is Gregoire (1986) describe some of the distortions re- flexible. garding trade between Niger and Nigeria. These in- In the case of the members of the CFA Franc Zone, clude different tax rates and the outright ban on the exchange rate is not flexible, and each country imports of some goods from some countries. As most can run an external imbalance, at least temporarily. trade distortions are impossible to enforce tightly in As a result of internal convertibility, any imbalance Africa, there are various illegal and semi-legal trans- in the parallel balance of payments will eventually actions involving goods and foreign exchange. show up in the official balance of payments, as agents Convertibility makes a difference in how the econ- buy (sell) the net flow of foreign exchange which they omy works when there are active parallel markets. need (have) from (to) the formal banking system. When a currency is convertible, its parallel market The contrast between convertible and inconvert- has to clear at each point in time through price flexi- ible exchange regimes has interesfing implications bility, i.e. by exchange rate adjustment. Azam and for the analysis of trade between a small CFA Franc Besley (1989) analyze the parallel market for the cedi, Zone member like Niger and its large neighbor, Ni- the Ghanaian currency. All the shocks affecting the geria, whose currency is inconvertible. Niger's econ- parallel foreign exchange market are reflected in the omy has probably run a sizeable surplus in its paral- exchange rate and hence in the prices of the goods lel balance of payments with Nigeria since 1983. We 66 Niger and the Naira: Some Monetary Conseqeces of Cross-Border Trade With Nigeria try in the following to offer some explanatory factors The number of observations is noted N and R2 is for this performance. the usual coefficient of determination. DW is the We analyze the currency flow below to provide for classic Durbin and Watson test of residuals autocor- Niger an indicator of the balance of payments on the relation and the numbers in parentheses below the parallel market. We show in the second section how estimated coefficients are the usual t-ratios. The de- Niger's economy can acquire convertible foreign ex- pendent variable loge is the log of the parallel ex- change by trading with Nigeria, despite the latter change rate of the naira in Zinder, Niger, expressed having an inconvertible currency. Although the price in terms of nairas per CFA franc. M is the quantity of of the naira can be regarded as exogenous for Niger, local currency held outside the banking sector in as it does not adjust to clear the foreign exchange Nigeria, as a proxy forcashbalances; q is the Nigerian market, it is not constant, and its fluctuations have producer price of cocoa; pw is the consumer price significant macroeconomic effects. The third section index for industrialized countries; Q is the official focuses on its impact on the consumer price index, imports into Nigeria, in real terms, as a proxy for and shows that Niger can be regarded as a small price import quotas; and qW is the world price of cocoa on taker. The fourth section shows how the price of the the London market. The value of the R2 shows that naira affects the demand for cash balances in Niger. nothing much is left to Niger for influencing this It is suggested that the increase in cash balances exchange rate. which occurred during a large part of the 1980s, The countries which participate in the naira mar- following the sustained depreciation of the naia on ket can run a surplus or a deficit in their parallel the parallel market, played a favorable role in the balance of payments with Nigeria, provided the rest improvement of Niger's external balance. The fifth of the market participants have a net balance of the section provides additional evidence in favor of an same magnitude with the opposite sign. Figure 7-1 asset approach to at least partly explaining Niger's illustrates this reasoning. On the left hand panel, the parallelbalance of payments surplus. The last section demand (NNd) for and supply (NN5) of nairas in Niger provides a conclusion. are pictured. The former is the sum of imports from Nigeria, investments by residents from Niger in Ni- Eanming Convertible Foreign Exchange on the geria, and increases in naira balances held by Ni- Naira Market gerien citizens. Its negative slope reflects the fact that an increase in the price of the naira reduces the As the naira is inconvertible, itsparallel market has competitiveness of Nigerian products and reduces to clear by price adjustment; no central bank inter- the expected capital gains on assets denominated in venes to peg the exchange rate. Hence, from the naira, for a given expected path of the exchange rate viewpoint of Nigeria, this market is in equilibrium. in the future. The supplyof nairas in Nigeris thesum All of Nigeria's neighbors belong to the CFA Franc of exports to Nigeria, investment in Niger by Nigeri- Zone, The CFA franc is pegged to the French franc; it ans, and reductions in naira balances held. The posi- is thus linked, via the latter, to all the currencies of tive slope is due to the increased profitability of the European Monetary System and hence to all the export to Nigeria when the price of the naira in- main currencies of the world. Therefore the CFA creases and, for given expectations, the reduced at- franc price of the naira is determined by Nigeria and tractiveness of naira-denominated assets. by the world economy. No other African economy on The right-hand panel of figure 7-1 represents the its own has a significant influence on this market. corresponding naira demand and supply curves for This is illustrated by the following equation, which the rest of the world, denoted byNRd and NR', respec- Azam and Vernhes (1990) have estimated on quar- tively. Theirslopesaredeterminedbythesameargu- terly data for the period from 1982.1 to 19872, using ments as those presented above. Units have been an extension of the model analyzed by Azam and chosen in such a way that the exchange rate between Besley (1989): Niger and the rest of the market is equal to one. If the naira market is unified by the arbitrage operations of (7-1) loge = -41.0 + 0.97 logM + 123 logq exchange brokers, the exchange rate clears the mar- (21.4) (4.93) (1.64) ket in such a way that any imbalance of the Niger + 5.79 logpW 0.06 logQ - 0.34 logqw segment of the market is offset by a net imbalance of (5.78) (1.38) (4.04) the same size and the reverse sign in the rest of the market. Evidence reported in Azam (1990a) supports N = 22, R2 = 0.99, DW = 1.70 this assumption of arbitrage. Comparing the CFA Franc price of the naira on the parallel market in 67 Niger and the Naira. Some Monetary Consequences of Cross-Border Trade With Nigeria Lome, Togo, and Zinder, Niger, using 86 monthly Figure 7-1. The Naira Market in Niger and in the Rest data points between 1980.08 and 1987.12 shows a of the World striking arbitrage performance. Excluding a few out- liers during the six months following the closure of e theborderby the Nigerian government in April 1984, NRs the mean premium on the Lome market is not signifi- cantly different from zero. Its value is -0.012 percent, and it has a standard deviation of 2.72 percent. Only Ns seven data points, besides the six post-April 1984 N outliers, show a premium larger than plus or minus 5 percent, with a maximum of 7.3 percent in Novem- S / ber 1984. Rd In the case pictured in figure 7-1, there is an excess N supply of nairas in Niger, which is offset by excess demand in the rest of the market. If arbitrage did not Niger Rest function properly, the price of the naira would go down in Niger and up in the rest of the market. Assuming perfect competition in the exchange bro- Assuing erfct cmpeitio inthe xchnge ro- Figure 7-2. Flow of BEAC Banknotes into Niger, 1980-89 kers business, any premium appearing in the rest of the world segment of the market triggers a flow of (100=1982) nairas away from the Niger market, until excess de- 600 mand is cut down to zero everywhere in the market. If we neglect transactions costs and the profits of the brokers, which vanish in a free entry equilibrium, 5 this flow of nairas must be compensated by a flow of currency from the rest of the market in the reverse 400 - - direction. The market for the naira clears globally. It serves as an indirect channel whereby Niger is con- I nected through arbitrage with the other economies 300 participating in this market. When Niger's parallel balance of payments shows a surplus, Niger can 200 / acquire foreign exchange from the international naira market. In the case of Niger in the 1980s, one can find an 100 indicator of such an inflow of currency via the naira market by looking at the series of banknotes from the o- BEAC monetary union purchased by the Banque 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 Centrale des Etats de l'Afrique de l'Ouest (BCEAO) Year in Niger. These CFA francs from Central Africa can be exchanged one for one against the CFA francs Source: BCEAO. from West Africa, without any transactions costs. However, they are not perfect substitutes for the mainly involve banknotes, as checking accounts are BCEAO banknotes because they are not legal tender. not widely used in this milieu, especially when deal- Theyare not generallyaccepted in the shops in Niger. ing with illegal transactions. The above analysis of the naira market may be The resulting net inflow of CFA francs from the simplified by assuming that the rest of the market BEAC into Niger would eventually be used to ac- comprises only Cameroon. Cameroon has an active quire BCEAO banknotes from the Central Bank and border trade with Nigeria and, through it, the other would then be recorded, thus contribufing to the member states of the BEAC zone (see Azam 1990a). officialbalance of payments (see Azam 1990b). These The case of figure 7-1, in which Niger runs a surplus banknotes would be acquired by the BCEAO in its and the rest of the market a deficit, would give rise three agencies in Niamey, the capital city, and in to a flow of nairas from Niger to Cameroon; BEAC Maradi and Zinder, the two main trading towns in CFA francs would flow in the reverse direction. It is the Hausa country north of the Nigerian border. On reasonable to assume that these flows of currency average the latter two towns account for more than 68 Niger and the Naira. Some Monetary Consequences of Cross.Border Trade With Nigeria 85 percent of this flow. This strengthens the interpre- (7-2) C(F,G) tation of this series as an indicator of the balance of cross-border trade between Niger and Nigeria, as whereFdenotes theamountof official transactions Maradi and Zinder are known as the focal points of and G denotes the amount of smuggling, and with this trade (Arnoult 1983, Gregoire 1986). Neverthe- derivatives CF < 0 and CG > 0. The negative impact of less, this indicator is somewhat crude and probably official transactions on the costs of smuggling is an involves quite large observation errors (see Azam assumption due to Pitt (1981) and discussed by 1990a for further discussion). Devarajan, Jones and Roemer (1989). It captures the The inflow of BEAC banknotes into Niger during "cover effect" whereby official transactions can cloak the 1980s is pictured in figure 7-2. There is a down- illegal ones. ward drift over most of the period and a sharp rise in The local price, p, and the border price, p*, of the 1983. The jump in 1983 corresponds probably to a imported good are expressed in local cur'ency, and t switch from a deficit position in the parallel balance denotes the rate of ad valorem tax on official transac- of payments to a surplus. Additional evidence pre- tions. Then the profit function of the smuggling firm sented in Azam (1990a) supports this explanation. is defined as: The low volume inflow of BEAC banknotes before 1983 reflects probably a normal level of trade with (7-3)R(p,p*,t) = max ((p-p*)G + (p-p*(l +t))F - C(F,G)). Chad or Cameroon, while its fivefold increase in 1983 is large enough to indicate a drastic change in the This definition entails the following arbitrage flow of payments. equations: It is reasonable to assume that the adjustment pro- gram launched in 1982 by the Niger govemment, (7-4) 0< p - p* = CG = p*t + CF < p*t. with the support of the International Monetary Fund and later the World Bank, played a large part in this This entails that official transactions are made at a external improvement (see Guillaumont et al 1988). loss, if CF is strictly negative, with the local price Real public expenditures were cut by more than 18 below the tax inclusive border price. The rationale is percent in 1982 and 1983. Moreover, there was a that official transactions have a positive effect on the drought during 1983 to 1985, when peasants and profitability of illegal transactions. The illegal profit herdsmen were selling their cattle. But it is not obvi- pays for the loss incurred by selling below the tax ous why such an improved position carried on until inclusive cost of the legally imported good. the end of the decade, while public expenditure Assumingthat free entryprevailsin the smuggling started to grow again. This is explained in part below, sector, then by analysis based on assets holding behavior. But first the relationship between Niger and Nigeria is (7-5) R(p,p*,t) = 0. analyzed in terms of the extent of cross-border mar- ket integration. This analysis sheds some light on the Hence, in equilibrium, we obtain a kind of pur- way the price level is determined in Niger. chasing power parity relation by inverting equation (7-5). This implies that the domestic price level p is Cross-Border Market Integration determined by the border price and the tariff rate, and does not depend on domestic monetary policy, The border between Niger and Nigeria does not etc. By totally differentiating profit function (7-5), the divide the Hausa country into two separate markets; following type of purchasing power parity relation- there is an integrated market across this border. The ship is obtained: price level in Niger is determined by theborder price and the tariff rate, and does not depend on monetary (7-6) d logp = (1 + at) (p*/p) d logp* + a(p*/p) dt, policy or aggregate demand management in general. where 0< a = FI(G+F) This is consistent with the theory of smuggling costs recently developed by Pitt (1981) and Devarajan, This typeof argument, whichwe have restricted to Jones, and Roemer (1989). It is assumed that private the case of illegal imports for the sake of simplicity, firms are engaged both in official trade and smug- could be reproduced for illegal exports, mutatis mu- gling. To keep things simple, let us concentrate on tandis. Moreover, one could assume the presence of one side of the trading business ordy, namely im- economies of scale in smuggling, so that illegal ex- ports. We assume that smuggling entails costs de- ports would entail a reduced marginal cost for illegal scribed by the function: imports (see Azam and BesIey 1989). Hence, provided 69 Niger and the Naira. Some Monetary Consequewecs of Cross-Border Trade With Nigeria the smuggling sector is reasonably competitive, with der trade do not make much use of bank accounts. approximately free entry, the prices of traded goods And second, during the period under study, the rules are determined by the border prices, irrespective of conceming the bank accounts of the mining firms monetary and budgetary policy, even if official im- were changed, with a significant impact on the quan- ports are restricted. Such an assumption seems rea- tity of money series. sonableforthecase of Niger, where nontraded goods We included in the equation several current and do not seem to play a significant part in the African lagged values of the monetary policy indicator, sepa- Consumer Price Index (CPI). (The European Con- rately and jointly. The most nearly significant impact sumer Price Index is not considered here.) appears in the following equation: The following econometric exercise shows this re- sult quite clearly, using monthly data from 1982.91 to (7-8) logp = 1.04 + 0.46 logp(-1) + 0.26 logp(-12) 1989.10: (1.77) (5.57) (4.14) + 0.07 logep - 0.03 logM(-12) (7-7) logp = 0.67 + 0.47 logp (-1) + 0.28 log p (-12) (4.36) (0.74) (2.29) (5.95) (4.95) - 0.24 AR (1) + 0.08 logep* - 025 AR (12) (2.14) (5.99) (2.40) N = 94, R2a = 0.75, DW = 1.90, F = 56.52 N = 94, R2 = 0.75, DW = 1.92, F = 70.8 Here M(-12) denotes the nominal quantity of cur- Here p is the African CPI in Niamey (Niger), e is rency, lagged one year. It is not significant and has the parallel market exchange rate in Zinder and p* is the wrong sign. Moreover, inclusion of this addi- the rural CPI in Nigeria. The one month and twelve tional variable in the equation has not modified no- months lagged values of p are denoted by p(-1) and ticeably the estimated coefficient of logep*. This con- p(-12), respectively. AR(12) denotes residuals auto- firms the intuitive result that e and p* are exogenous regression of the 12th order, estimated by the with respect to the macroeconomic policies of Niger. Cochrane-Orcutt method. The dynamic specification It also confirms the result reported in Azam (1990a), of equation (7-7) was selected by a simplification which is based on a shorter sample with quarterly procedure going from the general to the simple data. Niger is too small to have a significant impact (McAleer, Pagan, and Volker 1985). The numbers in on the market for the naira, which is presumably parentheses below the estimated coefficients are the dominated by the relationship between Nigeria and standard t-ratios. N is the number of data points the entire CFA Franc Zone. In other words, there is used, R2a is the usual coefficient of determination probably a causal relation from ep* to p. Figure 7-3 adjusted for degrees of freedom, DW is the classic illustrates the price deflation which occurred in Ni- Durbin-Watson test of residuals autocorrelation, and ger over the second half of the 1980s as a consequence F is the usual test of the estimated equation against of the naira having depreciated on the parallel mar- the assumption of no linear relationship at all. ket faster than the CPI increased in Nigeria. The fit is reasonably good for monthly price data. Therefore, it seems acceptable to regard Niger as a The price index of Nigerian goods, ep*, which is small open economy which is price taker in all its computed using the parallel market exchange rate, is markets. Nontraded goods do not seem to play a highly significant. Adding other lagged values of significant role in this economy. This entails as a logp and other AR terms does not reduce its signifi- corollary that the surplus of the parallel balance of cance. But the purchasing power parity-type hy- payments which we have presented in the previous pothesis implies not only that prices in Nigeria affect section cannotbe explained by applyingthe "elastici- the CPI in Niger, it also implies that the CPI in Niger ties approach". Competitiveness is not an explana- is not affected by the macroeconomic policies pur- tory factor if prices are determined abroad. More- sued in Niger. over, figure 7-3 illustrates that the CPI in Niger has To test this second part of the assumption, we have increased relative to the price index of Nigerian added to equation (7-7) an indicator of monetary goods. Therefore Niger has experienced, if anything, policy, namely 'currency held outside the banking a loss of competitiveness rather than an improve- sector" (circulation fiduciaire). This indicator is pref- ment. The divergence between the two series in fig- erable to a broader measure of the quantity of money ure 7-3 can be explained by the fact that Niger's CPI for two reasons. First, one can argue that the repre- is based on prices and consumption patterns in the sentative agents of the sectors involved in cross-bor- capital, whereas the rural CPI is used for Nigeria. 70 Niger and the Naira. Some Monetary Cosequences of Cross-Border Trade With Nigera Figure 7-3. Price Indexes in Niger and Nigeria Hence, another approach is required. In the devel- (1970=100) oped countries' literature, the elasticities approach 600 -_has been replaced by the asset approach. The latter seems capable of being adapted to the cases of less , CPIin Niger developed countries. However, one must take into 550 account some specific aspects of these countries, es- pecially in Africa. The next section is devoted to the soo - * ¢\ t\ I~ \J \ 1 development of an asset approach adapted to the structural features of the economy of Niger. An Adapted Asset Approach 400 - In the model presented in this section we attempt to capture as simply as possible a specific feature of 350 - Prices of Nigerian goods '', theNiger economywhich mustbe taken into account for adapting the asset approach to the balance of pay- ments in Niger. Cross-border trade with Nigeria in- 300 l l l v olves to a large extent storable goods, like cattle and 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 grains, which we refer to as African goods. These Afr- Year 0 ican goods can be held as assets; granaries and herds Source: BCEAO, Mnistry of Planning and World Bank are the bulk of the portfolios of most African people. Therefore, cross-border trade may be regarded as a channel whereby agents can swap physical assets Figure 7-4. Index of Herd Size and Cash Balances, like cattle against cash balances and vice versa. 1980-89 Niger also trades other goods with the rest of the world by official channels. We refer to these goods as (1982-100) European goods. Hence, we analyze a simple model 150 in which there are two goods. E denotes the quantity of European goods consumed and is meant to cap- . - hiture the flow of official imports, and A denotes the 125 Real cash balances - ' quantity of African goods consumed. African goods are assumed to be storable; their total inventory level . , is denoted by H. We assume this asset to be produc- 100 . tive, like cattle, so that the output of African goods is equal to F(H), F'O, F" S denotes the amount of African goods sold on the 75 \ market, which can be negative in the case of an acquisition. The accumulation equation for this asset can be written: Cattle herd size (7-9) dH/dt = F(H) -A - S. The alternative asset is real cash balances. As we 198 1981 1982 1983 1984 1985 1986 1987 1988 1989 have seen above, the market for the naira can be 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 regarded as an indirect channel for acquiring con- vertible currency. M denotes the nominal quantity of Source: BCEAO, anld Ministry of Agriculture, money held, and pE and pA are the money prices of the European goods and African goods, respectively. The following equation: Probably the former is affected by the price of goods which are not traded with Nigeria, but with Europe (7-10) p = p(pEpA), and the rest of the world. The prices of European goods have not fallen as have those of the goods defines the consumer price index p(.), which is in- traded with Nigeria. creasing, differentiable, and homogeneous of degree 71 Niger and the Naira: Some Monetary Consequences of Cross-Border Trade With Nigeria one. Accumulation of real cash balances (Mlp) is ginal product of the stock of African goods plus the governed by the following equation: expected capital gains entailed by an increase in their real price. The latter is made of the subjective value (7-11) d(M/p)/dt = (PAS PEE - (d logp/dt)M)/p. of the services provided by the cash holding, less the expected capital loss due to inflation. We assume that the representative agent is an un- Lastly, we get the standard Keynes-Ramsey for- constrained price taker in a large market, and so we mula, which equates the psychological rate of inter- treatHandM/pas "jump variables', i.e. as right-hand est to the real rate of interest. Here, it reads: differentiable functions of time. Equations (7-9) and (7-11) thus refer to the right-hand derivatives with (7-16) r - d logUAgdt = F'(H). respect to time for these two variables. Given these two differential equations and neglecting population Using again the homogeneity assumption, and (7- growth, we assume that the representative agent 14), we can invert (7-16) to get the optimal inventory seeks to maximize: of African goods as: 00 (7-12) f (U (E,A)+W (AVp) )e-rdt, (7-17) H' (d log(pA//pE)Idt) o = F' (r-h d log(pA/pE)/dt), H"' O, in which the utility function U(.) is twice differenti- able, increasing, concave and homogeneous of de- where h = -s(AUAdIWA) 0. This result is again int- gree one, and the wealth function W(.) is increasing uitive, and describes a simple speculative behavior twice differentiable and strictly concave. Including (assuming perfect foresight): an increase in the ex- real cash balances in the objective function, with pected rate of growth of the relative price of Afric- additive separability, is an assumption which dates an goods leads to an increase in the optimal inv- back to Sidrausky (1%7) (See Blanchard and Fischer, entory level (herd size, for example). The mechan- 1989, for a discussion of the Sidrausky model). The ism involved here isnotassetsubstitution. It isbased constant rate of time preference is denoted r. on the relation between saving and the return on Applying Pontryagin's Maximum Principle to this assets, as the argument of the function F1 (.) measur- problem and rearranging terms, we get three mar- es the real rate of interest. It can alternatively be ginal conditions. First, optimization of the consump- expressed as the positive effect of an expected im- tion bundle implies: provement of the terms of trade on investment in productive assets. (7-13) UA/UE = /A1PE Similarly, using (7-16) and the arbitrage equation (7-15), one can write the optimal cash balance as: where UA and UE are the partial derivatives of U(.) with respect to A and E, respectively. Because of the (7-18) (M/p)* -= W1 ((pUA/pA)(r-hd log(PA/pE)/dt homogeneity assumption, from (7-13) we can derive + d logpA/dt)) a negative relation linking the ratio of the quantity of the European goods to the quantity of the African Given the real rate of interest, which has a negative goods to their relative price: impact, this equation says that real cash balances will be lower, the higher the rate of inflation, which is here (7-14) E = AfpA/pE),fp 0. measured by dlogpA/dt. The argument in the second parenthesis may be regarded as a kind of monetary In the utility function, U(.), the elasticity of substi- rate of interest. tution between E and A is defined by s = (pA/pE)f'/f 0. We can modify this equation to make it more suit- Second, we have the fundamental arbitrage equation able for the econometric application presented be- of this portfolio model which reads: low, given our data limitations. Using the definition of the price index p(.) given in (7-10), we can change (7-15) F'(H) + d log( /p)/dt (7-18) into: = (PA/p)(WIUA) - d logp/dt. (7-19) (MIp) *..W1 pUA/pA)(r (Q/g;)J logqpA/p) This relation equates the real rate of return on /dt+d logpA/dt)) holding African goods, on the left-hand side, to the real rate of return on holding real cash balances, on where Og=pEppE/p is the share of European goods in the right-hand side. The fonner is made of the mar- the price index. 72 Niger and the Naira: Some Monetary Cons eqances of Cross-Border Trade With Nigeria We have been able to test a log-linear approxima- or if the elasticity of substitution between European tion of equation (7-10), using monthly data from goods E and African goods A (s) is not significantly 1984.04 to 1988.10. We use again circulation fiduciaire different from zero. Moreover, if these two elasticities as our monetary aggregate. The price variable repre- are significantly different from zero, but small, their senting the Nigerian CPI is the rural price index, product h is even smaller and may not be signifi- which is multiplied by the parallel exchange rate to cantly different from zero. Dropping this variable obtain the CFA franc price of Nigerian goods. There- does not alter much the other features of the equa- fore ep* is a good indicator of the price of African tion: goods. On the other hand, since our price index for Nigerien goods is the CPI for African consumption (7-21) log(M/p) = -3.67 - 021 Dlog(ep*) in Niamey, it is relatively more influenced by the (4.56) (2.77) prices of European goods. Then (ep*) corresponds to + 0.31 log(ep*/p) - 0.001 RF pA and (ep*/p) to (pA/p). (3.09) (3.16) The following equation is based on this conven- + 0.98 AR(1) tion: (36.01) (7-20) log(AMp) = -3.5 - 0.35 Dlog(ep*) N=67, R2a = 0.93, DW = 2.10, F=204.67 (4.74) (2.60) + 0.17 Dlog(ep*/p) The fit remains impressive and there is no trace of (1.26) residuals autocorrelation. However, the coefficient of + 027 log(ep*/p) D log(ep*) is reduced, as this variable now captures (2.46) to some extent the positive effect of Dlog(ep*/p). -0.001 RF + 0.98 AR(1) This mongrel coefficient remains significant, but this (2.98) (34-34) suggest that (7-20) is preferable, as it separates clearly the two effects. N=67, R2= 0.93, DW = 2.08, F = 165.6. These equations bring out quite forcefully the role of the price index for Nigerian goods, and hence of Here, D log(.) represents the rate of growth of the the parallel exchange rate of the naira, in determining c-orresponding variable computed over 12 months. the level of real cash balances in Niger. We have used RF is an average of the monthly rainfall in the towns D log(ep') and D log(ep*/p) as explanatory variables, of Maradi, Tahoua and Zinder, which should be rep- assuming implicitly static expectations. Various at- resentative of the climatic situation in the relevant tempts have been made at applying the rational ex- geographic zone. Its significant impact and negative pectations hypothesis with various leads, using the sign may be regarded as capturing the negative im- "error in variable' method advocated by Wickens pact of rainfall on the perceived uncertainty of the (1982). The results were much less satisfactory. near future. This shifts the W'(.) curve downwards, As we have seen above, the price of African goods, by reducing the precautionary demand for money. as defined here, decreased over most of the period AR(1) is the coefficient of first-order residuals autore- analyzed. This would explain the increase in real gression, estimated by the Cochrane-Orcutt method. cash balances which has taken place. The increase in Its value suggests that the residuals are generated by real cash balances has probably played an active role a random walk process, i.e. that shocks tend to be in sustaining the surplus of the parallel balance of permanent. As for equations (7-7) and (7-8) above, payments which we have presented above. However we have represented by N the number of observa- we have not been able to provide a direct test of this tions used, by R2' the usual R2 adjusted for degrees assumption, for lack of appropriate data. of freedom, etc. The t-ratios are in parentheses below the estimated coefficients. Further Evidence The rate of growth of the relative price of African goods is not significant at the 20% (two-tailed) confi- Figure 7- 4 presents indexes of the real cash bal- dence level, although it has the correct sign. This ances series and of the estimated herd size for cattle. might simplybe dueto the strongcorrelation existing Data for the latter are known to be very inaccurate, between D log(ep*) and D log(ep*/p), but in terms of but are the only evidence available on this topic. The our theoretical model, it can be rationalized as show- estimated herd size series shows a sharp drop in ing that (hig) is close to zero, which is the case if either 1984, corresponding to the drought, but no recovery AUA,4UA is close to zero, i.e. if UA is nearly constant, took place after that date. 73 Niger and the Naira. Some Monetary Consequences of Cross-Border Trade With Nigeria The real cash balances series shows sustained geria. The price series for Nigerian goods is charac- growth in the second half of the period. Before the terized by a downward slope, which explains the drought, herd size was increasing and real cash bal- price deflation which took place in Niger in the sec- ances were decreasing. After the drought, herd size ond half of the 1980s. Thisdownward trend nayplay decreased or remained at a low level, and cash bal- a part in explaining the above mentioned surplus, ances increased in real terms. The econometric exer- within an asset approach framework. cise presented above suggests that the continuing We have analyzedasimplemodel of asset accumu- depreciation of the naira on the parallel market, pull- lation adapted to the structural features of the Niger ingtheprice index of African goods downwardsboth economy. The main point is that cross-border trade in nominal and in real terms, played a large part in in this area involves mainly storable goods like cattle determining this change in portfolio composition. and grains, and can thus be analyzed as a means for The parallel exchange rate of the naira stabilized exchanging physical assets ("African goods") against with respect to the CFA franc in 1989. If this break in cash balances. We have derived from this model an the downward trend which dominated the 1980s econometric equation to explain the behavior of real would be confirmed in the future, it would remove cash balances, based on monthly data. The results the incentive to hold large and increasing real cash confirm that the increase in real cash balances which balances. There maybe a sudden drop in the amount occurred during the 1980s can be explained by the of real cash balances and this might result in the decrease in the price of African goods. Further evi- parallel balance of payments turning to a deficit po- dencewas added to strengthenthispointby showing sition. that cattle herd size did not recover after the drought, but continued to decline. Conclusion We have not been able to provide a direct test of the relationship between the improvement in the In this paper we have shown first how a small parallel balance of payments and the increase in real economy with a convertible currency like Niger can cash balances induced by price deflation. Such a test acquire foreign exchange via its cross-border trade would require a more complicated model with speci- with Nigeria, despite the latter having an inconvert- fication of the balance of official payments. In the ible currency. The reason for this result is that the case of a convertible currency like the CFA franc, a naira market is well unified by arbitrage and must surplus in the parallel balance of payments can be therefore be cleared globally, by adjustment of the used either to increase cash balances or to transfer parallel exchange rate, and not on a bilateral basis funds from the parallel market to the official one. In with all the countries participating in this market. Niger during the 1980s, the accumulation of cash Hence Niger can run a surplus in its parallel balance balances certainly played a part in the improvement of payments, provided the rest of the participants in in the parallel balance of payments. this market run a net deficit of the same magnitude. Hence, cross-border trade should be given some at- Notes tention in the design of macroeconomic policy in Nigeria's neighboring countries. 1. 'Nigerien' means "frum Niger", while 'Nigerian' We have found an indicator of such a surplus of means "from Nigeria". the parallel balance of payments by looking at the 2. The validityof this assumption dependson theinsti- series of banknotes issued in the BEAC monetary tutional features of the commercial sector. It would not be union and purchased by the BCEAO in Niger. It appropriate in its simple form for the smuggling of cocoa happens that most of these notes are bought in Ma- out of Ghana, for example, where the parastatal Cocobod radi and Zinder, which are known as the main cen- has a monopoly on the cocoa trade and a private agent ters for cross-border trade with Nigeria. This indica- transporting cocoa is smuggling. Any attempt at cloaking tor shows that Niger has had a sustained surplus this trade must be more indirect, using official transporta- since 1983. tion of other goods as a cover. There is no such monopoly We have shown that this surplus cannot be ex- now in Niger. plained by applying the elasticities approach and 3. This is at variance with the assumption tested by referring to the competitiveness of Nigerien goods. Azam and Besley (1989) for the case of Ghana, where the A sort of purchasing power parity relationship supplycurveofthesmugglingindustryisupwardsloping. passed econometric tests. This suggests that the price 4. One should qualify this statement by remarking that level in Niger is mainly determined exogenously by the African CPI in Niamey might not be very repre- the price of the goods which are tradeable with Ni- sentative. Its weights are based on a survey done in 74 Niger and the Naira Some Monetary Consequeencs of Cross-Border Trade With Nigeria 1961-62, with a sample of 317 households. In particular, Jones, eds., Markets in Developing Countries: Parallel, Frag- rents are not included, whereas a recent survey has shown mented and Black. San Francisco: International Center for that they account for more than 15 percent of the expen- Economic Growth. Forthcoming. ditures of urtban households in Niamey. But it seems like- Azam, J.-P. 1990b. *Marches Paralleles et Convertibilite: ly that the corresponding figures for rural households and Analyse theorique avec references aux economies afri- households living in smaller towns is much smaller. caines.' Revue Economique. Forthcoming. 5. We neglect other goods, such as European goods like Azam, J.-P. and T. Besley. 1989. 'General Equilibrium with Benson & Hedges cigarettes, which are smuggled into Parallel Markets for Goods and Foreign Exchange: The- Nigeria via Niger, and Nigeria-made Peugeot cars which ory and Application to Ghana." World Development 17, are smuggled into Niger, which play a not-so-negligible 1921-1930. part in the real world (see Azam 1990a). Azam, J.-P. and C. Vernhes. 1990. 'La determination des 6. This homogeneity assumption is meant to simplify taux de change paralleles en Afrique: modele the presentation as a shortcut through some dynamic com- macroeconomique et test econometrique (Ghana, Nige- plexities. It gives the model 'bang-bang' reactions. It zia, Zaire)." Economie et Prvision. Forthcoming. should not be pushed too far, as it removes from the model Blanchard, O.J. and S. Fischer. 1989. Lectures on Macroeco- the well known consumption smoothing behavior which nomics, Cambridge, Massachusetts: MIT Press. is found in most standard theories of consumption and Devarajan, S., C. Jones and M. Roemer. 1989. 'Markets saving (see footnotes 8 and 9 below). under Price Controls in Partial and General Equilib- 7. If the homogeneity assumption (see footnote 7) was rium." World Development 17, 1881-1893. given up, the growth rate of A or E would show up in Gr6goire, E. 1986. Les Alhazai de Maradi (Niger). Paris: Edi- equation (17). But it could be substituted out in the full tions de l'Orstom. blown dynamic analysis, to get (17). Guillaumont, P. et al. 1988. "La politique d'ajustement au 8. This residuals dynamic may be regarded as a si- Niger 1980-1987." Report to the Government of Niger, mplified way to take into account the dynamics squeez- forthcoming in an abridged version as "Ajustement ed out of the theoretical model by our simplifying assump- structurel, ajustement informel: le cas de Niger." tion. McAleer, M., A.R. Paganand P.A. Volcer. 1985. "What Will Take the Con out of Econometrics?" American Economic References Review 75, 293-307. Pitt, M.M. 1981. "Smuggling and Price Disparity." Journal of Arnoult, E.J. 1983. "Cross-Border Trade BetweenNigerand International Economics 11, 447-458. Nigeria.! In Elliot Berg and Associates, Joint Program Sidrausky, M. 1967. 'Rational Choice and Patterns of Assessment of Grain Marketing in Niger. 2, 1-12, Niarney, Growth in a Monetary Economy." American Economic Niger: USAID. Review 57, 534-544. Azam, J.-P. 1990a. "The Balance of Cross Border Trade Wickens, M.R. 1982. "The Efficient Estimationof Economet- between Niger and Nigeria (1980-87): Evidence from the ric Models with Rational Expectations." Review of Eco- Parallel Market for the Naira." In M. Roemer and C. nomic Studies 49, 817-838. 75 8 Comments on Parallel Markets Paul Collier and Vikram Nehru Paul Collier gality sometimes, but not always, involves cost. The other effect that is ignored in the simple quota model I found these to be three rather convincing papers. is that foreign exchange is an asset, so there will be But I want to step back from them and ask about the asset demand effects. All three papers are really go- framework which might be used to look at the rela- ing beyond the simple endogenous trade policy tionships between changing the exchange rate and model to incorporate either asset effects or illegality inflation, which is the core of two of the papers. effects or both. We come to this problem with quite a lot of intel- lectual baggage, a lot of which is inevitable, but The Effect of Relative Prices on Devaluation unfortunately dysfunctional. Let me sketch two of the models that we bring along. One is the orthodox Within the endogenous trade policy model, what monetarist model as applied to developed countries happens when we devalue? There are relative price where we have purchasing power parity. What hap- effects and there are price level effects. What happens pens when we devalue? The price of all tradable to relative prices? The devaluation permits a trade goods rises in domestic currency and that pushes up liberalization. That is the first effect, so that the do- the demand for money, so for a given money supply, mestic price of exportables rises relative to import- the balance of payments improves. That is what I ables. There are two relative prices in this endoge- would call abalance of payments improving devalu- nous trade policy model. One is the domestic price ation. I think it is not relevant in the present context. of exportables to importables. The other is the do- The second model, which is a structuralist model, mestic price of importables to nontradeables. is particularly relevant to the Latin American context. We are told that the IMF measure of the real ex- The main focus is that a devaluation causes cost plus change rate is a proxy for the relative price of import- inflation, partly through pushing up import costs ables to nontradeables. But the effect of a devaluation and partly through pushing up wages in response to in the endogenous trade policy model is that the import costs, which somehow causes an endogenous relative price of nontradeables to importables rises. increase in the money supply. In the typical African devaluation, the IMF measure I think that neither one of these models is a very of the real exchange rate goes down a lot, and the good framework for analyzing the devaluations in relative price of nontradeables to importables actu- Africa over the last decade. The simple view of the allygoes up. That isone reasonwhythe IMF measure exchange rate premium ignores two potentially im- of the real exchange rate is so extremely confusing in portant effects. One is that parallel market transac- the African context. The concept of the real exchange tions involving foreign exchange are illegal, and ille- rate (the IMF measure) predates the whole structural 76 Comments on Parallel Markets period. It was a concept devised basically for devel- firms the endogeneity of trade policy in response to oped economies, and I think it isbeingmisapplied in terns-of-trade shocks. the African context. In developed countries trade policy does not mat- Illegality and the Effect of Assets ter very much. It can be ignored in the first approxi- mation. There is only one relative price-the price of Two modifications of the simple endogenous trade tradeables to nontradeables. In Africa there are two policy model involve illegality and the effect of as- relative prices and one damn term-the real ex- sets. These can be discussed in the context of ex- change rate. And there is only one measure, which is change rate unification-unification of the parallel theIMFlookingat these CPIsin Africa and wholesale and the official exchange rate. price indexes in the US. The concept and its measure are not up to the job. Illegality The Inflationary Effect and Effects on the Price The effects of illegality on the premium cut both Level of a Devaluation ways. Illegality of exports tends to raise the pre- mium, which has to cover the costs of getting dollars. When the trade policy is endogenous, we would Illegality of imports tends to lower the premium. expect that the price level would fall as long as the This is where Kaufmann and O'Connell's work on money supply is constant. This is because trade lib- the own-funded imports scheme in Tanzania is par- eralization leads to some real income gains. If the ticularly interesting. I think they show that the ovn- money supply is constant and real income goes up, funded import scheme has indeed tended to raise the the price level goes down. So we would expect from premium. A policy inference from that might appear our simple model that devaluation in the African to be that we jump straight to unification of an ex- context would be counterinflationary, as long as the change rate and forget about an intermediate stage money supply is constant. where we have an own-funded import scheme be- Another rue to the devaluation is how it changes cause an own-funded import scheme pushes the pre- the money supply, which is a matter of how it mium up relative to what would happen with unifi- changes the budget and the politics of the off-setting cation. I think one qualification to that is that what effects. Tax revenue from exports goes down because Kaufmann and O'Connell are doing is true in a flow the overvaluation of the exchange rate is an implicit sense, but certainly in Tanzania (and I think most tax on exports. But off-setting that, the implicit sub- countries), prior to reforms there had been quite a lot sidy on imports, which is the quasi-expenditure of of build-up of illegally held dollars, so there isa stock the government, also goes down. Usually in Africa, of dollars held abroad. On introduction of the own- because imports exceed exports, the budgetary effect funded import scheme, that stock of dollars tends to of the reduction in the implicit subsidy exceeds the be repatriated. That repatriation of dollars tends to budgetary effect of the reduction in the implicit tax. depress the premium. Itis nota sustainable thing, but You can usually expect the budget to improve. There something that will happen over two or three years. are two reasons why we would expect devaluation So we might cash in on this short-run repatnation to be counterinflationary. The first is the effect of effect without having the premium overshoot the raising real income for a given money supply. The level it would reach were we to reunify the rate. second is the effect of the improvement in the budget and reduction in the money supply. Asset Effects The Chhibber-Shafik paper takes a structuralist perspective on the relationship between devaluation All three papers stress asset effects. The Kauf- and inflation. The authors say that they get counter- mann-O'Connell paper and Chhibber-Shafik paper intuitive results in that devaluation in Ghana has stress interest rate effects. The Azam paper ingen- actually reduced inflation. I am not at all surprised iously adds the real economy effects; cattle in the end by the result, which to me is entirely intuitive. It is are the real determining feature. I have one problem not counterintuitive if you start from a cost plus with the interest rates effects on the premium. The framework. In my opinion this just shows that the policy conclusion seems to be to raise nominal inter- framework is not right for Africa. The O'Connell- est rates. I am worried by that because certainly in Kaufmann paper comes to exactly the same result- the Tanzanian context, and I think perhaps in the that the devaluation of the official exchange rate in Ghanaian context, the government is heavily in debt Tanzania has been counterinflationary. It also con- to the banking system. In Tanzania the banking sys- 77 Comments on Parallel Markets tem is basically there to lend money to the governme- ease the interest rate on money, it is purely a transfer nt and parastatal organizations which are acquiring between some holders of money and other holders thebankloans. The private sectorisjust makingbank of money. The real return goes up in Tanzania for that deposits. So when we raise nominal interest rates, th- quarter of money holders who are holding their mo- ere is a direct pass-through onto government expen- neyin interest bearing money. The counterpart is that diture and so an increase in the budget deficit. Thus the real interest rate goes down for three-quarters of raising nominal interest rates just increases the infla- money holders who are holding non-interest bearing tion tax. The holders of money pay the inflation tax. money. I worry thatwe are getting these interest rates So increasing the nominal interest rate does not incr- effects in the model. I do not think we should. Vikram Nehru The three papers presented-although all on the sub- tion need to be related to return on financial assets ject of parallel foreign markets-were written with abroad. In other words, quite apart from its other different objectives in mind. The Chhibber-Shafik virtues, the existence of parallel foreign exchange paper on Ghana examines the consequences of these markets imposes a discipline on macroeconomic markets for inflation; the paper by Azam traces the management that can only improve the domestic cases of large purchases of BEAC bank notes in Nia- environment for investment and growth. mey, Niger; and the Kaufmann-O'Connell paper My second general observation is on the key issue focuses on the effect of Tanzania's macroeconomic of foreign exchange market unification mentioned in policies on the premium of the parallel foreign ex- both the Ghana and Tanzania papers. The evolution change market. I would first like to make two general of Tanzania's official and parallel foreign exchange observations and then proceed to examine some spe- markets appear to be identical to those of Ghana, but cific issues in the papers presented. with a two-to-three year time lag. In this regard, the My first general observation is that, on the basis of papers on Tanzania and Ghana have struck the right these papers, it is clear that channels for capital flows cautionary note with regard to the question of "uni- have become permanently etched in the African fication" of the official and parallel markets. Both landscape, well concealed to be sure, from the eyes papers are appropriately circumspect on the ques- of the law, but a reality nevertheless that no policy- tion of capital controls remaining an essential feature maker can ignore. These channels have served two in the official foreign exchange market. In moving functions: a defense by the private sector to maintain from its present system of foreign exchange alloca- its real wealth in the face of efforts by the State to tion through import licensing, Tanzania could use- redistribute national resources to the public sector- fully study Ghana's experience with the foreign ex- and a way to cope with uncertainties ensuing from change auction and its gradual liberalization since genuine concerns of economic and political instabil- September 1986. Only five weeks ago, Ghana intro- ity. All three papers have observed that a critical duced an extended interbank foreign exchange mar- factor explaining parallel foreign exchange rate ket supported by weekly wholesale auctions with a movements is private capital flows arising as a result system of capital controls which, to my mind, strikes of portfolio adjustments or through such means as a fine balance between caution and liberalization. over and under invoicing. What are the implications There is also a more practical and mundane pur- of this important observation? The first is that the pose behind maintaininga monitoring system for the marginal rate for foreign exchange is no longer an use of official foreign exchange-that of ensuring the exogenous variablebut an endogenous one that poli- documentation support for disbursements of pro- cymakers need to target for several reasons-the gram aid. Once all current transactions and amorti- most important being that the parallel rate effectively zation payments are covered by the official market, sets the relative price of tradables in the economy. the effective difference in coverage of the official and The second is that aggregate demand management is parallel foreign markets will then be entirely due to central for stability of the exchange rate. And the capital flows, which the analysis in these papers third is the elevation of monetary policy as one of the shows us to be highly influenced by the level of tools for achieving external balance, given that do- domestic deposit rates adjusted for expected depre- mestic deposit rates adjusted for expected deprecia- ciation of the exchange rate. The focus of policy 78 Comments on Parallel Markets would then shift to monetary management as a a significant part in the CPI does not appear to bome means to bringing about the "unification" of the two out by the results in equation 6. 1 would suspect that markets that most economists consider so desirable. adding money to equations 6 and 19 would provide more convincing results and a better understanding The Azam Paper of what exactly is going on in Niger. I found the paper by Azam on cross-border trade The Kaufnann-O'Connell Paper between Niger and Nigeria intriguing and fascinat- ing. Mr. Azam's hypothesis starts with the observa- I found little to disagree with in the analysis of the tionthatarealappreciationof theNigerienCFAfranc relationship between the parallel market premium vis-a-vis the Nigerian naira coincided with a cross- and key macroeconomic variables in the Kaufm- border surplus in favor of Niger as indicated by ann-O'Connell paper. However, the reduction of the purchases of BEAC banknotes by the BCEAO in large set of interrelated variables into the two re- Niger. The reason, according to Mr. Azam, is the duced form equations given in tables 3-1 and 3-3 left reduction in the price of all rural goods in Niger one somewhat confused. I was also puzzled as to following a similar decline in the rural Nigerian why the domestic money stock is expressed in ter- CPI-including the price of cattle, which is a part of ms of foreign exchange after being converted at the the Hausas' savings portfolio. We are then told that official exchange rate. Finally, I agree with the Niger's bilateral balance of payments surplus with authors' judgment that an important lacuna in the Nigeria can be attributed to sales of this "bovine model is the absence of a fiscal module. A key qu- portfolio" by the Nigerien Hausas to their Nigerian estion would be how changes in the official rate brethren across the border, because of the relatively would affect the budget deficit and in tum, through higher return to holding cash balances than to own- the money supply, have a bearing on the parallel ing cattle. This is an ingenious explanation, but one premium. Finally, I did not get a sense of what is that warrants some further examination. being done in monetary management and financial First of all, we should interpret with some caution, sector reform in Tanzania, and how this may have a theresultsofananalysiswhichusesdataontheown- bearing on the interest parity differential. In this ership of cattle for countries that share a thousand context, it would be useful if the paper ran some milecommonborderwithlittlemeaningtoHausaca- "what-if" simulations to highlight key policy op- ttle herders. But even were this data to be accurate, tions. figure 4 in the paper does not appear to support the hypothesis. Although herd sizes stabilized after 1986, The Chhibber-Shafik Paper the rise of real cash balances appears to continue un- abated. An alternative hypothesis, and one that dese- I am at somewhat of a loss to comment on the ves some examination, is that the real depreciation of paper by Chhibber and Shafik because the authors the Naira led to capital outflows from Northern Nig- graciously took on board some of the comments I had eria to Niger. With a fixed exchange rate, the conse- made on earlier drafts. I therefore look forward to a quences would be an increase in real cash balances, more objective assessmentby Paul Collier. Neverthe- which, presumably, were not drawn down because less, permit me to ask one, perhaps simple, question. of continuing uncertainties in Nigeria. Finally, on the Why is it that in the simulations, the general course Azam paper, I felt that the analysis could have bene- of inflation appears largely unaffected despite sig- fitted from a more comprehensive treatment of the nificant changes in key assumptions? Can the realtionship between the parallel exchange rate, do- authors point to certain fundamental factors driving mestic inflation, and the money supply in Niger. The inflation that appear impervious to large variations statement that nontraded goods do not seem to play in the official exchange rate or production? 79 9 Issues in Public Expenditure Policy in Africa: Evidence from Tanzania's Experience Laurean W Rutayisire In an effort to counter Tanzania's economic crisis, the public sector to increase capacity utilization, im- authorities launched a structural adjustment pro- prove labor utilization, and reduce unpro-ductive gram in 1982 and an economic recovery program in activities. Put into practice, this has involved reduc- 1986. These programs articulated the new objectives ing public sector employment, closing some public of economic policy which the country would pursue corporations, shelving some development pr-ojects, and the measures by which the objectives would be and decentralizing some central government func- achieved. The objectives have been the following. to tions to local governments. Complementary to these reduce the rate of inflation; to achieve balance of measures, the authorities have further resolved to payments adjustment so as to alleviate the existing increase the role of the market in allocation. Thus, extreme foreign exchange scarcity and the conse- interest rate and producer pricing policies havebeen quent underutilization of domestic production; to used to improve efficiency in the allocation of domes- achieve an increase in the productivity of parastatal tic resources. Exchange rate policy has been used to enterprises and improvement in public sector man- improve the country's international export competi- agement; to secure increases in food and cash crop tiveness and to provide export producer incentives. production by providing adequate producer incen- There have also been trade liberalization measures, tives, marketing and distribution systems, and which have included removal of restrictions on a resources available to agriculture; and to maintain wide range of imports and reduction in the number the already achieved equity in income distribution, of commodities under price control. including the provision of social services and other The economy of Tanzania is beginning to experi- basic needs to the majority of the population. ence signs of recovery, including the reversal of out- According to the structural adjustment and eco- put decline. However, the country's economic crisis nomic recovery programs, the implementation of the seems to have acquired an additional dimension. above objectives involved the following measures. There has been an increasing deterioration in the First, the government has constrained expenditure to quality of the country's basic economic and social reduce money supply growth and inflation. This has infrastructure, which is apparently causing a con- involved decreasing government's share in total do- straint on subsequent recovery efforts. There are two mestic credit, in order to minimize its crowding out major perspectives on explaining the causes of such impact on the availabilityof bank credit to the rest of deterioration in the quality of economic and social the economy. In order to reduce government expen- infrastructure. One line of argument is that the fiscal diture, various budget subsidies have been abol- and monetary policy measures adopted under the ished. The second set of measures adopted has con- structural adjustment and economic recovery pro- sisted of the rationalization of prod-uction in the grams have been perversely restrictive. The result 80 Issues in Public Expenditure Policy in Africa. Etidencefrom Tanzana's Expience has been inadequate funds for infrastructural reha- (9-1) D - G - - K - (P/P) GR-1 - ( e/e G ex - nPSFD) bilitation and maintenance.' An alternative view similarly maintains that the deterioration in the qual- The unsustainable or permanent deficit is denoted ity of the country's economic and social infrastruc- by D and govemment expenditure is denoted by G. ture is attributable to underfinancing. However, it Govemment revenue, is expressed in permanent in- relates the underfinancing of provisions to the over- come terms as a three year moving average; K de- grown size of government expenditure.In other notes nominal government capital expenditure; PXP is words, the underfinancing has been the result of a the actual inflation rate; 4'e represents change in the thin distribution of govemment expenditure over a nominal exchange rate; GR-I is government recurrent wide range of items.2 expenditure lagged one period; GCx is net foreign This paper attempts to explain the decline in the obligations denominated in local currency; and nP- qualityof Tanzania'sbasiceconomic and social infra- SFD is a growth adjustment term. PSFD denotes the structure. In order to identify the appropriate per- budget deficit and n denotes real GDP growth. spective, the literature on optimal and sustainable First govemment permanent income (M) is sub- level of government expenditure is first surveyed in tracted from overall government expenditure (G). the second section. The question of what can be Govemment permanent income should be the in- regarded as an optimal and sustainable level of gov- come which the government expects to eam on aver- emnment expenditure is also examined in this section. age and over a long period. This definition considers Section three proceeds to ascertain the optimal and current and discounted future streams of revenue sustainable level of government expenditure in Tan- from taxes, government property, and capital stock. zania and to what extent the current size of govern- Government permanent income hasbeen proxied by ment expenditure in the country has departed from a three year moving average of govemment tax reve- such a level. During the structural adjustment and nue, royalties on government natural property economic recovery programs, the level of govern- rights, dividends from public enterprises, and other ment expenditure has remained in excess of the op- government revenue. The approximation of penna- timal and sustainable level for the country. Therefore nent income by a moving average of income series the factors which explain such a persistently high has been used in a number of studies.3 level of government expenditure are explored. Some Government capital expenditure (K) is subtracted of the effects of increased govemment deficit spend- from overall govemment expenditure less govern- ing are examined in the fourth section. Suggestions ment permanentincome. The residual constitutes go- and conclusions are presented in section five. vernment recurrent expenditure, which takes into ac- count the productivity gain which may derive from On the Optimality and Sustainability of government capital projects. The sale of governm- Government Expenditure ent projects can also generate revenue with which the government budget deficit may be bridged. Gov- An optimal and sustainable level of government ernment recurrent expenditure is partly on account expenditure maximizes the government's welfare of servicing capital projects. However, it is largely improvement program subject to the constraints im- made up of general recurrentexpenditure (consump- posed by the government's permanent income, its tion spending) and debt service. Thus the equation target of reducing inflation, and the objective of mini- emphasizes the need to balance government recur- mizing the burden and crowding out impact of gov- rent expenditure and government pennanent in- ernment debt (Zee 1988, Buiter 1983). It has been come. argued that the inflation adjusted budget should be In the existing literature,4 the need to account for balanced.This would support government fiscal and the impact of inflation on the government budget monetary policy which is committed to bringing deficit has been discussed mainly in connection with down inflation and to creating conditions in which the government debt. The argument has been that sustainable economic growth can be achieved (Miller government deficit spending is liable to increase due 1982). Balancing the government budget adjusted for to inflation indexed interest payments. However, it inflation requires that government expenditure be has also been argued that inflation leads to real dec- held constant in real terms; it should be allowed to lineintheoutstandingstockof governmentcdebt.The increase only in nominal terms. excess of the latter impact over the former should be An optimum and sustainable govemment expen- considered as additional revenue due to the infl- diture level can be ascertained as in the following ation tax. The converse would cause a net increase in equation (Zee 1988, Buiter 1983b, Miller 1982): thegovemnmentbudgetdeficitasaresultof inflation. 81 Issues in Public Expendibare Policy in Africa: Evidencefrom Tanzania's Expience Although the inflation tax should be taken into in real GDP entail increases in govemment capability account, in practice govemment accounting is done through increased revenue with which to sustain a on a cash basis rather than on an accrual basis.There- higher level of expenditure. Thus, in summary, the fore, the adjustment in the above equation is for balancing of government recurrent expenditure and nominal increases in government expenditure due to govemment permanent income is adjusted for infla- inflation. This is in view of the fact that the inflation tion, exchange rate change, and real GDP growth. tax on outstanding debt is only an accrual item. However, the equation does not include a cyclical However, cost price interactions affect the cost of adjustment term. During the later years of the struc- public provisions as well. The adjustment for infla- tural adjustment program and the economic recov- tion which is provided for in equation (9-1) thus ery program, the authorities have abolished a wide considers both debt and nondebt government recur- range of budget subsidies and increased discretion- rent expenditure. ary taxes. These measures have tended to minimize Theissuebehindinflationadjustmentofthegover- the cyclical sensitivity of the govemment budget. nmentbudget deficit is the need to hold govemrnent Hence, a provision for this adjustment has not been expenditure at a constant level in real terms and only made in the equation. provide for nominal increases in govemment expen- Other studies have suggested that a sustainable diture due to inflation. Thus, in the above equation, govemment expenditure program should provide a nominal increase in govemment expenditure due forthegovernment'smajorroleinincomeredistribu- to inflation is provided for on govemment expendit- tion or the social contract deficit (Bossons 1986). Con- ure in the preceding period, which, according to the sidering the economic policy measures which were above argument, should be held constant. Holding summarized in the introduction, it is evident that the govemment expenditure at a constant level in re- during the structural adjustment and economic re- al terms is consistent with the policy target of bring- covery programs emphasis has been on securing ing down inflation. Moreover, inflation indexation functional income redistribution byadoptingappro- can enhance price stabilization efforts to the extent priate producer pricing and marketing reforms. that it will institute inertia in cost price interaction.5 Hence a provision for the govemment budget's in- The equation also provides for the impact of ex- come support has not been made in the equation. change rate change s/ on the government budget In addition to these qualifying remarks, this study deficit. In particular, the adjustment indicated pro- has concentrated on the sustainability of total gov- vides for the net direct impact of exchange rate emment expenditure rather than its composition. change on the govemment budget through foreign This perspective has been maintained in macroeco- exchange denominated receipts and payments, in- nomic stabilization policies. A treatment of the com- cluding the import content of the budget GeX. The position of govemment expenditure in Tanzania has equation provides for real GDP growth n. Increases been made elsewhere. Table 9-1. The sustainability of government expenditure in Tanzania.fiscal years 1982183 to 1987/88 (millions of shillings) Item 1982/83 1983/84 1984185 1985/86 1986/87 1987/1988 1 Overall government expenditure 19,276 23,918 26,728 33,220 55,481 77,326 2 Government permanent incomea 11,919 14,644 17,662 23,392 32,940 49,400 3 =1-2 7,357 9,274 9,066 9,827 22,542 27,926 4 Capital account expenditure 4,404 5,736 5,391 5,817 15,091 17,255 5 = 3-4 2,953 3,538 3,675 4,010 7,451 10,671 6 Inflation adjustment" 1,%0 2,231 2,727 3,201 4,110 6,059 7 = 5-6 992 1,307 947 810 3,340 4,613 8 Exchange rate change adjustment (net) 153 449 0 179 -3,139 -3,313 9 = 7-8 839 858 947 630 6,479 7,925 10 Adjustment for GDP growth 45 -233 -211 297 723 1,198 11 = 9-10 surplus (-) and defecit (+) 794 1,091 1,158 333 5,756 6,727 Notes: a Permanent income is computed as a three year moving average of government tax and nontax revenues excluding external import support grants. b In the course of government expenditure estimates the authorities have directed a 15 percent inflation adjustment, which is also used here. Source: The computations are based on the optmal and sustainable government expenditure level equation in the text (9-1). Data has been obtained from the Economic Survey, Bank of Tanzania Reports, and Government Expenditures Supply Notes Vol. II. 82 Issues in Public Expenditure Policy in Afic Evidencefrom Tanzania's Experience The Size of Governrent Expenditure in Tanzania eventual recourse to raising seigniorage revenue or increasing reliance on foreign borrowing. The cur- The above measurement of the sustainability of rent level of government expenditure cannot be sus- govemment expenditure has been applied to the case tainable unless there is an increase in the current and of Tanzania. The focus here is on the period covered expected future levels of government revenue and by the structural adjustment and economic recovery increased productivity in the public sector. programs. The results in table 9-1 proceed according In an attempt to examine the factors which expl- to the above equation. As is indicated by the results ain the level of government expenditure in a count- in row 11, the level of govenunent expenditure in r* it is useful to begin by looking into the rationale Tanzania entails abudget deficit which is not sustain- for govemmental intervention. Besides the stand- able within the constraints imposed by the country's ard market breakdown arguments and merit goods economic stabilization and recovery targets and the pro-vision, the control of the commanding heights of constraint imposed by the government's permanent the economy argument was important in the case of income. Tan- zania. Whatever the justification for govern- In table 9-2, the results are presented for the same mental intervention, concem has been expressed in measurement process with a change in the inflation the lit-erature about how governmental agencies adjustment term. In this case, the inflation adjust- have ifl-uenced government expenditure levels. ment in row 6 of table 9-2 provides for a full inflation Various oth-er factors have also been emphasised in indexation of government expenditure. As a conse- the havioral explanation of high govemment expen- quence of this change, the results in row 11 indicate diture.8 a budget surplus during the period covered by the Figure 9-1 shows the trend in government expen- structural adjustment program (1982/83 to 1985/86). diture during the period from 1970 to 1987 in Tanza- An unsustainable deficit is indicated during the pe- nia. There have been displacements9 in goverunent riod covered by the economic recovery program expenditure levels in 1977, following the break up of (1986/87 to 1987/88). the East African Community; in 1979, following the T1hese results indicate the extent of the conflict in Idi Amin war; and in 1981/82, following the onset of perspective which was reviewed in the introduction. the country's current economic crisis and the failure The major difference is in the level of inflation in- of the National Economic Survival Programme dexation accommodated by the governmentbudget. (NESP). However, overall there has been steady However, even after accommodating a full inflation growth in government expenditure. indexation, the results still show that the current level The displacement effects are treated as random of government expenditure in the country is unsus- shocks. Our focus addresses the institutional percep- tainable. Such a level cannot be sustainable without tion of the tax price of government expenditure and Table 9-2. The sustainability of government expenditure in Tanzania with full inflation indexation fiscal years 1982/83 to 1987/88 (millions of shillings) Item 1982/83 1983/84 1984/85 1985/86 1986/87 1987/1988 1 Overall government expenditure 19,276 23,918 26,728 33,220 55,481 77,326 2 Government permanent income5 11,919 14,644 17,662 23,392 32,940 49,400 3 =1-2 7,357 9,274 9,066 9,827 22,542 27,926 4 Capital account expenditure 4,404 5,736 5,391 5,817 15,091 17,255 5 = 3-4 2,953 3,538 3,675 3,936 7,451 10,671 6 Inflation adjustmentb 3,780 4,026 6,569 7,101 8,887 12,097 7 = 5-6 -828 -488 -2,895 -3,165 -1,436 -1,425 8 Exchange rate change adjustment (net) 153 449 0 179 -3,139 -3,313 9 = 7-8 -981 -937 -2,895 -3,344 1,703 1,887 10 Adjustment for GDP growth 45 -233 -211 297 723 1,198 11 = 9-10 surplus (-) and defecit (+) -1,025.9 -704 -2,684 -3,641 980 689 Notes: a Permanent income is computed as a three year moving average of govemment tax and nontax revenues excluding external import support grants. b In the course of government expenditure estimates the authorities have directed a 15 percent inflation adjustment, which is also used here. Source The computations are based on the optmal and sistainable government expenditure level equation in the text (D+...). Data has been obtained from the Economic Survey, Bank of Tanzania Reports, and Government Expenditures Supply Notes Vol. 11. 83 Issues in Public Expenditure Policy in Afrlc Evidence from Tanzania's Erpenence how that has influenced increases in government import content of the government budget and the expenditure, the accounting procedures used for domestic price level.11 public expenditure, and the cost price impacts on The above hypotheses have been tested for Tanz- government expenditure. According to the instit- ania. Theresults are shown in the followingequation: utional perception of the tax price of government expenditure (Buchanan and Wagner),10 deficit fi- (9-2) G1 = 0.032 - 0.531Ti + 1.394G1(.I) + 0.4381MP( 1) nancing, among other consequences, reduces the (0.119) (-1.118) (10.555) (2.280**) perceived price of public provisions. This, in turn, + 42.175POP1 induces higher demand for public provisions on the (1.508) part of voters and parliamentary members. The public agency behavior, or the incrementalist R2 = 0.963; R2= 0.950; DW = 1.88; hypothesis, postulates a systematic upward bias in F-Stat. = 77.9; d.f.= 12 government expenditure determination (Wildavs- * = significant at 2% and 5% respectively ky 1964). It asserts that in the budget making proc- ess only a few unusual departures of amounts in the Here Gi is overall government expenditure as a budget from those allocated in preceding years will percentage of GDP. The ratio of tax revenue to total receive scrutiny, and that a large number of average financing of government expenditure, Ti, has been increase items will be approved routinely. However, used as a measurement of the perception of the tax the hypothesis does not explain why this occurs. The price of government expenditure. G1 lagged one pe- experience of Tanzania seems to suggest that lack of riodGi(.l) represents the incrementalist hypothesis. proper procedures for costing of public provisions The percentage change in the import price index and distorted incentives which result in agency ex- IMP(_1) measures the input cost impact. The ratio of penditure maximization are among the main factors total population to total govemment expenditure behind the systematic public expenditure bias. POPI is a further measure of the cost price impact. There are several input cost effects on government The above results indicate that the incrementalist expenditure. One is the impact of population growth hypothesis and the input price impact have signifi- on government expenditure. The quality level of cantly influenced government expenditure increases services offered is limited by deteriorating capacity. in Tanzania. The tax price impact hypothesis argues The presumption is that government responds in that a reduction in the ratio Tlor increased reliance order to restore the quality of the services. Another on deficit financing'should lead to a reduction in the cost impact is the input price impact. Our emphasis perceived tax cost and conversely to a rise in govem- has been on import prices and their influence on the ment expenditure. In the above results, the coeffi- cient estimate has the expected negative sign but is insignificant. This could indicate the extent to which in Tanzania, 1970-87 the tax price of government expenditure has not had any significant price rationing impact on the demand Tanzanian shilling (millions) for govemment expenditure increases. 60,000 The coefficient of the ratio of population to govern- ment expenditure is positive as expected, but is also 50,000 / insignificant. The lack of expenditure response to counter the impact of population pressure on the quality of public service provisions would support 40,000 l the contention that the quality of services has dete- riorated. However, evidence of unsustainably high 30W 000 - /government expenditure may indicate that the de- cline in the quality of services reflects a thin alloca- tion of government finances over a broad expendi- 20,000 _ // ture commitment. 10,000 -- Some Implications of Government Deficit 10,000 X § E | | | | | Spending Increases 0 i I I I i A number of studies have shown the extent to 1970 1972 1974 1976 1978 1980 1982 1984 1986 which the monefization of deficit spending has been 84 Issues in Public Expendibure Policy in Afriw Evidece from Tanzama's E&eJenm among the main factors which have accounted for tionandfactorsupplyingentitiesbasetheirdecisions Tanzania's inflation.12 This section focuses on the onrelativepricesonly,thenchangeinrealoutputwill implications of government deficit spending and its vary with perceived relative prices and with its own possible crowding-out impact. lagged value. Excess government borrowing from the banking Lucas' specification of this assertion is the follow- system has resulted in quantitative credit rationing ing: in the rest of the economy. The testing of the above hypothesis has involved specifying and estimating (9-4) Yt = bi(Pt - E (Pt/lt-1)) + b2Yt-i1 the bank credit supply function. This includes a vari- able for governmentborrowing from thebanking sy- The variables denote real output, Yt; the actual infla- stem. A significant negative coefficient on this varia- tion rate, Pt; the expected inflation rate, E (Pt/Ilt . ble should indicate crowding out, while a significant and lagged real output, Yt.I.Lucas' conclusion is that positive coefficient should indicate crowding in. any policy response which purports to increase out- On the basis of the bank credit supply function put and leads to inflation, will only lead to output specified by Bryan and Carleton,13 the following increases as long as the actual inflation rate exceeds equation has been adopted: the expected inflation rate. This function has been fitted on data in developed and less developed coun- (9-3) CDRt =a, + alTDDRt + a2GBDRt + a3RL t+ Ut. tries.15 The Lucas supply function specified in the above The credit to total deposits ratio is denoted by equation has been extended to include a proxy for CDRt, for time period t. The ratio of time deposits to credit rationing as: total deposits, TDDRt, is introduced as a measure of the extent to which banks do not need to hold excess (9-5) Yt - bi(Pt - t( Pt/It-1)) + b2 Yt-1 + b3B1t reserves for the purpose of meeting the demand for cash. Given other factors, an increase in the ratio of Credit rationing in the rest of the economy, Bt, has time deposits to total deposits should facilitate in- been measured by the amount by which credit to the creases in credit advances by banks. Government rest of the economy has decreased as a result of borrowing from the banking system as a percentage government bank bortrowing. The variable has been of total deposits, GBDRt, has been introduced as a generated as a fitted value from the regression of measare of the extent to which credit ceilings are credit to the rest of the economy on government imposed on banks' credit to the rest of the economy borrowing from the banking system. as a result of excess government borrowing from the During the structural adjustment and economic banking system. The real lending rate, RLt, is in- recovery programs, restraints on government bor- cluded in the model as a measure of returns on bank rowing from the banking system have succeeded in lending operations. An increase in the lending rate reducing such borrowing to a level below the targets should be expected to induce banks to provide more set by the programs. Hence the period before the loan advances. structural adjustment program, from 1966 to 1982, The data for the variables have been computed has been considered to be more relevant in the analy- from the Bank of Tanzania reports. A weighted aver- sis of the crowding out effect. Results obtained using age of lending rates on loans of various maturities the ordinary least squares estimator are presented in has been computed. Hence, the real lending rate has table 9-3. been obtained as the weighted average of the nomi- As shown in equation (9-1), which refers credit to nal lending rates less expected inflation. The above agricultural production, the coefficient on govern- credLit equation has been fitted on bank credit to ment borrowing from the banking system is signifi- agricultural production, agricultural marketing, cantly negative. In the case of credit to agricultural manufacturing, public enterprises, and the private marketing, which is shown in equation (9-2), the sector (table 9-3). coefficient on government borrowing is significantly The estimation of the crowding out of credit to the positive. Thus, while credit to agricultural produc- rest of the economy can be regarded as one facet of tion has been crowded out, credit to agricultural the impact. A second facet relates to the impact of marketing has been crowded in. such a crowding out of credit on output. In order to Equation (9-3) indicates estimates of credit to estimate the consequent output implication, the Lu- manufacturing.The coefficient on government bor- cas supply function has been adopted. The function rowing is negative but insignificant. A crowding in postulates that, given a situation in which produc- of the public enterprise investment would be ex- 85 Table 9-3. The quantitative credit rationing impact of government borrowingfrom the banking system in Tanzania, 1966 to 1982 (i) Dependent variable: CADRt (credit to agricultural production as a percentage of bank deposits) Constant TDDRi GBDRt RLt Coefficient 0.072 0.001 -0.121 -0.0001 T-statistic (2.196) (0.01) (-3.075*) (-0.087) R2 = 0437 0.307 DW = 1.7 F-Stat = 3.4 d.f. = 13 (ii) Dependent variable: CAMDRt (credit to agricultural marketing as a percentage of bank deposits) Constant TDDRt GBDRt RLt Coefficient 0.264 0.401 0.373 -0.004 T-statistic (3.470) (-1.584) (3.016*) (-1.750) R - 0.469 R2 = 0.347 DW = 1.2 F-Stat = 3.53 d.f. = 13 (iii) Dependent variable: CAMDRt (credit tothe manufacturing sector as a percentage of bank deposits) Constant TDDRt GBDRt RLt Coefficient 0.145 0.079 -0.196 -0.001 T-statistic (1.074) (0.178) (-0.849) (-0.155) R = 0.45 DW = 1.63 F-Stat = 02 d.f. = 13 (iv) Dependent variable: CAMDRt (credit to the private sector measured by commercial bank assets held in the private sector) Constant TDDRt GBDRt RLD Coefficient 0.667 -0.487 -0.633 0.003 T-statistic (2.280) (-0.969) (-1.832**) (0.632) R2 = 0.382 R2 = 0.239 DW = 1.64 F-Stat = 2.5 d.f. = 13 (v) Dependent variable: CAMDRt (credit to public enterprises measured by commercial bank assets held in public enterprises) Constant TDDRt GBDRt RLD Coefficient 0.098 0.274 1.094 -0.007 T-statistic (0.430) (0.375) (3.956*) (-0.704) R2=0549 R2=0445 DW=0.8 F-Stat=5.3 d.f.=13 86 Issues in Public Expenditume Policy in Africa: Evidencefrom Tanzania's Experience Table 9-3. (continued) (vi) Dependent variable: LQAt (natural log of output of the agricultural sector) Constant LCPlt LQAt-l L8t Coefficient -0.008 -0.280 0.880 -0.042 T-statistic (-0.714) (-2.773) (2.767*) (-0.843) R2 = 0.653 'R2 = 0.513 DW = 1.83 F-Stat = 6.33 d.f. = 12 (vii) Dependent variable: LQMFt (natural log of output of the manufacturing sector) Constant LCPIt LQMFt-I LBt Coefficient -0.026 -1.045 0.888 -0.508 T-statistic (-1.555) (-6.382*) (7.653*) (-0.826) R = 0.923 = 0.904 DW = 2.07 F-Stat = 48.05 d.f. = 12 Note: The asterisks, * and ", indicate significance at 2 percent and l0 percent levels respectively. TDDR is the ratio of time deposits to total deposits; GBDR is government borrowing from the banking system as a percentage of total deposits; RL is the real lending rate; CPI is the consumer price index; B is a proxy for the crowding out impact of government borrowing on the rest of the economy's credit; and L indicates natural logs. pected, since manufacturing public enterprises have above results. Evidently, credit to agricultural mar- accounted for a large share of the country's invest- keting is as critical in importance as credit to agricul- ment through the period in question. The estimation tural production, because peasant producers are mo- of manufacturing sector credit in equation (9-3) ag- tivated by the ability to sell their crops. Similarly, gregates public and private manufacturing under- credit to public enterprises has a positive impact on takings; a disaggregation is made in equations (9-4) their output, due to liquidity constraints on working and (9-5). Equation (9-4) presents commercial bank capital. As for the output effect of the crowding out credit to private enterprises. The coefficient on gov- of credit to private enterprises, existing evidence in- emment borrowing is significantly negative. Equa- dicates that a large fraction of private enterprises' tion (9-5) relates to public enterprises' credit, and the investment finance has derived from alternative coefficient on government borrowing is significantly sources other than credit from the National Bank of positive. These results indicate the extent to which Commerce. government borrowing from thebanking system has crowded out credit to agricultural production and Conclusion private enterprises. The same borrowing has crowdedincredittoagriculturalmarketingand pub- This paper has addressed the size of government lic, enterprises. expenditure in Tanzania. Two conflicting concems In equations (9-6) and (9-7), the Lucas supply over the current level of govemment expenditure in function has been fitted on agricultural sector output the country have been identified. One is the concem and manufacturing sector output respectively. A thatthecurrentmeasurestobringaboutreductionin proxy of the crowding out impact of govemment the size of govemment expenditure have been un- borrowing on the rest of the economy's credit has duly contractionary. The other is the view that gov- been included in the equation. As shown, however, emient expenditure is still too high and over- the coefficient on the variable, though negative as stretched. expected, is insignificant. Ambiguity as to the output The model of optimal govemment deficit and debt effect of excess government borrowing from the has been applied to Tanzania. It has shown that the banking system may reflect the crowding in and current level of govemment expenditure is unsus- crowding out effects which are evidenced in the tainable within the constraints of the government's 87 Issues in Public Expenditure Policy in Afrka. Evidene from Tanzania's Experience average income and its economic stabilization tar- 4. See for instance Miller, op cit. gets under the structural adjustment and economic 5. For the debate see Begg (1982), ch.6. recovery programs. Hence, the prevalence of the un- 6. See Buiter 1983 and Rutayisire 1987. derfinancing of government provisions despite the 7. See Rutayisire 1988. unsustainably high level of government expenditure 8. For a review of the literature see Atkinson and corroborates the broadbased nature of government Stiglitz (1980). commitments. 9 The thrusting of the level of government expenditure Welfare improvement motivations and unex- to a higher platau of spending as a result of such a shock pected shocks partially explain the high level of gov- as war without reverting to thepre-war level has been ernment expenditure. However, the government's propounded by Peacock and WLseman (1961) as the dis- expenditure estimation procedure (costing) and its placement effect. tolerance of deficit financing have caused the gov- 10. See Buchanan and Wagner (1977); also the review ernment budget constraint to be less binding on the in Shibata and Kimura (1986). level of government expenditure. The cost price im- 11. For evidence see Rutayyisire and Mgonja (1989). pact has also been evidenced to have had a signifi- 12. Among others see Rutayisire and Mgonja (1989) cant influence on government expenditure increases. 13. Bryan and Carelton (1967). Government deficit spending has affected invest- 14. See Lucas (1973). ment in the rest of the economy. Agricultural produc- 15. Ibid. tion and private enterprise credit have been crowded 16. See Rutayisire (1990). out. Credit to agricultural marketing and public en- terprises has been crowded in.The net effect on out- References put has been shown to be indeterminate. There are several policy suggestions based on Atkinson, A.B. and J.E. Stiglitz (1980); Lectures on Public these findings. First, the sustainability of govern- Economics (McGraw-Hill). ment expenditure raises a case for further control of Begg, D.K.H. (1982); The Rational Expectations Revolution in government expenditure in the country. Overall re- Macroeconomics (Philip Allan). straints on government expenditure, such as those in Bryan, WR. and W.T. Carleton (1967); Short-run Adjust- the current macroeconomic stabilization policy ment of an Individual Bank, Econometrica, Vol. 35, No. 2, measures, need to be emphasized. This is especially (April), pp. 321-347. important in relation to eventual munetization and Bossons, J. (1986); Measuring the Viability of Implicit In- the debt service burden, and to increasing the bind- tergenerational Social Contracts in Public Finance and ing nature of the government budget constraint and Public Debt, Proceedings of the 40th Congress of the Interna- inducing prudence in the government budget. tional Institute of Public Finance (Wayne State University). Second, unless there is a dramatic increase in the Buchanan, J.M. and R.E. Wagner (1977); Democracy in Defi- government's current and expected future income, cit: The Political Legacy of Lord Keynes (New Yorkl Aca- the sustainability of government expenditure in Tan- demic Press, Inc.). zania will require reduction in government expendi- Buiter, H.W. (1983); The Theory of Optimum Deficits and ture. Measures to increase income would include a Debt, NBER Working Paper Series, Working Paper No. dramatic reversal of the inadequate performance of 1232 (Cambridge, Mass: NBER). public enterprises. Lucas, R.E. Jr. (1973); Some International Evidence on Third and finally, according to the results pre- Output-InflationTrade-Offs,American EconomicReview, sented, government borrowing from the banking Vol. 68, pp. 326-334. system can be used to some extent to obtain needed Mikesell, R.F. and J.E. Zinser (1973); The Nature of the crowding in effects. In this respect, government guar- Savings Function inDeveloping Countries: A Survey of anteesof credit to peasant producersandagricultural the Theoretical and Empirical Literature, Journal of Eco- marketing should be emphasized. nomic Literature pp. 1-26. Miller, M. (1982); Inflation-adjusting the Public Sector Fi- Notes rancial Deficit, in J. Kay The 1982 Budget(Oxford Basil Blackwell). 1. See ERP Uune 1986); Finance Minister's Budget Peacock, A.T. and J. Wiseman (1961); The Growth of Public Speech (une 1987); and Tibaijuka (1988). Expenditure in the United Kingdom (Princeton University 2. See Finance Minister's Budget Speech (une 1988). Press). 3. See Mikesell and Zinser (1973) for a review of the Rutayisire, L.W. (1990); Issues in Monetary and Credit literature. AUocation Policies in the African Context: Evidence from 88 Issues in Public Expendihtre Policy in Africa. Evidencefrom Tanzania's Epcnence Tanzania's Experience. A paper presented at the African organized by the African Economic Research Consor- Centre for Monetary Studies and National Bank of Ethio- tium (AERC), Harare, Zimbabwe. pia Seminar on Experience with Instruments of Eco- Shibata, H. and Y. Kimura (1986); Are Budget Deficits the nomic Policy, Addis Ababa. Cause of Growth in Government Expenditures? Public Rutayisire, L.W. (1988); Ways and Means of Improving the Finance and Debt. Proceedings of the 40th Congress of the Quality and Effectiveness of Public Expenditure Pro- International Institute of Public Finance (Detroit, Michi- gramming: A Case Study of Tanzania in Ways and gan: Wayne University Press). Means of Improving the Quality of Public Expenditure Pro- Tanzania, Economic Recovery Pmgramme aune 1986)., gramming, ECA/PHSB/BUD 188/7 (2.1) (a) (Addis Finance Minister's Budget Speech (1987 & 1988). Ababa: United Nations Economic Commission for Af- Tibaijuka, A.K., The Need to Monitor the Welfare Implica- rica, October). tions of Structural Adjustment Prograammes in Tanzania. Rutayisire, L.W. (1987); Measurement of Government Bu- A paper presented at the 5th Economic Policy Workshop dget Deficit and Fiscal Stance in a Less Developed Econo- (Dar es Salaam, May 1988). my: The Case of Tanzania, 1966-198, World Development, Wildavsky, A. (1964); The Politics of the Budgetary Process Vol. 15, No. 10/11 (Oxford: Pergamon Press), pp. 1337- (Boston: Little Brown and Co). 1351. Zee, H.H. (1988); The Sustainability and Optimality of Rutayisire, L.W. and G.S. Mgonja (1989); The Effect of Government Expenditure, IMF Staff Papers, Vol. 35, No. Devaluation on Government Budget Deficit: A Case 4, pp.658-685. Study of Tanzania. A paper presented at a workshop 89 10 Poverty Concious Restructuring of Public Expenditures Marco Ferroni and Ravi Kanbur Public expenditure choices play a fundamental role theextent towhich expenditures focus on, and reach, in poverty alleviation through their effects on the thepoor. Inthethird section, thelinksbetween public supply response of adjusting economies and their expenditure restructuring and poverty alleviation contribution to human capital formation. Public in- are explored in a framework which takes into ac- vestment and recurrent spending are important count human resource interactions and the multi- determinants of the qualityand quantityof economic dimensionality of the standard of living. and social infrastructurewhich, in turn, affect human The following section illustrates applications of resources and labor productivity, as well as produc- the framework with African examples. Three areas of ers' ability to take advantage of adjustment-related applications are explored. We first review a recent changes in relative prices. But government budgets attempt at deriving weights for various components have become tighter during the 1980s in sub-saharan of the standard of living (i.e. UNDP's Human Devel- Africa because of stagnating economic growth (par- opment Index). The question asked is whether the ticularly in the more easily taxed formal sector), social valuation of specified components of the declining commodity prices and a more restricted standard of living, which is implicit in the Human international credit environment. Resource con- Development Index, can be used as a guide to priori- straints and the need to control inflation by means of tize public expenditures. The second application more conservative monetary and fiscal policy are uses a production function approach to human re- motivating public finance reform in many countries sources. The question here is how basic needs inputs today. The objective of these processes of reform is (for example, health programs, education services) to bring spending more closely in line with revenue and the cost thereof in terms of public expenditure and to raise the efficiency and effectiveness of the are linked to specified achievements in human devel- government's participation in the economy in sup- opment such as levels of child mortality, life expec- port of a specified development path and re- tancy, and adult literacy. The third application fo- distributive (including poverty alleviation) goals. cuses on the analysis of consumer budgets of the The objective of this paper is to provide an overview poor. This is combined with data on public spending of public expenditure patterns in sub-saharan Africa and used to determine the degree to which govern- during the adjustment decade of the 1980s and to ment subsidies are reaching the poor in one particu- develop a theoretical framework for retargeting of lar country, Cote d'Ivoire. public spending toward poverty alleviation. The sec- The final section concludes the paper and suggests ond section is devoted to an analysis of trends in priority areas of research and data collection to sup- public expenditures, paying particular attention to port poverty conscious public expenditure reform. 90 Poverty Conious Resstrwutring of Public Erpenditures Public Expenditure Patterns in Africa During the tionary spending as a fraction of GDP declined in Adjustment Decade 1983 and 1984, stagnated during the following two years and recovered in 1987. The overall picture, In this section, recent trends in public expenditure then, is one of a set-back during 1983 and 1984 and pattems and, in particular, in spending benefitting moderate growth in the other years of the decade. the poor are analyzed. This is a necessary step toward As might be expected, GFS data point to consider- the forrnulation of recommendations, later in the able between-country variation around these aver- paper, regarding poverty conscious public expendi- age trends. In a thorough, GFS-based study of fiscal ture reform. We look at aggregate expenditures, the policy change in Africa, Sahn (1990:27-31) notes that intersectoral and intrasectoral composition of expen- discretionary government expenditures in low-in- ditures and the quality and effectiveness of programs come and oil importing countries expanded at a and services delivered by the government. much slower pace than in middle-income and oil exporting countries during the 1980s. The authors Aggregate and Sectoral Expenditures country-specific analysis also indicates that there has been little growth in discretionary expenditures in The study of public expenditure patterns in sub- C(A countries and east Africa, the mean trend of the saharan Africa is hampered by data deficiencies. lattergroupofcountriesbeingloweredbythe decline There are no comprehensive data on program expen- in Tanzanian real expenditures observed during the ditures (as opposed tobroad sectoral aggregates) and 1980s. On the other hand, growth has been more on regional and local expenditures. The IMF Govern- rapid in central and southem Africa (extreme cases ment Finance Statistics (GFS) are the only compre- in the latter group of countries are Botswana, which hensive data source on central government spend- exhibits rapid expenditure growth, and Zambia, ing. Even though the GFS data are in reality largely where real expenditures declined, but began to re- budget figures rather than actual expenditure data, cover during the second half of the 1980s). The coun- they are used in this paper for want of a published tries where discretionary public expenditures per alternative. The GFS figures are, however, comple- capita (expressed in constant 1980 US $) declined mented by evidence from the World Banks' recent during the 1980s include Kenya, Liberia, Madagas- public expenditure reviews in Africa, and this re- car, Malawi, Mauritius, Niger, Nigeria, Sierra Leone, source, as well as other knowledge makes it possible and Togo. They increased, according to GFS data, in to develop at least preliminaryconclusions regarding Burkina Faso, Cameroon, Ethiopia, Ghana, Mali, So- the poverty focus and poverty alleviation effect of malia and Zimbabwe, among other countries. public spending. Looking at the decade as a whole, rather than its It is often assumed that central government expen- first half, GFS data indicate, in Sahn's words (p. 29), ditures in Africa and spending on social services that there have been no "across-the-board reduc- (health, education) have plummeted during the debt tions' in real discretionary government spending and adjustment decade of the 1980s. GFS figures and that for most countries expenditures were 'on indicate that, on average, total real government the rise, orat least steady ... despite the proliferation spending (including interest payments) has grown of IMF and World Bank loans that often carry with during the 1980s with a flattening of the trend be- them conditions involving budgetary austerity." In tween 1982 and 1984 (see table 10-1, which was view of the 1983-84 set-back in real spending referred developed using data for 20 African countries chosen to above, it seems, in fact, that adjustment programs, on the basis of data availability for all spending which really proliferated after 1984, have helped re- categories and years considered in the table). The store real spending by lowering the effective rate of 1982-84 period was one of drought in the Region and inflation and raising foreign resource transfers and particular resource constraints which were later re- (through debt-rescheduling) discretionary budgets. laxed to a degree by an increase in donor finance and Table 10-1 shows that social and economic sector debt rescheduling in support of the adjustment proc- spending tended to increase during the decade under ess. Real per capita expenditures declined in 1983 review, both as fractions of total discretionary expen- and 1984, but grew in all other years and regained in diture and in real per capita terms. (There was a 1987 the level attained in 1982. Real discretionary decline in average economic service expenditures spending (i.e. government spending exclusive of between 1986 and 1987, but not in spending on the debt service payments) increased in the early 1980s, agricultural component thereof.) Trends in social sec- stagnated in the mid-80s (it declined in 1983and 1984 tor spending are closely correlated with trends in in per capita terns) and recovered thereafter. Discre- discretionary governmental spending in the data 91 Poverty Concious Restncturing of Public Expnditures used. While social expenditures fluctuated between tion of public spending, among other measures, as years, theywere higher in 1987 (the last year reported well as improvements in the factors which constrain in the table) than at thebeginningof the decade. This, the effectiveness of program delivery. of course, should not be taken to mean that essential social services were adequately funded. But there is Intrasectoral Resource Allocation no evidence, in table 10-1, of a decline in real re- sources devoted to social services. This expenditure Anti-poverty spending encompasses economic category does not seem to have been vulnerable to sector spending to raise the income opportunities of cut-backs in government expenditures in Africa or to the poor, 'essential' social services, and transfer pay- have borne the brunt of reductions in aggregate ex- ments or 'safety nets" to bolster the consumption of penditures where they have occurred. the chronically poor and, transitorily, the needy ip,- This supports the hypothesis formulated by Hicks ers under the adjustment process. Economic sector and Kubisch (1984) that governments tend to protect expenditures to raise the incomes of the poor and to social expenditures in times of economic difficulty. increasetheircontributiontogrowthincludeprojects In their study, social expenditures emerged as the and programs tobuild infrastructure and institutions most protected of five expenditure categories exam- in support of labor-intensive sectors and patterns of ined in 32 developing countries for the period 1972- production. Credit programs and interventions to 80. (The authors consider their conclusions prelimi- improve poor people's access to inputs and services nary, because their study is limited to consolidated and to improve the functioning and transparency of central government accounts, excluding expendi- the markets in which they trade are items of high tures of state and local governments.) Updating this priority. Agriculture (including smallholder agricul- analysis for 1979-84, Pinstrup-Andersen et al. (1987: ture) is a key sector to be promoted, despite the fact 75-81) found similar aggregate results for Africa but thatagriculturalsupplyresponsetakestime,sinceno considerable reduction in social expenditure in Latin other sector is capable of comparably participatory America. Sahn (1990: 56) found the elasticity of growth. Yet, macro price policy and development health (education) expenditures with respect to total spending have tended to favor industry, in keeping government expenditures in sub-saharan Africa to with the African post-independence tradition of in- be rather high, viz. 0.96 (1.08, respectively) for the ward-looking, industry-based thinking about devel- 1985-87 period. Elasticities were estimated for three opment. Within agriculture, public spending is often periods during 1974-87. The estimated values of elas- found to be biased in favor of state farms as opposed ticities increased with time, indicating that health to private holdings and commercial farming as op- and education expenditures were given progres- posed to the smallholder sector. Rural infrastructure sively higher priority as time passed. The elasticities development and agricultural research and exten- with respect to GDP were found to be above unity for sion are deemed to be underfunded in many African both health and education, indicating that increases countries (see, for example, World Bank 1989a, 1989b, in aggregate economic growth have led to more than 1989c). proportionate increases in government spending on Essential social services geared toward the poor the social sectors in Africa. are normally said to include primary and preventive These trends do not permit conclusions regarding health care (as opposed to hospital-based care, for the redistributive or poverty alleviation effect of pub- example) and primary, rather than higher-level, edu- lic spending. They actually appear to indicate that if cation. The focus on primary education is explained poverty alleviation effects were weak (as will be by the fact that this is the foundation of schooling, argued below), and to the extent that GFS data can that primary education is known to yield high pri- be believed, resource scarcity cannot, a priori, be in- vate and social rates of return relative to higher-level voked as the reason. Inequitable intrasectoral expen- education, and that primary enrollment rates have diture patterns, low effectiveness in program deliv- stagnated in sub-saharan Africa during the 1980s, ery and inadequate ratios of recurrent to capital and the quality of schooling has declined (World spending are the reasons why the poor capture a Bank, 1987). To judge the impact of public spending disproportionately small share of public subsidies on the poor it is, therefore, necessary to study the and services, as explained below. While increases in record regarding the intrasectoral allocation of re- financial resources to support anti-poverty programs sources and the effectiveness of program targeting are needed in Africa, raising the poverty focus of and delivery. This task cannot be carried out with governmental expenditures is also likely to require reference to the aggregate GFS, and there is no com- changes in the within and between sector composi- parably comprehensive, disaggregated source. The 92 Po"ery Concious Restructuring of Public Expenditures Table 10-1. Avaerage trends in public expenditures, 20 African countries (1980-87) (1980 U.S. dollars, indexed: 1980=100) Aggre'gatf government expendifures Discretionary as a Year Total Per capita Discretionary Per capita discretionary percentage of GDP 1980 100.0 100.0 100.0 100.0 24.7 1981 108.0 104.8 107.7 104.6 25.3 1982 117.7 111.0 115.7 109.1 25.5 1983 117.8 107.9 113.2 103.6 24.0 1984 114.0 101.2 107.3 95.2 22.8 1985 122.5 105.5 113.8 98.0 22.8 1986 129.5 108.1 117.5 98.2 22.5 Sectoral expenditures Share of discretionary Share of discretionary Per capita budget (percentage) Per capita budget (percentage) Health Economic services 1980 8.4 5.5 1980 41.6 5.5 1981 8.8 5.6 1981 38.2 5.6 1982 9.5 5.8 1982 39.7 5.8 1983 9.0 6.0 1983 37.7 6.0 1984 8.7 6.0 1984 37.6 6.0 1985 8.7 6.0 1985 39.2 6.0 1986 9.1 6.0 1986 43.0 6.0 1987 9.9 5.8 1987 39.7 5.8 Education of which agricultural services 1980 24.0 15.4 1980 13.5 8.1 1981 25.3 15.1 1981 11.9 7.6 1982 26.0 16.3 1982 13.9 8.4 1983 25.4 17.0 1983 11.7 8.7 1984 24.8 16.6 1984 12.6 8.4 1985 24.2 16.4 1985 12.5 8.4 1986 24.6 16.7 1986 13.3 9.7 1987 25.3 15.4 1987 14.7 9.5 Social services (health and education combined) 1980 32.4 19.9 1981 34.1 20.7 1982, 35.5 22.1 1983 34.4 23.0 1984 33.5 22.6 1985 32.9 22.4 1986 33.7 22.7 1987 35.2 21.2 ' See text for explanation. The countries are: Botswana, Burkina Faso, Cameroon, Ethiopia, Ghana, Kenya, Liberia, Madagascar, Malawi, Mali, Mauritius, Niger, Nigeria, Sierra Leone, Somalia, Swaziland,Togo, Uganda, Zambia, Zimbabwe. Source Caluculated on the casis of data from IMF, Government Finance Statistics (1989 Yearbook, Washington, D.C., 1989. 93 Poverty Concious Res trxturing of Public Expenditures evidence must be pieced together, and we do this on Tuming to the issue of transfers, the last of the three the basis of World Bank Sector Reports and recent broad areas of anti-poverty public spending identi- Public Expenditure Reviews, focusing in particular fied above, and focussing on food subsidies, one on social sector spending and food-linked transfer finds that govemments do not normally publish data payments. It is worth noting, however, that while regarding explicit financial subsidies. Real spending Public Expenditure Reviews contain much that is on food subsidies is assumed to have declined in relevant to our purpose, they do not focus explicitly many African countries during the 1980s (for Zam- on poverty, but on sector strategies and the implica- bia, this is documented in Pinstrup-Andersen et al., tions of expenditure levels and pattems for the over- 1987: 86), but the impact of this trend on the poor is all monetary and fiscal stance. not straightforward. It depends on the distribution In contrast with the identified priority areas of of the incidence of subsidies across the income spec- anti-poverty social spending, public expenditures on trum, adaptive responses of the poor to price in- health and education in Africa tend to be charac- creases (both on the income and the consumption terized by disproportionate resource allocation to side), and whether policies to compensate the poor higher order services which benefit the wealthier for price increases are put into effect. Since, as is well and, in particular, urban-based groups of society. For known, African food subsidies are not well targeted example, the budget of Ghana's Ministry of Health is on the poor, declines in real financial subsidies may directed predominantly at curative care for the one- not be particularly harmful to this group. Malnutri- third of the population living in urban areas (World tion is, of course, wide-spread (see Alderman, 1990, Bank, 1989d: 44). In Tanzania, hospital services ac- for evidence regarding the incidence of malnutrition count for 68% in the 1988/89 budget of the central in selected African countries, and an analysis of the Ministry of Health, whereas preventive services prevalence and determinants of malnutrition in claim 5.9% (World Bank, 1989b: 86). While this un- Ghana), calling for remedial action capable of im- derestimates the actual expenditure on preventive proving both income and human resource factors care, because hospitals also offer this kind of care (for (for example, mothers' education) which determine example, immunizations) and because the main re- nutritional status. Food subsidies clearly have a role sponsibility for preventive care is reported to be at to play among the envelope of policy options to the district level and reflected in the recurrent budget improve nutrition (Behrman, 1989: 101), but they of the Ministry of Local Government, this budget, should be narrowly targeted to the poor--a condition too, shows only a small percentage of expenditures which, for administrative and political reasons, it (4%) devoted to preventive care. Of total government may be difficult to fulfill. recurrent expenditures for health, including those A method to determine the poverty consequences channelled to district councils, only 8.8% was allo- of food price changes is proposed and illustrated cated to primary health facilities in 1987/88 (World later. It hinges on the recognition that, in order to Bank, 1989e: 86-87). reach the poor and to significantly affect their real In education, this pattern of discriminatory public incomes, government subsidies must be directed at spending finds expression in the fact that subsidies goods and services which are widely consumed by per user increase dramatically with the level of this group. To keep leakages of benefits to nonpoor schooling. Thus, public recurrent expenditure per groups small, the poor must account for a large pro- primary pupil in sub-saharan Africa has been es- portion of the total national consumption of subsi- timated at $48 in 1983 (15% of GNP per capita). It dized goods and services. The described pattern of amounted to $223 per secondary student (62% of intrasectoral resource allocation and the analysis of GNP per capita) and $2710 per tertiary student, or consumerbudgets suggest that neither of these con- eight times GNP per capita (World Bank, 1987: 141- ditions is usually fulfilled. 143). While post-primary education must, of course, be promoted, it would appear to be appropriate, on Effectiveness of Public Expenditures equity grounds, to increase students' participation in the financing of costs, combined with the devel- op- The efficiency of resource use and the effectiveness ment of an educational credit system and a sch-olar- of public service delivery have been found to decline ship fund to enable students from poor families to when government budgets contract (Gallagher and enroll. Public funds thus released could be div- erted Ogbu, 1989). This is in large measure due to the to primary education where there is normal ly be- circumstance that, in times of resource scarcity, gov- lieved to be little scope for cost recovery, but where errments tend to maintain the number of workers on there is great need to increase coverage and quality. their pay roll, and nominal salaries, at the expense of 94 Po"rty Conaous Restructuring of Public Expenditures non-personnel inputs. Nevertheless, real civil ser- investment expenditures appears to have increased vant pay has declined drastically in some African during at least part of the 1980s according to GFS countries (in Tanzania it declined by a factor of five data, although the ratio varied considerablybetween between the early 1970s and the late 1980s; World countries and years (Sahn, 1990; Table 7). The analy- Bank, 1989f: Vol. I, p. 3), leading to a crisis of motiva- sis of the implications of given functional composi- tion and ability to deliver on the part of the civil tions of expenditure and of the decline in capital service and prompting many competent officials to spending in terms of future growth and the ability to move to the private sector or engage in moonlighting deliver services requires a careful country by country activities. The frequently observed insufficiency of evaluation of the trade-offs involved. This is beyond funds for materials, operations and maintenance, the scope of this paper. Where investment expendi- and mobility of personnel has had the effect of un- tures have declined in the face of growing discretion- dermining government delivery systems in such sec- ary expenditures, additional resources to defray re- tors as transport, agricultural extension, and educa- current costs have become available. Nevertheless, tion and health. Inadequate operations and mainte- theBank's recent Public Expenditure Reviews which nance expenditures have diminished the productiv- were analyzed as part of the research undertaken for ity of past capital spending and have ushered in a this paper consistently point to a lack of funds to period (which began in the 1970s in many countries) cover the recurrent expenditures associated with es- of degradation of economic and social infrastructure. sential services. This would indicate a need to raise The implementation of donor-assisted development the budgetary resources of sectoral agencies provid- projects has been slowed down by the lack of match- ing essential services and infrastructure, while at the ing recurrent expenditures from governments. same time undertaking measures to correct imbal- In countries where data are available, wages and ances in the ratio of wage to nonwage recurrent salaries have typically absorbed from one-half to expenditure. Raising the budget of the identified three-quarters of recurrent expenditures. In Ghana's agencies for reasons of poverty alleviation is likely to health budget, for example, wages took 62% and 76% require resource reallocation, and decision criteria to in 1985 and 1986, respectively (Vogel, 1988:26). The guide this process are developed later in this paper. share of wages and personnel emoluments in Beyond inappropriate inputcombinationsand rat- Rwanda averaged 48% of recurrent spending in the ios of recurrent to investment spending, a crucial fac- health sector during 1982-88 and took over 80% in tor constraining the effectiveness and efficiency of education. Only 34% of recurrent spending in edu- public spending is the lack of adequate planning and cation was left for materials and less than 1% for management and, hence, of coherent public expen- maintenance (World Bank, 1989b). This pattern of diture strategies. Efforts to raise the productivity of budgetary allocation under adverse conditions, un- budgetary resources include the setting of realistic justifiable on economic grounds, but understandable goals (arguably a difficult task in the light of political for political and social reasons, has all but crippled pressures and rapidly expanding populations), the many government services in the Region due to explicit acknowledgment of recurrent cost implica- shortages of drugs and disposable supplies in health, tions of new projects and past investments, the ex- shortages of text books and school supplies in edu- ploration of opportunities to raise new revenue, the cation, and immobility of service delivery personnel. promotion of manpower competence in the civil The effectiveness (and the poverty alleviation im- service through training and management improve- pact) of a given evolution of real discretionaryexpen- ments, and the rationalization of donor assistance. It ditures canbe very different dependingon the differ- is now widely recognized that projects are often in- ential trend of recurrent and investment expendi- cluded in public spending portfolios because of the tures within that envelope. For example, an expan- availabilityof donorfunding, rather than theiragree- sion in discretionary expenditures coupled with a ment with articulated sectoral priorities. Similarly, rapid decline in the ratio of recurrent to capital ministerial management capabilities are taxed to spending could exacerbate the shortage of funds for theirlimitby the diversityof projects and partners. In nonwage recurrent inputs referred to above. This Zambia's agriculture sector, for example, there are was probably the case in Cameroon in the latter part about 150 projects supported by more than 20 donor of the 1980s since the average ratio of recurrent to agencies, each maintaining its own procedures and capital expenditures fell from 2.3 to 1.3 between 1978- requirements. "Overlaps and conflict between the 80 and 1986-87 in that country, while real wage ex- objectives of different projects are commonplace" penditures rose (Sahn, 1990; Tables 7 and 8). In most (World Bank, 1989a: Vol. II, p. 29). The solution to the African countries, however, the ratio of recurrent to problems arising from donor proliferation is, of 95 Poverty Concious Restructuring of Public Expenditure course, not to suggest a cut-back in donor assistance, Restructuring of Public Expenditures for Poverty but its progressive subordinat ion to government Alleviation: a Theoretical Framework priorities in the 1990s, although it is recognized that these priorities may not yet be fully spelled out in Our analysis of public expenditure patterns in the some countries. previous section focused on two policy variables, viz. the flow of funds to certain sectors and programs, Conclusion and the factors which determine the efficiency and effectiveness of expenditures (i.e. the functional com- Our discussion suggests that the poverty allevition position of expenditures and public expenditure impact of public expenditures in sub-saharan Africa management). In this section we will focus on the is low because recurrent expenditures are under- first of these variables. We hope to contribute to the funded, government subsidies tend to be directed to formnulation of poverty sensitive public expenditure higher order services and commodities not widely strategies through an analysis of public expenditure consumed by the poor, and because management choices and effects. This is complicated, among other problems and suboptimal input combinations are reasons,by thefact that thevarious dimensions of the diminishing the efficiency and effectiveness of ser- standard of living of households and individuals are vice delivety. This is also the conclusion of the Wor- affected individually and jointly by the different Id Development Report 1990, on poverty, which components of a govemmental spending program. finds that government spending "tends to be skew- Thus, public expenditure restructuring will have ed away from the people who need it most--the poor" complex effects with many interactions, and difficult (World Bank 1990a, p. 76). An important assumpt- choices will have to be made from the various trade- ion implicit in this conclusion is that the 'poor' are offs. The purpose of the analysis that follows is to separated from the "nonpoor" by a poverty line de- develop a framework in which some of these trade- fined as the level of income which cuts off, say, the offs can be clarified. lower 30% 40% of the income distribution. While this The multi-dimensionality of the standard of living is not the place to discuss the measurement of pov- hastobe faced head on in assessingtheconsequences erty or the conceptual and practical difficulties inher- of public expenditure restructuring. The standard ent in the notions of relative and absolute poverty, it focus on income or expenditure based measures of is obv- ious that if everybody, or a vast majority of welfare must be complemented by the concept of the population, were classified as being poor (it is 'basic needs" which was introduced in the 1970s. sometimes suggested in Africa that thisbe done), our The "basic needs" literature stresses a number of conclus ion regarding the poverty focus of public indicators (in particular, life expectancy, literacy, expen- ditures would not hold. The definition of a health, nutrition, and housing) as being complemen- poverty line which helps policy makers focus on the tary, if not superior, to the usual income /expenditure lower end of the distribution of income is a prereq- indicator, and argues for a strategy to increase the uisite for the fonnulation of public policy, including values of these indicators. A closer reading reveals public expenditure reform, seeking to alleviate pov- several arguments (sometimes not clearly separated) erty. for a basic needs strategy: If poverty alleviation is accepted as an explicit (i) Thinking of basic needs requirements as enter- objective of public expenditure strategies, it is neces- ing the standard of living directly, it is argued that the sary to teview and redefine the sectoral, intrasectoral standard of living of the poor can be raised more and functional distribution of funds in the three areas efficiently by focussing on basic needs. There are, in of essential economic and social services and infra- turn, two sub-arguments here. One is that for any structure, and transfer payments, identified earlier. poor person a dollar spent directly on basic needs It is, however, not obvious how this may be best will be better than a dollar spent directly on income done, because human capital inputs interact with raising (which will then, indirectly, influence basic each other and with directly income-enhancing eco- needs). Another is that basic needs spending could nomic sector expenditures, and these interactions are be better targeted toward the poor. difficult to quantify. Forexample, what is the poverty (ii) Thinking of basic needs achievements as being alleviation effect of an additional dollar spent on inputs into income generation, it is argued that the agricultural extension for women farTners compared rate of return to such investment is higher than that, with an additional dollar spent on primary health for example, in directly productive physical capital. care? A framework to analyze this problem is devel- Thus, even if income/expenditure were the ultimate oped in the next section. aim, a basic needs strategy is superior. There is once 96 Poverty Concious Restmuturing of Public Ependitures again a targeting argument to supplement and bol- Let us start by highlighting (i) - (iii) and focus, ster the basic case. therefore, on a typical (not necessarily poor) house- The 'social sectors" in a governmental expend- hold or individual. To put some of these issues into a iture program display all of these considerations, but common framework, consider the following simple they lie along a spectrum. In the case of housing we model. Basic needs achievement B is a function of come closest to the pure consumption end of the social expenditure E and income Yas follows: spectrum (at least, we have not seen studies that argue for housing in terms of its productivity enh- (10-1) B aO + aE E + ay Y ancing properties). In fact, many studies on poverty do indeed monetize housing consumption by im- Income on the other hand is a function of basic putatian through hedonic regressions and the like. needs achievement and "productive" expenditure I: Education is perhaps at the other end of the sp- ectrum, where the literature concentrates primarily (10-2) Y - (, + (BB + PII on its productivity aspects rather than on the con- sumption value. In principle, the decision rule for There is a budget constraint public resource allocation is straight forward-it de- pends on rates of return to different levels of edu- (10-3) G =E + I cation. Health and nutrition are located toward the mid- and the valuation of B and Y to give the "true' stand- dle of the spectrum. They enter the standard of living ard of living W is, directly as well as indirectly through productivity effects. The well-known controversy about the rela- (10-4) W = Yo + yBB + yyY tive merits of growth oriented versus basic needs oriented public action in Sri Lanka (Bhalla and The govemment faces the choice of restructuring Glewwe, 1986; Sen, 1981; Isenman, 1980; Anand and by changing the balance between E and I. In which Kanbur, 1990) can be seen to take the first route. All direction should it move? In order to answer this participants in the debate have agreed that infant question let us first solve (10-1) and (10-2) to give the mortality, for example, should be reduced because values of B and Y for any given values of E and I: this is a desirable goal by itself. The only question is + aEE + ay Pi whetherthisisbestachieved throughincomegrowth (1-5) Be I0++ aypB or through basic needs intervention. An issue which 1 - aY (3 is not addressed in the literature just cited is the feedback of improved health and nutrition on in- (10-6) Y . o + P8 ca + -- aE E + (3B I come growth. Another qusetir. wH- : not rAicd d -, aY P although it is implicit in the discussion, is whether Substituting these into (1-4) we get direct intervention (i.e. "social sector" spending) can be better targeted towards the poor. In this context, (10-7) W= - + ('E - cY )B however, it should be noted that the entire discussion 1 - aY ,B on Sri Lanka is based on national average indicators of basic needs fulfillment-the question of basic + a(Y y.B + yY PB atE) E + 6yB cw (3j+ yy PB) 11 needs fulfillment of the poor can therefore be ad- 1- QY(3B dressed only indirectly, if at all. At the margin, therefore, the choice between put- There seem, then, to be (at least) four complicating ting one more dollar in E versus I depends on the aspects of public expenditure restructuring- (i) One comparison component of public expenditure can affect several aspectsofthestandardoflivingofa typical househo- (10-8) aE(y+ yYPB) >< PI(Yy+ ysay) ld; (ii) Each component of the standard of living can affect other components of the standard of living for The comparison depends on a combination of pro- a typical household; (iii) The different components of ductivity and valuation considerations. Forexample, the standard of living have to be valued relative to suppose that we were interested only in basic needs each other for a typical household; and (iv) Some achievement, so that y, = o. Then the choice between components of public expenditure when passed E and I depends on through (i) - (iii) above, are more effective than others in raising the standard of living of poor households. (10-9) aE>< Pi ay 97 Poverty Conious Restructuring of Public Expenditures In other words, it depends on a comparison of The above discussion highlights some of the inter- direct versus indirect effects. This comparison is at actions between different categories of public expen- the heart of the debate on Sri Lanka's policies in the diture in their impact on the standard of living of a 1960s and 1970s. Anand and Kanbur (1990) have typical household or individual. We now turn to the estimated the relationship (1) for Sri Lanka on time extent to which the policy actions affect the poor. As series data, and conclude that for the infant mortality noted earlier, an implicit argument in some of the rate the direct effects of social expenditure greatly literature is that certain types of expenditure are to outweigh the indirect effects operating through in- be preferred because they are or can be better tar- vestment in income earning opportunities. But this is geted. This argument needs tobe made explicit, tobe a country specific finding which may not apply to made precise, and to be quantified. Africa. Suppose that there are n units in the economy, If basic needs had no productivity effects, so that indexed by i = 1, 2, ..., n. Out of the total expenditures PB = 0 then (10-8) collapses to E and I, let Ei and Ii be the amounts that reach unit i. Clearly, (10-10) aE YB > < 01 [YY + YB ay n or (10-12) Ei- E;I Ir. I (aE- PI ay) YB > < PI YY i-1 Thus a prerequisite for basic needs expenditure to The individual counterpart to (10-7) is thus given by be worth considering at all is that direct expenditure + TB (ao + a Po) + yy (P + PB ao) leads to a bigger effect (aE) on basic needs than the (10-13) WV? +Y 1-ayps indirect effect through income increase (pi yy). This is the Sri Lanka controversy again, but now there are YB a + YY PaE extra considerations in YB and yy. Even if the direct + Y a ) E effect was greater, the social valuation of basic needs 1 - aY ,> achievement would have to be sufficiently high for B aY Pi+ yy Pi basic needs expenditure to be worthwhile. + 1 - ay , B Finally, if basic needs had onlyproductivityeffects and there were no feedbacks via income to basic Equation (10-13) gives the effect on individual needs, so that ay = 0, and basic needs were not income of expenditures Ei and i reaching individual valued for themselves but only for the income they i. We now need to tormalize a focus on poor units, so generate, so that YB = 0, then (10-8) collapses to as to gauge the impact of restructuringon them. This is not the place to discuss in detail the drawing of (10_l ) CtE PB > < PI poverty lines or the formulation of poverty indices. There is now a large literature on this. Suffice it to say This is a straight-forward marginal productivity that given a poverty line z which delineates poor comparison between expenditure on education, say, from nonpoor, one index that is becoming quite com- and expenditure on other production activities. A monly used is that put forward by Foster, Greer and unit of expenditure diverted from these other activi- Thorbecke (1984). This is defined by: ties has opportunity cost Pi in terms of income fore- q gone. But it leads to an increase in education, which (10-14) Pa- j ( z - Wi)/Z ]; a2 0 in turn leads to an increase aE PB in income. i-i Thus we see that many of the strands of the argu- When a = 0, this index turns into the commonly ments surrounding basic needs fall out as special used head count ratio or incidence of poverty. When cases of (10-8). Moreover, the above analysis can be a = 1, it measures the normalized "poverty gap." As applied to the choice between any two categories of the value of increases, more and more weight is given public expenditure regardless of the level of sectoral to the poorest of the poor. This family of measures or intrasectoral aggregation-health and education, has proved useful in operationalising poverty meas- health and economic infrastructure, primary and urement, while allowing us to represent a range of higher levels of eduction. While the technical pa- valuejudgementsthroughtheabilitytovarythepar- rameters a and P are necessary and not easy to esti- ameter. mate, what seems to stand out is the need to arrive at Suppose now that a marginal budgetary shift from clear social weights y between different dimensions E to I occurs in the aggregate. How does this feed of the standard of living. through to individuals? One might entertain differ- 98 Poverty Connous Restnrctunng of Public Erpenditutrs ent possibilities. One is that individuals gain or loose sion as opposed to another in the social valuation of in proportion to their current levels of E and I. Let the standard of living, (ii) those that quantify the each Ei become links between public expenditure and achievements E1 (1+ 8 ) and each Ii become Is (1- a). Clearly, from along several dimensions of the standard of living, the budget constraint (10-3): and (iii) an assessment of what fraction of public expenditure in any given category reaches the poor. (10-15) OBE = al We now consider each of these parameters in turn. Totally differentiating (10-14) and using (10-15) we Social Weights on Different Dimensions of the Standard get of Living q (10-16) dPai - n 2 (/z) [(z-W*Yz]-'CE EidB- CiIi da] Attachingweights to the various dimensions of the i,l standard of living is a normative question. The topic q of weighting different outcomes in, say, health, edu- = -_ (-/z) [(z- W-)/z]--' [CE (Ei/E)-Cl (1l)] cation and income needs considerable thought, not i-i least from policy makers, so as to arrive at coherent Where, ways of assessing the normative evaluation of alter- CE YB aE + YY PB aB native policies. Otherwise, every policy can be justi- fied on some weighting or the other. C YB aY PI + Y PI Developing such a weighting is not an easy task, NE - aY AB and attempts to do so are fraught with danger. One 1 - ay PB recentattemptisthatofUNDP (1990),whoseHuman As might be expected, the impact on poverty de- Development Report advances a Human Develop- pends on the current shares of total expenditure of ment Index (HDI) as a measure of achievement that each type reaching the poor. Thus, when a =1, we get incorporates income and non-income factors. The HDI is defined as follows. First, we specify three (10-17) dP1 _ - (E/nz) [CE( EP/E) - C1 ( IP/I )J relevant indicators at the national level as its compo- nents-life expectancy (XI), literacy (X2), and the Where EP and IP are expenditures of each type reach- logarithm of real CDP per capita (X3). Looking across ing the poor. Thus, with this framework the "target- a range of countries, the maximum and minimum ing" case for certain categories of expenditure relies value for each indicator is established. A 'depriva- on high values of ratios of the type Ep/E and P/I. tion' index for the ith indicator and the fh country is These are in principle verifiable and quantifiable. then defined as Toward Implementation: Some African 4ij= (max1 Xri - Xtj) Illustrations max1 Xy - minj Xq) Clearly, l4 lies between 0 and 1. UNDP (1990) then It is convenient to think of the problem of poverty defines the deprivation index for country j as a sim- conscious public expenditure restructuring as fol- ple average of the three deprivation indexes for the lows. For a given total of public expenditure, the country: policy instrument available to us is to alter the com- 3 position of expenditure between relatively broad sec- (IA) ' Iq tors and programs within sectors. While the answer to the question 'how broad?" is country specific, we The Human Development Index" is then defined as: would like to retain the sense that we are not here discussing very fine micro management of individ- (HDI)j = 1-I ual programs. While this is important, our task here is to change the pattern of resource flows at a more The reader is referred to the original source, UNDP aggregative sectoral or intrasectoral level. Given the (1990), for the calculated values of the HDI. It is worth current structure of utilization of these resources, we noting that, in the report, 44 countries are defined as wish to determine how best to allocate expenditure having a "low" level of human development or an across different categories. The theoretical analysis HDI of less than 0.5. Of these, 32 are in sub-saharan highlights three sets of parameters that are crucial: Africa. However, while the HDI is an interesting (i) those that quantify the importance of one dimen- attempt at arriving at a unidimensional measure of 99 Poverty Condous Ratructaring of Public ?penditures achievement, it has to be viewed with caution. First, one can make for children's health. But quantitative the normalization assumption seems problematic. In estimates of the effect are rare. Would a dollar di- effect the index views achievement relative to the verted from primary health care to primary educa- best country in the sample. Thus, if Japan's life expec- tion for girls really lead to a long term decline in child tancy were to fall, Kenya's HDI would go up! That is mortality? If so, by how much? Within the health an odd sort of index to have. Second, even if we set sector, would a dollar diverted from drugs in rural aside this normalization problem, the index essen- health centers to improvingqualityof health person- tially gives equal weight to achievement along the nel in these same health centers lead to an improve- three dimensions. Is this an accurate reflection of ment in measured health indicators? If so, by how value judgements, and is it even clear what value much? Equally difficult questions lurk in the nutri- judgements this implies? tion-health area. The complementarities here are well To see how one might set about developing an recognized but it is their quantification that is prob- index from a coherent set of value judgements, and lematic and lagging behind. If the dreadful choice only as an illustration, let an individual's income at between maintaining a nutrition program or a vacci- time t be yt and his utility of income U(yt). Then nation program has to be made, what are the trade- his/her expected lifetime utility at birth is: offs? A badly nourished person is more likely to develop complications from a disease. But how much W- E I U(y,)e-dt more likely? 0 These are not questions to which we have answers, where f is a discount rate, T is lifetime (a random but we doubt that policy analysts have often enough variable) and E is the expectation operator. Then, if posed the question in this way, so as to force the lim- T is exponentially distributed with parameter X, ited econometric analysis that exists to speak to these ,, issues. One notable exception is a recent Bank study w . f u (y,) e- (6+ X }' dt on 'Health Care Cost, Financing and Utilization" in Nigeria (Vorld Bank, 1990b). This studyconcentrates Moreover, if yt = y for all t and U(y) = ln y, then on the problems and prospects for cost recovery in L health care, using as an example Ogun State in Nige- W, 1 * L [1ny] ria. Based on a household survey, the study first I 4:6L estimates the demand for health care in its various where L = 1 /k is the expected lifetime. Thus, if = 0, dimensions as a function of individual variables (in- we get come, education, etc.), price of care, as well as quality W = L. [In y] variables (for example, the availability of drugs and the physical condition of facilities). In this health The above expression is related to (at least two ele- production function (or reduced form demand ments of) the HDI, but is nowhere near it in actual- analysis) the frequency of visits to different types of ity-country rankings could be vastly different as health facilities is taken as the dependent variable. between the two. However, the index just derived at Such studies, on the intermediate or proximate de- least has the virtue of making its value judgements terminants of achievements are common enough. clear and transparent. With the HDI it is not clear Two recent studies of this type on the determinants what the judgements underlying the very precise of nutritional achievements in Africa are by Alder- implied tradeoffs are or where they came from. man (1990) and Sahn (1990). The explanatory variab- Thus, there does not seem to be a basis for recom- les in these studies include parental education and its mending the HDFs weighting of welfare compo- effect on children's nutritional status as measured by nents as a guide to public expenditure reform. anthropometric achievements. What is unknown in this approach is the influence of (past) public expen- Public Expenditures and Welfare Achievements diture strategies on today's level of parental educa- tion. The second consideration mentioned above (i.e. Taking frequency of visits to modem (private and the quantification of the links between public spend- public) health centers as the policy variable of inter- ingand welfare achievements) calls foran evaluation est, the Nigeria study then simulates the impact of of complementarities (or substitutabilities) between various cost-recovery related policy changes on the different types of interventions in affecting relevant outcome. Examples of this might be (i) an increase in outcomes. For example, it is often argued that im- the price of health services provided by the public proving mothers' education is the best investment sector, (ii) an increase in the price plus an improve- 100 Poverty Concious Restructuring of Public Expenditures ment in quality of facilities (suitably defined), and the total amount consumed in the economy. The (iii) improvements in drug availability, and so on. impact on the reduction of the poverty gap is (to a For 15 such policy options the study simulates the firstorderof approximation) proportional to the total impact on frequency of visits and on the public amount consumed by the poor. Hence the critical budget. The study illustrates what is required before ratio is that of the latter to the former--a poverty expenditure restructuring analysis can be done even conscious, and balanced budget, restructuring would on one quite narrowly defined outcome-frequency mean spending more on commodities that had a of visits to health centers. higher value of this ratio. While the simulation of the impact of these policy The above analysis can be complicated consider- changes on the frequency of visits uses the estimated ably, but the rule of thumb developed here has the demand for health functions discussed above, their virtue of simplicity. It can be applied to household impact on the public budget requires a different type income/expenditure surveys that disaggregate con- of analysis: inparticular, thecostsof qualityand drug sumption by commodities and permit the estab- availability improvements have to be estimated. This lishment of a poverty line. Let us start by analyzing requires a detailed cost analysis of the different com- each household's accounts in terms of sources of ponents of the public health system, including pers- income and destinations of expenditure. In fact, total onnel, drugs, and physical infrastructure. While this expenditure (per capita) will be our measure of indi- is presented in the Nigeria study, this type of analysis vidual welfare. The destinations of expenditure can is typically not available for African countries. It is be broken up into several basic categories: for exam- precisely this direction of research which is needed ple, consumer expenditure on food, consumption of to complement the growing literature on estimates of home produced food, consumption of home pro- household level health production functions. duced nonfood items, other consumption expendi- ture including nonfood items, and remittances paid Public Expenditures and the Consumption Patterns Of out. Each of these can in turn be disaggregated fur- the Poor ther to any level that the data will allow and the policy analysis requires. Thus, if the price of kerosene The third component of the framework developed is an important policy instrument this should appear in Section 3 is the assessment of what fraction of as a separate category in the disaggregation of con- public expenditure of a particular type is reaching the sumer expenditure on nonfood items. Similarly, if the poor. It may be surmised that this is more easily price of rice looms large in policy discussions of analyzed than the questions asked in the two preced- consumer subsidies, these should appear as a sepa- ing subsections. This is in fact correct, but (like the rate category. In any event, given a mutually exclu- estim,ation of the production function of basic needs sive and exhaustive categorization of expenditures, satisfaction referred to above) it relies on the avail- and given a poverty line, an expenditure decompo- ability of household survey data of a type that is not sition matrix can be drawn up. yet widely available in African countries. We will Under the total in each cell should be the per- illustrate the analysis with three examples from Cote centages it represents of the column total and of the d'Ivoire, a country which has undergone dramatic row total. These percentages are highly relevant for economic difficulties in the 1980s, and has been the analysis of the poverty impact of food subsidy forced to reconsider the level and composition of its changes. But they have to be used with care. Often public expenditure. one finds analysts using column percentages to cl- Let us start the discussion with a consideration of aim that the impact of a food subsidy reduction will consumer price subsidies and their impact on the or will not be large on the poor. Thus, if a particular public budget and on the poor. As between subsidies commodity accounts for a small fraction of poor on one commodity or the other, it can be shown that households' total consumption (i.e. a low column in expression (10-17) CE = Ci, so that in a poverty percentage), it is argued that this is an attractive conscious restructuring of subsidies the government commodity for a subsidy cut. However, this does not should move towards commodities that have a high take into account what the saving will be on the fis- value of the ratio of poor people's consumption to cal deficit account-which is why the subsidy is national consumption. This is shown formally in being cut in the first place. An appropriate question Besley and Kanbur (1988) but is intuitively obvious. to ask is: What is the povertyimpact perunit of defic- A unit reduction in the national subsidy on a com- it reduction? For this it is the row percentage that is modity saves the budget an amount proportional to relevant, as argued above. Of course one can make 101 Poverty Concious Restructuring of Public Expendihtres the analysis much more complicated, but in the op- Table 10-2. Rice and poverty in C6te d'Ivwire erational context a matrix such as the one suggested 1. Incidence of poverty among 35.7% here should prove useful as a first cut. farmers Often a price change will have an effect on produc- 2. Ratio of mean area of poor to 80.0% ers of the commodity also. A similar strategy can be mean area of all applied to the sources of income. Starting off from 3. Ratio of rice consumption by poor 8.7% very broad categories (employment income, agricul- to total tural income, nonfarm self-employment income, re- 4. Ratio of food consumption by 13.4% mittances received), we can disaggregate down as far poor tp total food consumption as the data will allow and the analysis requires. Em- Source: Kanbur (1990) ployment income can be further broken down by production sectors and agricultural income by crop. Nonfarm self-employment income can also be fur- the consumer price of rice and the price of food in ther disaggregated to production sectors where rela- general, rice would be the prime candidate. Likewise, tive price changes are occurring. where price reductions are concerned, food in gen- An illustration of the above analysis is the case of eral is preferred to rice. Similar indicators can be the price of rice in C6te d'Ivoire. From the Living Sta- calculated for other foods as policy issues emerge ndards Measurement Survey (1985) for C6te d'Ivoire, around them during adjustment. The theory devel- Kanbur (1990) calculated the indicators in table 10-2. oped in Besley and Kanbur (1988), and implemented The details of the calculations are given in the origi- here for rice, which is highly relevant to the Cote nal source and are not our concem at the moment. In d'Ivoire policy dialogue, has wide applicability and the third C6te d'Ivoire Structural Adjustment Loan can be utilized for other countries as comparable or Agreement (1986) explicit mention is made of the better data begin to become available. need tobring domestic prices of rice more in line with Another illustration of how household survey data international prices. The efficiency based welfare can be used in assessing public expenditure rest- economics of such policy refonn is well known--the ructuring priorities is in the area of housing in Cote sum of producer and consumer surplus increases d'Ivoire. In the Second Structural Adjustment Loan relative to the distorted equilibrium. However, there Agreement (1983), the government announced a ma- are distributional implications depending on how jor disengagement from the housing sector, noting prices are changed and in what direction. explicitly that it would pass on costs to public sector Table 10-2 shows that the incidence of poverty tenants in rental housing. What might be the conseq- among rice fanners is 35.7% (compared to 30% for all uences of this for the poor? Table 10-3 below prov- Ivorians). Rice producers thus have a special claim in ides some information on this. Using data from the a policy of poverty alleviation. The mean area farmed LSMS, it gives an indication of how much of the pub- by poor rice farmers is lower than that for all rice lic expenditure on rent was in fact going to the poor. farmers. Combining 1 and 2 we find that the ratio of The table shows that while out of all individuals in land farmed by poor farmers to total land farmed is the sample who rented, 27.3 percent rented from 28.6 percent. While the relevant ratio suggested by SICOGI/SOGEFIHA or some other public agency, theory is that of rice production by the poor to total only 6.9 percent of the poor did so. In fact, out of the rice production, the above is an adequate proxy. This 730 individuals in the sample who rented from these ratio is to be compared with ratio 3 in table 10-2. It agencies, only 14 were below the poverty line. This is seen from these that rice is notreallythepoorman's seems fairly conclusive evidence that the poor have food. Only 8.7 percent of total consumption is ac- notbenefited from the operation of SOCOGI/SOGE- counted for by the poor. Thus from the point of view FIHA; and table 10-3 goes on to confirm this by of poverty targeting, if the choice is between reduc- showing that of those for whom rent is subsidized or ing producer price or increasing consumer price, the paid by someone else, this subsidy comes from a indicators suggest that it is the latter which will do public agency 80.5% of the time, but there were no least damage to poverty at the aggregate national poor who fell into this category in the sample. level. Similarly, if the choice is between increasing Thus, the disengagement of the government from the producer price or reducing the consumer price, its present activities in the housing sector is unlikely the former has priority in a poverty focussed strategy. to be detrimental to the poor given that current inter- Finally, compare 10-3 and 10-4. Relative to food on vention is largely in favor of the top 70 percent of the average, rice is decidedly the consumption of the population and not the bottom 30 percent. However, non-poor. Thus, if the choice is between increasing a caveat is in order. It remains true that for the vast 102 Poverty Concious Rcstructuring of Public Expenditures Table 10-3. Rental housing characteristics by On the basis of the figures in table 10-5, Glewwe poverty group in C6te d'Ivoire and de Tray (1988) conclude as follows: Poor-AU " ...reductions in funding for university education Poor A]Z will have very little effect on the poor. If these funds 1. Own house (%) 91.9 74.4 were instead used to improve the quality and avail- 2. Of thosw who rent, rental 6.9 27.3 ability of primary education instruction, the poor from SICOGI/SOGEFIHA/ would likely receive substantial benefits. In terms of public agency scholarships received, the impact is roughly the same 3. Of those who rent, those for 6.9 12.8 among the poor and the non-poor measured relative whom rent iS paid by to household expenditure levels. However, since the someone else (%)welhehoshlshvmuhhgrexndte 4. Of those for whom rent is paid 0.0 80.5 wealthierhouseholdshavemuchhigherexpenditure by someone else, payment by levels in per capita terms, much more scholarship SICOGI/SOGEFIHA/other money is going to wealthier households (whose public agency members are more likely to be in school at high levels Source: Glewwe and deTray (1988) of education) than to poor households. Cutting scholarship money across the board will not dispro- majority of Ivorians, even more so for poor Ivorians portionately affect the poor, while targeting that in the rural sector, rental housing is not a concern. It money to improve the quality of primary education isthequalityofhousingandamenitieswhichmatter, is clearly to their advantage. Overall, policies for and it is to these that we now turn. Glewwe (1986) funding changes in education are more likely to has analyzed these in some detail. He notes that the benefit the poor than hurt them." amenities of the poor are relatively worse on every The above three examples, in the area of food count. Relatively more of them have no toilet access subsidies, housing and education, illustrate the role and of the people who do have such access, very few that comprehensive household surveys in Africa of the poor have access to a flush toilet or toilet inside could play in identifying poverty conscious public the house. Similarly, the source of drinking water is expenditure restructuring possibilities. The collec- a well (with or without pump) for most poor people tion of such data should clearly be a priority. and the next most important source is river, lake, spring, or marsh. These figures lay out the need for a Sunurnay and Conclusion restructuring of expenditure away from rental sup- port to programs of improvement of amenities to the In the foregoing pages, recent patterns of public poor. expenditures in sub-saharan Africa were reviewed, Finally, we look at the example of education schol- an analytical framework to guide marginal resource arships in Cote d'Ivoire. Table 10-4, adapted from allocation between spending alternatives was devel- Glewwe and de Tray (1988), illustrates the clear case oped, and the scope and some limitations of poverty for restructuring public subsidies away from gener- conscious public expenditure restructuring analysis alized scholarships. were illustrated by means of selected applications. It was concluded, that the poverty alleviation record of Table 10-4. School attendance and scholarships past public expenditure strategies is rather limited. in Cote d'Ivoire However, decisions regarding the reallocation of re- sources to directly poverty-reducing spending pro- 30% 70% All grams are complicated by the fact that the various Household members human capital inputs interact with each other and Househdi membeo s with productive sector expenditures targeted on the attending school Primary 11.7 19.7 17.3 poor in producing welfare outcomes. How is one to Secondary 2.4 8.4 6.6 prioritize the use of scarce resources for poverty re- duction? University 0.1 0.7 0.5 The answer suggested was that, given the multi- Scholarship money dimensionality of the standard of living, the choice Household 0.8 0.6 0.6 between social and economic sector resource alloca- expenditure tion at the margin (or indeed between any spending member (CFAF/year) 467.3 1417.6 1137.9 alternatives regardless of their sectoral affiliation and level of aggregation) depends on the comparison of Source: Glewwe and de Tray (1988) total expected welfare effects working through both 103 Pozvrty Condous Restncwuring of Public Expendituns basic needs and income mechanisms. A distinction The collection of household budget and consump- was made between direct and indirect expenditures, tion data should therefore be promoted. The reader the former taking place in the primary sector in is referred to World Bank (1991) for the description which one seeks to obtain improvements (for exam- of a prototype household survey which is designed ple, health), thelatteroccurringinother(forexample, to capture the information needed to perform the income-enhancing) sectors which will lead to im- typesof analysisidentified above. Onthesupplyside provements in health through interactive effects. The cost data on specified services and their input com- analysis implied that there is normally a case for a ponents, and data on service utilization, are requir- combination of direct and indirect expenditures. Ap- ed. Cost data are needed both to link expenditures proaches to theanalysis of the povertyalleviation im- and achievements and as abasis for the improvement pact of public expenditures were discussed. Three of public expenditure management. The collection pieces of information were identified as needed, viz. and analysis of these demand and supply side data the weight to be attached to various components of is a demanding task, but (as we hope to have dem- the standard of livin& estimations of the linkagesbe- onstrated) well worth the effort in terms of the in- tween expenditures and achievements, and knowl- sights gained to guide the restructuring of public edge regarding the fraction of expenditures reaching expenditures from the point of view of poverty re- the poor. The social valuation of the various compo- duction. nents of the standard of living is a normative matter. We have therefore recommended caution in the use References of weighting schemes, such as UNDP's Human De- velopment Index, as a basis for decisions reg- arding Alderman, H. (1990): 'Nutritional Status in Ghana and Its changes in the composition of public expenditure. Determinants,' Social Dimensions of Adjustment Work- However, much is to be gained from production ing Paper No. 3. World Bank. function approaches to the analysis of welfare ach- Anand, S. and R. Kanbur (1990): 'Public Policy and Basic ievements and from analyseswhichlinkpublicexpe- Needs Provision: Intervention and Achievement in Sri nditures and the consumption patterns of the poor. Lanka,' forthcominginJ. Dieze and A.K. Sen (eds) Hun- For example, an individual's health production func- ger and Public Policy, Oxford University Press. tion can be said to include, among other arguments, Behrman, J. (1989): 'Interactions Among Human Re- a set of health inputs (medical consultations, preven- sources and Poverty: What We Know and What We tive care, availability of health facilities) and a set of Don't Know,' PHR Department, Processed (February) household "public" goods such as sanitation fadli- World Bank ties, water quality and the like (World Bank, 1990c: Besley, T. and R. Kanbur (1988): "Food Subsidies and Pov- 91). These inputs are a function of public expenditure erty Alleviation,' Economic Journal. levels and patterns (unless they are provided privat- Bhalla, S.S. and Glewwe, P. (1986): 'Growth and Equity in ely), and their contribution to health outcomes can Developing Countries: A Reinterpretation of the Sri be estimated on the basis of appropriately designed Lank an Experience," World Bank Economic Review, Vol. 1, household surveys. The availability and prices of No. 1. medical inputs and household "public goods" are Foster, J., J. Greer and E. Thorbecke (1984): 'A Class of determinants of the demand for health. Changes in Decomposable Poverty Measures," Econometrica. public policy regarding these determinants (e.g. pric- Gallagher, M. and O.M. Ogbu, 1989. Public Expenditures ing or cost recovery policy; the expansion of health Resource Use and Social Services in Sub-Saharan Africa. facilities) will thus affect health care choices as gov- Draft, Africa Region Technical Department, The World erned by the parameters of the health production Bank: Washington DC. function. If combined with studies of the cost of Gertler, P. and J. van der Gaag (1988): "Measuring the public services and their components, this demand Willingness to Pay for Social Services in Developing side analysis allows one to trace changes in welfare Countries," Living Standards Measurement Study, achievements (or proxies thereof) to changes in pub- Worldng Paper No. 45. Washington DC: World Bank. lic expenditures as exemplified in World Bank Gertler, P., L. Locay and W. Sanderson (1987): "Are User (1990b). Fees Regressive? The Welfare Implications of Health In conclusion, then, it is noted that household sur- Care Financing Proposals in Peru," journal Econometrics veys are a prerequisite for policy research on the 36, pp. 67-80. determinants of basic needs satisfaction and the de- Glewwe, P. and D. de Tray (1988): "The PoorDuring Adjus- gree of overlap between the pattern of government tment: ACaseStudyofC6ted'Ivoire," LivingStandards subsidies and the consumption habits of the poor. Measurement Study, Working Paper no. 47, World Bank. 104 Pov"rty Condous Restucturing of Public Expenditures Hicks, N. and A. Kubisch (1984): "Cutting Government WorldBank(1987): "EducationinSub-SaharanAfrica: Po- Expenditures in LDCs," Finance and Development, Sept. licies for Adjustment, Revitalization, and Expansion," A Isenman, P. (1980): "Basic Needs: The Case of Sri Lanka, World Bank Policy Study. Washington DC: World Bank. World Development _ (1989a): Zambia: Public Expenditure Review, Wash- Jaeger, W. and C. Humphreys (1988): "The Effect of Policy ington DC: World Bank Reforms on Agricultural Incentives in Sub-Saharan Af- _ (1989b): Rwanda: Public Expenditure Program: An rica, American Journal of Agricultural Economics 70, pp. Investment in Economic Strategy. Report No. 7717-RW, 1036-43. Washington DC: World Bank. Kanbur, R. (1990): "Poverty and the Social Dimensions of _ (1989c) Mozambique: Public Expenditure Review, Adjustment in C6te d'lvoire," Social Dimensions of Ad- Washington DC: World Bank justment Working Paper No. 2, World Bank. . (1989d) Ghana: Population, Health and Nutrition Pinstrup-Anderson, P., M. Jaramillo and F. Stewart (1987): Sector Review. Report No. 7597-GH. Washington DC: "The Impact of Government Expenditure," in Cornia, A., World Bank Jolly, R. and Stewart, F. (eds) Adjustment with a Human _ (1989e) Tanzania: Population, Health and Nutrition Face. Protecting the Vulnerable and Promoting Growth (A Sector Review. Report No. 7495-TA. Washington DC: UNICEF Study). Clarendon Press, Oxford. World Bank Sahn, D. (1990) "Fiscal and Exchange Rate Reforms in _ (1989f) Tanzania: Public Expenditure Review, Wash- Africa: "Considering the Impact on the Poor," Cornell ington DC: World Bank University, Food and Nutrition Policy Program, Proc- _ (1989g) Malawi: Public Expenditure Review: Report essed (March), Washington DC. No. 7281-MAI. Washington DC: World Bank Sahn, D. (1990): "Malnutrition in Cote d'lvoire," Social _ (1990a): World Development Report 1990, Washington Dimensions of Adjustment Working Paper No. 4, World DC: World Bank Bank. _ (1990b): "Health Care Cost, Financing and Utilization Sen, A.K. (1981): "Public Action and the Quality of Life in in Nigeria," In preparation, Washington DC: World Developing Countries" Oxford Bulletin of Economics and Bank Statistics. , (1990c): "Analysis Plans for Understanding the So- UNDP (1990)" Human Development Report, Oxford Univer- cial Dimensions of Adjustment," Social Dimensions of sity Press. Adjustment Unit, Report No. 8393-AFR, Washington Vogel, R.J. (1988): "Cost Recovery in the Health Care Sector DC: World Bank Selected Country Studies in West Africa," World Bank , (1991): "The SDA Integrated Survey, Social Dimen- Technical Paper Number 82, Washington DC: World sions of Adjustment Unit, Washington DC: World Bank. Bank.22 105 11 Manufacturers' Responses to Infrastructure Deficiencies in Nigeria: Private Altemnatives and Policy Options Kyu Sik Lee andAlexAnas In many countries in Africa, infrastructure provision icing policies in order to induce active private sector suffers from two kinds of inefficiencies. The first is participation in infrastructure service provisions. the presence of a public sector with a relatively high The purpose of this paper is threefold. First, we level of capital investment in place but which re- document how infrastructural deficiencies affect mains non-performing or unable to provide steady manufacturing firms of different sizes in different and reliable infrastructural services. The second, a regions. Second, we describe how firms respond to consequence of the first, is that many users of the the deficiencies, identify the costs of these responses, public infrastructural services find it necessary to and estimate the extent of private cost as a measure provide their own facilities in whole or in part by of thewillingness of firms to payforreliableservices. incurring the much higher costs of private provision. Third, based on these observations, we offer altema- These two extremes, (i) the non-performing public tive policy options for improving infrastructure pro- sector and (ii) the private provision responses of visions in Nigeria. These policy options provide al- firms, are well known to exist in Nigeria. ternatives between the two extremes of the non-per- The solution to this problem of infrastructural def- forming public sector and the private provision by iciencies in Nigeria and other African countries is not individual manufacturers. The policy options dis- likely to be a technological one. It is generally under- cussed in this paper include: (i) regulatory changes stood that in these countries large additional capital for enabling fuller utilization of existing private pro- outlays or extensive rehabilitation programs cannot vision capacities, for example, by allowing the sale of be fully effective without progress in improving inst- excess private power supply; (ii) private sector par- itutional organization, logistical support services, ticipation in selected infrastructure support activi- and administration. Yet, it is these areas which are the ties, such as production, distribution, maintenance, least well understood and where progress has rem- metering or revenue collection; and (iii) altemative ained elusive, difficult and unpredictable. Thus, it is pricing and tariff strategies which exploit observed realistictoassumethatthepublicsectorwillcontinue variations in private provisions by firm size and to remain non-performing for some time to come and location. that the infrastructural deficiency problem will need Ouranalyses in this paper arebased on the empiri- to be addressed in a way which minimizes the social cal results from the survey of manufacturing estab- cost in a shorter timeframe. This would require fine- lishments conducted for this research project. The tuning regulatory regimes and the existing instituti- questionnaire was developed by the World Bank onal structure, and comingup with moreefficientpr- (supported by then the West Africa Regional Re- 106 ManufacbArers'Responses to Infrastructure Deficiencies in Nigeria: Private Alternatives and Policy Options search Fund) in collaboration with the Nigerian In- respond to changes in policies such as tariffs or regul- dustrial Development Bank. The field survey was atory constraints. The current paper sets the stage for implemented by Arthur Andersen & Co., Lagos. A such an analysis by identifying the appropriate re- stratified random sample was drawn from the sam- sponse pattems. The plan for an econometric analy- ple frame of manufacturing establishments provided sis and the associated measurement needs arebriefly by the Nigerian Federal Office of Statistics. The sam- mentioned in the concluding section. ple covered five states: Lagos, Anambra, Imo, Kaduna and Kano. The survey consisted of 36 pages The Extent and Causes of Infrastructure with 349 computer readable variables, and was com- Deficencies pleted in late 1988 for 179 manufacturing estab- lishments. The sample firms covered all manufactur- It is common knowledge that Nigerian manufac- ing industries (at the two-digit level of the Standard turers suffer from frequent interruptions of publicly Industrial Classification) and a continuum of firm provided services such as electricity, water, telecom- sizes as measured by employment. Infrastructural munications, transport, and waste disposal and by deficiencies and firms' private provision responses the poor quality of these services when and where are covered for five subsectors: electricity, water sup- they are available. A detailed discussion of these ply, transportation of freight and personnel, telecom- problems for each infrastructural sector is given in munications, and waste disposal. (The Nigerian In- Anas and Lee (1988; pp.3-8). dustrial Development Bank has completed the sur- The Nigerian Industrial Development Bank vey for an additional 66 establishments among its (NIDB), the collaborating local institution of this client firms. This data is still being processed and not study, has been particularly concemed about these included in our analysis in this paper.) problems, since financing industrial projects has T;he paper is organized as follows. The second been its main activity. According to NIDB's staff, section documents and discusses the extent, appar- frequent power cuts and voltage fluctuations have ent causes, and incidence of infrastructural deficien- forced almost every industrial establishment in the cies in Nigeria. We have drawn from World Bank country to undertakeextra investments ingenerators project reports, and institutional and other qualita- in order to avoid production losses as well as damage tive information on the state and causes of deficien- to machinery and equipment. For similar reasons, cies in selected infrastructural subsectors in order to extra investments are also made in sinkingboreholes complement the information from the establishment and installing water treatment plants. Such extra survey data collected. The third section focuses on investments raise industrial costs and make it diffi- the alternative private provision responses of manu- cult for local industrial products to compete in price facturers. Prior to the survey of establishments, our with their imported counterparts. By unduly enlarg- knowledge of private response options of the firms ing the overhead and running costs, they lengthen was based on rough aggregate figures, anecdotal the gestation period of industrial projects. descriptions of selected cases, or specific field inter- State monopoly enterprises such as the Nigerian views of several firms. The survey findings clearly Electric Power Authority (NEPA) or the Nigerian document the presence of a wide range of responses, Telephone Company (NITEL) have a large amount and the frequency of their occurrences, and also their of capital investment already in place but fail to incidence and costs by region, size of firrn, and other deliver their services at the level required to meet the characteristics. The fourth section presents the esti- demand. Such failures not only result in the waste of mates of capital costs of various private provisions sarce resources but also significantly affect manu- and analyzes the private cost as a measure of the facturing and other productive activities in the Nige- willingness of firms to pay for reliable services. The rian economy. Therefore, it is important to emphasize fifth section discusses several policy options devel- that infrastructure services are intermediate inputs oiped and their economic rationale. To the extent used in producing final goods and services and that possible, we use the survey results to give a prelimi- the inadequate supply of these services will ad- nary empirical justification of the potential benefits versely affect the productivity growth of industries of the policy options considered. and economic development in general. To make quantitative estimates of the benefits of The causes of infrastructural failures may be the suggested policies it will be necessary to imple- grouped into two kinds. The first is relatively well ment empirically the analytical framework devel- understood and relates to shortcomings of the tech- oped in Anas and Lee (1988), by estimating the de- nologyusedby the public sector, includingproblems gree by which individual firms in the sample will in the day to day management, and operation and 107 Manufacturers' Responses to Infrastructure Defkicencies in Nigeria: Privatt Alteratives and Policy Option. maintenance of the facilities. The second is more erable at any given time. However, operable gener- complex in nature and less well controlled, and re- ating capacity is still considered substantial and es- lates to general problems with administration, bu- sentiallyadequate. Most powerinterruptions(nearly reaucracy, planning, metering, billing for services two thirds) are a result of bottlenecks on the trans- delivered, revenue collection, personnel training in mission and distribution networks. These recurring the public sector, and lack of appropriate incentives transmission problems are believed to be due to the for management and personnel in part because of lack of spare parts or the delays in obtaining them. In civil service pay ceilings. This second set of factors addition, shortages of materials, vehicles and foreign has remained the key problem over the years because exchange have been the key factors which have con- further investments in additional facilities is easily strained the expansion of the distribution system. In rendered ineffective if the institutional organization recent years these factors have been aggravated by and logistical support systems are lacking. the sharp fall in the price of oil which has reduced the Assessing the actual burdens imposed by the cur- public budget, as well as by the sharp devaluation in rent inadequacies and the costs of ongoing adjust- the value of the naira which makes imported spare ments will be useful as the government continues to parts even more expensive. A persistent problem has make strategic investment choices which involve the been the frequent overloading of transformers. The following types of trade-offs: fact that only 400 to 500 of NEPA's fleet of 3,000 * among different users of the infrastructural vehicles are operational has systematically ham- services such as residential versus manufac- pered routine maintenance of the distribution net- turing; work. Similarly, lack of properly trained personnel is * betweenadditional capital investments versus the apparent cause of failures to maintain circuit maintenance and rehabilitation of existing fa- breakers on the transmission network. Another area cilities or the training and recruitment of per- which contributes to these problems is the inade- sonnel; quacy of NEPA's monitoring facilities in its National * among different infrastructural subsectors System Control Center which is supposed to track such as electric power versus telecommunica- and quickly service failing components on the na- tions; (iv) between as well as within regions tional network (World Bank 1989). and cities; Most recent studies have paid attention to the * between alternative pricing and tariff struc- nontechnological factors contributing to power in- tures; terruptionsand failures, and thecurrentgovernment - between assisting the private sector in its self- efforts to partially commercialize a number of para- provision efforts versus supporting further statals have also included NEPA (World Bank 1989). the public infrastructure sectors; and NEPA will therefore have more autonomy in wage - between different organizational and struc- and compensation policy, tariff setting and in deter- tural reforms focused on deregulation, com- mining its own capital expenditure program. It is mercialization, and the partial privatization of generally recognized that current NEPA tariffs for selected infrastructure related functions in in- electricity have essentially no relationship to eco- dividual subsectors. nomic opportunity costs. Forexample, electricity tar- iffs remain unchanged since 1979 when they were Causes of the Deficiencies raised to 7 kobos per kWh. At this level, it is estimated that they are about one seventh of the long run World Bank studies and project work in the last marginal cost of supply and do not even cover the decade have documented the extent and causes of cash operating costs of generation, transmission and infrastructural failures in each sector. Taking electric- distribution. It is known that in most of developed ity generation as one example, the current situation and advanced developing countries such as Korea, can be gleaned from two project appraisal reports the tariff per kWh is about 7 US cents (52.5 kobos). which are eight years apart (World Bank 1981 and With the already established commercialization of 1989). The basic types and causes of failure remain the Nigerian National Petroleum Company (NNPC) essentially unchanged over the entire decade of the which supplies gas fuel to NEPA, gas prices are going 1980s. Technological causes of failure in this sector up and NEPA would have to raise its electricity tariff are primarily related to transmission and distribu- soon, as NEPA becomes subject to a higher degree of tion. The ratio of the available capacity to that of the market discipline. In addition, it has been recom- installed capacity is generally low and as much as 50 mended that the tariff be raised in stages in the next percent of installed capacity maybe essentially inop- several years (World Bank 1989). 108 Manufacturers' Responses to Infrastructure Deficiencies in Nigeria: Private Alternatives and Policy Options The problems of underpricingare also observed in lows. Table 11-1 shows that only 14 out of the 179 water supply. The Lagos State Water Commission firms, or 7.8 percent do not have their own electricity (LSWC) since 1986 is operating under a new tariff generators. Twelve of these firms are in Anambra and which raised water prices for manufacturers by ab- Imo and two are in Lagos while all firns in Kaduna out 40 percent and a further increase of 270 percent and Kano have their own generators. For the firms was due forapproval. Avendorlicensingsystem au- that do not have their own generators (or "captive thorized to levy direct charges for water sold at pub- firms"), the supply is not 100 percent reliable. Table lic standposts is under discussion (World Bank 1988). 11-2 shows that the captive firms are generally small Since industrial wateruse isbeginningtoexhaust the ones. Moreover, table 11-3 shows that the smaller groundwater supplies of the Lagos State region, it is firns are subject to the bulk of the power failure reasonable to expect that tariff increases for industr- incidents. Some small firms do not have their own ial use of water may become more feasible in the fu- generating equipment, not because the burden of ture. (More detailed discussions of the causes of de- poor electricity supplies is less per unit of output for ficiencies appear in Lee, Stein, and Lorentzen 1989.) them, but rather because the generation cost per unit of electricity is higher for them because of economies The Incidence of the Deficiencies by Firm Size and of scale in electricity generation. Region Small firms are the ones that cannot afford capital investments for boreholes, vehicles for the shipment Our data reveal that there are large variations in of products, motor cycles and couriers, and radio the availability and quality of public infrastructure equipment. Compared to the other two regions, services and in firms' private provision responses Anambra-Imo has a higher concentration of small across regions and firm sizes. Such observations im- firms and the burden of inadequate infrastructure plyan important role in government strategy regard- seems to be more serious there. ing infrastructural policy reform. Variations in pri- The heavy incidence of infrastructural failures vate provision patterns can be summarized as fol- among small firms has an implication for the growth of industries and the generation of employment. Ac- Table 11-1. Distribution of manufacturing cording to the "incubator hypothesis" that was tested establishments by region and source of electricity in the earlier World Bank research on industrial loca- Source of electricityfor production operation tion in Bogota (Lee 1989a) and in Seoul (Lee 1985), it NEPA NEPA Own Own was observed that small new firms spend their early Region only main gen. mingen.only Total years near the city center or in an old industrial area with easy access to good utilities and other essential services. They do so because it is prohibitively expen- frequency 2.00 68.00 10.00 2.00 82.00 sive for small firms to operate in outlying areas row% 2.44 82.93 12.20 2.44 100.00 where infrastructure services are poor. As they grow column% 14.29 48.57 50.00 40.00 45.81 and become more independent, they tend to move Anambra/Imo out of the central area for more space. In Nigeria and frequency 12.00 22.00 1.00 1.00 36.00 perhaps in most African countries, large cities with row% 33.33 61.11 2.78 2.78 100.00 poor infrastructure cannot offer the incubator func- column% 85.71 15.71 5.00 20.00 20.11 tion for small new firns. Since small firms cannot Kaduna/Kano afford their own generators and boreholes and other frequency 0.00 50.00 9.00 2.00 61.00 facilities, the burdens of inadequate public infra- now% 0.00 81.97 14.75 3.28 100.00 structure services are especially severe for the small column% 0.00 35.71 45.00 40.00 34.08 firms which start and grow in those cities. This has a Total serious negative implication for the birth and growth frequency 14.00 140.00 20.00 5.00 179.00 of small firns and for the generation of employment row%c 7.82 78.21 11.17 2.79 100.00 and income. The studies mentioned above (Lee 1985, column% 700.82 780.21 110.0 17 2.79 100.00 1989a) showed that small new firms generate be- tween 60to 80 percentof thenewjobscreated in large Ncte: a. NEPA only = using 100% from NEPA, NEPA main = cities in Asia and Latin America. This implies high NEPA as the main source and own generators as standby; Own returns in Nigeria to selectively improving infra- gen. main = NEPA as standby; Own gen. only = 100% from own structure service provisions for particular users at Sourcet NrsD/.BRD Establishment Survey,1988. particular locations, since the observed service reli- 109 Manufact*rers' Responses to Infrastructure Dficiendes in Nigeria: Private Altenatives and Policy Optons Table 11-2. Distribution of manufacturing estalishments' source of electricity by firm size Firm size Source of electricity 1-19 20-49 50-99 100-199 200499 500-999 1,000 & over Total NEPA only frequency 11.00 3.00 0.00 0.00 0.00 0.00 0.00 14.00 row% 78.57 21.43 0.00 0.00 0.00 0.00 0.00 100.00 column7% 68.75 8.57 0.00 0.00 0.00 0.00 0.00 7.82 NEPA main frequency 3.00 26.00 35.00 30.00 25.00 13.00 8.00 140.00 row% 2.14 1857 25.00 21.43 17.86 9.29 5.71 100.00 column% 18.75 74.29 79.55 85.71 96.15 86.67 100.00 78.21 Own gen. main fiequency 2.00 4.00 8.00 5.00 0.00 1.00 0.00 20.00 row% 10.00 20.00 40.00 25.00 0.00 5.00 0.00 100.00 column% 12.50 11.43 18.18 14.29 0.00 6.67 0.00 2.79 Own gen. only frequency 0.00 2.00 1.00 0.00 1.00 1.00 0.00 5.00 row% 0.00 40.00 20.00 0.00 20.00 20.00 0.00 100.00 column% 0.00 5.71 2.27 0.00 3.85 6.67 0.00 2.79 Total frequency 16.00 35.00 44.00 35.00 26.00 15.00 8.00 179.00 rWo% 8.94 19.55 24.58 19.55 14.53 8.38 4.47 100.00 column% 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 Source: NDIB/IBRD Establishment Survey, 1988 Table 11-3. Distribution of manufacturing estalishments'power outages byfirm size Firn size Average number of pouer outages per week 2.19 20-49 50-99 100-199 200499 500-999 2,000 &over Tofal Less than 5/week frequency 7.00 11.00 15.00 8.00 7.00 7.00 3.00 58.00 row% 78.57 21.43 0.00 0.00 0.00 0.00 0.00 100.00 columnn% 68.75 8.57 0.00 0.00 0.00 0.00 0.00 32.40 5-10/week frequency 9.00 26.00 22.00 19.00 15.00 4.00 2.00 91.00 row% 2.14 18.57 25.00 21.43 17.86 9.29 5.71 100.00 column% 18.75 74.29 79.55 85.71 96.15 86.67 100.00 50.84 More than 10/week frequency 2.00 4.00 7.00 8.00 4.00 4.00 3.00 30.00 row% 10.00 20.00 40.00 25.00 0.00 5.00 0.00 100.00 column% 12.50 11.43 18.18 14.29 0.00 6.67 0.00 16.76 Total frequency 160.00 35.00 44.00 35.00 26.00 15.00 8.00 179.00 row% 8.94 19.55 24.58 19.55 14.53 8.38 4.47 100.00 column% 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 Source: NDIB/IBRD Establishment Survey, 1988 110 Manufacturers' Responses to Infrastructure Deficiencies in Nigeria: Private Alternatives and Policy Options ability problems are to an extent location and user production in favorof those inputsand raw materials specific. which are less infrastructure intensive. For example, if a firm has a choice between a labor intensive and a Alternative Responses of Manufacturers capital intensive process and if the labor intensive process relies less on infrastructure than the capital There are essentiallyfourways in which firms mig- intensive one, the firm's strategy would be to substi- ht respond to infrastructural deficiencies. These are: tute labor for capital thus reducing the quantity of * relocation, infrastructure inputs. The various private provision * factor substitution, activities with large capital expenditures undertaken * private provision, and by the Nigerian firms indicate that their ability to * output reduction. adjust to the relative prices of labor, machines, mate- rials, or various infrastructure service inputs is rather Below we discuss the economic rationale behind constrained by the current technologies in use. Since each of these responses, and why they are or are not such input substitution possibilities are limnited, the observed in Nigeria. finns operate inefficiently by providing their own infrastructure services when these are crucial for Relocation their operations. In case of a milk processing plant, for example, even if the public power supply were The firm may relocate to a site with better infrast- available at proper voltage for as much as 90 percent ructure services. Such relocation can occur within a of the time, the firm could not afford to eliminate its city or from one region to another. Our survey results own generators with 100 percent capacity because show that Nigerian firms do not move to other locati- anyvoltage surges and drops at a critical time would ons from the initial site. Even though 50 percent of threaten key equipment in the production process the firms had been at their present location since and result in much waste. 1980, only two out of the 179 sample firms indicated that they had relocated from another location. This Private Provision absence of mobility is striking considering that the average annual moving rate observed in large cities As already mentioned, numerous strategies are in other developing countries is about 5 percent. The available for the firms to provide their own infra- relative immobility of Nigerian firms is consistent structure services. The fact that the vast majority of with the fact that the capacity, regularity and quality firms do so even when the publicly provided infra- of infrastructure vary from bad to worse within and structure services are extremely inexpensive, indi- across Nigerian cities. This tends to limit the gains in cates the importance of having reliable infrastruc- infrastructure quality that can be achieved by mov- tural inputs. Private provision as a strategy is not ing to new locations. Nigerian firms instead under- entirely separate from factor substitution. In fact, by take their own extensive capital expenditures and providing their own infrastructural services, firms incur regular operations and maintenance outlays to are substituting internal capital in the form of equip- provide their own services. The high setup cost with ment, machinery as well as labor in the form of a large amount of initial capital investment for own maintenance personnel for the publicly provided in- service provision would make it difficult for the firms frastructure services which are not forthcoming. As to move. documented in Anas and Lee (1988), Nigerian firms Another problem with relocation is that it often are observed to pursue four different private re- involves trading one infrastructural deficiency for sponse strategies. These are: another. For example, a firm that moves into the * Self-sufficiency: The firm provides its own in- Lagos area because it is much cheaper to sink bore- frastructural services to the point where it holes there (since the water table is high), might does not need any public inputs. For example, better its water supply, but the finn may face new table 11-1 shows that only 5 out of the 179 problems such as losses in production time due to the surveyed firms are in this mode with respect commuting delays of employees. to electricity generation. Standby private provision: The firn has its Factor Substitution own infrastructural facilities in place and swit- ches to these facilities when the quality or rel- The firn may substitute away from the use of the iability of the public services falls below a crit- poorly provided service by adjusting its mode of ical level. From table 11-1,140 firms or 78 perc- 111 Manufactures' Responses to Infrastructum Deficiencies in Nigeria: Private Atternatives and Policy Options ent of those surveyed are in this situation with even back to NEPA. 'Satellite behavior" is the other respect to power supply. side of the coin with respect to joint production. A * Public source as standby: The finn relies pri- satellite firm is one which purchases infrastructure marily on its own facilities but switches to the services from another firm that has surplus infra- public supply during those times of the day structure services to sell. At times of power interrup- when the public source delivers a high quality tion, for example, a satellite firm would switch from service. Again, from table 11-1, twenty firns NEPA to the generators of a nearby private producer. or about 11 percent of the surveyed firms re- "Shared production" refers to the possibilityof firms ported such behavior. coming together in a club type of arrangement called e Captivity: The firm continues to rely on the "utilitypool'tosharethecostofinfrastructuralcapi- public source exclusively despite the very low tal inputs by building their own facilities. (A theoreti- reliability of such a service. It is reasonable to cal framework for the club type arrangement is dis- expect that captivity will be the dominant cussed in McGuire 1974.) The above typology of mode among the very small firms who cannot private provision alternatives is applicable to all five afford infrastructural capital investments. On- infrastructure subsectors considered in this study. ly 14 or 7.8 percent of the surveyed firms re- ported such behavior in the case of electricity. Output Reduction Anas and Tee (1988) argued that there are eco- This response to infrastructural deficiencies is also nomic incentives for three additional regimes of pni- common. Firms which are captive or use their own vate provision which are not observed in Nigeria standby equipment are subject to output reduction because of government regulations on the supply either on a regular basis or when their own equip- and trading of infrastructure services by private en- ment fails to operate propedy. However, the chief tities. These regimes are: impact of output reduction necessarily falls on small • joint production; firms which find it too expensive to pursue another • satellite behavior, and response, or on very large power intensive firms e shared production. which cannot find appropriate size equipment (e.g. generators) to meet their service needs. It is difficult "Joint production" refers to the case where a firm, to observe, but it undoubtedly happens that many typically a large one, which has already made a small firns in Nigeria have either shut down or have substantial investment in infrastructural capital failed to grow to any critical size because of infra- finds it profitable to sell part of its infrastructural structural deficiencies. Also, births of new firms will output to other firms. With few exceptions, this has be reduced if many must shut down soon afterbirth not been possible in electricity production in Nigeria, because of infrastructural inadequacies. because private producers of electricity are not nor- mally allowed to sell surplus power to other firms or Costs of Private Provision Table 11-4. Values of private infrastructure Capital Costs and their Incidence provision as a percentage of the total value of machinery and equipment The firms that we have surveyed provide a telling Private provision S ,all firms Lrge fir Total story of the incidence of private provision which is by far the most dominant response among Nigerian Generators 24.78 10.06 10.42 manufacturers. Table 11-4 summarizes the data on Borholes 2.81 1.91 1.91 average current market values of various equipment Vehicles for workers 5.49 2.84 2.86 and facilities used for own service provisions and Vehicles for shipments 10.95 4.47 4.62 their share of the total value of the firm's machinery of goods and equipment for production. We find that the capi- Vehicles for garbage 0.15 0.48 0.48 tal value of generators and support facilities such as disposal the swikhes and transformers is on the average 25 Radio equipment 1.48 0.59 0.59 percent of the total value of machinery and equip- Note: a Establishments with less than 50 employees. TIhe ment for small firms (with less than 50 employees) values of generators, borehole, and radio equipment are included in the total values of machinery amd equipment, but and 10 percent for large firms. This share varies those vehicles are not included. widely across the five states and by firm sizes, from Source: Anas and Lee (1988). 4 percent for large firms in Imo to 36 percent for small 112 Manufactuwrs'Responses to Infrastructurm Deficiencies in Nigeria: Private Alternatizvs and Policy Options firms in Anambra. The average value of capital for inability to get them to the factory on time. In the electricity generation including all firms is 954,000 shipment of goods, 63 percent of the surveyed firms naira (about 130,000 U.S. dollars). This value is al- had their own vehicles. These vehicles make up 11 most four times larger than the share of capital for percent of total capital equipment for small firms but boreholes and treatment facilities. The average value only slightly more than 4 percent for large firms. The is 260,000 naira for all firms with boreholes, which is average capital value of these vehicles was 387,000 about 2 percent of the total value of machinery and naira for each firm. Capital expenditures such as equipment. This share value varies from 0.5 percent radio equipment and motorcycles for couriers are in Kano to 2.1 percent, or six times higher, in Lagos. small compared to generators and boreholes, but Although water supply takes up a much smaller returns to these investments are extremely high. share of equipment and machinery than does elec- About 37 percent of the firms have radio equipment tricity, the share is again higher for small firns than and its share in the total value of machinery and it is for large ones, by about 50 percent. equipment is nearly three times higher for small Although only about 15 percent of the firms pro- firms. On the average, managers spend more than 10 vide transport for their workers, the share of these hours per week on the road to deliver messages or vehicles in total capital equipment is 5.5 percent for hold conversations that could be handled in mo- small firms and just under 2.8 percent for large firms. ments over a working phone line. The low ratio of self-provision observed in transport- ing one's own workers mean that, at least in Lagos, The Private Cost as a Measure of Willingness to Pay a great deal of production time is lost because of the for Reliable Services late arrival of workers. When firms choose not to make capital expenditures for their own provision of As documented in the above section, manufactur- certain services, they often incur comparable costs in ers incur high capital cost in installing own facilities other forms such as in lost production time. In Lagos, for providing their own services. In the case of elec- long commuting time is not due to the distances tric power generation, the survey reveals that nearly between residences and workplaces but due to long all standby firms have installed capacity sufficient to waiting times for buses. Savings from employing run the entire plant during a period of NEPA power workers with lower wages are limited by the firms' interruption. The data also indicates that the sample firns as a whole 25 percent of all power used by them Figure 11-1. Proportion of Electric Power Supply during 1987 came from their own generators and 75 from Own Generators, 1987 percent from NEPA. (The breakdown by firm size is Percent shown in figure 11-1) Because the typical installed §0_ private generation capacity is approximately suffi- cient to run the entire plant (with some reserve for maintenance), this means that about 75 percent of the - 41.45 generation capacity remains idle. This idle capacity 40 - results in extremely high total average cost of private 35.46 power generation as shown below. The high cost of _32.33 2975 private provision sustained by the firms is the im- 30- - _ _ _ _ X 29.75 plicit value of service reliability that the firms are willing to pay for. A precise measure of willingness 23.61 to pay can be determined by calculating the average 20- - cost (per kWh) of electricity produced by the firm's own generators (as the lower bound). When the av- erage cost of the privately produced power is higher than the price charged by NEPA, the difference be- 10_ tween these two gives the premium which manufac- turers are willing to incur in order to insure them- selves of an uninterrupted power supply at all times. 0 20 , Tables 11-5 and 11-6 show two such sets of com- 1-9 10-19 2049 50-99 1 20-099200499 5009991000 + putations on the average cost of pnvate power gen- Firm size eration. Table 11-5 shows the average cost computed Source: Based on data from NIDB/IBRD Project Establishment Suovey, using each firm's reported power consumption from 1988. own generators during 1987 for different firm size 113 Manufacturers' Responses to Infrastructum Deficencies in Nigeria: Private Atternatizvs and Policy Otions categories (25.48 percent of the total consumption for US dollars. (During the 1980s the average inflation the sample firms as a whole). Thus, these figures ratewasaboutl2percentandthecurrentcommercial reflect the cost of holding idle generating capacity. lending rate is about 20 percent.) Table 11-6 shows the average cost of electric power Table 11-5 shows that at the actual average utiliza- generation for different firm size categories assum- tion rate of 25 percent of the generating capacity, the ing that 100 percent of power supply comes from average cost per kWh is 4.61 Naira, which is 66 times own generators. In both tables 11-5 and 11-6, the the present NEPA price of 7 kobos. Suppose that the capital recovery cost is computed by annualizing the NEPA tariff were to be adjusted to 30 kobos, a rate current market value of the firm's generators and currently charged by a private supplier in Lagos. The accessories using the remaining service life. The re- average firm would still be incurring 15 times the curring costs of fuel, maintenance, and labor, are new NEPA price at the actual utilization rate of 25 added to the capital cost. In table 11-6, these reported percent. Even under the assumption of 100 percent recurring costs are appropriately adjusted for the full supply from own generators, the average cost of 1.41 utilization case as explained in footnote (a) to table Naira for all firms (table 11-6) will be five times 11-6. Theaveragecostschedulebyfirm sizehasbeen higher than NEPA's 30 kobos. The premium is high- calculated with different sets of assumptions on (i) est for the 20-49 person firm size category with a the real rate of interest and (ii) the exchange rate. In factor of 6 while for the largest size category of 1,000 our discussion below, we refer to the average cost or more persons the premium is a factor of only 1.3. schedule computed with the 10 percent real interest Small firms pay for a higher premium because of rate and the current exchange rate of 7.5 Naira per economies of scale in electric power generation. Table 11-5. Average cost of electric powergener- Table 11-6. Average cost of electic powergener- ation by firm size: underutilization casea ation by firm size: full utilization case' Average cost Average cost (Naira/kWh) (Naira/kWh) Finn size Interest rate b 5% 10% 15% Firm size Interest rate b 5% 10% 15% (1987 exchange rate US$14.0 Naira) (1987 exchange rate US$14.0 Naira) All firms 2.540 2.834 3.150 All firms 0.959 1.021 1.086 0-9 0.374 0.426 0.483 0-9 0.143 0.155 0.169 10-19 0.698 0.781 0.871 10-19 0.453 0.463 0.493 20-49 3.336 3.740 4.171 20-49 1.101 1.180 1.263 50-99 2.698 3.009 3.346 50-99 1.045 1.122 1.206 100-199 2.573 2.936 3.328 100-199 1.023 1.091 1.163 200-499 2.357 2.564 2.780 200-499 1.018 1.060 1.104 500-999 1.442 1.611 1.793 500-999 0.675 0.712 0.752 1,000 + 2.327 2.439 2.556 1,000 + 0.314 0.326 0.339 (2989 exchange rate US$1 = 7.5 Naira) (1989 exchange rate US$1 = 7.5 Naira)c All firms 4.061 4.612 5.204 All firms 1.291 1.407 1.530 0-9 0.634 0.732 0.838 0-9 0.205 0.228 0.252 10-19 1.086 1.243 1.412 10-19 0.568 0.621 0.677 20-49 5.701 6,457 7.267 20-49 1.106 1.752 1.908 50-99 4.191 4.775 5.407 50-99 1.380 1.525 1.682 100-199 4.196 4.876 5.611 100-199 1.444 1.572 1.708 200-499 3.718 4.106 4.512 200-499 1.243 1.322 1.406 500-999 2.063 2.379 2.721 500-999 0.821 0.890 0.996 1,000 + 3.105 3.315 3.534 1,000 + 0.314 0.339 0.423 Note: a The average utilization of installed geberating Note a Assumed 100% of electric power supply comes from capacity was 25.48%. b Interest rates represent hypothetical real own generators. Fuel consumption and maintenence cost are rates. c Adjusted for the values of the generators and accessories adjusted accordingly. Fuel by a factor of 4 and mainlenence arvd only. parts by 3, when utilization rate increases from 25% to 100%. Source: NIDB/IBRD Project Establishment Survey, 1988. Interest rates represent hypothetical real rates. ' Adjusted for the values of the generators and accessories only. Source NIDB/IBRD Project Establishment Survey, 1988. 114 Manufacturrs'Responses to Infrastructurm Deficiencies in Nigeria: Private Altematives and Policy Optons From the 20-49 person firm size category the average tensive equipment is fully utilized, the fall in average cost declines exponentially with the firm size. The costishigherinthecaseof fullerutilization.From the cost schedules in tables 11-5 and 11-6 have been above analysis, we can conclude that the premium fitted to semi-log and double-log regressions as re- over the NEPA price declines with an increase in firm ported in table 11-7. The slope coefficients are all size and that even after a hypothetical tariff increase statistically significant. The average cost values to 30 kobos per kWh, the fuller utilization case pre- shown in tables 11-5 and 11-6 are plotted in figure miums would still be larger than the NEPA price for 11-2 (excluding the values for firms with less than 20 all firm sizes. Of the average total cost of 4.61 naira employees). in the case of underutilization, the average variable The premium paid by firms varies with firm size. cost is 80 kobos for the sample firms as a whole (table Such variation should be a central concern in the 11-8). A NEPA price of 30 kobos will be only about a design of appropriate policies forboth efficiency and third of the average variable cost of self-generation. equity reasons. The smallest group with less than 20 In some developed countries, gas turbine generators employees shows an average cost that is lower than are widely used and they do not manifest economies the sample mean. This is not because they can gener- of scale. This technology is seldom used in Nigeria as ate electric power at lower cost however. Rather it is yet. The minimum size for gas turbine generators because they cannot afford to make the expensive however is likely to be too large for the need of most capital investment to meet the required power need. individual firms. They may be able to generate enough power to sup- port the lighting and other critical elements. Developing Policy Options for Improving Service The evidence of the presence of economies of scale Provision in electric power generation is clear from the 20-49 size category as mentioned above. The cost of pro- As explained in the introduction, in Nigeria two ducing 100 percent of power supply from the in- extreme cases of inefficiency in the provision of infra- stalled generating capacity falls by a factor of 4.4 structural services are observed: first, the non-per- (from 1.752 to 0.399 in table 6) as firm size increases forming public sector with heavy capital invest- from "20-49" to "1,000 and over." When the cost of ments; second, the costly provision of services by idle capacity is included, the average cost in the same individual firms themselves. The self-provision re- range of firm sizes falls by only a factor of 1.9 (from sponse has developed over the years because of non- 6.457 to 3.315 in table 11-5). Since large firms can performance in the public sector. Without the exten- achieve great scale economies when their capital in- sive private provision responses, the total welfare Table 11-7. Regression of the average cost of electric powergeneration firm size. Semi-log Double-log 5%a 10% 15% 5% 10% 15% Fuel utilization case" Constant -0.172 -0.095 -0.005 1.153 1.323 1.432 (1.50) (0.83) (0.04) (2.41) (2.77) (3.06) Slope -0.000619 -0.000625 -0.000636 -0.306 -0.325 -0.328 (3.25) (3.39) (3.39) (3.19) (3.39) (3.49) R2 0.0669 0.0670 0.0720 0.0649 0.0721 0.0761 N 149.Oc 150.0 150.0 149.0 150.0 150.0 Underutilization cased Constant 0.646 0.773 0.870 1.629 1.830 1.921 (4.89) (5.86) (6.46) (2.95) (3.33) (3.41) Slope -0.000503 -0,000522 -0.000513 -0.229 -0.246 -0.240 (2.30) (2.39) (2.29) (2.07) (2.23) (2.16) R2 0.0355 0.0379 0.0349 0.0289 0.0332 0.0310 N 146.0 147.0 148.0 146.0 147.0 148.0 Note: a Interest rates represent hypothetical real rates. b Assumed lOO% of electric power supply comes from own generators. Fuel consumption and maintenence cost are adjusted accordingly. Fuel by a factor of 4 and maintenence and parts by 3, when utilization rate increases from 25% to 100%. b Interest rates represent hypothetical real rates. c The total number of observations may not be the same because the log of negative values is not defined and they are treated as missing values. d The average utilization rate of installed generating capacity was 25.48%. Source. NIDB/IBRD Project Establishment Survey, 1988. 115 Manufacturers'Responses to Infrastructure Deficiencies in Nigeria: Private Alternatives and Policy Options loss resulting from the public sector failures would strengthening those already existing markets for the have been much higher in Nigeria. private supply of infrastructure services or creating At best, public sector performance is likely to im- new ones such that the costs of private provision are prove very gradually. In addition, improvements in significantly reduced and more efficient private pro- public sector performance will be accompanied by vision alternatives are offered. Policy options can be considerable upward adjustment in pricing and tar- grouped into three categories: (i) regulatory changes iffs. Such adjustments which are necessary for long which will induce fuller utilization of existing pri- run efficiency howeverarebound to create hardships vate provision capacities; (ii) private sector participa- in the short run, as firms of all sizes and in all sectors tion in selected subactivities; and (iii) changes in and regions make their own adjustments. For these pricing and tariff structures. We will discuss below reasons, the correct policy perspective for Nigeria is each of these policy areas as illustrations for possible not to stress improvements in public sector perform- policy options drawing on the survey results. More ance to the exclusion of private sector incentives. definitive policy recommendations will be made Rather, the challenge is to find feasible intermediate later in the study based on formal empirical analyses term policy options which bridge the gap between to be conducted. the above mentioned two extreme cases of ineffi- ciency, namely, the nonperforming public sector and Regulatory Changes for Fuller Utilization of Private costly private provision by individual firms. Provision Capacities As discussed in Anas and Lee (1988), there are numerous opportunities that can be exploited for Some minor regulatory changes can generate sig- nificant benefits to individual firms. As noted earlier, most firms have standby generators which stay idle Figure 11-2. Average Cost of Electric Power Generation, about 75 percent of the time. These firms however are 1987 not allowed to sell the excess power they produce to Naira/kWh NEPA or to other firms. The potential cost savings 6- ________________________________ from allowing such transactions can be large. The Interest rate current regulations inhibit the regimes of "joint pro- \ 15% duction", "satellite behavior", and "shared produc- ------ 10% 0/0tion." Indeed, the efficiency gains of allowing large 5 - 5% firms with a high level of installed capacity to exploit fully their scale economies and to compete with Underutilization case NEPA by supplying smaller satellite firms could be 4 _ \ ' . \ significant. The presence of economies of scale was shown in the case of electric power generation. -. \ ~~~~~Such regulatory changes could also motivate -~ - - . \"shared production" whereby private manufactur- 3 - \ ers join forces to form certain types of "utility pools" to exploit economies of scale in the provision of each type of infrastructural service and economies of scope in the provision of several different infrastruc- tural services at the same time. Utility pools should be quite feasible in the existing industrial estates or Full utilization case in areas with a relatively high concentration of indus- - tries. The participation of large firms in the infra- structure production process and competition with the public suppliers broadens the choices available to small firms and especially the "captive firms." o- + | + l Small firms in such an environment can become 0 500 1000 1500 2000 2500 satellite firms or can join in utility pools. Small firms Firm size have very high willingness to pay for reliable electric Note: The plotted values are from the regression of the values in power supply. Thus, they would be motivated to join Tables 11.2 and 11.3 excluding establishments with fewer than 20 employees (using the exchange rate of USS1.00+7.5 naira). a utility pool or to become satellites to larger firms. Source: Based on data from NIDB/IBRD Project Establishment Survey A good example of a "utility pool" in place is the 1988. central effluent collection and treatment facility in 116 Manufacturm' Responses to Infrastructure Deficiencies in Nigeria: Private Alternatives and Policy Options the Agbara Industrial Estate which was established serves theentire campus. Anote prepared forarecent by a private developer. This central facility is oper- Industrial Sector Study (Lee 1 989b) further discusses ated by a management company. As the government such possibilities for the existing industrial areas in attempts to tighten industrial pollution control, treat- Nigeria. ing the effluent within individual firms would be prohibitively expensive, especially for small firms. Private Sector Participation in Contestable Markets for Similarly, in industrial layouts in Lagos, the central the Supply of Infrastructure Related Services collection and treatment of effluent by the manage- ment board should be technically feasible and will Although more efficient pricing systems combin- induce economies of scale. Such a management ed with appropriate relaxation of regulatory con- board could be further empowered to operate and straints canbe introduced to induce improved public manage "utility pools" which include a wide range sector performance and to minimize the adverse im- of services such as electric power generation, garbage pacts of infrastructural deficiencies on manufactur- collection, and the shipment of goods. Another ex- ers, these strategies alone are unlikely to significantly ample of a central facility in place is the six megawatt improve the current situation in the short run or even standby generator of the University of Ibadan which in the medium term, because of the various ineffi- ciencies in administration, financial management, Table 114. Average fixed and variable costs of and the operation and maintenance practices of the ouning own electric power generation per kWh public agencies. Based on what is observed in Nige- Firm size Fixed cost Variable cost Total ria, a sensible way of breaking this inertia seems to be the encouragement of private sector participation All firms in various infrastructure related functions and sub- Naira 3.180 0.803 4.612 activities. percent 82.60 17.40 100.00 Indeed, in Nigeria we observe that some private 0-9 firms are already engaged in certain types of infra- Naira 0.655 0.077 0.732 structure related subactivities. Recently, NEPAbegan percent 89.50 10.50 100.00 subcontracting certain segments of its operations, 10-19 such as maintenance for a power station and trans- Naira 0.990 0.253 1.243 mission facilities, to private firms. Many foreign percent 79.62 20.38 100.00 firms including Siemens and ITT have already had 2049 maintenance contracts with the Nigerian Telephone Naira 5.824 0.634 6.457 Company (NITEL). The government allowed a pri- percent 90.19 9.81 100.00 vate firm, DHL, to operate in Nigeria. DHL charges 50-99 a much higher fee than the Nigerian Postal Service, Niara 3.784 0.991 4.775 but it is faster and more reliable, and thriving with perc 3.74 20.99 4.77 good business. This is additional evidence that users percent 79.24 20.76 100.00 of services have high degrees of willingness to pay 100-199 when reliable services become available. This was Naira 4.157 0.719 4.876 also observed in the Maroko low income area in percent 85.26 14.74 100.00 Victoria Island. This area, which NEPA never in- 200-499 cluded in its network, has been served by a private Naira 3.305 0.801 4.106 entrepreneur who charges 30 kobos per kWh, four percent 80.49 19.51 100.00 times higher than NEPA's 7 kobos. But this rate of 30 500-999 kobos is still many times lower than the average cost Naira 1.646 0.733 2.379 of own power generation. Another example is the air percent 69.19 30.81 100.00 freight and passenger transport sector. In this area, a 1,000 + number of small privately owned domestic airlines N:iara 1.877 1.438 3.315 provide stiff competition to Nigerian Airways be- pebrcent 56.62 43.38 100.00 cause they supply more reliable service. Railroads, perce: 'Ant ua izedcapitalvalueofgenentorsandwhere the high sunk costs associated with the capital Atte Annualized capital value of generators and faiite mak th inutyls.otsal,cno accessories. b Include fuel, maintenence, parts and labor. For the faclities make the mdustry less contestable, cannot sample firms as a whole, 25.48% of electric supply came from as easily benefit from such private competition, but own ,generators. truckinghas emerged asa veryviablealternate trans- Source: Anas and Lee (1988). port mode. 117 Manufacturers' Responses to Infrastructure Deficiencies in Nigeria: Private Alternatives and Policy Options In Nigeria, abroad continuum of options exist bet- structural services are contestable (Baumol, Panzar ween the two extremes of inefficiency characterized and Willig 1982) because there are no large sunkcosts above. These optionsamountto providingincentives involved in capital facilities, it should be feasible to for private entrepreneurs to engage in the supply of liberalize restrictions against the setup and operation certain infrastructure services, thus creating appro- (entry and exit) of private firms. priate market mechanisms. Such markets can be spe- There are a number of situations where such a cialized to infrastructural services in the areas of strategy can be successful. A good example is the production, distribution, maintenance, administra- utility pool already discussed above. Individual tion, metering and monitoring, or bill collection. The firms in a pool may prefer to have a private infra- feasibility of creating and expanding such markets structure provider who will manage and operate a for the supply of these services by the private sector pool with shared facilities such as vehicles and waste lies in the fact that the government fails to provide collection equipment. This would allow the pool to adequate services whereas the users are willing to take advantage of the economies of scale and scope, pay for more reliable services when such are avail- as well as to pass the transaction costs of administra- able, for example in the case of electric power supply. tion and management to the private entrepreneur A recent Bank case study by Whittington, Lauria who would be self-financing by levying charges on and Mu (1989) documents how high willingness to the pool members. pay for water has led to the emergence of a complex As mentioned earlier, power generation is an area web of private market mechanisms for water distri- where private participation can be greatly increased bution in Onitsha, a Nigerian town of 700,000. In this by allowing private entrepreneurs to set up power town, the private sector operates about 275 tanker plants which compete with NEPA. A successful ar- trucks which purchase water from about 20 privately rangement exists at Jos where a privately owned owned boreholes and sell it to businesses and house- power plant which was setup in colonial times has holds with storage tanks. Many of the households been allowed to operate. This firm supplies much of purchasing such water in turn sell it to individuals the local power needs and sells its excess power to who are not equipped to store in large quantities, or NEPA. Additional private power providers are likely to thousands of small mobile private vendors. The to emerge throughout Nigeria if the existing regula- private vendors provide two times more water on the tory constraints were relaxed. If this were to happen, aggregate compared to the public utility and collect NEPA could stiffen its tariff structure since users 1 ()times the revenue in rainy season and 24 times the would have the freedom to switch to the private revenue in dry season. Households pay these private suppliers. NEPA's transmission and distribution vendors over twice the operations and maintenance grids should be made accessible to such private costs of piped water, a strong indication of the will- power companies which can be required to pay ap- ingness to pay for reliability, and clearevidence of the propriate access fees which reflect the marginal costs private sector's ability to compete with the public of serving them. Allowing access to the grids makes sector. the generation of power a contestable activity which To operationalize a workable framework for pro- greatly increases the incentive for private particip- moting private participation in the infrastructure ation. The levying of efficient access fees by NEPA subsectors, the following strategies in three key areas would provide a source of revenue which aids in cost need to be considered. recovery while reducing some of NEPA's own pow- er generation costs. This approach has been follow- Regulatory regimes and market mechanisms. The first ed in Britain with respect to both thepowerauthority step is to improve the present regulatory regimes to (Henney 1987) and British Telecom (Beesley 1981). A provide a more favorable environment for private wide range of options for private sector participat- investors so that they can enter the market for a ion have also been considered in the past. These specific service and offer alternative sources of sup- include, for example, farming out distribution func- ply. Manyof thepublicsectorfailures inNigeriastem tions to private firms (World Bank 1983b, Coyaud from the fact that most infrastructure services are 1986). provided by strongly centralized government mo- Organizational and institutional mechanisms. To ind- nopolies. As discussed in Anas and Lee (1988), how- uce the development of appropriate market mechan- ever, even some services which have the charac- isms forprivate sector participation in infrastructure teristics of public goods can be supplied with the supply, it will be necessary to allow appropriate inst- participation of the private sector (also see Roth itutionalarrangementssuchassubcontractingorfra- 1985). To the extent that the markets for certain infra- nchising to carry out a particular type of infrastruct- 118 Manufacturers' Responses to Infrastructumv Defiacncies in Nigeria: Private Alternatives and Policy Options ural service. Such mechanisms will tend to vary from tions currently under government control are dec-en- sector to sector and will depend on the strength of in- tralized and privatized, it will be important to re- centives which are needed and the efficiency gains define the role of the government for appropriate which will occur from private sector participation. monitoring and supervision of efficient market oper- A good illustration is available in thewastecollect- ations. For example, if a subactivity such as bill col- ion and disposal subsector in Nigeria where a numb- lection or garbage pickup is contracted out to priv- er of alternative institutional responses havebeen ob- ate firms, it will be necessary to monitor their suc- served in recent years (Sulu 1987). While Lagos app- cess with revenue collection or quality of service in roached the problem of solid waste disposal by auth- garbage pickup. Their contract renewal could be de- orizinglarge capital expenditures (World Bank 1985), termined by a periodic competitive bidding process. Ibadan implemented a citizen participation proced- In sum, the government will play an important ure in which private firms haul their garbage to des- role in implementing the new institutional setups re- ignated points to be picked upbyprivate licensed su- sulting from the policy options and reforms that mig- bcontractors orby the public sector. In Owerri the sol- ht be adopted. More systematic analyses of economic ution was to enter into a subcontract with the Germ- and institutional feasibility will follow in this study an firm SULO A.G., which made an unsolicited offer. Luger (1989) in a recent World Bank discussion Congestion, System Failures, and Pricing Policy brief argues for more private sector participation in solid waste collection in the Lagos area to increase its The fluctuations in the quality of public infrastruc- share of industrially generated waste up from the ture services observed in Nigeria are, in part, a result cunrent 7 percent. Luger breaks down solid waste of congestion in the use of the system. While the collection into the following subactivities: demand for the service from a public agency such as * pickup at the source and delivery to process- NEPA is a function of quality, the quality itself is a ing plant or transfer station; declining function of the quantity demanded due to * pickup at processing plant or transfer station congestion effects. The public agencies must con- and delivery to tipping site or resource recov- sider the trade-offs between the quantity supplied ery facilities; and the quality (and reliability) of services in deter- * transfer points, tipping sites, processing facili- mining the pricing policy, especially in the short run ties, or incinerators; when the ability to expand the system is limited. e maintenance of various facilities; and Treating congestion as endogenous is common in 0 administration including bill collection. transportation and other urban infrastructure sys- tems, and congested situations require the levying of While the Lagos State Waste Disposal Board an optimallyset congestion toll which will reduce the (LSvDB) could continue to maintain control over load and congestion to a socially optimal level. regulation, the remaining subactivities are candi- As an illustration, consider the electric power pric- dates for various forms of privatization on a case by ing by NEPA. Most power interruptions (nearly two case basis. For example, the private sector could be thirds) are a result of bottlenecks on the transmission induced to set up landfill sites or resource recovery and distribution networks. It is commonly observed facilities if they are allowed to produce gas, energy, that in the industrial areas in Lagos when large en- or compost which can be sold profitably. In finance, ergy intensive manufacturing plants such as steel bill collection can be contracted out, where the con- mills start operating, the resulting voltage surge tractor's payment is based on the percentage of out- often damages machinery and equipment of smaller standing revenues that are collected. Such a private firms located in the vicinity. Large energy intensive collecting entity would be more motivated than the firms place heavier loads on the system, thus tying existing bureaucracy to achieve full revenue accrual. up more operable transmission capacity. However, In the areas of pickup and delivery to intermediate these large firms are the ones which can afford to points, there is a variety of available options includ- have own generators, have a greater amount of un- ing direct delivery by the manufacturing estab- used generating capacity, and can produce electric lishment's own vehicles, pickup by private entrepre- power at a much lower average cost than small firns. neurs on a demand activated basis, or pickup by a In the case of NEPA, the congestion is so severe that private entity licensed to operate as a spatial monop- the system tends to fail completely resulting in fre- olv within a particular district. quent power outages. In such a situation, it would be Monitoring mechanisms for market operations and desirable to raise the tariff to a sufficiently high level service quality. As various infrastructure related func- to clear the market. For example, at a NEPA tariff of 119 Manufacturers' Responses to Infrastructure Deficiencies in Nigecia: Private Altenatives and Policy Options 50 kobos per kWh, large firms may find self-generat- ers; to observe the responses of the manufacturers to ion cheaper and use their own generators more fully, these deficiencies; and to develop viable policy op- thereby reducing congestion. Deregulation, to allow tions based on the observations from the data col- those firms to sell excess power, should provide add- lected. The results of the establishment survey re- ed incentives to own generation of power. Small fir- vealed general patterns of deficiencies and self-pro- ms will then have better access to the system. Public vision responses by manufacturers which cut across supply quality is expected to improve at the higher all five infrastructure subsectors included in the NEPA price which smaller users may find still lower study. In particular, in nearly all infrastructural ac- than the cost of self-generation. We have requested tivities, small firms face higher unit costs than larger NEPA to provide us with the necessary data to docu- firms do and the patterns of self-provision by firms ment statistically the correlation between loads on differ a great deal by region within the country as the transmission network and the frequency of pow- well as by type of firm. er failures, in order to measure the quality improve- Our main thrust in developing policy options is ments that can be expected from inducing firms with that the ongoing structural adjustments in Nigeria, different private provision capacities to reduce their including changes in pricing, regulation, and institu- use of the public supply in response to higher prices. tional structure in most sectors, need to be extended A more comprehensive study of the market struct- to managing and accommodating the costs of the ure, including NEPA's costs and variations in dem- widespread private provision of infrastructure serv- and by user types and locations is needed to determ- ices resulting from public sector failures. Because inetheorderofthepricethatwillremovecongestion. improvements in public sector performance are Producing specific tariff systems for individual likely to remain slow in the short and intermediate subsectors such as electric power, water, and teleco- terms, manufacturers and especially small firms will mmunications is beyond the scope of this study. In continue to bear the costs of self-provision. Further- this research project, however, we intend to quantify more, with the ongoing upward adjustments in tar- relative efficiencies of alternative pricing regimes by iffs the burdens of the deficiencies which are borne simulating the responses of different types of firms by small firms will increase. To ease these private to such regimesthat reflect particulartypes of market burdens and to improve the overall infrastructural structures. Possibilities for considering variations of provision in Nigeria, we have considered plausible the "two-part tariff", for example, were discussed in policy options in the following three areas: detail in the framework paper (Anas and Lee 1988). a Regulatory reforms such as the relaxation of Bahl and Linn (1989) present an excellent review of regulatory restrictions against the trading of pricing urban services. A recent paper by Heady infrastructural services among manufactur- (1988) stresses the role of public sector prices as instr- ers. uments of cost recovery and explains the Bank's two- * Private sector participation in contestable ma- step practice in setting public sector prices. The first rkets for the supply of infrastructural services, step calculates the marginal cost; the second step adj- wherever appropriate for selected subactivit- usts marginal cost to take account of other factors su- ies such as production, delivery, maintenance, ch as revenue shortfalls, market distortions, and dist- revenue collection, and finance, by means of ributional effects. All these factors are relevant to the various institutional mechanisms such as sub- Nigerian situation. Another recent paper by Julius contracting franchising, and districting. and Alicbusan (1988) documents the two-step appro- * Alternative pricing policies taking into ac- ach in more detail and surveys the useof such pricing count the capacity limitation and congestion policies in many countries and various public secto- effects on the service facilities. rs. A clear discussion of short-run marginal cost prici- ng, economic user charges, and budget deficits is giv- A set of more definite policy recommendations en in Meier (1983, pp.192-203) which is reprinted fr- will be provided later in the study on the basis of the om Walters (1968) and Bennathan and Walters (1979) formal empirical analysis to be conducted with the who also discuss nonlinear "two-part tariff' pricing. establishment survey data. In particular, econometr- ic work outlined in the framework paper (Anas and Conclusions and Further Empirical Study Lee 1988) and Verma and Lee (1988) will enable us to estimate key production and cost function paramet- The main objectives of this paper were to docu- erswhichwillprovidefirmquantitativebasesforpo- ment the extent, causes, and incidence of infrastruc- licy analyses. Such econometric models can be used tural deficienciesas they affect Nigerian manufactur- to simulate the responses of selected firms to various 120 Manufacturers' Responses to Infrastructurm Deficiencies in Nigeria: Private Alternatizvs and Policy Options policy, changes. Such simulations are essential in ord erations and Maintenance." The World Bank, (mimeo). er to obtain better insight about the probable econ Luger, M.I. (1989), "Privatization Options for the Solid omicbenefits that are likely to result from the policy WasteSectorinNigeria."lDiscussionBrief, Infrastructure options and the implementation strategies which we and Urban Development Department, The World Bank. have dliscussed. McGuire, M. (1974), "Group Segregation and Optimal Ju- risdictions.' Journal of Political Economy, Vol.82: 112-132. References Meier, G.M. (1983), Pricing Policy for Development Manage- ment. The Johns Hopkins Press. (Published for The Eco- Anas, A., and K.S. Lee. (1988), "Infrastructure Investment nomic Development Institute of the World Bank.) and Productivity: The Case of Nigerian Manufacturing, Meyer, J.R., and M.R. Straszheim. (1971), PricingandProject A Framework for Policy Study." Infrastructure and Ur- Evaluation. The Brookings Institution. bar, Development Department, Discussion Paper, Re- Munasinghe, M. (1979a). Economics ofPower System Reliabil- port I NU 14. The World Bank ity and Planning. Johns Hopkins Press. Bahl, R.W., and JF. Linn. (1989), Urban Public Finance in Munasinghe, M. (1979b), "Electric Power Pricing Policy,' Developing Countries. The World Bank. (mimeo) World Bank Staff WorkingPaper. No. 340. TheWorld Bank Baumol, W.J., and D.F. Bradford. (1979), "Optimal Depar- Munasinghe, M. (1988), "Energy Economics in Developing ture from Marginal Cost Pricing." American Economic Countries: Analytical Framework and Problems of Ap- Review, 256-283. plication." The Energy Journal, Vol. 9, No. 1. (Available as Baumol, W.J., J.C. Panzar, and R.D. Willig. (1982), Contest- World Bank Reprint Series: Number 421.) able Markets and the Theory of Industrial Structure. Har- Roth, G. (1987), Private Provision of Public Services in court Brace Jovanovich. Developing Countries. Oxford University Press. Beesley, M.E. (1981), Liberalization of the Use of British Tele- Sulu, S.A. (1987), "Development of Maintenance Strategy communications Network. Department of Industry, Re- for Urban Solid Waste Collection." A Paper Presented at port to the Secretary of State, Her Majesty's Stationery the 1987 National Workshop on Engineering Mainte- Office, London. nance Organized by the Nigerian Society of Engineers. Bennathan, E., and A.A. Walters. (1979), Port Pricingand In- Verma, S.K. and K.S. Lee. (1988), "Measuring the Produc- vestment Policy for Developing Countries. Oxford Univer- tivity of Infrastructure Services in Nigerian Manufactur- sity Press. ing: A Theoretical Framework." The World Bank Coyaud, D. (1986), "Private and Public Altematives for Pr- (mimeo). ovidingWater Supply and Sewerage Services," in Mana- Walters, A.A. (1968), The Economics of Road User Charges. gement Optionsfor Urban Services. The W'orld Bank. (Rep- The Johns Hopkins Press. ort on Seminar held at Cesme, Turkey, November, 1985.) Whittington, D., D.T. Lauria, and X. Mu. (1989), 'Paying Heady, C. (1988), "Public Sector Pricing in a Fiscal Con- for Urban Services: A Study of Water Vending and Will- text." Policy, Planning and Research Working Papers, ingness to Pay for Water in Onitsha, Nigeria." Infrastruc- WPS 179, The World Bank, April 1989. ture and Urban Development Department, Case Study, Henney, A. (1987), "The Operation of a Power Market." (A Report INU 40. paper prepared for the government and the electricity World Bank. (1981). Nigeria: The Power Transmission and supply industry as a contribution to the debate on how Distribution Project. Report No. 3041-UNI. to promote competitive power generation), London. World Bank (1983a). Nigeria Macroeconomic Policies for Jutius, D. and A.P. Alicbusan(1988), "Public Sector Pricing Structural Change. Report No. 4506-UNI. Policies: Areviewof Bank Policy and Practice." Policy, World Bank (1983b). Nigeria Issues and Options in the En- Planning and Research Working Papers, WPS 49, The ergy Sector. Report No. 4440 UNI. World Bank. World Bank. (1983c). Nigeria Lagos Urban Sector Review. Lee, K.S. (1985), "An Evaluation of Decentralization Poli- Report No. 4479-UNI. cies in Light of Changing Location Patterns of Employ- World Bank (1985). Nigeria Lagos Solid Waste and Storm ment in The Seoul Region." Urban Development Discus- drainage Project. Report No. 6375-UNI. sion Paper, UDD-60. The World Bank. World Bank (1986). Trade Policy and Export Development Lee, K.S. (1989a), The Location of Jobs in a Developing Metropo- Loan. Report No. P4358-UNI. lis: Patterns of Growth in Bogota and Cali Colombia. Ox- World Bank. (1988). The Nigerian Structural Adjustment Pro- ford University Press. gram: Policies, Impact and Prospects. Report No. 6716-UNI. Lee, K.S. (1989b), "Infrastructure Constraints on Industrial World Bank (1988). Nigeria Lagos Water Supply Project. Growth in Nigeria." The World Bank, (mimeo). Report No. 6375-UNI. Lee, K.S., J. Stein, and J. Lorentzen. (1989), "Urban Infrastr- World Bank (1989). Nigerix Power System Maintenance and ucture and Productivity: Issues for Investment and Op- Rehabilitation Project. Report No. 7607-UNI. 121 12 Comments on Fiscal Deficit and Expenditure Policy Yao Kouadio and Nii Kwaku Sowa Yao Kouadio the authors build a very simple, yet instructive, model in which the standard of living is a linear T-he Ferroni and Kanbur Paper function of social expenditures and the level of in- come, and income is a function of basic needs and The main objectives of Ferroni and Kanbur's inter- productive expenditures. The interesting feature of esting paper are to present an overview of public the model is that it allows for feedback effects as far expenditure patterns in Sub-Saharan Africa, to de- as basic needs and income are concerned. velop a theoretical framework for retargeting public Given the budget constraint, the authors attempt spending toward poverty alleviation, and to under- to provide some simple decision rules as regards take some case studies based on the rules derived public expenditure restructuring. The major conclu- from the theoretical section. sion that emerges is that a necessary condition for I would like to indicate at the outset that I was a favoring basic needs expenditures hinges on bit surprised by the finding according to which ad- whether or not the direct effects of those expendi- justment programs in Sub-Saharan Africa have not tures exceed the indirect effects that are induced by really led to a reduction in real public expenditures. increases in income. As correctly emphasized by the Likewise, it seems odd that at the sectoral level, in authors, this is clearly not a sufficient condition for opposition to what might be expected, social and the social valuation of basic needs and income. economic sector public expenditures have not de- Obviously, one could question the linearity ass- clined despite the need to adjust most economies umption that is put forward in the theoretical model. through a reduction in domestic absorption. But this is not the crux of the matter, since most of the Whatever the actual trend in overall spending, the implications remain valid even under nonlinearity. authors point out that intrasectoral expenditure pat- What is more important is related to the issue of terns remain inadequate in the sense that there is a estimating social weights among the different di- disproportionate resource allocation toward higher- mensions of the standard of living. Carrying the order services, that is those going to the wealthy. analysis one step further, Ferroni and Kanbur con- Thus, the poverty alleviation impact of public expen- sider the effect of restructuring public expenditures ditures is less than it could be. on poverty using Foster, Greer and Thoebecke's pov- Following this overview of public expenditure erty index. The implications of their analysis are pattems is a very fine section on retargeting of public contained in two formulae that are to be compared. expenditures, which raises a number of challenging The basic finding here is that the impact of public questions of a normative nature. Recognizing that expenditure restructuring on poverty depends on the standard of living is a multidimensional concept, the current shares of total expenditures reaching the 122 Comments on Fiscal Deficit and Expenditau Policy poor. This insight is probably the major contribution and is therefore to be adopted given the current size of this paper. of the public finance deficit. After some digression on the Human Develop- I do not have any quarrel with the above analysis ment Index, the usefulness of which is surely debat- if ne accepts, as a primary goal, poverty alleviation able, the paper introduces some empirical evidence instead of some otherobjective. It shouldbe made cl- based on some African countries. Thus, regarding ear, however, that the conclusions do depend on the C6te d'Ivoire and the issue of rice pricing policy, data set and that one should perhaps undertake sens- Ferroni and Kanbur suggest, in line with the theoreti- itivity analysis, particularly as far as the poverty line cal prescriptions, that increasing the consumer price is concerned. would be preferred to reducing the producer price if On the whole, I find this paper very stimulating, poverty alleviation is a priority for policymakers. although one feels frustrated at the difficulty of con- Still, regar- ding housing policy in that country, the structing true standard of living indices in order to conclusion reached by the authors is that public dis- facilitate the assessment of public expenditure re- engagement is unlikely to be detrimental to the poor structuring. Nii Kwaku Sowa My comments on the three papers in this section have issue of how inefficiencies in infrastructural services to do with the lessons that African policymakers can get translated into unnecessary costs for manufactur- derive from them to help shape up their public ex- ers, and Ferroni and Kanbur provide solutions to penditures. expenditure management. Most African governments, soon after independ- ence, embarked on various programs aimed at mod- The Rutayisire Paper ernizing or developing their economies. A number of them adopted industrialization as a development go- In the paper on Tanzania, Rutayisire sets out to al and, in the process, established several import sub- resolve what he thought was a conflict of opinion on stituting industries. These industries, which were govemment expenditure in that country. On the one mainlystate owned, were supported byweakinfrast- hand is the concern about the reductions in the size ructural bases and have weak forward and backward of government expenditure, and on the other hand is lin'kages. Revenue generated from most of these ent- the view that the level of govemment expenditure is erprises could not support them and some had to rely still too high. In my opinion, there is no conflict of on government (subvension) to pay their workers. opinions: one can be cutting down on government The state, in most African countries, also was the expenditure, but so long as it has not come down to provider of basic utility services such as electricity, desirable levels, it can be said to be too high. water and waste disposal. These services were pro- Rutayisire sets out to determine the optimal and vided at very subsidized rates to make them afford- sustainable level of government expenditure for able to the poor. Tanzania. His results show that theyears of structural In the face of mountingexpenditures, African gov- adjustment and economic recovery (1981 to 1988) ernments have had to contend wi th very narrow and have been years of unsustainable government spend- weak tax bases, which have led to budgetary prob- ing. He unfortunately did not extend his analysis to lems. By the beginning of the 1980s, all countries in pre-structural adjustment and economic recovery Sub-Saharan Africa, except a few like Botswana, program years for us to compare. However, this re- were saddled with budget deficits which had to be sult by itself resolves what he thought was a conflict financed by governments printing money. These led of opinion. Structural adjustment and economic re- the countries into other economic problems like in- covery programs call for restraint on government flation and balance of payments difficulties. spending and yet, still public expenditure in The three papers by Rutayisire, Ferroni and Tanzania remains unsustainably too high. Thus, Kanbur, and Lee and Anas look at the fiscal problems either Tanzania expands its tax base or there must be of Africa from differentangles. Rutayisire looksat the further cut-back of government spending with problem of budgeting, Lee and Anas examine the proper targeting (Ferroni and Kanbur). 123 Comments on Fiscal Deficit and Expenditume Policy The issue of sustainability of government expen- longer term solution will be for proper layout plan- diture will, however, have to be looked at carefully ning before industries are allowed to set up. since it depends on certain factors which may not be Lee and Anas' calculations also show that firmsare properly measured. Budgetary procedures in most already sharing high costs in trying to supply infra- African countries are haphazard. Estimates of gov- structural services themselves. It follows that firms emnment spending do not follow proper costing pro- will be prepared to pay higher prices now for better cedures and rely mostly on percentage increases. public service. Governments are, however, reluctant This can either overestimate or underestimate gov- to increase the price of these utility services which emnment spending. In addition, inflationary expecta- they consider as part of the basic needs package for tions, if not well anticipated, may distort calculations the citizenry. for sustainability. In any case, even if government spending is unsustainable and calls for budgetary The Ferroni and Kanbur Paper cuts, care should be taken not to cut essential serv- ices. Cuts in sectors which provide infrastructural If poverty conscious expenditure is what the gov- support services can lead to serious repercussions in emient is thinking of, then Ferroni and Kanbur's other sectors of the economy. tools become useful. A closer examination may show that the proportion of the poor who use these utili- The Lee and Anas Paper ties-electricity, water (pipe-borne) and so on-is quite small. Thus subsidies on any of these utilities Lee and Anas, with Nigeria as a case study, look at may be favoring the rich rather than the poor. the effect of inefficiencies in infrastructure provision Some allusion has been made already to the use- on manufacturers and the latter's response to such fulness of Ferroni and Kanbur's paper. The follow- deficiencies. This study is particularly interesting in ing comments are directed to specific issues in the the way it unearths a problem which can be found in paper. most African countries. The paper considers the de- * The first part makes some generalizations ficiencies in the provision of services such as electric- about public spending in Africa. A doser look ity, water, telecommunication, transport, and waste at some countries will reveal that, contrary to disposal. It then examines the cost of these externali- what was expressed in the paper, real govern- ties to the manufacturers. It observes that because of ment expenditure has fallen in some countries. frequent interruptions and the inefficiencies in the With current efforts at cost recovery in both provision of these services, most manufacturers have health and education, certainly expenditure stand-by units to support them. Firms which do not on these sectors would have to be dropped. have any stand-by units suffer production loss not * It is true, as the paper observes, that anti-pov- through their own inefficiencies, but by picking up erty spending must encompass economic sec- external diseconomies from public corporations. tor spending as well as provision of basic Policies recommended by Lee and Anas include needs. But for most people in Africa who live changes in the regulation for the provision of these on the brink of survival, direct expenditure on services to allow private entrepreneurs to enter the basic needs leads to a bigger effect on their market. This move is already being implemented in standard of living than an indirect effect most countries under structural adjustment pro- through income increase. grams. However, the nature of these services some- * The paper attempts to develop a social welfare times dictates the existence of a natural monopolist- achievement index which is intended as an im- in Africa the State. provement on the UNDP's Human Develop- There is also the suggestion that the market be ment Index (HDI). The criticism against the deregulated to allow firms with stand-by units to sell HDI is that it weights each of the social indica- their services to firms without stand-by units. This tors equally. The Ferroni-Kanbur indicatorhas can only be a short-run solution, since a private a stronger theoretical foundation, but ends up company with a generator will sooner or later run with equal weights for the social indicators. into worse problems than the public company. There Moreover, whereas the HDI has three social was a hint that sharing of generators by firns will indicators, the Ferroni-Kanbur index has only ease congestion on the public lines. A better and two. 124 13 The Informal Financial Sector and Markets in Africa: an Empirical Study ErnestAryeetey and Mukwanason Hyuha Difficulties in comprehending the way the informal sector include moneylenders, rotating savings and sector operates have led to inaccurate speculation credit associations (ROSCAs), mobile bankers and about how informal financial activity influences the susu operators, mutual assistance groups, landlords, impact of monetary and other economic policies in neighbors, friends and family members, etc. many African economies. The potential of the infor- Ideally, analysis would be based on a disaggrega- mal sector is usually underestimated; this sometimes tion of informal financial activities in order to capture leads to the pretence among policymakers that this more accurately the functioning of informal groups, activity is inconsequential. Even where it is acknow- and their separate impacts on the economy. How- ledged that a substantial infonnal financial sector ever, this is not possible presently in view of the exists, policymakers are usually at a loss with regard limited information available. To amend the situ- to capturing its activity within a policy framework. ation, even if partially, most of the illustrations here The sector thrives and affects growth processes in are based on dominant activities and groups. Thus, several ways; however, it is either unaccounted for for example, the study of infornal savings mobiliza- or only improperly accounted for in the national tion is Ghana dwells mostlyon thesusu system using accounts. the single collector. This system involves more than Here informal financial market refers to all finan- 95 percent of the market women who save informally cial transactions that take place beyond the func- (i.e. 76 percent of the total sample), and the market tional scope of various countries' banking and other women are considered to be the largest group of financial regulations. Those that are covered by these informal savers in Ghana (Aryeetey and Gockel, regulations belong to what are called here the formal 1990). financial sector and these are "by definition, bound Current interest in what role the informal financial by the laws and regulations that governments have sector can play in the economic transformation of enacted for the control, supervision, or facilitation of many African countries is derived from the dearth of that sector" (USAID, 1989, page 7). Thus the formal financial resources forthecapitalizationprocess. Evi- sector includes the commercial banking system, dence of the inadequacy of existing resource alloca- near-banks, insurance companies, housing develop- tion and resource allocating mechanisms is apparent ment banks, investment banks, etc. The informal in the severe economic crises the region has suffered financial sector includes a wide range of financial since the mid 1970s. The crises reached alarming activity whose extent may differ from country to levels in the early part of the 1980s when the econo- country. Institutions and activists in the informal mies of most African countries were epitomized by 125 The Informal Financial Sector and Marnfts in Africa an Empirical Study severe stagnation and decline. These problems arose expressed about the nature and workings of the in- from deep-seated structural imbalances in most of fonnal system. An attempt is then made to hypo- the economies of the region. thesize within a very static framework the relation- To arrest the situation, many African countries are ship between the formal and informal segments of now pursuing various structural adjustment pro- the financial market in the fourth section. A summ- graams aimed at economic recovery and growth. In ary and concluding remarks appear in the final sec- view of the austerity posed by many of these pro- tion. grams, governments have had to follow tight fiscal and monetary policies. These are reflected partly in Characteristics of the Informal Financial Sector in reforms of the formal financial sector and public Five Selected African Countries expenditure policy. These reforns invariably imply substantial real reductions in the amount of credit This section presents a review of the informal fi- available to the private corporate and household nancial sector in five African countries-Ghana, sectors, which thereby forces more and more people Malawi, Senegal, Tanzania, and Zaire (see table to finance various economic activities informally. But 13-1). The five cases address the following issues: relatively little is known of the mechanisms inherent a the existence, spread, and estimated size of the in this activity. sector, Recent studies (Aryeetey and Gockel 1989; Hyuha, e the types of institutions operating for savings Ndanshau and Kipokola 1989; Chipeta and Mkan- mobilization, credit extension, and safekeep- dawire 1989) in Ghana, Tanzania, and Malawi have ing functions; helped to throw some light on the internal workings 9 the loans extended by the informal financial of informal markets, as well as on some of the rela- sector and the average operating costs of the tionships between them and the established or for- sector, mal financial market. This paper attempts to stream- * interest rates and other charges on loans; line the various country studies and map out com- * the relationship between the informal and for- mon areas of operations and similarities in the vari- mal financial sectors. ous informal markets. The information obtained from the primary studies in the three countries is Ghana supplemented with infonnation from recent studies in Zaire and Senegal by the US Agency for Interna- Under the original forms of land tenure systems in tional Development (1989). Ghana, rich landowners allocated parcels of their It is argued in this paper that the infonnal financial land to squatters or migrants for farming, with the sector in Africa is relatively large and plays a crucial condition that one-third or one-half of the total pro- role in savings and investment. Thus, for optimal ducewould begiven to the landowner. Followingthe financial resource allocation, informal financial insti- monetization of the economy with modem admin- tutions and markets need to be given due considera- stration, this transaction was increasingly made in tion in policy design. The paper presents the charac- monetary terms. This practice eventually dev- teristics of the infonnal financial sector in five Afri- eloped into the modern money lending business in can countries. It then looks at some of the ideas ruml and urban communities. Aside from money Table 13-1. Some macroeconomic features of the five countries Gross Gross Gross Gross Average national domestic domestic domestic M2 annual savings savings investment investment money inflation as a as a as a as a supply as a Population GNP per rate percentage percentage percentage percentage percentage (millions) capita (percent) of GNP of GDP of GNP of GDP of GDP Country 1987 (dollars) 1980-87 1980-87 1987 1987 1987 1987 Ghana 13.6 390 48.3 2.0 4 7.1 11 11.7 Malawi 7.9 160 12.4 7.0 12 19.0 14 25.0 Senegal 7.0 520 9.1 -2.8 6 15.5 13 23.5 Tanzania 23.9 180 24.9 9.0 -6 18.7 17 25.9 Zaire 32.6 150 53.5 4.9 10 14.3 13 8.8 Source: World Bank,World Development Report 1989, Oxford University Press. 126 The Informal Financial Sector and Markets in Africa an Empirical Study lending, other forms of informal financial activity is 6 months at an interest rate of 50 percen. They are include the single collector susu system or mobile patronized in spite of the relatively igh interest rates banking, the rotating susu system or ROSCAs, mu- because those who require tnem usually have no tual assistance groups or clubs, and credit unions. other choice, and it is relatively easy forbrrotwers to The most widespread of these forms of saving and acquire funds. In many cases, the borrower's expec- borTowing is the single collector susu system, which tations of a specific income within an agreed period, is believed to have begun in Ghana with migrant an introduction by someone known to the lender, Nigerian traders (IPC 1988). With this activity, a col- and a written undertaking are enough to secure the lector (usually male) visits his clients at their work borrower credit. The default rate is said to be quite places, market stalls, shops, and homes and collects low, but is believed to have gone up in the last funds towards a savings plan. Savers deposit fixed decade, when personal incomes suffered greatly. amounts for an agreed period of time, after which Credit unions and mutual assistance groups often their savings, less an agreed sum, are returned. have longer tenn facilities. Over 76 percent of 1,000 market women inter- The ratio of deposit money to M2 money supply viewed by Aryeetey and Gockel (op cit) in Accra, and the ratio of the banking sector's claims on the Kumasi, and Takoradi (the three largest centers in private sector to GDP have been used to measure the Ghana) saved in this manner. The study of rural hou- size and relative significance of the informaP market seholdsbyIPC (op cit) also showed that this was the (Aryeetey and Gockel op cit and Wai 1957). They most prevalent form of financial savings in those have both shown a relative decline in the strngth of areas. What is interesting about this form of savings the formal banking system and therefore significant is the fact that no interest is earned on deposits. Rat- growth in the informal sector. For rural Ghana, IPC her, the saver pays a service charge through the sav- (op cit) estimated that 60 percent of all financial ings retained by the operator at the end of the agreed savings would be done informally. It was estimated period of the savings plan, usually one month. Thus, that an association of 30 susu collectors in Tamale for a thirty-day plan, a day's savings is retained. mobilized about C30 million in savings in a month. The susu system permits access to credit for vaA- This contrasted with the mid-year total savings mo- ous purposes. For as many as 31 percent of the mar- bilizedby thebanks in the area of Cll milhlon in 1988. ket women saving with susu collectors, access to The informal sector was more significant for private credit on flexible terms was the most important rea- household sector savings mobilization and credit son for saving. A major characteristic of these loans, granting than was the formal financial sector. The however, is the fact that they are usually of a short fact that a substantial portion of savings accumu- term nature. They often must be repaid within three lated in the susu system is used to finance working months. In 25 percent of the cases studied, they had capital makes financing through the sector essential to be repaid within one month. In most cases no for small-scale investment. interest was levied on credit acquired this way. There are significant links between the formal and If the credit facility should attract interest at all, informal financial sectors on the deposit and lending interest payments in this widespread informal activ- sides. It has been noted that, "most of the 'susu' ity are as in the following example. A woman who operators deposit at the end of each day the total normally deposits one hundred cedis (ClO0) daily savings mobilized from the markets in one of the towards a 90-day savings plan would accrue C9,000. banks near the market" (Aryeetey and Gockel, op She maybe permitted a loan of C4,500 in addition to cit). This behavior was attributed to the security it her bulk sum of C9,000 (less the service charge). She affords the operators. Most of the bank deposits then pays C120 dailyto the collector for the next three made by the susu operators are of the demand de- months. Half of her normal deposits (C4,500 less posit type since the operators need to stay liquid at service charge) is returned to her, the other half plus all times to make payments at short notice. Similarly, C1,800 are used to repay the principal and interest on it is observed that many moneylenders operate bank the loan. The market women have very little problem accounts. They sometimes go as far as borrowing with this arrangement since the 40 percent interest from banks, if possible, and relending to their cu- rate over three months is paid daily in relatively stomers. In view of these links between the two sec- small amounts (C20 in this example). This may be tors, it is indeed likely that, through a proper assess- compared to an average lending rate of 30 percent ment of the actual cost and demand structures of the per annum charged by banks. two sectors, the infornal sector couldbe made comp- Moneylenders often make credit available for be- letely subject to various monetary and fiscal policies tween 3 and 12 months. The average maturity period of the government, without losing its character. 127 The Informal Financial Sector and Markets in Africa an Empirical Study Malawi The 'Katapila' charge high interest rates basically because they evolved customarily within such a tra- According to Chipeta and Mkandawire (1989), in dition. The scarcity of funds and expectation of high Malawi the infornal financial sector is larger than the default rates should also explain the high rates pay- formal financial sector. Using the ratio of informal to able to the moneylenders. formal lending to the private sector, they estimated The formal and informal sectors are closely linked. that the informal financial sector was about three Chipeta and Mkandawire (1989) describe this rela- times as large as the formal. tionship as follows: Institutions in the informal sector range from mon- "In general, the activities of the informal financial eylenders (Katapila) to friends and neighbors. Table sector complement those of the regulated formal fi- 13-2 shows the types of informal financial institu- nancial sector. The informal financial sector finances tions that exist in Malawi and some of the interest business enterprises that the formal financial sector rate charges on loans extended by the sector. Estate does not finance. It also finances consumption of owners, friends, and employers are significant insti- households that do not have access to formal finan- tutions in the informal financial sector (Chipeta and cial institutions. But despite its success, the informal Mkandawire 1989). financial sector does not meet all the demand for Informal sector loans are extended for consump- loans." tion and investment purposes. Consumption loans It is also suggested by the two researchers that in usually come from Rotating Savings and Credit As- view of its size, the informal financial sector plays a sociations (ROSCAs or Chiperegarii) and may be major role in the savings and investment process in spent on consumer durables. The loans are given Malawi. As such the formulation and implementa- withoutmuchbureaucracyand promptly, withmini- tion of policy in the country should take account of mal or no paperwork. Except for the moneylenders, the informal financial sector. interest rates on loans are either nonexistent or very low compared to those in the formal sector, as table Senegal 13-2 shows. The loan repayment rate is quite high despite the absence of fonnal collateral. The United States Agency for International Devel- On the low interest rates, Chipeta and Mkan- opment (1989) study of Senegal covered mainly ur- dawire (1989) write the following- ban areas. The relatively rich groundnut region "Most of the traders, farmers, friends, relatives and and a few other rural areas were also included. The neighbors do not seem to charge interest on the loans report maintains that: that they grant. Traders may be looking at free loans 'The informal sector, defined as that portion of as a means of maintaining and improving relations economic activity which is for the most part outside with their customers, while the rest may be looking the domain of the formal legal system and govern- at it as a means of maintaining solidarity." ment, is very large...The team believes that approx- Table 13-2. The infonnalfinancial sector in Malawi: structure and interest rates Nominal rate (annual percentage) Real rate (annual percentage) Institution Deposit Lending Deposit Lending Moneylenders n.a. 300-1200 n.a. 285-1185 Estate owners n.a. 2-25 n.a. -13-10 Traders n.a. n.a. n.a. n.a. Grain millers n.a. n.a. n.a. n.a. Smallholder farrners n.a. n.a. n.a. n.a. Savings & credit associations 94-100 24-720 79-85 9-705 Cooperative savings associations 0 0 0* 0* Community funds n.a. 12-600 n.a. -3-85 Employers n.a. 4.5-14 n.a. -10.5-1 Friends n.a. 0 n.a. 0* Relatives n.a. 0 n.a. 0* Neighbors n.a. 0 n.a. 0* Real rate is zero minus rate of inflation. Source Chipeta and Mkandawire (1989), Annex table 4, p. 30. 128 7Te Informal Financial Sector and Markets in Africa an Empirical Study imately 60 percent of the population in Senegal ...earn Table 13-3. Major informal units and the uses of their living primarily in this sector of the economy their credit in Senegal Information regarding the amount of income or fu- Institution _Uses of its credit ll-time employment generated by the inforrnal sect- Rural or is not available, but it is clear that most people Village savings associations Relgious feast participate in the infonnal sector at least part of the (Caises Villageois) Pilgrimage to Mecca time. Informal sector activities touch virtually every- Mutual aid funds one." Communal projects, etc. Table 13-3 shows the types of informal financial Family savings Family use (self-finance, institutions that exist in Senegal. Loans from the etc.) informal financial sector are usually of a short term ROSCAs (tontines) Purchase of durables nature and in general the interest rates applied are Otherconsumption very low. As in Ghana and Malawi, the operating Business finance costs per unit are relatively low, thus making the Urban informal financial sector units low cost producers of ROSCAS (tontines) Purchase of durables financial services. This is attributed to the absence of Cotrsumption red tape. The possession of collateral security is not Market rotating savings investment required and literacy does not play an important role Consumption in loan acquisition. Credit from the sector is available Market savings groups Mainly investment for investment and consumption purposes. (banks) In Senegal the informal financial sectoris generally Source. USAID; Informal Financial Markes: Senegal and regarded as a necessary complement to formal serv- 7aire (1989). ices. There is a direct link between the two sectors in that members of the market savings group actually bank their savings with the formal system. It is there- largest amounts of credit go to crop finance, food. fore likely that policies initiated for formal financial and business finance. These loans are of a short term development will have some effect on the informal nature. The informal financial sector appears to de- sector. In 1989 itwas reported that the informal sector rive at least some of its patronage from borrowers was experiencing a credit squeeze and a liquidity who find the formal sector too complex: problem following structural adjustment policies in- . . .the most impending factors can be summarized itiated by the government to affect primarily the as follows. First, the formal financial institutions are formal system (USAID, 1989). very complex and amorphous in the eyes of the peasants... The second factor is the time involved in Tanzania the process of dealing with the formal financial insti- tutions. In fact, up to 32.4% of the respondents H-iyuha et al (1989) generated inforrnation on the reported that it is too time-consuming to deal witht informal financial sector in Tanzania in a survey theinstitutions.Third,peasantsandsmall-scalebusi covering262 households in seven regions during the nesses normally do not have the right type of last half of 1989.1 There is no estimation of the size of collateral security demanded by the fornal financial Tanzania's informal financial sector, however, it is institutions. Fourth, the process is not only complex likely to be as large as the formal financial sector. In and time-consuming but also bureaucratic. Lastly, general, it appears that there is complementarity be- the formal financial institutions are urban-based... tween the two financial sectors. The major link be- (Hyuha and Ndanshau, 1990, p. 11-12). tween them is the semiformal cooperative societies, Interpersonal relationships (friend, ethnicity, which obtain credit from the formal sector and lend neighbor, etc.) explain the nature of interest rates, to the informal sector. which are usually very low in the infonnal financiai Table 13-4 shows the types of informal financial sector. Further, a lot of saving and lending goes on in, units in Tanzania and the share of total households kind, rather than in money fonn. using each. The most common informal financial unit is the semi-official cooperative society. Table Zaire 13-5 shows the major purposes of informal sector borrowing in Tanzania, by share of the total sample. The size of the informal sector in Zaire has not Y, Informal sector loans are used for consumption and been ascertained. It is estimated to be considerabil investment. According to the sample evidence, the large since about "85 percent of the population in 129 The Informal Financial Sector and Markets in ApTica an Empirical Study Table 13-4. Types of informalfinancial units in Table 13-5. Purpose of infonnal sector bovuowing Tanzania by households in Tanzania Share of total households Share of total sample Type of informal unit (pe-rcftage) Expenditure iten (perenlage) Moneylender 1.6 Cooperative society 35.7 Food consumption 202 Landlord 2.2 Crop production' 24.6 Shopkeeper/trader 4.9 Purchase of farm machinery 13.4 Family member 19.9 Purchase of livestock 3.7 Neighbor 11.1 Business/trade finance 17.1 Household in the village 9.5 Building of house 5.0 Househod outside the village 1.5 Children's education 9.1 Church/mosque 0.6 Weddings 6.1 Other 10.6 Source: Hyuha et al (1989) Note. Includes financing weeding and harvesting activities and machinery (ox-ploughs, tractors, etc.). Source Hyuha and Ndanshau (1990), Table 5, p.22. Zaire earn their living primarily in this sector" Review of the Literature and Some Theoretical (USAID 1989, p. 95). Underpinnings for the Existence of Informal The informal financial sector in Zaire is dominated Financial Markets by semiformal cooperative societies: "Zaire's semiformal financial markets are domi- Many economists interested in African develop- nated by cooperatives which lend primarily to the ment are beginning to realize that the informal finan- national government by purchasing its treasurybills. cial sector has a very important role to play in the gro- The small portion lend for commercial and personal wth and development process of the continent. This purposes is shrinking further in relative importance. is especially true in rural areas (Rahman 1989). Sev- Cooperatives receive substantial deposits from eral propositions have been made in explaining why members, mainly for safe-keeping. Where they exits informal financial markets continue to thrive despite the co-ops are well-located and acc- essible to small the official encouragement and support of a more urban savers, but pay heavily negat- ive real interest modern and sophisticated formal financial market. rates, are poorly supervised and sub- ject to frequent fraud and theft (USAID 1989, p. 92). Economic Dualism Other units include mutual aid societies, ristour- nes (tontines), private banking agents, cooperative Ever since Wai's (1957) pioneering study of the credit unions, mutual savings associations, and pri- informal financial markets, most of what has been vate banks. They are utilized mainly for safekeeping done in this area has been mainly descriptive. In most purposes; real interest rates are highly negative. of the descriptions of informal financial markets, The USAID study covered only the Shaba region however, their links with formal financial institu- of Zaire. Nevertheless, it is evident that loans are of tions have been portrayed as weak (Ghatak 1981). a short term nature and interest rates charged on the Some of these suggestions of weak links find their loans are very low. The loans are for consumption roots in earlier theories of economic dualism. These and investment purposes. suggest that, in many African and other developing There is evidence of some interrelationship be- countries, the modem sector fails to absorb the tradi- tween the formal and informal sectors. The coopera- tional sector completely largely due to inherent so- tives sprang up and have grown tremendously in the ciocultural reasons. It is difficult to accept this argu- last five years, mainly because of the near collapse of ment, however, in view of the fact that modem bank- the formal sector. As in Senegal, structural adjust- ing facilities cannot be found throughout the econ- ment policies have not checked inflation and have omy in most African countries. led to a serious cash scarcity (liquidity squeeze) The earliertheoriesof economic dualism are linked which favors urban over rural areas. Thus, macroeco- to the expectation that, as the economies of less de- nomic policies directed at the formal institutions and veloped countries grew and became modernized, the activities also affect the informal market. fonnal financial sector would tend to absorb the 130 The Infonnal Financial Sector and Markets in Africa: an Empirical Study informal sector (Ghatak 1981). The contrary is sug- system does not function adequately, as in rural ar- gested by evidence from Ghana. In Ghana, the infor- eas, the informal sector provides a very much appre- mal financial sector has continued to grow even after ciated substitute. Many people continue to patronize reforms to improve efficiency and strengthen capi- the facilities of the informal sector, especially in rural talization levels were introduced into the formal sec- areas, in view of the monopoly power enjoyed by tor by the government in 1985. Operators in the single units of the sector. Miracle etal (1980) think the informal sector have simultaneously introduced sev- use of formal banks is not as widespread as is usually eral innovations in order to keep up with the compe- imagined. tition. "The great bulk of the African population makes little or no use of formal savings and lending institu- Inadequacy of Formnl Institutions tions. There are few banks in most areas,... (1n) African countries over 70 percent of the population The orthodox explanation of the need for credit by is rural, and banks are found almost exclusively in small-scale entrepreneurs and peasant farmers the larger urban centers. However location is only hinges on poverty. Therefore, many formal financial part of the problem. Even the portion of the popula- institutions have been set up by African govern- tion near urban centers makes strikingly limited use ments to provide the required external finance and of formal financial institutions." (Miracle et al 1980, to pool together the meager domestic financial re- p. 701-2). sources in the countries. However, these develop- ment banks have not adequately satisfied the needs Complementarity of large sections of the intended clientele (Mauri 1987). The bureaucratic, complex, amorphous, and The Aryeetey and Gockel work (1990) makes it unfamiliar aspects of these institutions has distanced clear that the informal financial sector has managed them from the targeted clientele. They are thus seen to develop self-sustaining mechanisms that have to be beyond the reach of peasant farmers, small- helped it to achieve a greater level of product diff- scale entrepreneurs, and ordinary household savers erentiation under term conditions. These have been (see Ghatak 1981). Formal financial institutions are made possible through innovations in savings mo- regarded asbenefitingonlysomeurban dwellersand bilization techniques. Also, a lot of the recent lit- large-scale farmers and entrepreneurs, particularly erature based on empirical studies of informal finan- those in the commercial and trade sectors. In anycase cial markets appears to support the view that these available fonnal credit is inadequate, difficult to ob- markets often complement the activities of formal tain, and urban-biased. financial institutions. This is especially true for co- In view of the difficulties associated with formal mmercial banks. In areas where such banks exist, credit arrangements, the credit starved small peasant people demanding credit use both markets simul- farmers, small-scale entrepreneurs, and households taneously (Hyuha, Ndanshau, and Kipokola 1989, have had to make alternative arrangements for fi- and Chipeta and Mkandawire 1989). The informal nancing their expenditures. Many people have had financial sector's self-sustaining mechanisms, it is to resort to the informal financial sector. Over time argued, assist it in performing this complementary the informal financial sector has grown in size and role. importance. It is believed to be larger than the formal Seibel (1986) is also of the opinion that both sectors sector in savings mobilization and credit allocation have their own peculiar strengths which guarantee to households in some countries (as observed in them distinct markets. This may be corroborated by Malawi and Ghana). Mauri's (op cit) study of the financial markets in Ethiopia. What comes out clearly from the literature Substitution is the existence of two sectors, which are propelled by different forces. They are not entirely exclusive, Aryeetey and Gockel (1990) observed a strong cor- yet manage to benefit from each other's inadequa- relation between the loss of confidence in the bank- cies. ing system in Ghana and the expansion of the infor- mal financial sector. They suggest that the informal New Structuralist Argument sector thrives on the deficiencies in the formal finan- cial sector, which would support the substitution or Some of the works of the new structuralists contest competition argument. It is often argued that in areas the financial liberalization propositions of where no banks can be found or where the banking McKinnon (1973) and Shaw (1973). The new struc- 131 The Informal Financial Sector and Markets in Africa an Empirical Study turalist attempts to introduce considerations of the ever, Owen and Solis-Fallas (op cit) disagree with the informal sector into the analysis have assumed a assumptions underlying the new structuralist mod- significant relationship between the two markets (see els, such as that of higher allocative efficiency. Owen Van Wijnbergen 1982, 1985; Taylor 1983). As Owen and Solis-Fallas only see the activities of individual and Solis-Fallas (1989, p. 342), who are very critical moneylenders and their clients as making up the of the new structuralist analysis, point out, "a key entire infornal financial market when they insist that feature of the new structuralist critique is the empha- "loans are of relatively short maturity... (and that) sis on infornal credit markets as an important source moneylenders operate under quasi-monopoly con- of residual financing." ditions with each lender active in a small-scale spa- The new structuralist argument is that increases in tially defined submarket in which there are limited interest rates will push upward interest costs in- opportunities for maturity transformation or econo- curred in financing working capital. This will lead to mies of scale in risk pooling, administration of loans, a cost-push effect on prices which may then lower etc." (Owen and Solis-Fallas, op cit, p. 347). They output under monopolistic conditions (Owen and overlook the existence of cooperatives and many Solis-Fallas, ibid). The growth expected by other actors on the informal scene who maybe more McKinnon and Shaw will notbe achieved in the short interested in term loans than earlier literature would run and a form of stagflation maybe experienced. suggest. Van Wijnbergen (1982, 1985) and others argue that the effects of increasing real bank deposit rates on Sore Hypotheses on the Relationship Between short run portfolio allocation will depend on the the Formal and Informal Financial Sectors: A extent to which households can substitute between Microeconornic Consideration bank deposits, loans on the informal financial market (curb markets), and unproductive assets (i.e. cash, This section offers a tentative explanation for the gold, and commodity stocks). Households (which higher prices in the informal market. To determine are assumed to be surplus units) allocate their real the demand and supply structures for the two sec- wealth, W, in such a manner that lending on the tors, and hence their relafionships, it must be deter- informal market, L, is characterized by the demand mined a priori whether the two sectors operate function: within a single financial market or separate markets. This is related to the issue of whether the two sectors (13-1) L =f( P, i, rtd, y) W offer a homogeneous product or a heterogeneous set of financial products, and whether their products are where P is the inflation rate, i is the nominal informal complements or substitutes. There are ambiguities in lending rate, rtd is the nominal time deposit rate, and the internal workings of the informal sector across y is real income. countries, and wide variations within countries for There is very little empirical evidence of informal the different operators, such as mobile bankers and lending or other portfolio allocation being depend- moneylenders. The tendency usually has been to ent on formal nominal deposit rates in Africa. The make the joint activities of moneylenders repre- link between the formal and informal financial sec- sentative of informal financial sector activities. tors is not via the deposit and lending rates; it is via It is useful to compare the financial products of- other factors, such as security in the case of susu fered by the formal and informal sectors on the asset collectors. Aryeetey et al (1990), in testing the rele- side. The formal financial sector is dominated by vance of McKinnon-Shaw postulates for Ghana, ob- commercial and developmentbanks in manyAfrican served that real deposit rates were insignificant in countries. These offer short and long term credit, explaining changes in domestic savings in Ghana. In although evidence from most countries suggests that an inflationary environment, potential savers will term lending is not very significant in view of inter- reallocate their savings to the forrnal sector only if nal management problems (see Aryeetey et al 1989 they can be sure of obtaining credit within a set and Ghatak, 1981). In the informal financial sector period. Rising nominal interest rates do not guaran- there is greater emphasis on short term credit for tee this. working capital and consumption. The informal and The new structuralists conclude that the informal formal sectors offer a similar product, which is prob- sector helps to sustain the moneymarket by reducing ably not entirely homogeneous. Based on this crite- the contractionary effects of increased formal nomi- rion of product differentiation on the asset side, one nal rates. This appears to support the empirical evi- may assume an oligopolisfic relationship with some dence from various African country studies. How- minor product differentiation. Within the informal 132 The Informal Financial Sector and Markets in Africa an Empirical Study Figure 13-1. Marginal Cost Curves for the Formal sector does not gain from economies of scale. Thus and Informal Financial Sectors the marginal cost curves of the two sectors would tend to converge as the activity expands (see figure 13-1). P A The closeness of the two marginal cost curves de- MCF pends partly on the estimation of the risk element MCI involved in informal sector lending. The marginal cost curve for the informal sector depends on the relative strengths of the monopolistic and perfectly competitive elements. It is, however, likely that a huge chunk of monopoly profit goes into individual estimations, especially in rural areas. On the demand side, the two sectors cater to the credit needs of different target groups. Banks satisfy the demand of large established firms engaged in specific economic activities. Credit from the informal Q sector goes to individuals, small-scale businesses, ,' w and the agricultural sector. The two types of demand are not exclusive, however, as demand for the prod- ucts of each sector has specific and non-specific ele- ments. The two sectors have different demand curves in any case2 (see figure 13-2). In most rural areas Figure 13-2. Specific, Nonspecific, and Potential where the demand for credit cannot be satisfied by Total Demand formal channels, the demand from the informal sec- tor is specific or exclusive and this facility is an obvious subsfitute for what the formal sector would p P P have offered. In urban areas much of the demand for (a) (b) (c) credit is non-specific: many people interact relatively freely in both sectors and the issues of complemen- \,---- ---- \tarity and substitution are not clear. For example, the owner of a small processing plant DS DN D mayobtain credit from a commercial bank to expand his plant size. He may then borrow from friends to pay for new raw materials. In this case the products of the two sectors complement each other. If the bank 0 Q 0 Q 0 Q had rejected his application for credit, and he had to invest in the expansion with credit from friends or a moneylender, such credit would be seen as a substi- Source: Hemmer and Mannel, (1989). tute for the unavailable formal credit. The issue then becomes which of the two sectors is predominant in urban and rural areas. The greater the element of substitution in the de- sector itself there are monopolistic tendencies, as in mand, the more specific that demand is. Similarly, the case of some moneylenders, and nearly perfectly the greater the element of complementarity the more competitive practices, as in the case of some susu significant is the element of non-specificity of de- operators. mand. It would seem that demand in the formal On the supply side, the two sectors face different market would have a greater proportion of non- cost structures. This is important in explaining the specificity (in view of the ease with which urban different prices for their products in an oligopolistic dwellers and other patrons of the sector can switch situation. It is usually suggested that the operational from one sector to the other) than demand in the costs of the individual actors in the informal sector informal market. Most of our studies show that it is are much less than those in the formal sector (US AID easier for normally formal credit users to move from 1989). If costs were aggregated, however, it is likely formal borrowing to informalborrowing than it is for that the difference is not large because the informal normally informal borrowers to switch around. In 133 The Infonnal Financial Sector and Markets in Africa an Empirical Study Figure 13-3. Total Demand Curves for the Formal Figure 134. Pricing in the Formal and Informal and Informal Financial Sectors Financial Markets P P Mc D D Q QF Q,~~~~ the determination of prices (interest rates) in both cific elements in both demands. This can only be markets, therefore, the structure of the demand explainedbytheexistenceofstronginstitutionalbar- curves will play some role. The linkbetween the two riers vis-a-vis the demand for collateral and other markets is established by the non-specific element in institutional barriers. Thus, barriers to entry into the both markets. The interest rate at any given time in formal market would always lead to segmentation either sector would, if determined freely, depend on and probably provide an additional motivation for the proportions of specific and non-specific elements the existence and growth of an informal financial in the demand they face since these proportions de- market, either in a complementary or competitive termine elasticities. role. The evidence gathered so far suggests that de- mand for informal credit may outweigh demand in Sumniary and Concluding Remarks the formal sector for several reasons. The former, in an extreme case, may involve total rural demand and A large and significant informal financial sector a large part of urban demand. The existence of credit exists alongside the formal financial sector in many rationing in the formal sector is important in the African countries. Evidence from Ghana, Malawi, analysis. These observations lead us to suggest de- Senegal, Tanzania, and Zaire has been examined in mand curves for the sectors as in figure 13-3. The link this paper. It supports the argument that the infor- between the two markets by way of the non-specific mal sector is too large in these countries to be ignor- element in both demands is not portrayed here in ed in poiicy-making. This is even more so since ac- view of the fact that this link has not yet been studied tions taken in one sector have repercussions in the closely. Making allowances for this anomaly, super- other. imposing figure 13-1 on figure 13-3 yields figure The informal financial sector has similar charac- 13-4, which indicates the prices (interest rates) at teristics in the five countries studied. This was by which credit would be obtained from both markets way of the observed types of operators and activists, at any given time. The higher interest rates in the relative size, relative operating costs, and relative informal sector may be explained more by supply pricing. Differences arise on the question of which side conditions following the relatively rapid rise of informal units dominate the market. Thus, for exam- average and marginal costs in that sector, illustrated ple, while the susu appears dominant in Ghana, co- by MCi approaching MCf. This is also a trend we operative societies dominate informal financial ac- intend to investigate more closely. tivity in Tanzania. In this static presentation, Pi cannot approach Pf Only recently has the importance of the informal despite our earlier assertion that there are non-spe- sector to growth and development been highlighted 134 The Informal Financial Sector and Markets in Africa: an Empirical Shfdy in the literature, as per the new structuralists. How- Chipeta, C. and M.L.C. Mkandawire (1989); The Informal ever, some of the assumptions underlying the rela- Financial Sector in Malawi Scope, Size andRole, an interim tionship between the two sectors-such as the as- research report presented at an AERC workshop, sumption that they may relate via interest rates-ap- Harare, 4-8 December. pear overstretched. That assumption is not sup- Due, J.N. (1980); Costs, Returns and Repayment Experience in ported by the available empirical evidence. Ujamaa Villages in Tanzania, 1973-1976, University Press It has also been shown here that the informal finan- of America, Washington. cial sector acts both as a complement and a substitute Ghatak, S. (1981); Monetary Economics in Developing Coun- to the formal financial sector. It is this substitution ef- tries, the MacMillan Press Ltd., London and Basingstoke. fect that makes a relatively greater proportion of de- Hemmer, H-R and C. Mannel (1989); "On the Economic mand in this sector specific. A higher level of non- Analysis of the Urban Infornal Sector,' World Develop- specificity in demand would work to push down menat,iVol. 17, No. 10, pp. 154rS1552. prices in the sector in the long run in the absence of mynt, V., NO. 10, pp. 1 . institutional barriers. In view of this, attempts at forma, M, M.O. Ndanshau, and J.P. Kipokola (1989); In- forging~~~~~~~~~~~~~ a clsrln ewe h w akt ol formal Financial Markets in Tanz4nia. Scope, Structure and havetoinclude, amongotherthings, atconcentrati n Policy Implications: Evidence from a Pilot Survey, paper onve redingute, level ofher non-gspecit onc ftrmai presented at an AERC workshop, Harare, 4-8 December. On reducing the level of non-specificity in formal sector demand as a means of aligning demand in Hyuha, M. and M.O. Ndanshau (1990); Economic Adjust- both sectors and reducing the wide gap in interest ment Packages in Africa: The Informal Financial Sector as the rates. This may come about through expansion and Forgotten Half in Financial Reform in Tanzania, mimeo., improvements in the facilities which fornal financial Department of Economics, University of Dar es Salaam. institutions have to offer. IPC (1988); Rural Finance in Ghana, a research study pre- pared for Bank of Ghana on behalf of the Ministry of Notes Economic Cooperation of the Federal Republic of Ger- many, Bonn. The authors would like to express their appreciation for Lundahl, M. et al (1987); "Agricultural Credit in Tanzania: the financial and technical assistance provided by the Afr- A Peasant Perspective," Savings and Development, Vol. 11, ican Economic Research Consortium (AERC), particularly No. 4. Jeff Fine and Benno Ndulu, towards the preparatory work Mauri, A. (1987); The Role of Financial Intermediation in the for this paper. Mobilization and Allocation of Household Savings in Devel- The use of material made available by our colleagues C. oping Countries: Interlinks between Organized and Informal C'hipeta and M.L.C. Mkandawire (Malawi), M. 0. Ndan- Circuits, The Case of Ethiopia, paper presented at the In- shau and J. P. Kipokola (Tanzania) and F. Gockel (Ghana) temational Experts' meeting on "Domestic Savings Mo- is also acknowledged. This becamepossible by virtue of the bilization through Formal and Informal Circuits: Com- high level of cooperation within the AERC network on parative Experiences in Asian and African Developing Informal Financial Markets in Africa. Countries," Honolulu, 2-4 June. 1. See also Lundahl et al (1987), Msambichaka, Ndulu, Miracle, M.P., D.S. Miracle, and L. Cohen (1980); "Informal and Amani (1986), and Due (1980). Savings Mobilization in Africa," Economic Development 2. We borrow here the illustration of specific and non- and Cultural Change, Vol. 28. specific demand in the formal and informal sectors from Msambichaka, L.A., Bj. Ndulu and H.K.R. Amani (1986); Hemmer and Mannel. Agricultural Development in Tanzaniax Policy Ewlution, References Performance and Evaluation, Friedrich Ebert Foundation, Bonn. Aryeetey, E. and F. Gockel (1990); Mobilizing Domestic Re- McKinnon, R.I. (1973); Money and Capital in Economic Devel- sourcesfor Capital Fornation: The Role of Informal Financial opment, The Brookings Institute, Washington. Markets in Ghana, paper presented at a workshop of the Owen, PD. and 0. Solis-Fallas (1989); "Unorganized African Economic Research Consortium (AERC), Nai- Money Markets and 'Unproductive' Assets in the New robi, 26-30 May. Structuralist Critique of Finandal Liberalization" Jour- A ryeetey, E., Y. Asante, F. Gockel, and A.Y. Kyei (1990); nal of Development Economics, 31 p. 341-355. Mobilizing Domestic Savings for African Development and Rahman, F.H. (1989); "Rural Savings: A Neglected Dimen- Diversification: A Ghanaian Case, paper presented at a sion of Rural Development," The Courier, Journal of the workshop of the International Development Centre, ACP-EEC, No. 115, May-June, pp. 70. University of Oxford, Oxford, 16-20 July. Seibel, H-D (1988); Financial Innovationsforhiicroenterprises: 135 The Informal Financial Sector and Markets in Africa an Empirical Sthdy Linking Formal and Informal Financial Institutions in Africa South Korea, Journal of Development Economics, 10, April, and Asia, paper presented at the World Conference on pp. 133-169. "Support for Microenterprises," Washington, 6-9 June. van Wijnbergen, S. (1985); "Macroeconomic Effects of Shaw, E.S. (1973); Financial Deepening in Economic Develop- Changes in Bank Interest Rates: Simulation Results for ment, Oxford University Press, New York South Korea, Journal of Development Economics, 18, Au- Taylor, L (1983); Structuralist Macroeconomics: Applicable gust, pp. 541-554. Modelsfor the Third World, Basic Books, New York Wai, U Tun (1957); 'Interest Rates Outside the Organized van Wijnbergen, S. (1982); "Stagflationary Effects of Mone- Money Markets of Underdeveloped Countries,' IMF tary Stabilization Policies: A Quantitative Analysis of Staff Papers, Vol. 6, No. 1. 136 14 Mobilizing Domestic Resources for African Development and Diversification: Structural Impediments to Financial Intermediation Machiko Nissanke As the deep economic crisis which crippled Sub-Sa- of the structural characteristics and impediments in haran Africa throughout the 1980s wears on, official Sub-Saharan economies. It is imperative to examine donors and African recipients are increasingly con- in detail the prevailing structure of the financial sys- cerned about the prospect of foreign aid becoming tem and the economies at large, in order to design structurally embedded in the economies of the re- carefully the transition process from repression to gion. After several years of policy reforms tied to liberalization. It is also important to assess the effects structural adjustment loans, it is apparent that the of prevailing macroeconomic and financial policy foundations necessary for self-sustained growth regimes on domestic savings mobilization, and to have yet to be laid down. Many countries have sadly grasp the financial dynamics of structural adjust- experienced institutional disintegration, infrastruc- ment programs. tural erosion, and the accelerated deterioration of This paper tries to identify key motivational, insti- systematic capacity during the crisis. tutional, and policy bottlenecks inhibiting the mobi- The financial policies embodied in structural ad- lization of domtestic savings. It explores means to justment programs derive most of their intellectual develop the financialization of domestic savings and content from the financial repression hypothesis.1 It the intermediation between savings and investment is argued that a strategy of financial liberation will for economic diversification and development. The deepen the financial structure and renew economic discussion presented here is preliminary and is based development. However, little is known about how largely on studies of Ghana (Aryeety et al 1989) and the dynamic mechanism of liberalization impacts on Malawi (Chipeta and Mkandawire, 1989). The sec- the financial structures of fragile low-income coun- ond section summarizes the patterns of financial and tries. Institution building is needed to achieve suffi- savings behavior of the household sector, corporate cient market depth to enable market forces to operate and small-scale enterprises, and the public sector. beneficially. The sequence and pace of liberalization The third section evaluates the structure and func- mustbe very carefullyworked out and must take into tioning of formal financial institutions and markets. account the speed of market deepening and chang- Several policy implications are discussed in the con- ing macroeconomic conditions. cluding section. Moreover, in low-income economies institutional and motivational bottlenecks might be more binding The Pattern of Saving and Financing than the policy bottlenecks which are singled out by the financial repression hypothesis. Viable financial The analysis presented in this section is based policies must be based on improved understanding partly on recently conducted field work. Unfortu- 137 Mobilizing Domestic Resources forAfrican Development and Divcrsification.: Structural Impedimtnts to Finandal Intermediation nately, lack of data hinders the analysis of the pattern tional bottlenecks and weakness of the formal sector, of enterprise financing. Some interesting findings and the lack of access to its credit facilities in the regarding the household sector are emerging from absence of required collateral have created a vibrant current research. The savings and financing patterns informal financial sector. In the informal sector the of the public sector are discussed here mainly in cost of credit can be much higher, but even low-in- relation to those of the private sector. come groups can have access to liquidity.6 The striking degree of malfunctioning of the for- The Household Sector mal financial institutions and the system as a whole7 acts as a structural hindrance to more effective sav- Detailed and reliable estimates of savings rates by ings mobilization through the formal financial sys- region and type of household are hard to obtain. tem. Commercial banks do not have the motivation However, the available sample survey and previous to overcome their operational bottlenecks and make studies2 indicate, for example, that total savings in efforts at savings mobilization. This is due largely to rural households in Ghana could constitute as much the problem of excess liquidity. This has led to the as half of estimated annual income. It is estimated striking situation that "in a rural area of Ghana (ra- that only 20 percent of these savings is accounted for male), one association of 'susu' collectors was re- by financial savings. While various forms of asset ported to have mobilized C30 million in a month, holding as stores of value indicate a large scale of amounting to C360 million ($2.45 million) in a year" efficiency loss to the economy,3 this is undoubtedly (Bentil et al).8 This compares with the average total also a reflection of the insufficient degree of moneti- savings deposits mobilized by commercial banks in zation of the economy. Hence, analysis of rural sav- that areaof Cll million ($75,000) bymid-year in 1988. ings should take into account the fact that saving Funds mobilized through one informal sector behavior in a nonmonetized peasant economy canbe group are seldom used to finance other groups in the distinctly different from that in a monetized econ- informal sector. Thus, the potential use of savings omy (see, for example, Alamgir 1976 and Ghosh mobilized through the informal financial sector for 1986). The rate of monetization of economic activities economy-wide diversification remains unrealized. A and the degree of financialization of savings are con- process of industrialization and development based siderably higher in urban areas. on informal small production units has not yet taken Most of the financial savings (between 80 and 95 off the ground.9 Few informal financial institutions percent) of the household sector have been mobilized provide long term finance in either country studied. through informal financial institutions and markets. Evaluation of the current and potential links be- In Ghana, informal financial savings (excluding per- tween the formal and informal sectors can be perti- sonal loans) are estimated to constitute about two nent for future policy formation on savings mobili- percent of GDP. In Malawi the total credit extended zation. The two country studies emphasize the com- by the informal financial sector (including estate plementarity of relationships between the two sec- owners and personal loans among friends and fami- tors. However the facilities directly available to the lies) has been estimated at 6.5 percent of GDP;4 in- population through the informal sector are to some formal sector lending to the private sector was at least extent presently substituting for the operations of three times as great as that of the formal sector.5 formal institutions. A sizeable proportion of the de- Many of the traditional forms of saving and infor- mand deposits of banks is mobilized through infor- mal financial activities have long existed in African mal sector operations. This is not reflected in the countries. The informal financial sector includes growth of bank liabilities because informal sector credit unions, which could be described as semi-for- operators keep deposits in current demand and sav- mal institutions; the system of rotating savings; sav- ings accounts, i.e. in highly liquid forms. As the ings and credit associations; cooperative savings as- income level remains the dominant variable in deter- sociations; the single collector system; and money- mining the personal savings rate, the terms of trade lenders. The nominal interest rate charged on credit between the agricultural and other sectors has a cru- varies enormously across these heterogeneous op- cial role to play in rural savings mobilization. In erations. In Ghana and Malawi informal credit is addition, our study indicates the need for recycling often granted with very flexible repayment terms in rural savings in the local vicinities on a visible scale. light of debtors' personal circumstances. Savings mobilization and credit granting facilities Numerous factors inhibit the majority of the popu- must be made accessible and utilized actively within lation from using formal institutions. The institu- local communities. 138 Mobilizing Domestic Resources forAfrican Development and Diversificatiow Steucturat Impediments to Financal Intermediation The Pattern of Enterprise Financing this sector which is vital for Malawi's future diversi- fication.10 lJnfortunately, few comprehensive data are avail- able on the financing pattern of small-scale enter- Savings and Public Sector Finance prises in Ghana and Malawi. It is, however, clear that most private entrepreneurs have financed their ac- This section presents data on public and private tivities with loans from relatives and friends, infor- savings and public sector finance in Ghana and mal sector loans, and their own savings. This has Malawi. limited the scale of activities and investment. Aryeetey et al analyzed the available balance Public and Private Savings. The savings rates for sheets submitted by a limited number of established Ghana and Malawi during 1969 to 1988 are prestend corporations in Ghana. The data show that, on aver- in table 14-1. Table 14-1 presents an approximate age, equity financing is more dominant for public breakdown of gross national savings into public sec- corporations than for private corporations, and the tor savings and private sector savings. Here public ratio of external finance to internal funding and the savings refers to government savings defined as the loan/equity ratio are higher for private corporations. net excess of total revenue and grants over current The most common source of finance for the private expenditure, and private savings is calculated a re- sector is short term credit such as overdraft facilities, sidual after public savings have been deducted from trade credit, and supplier's credit. gross national savings. [n Malawi there has been virtually no commercial In Ghana, the gross domestic savings and gross bank lending to small- and medium-scale business- national savings rates have been at a very depressed es. Within the private sector, thebanks prefer to deal level (one of the lowest levels in the region). During with the estate sector. At any rate, the commercial 1972 to 1983, the Ghanaian government ran a persist- banks do not provide long term or development ent current budget deficit (see table 14-2), which loans. In order to address these problems and to further depressed the aggregate savings rates. The faclitate the development of small and medium en- current account balance stayed in deficit throughout terprises, a number of specialized institutions have the period, despite continuous downward adjust- been established in Malawi during the past 10 yea- ments in current expenditure (see table 14-2). The rs. These specialized institutions suffer acutely from root cause of the chronic deficits should be found in inadequate funds to carry out their mandate, and the rapidly deteriorating tax base, which declined at very limited administrative capacity to manage an a rate much faster than that of expenditure. As a extensive credit network. The agro-industry, which percentage of GDP, revenue dropped steadily from uses the agricultural base as a source of raw materi- over 16 percent in 1975 to 6 percent in 1983. als, is regarded as an especially promising area for In 1983, Ghana began reducing the deficit in its diversification. At this point, the supply of credit current budget and in 1985 began to generate a sur- would be a binding constraint on the expansion of plus with increased external grants. Current revenue, Table 14-1. Public and private savings (percentages of GNP) 1972-73 1974-78 1979-83 1984-88 Gross national savings rate Ghana 12.1 8.1 3.4 5.2 Malawi 8.7 12.1 2.4 6.1c Public Savings ratea Ghana -1.2 -2.7 -3.3 1.8 Malawi 0.5 3.1 3.4 1.9c Private savings rateb Ghana 13.3 10.8 6.7 3.4 Malawi 8.2 9.0 -1.0 4.2c iVotes: a. This is the current account balance. b. This is the gross national savings rate less the public savings rate. c. 'For 1984-86 only. Source: IMF International Financial Statistics Yearbook and Government Finance Statistcs Yearbook. 139 Mobilizing Domestic Resources for African Development and Divers Qicatio- Structural Impediments to Finandial Intermediation includinggrants, increased sharplyfrom8percentof the domestic financial system in Ghana and GDP in 1984 to 20 percent of GDP in 1987, whereas Malawi.11 The size of the deficit is much larger in expenditure was increased from 9 to 21 percent of bothcountrieswhencapitalexpenditureistakeninto GDP in the same period. The increase in public sav- account. Although there are interesting divergences ings since 1984 implies that the private sector savings between the two countries, the question of crowding rate, calculated as a residual, has continually fol- out is likely to take a central place in the discussion lowed a dwindling trend. of public sectorfinance, given thepredominant share In Malawi, the trend in gross savings rates has been of the public sector in domestic credit allocation. characterized by a distinct break between the 1970s The government of Ghana financed its deficit al- and the 1980s, and by large yearly fluctuations, as most exclusivelybyborrowing from the centralbank shown in table 14-1. Savings rates increased during and the banking institutions up to 1983, in the near the 1970s and deteriorated significantly after 1979. absence of external finance. Capital expenditure in- The government ran a surplus on its current budget creased after 1984, when the government accepted a account on average throughout the 1970s and 1980s. structural adjustment loan. The mode of financing In contrast to Ghana, this was achieved while keep- has changed dramatically since 1984 with a large ing the level of current expenditure between 15 and inflow of external funds. The share of the public 20 percent of GDP throughout the period. The differ- sector in total domestic credit increased enormously ence in the fiscal situations in the two countries is in the 1970s; since 1984 it has been decreasing slowly. explained by the revenue side: in Malawi, there was In Malawi, the governmenthas kept its currentbu- no deterioration in the domestically generated reve- dget balance in surplus. However, the overall budget nue base, and grants of considerable size consoli- deficit has been as large as 10 percent of GDLP, which dated the situation. indicates the consistently high level of capital expen- diture. Hence, the government has continually been Public Finance. Tables 14-2 and 14-3 show sum- faced with borrowing requirements of considerable mary data on public finance and its relationship to size. In contrast to the Ghanaian case, the governm- Table 14-2. Publicfinance 1972-73 1974-78 1979-83 1984-88 Percentage of GDP: Current account blance Ghana -1.6a -3.2 -3.7 3.0c Malawi 0.2b 3.0 3.2 1.7c Overall budget surplus/ defecit Ghana -6.Oa -9.8 -5.8 -1.1c Malawi -6.0b -6.9 -10.3 -7.6c Government borrowing from the banking system Ghana 4.2a 9.6 5.3 0.5d Malawi 1 b 1.6 4.8 3.2c Domestic credit Ghana 27.7 31.9 21.8 22.9 Malawi 12.1 23.2 37.7 31.8 Percentage. Growth in credit to public sector** Ghana 10.2 56.9 41.0 43.9 Malawi 37.5 64.7 29.2 8.2 Notes: a. This is the current account balance. b. This is the gross national savings rate less the public savings rate. c. For 1984-86 only. *Credit extended by the Central Bank and deposit money banks, i.e. those included in the monetary survey. **ln Ghana it includes the central government and non-financial public enterprises only. In Malawi it includes the central government and all official entities. Source IMF International Financial Statistics Yearbook and Govemment Finance Statistics Yearbook. 140 Mobilizing Domestic Resources for African Development and Diversification: Structural Impediments to Financial Intermediation ent resorted to heavy external financing throughout undertaken by the World Bank (Neal 1988), Ghana the period. While this strategy invited its own burde- ranks second only to Zaire as having the lowest ratio nsome problem of external debt management, it imp- of M2 to GDP in the world. lies that government has a less dominant position as In Malawi, the ratio of currency as a percentage of borrower in the domestic financial system in Malawi. M2 was 36 percent in 1965, declined to 20 percent in 1980, and remained at just under 20 percent through- The Formnal Financial System out the 1980s. The currency-to-M2 ratio in Malawi is about one-third of that in Ghana. This reflects the Aggregate financial indicators, the structure of the difference in the stages of financial development in formal financial system, the operational charac- the two countries. There has been a steady increase teristics of financial institutions, and the constraints in financial deepening in the Malawian economy and scope of financial markets are discussed in this during the 1967 to 1988 period. section. The Structure of the Formal Financial System Aggregate Financial Indicators Ghana's formal financial system is comprised of The time series data for MI, M2, and M3 as percen- the Central Bank (the Bank of Ghana), four large tages of GDP in Ghana and Malawi during 1969 to commercial banks, three development banks (DFIs), 1987 are used to show the extent of financial deepen- two merchantbanks, two smallerbanks, and 20 small ing.InGhanaall the ratios reached apeakin1976 and rural banks. The total assets of the banking system thereafter declined sharply. Currency accounted for (excluding the central bank) were 231 billion cedis at 54 percent of M2 in 1960, between 35 and 38 percent the end of 1988. This was about a quarter of GDP. Of of M2 during the mid 1960s to the mid 1970s, and inc- total assets, some 71 percent was accounted forby the reased gradually after 1976 to 54 percent during 1983 four large commercial banks, 20 percent by DFIs, and to 1986. These statistics reveal the process whereby 6 percent by merchant banks. movement toward financial deepening was reversed Bank density in Ghana increased from 1.9branches during 1976 to 1983. According to a recent study per 100,000 inhabitants in 1976 to 3.4 in 1988. This Table 14-3. Percentage of public sector in domestic credit 1972-73 1974-78 1979-83 1984-88 By the central banks: Total Ghana 95.9 95.1 99.0 89.8 MNalawi 77.1 92.3 94.7 100.0 Claims on government Ghana 71.3 85.2 81.0 71.3 Malawi 77.1 47.1 72.7 83.8 Claims on public enterprises Ghana 24.6 9.9 18.0 18.5 Malawi - 56.5 22.0 16.2 By the banking institution*: Total Ghana 65.2 80.3 90.0 83.2 Malawi 20.2 45.8 51.3 68.4 Claims on government Ghana 46.0 62.3 69.8 71.0 Malawi 17.7 24.4 36.4 53.6 Claims on public enterprises Ghana 19.2 18.0 20.2 12.1 Malawi 2.5 21.4 14.9 14.8 Notes: a. 1975-78. 'Includes central banks and deposit money banks. Source IMF International Financial Statistics Yearbook and Government Finance Statistics Yearbook. 141 Mobilizing Domestic Resources for African Devlopment and Diversifcation Structural Impediments to Finandcal Intermediation was due to 25 percent growth in the number of matching the maturity structure of assets with the primary commercial bank branches between 1976 existingmaturities of liabilities. Thebanks argue that and 1988 and the establishment of 120 rural banks medium and long term lending on a greater scale since 1976. There are currently 466 bank branches would be improper because they are forced to ma- spread unevenly over the country. intain a sizeable proportion of their assets in highly Malawi's financial system comprises the Central liquid form. Instead they lend to a small number of Bank (the Reserve Bank of Malawi, RBM); two com- well-established and financially strong private and mercial banks; three development finance institu- parastatal companies, which have been the main tions; two finance houses; the New Building Society beneficiaries of institutional credits offered at sub- (NBS); the Post Office Savings Bank (POSB); a num- sidized interest rates. The credit ceilings imposed for ber of insurance companies; and the Malawi Union the purpose of macroeconomic stabilization by the of Savings and Credit Cooperatives (MUSCCO), International Monetary Fund programs are sing- which is the umbrella organization for 58 credit un- ularly ineffective in an environment in which port- ions. The financial system's total assets at the end of folio management is conservative and undynamic. 1987, excluding RBM, amounted to 39 percent of In Ghana there are four large commercial banks: GDP. Of the total assets, 68 percent was in the com- the Ghana Commercial Bank (GCB), Barclay Bank of mercial banks, 12 percent in the insurance compa- Ghana (BBG), Standard Chartered Bank of Ghana nies, 10 percent in the POSB, 4 percent in the New (SBG), and the Social Security Bank (SSB). At the end Building Society, 3 percent in the DFIs, and 2 percent of 1988, these banks accounted for about 77 percent in the financial houses. of total deposit liabilities of financial institutions, 67 The number of branches and state agencies of com- percent of the loans market, and 76 percent of private merial banks and the POSB increased from 185 in bills and securities. They held about 15 percent of 1967 to 331 in 1988. However, because of Malawi's their total deposit liabilities as excess reserves in high population growth rate, the ratio of people to 1988; the share of liquid assets was 60 percent of total banks deteriorated from 22,300 in 1967 to 25,000 in deposit liabilities. 1988. (The comparable number for Ghana in 1988 Excess liquidity and the high frequency savings was 29,200.) The bank density is much higher in cycles of the private sector have resulted in a very sh- urban centers than in rural areas. ort term liabilitystructure. About 70percentof liabili- ties and between 85 and 90 percent of deposit liabili- Operational Characteristics of Financial Institutions ties originate from demand and savings deposits, the latter of which are operated almost like current acc- This section examines the functions and opera- ounts. The balance sheets also reveal that the comm- tional characteristics of the formal financial system's ercial banks in Ghana are typically undercapitalized. main subsectors-commercial banks, development Commercial bank loans in Ghana are predom- finance institutions, rural banks in Ghana, and the inately of short term maturity, and usually are made Post Office Savings Bank in Malawi.12 by extending overdraft facilities. It was estimated that in 1987 only 15 percent of commercial bank lo- Commercial Banks. Analysis of the balance sheets of ans had a maturity of over three years. Only through commercial banks in Ghana and Malawi reveals two continuous rollovers do the banks, in effect, provide strikingly similar features of their operational char- some long term finance to their more credit worthy acteristics: excess liquidity held above the required customers. Commerce and the construction sector level on a voluntarybasis, and a heavyconcentration are favored by the commercial banks, although they of loan portfolios on the short end of the market. are required to hold at least 20 percent of their total These banks hold assets far above the required statu- loan portfolio as credit to the agricultural sector. torylevels in cashwith no return, and in treasurybills In Malawi commercial banking activity is under- and government stocks with very low returns. The taken by two banks, the National Bank of Malawi absence of viable projects to which to lend is cited as (NBM) and the Commercial Bank of Malawi (CBM). a prime reason for the excess of loanable funds. Faced The former has 76 percent of commercial bank depos- withexcessliquidityandlimitedlendingopportunit- its and 60 percent of assets. The two banks have ies, the banks lack incentive to mobilize additional accounted for around 80 percent of the total deposit savings. liabilities of financial institutions throughout the Given the short term liability structure and the post-independence years. weak capital base of both countries, the portfolio Excess liquid assets held by Malawi's commercial management of commercial banks is geared to banks amounted to 20 percent of total deposits in 142 Mobilizing Domestic Resources for African Development and Diversificatiow Structural Impediments to Finandal Intennediation 1975. In 1988 excess liquid assets were 38 percent Three development finance institutions were es- above the minimum reserve ratio and accounted for tablished by the central bank in Ghana: the National 68 percent of total deposits.13 A salient shift in the Investment Bank, in 1963; the Agricultural Develop- composition of deposits appears to have taken place ment Bank, in 1965; and the Bank for Housing and in favor of time deposits since the early 1970s. At the Construction, in 1973. In 1988 the three banks had 26 National Bank of Malawi, the share of time deposits percent of the market share of loans and advances in was over 50 percent in the mid 1980s. However, more the banking system. than 80 percent of the time deposits have a maturity The development finance institutions were depen- of less than one year (Chipeta and Mkandawire). dent on term loans from external sources involving Due to increasing excess liquidity, the commercial exchange and interest rate risks. They assumed large banks have actively discouraged people from depos- financial risks by specializing in long tenn lending in iting money in interestbearing fime deposit accounts high risk areas such as agricultural and industrial fin- with maturities of more than one year. The cost of ance. They have experienced acute illiquidity and keeping the deposits in long maturities has been insolvency problems and have suffered from a sub- considered too high because of the absence of good stantial number of nonperforming portfolios with investment opportunities and the wide margin be- high default rates and bad debts. They are often not tween deposit and lending rates. This suggests that able to meet the Central Bank's reserve requirements interest rate liberalization alone is ineffectual for ef- and have had to be frequently refinanced and recapi- ficient savings mobilization when the financial inter- talized by the Central Bank or the government. mediaries are awash with excess liquidity. Due to the insufficiency of funds available from The commercial banksinMalawiare characterized external and government sources, Ghana's develop- by their extreme risk aversion. Banks are known to ment finance institutions have mobilized savings. At extend credit only to well-established large enter- the end of 1988, they had 13 percent of the market prises. The large proportion of commercial bank share of total deposits in the banking system. Faced lendin,g is for capital over short periods and is in the with chronic insolvency problems, they emphasized form of fluctuating overdrafts. short term self-liquidating loans and advances. Con- In the sectoral credit distribution of the commerc- sequently, even the development finance institutions ial banks in Malawi, agriculture has accounted for have begun to shift away from the manufacturing the largest proportion (over 50 percent throughout and nontraditional agricultural sectors. the 1980s). Most agricultural credit has been ab- There are three development finance institutions sorbed in financing the tobacco sector. The whole- in Malawi: the Industrial and Development Bank sale and retail trade has the next largest share in (INDEBANK), the Industrial and Development credit distribution. Fund (INDEFUND), and the Small Enterprise Devel- While the private sector in Malawi remains the opment Organization of Malawi (SEDOM). These dominant borrower (with between 60 and 70 percent development finance institutions rely totally on for- of bank claims), the central government's share of eign resources and do not mobilize domestic sav- claims has increased significantly in recent years and ings.14 INDEFUND and SEDOM have difficulties in accounts for between 25 and 30 percent of bank carrying out their functions due to inadequate re- credit.. Smallholders, smallholder tenants on estates, sources and staff shortages. In-house technical and nontraditional exporters, and small and medium en- administrative capacity is reported to be grossly in- terprises have received very little direct credit from adequate. It is reported that INDEFUND has consis- commercial banks in Malawi. tently made losses in its operations. SEDOM is more effici- ent in its operations, but also faces a risinglevel Development Finance Institutions. The inadequa- of arrears in its overall portfolio. In view of the vital cies cf conventional commercial banking practice importance of their functions for the country's diver- and the dearth of medium and long term credit to the sification prospects, it is critical to reduce the finan- priority sectors of the economy necessitated the es- cial constraints on the development finance institu- tablishment of development finance institutions in tions and to improve their resource base and opera- both countries. They were typically founded on te- tional efficiency. rm loans from external sources and have been co- ntinuously financed either directly by external Rural Banks in Ghana. The rural banks in Ghana are sources or indirectly through the government, often unit banks that have been established since 1976 to through refinancing schemes evolved by the central mobilize resources and extend credit locally. They are bank. owned and managed by their local communities and 143 Mobilizing Domestic Resources for African Development and Diversification: Sfructral Impediments to Financial Intermediation assist rural based industrialization through loans to Financial Markets: Constraints and Scope cottage industries. The central bank holds preference shares equivalent to one-third of the initial share Ghana and Malawi have not been able to generate capital. However, the major source of funds is de- long term credit and loan provisions for self-sustain- posit liabilities, which made up 70 percent of total able diversification. In spite of high liquidity in the resources in 1988. The rural banks serve small bor- overall banking system, the potentially productive rowers and depositors and have been quite success- sectors in the economy are not able to receive ade- ful at savings mobilization. quate institutional credit. In view of the acute need Between 1977 and 1988, the share of agriculture in for export diversification, the lack of availability of total credit from rural banks was halved. Credit to export financing for nontraditional exports should agriculture amounted to 54 percent of total credit in be regarded as a critical bottleneck. 1985, but had dropped to 34 percent by 1988. The Segmentation of the formal financial system in share of trading activities increased from 13 percent these countries has become structurally embedded in 1985 to 26 percent in 1988. into their economies. There is very little liquidity Moreover, the rural banks have accumulated ex- management involving interbank credit arrange- cess liquidity. They had a primary reserve ratio of 25 ments, whereby resources could be shifted from the percent and a secondary reserve ratio of 26 percent surplus sector to the deficit sector among financial as of mid-June 1988, which were far in excess of the intermediaries. This section presents the current state mandatory requirements of 10 and 20 percent in the of, and scope for, developing money and capital respective reserve categories. It is reported that man- markets in Ghana and Malawi. agers of the rural banks have resorted to low risk investment in government. Meanwhile an increasing Money Markets. The mismatch of liquidity posi- number of households are obtaining credit from the tions within the banking system calls for the creation informal financial sector. of an active interbank money market and the estab- lishment of a discount house that trades in a variety Post Ofice Savings Bank in Malawi. Nearly 90 per- of short maturity financial products. cent of Malawi's population lives in rural areas. The In Ghana the Consolidated Discount House Lim- network of the Post Office Savings Bank (POSB) has ited (CDHL) was formally institutionalized in 1988 the most extensive coverage for savings mobilizat- as an interbank intermediary for short term assets. It ion among all formal institutions in the rural areas. is jointly owned by eight banks and six insurance POSB has 286 deposit collecting points and has mo- companies. However, large portions of the banking bilized over 12 percent of total savings by formal system are still characterized by idle funds, while the financial institutions. The interest earned by its de- other financial institutions and the industrial and positors is tax-exempt, which gives POSB a comp- commercial sectors face severe liquidity problems. etitive edge in mobilizing resources. However, be- The CDHL's principal portfolios are in Bank of cause it primarily serves low-income groups, it has a Ghana bills, treasury bills, cocoa bills, grain bills, low minimum deposit balance requirement and a bankers acceptances, negotiable certificates of de- ceiling is imposed on deposits for individuals and posit, and government stocks. Commercial papers corporations. Its deposit liabilities are noticeably af- are not yet featured. The CDHL has functioned as an fected by changes in rural cash income and by remit- important source of short term borrowing for the tances of deferred payments by Malawian migrant government. workers. In Malawi, no formal money market exists. The The POSB lends exclusively to the government, as interfinancial institution claims are at present mainly its statutes require, through the purchase of govern- in the form of deposits of NFBIs at commercial banks ment securities. Chipeta and Mkandawire argue that and at the Central Bank. The existing range of non- the lack of lending facilities discourages depositors, government short term paper is very limited. Three- despite the attraction of nearness to their homes, and month treasury bills remain the major short term drives small depositors toward the informal financial debt paper in the economy. By far the largest propor- market. Drawing upon other countries' experiences, tion of treasury bills has usually been held by com- they suggest granting management and financial mercial banks. Hence, the development of the money autonomy to make the POSB a more important in- market for short term assets in Malawi appears to be strument for resource mobilization and credit alloca- presently hampered by the very limited number of tion to small-scale rural traders and manufacturing financial products on offer and prospective transac- enterprises. tors on the demand and supply sides. 144 Mobilizing Domestic Resources for African Development and Diversification: Structural Impediments to Financial Intermediation hasbeen pointed tobyDeaton (1989). Such saving is Capital Markets. The acute shortage of term loans high frequency and aimed at intertemporal con- with long maturity in both countries requires the sumption smoothing, in contrast to intergenerational urgent development of an active and broad based or life-cycle hump saving. It provides a buffer be- capital market. A capital market could transform and tween uncertain and unpredictable income and an lengthen credit maturities and considerably reduce already low level of consumption. In addition, the the credit risks inherent in transactions among indi- combination of high risks and high transaction costs viduals. It would enable corporate issuers to access acts as a severe constraint on viable large-scale finan- long term funds provided by investors and to offer cial intermediation by the informal financial sector. liquidity in the secondary market for shares and The infonnal sector currently plays an important role bonds. only in savings mobilization and not in efficient ac- In Ghana there is no capital market despite many tual intermediation. attempts to establish one and the presence of some In view of the extent of the segmentation of the 18 public companies. The National Trust Holding formal system, it is necessary to reorganize some of Company was established in 1976 as the nucleus of the financial institutions and to encourage the crea- a capital market to provide brokerage services to the tion of effective moneyand capital markets for chan- investing public and shareholders. Very few shares nelling funds into productive activities. However, have been issued, however, despite the abundant the observed mismatch of liquidity positions within potential supplyof funds from the commercialbanks the formal financial system in these countries is a and the Social Security and National Insurance Trust. structural phenomenon. While short term money In Malawi, in the absence of an active stock ex- abounds in the form of excess liquidity in one seg- change, the capital market is extremely shallow. It ment of the banking system, the economy as a whole only deals in securities involving central govern- is characterized by an endemic shortage of long term ment, locally registered stocks. The major institu- loan provisions for productive investment and diver- tions that have so far participated in these dealings sification. The financial system currently lacks the are the financial intermediaries and the three holding ability to take on the risks associated with the matur- organizations, namely MDC, ADMARC, and the ity term transformation which is indispensable for Press Corporation. the active intermediation of financial flows between surplus and deficit units of the economy. Thoughts on Policy Implications The fundamental problem of high risk in these fragile African economies should be addressed in a The findings presented in this paper have led us general program of financial development. Such a tentatively to conclude that the widely accepted wis- program would take into consideration the close dom that general capital shortages are a core problem correlation between the motivation for efficient sav- for development may be an illusion. The real bottle- ings mobilization and the investment climate. Evi- necks and binding constraints must be determined dence presented here suggests that interest rate ad- for each country at each state of economic develop- justment, which is a key instrument of financial lib- ment. eralization programs, has had little effect in improv- One of the core problems in Africa is the frag- ing savings mobilization and the efficiency of capital mented state of domestic resource mobilization. The allocation.15 In Africa, the financial market is still too system has been segmented into formal and informal shallow to make interest rates a key equilibriating sectors, and little interaction has taken place among variable between savings and investment. Under ex- the formal institutions. There is great potential for isting conditions, the freeing of interest rates is not savings mobilization for small-scale productive ac- likely to induce financial intermediaries to assume tivities through the development of the informal sec- greater risks. tor and the strengthening of potential links between Measures to be taken should include efforts to the formal and informal sectors. However, the limi- build up the local capabilities in risk and asset and tations of the present state of resource mobilization liability management, and in the government's mac- and intermediation are critical. roeconomic management. Risk sharing or credit risk Savings mobilized in the household sector through insurance schemes of some sort are urgently re- infornal agents and organizations are small in size quired.16 It must also be explicitly recognized that in per unit of transaction as well as short term in fre- a monocultural economy risk covariance is ex- quency. This must be in part a reflection of the nature tremely high. Financial intermediaries cannot func- of saving in low-income developing countries, which tion efficiently without appropriate devices and pol- 145 Mobilizing Domestic Resources for African Development and Diversificatiow Structural Impediments to Financial Intermediation icy instruments in place to deal with this problem. 6. Aryeeteyet al (1989) note thatthe liquidityproblems The mechanism of reducing the general financial risk of the rural population stem partially from their holding inherent in commodity based economies includes, at their savings in iliquid real assets. While they appear to the national level, improvements in the design and be sensitive to thecost of holdingfinancial assets, theyoften administration of macroeconomic and financial polic are less sensitive to the cost of borrowing to ease liquidity ies. A stable macroeconomic environment is a funda- problems. mental precondition if the interest rate is to function 7. As Germidis et al (1989) note, financial dualism can as an efficient instrument of savings mobilization. be ascribed as much to the shortcomings of the formal Issues related to the improvement of the interna- sector as to the intrinsic dualism of economic and social tional monetary, financial, and trade systems should structures in developing countries and the population's be reexamined. External official and private financial attachment to traditional modes of behavior. flows and foreign investment might assist in lessen- 8. Due to the large seasonal fluctuations characteristic ing the risks which are so detrimental to domestic of rural household income, however, it is necessary to savingsmobilization and hence to longterm, sustain- apply some caution in interpreting the annual estimate able development in low-income developing coun- presented by Bentil. tries. Above all, unless the risks associated with for- 9. An exception to this can be found in the case of eign exchange, interest rates, and commodity prices Malawi, where it has been estimated that one quarter of the exogenous to these economies are greatly reduced, total credit generated by the informal sector, K63.1 million, problems in domestic finance will remain. was used to finance trade and manufacturing in 1988 (Chipeta and Mkandawire, 1989a). Notes 10. Chipeta and Mkandawire (1989b) report that the government of Malawi has recently considered a number This paper is part of our ongoing research project on of new measures to expand credit resources for smallhold- "Domestic Resource Mobilization for African Development ers and small and medium enterprises. In view of the scope and Diversification' in which comparative empirical stud- for the expansion of informal financing of trade and diver- ies have been undertaken of the financial system of five sification, Chipeta and Mkandawire argue for including African countries - Ghana, Kenya, Malawi, Tanzania and the informal financial entities in the country's plan for Zambia. The project is set to look systematically and in a promoting small and medium enterprises. substantial way into the micro-macro links of the issues on 11. While in Table 1 the current account balance is finance and development. Due to the incompleteness of the expressed as a percentage of GNP to calculate the public study at this stage, this paper is very much preliminary and savings rate, in Table 2 it is expressed as a percentage of based largely on the interim reports on Ghana and Malawi. GDP in order tobe compatible with the rest of the indicators The author wishes to acknowledge the contribution of in the table. Ernest Aryeetey, Yaw Asante, Alexander Kyei, Fritz Gockel 12. Due to limited space, centralbanks, monetary policy, (Ghanaian team) and C. Chipeta and M.L.C. Mkandawire and other financial institutions are not included in this (Malawian team) whose research findings have been discussion. drawn on extensively in the writing of this paper. The 13. It is reported that the trade liberalization measures author would also like to thank the Gatsby Charitable taken in 1989 subsequently eased the banks' excess liquid- Foundation, SIDA and DANIDA for their generous finan- ity problems to some extent. cial support. 14. In orderto improve thesituation, Chipetaand Mkan- 1. For a critical review of the recent theoretical debate dawire (1989b) suggest that SEDOM and INDEFUND on finance and development, see Nissanke (1990). should be allowed to participate in domestic resource mo- 2. Aryeetey et al (1989) cite estimates made by Bentil bilization by issuing long term paper on which higher (1988). interest rates could be paid than those on bank deposits. 3. Aryeetey et al (1989) suggest that what may appear 15. Cho reports findings which suggest that 'in the as an investment may actually be saving as a hedge against presence of capital market imperfections, elimination of inflation. interest rate ceilings is no longer thought beneficial" (Cho 4. Since the coverage of the informnal financial sector 1986). AndersonandKhambata(1985)arguethat"theright differs between the two country studies, the size of the combination of policies is a relaxation of the controls com- informal sector presented here is not strictly comparable. bined with interventions to address shortcomings, It can be noted that both studies exclude credit unions. essentially of an institutional nature, in the capital markets' 5. As Chandavarkar (1985) argues, the existence of in- (p. 350). formal finance of this size may have significant 16. However, as Meyer and Cuevas (1990) note, many implications for the efficacy of monetary policy. conventional devices of loan insurance and guaranteed 146 Mobilizing Domestic ResourcesforAfrican D)velopment and Diversification Stuctural Impediments to Financial Inten,diation funds have failed to live up to expectations in terms of report presented at Workshop on Domestic Resource additionality in lending. This points to the need for more Mobilization in Harare, December 1989. innovative schemes. Ragazzi (1981) suggests establishing Chipeta, C. and M.L.C. Mkandawire (1989b), Thelnformal stabilization funds or other devices to cushion the impact Financial Sector in Malawi: Scope, Size and Role', pro- of adverse circumstances on small and medium enterprises gress report presented at Workshop of the AERC, Harare, and farmers. December 1989. Cho, YoonJe (1986), 'Inefficiencies from Financial Liberali- References zation in the Absence of Well-Functioning Equity Mar- kets", lournal of Money, Credit and Banking, vol. 18, no. 2, Alamgir, Mohiuddin (1976), 'Rural Savings and Invest- pp. 191-199. ment in Developing Countries: Some Conceptual and Deaton, Angus (1989), 'Saving in Developing Countries: Empirical Issues', Bangkldesh Development Studies, Janu- theory and review", paper presented at the First Annual ary 1976, pp. 1-48. World Bank Conference on Economic Development, Anderson, Dennis and Fardida Khambata (1985), 'Financ- Washington, D.C., April 1989. ing Small-Scale Industry and Agriculture in Developing Germidis, Dimitri, Dennis Kessler and Rachel Meghir Countries: The Merits and Limitations of 'Commercial' (1989), "Mobilizing Domestic Savings for Development: Policies", Economic Development and Cultural Change, pp. What Role for the Formal and Informal Financial Sec- 349-371. tors?", OECD, summary note on the Development Cen- Aryeetey, Emest, Eric Asante, Fritz Gockel and Alexander tre Study, Paris. Kyei ('1989a), 'Mobilizing Domestic Savings for African Ghosh, Dipak, (1986), 'Savings Behaviour in the Non- Development and Industrialization: A Ghanaian Case", Monetized Sector and Its Implications", Savings and De- an interim country report presented at Workshop on Do- velopment, no. 2. mesticResourceMobilizationinHarare,December 1989. Malawi/USAID (1987), "New Directions for Promoting Aryeetey, Emest and Fritz Gockel (1989b), 'Mobilizing Do- Small and Medium Scale Enterprises in Malawi: Con- mestic Resources for Capital Formation: The Role of straints and Prospects for Growth". Informal Financial Markets in Ghana', preliminary re- Meyer, Richard L. and Carlos E. Cuevas (1990), "Reducing port presented at Workshop of the AERC, Harare, De- the Transaction Costs of Financial Intermediation: The- c*ember 1989. ory and Innovations", paper prepared for the Interna- Bentil, B. et al (1988), "Rural Finance in Ghana', a research tional Conference on Savings and Credit for Develop- study prepared for the Bank of Ghana on behalf of the ment, Copenhagen, May 28-31. Ministry of Economic Co-operation of the Federal Re- Neal, Craig R. (1988), "Macro-Financial Indicators for One public of Germany. Hundred Seventeen Developing and Industrial Coun- Chandavarkar, Anand G. (1985), 'The Non-Institutional tries", December 1988. The World Bank, Washington, Financial Sector in Developing Countries: Macroeco- D.C. nomic Implications for Savings Policies", Savings and Nissanke, Machiko (1990), "Theoretical Issues on Finance Development, no. 2. and Development: a critical literature survey", Queen Chipeta, C. and M.L.C. Mkandawire (1989a), 'Mobilizing Elizabeth House, Oxford University discussion paper. Dom.estic Savings for African Development and Indus- Ragazzi, Giorgio (1981), "Savings Mobilization in Africa", trialization: A Case Study of Malawi', an interim country Savings and Development, no. 1. 147 15 Monetary Cooperation in the CFA Zone Patrick Honohan There has been considerable discussion over the past paper is not an encouraging one. Despite the fixed few years on the merits for small countries of joining exchange rate and an elaborate set of rules for avoid- a monetary union. For the developing world, this ing excessive credit expansion, the CFA zone has debate has centered on the experience of the CFA or almost foundered in widespread bank insolvency. franc zone in Africa. This zone is the largest and most Although the zone's institutional set-up seems equi- enduring of currencyblocs. Most of the recent policy table, in practice the burden of paying for the losses discussion on the CFA zone has focussed on three involved is likely to fall disproportionately on the types of question: competitiveness, the mechanics of poorer countries, while many of the non-performing bank restructuring and the comparative growth per- credits have been made in some of the more prosper- formance of CFA versus non-CFA countries.' The ous countries of the zone. focus of this paper is rather different. It is designed The paper is organized as follows. Section 2 de- as a guide to understanding how the zone works as scribes the member countries of the zone and points a mechanism for monetary cooperation. out in particular the diversity not only in their eco- When consideringthe merits of joininga monetary nomic structure but in the state of the financial sys- union small countries naturally value the credible tem. Some countries are net borrowers within the commitment to exchange rate and price stability zone, some net lenders. This might suggest that some which membership represents and which would be efficient mechanism is in place to channel investable hard to sustain through unilaterally pegging one's resources to where they are most needed, but this is currency.2 There is also a potential saving through not the case. Section 3 shows that there is no effective pooling of the external reserves of member curren- regional money market and considers how this cies. might be remedied. In the absence of a market to But there might also be other advantages. Capital determine the regional distribution of credit, this might flow more freely to where it was most needed must be done administratively. Section 4 discusses within a monetary union. Provided the distribution how credit has been distributed between the member of benefits of the union was reasonable, this could be countries and highlights the asymmetries which to the advantage of all members, even those whose have arisen. While temporary fluctuations in export low capital productivity resulted in their becoming receipts have influenced part of the distribution, the net lenders within the union. Furthermore the opera- bulk of available credit has not been used as a revolv- tion of monetary policy and the prudential supervi- ing fund to meet such temporary needs, but has sion of the banking system might be more effective if increasingly gone to a few favored countries. Again the resources of several small countries were pooled this cannot be explained as a flow to where funds are in a strong and independent central bank. best rewarded, as much of the credit has gone to This paper examines whether such additional ad- banks which have subsequently failed. The costs of vantages have been realized in the CFA zone. Unfor- these bad debts affects the distribution of the seignor- tunately, the experience which is described in this age benefits of the zone; section 5 explains how seig- 148 Monetary Cooperation in the CFA Zone norage is divided in the unions and shows how the * an annual monetary programming exercise losses may fall disproportionately on non-favored which determines the planned growth in do- countnies. Section 6 concludes by considering how mestic credit in each member country. the recentbankingcrisis has revealed the deficiencies 0 implementation of this monetary program in the zone's arrangements. through credit ceilings to each government and, in some cases, to the private sector, or The CFA Countries through ceilings on central bank refinancing of private sector credit. The CFA zone is made up of two separate unions, * administered interest rates, including prefer- the West African Monetary Union (UMOA) and the ential rates for priority sectors. countries which have formed the Bank of the Central African States (BEAC).3 Though not all of the former The total land area of the CFAzoneamounts to 31 % French colonies in Sub-Saharan Africa (SSA) joined of SSA, though its population is only 16% and its total one or other of the two unions, and though there have GNP about 22% of that in SSA. The members include been defections4 the current situation is that thirteen some of the poorest countries in the world (chief countries with a total population of some seventy economic indicators for the members are shown in million share just two central banks and two curren- table 15-1). All of them rely heavily on exports of a cies. The currencies, each of which is known as the variety of primary products for the foreign earnings CFA franc, are both valued at one-fiftieth of the needed for imports of necessities. Most are also heav- French franc, a peg theyhave retained for fortyyears. ily dependent on official transfers and concessional Along with the fixed exchange rate, the members of lending.5 the CFA zone have traditionally retained an open It is striking how varied the CFA countries are. intemnational capital market, with statutory freedom Some are open to the sea, some are landlocked. In of capital movements among the member countries terms of surface area, each of the largest countries and with France. (Chad, Niger and Mali) is almost fifty times the size The monetary mechanism in the two zones has of the smallest, Equatorial Guinea. The density of been characterized by: population in Togo is almost fourteen times that in Table 15-1. CFA Zone-general indicators GNP GNP (per Life GDP (per capita GDP GDP Avg. Avg. expect- Population Area (1986 capita growth growth growth inflation inflation ancy M2/ MU (millions) ('OOs $ 1986- 1965 1965- 1980- 1965- 1980- years GDP GDP Country mid. 1986 sq. km.) million) dollars) 1980 1980 1986 1980 1986 1986 1965 1985 UMOA Benin 4.2 113 1,320 270 0.2 2.3 3.6 7.4 8.6 50 10.6 22.8 Burkina Faso 8.1 274 930 150 1.3 3.5 2.5 6.2 6.3 47 9.3 22.1 C6te d'lvoire 10.7 323 7,320 730 1.2 6.8 -0.3 9.3 8.3 52 21.8 29.4 Mali 7.6 1,240 1,650 180 1.1 4.1 0.4 - 7.4 47 - 23.0 Niger 6.6 1,267 2,080 260 -2.2 0.3 -2.6 7.5 6.6 44 3.8 15.9 Senegal 6.8 196 3,470 420 -0.6 2.1 3.2 6.5 9.5 47 15.3 24.5 Togo 3.1 57 980 250 0.2 4.5 -1.1 6.9 6.7 53 10.9 45.3 UMOA 47.0 3,470 18,020 358 0.0 4.0 1.0 8.0 8.0 49 16.0 26.0 BEAC C.A.R. 2.7 623 900 290 -0.6 2.6 1.1 8.5 11.5 50 13.5 17.4 Chad 5.1 1,284 - - - 0.1 - 6.3 - 45 9.3 25.5 Cameroon 10.5 475 11,280 910 3.9 5.1 8.2 9.0 11.0 56 12.5 19.4 Congo 2.0 342 2,000 990 3.6 5.9 5.1 7.1 7.5 58 16.5 20.1 Gabon 1.0 268 3,190 3,080 1.9 9.5 1.5 12.7 4.8 52 16.5 26.3 BEAC 21.0 2,992 17,370 950 3.0 5.0 6.0 9.0 9.0 55 14.0 - Sub-Saharan Africa 424.0 20,895 165,990 370 1.0 6.0 0.0 13.0 16.0 50 - - Sources: IMF and World Bank data 149 Monetary Cooperation in the CFA Zone Chad. Per capita income in oil-rich Gabon is esti- of the UMOA, a listing of the products which directly mated at fifteen times the figure for Burkina Faso. or indirectly account for at least five per cent of Only in regard to population size and inflation rates exports (to industrial countries) involves fourteen does the range of CFA experience fail to encompass quite distinct products (table 15-3). the Sub-Saharan Africa average (table 15-2).6 Even Since the member countries of both unions (but here, there is a wide divergence within the CFA zone especially the UMOA) depend on different products with individual country populations from less than for the bulk of their exports, and since the prices of a half a million to over 11 million; and, though long- these commodities are not very closely correlated,7 term inflation trends converge around the French there is substantial scope for pooling the risks of norm (see Honohan, 1990), cumulative 1980-86 infla- fluctuations in export markets. tion varied between 32% and 92% . Reflecting the wide differences in level of develop- Exports of CFA countries are overwhelmingly ment, the range and number of financial institutions (over 90 per cent in each of the countries) composed varies from country to country within the zone. The of primary products or of lightly processed primary dominant financial institutions are banks; the non- products such as vegetable oils, cotton fabrics, bank financial sector is negligible. Most of the largest wooden veneers and fertilizers. However, it is signifi- commercial banks are at least partly government- cant that different countries depend on different owned, often with key management and systems product mixes. For example, for the seven countries provided by French banks which are minority share- holders. There are also large government-owned de- Table 152. The CFA countries in Sub-Saharan velopment banks, some of which accept deposits. As Africa has already been mentioned and is further discussed below, banks in most of the countries have been SSA found to be insolvent. This applies to a substantial average CTA lo C-A hi fraction of the banking system in Cameroon, C6te Populatio)0density d'Ivoire, Senegal, Benin and Chad. (pop./kmD) 20 4 The monetary structure of the UMOA countries Population (million) 12 <1 11 also displays wide variations from country to coun- Area (km2) 597 28 1,284 try. As shown in table 6, many of the CTA banking lnflation average.. (1980-86)% 16 5 12 systems are heavily indebted to the central banks. Life086% expectancyThe average indebtedness of the commercial banks at birth(yrs) 58 to the central bank throughout the zone in recent Gtrowth average years has been about one-third of their non-govern- (1980-86) % 0 -3 8 ment lending. There are just two countries (Togo and Percent share in GDP: Burkina Faso) at the other extreme, with the commer- Agriculture 36 10 50 cial banks accepting deposits so far in excess of what Industry 25 13 54 they can usefully lend that they place substantial Savings (GDS) 11 7 30 deposits with the regional central bank. But in most Investors 14 9 37 of the countries, and especially in C6te d'Ivoire, Sene- Export share 19 14 47 gal and recently Cameroon, lending by the commer- Expovernmt shre ms 19 14 47 cial banks far outstrips deposit resources, with the Governm ent ronucers 13a9 26 shortfall being made up through refinancing by the Primary products as a 21 96 central bank. In all of the countries, direct lendingby Percentage of expofacturts 5the banks to government is small. Percent of manufactured exports to industrial centers 34 11 82 The central bank in each of the two unions thus Persons per physician 25 3 56 provides an intermediation function between the fi- School enrollment: nancial systems of the surplus and deficit countries. Primary % 75 23 >100 One way of looking at this is to observe that, while Secondary % 23 5 25 the central banks' total net claims on commercial Agricultural share in the 75 62 87 banks is of the same order of magnitude as the total workforce% currency issue, three countries in UMOA and one in Urbanization % 23 8 4 BEAC have commercial banking systems indebted to Infant mortality % 113 74 144 the central bank in an amount in excess of the na- tional currency holdings. In effect, some of the re- Source: World Development Report 1988, The World Bank, sources mobilized through currency issue in the Data refers to 1986 or a recent year. 150 Monetary Cooperation in the CFA Zone Table 15-3. 7UMOA: Commodity composition of exports Product Benin Burkina Faso C6te d'Ivoire Mali Niger Senegal Togo Cotton 21 35 - 68 - - 5 Coffee 7 - 25 - - - 40 Cocoa 16 - 36 - - - 14 Uranium 5 - - - 97 - - Phosphates - - - - - 16 Fruit and - 7 7 - - - _ Vegetables Groundnuts - - - _ _ 44 Karite - 36 - Almonds Petroleum 35 - - Fish - - - - - 28 Wocd - - 15 - - - Gold - - - 14 - - Hides - 7 - Note: Table shows percentage of each country's exports to industrial countries of varios commodities and products manufactured from. those commodities. Only commodities accounting for more than 5% of total exports are included. Source: World Trade Statistics, UN. other countries are being channeled to the banking about counterparties associated with language dif- systems in the other four countries. (The central ferences and limited trading relations.8 Some of these banks also lend to the government, their overall in- problems are well dealt with by the CFA unions, but termediation role is discussed further below). some are not and the net result is that no effective union-wide regional financial market exists. A Union-Wide Money Market? Where the CFA unions do function reasonably well is in the area of exchange controls and foreign ex- In a single country the financial system is relied change risk, though even here there are problems. In upon to intermediate financial flows from surplus to particular it has not always been as easy in practice deficit entities. Bank and non-bankfinancialinterme- to transfer funds from CFA countries to France in diaries and securities markets are the means by wh- recent years as it has been in theory.9 Transfers of ich savings flows reach those who need to borrow in funds between the CFA countries are much freer order to invest in real resources or in order to meet a especially since there are only two currencies in- temporary excess of consumption needs over in- volved: transfers between two countries of the come. This is not always done with perfect efficiency. UMOA involve no foreign exchange transaction; Various market and institutional obstacles stand in likewise between two BEAC zone countries; never- the way. For example informational asymmetries theless, there is always the residual risk that a mem- and the incentive for abuses mean that savers do not ber country might, at a later date, impose restrictions have perfect confidence in those who wish to use the- on foreigners attempting to repatriate funds. The ir funds. Standard financial contracts (such as the de- fixed exchange rate reduces but does not eliminate bt contract) do not always divide the risks between the risk of foreign exchange losses: balance of pay- lender and borrower in an ideal manner; contracts ments difficulties have, from time to time, generated which mightbebetter from that point of view can pr- market fears of a change in the parity. For capital ove to be unenforceable. All this is common to all co- between UMOA or BEAC countries those risks are untries. The challenge for the monetary union is to smaller. after all, one is dealing with a single cur- attempt to ensure that financial intermediation wo- rency and the scenarios under which repatriation of rks as smoothly across international frontiers within the funds to another African country within the un- the zone as it works within each member country. ion could entail an exchange loss are more remote. National frontiers could create several types of Cotonou in Benin is about 2,500 km from Dakar in obstacle to financial intermediation, including con- Senegal. This figure, together with the fact that the trols on foreign exchange movements, risk of UMOA is larger than India, gives some idea of the changes in foreign exchange rates, lack of knowledge geographical scale of the unions we are concerned 151 Monetary Cooperation in the CFA Zone Table 15-4. ULMOA: Variability of export prices Country Period Benin Burkina Faso C6te d'lvoire Mali Niger Senegal Togo 1980 21 35 - 68 - - 5 1981 7 - 25 - - - 40 1982 1 - 36 - - - 14 1983 5 - 97 - - 1984 - - - - - 16 - 1985 - 7 7 _ _ - _ 1986 - - - - - 44 - 1987 - 36 - - - - _ 1988 35 - - - - - - Variance - - - - - 28 Matrix of correlations Burkina Faso C6te d'Ivoire Mali Senegal Togo Benin 0.067 0.415 0.667 0.625 0.711 Burkina Faso - 0.927 0.206 0.130 0.474 C6te d'lvoire - - 0.385 0.292 0.693 Mali - - - 0.679 0.681 Senegal - - - - 0.608 Source: World Bank data about. Over such large scales the information flows Furthermore, the gap between borrowing and de- necessary to build sufficient confidence to allow bor- posit rates in the marche is as wide as one per cent rowing and lending to take place efficiently require per annum-far higher than interbank spreads in a the use of trusted intermediaries. In the present stage developed money market. of development of CFA financial markets this means The situation in the BEAC zone is no better, there that savings flows must be intermediated by the is no "marche monetaire", though the BEAC does banking system. Thus a saver in Dakar may trust his accept deposits and makes loans. Once again the local banker, a bank in Cotonou may have a credit- loans are quite discretionary and depend on policy worthy client. Unfortunately, international loans considerations other than pure banking prudence. within the unions, even between banks, are virtually The interest rates on loans bear no relation to market- unknown. (Even within a single member country as clearing rates. developed as Senegal or Cote d'Ivoire the interbank In short, it has to be said that there is no effective marketisverythinand confined to the leadingbanks regional money market in the CFA zone. Under the only.)'I Accordingly, it is unlikely, in the posited situ- present circumstances, the easiest way to encourage ation, that the Dakar funds will find their way to the the development of one would be for the central borrower in Cotonou. banks to act as intermediaries matching the needs of The BCEAO does operate what is known as the sound deficit banks with the supply from surplus 'Marche Monetaire", but this is little more than the banks. In view of the fragile state of the banking combination of a deposit facility for banks, and a system it ismore than unlikelythat anyprivate inter- facility for secured borrowing by banks from it and mediary would undertake this activity. After all, at its discretion. Thus, if a bank in Benin for example, three large French banks have affiliates in most of the has a demand for liquid funds it may not be able to CFA countries, yet they do not undertake interbank meet this demand from the marche monetaire even transactions even between affiliates in different if it is able to provide adequate security, and even if countries. Only the central bank has sufficient infor- a bank in Senegal, for example, has just placed sur- mation to judge the soundness of participant banks. plus funds with the marche monetaire. But each central bank must also manage monetary The BCEAO has not attempted to adjust interest policy, and this is also done by means of influencing rates to clear the domestic market forbankliquidity.11 the quantities of bank liquidity in the different coun- 152 Monetary Cooperation in the CFA Zone Table 15-5. Matrix of trade patems of CFA countries (Exports % of total exports of each country) Burkina C6te Camer- Total French Benin Faso d'lvoire Mali Niger Senegal Togo oon C.A.R. Chad Congo Gabon CTA franc Benin 0.0 0.0 0.0 0.1 1.1 0.1 0.0 0.0 0.0 0.0 0.1 0.0 0.1 0.0 Burkina Faso 0.1 0.0 0.0 0.1 0.3 0.0 0.2 0.0 0.0 0.0 0.0 0.0 0.1 0.0 C6te d'lvoire 3.3 28.9 0.0 23.3 9.6 4.2 5.4 1.1 0.0 0.0 0.0 1.4 4.2 0.4 Mali 0.0 0.4 0.2 0.0 0.1 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.1 0.0 Niger 0.6 0.0 0.1 0.2 0.0 0.0 0.1 0.0 0.0 0.1 0.0 0.0 0.1 0.2 Senegal 0.9 1.2 1.6 6.4 0.8 0.0 0.9 0.9 0.2 0.2 0.8 0.2 1.1 0.1 Togo 0.1 0.3 0.0 0.1 0.6 0.0 0.0 0.0 0.0 0.1 0.0 0.0 0.1 0.0 Cameroon 0.0 0.0 2.8 0.0 0.0 0.0 0.0 0.0 0.0 19.5 0.0 0.8 1.0 0.4 C.A.R. 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.2 0.0 0.0 0.0 0.0 Chad 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.4 0.0 0.0 0.1 0.0 0.1 0.0 Congo 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.9 0.0 0.0 0.0 0.1 Gabon 0.0 0.0 1.4 0.0 0.0 0.5 0.0 0.0 0.0 0.0 0.0 0.0 0.4 0.4 CFA 4.9 31.5 6.3 30.2 12.5 4.7 6.6 2.4 0.3 21.0 1.0 2.4 0.0 1.7 French franc 21.6 34.3 35.3 29.3 39.3 34.9 27.5 43.0 59.3 38.0 53.4 55.0 38.7 0.0 CFA franc 26.4 41.6 41.6 59.9 51.9 39.6 34.2 45.4 59.6 59.0 54.4 57.4 45.9 1.7 World 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 BenirL 0.0 0.3 0.7 0.0 0.3 0.9 0.2 0.0 0.0 0.0 0.0 0.0 0.3 0.1 Burkina Faso 0.1 0.0 4.3 0.2 0.7 0.8 2.0 0.0 0.0 0.0 0.0 0.0 1.3 0.1 C6te d'lvoire 0.3 1.8 0.0 3.6 0.6 5.7 0.0 2.3 0.0 0.0 0.0 1.6 1.3 0.5 Mali 0.2 0.5 4.2 0.0 0.2 5.1 0.2 0.0 0.0 0.0 0.0 0.0 1.5 0.1 Niger 1.5 0.8 1.2 1.3 0.0 0.4 0.7 0.0 0.0 0.0 0.0 0.0 0.4 0.1 Senegal 0.0 0.0 1.8 0.3 0.0 0.0 0.0 0.9 0.0 0.0 0.0 0.3 0.6 0.3 Togo 0.2 0.5 0.3 0.1 0.1 0.7 0.0 0.0 0.0 0.0 0.0 0.0 0.1 0.1 Cameroon 0.0 0.0 0.8 0.0 0.0 2.7 0.0 0.0 0.0 10.6 0.0 0.1 0.5 0.6 C.A.R. 0.0 0.0 0.0 0.0 0.0 0.0 0.1 0.0 0.0 0.0 0.0 0.0 0.0 0.1 Chad 0.0 0.0 0.0 0.0 0.0 0.1 0.1 1.2 0.3 0.0 0.1 0.0 0.3 0.0 Congo 0.2 0.0 0.1 0.0 0.1 0.0 0.1 0.0 0.0 0.5 0.0 0.0 0.1 0.2 Gabon 0.0 0.0 0.5 0.0 0.0 0.3 0.0 0.3 0.0 0.0 0.0 0.0 0.2 0.4 CFA 2.6 3.8 13.9 6.4 2.0 17.7 3.5 3.9 0.3 11.2 0.1 2.0 0.0 2.5 French franc 10.8 38.1 24.9 14.1 88.0 34.8 17.0 26.4 18.1 12.6 12.8 35.9 38.7 0.0 C(A franc 13.4 42.0 38.8 20.5 90.0 52.5 20.4 30.3 18.3 23.8 12.9 37.9 45.9 2.5 World 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 Source: World Bank data tries. An institutional framework for resolving such the monetary policy arm. The latter would adjust the conflicts would need to be devised. interest rate on its loans to ensure that the market For example, in order to ensure a separation of the cleared, i.e. that the intermediation arm was not bor- functions of monetary control and intemational in- rowing more than the planned ceiling. Having a termediation, one arm of the central bank (call it the market-clearing interest rate is the key to satisfactory "monetary policy arm") would, as at present, be operation of international intermediation within the charged with planning the available amount of cen- union. Sound banks will not over-borrow if the inter- tral bank credit for the union as a whole. Anotherarm est rate is not subsidized.12 (the "intermediation arm") would match supply of To summarize, the international flow of funds and demand for bank liquidity on the basis meeting within the UMOA and BEAC unions is not supported any demands from sound banks presenting accept- by an adequate institutional infrastructure. The pre- able paper as security. The intermediation arm vailing risks are too great for private markets to would accept deposits and would also borrow from perform this function given the information at their 153 Monetary Cooperation in the CFA Zone Table 15-6. CFA countries monetary survey (end-1987, in billions of CFA francs) Clains of central bank Net external Domtestic on commercial assets credit Money (M2) Currency bank Benin -55 106 97 20 49 Burkina Faso 68 68 127 44 8 C6te d'Ivoire -378 1,363 930 305 511 Mali -43 189 130 61 32 Niger 19 104 114 36 30 Senegal -206 539 332 101 156 Togo 65 97 164 48 5 UMOA -530 2,466 1,894 615 780 Cameroon -150 939 678 171 342 Central African Republic 18 48 60 41 13 Chad 9 78 76 47 39 Congo -32 219 138 55 42 Equatorial Guinea 4 11 9 7 4 Gabon -33 384 238 49 29 BEAC -192 1,679 1,199 370 467 Source: IMF Finance Yearbook disposal. The central banks of each union are better on a simple standard fixed exchange rate quantity positioned to channel funds from surplus to deficit theory framework. Credit creation in excess of the countries. At present they do this without explicit increase in money demand will spill over into the reference to market criteria; they should move to a balance of payments and deplete the union's re- market-based system of allocating bank liquidity serves. Therefore the credit plan is in principle de- throughout their respective zones, while preserving signed to limit such excesses, both for individual monetary control. countries and for the union as a whole. Broadly speaking, an expansion in central bank Monetary Policy refinancing of bank credit leads to an expansion in overall bank credit, franc for franc. Although the Ideally, the available credit in the union should go money stock responds to such an increase at first, to where it yields the highest return. This would thereisnosustained increase in moneydemand: over presumably involve flows to support domestic ex- time the credit expansion leaks out into the balance penditure in member countries faced with a transi- of payments.14 tory balance of payments difficulties resulting from Because there are no private banking flows be- adverse export shocks. On a longer term basis, there tween member countries, central bank refinancing in should also be a tendency for members where the one country will result in an increase in bank credit marginal product of capital was low to be net lenders in that country, and not in the other countries. This towards those where the marginal product of capital allows the central bank to pursue independent credit was high. In this section the process of credit alloca- objectives for each member country, as is often re- tion is reviewed to assess whether it flows in accord- quired by IMF programs. (If a true international ance with these criteria. money market were established in the unions then In order to protect the external position of the union-wide liquidity might have to be adjusted in union, and ultimately to ensure the sustainability of response to an emergent overshoot of the credit ceil- the fixed parity, the monetary authorities avoid cre- ing in one country).15 ating too much domestic credit. In practice, there is Among the factors apparently influencing the al- an annual credit planning exercise for each country location of central bank credit between countries, at separately, with the individual plans being simply least in the short-run, appears to be each country's added together to obtain the monetary plan for each export receipts. In particular, UMOA countries who union. The modelling underlying the plan is based experienced a decline in the value of exports (relative 154 Monetary Cooperation in the CFA Zone to the average of the union experience) have tended ate share in the regional distribution of credit. The to obtain a higher share of central bank credit. To this major shifts in the distribution of credit over the years extent, the members have been able to benefit from in favor of these countries have shown no tendency the risk-pooling afforded by the diversity of export to be reversed. Figure 15-1 shows lending to banks products in the union. However, efficient pooling of by the BCEAO expressed as a percentage of bread these risks requires the available credit to be used as money. The upper panel of the figure shows the th- a revolving fund, with borrowings to cover short- ree countries, C6te d'Ivoire, Senegal and Benin, term disturbances being repaid in better times. which have received the greatest proportionate share In fact, not all countries appear to treat central bank ris ing to more than 60 percent of each country's wide credit to the banking system as a revolving fund. moneystock thelowerpanelshowtheotherthree,17 Somehow, at least in the UMOA, the larger and more none of whom ever received more than about 40 prosperous countries have gained a disproportion- percent of wide money (less than 20 per cent for Figure 15-1. BCEAO lending to banks as a percentage Figure 15-2. Net Foreign Assets as a percentage of M 2, of M2, 1970-88 1970-88 Cote d'lvoire, Senegal and Benin Cote d'Ivoire, Senegal and Benin Percent Percent 70 120 O -100 60 80 50 ~~~~~~~~~~~~~~~~~60 50~~~~~~~~~~~~~~~~~4 40. - ,, . .- ~~~~~~~~~~~~~20 30 0~~~~ --30ni Co e'l o r - 0 - -Sega---BnnCtdlor - - - -Sega 70 \.. -20 S 20 8 40 10 -80 - 0 - 1 I l l ll l l l l i 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 -- - - Benin - Cote d'lvoire - - - - Senegal - -- Benrin Cote d'Ivoire - Senegal Source: World Bank data. Source: World Bank data. Burkina Faso, Niger, and Togo Burkina Faso, Niger, and Togo Pe-rcent Pe-rcent 70 120 100 60 8'' 40 40 /~~~~~~~~~~~~~~~~~4 40 2~~~~~~~0. -' 30 ~~~~~~~~~~~~~~~~~~~-20 20 -40 -60 10 ~~~~~~~~~~~~~~~~~~~-80 0 - -100 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 19701972197419761978198019821984198619881970 1972 1974 1976 1978 1980 1982 1984 1986 1988 Burkina Faso - Niger -- - Togo - - . Burldna Faso - Niger - - - - Togo Source: World Barnk data. Source: World Bank data. 155 Monetary Cooperation in the CFA Zone Togo) and whose credit has tended to revert to low central bank dividends-or seignorage-will have levels. been seriously eroded by these losses. It is to these BCEAO direct lending to governments has also financial returns that we now turn. been important, even when one leaves out borrow- ing from the IMF channeled through the Central The Distribution of Seignorage Bank. The same country pattern is true of total BCEAO lending, including lending to governments. Membership of a monetary union is not a zero-sum Since excessive credit expansions have leaked out game; some of its benefits, such as those of price through the balance of payments, this skewed distri- stability, are public goods that can be enjoyed by all. bution of credit has resulted in a skewed pattern of The benefits of seignorage are divided among the external deficits and external reserves, with the three members. This section explains how seignorage is larger countries holding disproportionately low divided and shows that the division is not as equita- gross external reserves. ble as it may appear at first sight. Cote d'Ivoire's share of union GDP has fluctuated Broadly speaking, seignorage is the benefit which between about 40 per cent and 55 per cent in the accrues to theissuerof moneybyvirtueof beingable period 1970-86 (with its share of union population to issue interest-free currency rather than having to growing from 23 per cent to over 27 per cent). Before incur interest charges on borrowing. The total value 1980 C6te d'Ivoire's share in union external reserves of the currency issue in the UMOA is about $2 billion, fluct- uated between 22 per cent and 62 per cent, but in the BEAC zone it is about $1.3 billion. Most of the in that year it fell below 7 per cent and has been as benefit of the seignorage on this currency issue low as 1 percent, it share has never recovered to should flow to member governments through the pre-1980 levels. A similar pattern is noted for Senegal distribution of BEAC profits, but some is transferred and Ben- in. The fall in these three countries' shares to those who are given subsidized credit from the of external reserves was taken up largely by Togo. Central bank. AlthoughTogo's share of unionGDP hasbeen falling The easiest way to trace the flow of seignorage is from only about 7 per cent to about 5 per cent, and to consider the accounts of the central banks. Taking although its share in union external reserves never the BCEAO 1986 balance sheet (table 15-7) for exam- exceeded 20 per cent before 1980, since 1981 its share ple it can be seen that currency is the largest single in union external reserves has never fallen below 40 item19 representing over half of the BCEAO's liabili- percent. ties. The principal corresponding assets are claims on Tlhis dramatic shift in relative shares of the union's banks and on the national treasuries, the latter being gross external reserves corresponds to a dispropor- little more than one-third as important as the former. tionate allocation of central bank credit to Cote These claims are interest bearing, while the currency d'Ivoire, Senegal and Benin. It is also reflected in the is not. This offers a margin which, after the operating net external assets of the monetary system (figure 2). costs of the BCEAO are deducted, represents most of The excess credit leaked out into the balance of pay- the flow of benefits to member governments arising ments resulting effectively in these countries indi- out of seignorage. The remainder of the seignorage rectly borrowing from (poorer) Togo to finance im- is passed through the implicit interest subsidy on ports. This would be unsurprising on a temporary government borrowing and on some of the commer- basis, but as it has been sustained for many years it cial bank borrowing from the BCEAO.W suggests that there is no adequate mechanism for Themostnatural distributionofseignoragewould ensuring that the resources of the union are treated be in proportion to each member's national currency as a short-term revolving fund. The forces for repay- issue. In accordance with the UMOA statutes, how- ment are extremely weak.18 ever, the BCEAO's dividend is distributed equally As already suggested, a perrmanent pattern of per- among the member states. Therefore were it not for sistent borrowers and patient lenders might be justi- the subsidy on BCEAO lending rates it would seem fied on the basis of a higher marginal product of that we could say that the seignorage is divided capital in the borrowing countries, in which case the equally among the countries and thus represents a lending countries might achieve a better return than progressive redistribution from the large (and rich) if they had increased domestic expenditure. How- to the small countries).21 But there is an important ever, the fact that much of the credit has gone to caveat. banks which have failed proves that credit has not The caveat arises because a large part of the sought out high-yielding activities. Indeed, the fi- BCEAO's lending to banks has become non-perform- nancial return to smaller members in the form of ing. In Benin, Senegal and C6te d'Ivoire in particular, 156 Monetary Cooperation in the CFA Zone the BCEAO holds especially large claims on insol- -the three most prosperous-have taken the bulk of ventbanks. Althoughit falls to each national govern- central bank credit and apparently squandered it. ment to meet BCEAO claims on insolvent banks, the Between them, banks in these three countries have BCEAO has already decided to consolidate these run up a net capital deficiency approximately equal debts at very concessional rates. As the total of the to the average currency issue for the whole union-a claimsinquestionareof thesameorderof magnitude sum not far short of $2 billion or ten per cent of the as the currency issue, such concessions will severely union GNP for one year. It is as if the central bank dent the flow of seignorage benefits coming through had printed the entire currency issue only to hand it the BCEAO dividends. In fact, by making some very to insouciant borrowers in these three countries. rough calculations on the distribution of seignorage How does it arise that the three countries with the after retrospective account is taken of the annualized largest bank insolvencies are also the countries which cost of these concessions, it appears that close to one have received the lion's share of central bank credit? hundred per cent of the seignorage of recent years It is difficult to avoid the inference that, for all the talk has benefitted C6te d'Ivoire, Senegal and Benin.22 of monetary programming, the inter-country alloca- The full story about whobenefits from the seignor- tion of credit has been driven by pressure to finance age therefore becomes a combination of equal distri- what has proved tobe unsound lendingby thebanks bution of central bank profits, subsidies on treasury in the most influential countries. Perhaps envisaging borrowings, and concessions made to individual such an eventuality, the statutes of the union require countries in respect of bank insolvencies.23 Though each national government to make good any claims the smaller member countries have benefitted over of the BCEAO on an insolvent bank.25 Despite this, the years by a larger distribution than they would the governments involved have obtained generous have received had the distribution been proportion- rescheduling terms from the BCEAO and are also ate to currency issue, they will now suffer by the hoping for substantial additional concessionary as- equal distribution of losses on the BCEAO's portfo- sistance from bilateral and multilateral donors to lio. In the end it may be that they will effectively lose finance the debts. all of the seignorage. If one is to think in terms of winners and losers, so For the BEAC, the seignorage analysis is more far the winners are those who benefitted from the complex. The distribution of BEAC profits is not equal, but depends on a formula designed to take Table 15-7. BCEAO balance sheet account of the actual contribution of each member (September 1986) In billions CFA Francs country to the profits of the BEAC.24 Whatever about Assts the merits of the formula in normal circumstances, it Gold 21 is not clear how it will be applied if it becomes Gd16 necessary to take account of heavy write-offs on IMRreserve positionandSDRs16 BEAC loans to weak banks. Therefore it is too early Foreignexchange 35 to judge the likelydistributional impact among coun- (of which operations account) (95) tries of such losses. Claims on banks and OFls 596 The overall conclusion must be that the arrange- Claims on national treasuries 431 ments for distributing seignorage have been vitiated (of which IMF and Trust Fund) (399) by the failure of the central banks to ensure that the Fixed assets, etc. 47 resources mobilized by the currency issue were in- Suspense accounts and miscellaneous 341 vested in a sound manner. Total assets-total liabilities 1,809 Liabilities Concluding Remarks: the Banking Crisis Current 528 Claims on foreign banks, etc. 57 Previous sections have shown that, although the Claims of local banks and OFIs 234 CFA zone has ensured a degree of price stability over Claims of national treasuries, etc. 80 the long-run, money and credit policy has not been Transfers being executed 10 without problems. The distribution of credit among Obligations in foreign exchange 0 different member states has not followed market anditreignd ° principles but has been determined administratively. IMF, SDRs, and trust fund While the administrative arrangements are not obvi- Capital and reserves 107 ously inequitable in themselves, the UMOA has slid Suspense accounts and miscellaneous 340 into a very lop-sided situation where three countries Source: BCEAO. 157 Monetary Cooperation in the CFA Zone unwise and un-repaid loans made by the failed Cdte d'Ivoire and Senegal might, in that wider banks. The four poorest countries of the UMOA are, grouping provide a check on the others. so far, losers; their claims on future dividends from the BCEAO are compromised by the uncertain value Notes of its claims on the failing banks and the three less- poor governments. The tentative calculation above 1. See,forexample,thereferencesbyDevarajanandde suggested that these countries have lost seignorage Melo (1987a and b), Dittus (1987), Guillaumont & Guillau- equivalent in capitalized terms to the value of cur- mont (1984 and 1988), Krumm (1985 and 1987), Ouattara rency in circulation. (1986). Perhaps over-emphasis on the question of main- 2. By embodying the exchange rate decision in a multi- taining the parity led to preoccupation with macro- country institutional set-up involving a joint central bank, economic aggregates, such as the total amount of each member country commits itself to an arrangement central bank credit in the system to the neglect of the form which would be rather costly to exit. This pre-com- prudent distribution of this credit, and the adequate mitment enhances the credibility of the fixed exchange rate management of banks in the union. It is even possible arrangement for each member. Pressures on any one mem- that the device of a statutory ceiling on advances to ber to adjust the exchange rate will be more easily resisted government (designed to facilitate the credit restraint by the authorities because of the sunk costs of union mem- necessary to assure the exchange rate) provided a bership. The credibility of the exchange rate commitment false sense of security and may have perversely led could then lead to a more stable monetary and inflation the central bank into a complacent approach to the evolution. A companion paper (Honohan, 1990) provides growth of parastatal and government influenced evidence that franc zone inflation has been largely deter- credit. mined by French inflation. On any reckoning this is a very poor performance 3. While much of what is said in this paper applies to of the union. It leads one to ask whether alternative the whole CFA zone, some of the evidence provided is statutory framework could have led to a more favor- based only on experience in UMOA. able outcome, with the resources of the union being 4. Guinea (Conkary) abandoned the French franc distributed in a genuinely revolving fund to meet shortly after independence; Mauritania and Madagascar fluctuations in external needs of individual countries stayed with the French franc until 1973, though neither rather than being systematically channeled to a select joined a union. Mali left the Franc zone after independence, few. The establishment of a more effective and mar- only to rejoin it in 1967, and to become a member of the ket-based mechanism for channeling surplus funds UMOA in 1984. Togowas notafoundermemberof UMOA would help: a true "marche monetaire" where sol- either, but joined after a change of government in 1963. In vency of the borrowing institution would be the sole 1985 Equatorial Guinea became both the smallest member criterion of whether it could have access to funds at of the (FA zone and the first member not to have been a a market-clearing rate. former French colory. Monaco and the Comoros Islands In the BEAC there are also serious problems of are the remaining non-CFA members of the franc zone bank insolvency, though in this case the evolution of apart from France and its dependencies. central bank credit has changed sharply in the last 5. Official development assistance to the eight poorest couple of years, largely as a result of the fall in oil members averaged over 12 percent of GNP in 1987. prices, so that it is too early to speak in terms of 6. And even for inflation the median of Sub-Saharan long-term structural inequalities between countries African countries is within the CFA range. in the BEAC zone. 7. Table 15-4 shows that the variability of each UMOA One must fear that the very unequal distribution country's typical export basket has not been strongly cor- of economic power of the member countries will still related across the countries in the 1980s. give C6te d'Ivoire, and to a lesser extent Senegal, in 8. Note, however that the reliability of official statistics the UMOA, and Cameroon in the BEAC, the ability which indicate a rather low level of intra-regional trade and incentive to continue to abuse the system when have been questioned by Yeats (1 989b). In line with some it has been recapitalized. Perhaps the power struc- studies of colonial relationships (cf. Austen (1987)), the ture of the CFA zone would be more diffuse if the two substantial trade with France has prompted some to sug- unions were to merge. The dominance of C6te gest that the existence of the zone provides extensive d'Ivoire would certainly be tempered, and the likeli- benefits to France. A recent study on these lines is Yeats hood of a four-way coalition of the prosperous coun- (1989a). tries working to the disadvantage of the rest would 9. There are some limitations which mean that capital seem somewhat remote. Each of Cameroon, Gabon, is not fully free to flow to France. First, transfers are increas- 158 Monetary Cooperation in the CFA Zone ingly being delated or even (in some countries) refused expansionof credit.This mightbethebestsolutionifitwere except where thebank can provide the funds abroad itself. to prevail only for a short while. Sustained use of such This is probably the most important limitation. Second, devices would serve to recreate a wedge between national banks are not permitted to hold more than working bal- credit markets and impede the market allocation of credit. ances outside their own country. In some cases this has 16. As indicated by the negative correlation between been stretched to imply that banks must always use to the country shares in monetary authorities claims on banks full all foreign lines of credit available to them. Third, and government and shares in union-wide exports. insurance companies are only allowed a very restricted 17. Mali is not shown; it joined the union in 1984. rangeofinvestmentpossibilitiesoutsidetheirowncountry. 18. In contrast, the union as a whole tends to repay Fourth, a transfer fee of 0.5% has been payable to the drawings from its overdraft facilityat the FrenchTreasury. BCEAO on all transfers (for current or capital payments) 19. Excluding the IMF-related items and the miscella- made to France. This hardly discourages flows of a long- neous items "comptes d'ordre et divers." term character, and is said to be bypassed by many banks 20. There is not enough published information to make wherever possible. Fifth, export receipts must be repatri- a precise calculation of each of the elements in the seignor- ated to the exporters' country. (This serves to block one age. Rough approximations based on 1987 balances and possible way of avoiding the transfer fee). Sixth, there have interest rates suggest that about one-third of the flow of been requirements to declare export of CFA currency notes; UMOA seignorage benefits for that year cam through the these however are largely ineffective as the currency notes interest subsidies. For the BEAC the proportion was more have been freely acceptable in Europe. like three-fifths. 10. The main reason for this has been the uncertain 21. The rough calculations based on 1987 suggest that solvency condition of several banks and the strained li- C6te d'lvoire could have taken almost 30% of the UMOA quidity position. In recent months evensome of thelargest seignorage flow and Senegal 16%. In each case the take was banks in C6te d'Ivoire have found themselves unable to not greater than these countries' share (50% and 16%) of meet their obligations on the due date, a development total currency in circulation; but it was greater than would which has encouraged several banks to withdraw alto- be implied by equal distribution (14%). Benin's estimated gether from interbank lending. take (12%) was the third largest, even though it had the 11. For example, during a very tight period for the lowest share of currency in circulation. Ivorian banks in January 1988, the marche monetaire inter- 22. Adjustingtheestimatesmentionedin previous foot- est rates were actually lowered in line with interest rate notes gives these three countries 46%, 26%, and 16% movements in Paris, despite the persistence of excess de- respectively. These calculations are very tentative as the mand for loans from the marche monetaire. scale of the losses is not yet fully determined. 12. There is a long-standing debate over the appropri- 23. The country-by-country assignment of interest ate relationship between CFA interest rates and those in charges on the BCEAO's operations account borrowing at Paris. As discussed in Honohan (1990) the policy-deter- the French treasury could also be relevant; but so long as mined CFA rates have tended to track Paris rates, as would these remain mostly as unpaid book entries between the be the normal market tendency considering the fixed ex- members and the BCEAO they need not have any impor- change rate and the relatively open capital market. tance and are not therefore discussed here. However, because of the residual risks perceivedby foreign 24. Some 70% of the profit distribution is divided on banks, capital will not always flow in plentifully in re- thatbasis. A further 15% is distributed inproportions to the sponseto a small interest differential. Higher interest rates currency issue and the remainder (15%) is distributed may sometimes be necessary to allow the market to choose equally among the members. In calculating the contribu- between competing uses for the limited funds available tion of each member country to profits, interest payments within the union. to the BEAC from a given country are regarded as part of 13. In view of the small degree of economic interaction that country's contribution to profits. Interest earnings between the member countries, there are no substantial from outside the union are allocated in proportion to each spillcver effects of credit creation in one country on de- member country's external reserves. A detailed critique of mand conditions in the others. the formula and its application is not attempted here, but 14. This process is documented in Honohan (1990). itisclearthatitdoesnotsuccessfullymimicthedistribution Other factors can also influence credit growth; the banks' of seignorage earnings which would prevail if each country borrowings from correspondents have been an important operated with an independent currency issue. A rough factor in the past. calculation suggests that the Central African Republic and 15. Alternatively, if on or more country credit ceilings Chad (bothof whichhavecomparatively smallborrowings were being exceeded, country-specific reserve require- and are among the poorest of the BEAC countries) are the ments could be imposed in those countries to choke off the most disadvantaged by the formula. 159 Monetary Cooperation in the CFA Zone 25. This provisions may actually have encouraged Guillaumont, P. And S. (1984), Zone franc et developpe- complacency on the part of the central bank in making ment africaine, (Paris, Economica). loans to unsound or doubtful banks, as it divorced respon- Guillaumont, P. And S., eds. (1988), Strategies de develop- sibility for losses from the decision to lend and meant that pement comparees: zone franc et hors zone franc, the central bank had no clear incentive to ensure that the (Paris, Economica). banks to which it lent were sound. As a result, the central Guillaumont, P., S. Guillaumont and P. Plane (1988), "Par- bank operatedbanksupervision as if it were a disinterested ticipating in African Monetary Unions: An Alternative party, merely reporting its findings to the national authori- Evaluation', World Development, Vol 16, No. 5, pp. ties and taking no steps directly to curb bad management 569-76. practices. On the other hand the provision it has certainly Honohan, P. (1983), "Measures of Exchange Rate Variabil- not been effective in discouraging governments from ar- ity for One Hundred Countries", Applied Economics, ranging for loans to made for political rather than sound Vol. 15, pp. 583-602. economic purposes. Honohan, P. (1990), "Price and Monetary Convergence in Currency Unions: The Franc and Rand Zones", World References Bank, Washington, D.C. Krumm, K (1985), "Adjustment policies in the Ivory Coast Austen, R. (1987), African Economic History: Internal De- in the framework of UMOA", World Bank CPD Discus- velopment and External Dependency, London, James sion paper 1985-6, Washington, D.C. Currie Publishers. Krumm, K L. (1987), "Adjustment in the franc zone: focus Bhatia, R.J. (1985), The West African Monetary Union: An on the real exchange rate", World Bank CPD Discussion Analytical Review, (Washington, D.C., International Paper 1987-7, Washington, D.C. Monetary Fund). Ouattara, A. (1986), The Balance of Payments Adjustment Devarajan, S. And J. de Melo (1987a), "Evaluating Partici- Process in Developing Countries: The Experience of pation in African Monetary Unions: A Statistical Analy- Ivory Coast, Report to G-24 (UNDP), UN, New York. sis of the CFA Zones", World Development, Vol 15, No. Yansane, K (1984), Controle de l'activite bancaire dans les 4, pp. 483-96. pays africains de la zone franc, Paris, Librarie Generale Devarajan, S. And J. de Melo (1987b), "Adjustment with a de droit a de Jurisprudence.. Fixed Exchange Rate: Cameroon, Cte d'lvoire and Sene- Yeats, A. (1989a), "Do African Countries Pay More For gal", The World Bank Economic Review, Vol.1 No.3, pp. Imports? - Yes", PPR Working Paper 270, World Bank, 447-87. Washington, D.C. Dittus, P. (1987), "Structural adjustment in the franc zone - Yeats, A. (1989b), "On the accuracy of economic observat- the case of Mali", World Bank CPD Discussion Paper ions: Do Sub-Saharan Trade Statistics Mean Anything?", 1987-11. PPR Working Paper 307, World Bank, Washington, D.C. 160 16 Comments on Financial Sector Policy FM. Mwega and Diery Seck FM. Mwega model to explain this. Yet this argument is to a large extent only valid for money lenders who constitute a My discussion focuses on the paper presented by small component of the informal financial sector. Earnest Aryeetey and Mukwanason Hyuha entitled The second issue analyzed in the paper is the size The informal Financial Sector atd Markets in Africa: An of the informal sector. In a number of country case st- Empirical Study and the paper by Machiko Nissanke udies, such as those of Ghana and Malawi, some evi- entitled Mobilizing Domestic Resources for African De- dence is provided that shows the informal sector to velopment and Diversification: Structural Impediments be very large. Indeed, the paper argues that the infor- to Financial Intermediation. mal financial sector is larger than the formal financial sector. In other country case studies however, this ev- The Aryeetey and Hyuha Paper idence is not convincing. In the case of Senegal and Zaire, the paper cites evidence that is interpreted to I found the paper by Aryeetey and Hyuha very mean that 60 to 85 percent of the population eam stimulating and think it makes an important contri- their livelihood in the informal financial sector. This bution to ourunderstandingof the informal financial does not seem plausible. In these two cases (Senegal markets in Africa. My presentation is therefore in the and Zaire), it is necessary to distinguish between the nature of raising issues for discussion and further infornal sector broadly defined and the informal fin- research. ancial sector. In general, there is need for more re- The paper addresses itself to a number of issues. search on the size of this sector relative to the formal The first issue is the anatomy of the informal financial financial sector. This is because of the amount of pol- markets in Africa. The paper shows that the informal icy attention the sector should receive is to a large ext- sector consists of a wide range of activities which ent, a function of its size. There is also need to study vary from country to country. While the savings and the internal dynamics of the sector, how it has evol- cred!it societiesareforexample dominant in theinfor- ved over time and what factors have influenced this. mal financial markets of Tanzania and Zaire, the so- A third issue that the paperanalyzes is thelinkages called 'susu' system is dominant in Ghana. In Mala- between the informal financial sector and the formal wi, estate owners, friends, relatives and employers financial system. The paper concludes that the two are dominant. It is therefore very difficult to derive a sectors are, to a large extent, complementary. In the general theory of how operators in the informal fi- rural areas, informal institutions fill the gap created nancial market behave, unless the sector is disaggre- by inadequate formal financial institutions, while in gated by the type of economic activity. It is in this urban areas the informal financial sector coexists light that the models developed in the paper should side-by-side with the formal financial sector. Further be seen. One part of the paper, for instance, argues issues however can be raised about this relationship. that interest rates in the informal sector are higher To what extent is it a dualistic relationship in which than in the formal financial sector and develops a a modem sector coexists with a traditional sector? 161 Comments on Financial Sector Policy Lastly, the paper tackles the policy implications of a major problem in mobilizing household savings, having a large informal financial sector. The basic especially in rural areas, is that a large component of argument by the paper is that since the sector is very this is in-kind, making it difficult to mobilize them large, it should be taken into account in policy form- through the financial system. At the same time, a ulation. Issues for further investigation then include large proportion of the financialized savings are mo- the following. What kind of policy interventions are bilized by the informal financial system, which as required? Does the formal financial sector reduce the stated above, may be larger than the formal system effectiveness of monetary policy? Does the sector in terms of the mobilized deposits and the credit create money in the same way as commercial banks? given. The sector may have a limited influence on theeffe- Similarly, the paper cites evidence that the formal ctiveness of monetary policybecause a number of the financial system does not seem to reach the small and operators such as money lenders. 'susu' collectors, medium sized firms so that they have mainly to util- etc. in the informal financial sector deposit their fu- izetheirown savings. Thesesavingsare however, not nds with formal financial institutions. Itisonlywhen adequate. Formal sector credit tends to go to para- they need to make payments that they withdraw the statals and to large firms while institutions that have money or write cheques. The activities of these op- been set up to provide credit to the small and me- erators can therefore be controlled for monetary pol- dium sized firms do not seem to be effective for vari- icy purposes through the regulation of formal finan- ous reasons. The paper tells a similar story about the cial institutions. Indeed, the paper states that in the difficulties of mobilizing public savings in African case of Senegal, a squeeze in the formal financial sect- countries using Ghana and Malawi as case studies. or leads to a squeeze in the informal financial market. The paper raises a number of issues. First, how do The paper argues that interest rates in the formal the various sources of savings interact with one an- financial sector do not influence those in the informal other? There is for example, a large literature that financial markets. If major operators in the informal shows that an increase in public savings that is financial sector deposit their funds with the formal achieved through a reduction in budget deficits may financial institutions, and sometimes borrow from not increase total savings much if this depresses the them to on-lend in the informal financial market, it is economy and hence reduces private savings through nor clear why the authors come to this conclusion. a Keynesian transmission mechanism. Second, what is the degree of market segmentation The Nissanke Paper of the formal financial system? The paper argues that there is a high degree of segmentation of the formal I also found the paper by Machiko Nissanke very financial system in which commercial banks keep useful. It is quite true that African countries have to excess reserves while the other institutions face a increase domestic savings if they are to reduce their shortage of funds to lend. But to what extent are dependence on net external capital inflows which in commercial banks' excess reserves really excessive? anycase havebeen going down over time. Thepaper In discussions with bankers while doing a study of discusses the potential of mobilizing household, en- monetary policy in Kenya, they said the reserves they terprise, and public savings and the shortcomings of keep are not excessive and are approximately what the formal financial system in allocating the mobi- they would keep if there were no minimum required lized savings efficiently to investment. The paper reserves ratio. argues persuasively on the need to take into account Lastly, in the kind of situation described in the the structure and the institutions of a country. paper, what would be the effectiveness and the ap- According to the paper, household savings comp- propriate sequencing in the introduction of financial rise a large proportion of total savings in Africa and reforms? Diery Seck The Aryeetey and Hyuha Paper recently had attracted little attention on the part of students of development economics-namely, the The paperby Aryeetey and Hyuha can be seen as informal financial sector in Sub-Saharan Africa. In part of a growing literature on a topic that until order to tackle this challenging issue, the authors set 162 Cvmmnts on Financial Sector Policy the following three goals in their paper. a) give a The findings of the paper can be summarized as review of the empirical evidence on the size, nature follows. Empirical research in several African coun- and workings of the informal financial sector in five tries has shown than more than half the financial different Sub-Saharan African countries, b) discuss transactions are done in the informal financial sector, briefly the literature aimed at explaining the growth the informal financial sector is in fact a collection of of informal financial markets in Third 'Wlorld coun- different financial schemes with different types of tries in general and in Africa in particular, and c) participants and markedly different interest rates propose a rationale that establishes a relationship prevail at the same time in the same geographical between the fonnal and irformal financial sectors, markets although quoted by different types of mar- based on, a micro-economic approach. ket participants. It is also found that the informal Research on this topic is timely and important for financial markets success is at least partly due to the a number of reasons, some of which being: failure of the formal market to meet the needs of * lack of success of most economic policies that segments of the population and that interest rates are have been implemented by African countries a weak channel if any for a possible linkbetween the over the last thirty years and that had the two markets. WVhether the two markets are comple- common characteristic of ignoring the infor- mentary or in competition with one another is still an mal sector or taking its compliance with stated open question. policies for granted; * failure of the banking system in most African The Size and Nature of Informal Financial Markets in countries to perform its asset transformation Sub-Saharan Africa function in a manner that is satisfactory even by the standards of policymakers and condu- Given the structure of the paper, the most appro- cive to capital formation and growth; priate way to discuss it is to examine its sections one * growing budget deficits and the realization at a time. The second section analyzes the informal that more than half the economic activity, be- financial sector in 5 countries. The definition of the ing informal, may be beyond the reach of eco- informal financial sector is a catch all statement that nomic policies, at least of the tax collector. is arbitrary and does not allow for an accurate meas- urement of the size of the market. Clearly, there is Without a doubt, understanding the informal sec- need for a definition of the informal financial sector tor, financial or otherwise, is quite an intellectual that is better suited for rigorous theoretical analysis challenge and may have become a prerequisite to the as well as tractable empirical work. formullation of viable economic policies for Sub-Sa- For instance, despite the marked differences in the haran Africa. Why? Because, as has been docu- activities of the sector in the uses of loaned funds, the mented by Aryeetey and HIyuha, it is the economic types of contractual arrangements and above all in sphere where more than half the people seek alterna- the prevailing interest rates, the authors make bla- tives to what is being offered by the formal sector of nket statements such as ".. the informal and formal the economy and often shield themselves from the sectors of fer a similar product, which is probably not effects of government policies and regulations. Any- entirely homogenous" (in the fourth section of the thing that we learn about informal markets in Africa paper), or ".. through a proper assessment of the is likely to teach us something about the nature and actual cost and demand structures of the two sec- extent of the imperfections that undermine the nor- tors, the informal sector could be made completely mal functioning of African markets and conse- (my emphasis) subject to various monetary and fis- quently point the direction for needed policy cal policies of government" (in the section on changes. In that sense the authors' effort should be Ghana). commended. The authors' definition also fails to give insights The paper should be seen as a continuing effort by into the classification of the various segments of the Mr. Aryeetey and others-see the paper's bibliog- informal financial sector. These segments canbe cate- raphy-at gathering data on and analyzing informal gorized according to different criteria namely, by financial markets in Africa with, the admitted desire type of contractual arrangement, by type of underly- to rationalize the body of knowledge on the topic. ing motives of the participants, by type of timing of This partly explains the structure of the paper where cash flows, etc. Proper analysis of the informal finan- the three major sections are loosely related and could cial sector must, as a prerequisite, acknowledge the have been developed into full papers in their own diversity of the activities that are performed under right if the authors had chosen to do so. the umbrella of the sector and perhaps be limited to 163 Comments on Financial Sector Policy the confines of each major segment of the market to informal financialsector, theyexplain, isbased on the be really useful. need for people who were left out of the formal sector While the literature on the informal financial sector to find alternative ways to satisfy their need for is still in its infancy, there is a crying need for better financial services. As a result, the authors implicitly techniques of measurement of the volume of trans- side with proponents of substitution in the comple- actions in the informal sector. Use of ratios such as mentarityvs substitution controversy. This approach Deposit money/M2 or Banking sector claims/GDP isbiased and onlyinvitescounter-examples. Further- only result in doubtful comparisons. Appropriate more, it ignores the fact that some segments of the measures should allow the direct assessment of the "modern" informal sector pre-date the era of the magnitude of each type of activity in the informal modernbankingfacilities, even if one readilyadmits financial sector and the technique that consists in that the informal sector may take advantage of the estimating the total amount saved by market women misfortunes of the banking sector. Some forms of through the "susu" scheme by way of interviews is a activity in the informal sector would persist even if step in the right direction.1 the formal sector performed ideally. In short, it takes One issue that begs examination is why lending more than historical accounts to settle the dispute rates differ so much from one segment of the infor- that is raging between the complementarity camp mal sector to another. Does the interest rate differen- and the substitution camp. tial imply market segmentation within the informal In their discussion of the possible relation between sector and if so, how to account for the persistence of the formal and informal sectors, the authors exclude such segmentation? The authors mention that mon- a link via interest rates, based on regression results eylenders in Ghana and co-operatives societies in obtainedbyAryeeteyandGockel (1989). Theirposit- Tanzania borrow from the bank to lend to the infor- ion is hasty and inappropriate because the empirical mal sector. Research into how widespread the phe- evidenceprovidedbytheregressionestimatescanbe nomenon is and its potential for expanding the reach challenged on several grounds. First, although the of the formal sector to hitherto unexploited segments authors acknowledge the presence of collinearity of the population might be fruitful. among the explanatory variables, its possible effect on the sign and significance of the regression coeffic- The Informal Sector: Growth and Links with the Fornmal ients is unaccounted for. Second, a single cross sect- Sector ion regression in a single country is clearly not en- ough to reject the link via interest rates of the formal In the third section of their paper Aryeetey and and infornmal sectors. Consequently, the issue is still Hyuha address the issue of duality in African finan- unresolved and calls for further empirical evidence. cial markets and that of complementarity or substi- The third section of the paper has left unanswered tution between the informal and formal sectors. On several key questions in the understanding of the the duality issue, they reject the view that socio-cul- duality of African financial markets, and more im- tural attachments of participants to the informal sec- portantly, the internal mechanisms of the informal tor help explain the durability of the informal mar- sector. For example, what is the explanation for the ket. Their reason is that "modem banking facilities lack of unified interest rates, given that customers cannot be found throughout the economy in most have access to various segments of the informal mar- African countries". ket? What are the competitive pressures within the I fail to see the logic of their argument. Moreover, informal market and why do they fail to establish Aryeetey and Gockel (1989) give empirical evidence parity of the interest rates quoted by the different that shows that for market women in Ghana, the segments of the informal sector? Are some segments proximity of a banking facility does not constitute an of the informal financial sector more closely related incentive to use the services of the formal banking to the formal sector and what are the prospects for sector, in fact proximity is found to be a disincentive! some of the segments of the informal sector being While the cultural attachment factor cannot be unifiedwiththeformalsectorwithrespecttointerest proved right or wrong based on available data, the rates, especially in urban areas? counter-argument opposed by Aryeetey and Hyuha is certainly not valid. Relative Pricing and Cost Structures of the Two Sectors To account for the duality of the financial sector, the authors use the historical evidence on the inabil- In the fourth section of the paper, Aryeetey and ity of the formal financial sector to satisfy the needs Hyuha useamicro-economicapproach to explain the of portions of the population. The success of the higher pricing in the informal financial sector. The 164 Comment.s on Finandal Sector Policy rationale that they propose is interesting and pro- The Honohan Paper mising. However further clarification is needed with respect to the market structures of the informal and Mr. Honohan claims that his paper is designed as formal sectors. Their claim that "one could talk of an a guide to understanding how the CFA Zone works oligopolistic relationship with some minor product as a mechanism for monetary cooperation. Taking a differentiation should be backed by a stronger theo- cost-benefit approach, he lists the potential gains that retical demonstration. As for their admission that can accrue to small countries when they are part of a there are monopolistic tendencies and near-perfectly monetary union, namely: competitive practices in the informal financial sector, e credible commitment to exchange rate and a statement that is not taken into account in the rest price stability, of the paper, it clearly indicates the urgency for more * pooling of reserves basic research directed at discovering the informal * larger capital market with free movement of financial market. funds within the zone and The authors mention that theoperatingcosts of the * a more effective and independent monetary individual actors in the informal sector are lower policy by a unified central bank. than those of the formal sector. I do not see why simple aggregation of the participants in both sect- The paper's intent is to determine if the last two ors would result in closer operating costs for the two potential gains are verified in the CFA Zone. sectors given that actors in both markets operate on The conclusion of the paper can be summarized as an individual basis. Moreover, the authors state th- an indictment of the way in which the BCEAO ma- at the informal sector does not gain from economies nages the money market in the Zone. In the author's of scale which explains the convergence of the mar- opinion, despite an effective policy of fixed exchan- ginal costs of the two sectors as the activity expands. ge rate and exchange control and an elaborate set of Such a view can be challenged because no in-depth rules to avoid excessive credit expansion, the CFA analysis of the cost functions of the various se- Zone almost collapsed because of widespread bank gments of the informal sector has been undertaken insolvency. According to his rationale, the poorer co- yet. untries of the union will pay more than their share of The authors' claim that there is a tendency for the the losses sustained as a result of the failure of the marginal cost curves of the two sectors to converge Zone's banking system. He gives two reasons for needs stronger substantiation. As for the authors' that: pessimistic note with respect to the likelihood of the * According to the statutes of the UMOA, losses two markets coming closer together because of bar- of the banking sector should be covered by the riers of entry into the formal sector, one can onlyhope respective home govemments of the banks. that the recent developments in the financial markets Yet, the BCEAO has decided to cover all the and monetary policies of many African countries will losses. Consequently, all the member coun- irreversibly turn barriers of entry into a thing of the tries of the monetary union will equally share past. the burden of those losses; Aryeetey and Hyuha use the concepts of specific The countries where the banking sector suf- and non-specific loans to analyze the issue of comp- fered most losses happen to be the three richer etitive asset pricing in the two financial sectors. Wh- ones, i.e. C6te d'Ivoire, Senegal and Benin and ile these concepts are helpful in understanding the coverage of losses of the banks by the central pricing mechanism suggested in the paper, the auth- bankamounts to a skewed distribution of seig- ors have not used them to develop a full-fledged the- norage profiting those richer countries alone. oretical asset pricing model. An equally valid app- roach is to view the formal financial sector as segme- My discussion of the paper is organized as follows. nted and the informal financial sector as integrated. First, a few general comments will underline my Equilibrium asset pricing models for segmented and appreciation of Mr. Honohan's study. Second, the integrated financial markets are proposed by Eun points about which the views expressed in the paper and Janakiramanan (1986), Subrahmanyam diverge from mine are discussed. Third, I will at- (1975a,b) and Stapleton and Subrahmayam (1977). tempt a brief extension of the issues under discuss- The authors may wish to make use of this literature ion in the paper. A few suggestions for further re- in their future research on asset pricing in Sub-Sa- search will serve as the concluding note to this dis- harain Africa's informal and formal financial sectors. cussion. 165 Comments on Financial Sector Policy General Comments banks to operate with low volumes of deposits and consequently to borrow substantially from the cen- The second section of the paper constitutes an tral bank. Central governments of the UMOA borrow excellent and well documented presentation of the from the central bank to finance their growing CFA Zone and of its member countries. The similari- budget deficits. ties and differences of those countries among them Furthermore, recent trends show that the central on the one hand and with non-member countries of governments have engaged in a policy of crowding Sub-Saharan African on the other hand are expertly out that has been debilitating for the private sector. illustrated. While it is not a part of the main discus- As is shown in the paper, the share of the central sionof thearticle,Mr.Honohan'sunderscoringof the bank's claims on the economy going to the central potential for export revenue diversification-be- government has increased dramatically between cause of low correlations between countries' export 1979 and 1988 for all the countries except Benin and revenueindexes-through riskpoolingisan interest- Mali. The data series are not complete for those two ing point. It might be an avenue for fruitful future countries. In addition to this direct form of crowding research. The author gives ample statistical evidence out, one must keep in mind that a considerable to support his views. amount of the central bank's credit extended to com- In a way, Mr. Honohan describes how, in the case mercial banks is loaned by the commercial banks to of the UMOA, the attempt by a central bank to play state firms. Overall, central bank's funding of the a major role in financial intermediation has failed private sector through commercial banks is dwarfed miserably. The main reasons for the failure are: by its direct and indirect funding of the public sector. * credit ceilings to governments and to the pri- The author's focus on the shortcomings of the vate sector, money market-he calls it by its French name: mar- ) inefficient domestic inter-bank money market che monetaire-has led him to advocate a major role • little or no inter-country capital flow within for the central bank in that money market. Such a the Zone; view can be challenged on several grounds. While * passive supervision of commercial banks; the inter-bank money market is needed to recycle * failure by the central bank to use market clear- excess bank liquidity, what is more important is the ing interest rates in the money market and in emergence of an organized capital market. Failure to re-discounting. mention it may lead to the false assumption that an adequate capital allocation mechanism can be in- In a word, the central bank has failed to achieve sured by the commercial banks or the central bank. allocative efficiency-best loan prospects get fund- A separate and well-functioning organized capital ing at lowest rate-and operational efficiency-nar- market that is autonomous from monetary authori- row spreads in inter-bank money market. As a corol- ties is needed because: lary to the author's findings, the monetary union * there is a situation of chronic crowding out; could possibly break down, not because it will no * the banking sector is in crisis and may fail to longer be able to manage the monetary affairs of the accomplish its asset transformation function member countries, but because the poorer countries properly; may justifiably feel that while the gains of the mone- * there is a need for a long term substitute for tary union are equally shared, they pay a dispropor- external funding that is part of the countries' tionate share of the cost of the functioning of the respective structural adjustment programs, monetary union. and; * there is a need to restore savings and invest- A Few Notes of Dissent ment to their pre-crisis levels. The author claims that, with the exception of Togo There are other reasons for not entrusting the cen- and Burkina Faso, the commercial banks of the tral bank with the organization of the money market. UMOA are heavily indebted to the central bank. I The first one is that, given the bureaucratic tradition have used the same data and found that, given the of the BCEAO, its ability to run side by side a "mone- role played by the central bank in capital allocation tary policy arm" and an "intermediation arm" that within the Zone and its consistent policy of negative are independent from one another, as is suggested by real rates of interest, it is not the commercial banks Mr. Honohan, can be questioned. The second reason but the central governments that are heavily in- lies in the contradictory situationwhereby thecentral debted. Negative real deposit rates force commercial bank would run a money market within the Zone, 166 Comments on Financial Sector Policy Table 16-1. The respective shares of the centralgovernment and commercial banks in the claims of the central bank Benin Burhina Faso Government Banks Government Banks 1979 n.a n.a. 8.9 91.1 1980 n.a. n.a. 11.7 88.3 1981 n.a. n.a. 35.3 64.7 1982 n.a. n.a. 41.4 58.6 1983 n.a. n.a. 31.5 68.5 1984 n.a. n.a. 48.6 51.4 1985 n.a. n.a. 48.8 51.2 1986 21.8 78.2 56.1 43.9 1987 21.5 78.5 60.8 39.2 1988 16.5 83.5 74.8 25.2 C6te d'lvoire Mali Government Banks Government Bank 1979 7.1 92.9 n.a. n.a. 1980 24.9 75.1 n.a. n.a. 1981. 30.8 69.2 n.a. n.a. 1982 28.1 71.9 n.a. n.a. 1983 37.4 62.6 n.a. n.a. 1984 42.0 58.0 87.8 12.2 1985 43.5 56.5 82.4 17.6 1986 42.0 58.0 70.7 29.3 1987 34.0 66.0 73.6 26.4 1988 35.8 64.2 74.8 25.2 Niger Senegal Government Banks Government Banks 1979 -85.7 185.7 16.0 84.0 1980 1.5 98.5 18.1 81.9 1981 -24.3 124,3 26.7 73.3 1982 23.0 77.0 36.2 63.8 1983 19.5 80.5 40.1 59.9 1984 40.8 59.2 45.5 54.5 1,985 52.4 47.6 45.8 54.2 1986 42.4 57.6 49. 50.3 1987 25.1 74.9 50.7 49.3 1988 35.8 64.2 45.6 54.4 7Tgo Governmnent Banks 1979 55.8 44.2 1980 47.9 52.1 1981 74.5 25.0 1982 75.1 24.9 1983 79.4 20.5 1'984 77.7 22.3 1985 88.3 11.7 1986 87.9 12.1 1987 89.1 10.9 1988 91.5 8.5 Source: BCEAO 167 Comments on Financial Sector Policy using market-determined rates and still enforce cre- in tablel6-1, except for Benin and Mali, which have dit rationing schemes such as assigning credit ceil- incomplete data series, the historical evidence of the ings to each bank in the banking system. While a mo- period from 1979 to 1988 points to a trend of increas- ney market operating under market clearing condi- ing crowding out for all the countries of the UMOA. tions fosters competition, credit ceilings undermine In absolute terms, for Cote d'Ivoire the net credit itbecausetheytendtomaintainbanks'marketshares granted by the central bank to the central govern- fixed, irrespective of their relative performances. ment has been multiplied by a factor of 18.5 while Mr. Honohan dispels the possibility for private that of the commercial banks is 2.52 The correspond- intermediaries to organize the money market be- ing numbers are for Senegal: 10.8 and 2.5, for Togo: cause of the fragile state of the banking system and 3.3 and 0.4, for Burkina Faso: 11.7 and 0.4 and the of the unacceptable prevailing levels of risk sur- government of Niger changed from being a net roundingthebankingsector.Itcanbearguedthatthe lender to the central bank in 1979-6.3 billion perceived risk level is a direct consequence of the francs-to being a net borrower in 1988-16.2 bil- mannerinwhichthe centralbankhasbeenoperating lion-while the claims on the commercial banks the money market, namely a deliberate policy of grew by a factor of 2.1. geographical segmentation across member coun- Clearly, during the difficult years of 1982 to 1988, tries, lack of active supervision of commercial banks, for most countries of the UMOA, the central bank and failure to use market determined lending terms chose to fund governments' budget deficits rather in the money market. than the private sector with adverse consequences on the liquidity of the bankingsector and on the level of Extension to Mr. Honohan's Study real economic activity. The national private sectors that were penalized the most are those of Togo, Burk- Mr. Honohan has shown that the CFA Zone has ina Faso and Niger, those very countries that, Mr. failed to provide member countries with an efficient Honohan shows, have paid a disproportionate share and well organized money market with free move- of the losses incurred by the union's banking system. ment of funds across countries of the Zone. He has To what extent the central bank has played a role in also demonstrated that, while they use the credit the current crisis of the banking sector is a research facilities of the central bank the least, the poorer topic worth pursuing. countries of the union pay a disproportionate share Negative real deposit rates of interest have comp- of the losses caused by the massive failure of the lex implications because they determine the extent to banking system. A natural extension of the author's which commercialbanks seek funding from the cent- analysisisto examine thegainsthat the private sector ral bank, which limits their operational autonomy. In of the monetary union derived from monetary coop- other words, negative deposit interest rates give the eration within the Zone. central bank non statutory powers over the banks by In that respect, two issues pertaining to the effect way of conditionality on re-discounting operations of monetarypolicyof theprivate sectorwillbebriefly e.g. credit ceilings, mandatory loan portfolio compo- discussed: the extent and impact of crowding out and sition for banks etc.. They also allow the central bank of negative real deposit rates of interest on private fir- to have a direct control on the level of growth of the ms. How well does the private sector fare in its coop- productive sector since it can vary over time the eration with the public sector with respect to aggreg- minimum quality of the paper that it can accept for ate credit allocation by the central bank? As is shown re-discounting. Table 16-2. Average real deposit rates of interest and differential between deposit rates of B CEAO and France Period Dif' Burkina Faso C6te d' Iwire Niger Senegal Togo 1970-88 -0.32 -1.19 -3.45 -1.71 -3.02 -1.96 1985-88 0.35 4.48 1.83 9.01 2.63 5.38 1980-84 -0.87 -2.01 -1.21 -3.15 -4.09 -2.81 1975-79 -0.59 -6.74 -10.12 -8.68 -5.46 -6.02 1970-74 -0.04 0.66 -3.24 -1.86 -4.04 -2.92 Difference between the nominal deposit rate of interest of the BCEAO and France. Source: International Financial Statistics, IMF 168 Comments on Financial Sector Policy Has the central bank sought to implement policies growth whether the money market is organized by of low interest rates? Table 16-2 provides part of the the central bank orby private intermediaries? answer. The column labelled DIFF displays the dif- The UMOA has recently created a supranational ference between the nominal deposit rate of interest supervising committee that will monitor the integr- of the BCEAO and that of France. The historical ity and quality of the management of banking insti- differences are not in UMOA's favor except for the tutions in the zone. What will be its level of effective- last four years. The real deposit rates follow the same ness compared to an insurance deposit scheme wh- historical pattern in that they have been negative on ereby insured banks pay premia that are commensu- average through 1984. Theybecome positive starting rate with the quality of their respective portfolios as 1985 because member countries experienced low or opposed to their levels of deposits or loans? negative rates of inflation. Capital flight or use of the informal financial sector Notes are likely responses to consistently low deposit inter- est rates. For the productive sector, what seemed to 1. See Aryeetey, E. and Gockel, F., MobilizingDomestic be a good deal as far as bornowing rates are con- Resources for Capital Forrmation: The Role of Informal cerned turned out to be a bad deal with respect to the Financial Markets in Ghana, Paper presented at the Work- aggregateamountofcreditsupply.Thwartedgrowth shop of the African Economic Research Consortium and a higher incidence of bankruptcies may have (AERC), Harare, 4-8 December 1989. been the inevitable consequence for firms that were 2. Credit granted to other financial institutions is not in need of new funding. It is noteworthy that the included but is negligible. countries that have had the lowest real deposit rates of interest are the richer and economically bigger References countries of the monetary union, which gives an indication of the extent to which the central bank had Aryeetey, E. and Gockel, F., 1989, Mobilizing Domestic Re- to step in and fund the private sector. sourcesfor Capital Formation: TheRole of Informal Financial Markets in Ghana, Paper presented at the Workshop of the A Few Concluding Suggestions African Economic Research Consortium (AERC), Harare, 4-8 December. Mr. Honohan's paper has been very illuminating Eun, C. and Janakiraman, S., 1986, "A Model of Interna- in terms of how the benefits and costs of being part tional Asset Pricing With a Constraint on the Foreign of a monetary union can be inequitably distributed Equity Ownership", Journal of Finance, 41, 897-914. among the participants of the union. Several ques- Stapleton, R. C. and Subrahmanyam, M., 1977, "Market tions still remain unanswered and could be the focus Imperfections, Capital Market Equilibrium and Corpo- of future research. In that regard three issues worth ration Finance", Journal of Finance, 32, 307-319. investigating are as follows. Subrahmanyam, M. 1975a, "On the Optimality of Interna- What is the impact of crowding out on the private tional Capital Market Integration", Journal of Financial sector of a monetary union such as the UMOA and Economics, 2, 3-28. are there ways to mitigate it if the crowding out is to Subrahmanyam, M. 1975b, "Intemational Capital Market persist? Equilibrium and InvestorWelfare With Unequal Interest What is the potential contribution of using market Rates", in E. J. Elton and M. J. Gruber, eds.: International determined interest rates in capital formation and Capital Markets (North Holland, Amsterdam). 169 17 Commodity Exports and Real Income in Africa: a Preliminary Analysis Arvind Panagariya and Maurice Schiff This paper is part of a project whose principal objec- why the world demand is unlikely to be a binding tive is to address the frequent concern that a constraint in the future. simultaneous expansion of exports by several devel- The general case for export pessimism has lost oping countries is likely to lead to a decline in their much of its force, at least for now. Concems have terms of trade, export revenues and real income. This remained very much alive, however, with respect to concern dominated the writings of development some specificcountriesand commodities.Thus, fears economists during the 1950s and has kept resurfac- continue to be expressed that a simultaneous expan- ing inonecontextoranothereversince.Theessential sion of exports of certain commodities (e.g., cocoa, argument as expounded in the early writings-e.g., coffee and tea) by several African countries, most Nurkse-was that exports of developing countries recently resulting from the adoption of structural consisted mainly of primary products for which adjustment programs, may lead to a decline in the world demand was inelastic. Therefore, any produc- real incomes and exports revenues of the countries. tivity gains in exportables were likely to be passed on These fears have been sufficiently serious that since to importing countries via a change in the terms of 1968 the World Bank has had special lending guide- trade favorable to the latter. There was little to be lines for commodities such as coffee, cocoa and tea. gained by relying on exports as the engine of growth. Bank policy is to deny lending for output expansion A number of economists, including Krueger of these commodities, unless the country has no vi- (1961), Cairncross (1962), Keesing (1967), Balassa able economic alternative or if the country has suf- (1978) and Bhagwati (1978, 1988), refuted the wis- fered a recent loss in production due to climatic or dom of this export pessimism. In particular, the re- other reasons. cent paper by Bhagwati (1988) provides a compre- Concerns regarding possible harmful effects of hensive review of the controversy and makes a con- commodity export expansion have reemerged re- vincing case that the fears of elasticity pessimists, old cently in the context of structural adjustment both as well as new, are ill founded. Among other things, within and outside the Bank. For instance, such con- he points to the phenomenal growth of exports and cerns were raised recently in the German parliament incomes of many East Asian countries to counter which, in turn, prompted the following query to the export pessimists. He also argues that the dramatic Chief Economist of the World Bank for Africa: shift in the export composition of developing coun- "In the framework of structural adjustment pro- tries toward manufactures and the potential for in- grams, manv African countries endeavor to increase traindustry specialization provide further reasons their exports of agricultural products. An increased 170 Co%odiygy Epcvts and Rcal Tncome in Africa: a Preliminary Analysis supply of goods may soon lead to price declines of ever, is whether the current trade taxes adequately the,correspondent products, so tha; additional reve- handle the problems which arise from low demand nues may not be realized. How does the World Bank elasticities in the world market. justify its correspondent policy advice? What can be There are two analytically distinct sources of terms done to avoid the negative results?" of trade deterioration: exogenous deterioration due A similar sentiment has been expressed by Profes- to low income elasticity of demand and endogenous sor H.W. Singer in an exchange with Mr. C. Hum- deterioration resulting from increased productivity. phreys in the December 1989 issue of Finance and In the latter case, increased productivity may be ac- Development. Professor Singer asserts that the "ad- companied by a decline in income and export reve- verse changes in terms of trade, far from being 'ex- nues if the price elasticity of demand is low. Our ogenous', are related to the export expansion. . . concern here is solely with the problems which arise recommended simultaneously to indebted coun- from a low price elasticity of demand in the world tries". In response, Mr. Humphreys writes that the market and their implications for trade policy. Prob- argument of the negative effect on the terms of trade lems which arise from low income elasticity do not of export expansion is weak, and suggests Africa require policy intervention and will not concern us should raise productivity in its traditional exports, here.2 diversify its export base, and expand production of Given a low price elasticity, problems can arise at imported commodities. Perhaps there is some truth both the national and international level. Thus, ex- in both views but without systematic analysis we port expansion by one country affects not only its camnot reach a consensus. own income but that of the other countries as well. 'The recent World Bank Board paper "Strengthen- In particular, if export expansion is the result of in- ing Trade Policy Reform' also emphasizes the need creased productivity, the country expandingexports for studying the effects of increased commodity ex- is likely to gain while the other exporters will lose. ports on the tenns of trade and real incomes in Sub- The project will study the effects at both levels. Saharan Africa. Similarly, the study on Africa by In this vein, we would like to seek answers to the Landell-Mills, Agarwala and Please, "Sub-Saharan following questions in this project: Africa: From Crisis to Sustainable Growth" notes that What is the likelihood that export expansion some of the poorest African countries have been hit resulting from better and fuller use of re- hardest by adverse tenrs of trade changes over the sources can lead to a decline in real incomes past 30 years. and export revenues in African and non-Afri- These examples show that the fears raised by elas- can countries? For which commodities is this ticity pessimists are very mruch alive in the context of outcome plausible? What are the key parame- African commodity exports. Tne problem is viewed ters determining the impact of export expan- as being especially serious for countries whose ex- sion on the terms of trade, export earnings and ports are concentrated in commodities that are said real incomes? to exhibit low import demand elasticities (e.g., Cote Will further reductions in export taxes lead to d'lvoire which exports coffee and cocoa). Rhetoric an increase or decrease in real incomes? What has been strong on the part of proponents as well as will be the impact of tax reductions on tax opponents of pessimism. But evidence provided to revenues and output quantities? date on either side is sketchy. We believe that there is Empirically, how important is the issue of in- a real need to study the issue in depth and under- terdependence? For instance, while consider- stand whether the export pessimists are justified and ing a further tax reduction on cocoa, should if so what can be done to maximize these countries' Ghana pay attention to policy changes in C6te gains from exports. The present paper is a first effort d'Ivoire and Malaysia? In which commodities, in this direction. if any, does interdependence play an impor- A central premise behind the proiect is that domes- tant role? tic-policy instruments should be employed to pro- * Can countries realize most of the gains from mote efficiency at home while trade policy instru- trade by choosing taxes optimally in an inde- ments should be used to deal with problems related pendent fashion? to foreign demand.2 Thus, the policy measures de- signed to correct domestic price distortions and real In terms of development of theory, we should note wage and exchange rate misalignment are well ad- that the issue of income- and revenue-maximizing vised. The issue which deserves closer scrutiny, how- exports taxes when two or more countries compete 171 Commodity Exports and Real Income in Africa: a Preliminary Analysis against each other in the world market for the same The Model commodity has simply not been studied. The tradi- tional literature deals with the situation in which two Our model consists of a multi-country demand- countries choose taxes on each other's exports so as supply system for the commodity examined. The to maximize income (welfare) or revenue.3 effects of adjustment programs are captured by pa- rameters of the supply function. For instance, im- Relevant Comrnodities and Countries provements in productive efficiency via fuller and better use of resources are represented by shifts in the Commodities that concern us must satisfy two im- intercept and/or slope parameters.4 portant criteria. First, they must account for a signifi- Within a partial equilibrium framework, the prin- cant share of exports of one or more African coun- cipal theoretical issue is the determination of income tries. Second, African countries must have a large sh- -and revenue- maximizing export taxes when two are in the world market for those commodities. The or more countries compete against each other in the six most important commodities based on these cri- world market. Although there is a large body of teria are cocoa, coffee, tea, cotton, tobacco and grou- literature on optimal trade taxes, the issue of how ndnuts. Of these, the first three-cocoa, coffee and such taxes are determined when two or more coun- tea-are exported exclusively by developing coun- tries compete against one another has simply not tries while the last three are exported by both devel- been addressed. oping and developed countries. Developing coun- In this set-up strategic considerations inevitably tries do not import cocoa at all and account for only come into play. Fortunately, there have been impor- 7.9 percent of net imports of coffee. The remaining tant developments in the area of strategic trade pol- commodities are imported by developing countries icy in recent years which allow us to address the in large volume both in absolute and relative terms. problem in a reasonable way.5 Africa's share in the world market for cocoa is Theanalytic solution forincome-and revenue- larger than that for any of the other commodities. maximizing export taxes for several countries who Cote d'Ivoire and Ghana each has a large share in the compete in the world market is provided in Pana- world market and depends heavily on this commod- gariya and Schiff (1990). Below, we present a brief ity for export receipts. The two countries also raise diagrammatic exposition of the basic model and its substantial tax revenues from this commodity. Coffee application to cocoa. In the following section, we is the next most important commodity for Africa. present the simulation results with actual and in- Once again, Cote d'Ivoire is the largest African ex- come-maximizing taxes. Simulation exercises with porter. In terms of export share for Africa, tea was not revenue-maximizing taxes will be taken up at a far behind coffee in 1986, with Kenya being the larg- later stage. est African exporter. The basic structure of the problem can be ex- Our focus in this paper is on cocoa. Cocoa is not plainedconvenientlywiththehelpofathree-country only the most important commodity export from setup. Denote the three countries by A, B and C and Africa, but it also provides a clean case in that it is assume that the former two export cocoa to the latter. exported exclusively by developing countries and Assume that exporters do not consume and the im- imported exclusively by developed countries. Tea porter does not produce cocoa. Individual producers and coffee, on the other hand, are also imported by and consumers are perfectly competitive. Each ex- some developing countries and raise complex dis- porting country's government chooses the export tax tributional issues within developingcountries. These so as to maximize the country's profits taking the commodities will be examined at a later stage. other exporting country's tax rate as given. This be- Cocoa is exported in large amounts by non-African havioral assumption leads to what is called the Nash countries as well. We will incorporate explicitly these equilibrium in the game theory literature.6 non-African countries into the analysis. This will Consider country A's problem. For a given export allow us to examine the impact on African countries tax by B, say tB, A's government must choose tA so as of changes occurring outside Africa, and vice versa. to maximize profits from exports. We can obtain the The remainderof thepaperis organized as follows. excess demand facing A by subtracting B's supply In Section II, we outline the model and its application from the world (i.e., country C's) demand. This ex- to cocoa. In Section III, we present results of a number cess demand curve is represented by DA (t8) in figure of simulations. Finally, in Section IV, we conclude the 17-1. Coresponding to DA, we can draw a marginal paper. revenue curve, MRA. 172 Commodity Exports and Real Income in Africa: a Preliminary Analysis We assume that the supply curve, SA, reflects the curve RA, will be t'A. We refer to t1A as A's "myopic" true marginal social costs of producing cocoa in A. optimal tax to emphasize that t'A is optimal only if A Then the country's income-maximizing equilibrium expects B to continue to hold its tax at tIB. As RB will coincide with the profit-maximizing equilib- shows, when A's tax is at tVA, B's income - maximiz- rium. This equilibrium is given by the intersection of ing tax is lower than t1B so that the latter is unlikely SA with the MRA curve. The corresponding world to hold its tax rate at t'B. and domestic prices are given by pw and PA and the For any given value of tA, we can also define a difference between them, pw - PA = tAPW, equals the myopic optimal tax rate for country B. Thus, if tA = per-unit export tax. Tax revenue and producers' sur- t2A, B's myopic optimal tax rate is t2B. As in the plus, respectively, are given by PWNMPA and PAME. previous case, if B fixes its tax at t2s, A will want to In this paper, we will refer to the sum of these areas choose a lower tax than t2A. It is evident that only at as the country's profits from cocoa exports. N is A's tax rate optimal given B's tax rate and vice An increase in ts shifts B's supply curve (not sho- versa. In this sense, N is the Nash equilibrium. wn in figure 17-1) to the left and hence A's excess A final point which deserves noting is that it is demand curve to the right. This causes the optimal tA entirely possible that the profits of both A and B can to change. It is easily shown that as tB rises, the opt- be higher at arbitrarily chosen tax rates (t2A, t1B) than imal tA also rises. Moreover, A's profits associated wi- at Nash equilibrium. The problem with these arbi- th the higher (tA, tB) combination are higher as well. trary rates is, of course, that at least one country can In figure 17-2, curve RA shows countryA's optimal increase its income by changing the tax rate. There- tax rate for different values of tB. In Panagariya and fore, the arbitrary rates are not sustainable under the Schiff (1990), we demonstrate that if demand and Nash assumption. The lower profits at N than at supply curves are linear, the shape of RA must be as some higherrates is simplytheresultof "tax-compe- shown in figure 17-2. Analogously to RA, we can tition" between the two countries. derive RB which shows the optimal values of tB for In our simulations, we extend the above model to different values of tA. We will refer to RA and RB as allow for nine exporters. These exporters are Came- reaction curves and to the point of their intersection, roon, C6te d'Ivoire, Ghana and Nigeria in Africa, N, as the Nash equilibrium. Brazil and Ecuador in Latin America, Indonesia and Suppose that B's tax happens to be t1B. The corre- Malaysia in Asia, and Oceania. Markets are assumed sponding optimal tax for A, as shown by its reaction to be competitive in each country, and domestic con- Figure 17-1. Equilibrium Price Determination Figure 17-2. Optimal Tax for Country A Given Taxes in Country B Pnice tB (1- tA) PW-PA - M;/C \\ s tB~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ | MRA DA(t B ) / B o 0 N 1tA 7 tA tA tA 173 Commodity Exports and Real Income in Africa: a Preliminary Analysis sumption is assumed to be zero. We derive both the should be understood as generating comparisons myopic optimal export taxes and Nash export taxes between alternative long-run equilibria after all rele- for each country. vant planting changes have taken place. We also employ the model to perform several com- Moreover, since we are concerned with the long- parative statics exercises. We simulate the effects of run we abstract from the demand for stocks and improvements in production efficiency in African focus on consumption demand. We assume for sim- and non-African countries on real income and tax plicity that the residual output exported by the small revenues in the nine countries. The simulations are producers-those not included in our group of nine donebothwhentax rates are heldconstantand when major producers-is given and does not respond to they are altered endogenously (Nash taxes). changes in the world price. We then set the demand The world demand and country supply curves at the world price equal to the exports of the nine have been linearized by using existing elasticity esti- major producers. Finally, we abstract at this stage mates and 1986 prices and quantities. Knowledge of from the possibility of smuggling, say between the elasticities is not sufficient to solve for the tax Ghana and Cote d'Ivoire. This added constraint on rates. The exact form of the various functions also the power of some countries to set export taxes will matters. For instance, several studies have estimated be taken up at a later stage. demand functions in log form, and have obtained Table 17-1 presents some of the basic data used in elasticity estimates smaller than one. Clearly, the the analysis. The output levels (in thousands of met- elasticity cannot be smaller than one along the entire ric tons)and the shares correspond to 1986, with the demand curve as this would imply an infinitely high C6te d'Ivoire at 585 (36%), Brazil at 329 (20%), Ghana export tax at the world level. at 219 (13%), Malaysia at 125, (8%), Cameroon at 118 The functional form is a matter which we plan to (7%), Nigeria at 110 (7%), Ecuador at 85 (5%), Indo- take up in the future. At this stage, we have assumed nesia at 32 (2%), and Oceania at 30 (2%). Table 1 also linear curves, which in the case of the demand func- shows the 1986 world price, tax rate8, and domestic tion implies finite export taxes. Our results must thus prices, the long-run supply elasticities and the corre- be regarded as being preliminary. More definite re- sponding slopes and intercepts of the long-run sup- sults must await a more careful examination of the ply curves. precise form of the demand and supply functions. The highest export tax rate (70%) corresponded to Trade policy interventions should be designed not Ghana, followed by Nigeria (50%), Cameroon (40%), to serve short-term stabilization objectives, but Cote d'Ivoire (25%), Brazil (20%), and a zero tax rate rather should be implemented in order to improve for Malaysia, Ecuador, Indonesia and Oceania. The the long-term efficiency of resource allocation in more recent producers, Malaysia and Indonesia, those specific cases where the market fails to do so.7 have a high long-run supply elasticity (3), while Hence, long-term elasticity estimates were used as a those of traditional producers-like Ghana (0.71), basis for linearizing the supply functions. Thus, the Cote d'Ivoire (1.15), Nigeria (0.45) and Brazil (0.58) - simulations performed in the following section are significantly lower. Table 17-1. Basic data Output Output share Export tax' Domestic price Intercept (OOOMT) (%) (%) (US$/MT) Elasticity Slope' (OOOMT) C6te d'Ivoire 585 35.8 25.1 1550 1.15 0.434 -87.7 Ghana 219 13.4 70.0 621 0.71 0.250 63.5 Cameroon 118 7.2 40.0 1242 1.81 0.172 -95.6 Nigeria 110 6.7 50.0 1035 0.45 0.048 60.5 Malaysia 125 7.7 0.0 2070 3.00 0.181 -250.0 Indonesia 32 2.0 0.0 2070 3.00 0.046 -64.0 Oceania 30 1.8 0.0 2070 3.00 0.043 -60.0 Ecuador 85 5.2 0.0 2070 0.28 0.011 61.2 Brazil 329 20.1 20.0 1656 0.58 0.115 138.2 Note: a The non-zero export tax rates are from Imran and Duncan, Table 7, page 21, and refer to 1982 and 1983 (for Brazil). b The long- run elasticities for Brazil, Cote d'lvoire, and Malaysia were obtained from Akuyama and Bowers, page 25. They apply to ten-year periods, using the highest production levels to obtain those values. We assume that the elasticities of Indonesia andOceania are equal to that of Malaysia. The other elasticities are from Behrman. cThe slope is the change in metric tons for a one US dollar change in the domestic producer price. 174 Commodity Exports and Real Income in Africa: a Preliminary Analysis World supply is 1633 and the world price is 2070. sands of metric tons, and those on profits (and reve- The elasticity of demand is 0.4, the intercept is 2286.3 nues in theother tables) are in millions of U.S. dollars. and the slope is -315.6 M.T. per U.S. dollar. The actual world price is 2,070, while under free trade it is 1,562 or 24.5 percent lower. Hence, output Simulation Results falls for countries with an initial tax rate below 24.5 percent (all five Non-African countries) and rises in In this section, we perform five sets of simulations. the countries with an initial tax rate above 24.5 per- This will allow us to examine the effects of some of cent (all four African countries). The fall is largest in the recent changes in the cocoa market under actual the countries with high elasticity (by 74% in Malay- taxes and compare them with the effects in the case sia, 75% in Indonesia and 77% in Oceania). Output where countries choose Nash taxes. We also examine increases by over 100 percent in Ghana (the domestic the effects under "myopic" taxes. producer price increases by over 150%, from 621 to The simulations are as follows: 1,562, and the elasticity is .71). Africa's output in- * First, we compare the initial results with actual creases by over 30 percent (from 1032 to 1,349), world taxes with those in the case of free trade, Nash output increases by about 10 percent (from 1,633 to taxes and myopic taxes. This is shown in tables 1,793), and Africa's share increases from 63 percent 17-2 and 17-3. to 75 percent. * Second, we simulate an exogenous outward Profits, on the other hand, fall everywhere. They shift in Ghana's supply curve of 100,000 tons fall most for the small producers with elastic supply (M), and examine the results under existing (by 93% in Malaysia, by 93.6% in Indonesia and by and Nash taxes (see table 17-4). 94.5% in Oceania). The fall in profits is also signifi- * Third, we simulate an exogenous shift in Ma- cant in C6te d'Ivoire (42.6%), in Cameroon (37.7%), laysia's supply curve of 100,000 M.T. (see table in Brazil (31.9%), in Ecuador (27.9%) and in Nigeria 17-5). (24.8%). The only country where the fall in profits is * Fourth, we simulate the impact of C6te negligible is Ghana. Both the actual tax rate of 70 d'Ivoire reducing its tax rate to zero (see table percent and the zero tax rate are sub optimal. In fact, 17-6). we show below that the optimal (myopic or Nash) * Finally, we simulate the impact of 24 above tax rate lies between these two extremes. simultaneously (see table 17-7). Profits for Africa fall from 1,443 to 1,043 (or by 27.7%), and world profits fall from 2,182 to 1,512.3 (by 1. Table 17-2 shows the tax rates, outputs and total 30.7%). Thus, even though Africa's output (and share profits (or income) with actual taxes and under free in world output) would rise under free trade, it trade. The results on output are presented in thou- would come at the expense of a reduction in profits. Table 17-2. Initial results with actual, Nash and myopic tax rates Tax rates Output Profit Revenue (%) (OOOMT) (millions of USS) (millions of USS) Country Actual (2) Nash (2) Myopic (2) Actual (3) Nash (4) Actual (5) Nash (6) Myopic (6) Actual (7) Nash (8) C6te d'Ivoire 25.1 25.2 29.9 585 490 698 496 705 304 220 Ghana 70.0 19.5 20.5 219 421 405 493 544 318 146 Carneroon 40.0 8.2 9.8 118 184 138 125 172 98 27 Nigeria 50.0 5.7 6.3 110 139 202 182 221 114 14 Africa 1032 1234 1443 1296 834 407 Malaysia 0.0 2.8 5.1 125 63 43 14 44 0 3.2 Indonesia 0.0 0.7 1.2 32 17 11 3.4 11.3 0 .2 Oceania 0.0 0.6 1.3 30 16 11 3.2 11.1 0 .2 Ecuador 0.0 3.2 3.3 85 80 151 126 152 0 4.6 Brazil 20.0 13.4 14.7 329 315 523 424 525 136 75 World 1633 1725 2182 1866.6 970 490.2 Note: World price (U.S.$/MI) Actual: 2,070 Nash: 1,779. Profits are defined to include producers' surplus and government revenue. Actual profits are derived by assuming that the calibrated demand and supply curves are true demand and supply curves. These profits will be different in general from actual observed profits (inclusive of tax revenues). 175 Commodity Exports and Real Income in Africa: a Preliminary Analysis Table 17-3 presents the initial results with actual, expected, profits under Nash taxes are larger than Nash and myopic taxes. Let us first consider myopic under free trade for all countries (compare column taxes. For a given country, this tax is derived by (5) in table 17-2 and column (6) in tablel7-3). maximizing profits under the assumption that other The world price under Nash taxes is lower by U.S. rates are held fixed at their current (actual) levels. In dollars 291 per metric ton (MT) than under actual terms of figure 17-2, if B's tax is frozen at t1B, A's taxes, falling from 2,070 to 1,779 or by 14.1 percent. myopic tax is given by t'A. In table 17-3, if tax rates This is due to the significant fall in tax rates for of countries other than C6te d'Ivoire are kept at the several producers: from 70% to 19.5% for Ghana, levels shown in column (1), C6te d'Ivoire's (myopic) from 50% to 5.7% for Nigeria, and from 40% to 8.2% optimal tax is 29.9% (column 2). Similarly, if tax rates for Cameroon. These lower Nash tax rates result in of countries other than Ghana are held at the levels significantly higher output for Ghana (from 219 to shown in column (1), Ghana's myopic optimal tax is 421), Nigeria (from 110 to 139), and Cameroon (from 20.5% (column 2). 118 to 184). In contrast with myopic taxes, Nash taxes allow The Nash taxes are somewhat lower than actual other countries to adjust their taxes optimally. Thus, taxes for Brazil (13.5% versus 20%), they are not for each country, the tax rate in column (2) is optimal significantly different from actual taxes for C6te when other countries choose tax rates at levels shown d'Ivoire (25.2% versus 25.1 %), and they rise from zero in the same column. In contrast to myopic taxes, to 3.2% for Ecuador, to 2.8% for Malaysia, to .7% for Nash taxes do not leave room for profit-increasing Indonesia and to .6% for Oceania. C6te d'Ivoire's changes in tax rates for any country as long as the output falls by 16.3% (from 585 to 490) because of the others keep their taxes at Nash levels. 14.1 % fall in the world price, combined with a long- Not surprisingly, as seen by comparing columns run supply elasticity slightly larger than one. The (5) and (6), each country taken one at a time can outputs of Malaysia, Indonesia and Oceania fall by increase its profits under myopic taxes. Butwhile this about 50 percent (a combination of the 14.1% fall in country maximizes profits, others do not do so under the world price, a small rise in the tax rate and an the myopic tax assumption. Reactions by others elasticity of 3). eventually lead to the Nash equilibrium. World output rises from 1,633 to 1,725 (by 5.7%). The most interesting comparison isbetween actual Hence consumers in developed countries gain from and Nash taxes. A striking result here is that profits the lower world price and higher output. Africa's under actual taxes are higher than under Nash be- output under existing taxes is 1,032 and its share is havior for all countries except Ghana. With as many 63.2 percent. Under Nash taxes, Africa's output rises as nine participants in the market, Nash behavior to 1,234orby19.6%,anditssharerisesto71.5percent, leads to excessive tax competition and results in or by 8.3 percentage points. One of the concerns for lower profits for all participants except one. Also, as Africa has been the fall in its share of world cocoa Table 17-3. Initial results with actual, Nash and myopic tax rates Tax rates Output Profit Revenue (%) (OOOMT) (millions of US$) (millions of US$) Country Actual (2) Nash (2) Myopic (2) Actual (3) Nash (4) Actual (5) Nash (6) Myopic (6) Actual (7) Nash (8) C6te d'lvoire 25.1 25.2 29.9 585 490 698 496 705 304 220 Ghana 70.0 195 20.5 219 421 405 493 544 318 146 Cameroon 40.0 8.2 9.8 118 184 138 125 172 98 27 Nigeria 50.0 5.7 6.3 110 139 202 182 221 114 14 Africa 1032 1234 1443 1296 834 407 Malaysia 0.0 2.8 5.1 125 63 43 14 44 0 3.2 Indonesia 0.0 0.7 1.2 32 17 11 3.4 11.3 0 2 Oceania 0.0 0.6 1.3 30 16 11 3.2 11.1 0 .2 Ecuador 0.0 3.2 3.3 85 80 151 126 152 0 4.6 Brazil 20.0 13.4 14.7 329 315 523 424 525 136 75 World 1633 1725 2182 1866.6 970 490.2 Note: World price (U.S.$/MI) Actual: 2,070 Nash: 1,779. Profits are defined to include producers' surplus and government revenue. Actual profits are derived by assuming that the calibrated demand and supply curves are true demand and supply curves. These profits will be different in general from actual observed profits (inclusive of tax revenues). 176 Commodity Exports and Real Income in Africa: a Preliminary Analysis exports, and the rise in the share of Malaysia and other countries as well as their possible reaction to Indonesia. A Nash strategy would have resulted in a the recommended policy change. This would help higher share for Africa, as well as in a fall in the share avoid a potentially undesirable outcome such as the of Malaysia from 7.7% to 3.6%, and in a fall in the Nash solution where most countries end up worse share of Indonesia from 2.0% to 1.0%. Such an expan- off despite theexpectation of a welfare improvement. sion for Africa would have come at the cost of lower Even if some countries gain-such as Ghana in this profits, however. case-the actual gain would be substantially lower Profits under Nash taxes, in millions of US dollars, than what might have been expected based on the increaseby88 inGhana(orbyover20%), and fall by assumption that myopic taxes in various countries 202 (29%) in C6te d'Ivoire,by99 (1992%) inBrazil,by are mutually consistent. 29 (67.4%) in Malaysia, by 25 (16.5%) in Ecuador, by 2. The effects of a 100,000 M.T. outward shift in 20 in Nigeria (10%), by 13 (9.4%) in Cameroon, and Ghana's supply curve under both sets of taxes are by about 8 (69.7%) in Indonesia and Oceania (70.7%). shown in table 17-4. The difference in the effects of Overall profits fall by 315 or by 14.4 percent. Profits actual and Nash taxes in this case are very similar to in Africa fall from 1443 to 1296 or by 10.2 percent. those in the original case (table 17-3). Hence, we first Africa'sshareintotalprofitsrisesbythreepercentage compare the results in the original case and in the points (from 66.5% to 69.5). present case when actual taxes are levied. The com- The beneficiaries from Nash taxes rather than ac- parison of results when Nash taxes are used is similar tual taxes are the cocoa consumers and Ghana, while and is briefly discussed later. the other producers lose. The main losers, in terms of Clearly, Ghana gains from the exogenous produc- the proportional fall in profits, are Oceania, Indone- tivity increase. Its profits rise from 405 to 579 (see sia, Malaysia and C6te d'Ivoire. column 5 in tables 17-3 and 17-4) under actual taxes. Government revenues increase slightly in Ecua- The increased productivity in Ghana leads to a fall in dor, Malaysia, Indonesia and Oceania, and they fall the world price, from 2,070 to 1,993 (or 3.7%). everywhere else. They fall by 87 percent in Nigeria, The lower world price implies a loss in profits in 72 percent in Cameroon, 53 percent in Ghana, 46 all other countries. The loss is 56 in C6te d'Ivoire percent in Brazil and 27 percent in C6te d'Ivoire. The (8.%), 28 in Brazil (5.3%), 17 in Cameroon (12.3%), 11 proportional fall in revenues is directly related to the in Nigeria (5.7%), 7 in Ecuador (4.6%), 9 in Malaysia reduction in tax rates. (20.9%), 2.7 in Indonesia (24.5%) and 3.3 in Oceania These results are preliminary. Assuming that they (30%). Interestingly, the overall profits for the nine hold after careful empirical analysis is done, it would countries change very little. They rise only from 2,182 lead us to conclude that when providing policy ad- to 2,222, or by 1.8 percent. In other words, even vice to any one country in the case of commodities though Ghana gains from the increase in its produc- such as cocoa, the various bilateral and multilateral tivity, the losses by the other countries are such that donors should take into account both the effect on overall industry profits remain practically un- Table 17-4. Increasing Ghana's intercept by lOO,OOOMT Tax rates Output Profit' Revenue (%) (OOOMT) (millions of US$) (millions of US$) Country Actual (1) Nash (2) Actual (3) Nash (4) Actual (5) Nash (6) Actual (7) Nash (8) C6te d'lvoire 25.1 25.3 560 470 642 460 280 205 Ghana 70.0 23.5 313 492 579 631 437 200 Caxneroon 40.0 8.1 109 175 121 114 87 24 Nigeria 50.0 5.8 107 137 191 173 107 14 Africa 1089 1274 1533 1378 911 443 Malaysia 0.0 2.5 111 53 34 10 0 2.3 Indonesia 0.0 0.6 28 15 8.3 2.5 0 .2 Oceania 0.0 0.6 26 14 7.7 2.3 0 .1 Ecuador 0.0 3.3 83 79 144 122 0 4.6 Brazil 20.0 13.7 322 309 495 405 128 73 World 1659 1744 2222 1919.8 1039 523.2 Note: World price (U.S.$/MT) Actual: 1,993 Nash: 1,722. 177 Commodity Exports and Real Income in Afrca: a Preliminary Anal sis changed. Ghana gains 174, while the other countries ing of investment projects in commodities such as lose 134, so that overall profits increase by 40. Hence, cocoa, bilateral and multilateral donors should take the principal gainers from Ghana's increase in pro- into account the negative impact on the other pro- ductivity are Ghana and the consumers who benefit ducers. Moreover, simultaneous expansion of output from a lower price. in several countries would not generate the retums Africa's profits increase from 1,443 to 1,533, and its on the investment projects which would result from share in total profits rises from 66.1 percent to 69 output expansion in one country only. These points percent. Non-Africa's profits fall from 739 to 689, and were raised in an a somewhat different context in an its share falls from 33.9 percent to 31 percent. early paper by Goreux (1 972).9 The World Bank pol- The lower world price leads to a reduction in out- icy of denyinig lending for output expansion in com- put in all countries except Ghana (given the un- modities such as cocoa (except in special circum- changed tax rates), and world output increases by26 stances) is based on the above considerations. (from 1,633 to 1,659, see column 3 in tables 17-3 and 3. The effects of a 100,000 M.T. outward shift in 17-4), i.e., by one quarter of the shift in Ghana's Malaysia's supply curve are quite similar to those supply curve. Total revenues fall for all countries when the shift occurs in Ghana's supply curve, and with positive tax rates other than Ghana because of are shown in table 17-5. In the case of existing taxes, a lower output and lower world price. the effects are iderntical for all countries other than Interestingly, in the case of Nash taxes, producers Ghana and Malaysia. World supply in both cases is as a whole are able to retain more profits from 1,659andtheworldprice is1,993.lnthepresentcase, Ghana's productivity increase than in the case of Ghana's profits fall from 405 to 380 (and its output existing taxes, even though world prices are lower in falls form 219 to 213), while Malaysia's profits rise the former case. Under Nash taxes (see column 6, from 43 to 123 (and its output rises from 125 to 211). tables 17-3 and 17-4), overall profits rise by 53.2 World profits are lower (2,112) in this case than in (from 1,866.6 to 1,919.8), while under existing taxes the case where Ghana's supply shifts outward (222). (column 5, tables 17-3 and 17-4), they onlyrise by 40. The reason is that in the present case, Ghana's output Also, Nash tax rates in this case are similar to those islower(thanwhenthesupplyshiftoccursinGhana) in the original case for all countries except for Ghana, by 100, while Malaysia's output is higher by 100 (see whose Nash tax rate increases from 19.5 percent to tables 17-4 and 17-5J column 3), and Ghana is the 23.5 percent. lower-cost producer. As is shown in table 17-1, at the In this case, under existing taxes, Africa as a whole actual outputs, Ghana's producer price (and mar- gainsabouthalf (90) of whatGhanagains (174), while ginal cost) is 621 while that of Malaysia is 2,070. all producers gain about 23 percent (40) of what In fact, the outward shift in Malaysia's supply Ghana gains. These findings are again preliminary. If curve leads to overall immiserization for the export- theyhold undercloserempirical scrutinytheywould ing countries as a whole. Total profits fall to 2,112, lead us to conclude thatwhenconsideringthe financ- compared to 2,182 in the actual case. Hence, even Table 17-5. Increasing Ghana's intercept by lOO,OOOMT Tax rates Output Profit Revenue (%) (OOOMT) (millio ns of US$) (millions of US$) Country Actual (1) Nash (2) Actual (3) Nash (4) Actual (5) Nash (6) Actual (7) Nash (8) C6te d'lvoire 25.1 25.2 560 470 642 457 280 203 Ghana 70.0 19.7 213 408 380 463 297 138 Cameroon 40.0 8.1 109 174 121 113 87 24 Nigeria 50.0 5.8 107 136 191 172 107 14 Africa 989 1188 1334 1205 771 373 Malaysia 0.0 6.6 211 140 123 70 0 16 Indonesia 0.0 0.6 28 14 8.3 2.4 0 .1 Oceania 0.0 0.6 26 13 7.7 2.2 0 .1 Ecuador 0.0 3.3 83 79 144 12-1 0 4.5 Brazil 20.0 13.7 322 309 495 404 128 73 World 1659 1743 2112 1804.6 899 467 Note: World price (U.S.$/MI) Actual: 1,993 Nash: 1,717. 178 Commodity Exports and Real Income in Africa: a Preliminary Analysis Table 17-6. Comparing the initial equilibrium and the effect of fixing Cote d'Ivoire's tax at zero Tax rates Output ProfitI Revenue (%) (OOOMT) (milions of US$) (millions of US$) Country Original (1) New (2) Original (3) New (4) Original (5) New (6) Original (7) New (8) C6te d'lvoire 25.1 0.0 585 739 698 629 304 0 Ghana 70.0 70.0 219 206 405 352 318 275 Cameroon 40.0 40.0 118 100 138 105 98 76 Nigeria 50.0 50.0 110 105 202 179 114 100 Africa 1032 1150 1443 1265 834 451 Malaysia 0.0 0.0 125 95 43 70 0 0 Indonesia 0.0 0.0 32 24 11 2.3 0 0 Oceania 0.0 0.0 30 22 11 2.1 0 0 Ecuador 0.0 0.0 85 82 151 120 0 0 Brazil 20.0 20.0 329 313 523 396 136 119 World 1633 1686 2182 1855.4 970 570 Note: World price (U.S.$/Ml) Actual: 2,070 Nash: 1,905. though Malaysia's profits rise by almost 200 percent, are extremely close to the tax rates when the supply total industry profits fall. Comparing the present ca- shift occurs in Ghana. The tax rates for Ghana and se with that when Ghana's supply shifts outward Malaysia are quite different in each case. In the pre- (with world output and price being the same in both sent case, Ghana's Nash tax rate is 19.7 percent and cases), the reason for which profits are only 2,112 that of Malaysia is 6.6 percent. When Ghana's supply rather than 2,222 is that a larger share of the given tot- shifts outward, the Nash tax rates are 23.5 percent for al output is produced by a higher-cost producer (Ma- Ghana (i.e., 3.8 percentage points higher) and 2.5 laysia) rather than by a lower-cost producer (Ghana). percent for Malaysia (4.1 percentage points lower). Africa's profits in this case are 1,334 out of a total WithNashtaxes,profitsforallcountriesotherthan of 2,112, so that its share is 63.2 percent. That com- Ghana and Malaysia are extremely close in both pares with the higher profits (of 1,443) in the actual cases. Malaysia's profits rise from 10 (when Ghana's case (and a share of 66.1 %), and profits of 1,533 (share supplyshifts outward) to 70 in the presentcase, while of 69%) when Ghana's supply curve shifts outward. Ghana's profits fall from 631 to 463. Total industry In the case of Nash taxes, tax rates in the present profits in the present case are 1,804.6 rather than case in all countries other than Ghana and Malaysia 1,866.6 in the original cae, i.e., they fall by 62. How- Table 17-7. Comparingthe initial equilibrium and theeffectoffixingCoted'lvoire's taxatzero and increasing the intercepts of Ghana and Malaysia by lOO,OOOMT Tax rates Output Profit Revenue (%) (OOOMT) (milions of USS) (millions of US$) Country Original (1) New (2) Original (3) New (4) Original (5) New (6) Original (7) New (8) C6te d'lvoire 25.1 0.0 585 673 698 522 304 0 Ghana 70.0 70.0 219 295 405 483 318 362 Cameroon 40.0 40.0 118 84 138 80 98 59 Nigeria 50.0 50.0 110 102 202 160 114 89 Africa 1032 1154 1443 1245 834 510 Malaysia 0.0 0.0 125 167 43 77 0 0 Indonesia 0.0 0.0 32 17 11 3.0 0 0 Oceania 0.0 0.0 30 15 11 2.8 0 0 Ecuador 0.0 0.0 85 80 151 124 0 0 Brazil 20.0 20.0 329 299 523 412 136 105 World 1633 1732 2182 1863.8 970 615 Note: World price (U.S.$/MI) Actual: 2,070 Nash: 1,754. 179 Commodity Exports and Real Income in Africa: a Preliminary Analysis ever, with existing taxes, they fall from 2,182 to 2,112 miningtheequilibrium tax rates, outputs, profits and or by 70. Hence, the use of Nash taxes enables the revenues. In future work, we plan to closely examine industry as a whole to suffer smaller losses from the the form of the relevant functions. We will also exam- supply shift than with existing taxes. ine the properties of other Nash equilibria. These in- Thus, Africa aswell as the industry as awhole loses clude revenue-maximizing Nash taxes, and usingex- both with actual and Nash taxes. This would seem to port quotas rather than the export tax as the policy reinforce the point made earlier that donors should variable. Our findings so far seem to suggest that in take into account the interdependencies among pro- providing policy advice and support of investment ducing countries when assessing investment pro- projects in the case of commodities such as cocoa, the jects. donor community should take into account the effe- 4. Up to mid-1989, the Caisse de Stabilisation in the cts on and possible reactions of the other producing C6te d'Ivoire did not pass on the significant reduct- countries. ion in world cocoa prices to its producers. The Caisse was still paying the producers more than the price it Notes obtained (net of marketing costs). It was paying 400 CFA /MT, implying a cost for the Caisse of well over We would like to thank Shantayanan Devarajan for use- the FOB price of 500 CFA /MT. We simulate the red- ful comments and Brendan Kennelly for excellent research uction in C6te d'Ivoire's export tax rate by setting it assistance . The views expressed in this paper are the equal to zero and keeping the tax rates of the other authors' and do not necessaily reflect those of the World countries at their initial level. This is shown in table Bank or its affiliated organizations. 17-6. 1. This viewderives from the theory of the second best We notice a dramatic increase in C6te d'Ivoire's which emphasizes that distortions should be attacked at output, from 585 to 739, orby 26.3 percent. The world the source. See Bhawati (1971) for further details. price falls from 2,070 to 1,905 or by 8 percent, so that 2. See Bhagwati (1988) on this issue. all countries other than C6te d'Ivoire lose. However, 3. For example, see the classic article by Johnson (1954) C6te d'Ivoire also loses as its profits fall from 698 to and the more modern treatment in Dixit and Norman 629, orbylOpercent. Hence, thestrategyfollowedby (1980, Ch. 6). C6te d'Ivoire was certainly not optimal from its own 4. If we feel that general equilibrium analysis may be viewpoint (unless the Caisse aimed to stabilize the useful, we will engage in the future. This may be the case producer price and believed that the changes leading if we wnat to include the implicit tax on exports due to to a fall in the world price were transitory). With its tariffs on imported manufactures (Lemer Symmetry Theo- large share in the world market, a zero tax rate does rem). not maximize profits. 5. Forasurveyof recent game theoreymodels, seeDixit 5. Finally, we examine the impact of the three (1986) and the references therein. previous simulations simultaneouslr, i.e., a 100,000 6. Weassumethattheimporter,C,doesnotlevyimport M.T. supply shift in Ghana and Malaysia and a zero tariffs. export tax rate in C6te d'Ivoire. The results are shown 7. Power to affect the price in the market for a country's in table 17-7. The output of C6te d'Ivoire rises from exports of a competitively produced commodity is a case 585 to 673 because of the fall in its tax rate (but rises where optimal export taxes will raise welfare for the ex- by less than under the previous simulation), and it porting country. rises for Ghana and Malaysia because of the supply 8. The tax rates correspond to 1982 and 1983. reliable shift. Output falls in all other countries. Total output estimates of export tax rates for 1986 were not available. rises from 1633 to 1732, and the world price falls from 9. He examined the returns on investment projects alter- 2,070 to 1754 (by 15.3%). Profits fall in all countries natively from the viewpoint of the firm of the investing except Ghana and Malaysia. Total profits and Af- country, of all producing countries, and of the world as a rica's profits also fall. whole. Concluding Comments References The results presented in this paper are preliminary. Akiyama, T. and A. Bowers (1984); "Supply Response of The demand and supply functions were linearized Cocoa in Major Producing Countries," Division work- based on point estimates of the relevant parameters. ing paper No. 1984-3, Commodities Studies and Projec- The exact form of these functions is crucial in deter- tion Division, EPD, World Bank. 180 Commodity Exports and Real Income in Africa: a Preliminary Analysis Balassa, B. (1978); "Exports and Economic Growth: Further Humphreys C. and W. Jaeger (1989); "Africa's Adjustment Evidence" Journal of Development Economics (une): 181- and Growth," Finance and Development, June. Hum- 89. phreys C. (1989); "Response to Singer," Finance and Behrman,J.R. (1968); "Monopolistic Cocoa Pricing". Ameri- Development, December p. 52. can Journal of Agncultural Economics. 50: 702-19. Imran, M. and R. Duncan (1988); "Optimal Export Taxes Bhagwati, J.N. (1971); "The Generalized Theory of Distor- for Exporters of Perenial Crops," Division Working Pa- tions and Welfare," in J. Bhagwati, et al., Trade, Balance of per No. 10. International Commodity Markets Division, Payments and Growth, Amsterdam: North-Holland. WPS 10, World Bank. Bhagwati, J.N., (1978). Anatomy and Consequences of Ex- Johnson, H.G. (1954); "Optimum Tariffs and Retaliation. change Control Regimes. Cambridge, Mass.: Ballinger Review of Economic Studies 21,142-53. Bhagwati, J.N., (1988). "'Export-Promoting Trade Strategy: Keesing, D. (1967); "Outward-Looking Policies and Eco- Issues and Evidence". World Bank Research Observer 3 (1): nomic Development". Economic Journal (June):303-20. 27-57. Krueger, A.O. (1961); "Export Prospects and Economic Cairncross, A. (1962); Factors in Economic Development. Lon- Growth: India: A Comment' Economic Journal (une): don: Allen and Unwin. 436-42. Dixit, A. and V. Norman (1980); Theory of International Landell-Mills, P., Agarwala R. and S. Please (1989); "Sub- Trade, London: Cambridge University Press. Saharan Africa: From Crisis to Sustainable Growth - a Dixit, A., (1986), "Comparative Statics for Oligopoly," In- Long-term Perspective Study", World Bank. ternational Economic Review, 27(1): 107-22. Panagariya, A. and M. Schiff (1990); "Optimal and Reve- French, B.C. and J.L. Matthews (1971); "A Supply Re- nue-Maximizing Export Taxes in a Multi-Country sponse Model for Perennial Crops," American Journal of Model". World Bank, CECTP. Agricultural Economics 53. Singer, H.W., (1989); "African Adjustment," Finance and Goreux, L.M. (1972); "Private, National and International Development, December: p. 52. Returns; An Application to Commodity Lending". Eur- World Bank (1989); "StrengtheningTrade Policy Reform", pean Economic Review. 3(2): 131-80. SecM89-1454. 181 18 Response to Relative Price Changes in C6te d'Ivoire: Implications for Export Subsidies and Devaluations Victor Lavy, John L. Newman, Raoul Salomon, and Philippe de Vreyer The ability of firmns to expand their output and ex- export subsidy alone might be thought sufficient for ports in response to changes in relative prices has expanding exports and improving the balance of been central in the debate regarding the use of export payments, such a policy would not increase the price subsidies as a trade policy instrument. Since the early of imported goods and, thus, would not have as large 1980s, the introduction of export subsidies has been effects on relative prices as that of a devaluation. In proposed as a way to counteract the adverse effects order to address the major concern of expanding of an exchange rate overvaluation among member output and employment in the tradeables sector, pro- countries of the West African Monetary Union. De- ponents of the tariff cum subsidy scheme argued that spite deteriorating terms of trade and mounting it was necessary to mimic the effects of a devalu- external debt, these countries and France have opted ation more completely and increase import prices as not to change the CFA exchange rate with the French well. Franc. Even with the increase in both export subsidies Faced with an exchange rate out of equilibrium, and import tariffs, this policy will still not mimic the economy must adjust-either through adjust- exactly a nominal devaluation. A devaluation would ments in quantities (output and employment) or in affect services and capital movements as well. More- prices. As quantity adjustments typically involve over, export subsidies will lead to higher export higher foregone consumption during the adjustment prices only if firms choose to participate and take period, adjustments generated by changing the rela- advantage of the subsidy program. If, due to lack of tive price of traded to nontraded goods are generally credibility of the program or high transaction costs, preferred. In the absence of a nominal devaluation, firms do not participate, their export prices will not one way to attempt a real devaluation would be to rise. Finally, the budgetary consequences of the sub- reduce nominal wages through tight monetary and sidy cum tariff policy could be substantiallydifferent fiscal policies. from that of a devaluation. An alternative means of altering the relative price Opponents of the tariff cum subsidy policy of traded to nontraded goods is to attempt to mimic claimed that it is doomed to fail because: a devaluation by raising import tariffs and export a producers in Africa are slow to adjust to a subsidiesby the same magnitude. This is, indeed, the change in market signals policy that has been proposed for C6te d'Ivoire. * the short run export supply elasticity is very There has been some disagreement over whether the small since African exporters cannot compete plan has been implemented as proposed. While an and capture international markets due to their 182 Response to Relativ Price Changes in Cdtc d'Ivoire: Implications for Export Subsidies and Devaluations high labor cost compared to their competitors making the domestic price endogenous from the and to the protectionist policies of the devel- viewpoint of the industry. Allowing for the endoge- oped countries neity of the domestic price requires a model of the * the export subsidies will lead to a comparable domestic demand for the output of the manufactur- increase in domestic prices, leaving unaltered ing sectors. Thus, we model jointly firm output sup- the relative profitability of exporting ply and input demand functions and domestic de- * the import surcharges will not raise enough mand for domestically produced goods. revenue to finance the export subsidies, imply- The short-run focus of this study is motivated by ing a heavy fiscal burden on other revenue methodological as well as policy considerations. sources. The expected long run nonsustain- First, analyzing the long run response implies treat- ability of this deficit will undermine from the ing the capital stock as endogenous and modeling beginning the credibility of the program and investment behavior and credit markets. This task is will discourage firms from responding to it. further complicated by difficulties in measuring the price of capital services. Second, from a policy point These views have not been based on extensive of view, the uncertainty associated with this particu- empirical evidence, as evidence on export and out- lar policy reform in C6te d'Ivoire suggests that a put supply response is even more scarce for African sensible approach is first to evaluate the immediate countries than it is for other developing countries. short-run effects. One of the few studies is by Balassa (1987) who On the supply side, our empirical model follows presents results suggesting that the response of ex- closely the production theory framework of Diewert ports of goods and agricultural products to price and Morrison (1988) (see also Kohli, 1978). The major changes is actually greater in Sub-Saharan African new element that we introduce to theirframeworkis countries than in other developing countries. (for the modeling of the demand side. Recent studies evidence on other developing countries see, for ex- such as Zilberfab (1980), Aspe and Giavazzi (1982), ample, Balassa et al 1986; Nogues 1989; Milanovic and Faini (1988) have discussed the role of domestic 1986; Artus and Rosa 1978; Goldstein and Khan 1978; market conditions and relative prices in modeling Bauman and Braga 1988). export supply, but have not treated domestic prices The objective of this paper is to address the first as endogenous and identified empirically the influ- three issues above and to provide empirical evidence ence of exogenous export and import prices on do- on export and output supply responses that could mestic prices. Our work also differs in that we use inform policy discussions on the tariff cum subsidy sector-level data aggregated from individual firm scheme introduced in C6te d'Ivoire in 1986. We base records, whereas Diewert and Morrison use aggre- the empirical work on data from two sources for the gate data. We also allow for non-constant returns to six years that preceded the implementation of the scale. program (1980-85). Information on sector level out- We compare results from a model that estimates put, exports, domestic sales, capital, and variable only the supply side with one that estimates jointly input use is obtained by aggregating information supply and demand. We simulate the model to yield from individual firm-level panel data obtained from domestic supply and export supply elasticities, as the Banque des Donne es Financie res. The individual well as measures of the sensitivity of the domestic data was aggregated because export, import, domes- price and domestic demand to exogenous variation tic output, and input price indices obtained from the in export and import prices. The empirical results are Ministere du Plan et de l'Industrie are available only very sensitive to the assumption regarding the endo- at the sectoral (3 digit) level. geneity of the domestic price and to the inclusion (or The paper models the short-run response of firms exclusion) of the demand function in the model. Not to exogenous changes in export and import prices, allowing for domestic prices to change and affect the taking into account the possibility that firms may sell relative profitability of exporting versus selling do- to both domestic and foreign markets. In such a mestically leads to a large bias in the estimate of the framework, firms may alter sales in response to the export elasticities. relative profitability of the two markets. The net out- The main and somewhat surprising result is that put response depends on the response in both mar- manufacturing producers in COte d'Ivoire are able to kets. In world markets, Ivorian industries are as- expand their exports in the short run in response to sumed to be price takers. In the domestic market, an increase in export prices. However, most of this even if individual firms face exogenous prices, the expansion comes at the expense of sales in the do- industry demand curve will be downward sloping, mestic market. The net short run output and employ- 183 Response to Relative Price Changes in Cote d'Ivoire: Implications for Export Subsidies and Devaluations ment responses are small. A second important and supplied to domestic and foreign markets vary as fu- less surprising result is that the domestic supply nctions of the instrumented domestic prices and ex- curve is much more sensitive to price changes than ogenous export and input prices. In the second ap- the export supply function. This implies that any proach,weexplicitlymodelthedomesticdemand for exogenous shocks that lead to an increase in domes- domestic output and for imports. Domestic demand tic demand, such as an increase in import prices, will and import demand functions are then estimated have a sizable effect on domestic sales and output, and a contractionary effect on exports. Finally, in- Figure 18-1. Bias in Estimated Net Export creases in export prices alone were estimated to have and Output Responses a much smaller effect on output and employment than would increases in both export and import The exportmarket prices, as would occur in a devaluation or in the tariff x' cum subsidy program. Empirical Specification As has been observed elsewhere, the evidence fr- p om C6te d'Ivoire indicates that firms tend not to spe- , cialize in exporting but to sell their products in both domestic and foreign markets. Between 1980 and 19- px 85 about a third of all manufacturing firms exported , A to foreign markets. Of those who exported, roughly sixty percent of the value of their sales was to the do- mestic market. We therefore model firm behavior by allowing firm and industries to produce two differe- nt products in a joint production process. The produ- cts could be identical or differentiated from each oth- er. Often export products willbe similar to the produ- ct destined for the domestic market, but distingui- - shed by being of a higher and more uniform quality. Our empirical specification of joint production is E0 E2 E1 based on the production theory approach employed by Diewert and Morrison (1988) in their study of The domestic market aggregate export supply and import demand func- tions in the U.S. In their approach, one specifies a single profit function and derives supply functions s (p) to domestic and foreign markets by partially differ- \ entiating the profit function with respect to the out- s ( put price in the respective market. Our work differs in two main respects. First, we work with sectoral rather than aggregate data. It is not our intention to model the complete trade balance, but rather to model short run responses to price changes among Pi --------- firms in the manufacturing sector. Second, we model the domestic demand for domestically produced Po goods and imports. Firms are assumed to operate under perfect com- petition in factor markets. Thus, all input prices are treated as exogenous. Export prices are also assumed . to be exogenously determined. We consider two al- / temative ways of accounting for the endogeneity of domestic prices. In the first approach, we simply instrument for the endogenous price of domestic go- s 1 s2 so ods and estimate how input demands and outputs 184 Response to Relative Price Changes in COte d'lvoire: Implications for Export Subsidies and Devaluations jointly with the domestic output supply, export sup- If the domestic price is forced to remain atP, then ply, labor demand and intermediate goods demand the export and domestic responses can be seen to be functions. The domestic market equilibrium condi- upwardly biased (cf. El versus E2 and Si versus S2) tion for domestically produced goods determines the and the increase in total output will be downwardly domestic price level and the quantity sold. The ad- biased. The magnitude and the importance of these vantage of the first approach is that the estimation biases will be demonstrated later in the discussion of procedure is simpler, being very similar to the proce- the empirical results. dure under the assumption of exogenous domestic Despitethebias, it isstill usefulto considerthe sim- prices. Using instrumented domestic prices purges pler model as it provides an upper bound to the ex- the system of simultaneous equations bias and al- pected size of the export supply response to export lows one to estimate the model as a system consisting price changes. If there is no observed response of ex- only of output supply and input demand equations.1 port supply to changes in export prices when domes- The main disadvantage of this approach is that the tic prices are held constant, it is unlikely that there simulations obtainable from this approach do not tell will be any response once the influence of export pr- us exactly what we would like to know. Either under ices on domestic prices is taken into consideration. theassumption of exogenous domestic prices or with Thus, we first specify the model using only the prod- instrumented prices, it is only possible to obtain es- uction relations and then indicate how this approach timates of the change in domestic supply or export is modified to take account of the demand side. supply with a change in export prices, holding do- mestic prices constant. However, one would not ex- 7he Supply Side pect domestic prices to remain constant. Export supply and domestic supply functions are We derive econometric specifications of producer derived holding the output price in the other market supply and input demand functions, consistent with constant. Thus, an increase in export prices that in- profit-maximizing behavior, from a generalized Mc- creases the relative profitability of exporting will Fadden restricted profit function of the type em- shift the domestic supply function to the left. If do- ployedbyDiewertand Morrison (theirequation 8.4). mestic prices are assumed exogenous, the leftward This form of the profit function is one of the class of shift of the domestic supplycurve willnot lead to any flexible functional forms and can be specified with or change in domestic prices. However, if as argued without an assumption of constant returns to scale.Z3 earlier, it ismore reasonable to assume thattheindus- For a nonconstant returns to scale technology, the try demand curve is downward sloping, then the restricted profit function is:4 shift in the domestic supply function will lead to an increase in domestic prices. As domestic prices (18-1) lI' (PdtPe,,wt,Ki )bssPdtKt + bluwItKt + bggwfKt change, this would call forth a leftward shift of the export supply function. The net result after the do- + o PIt + tx 14t + aw Itt + CEgW ft mestic price adjustment would be a smaller substitu- tion away from domestic sales to exports and, conse- +bstPttKi + bxtPtetKt + butwlttKt + bgtWftKtJ quently, a smallerincrease in exports than if there had been no adjustment. +bPtd + bxPt' + biwlt + bgwuI The bias in the estimat of the net ex[ort and output p +xpe + responses from neglecting to account for the adjust- +PS 2Kt' + tKt' + PlwtKt' + Pgv.Kt ment of the domestic price is illustrated in figure 18-i. An increase in the export price from PO to Pl + ysIt2Kt + yxt1t2Kt + ylwItt2Kt + ygwft2K, leads to firms to move along the export supply func- 1 (pd)2 1 (wI)2 2 tion and to reduce the amount they are willing to + bsx e ) Kt+- bx Kt 1+-, (i) Kt supply to the domestic market at a given price, a 2 Pt x t leftward shift in the domestic supply function from d l S to 5'. Even if all firms take prices as given, the b ) Kt+b tl'sg(-) Ki+big t lKt reduction in the industry supply curve will lead to pte p1 pte an increase in the domestic price if the industry de- where:St = output supplied to domestic markets; Xt mand curveis downward sloping. Theincrease in the = output supplied to foreign markets; Lt = real ex- domestic price leads to a shift of the export function penditure on labor; Gt = real expenditure on from X to X'. The new domestic priceand output will intermediate goods; wlt = the cost of labor; wu = the be at Pi and S2, while exports will be at E2. cost of intermediate goods; Pj = the domestic price 185 Response to Relative Price Changes in Cote d'lvoire: Implicationsfor Export Subsidies and Devaluations of domestically produced goods; Pi = the foreign S 1 1 priceof domesticallyproduced goods; Kt =thecapital = 2 bss + bstt + ast (K2) + b5 (-) + PsKt + y5t2 stock; t = time; Theb's a's P's and y's are parameters Kw to be estimated. + bsx (-) + b8, (-) + bsg (-) + Es e is linearly homogeneous in prices. A constant Pf PE returns to scaleprofit function would be obtained by The exact forms of the output supply and input de- imposing the restrictions that bi = 0, ai = 0, pi = 0, and mand equations were determined by limitations of yi = O, i = s,x,l, and g. the available data. While we had information on pro- As Diewert and Wales (1987) indicate, in order for duction, domestic sales, exports, and input use of in- the profit function to satisfy the theoretical curvature dividual firms in the manufacturing sector, we did properties, namely, that the profit function be convex not have information on the output or export prices with respect to prices, the matrix of second deriva- of the individual firms. Information on prices was tives of the restricted profit function with respect to available only in the form of sous-branche level price prices must be positive semidefinite. This will be indices, corresponding to a three digit SITC classifi- satisfied if and only if the matrix cation. In the absence of output prices of individual firms, bs, bs, bsg we estimated the model at the sous-branche level to (18-2) B = b bl, b1, ensureaclosecorrespondencebetweenthepricedata bsg big bg. and theoutputsupplyandinputdemandfunctions.5 To arrive at sous-branche level values of domestic is positive semidefinite. We did not impose positive sales, exports, and expenditure on inputs we semidefiniteness in the estimation, but did check summed the values of individual firms in the sous- whether the conditions were satisfied. branche. We estimated our models for two different Partially differentiating the restricted profit func- samples-one where the sous-branche level values tion with respect to input and output prices yields: were formed by summing the individual values over (1) a domestic supply function; (2) (minus) the labor all firms in the sous-branche and the other formed by demand function; (3) (minus) the intermediate input summing over only the firms that had exported in demand function; and (4) the export supply function. that year. The resulting solutions are analogous to equations We considered two samples because it was not ob- (8.5-8.9) in Diewert and Morrison (1988). As they vious what aggregation would lead to the closest suggest, we divide by Kt to make the assumption of connection between firm decisions and the relevant homoskedasticity of the error terms more plausible. price data. The export price index, based on quantity weights of year t, does not capture changes in export Y18-3) xt 1xx 1 bxtt + axt + Yxt2 prices of firms who happened not to export during (18-3) K = + bet + axt ) + bx(Rt-) + PxKt + yxt2 year. However, these unobserved price changes mi- Kt t14' t ~ ght be expected to affect the observed sous-branche 1 P? I wl 1 bgU changes in output and input demands from year t-1 b(sx 22 blx bgx ( )2 to year t. Limiting the sample to exporting firms and sxmj ix ~ correcting for the selectivity bias in forming that (P9wI) (PYw#) (wiwi) sample might present a closer connection between - (pf2 (pf)2 - pg)2 + EX the export price and the output of the firms. How- ever, limiting the sample to exporting firms weakens 14 1 the connection between the domestic price index and = bl + blt + alt () + b (K) + Kt + yet2 output decisions since we do not have access to a Kt Kt ) domestic price index specific to exporting firms. Pt, WI Wi A second feature of the data was that because the + b() + b1x (NI) + big (p) + El base used in calculating the published price index changes every year, it was impossible to use a con- Gt 1 1 stantbase-periodweightedpriceindicesintheanaly- = bgg + bgtt + CLgt (K) + bg + i +gKt + ygt2 sis. Because the value of the published price in- dex K1t t in year reflects the change in prices between year -1 Pi + bi and, maintaining a close connection between the + bsg (p) + big (p) + bgx (-) + £g theoretical model and the econometric specification Pf Pf Pf implies that the estimating equations should relate 186 Response to Relative Price Changes in Cdte d'Ivoire: Implications for Export Subsidies and Devaluations changes in output supplies and input demands from mestic demand for domestically produced goods year t-1 to t to the value of the price index in year t. and the demand for imports. Again, given the nature Therefore, we estimated the model in differenced of the price indices, we estimate the model using formn. This also has the advantage of eliminating any differences in demands. Since the equilibrium condi- fixed unobserved sous-branche specific effects. tion is that the domestic supply equal the domestic demand of domestically produced goods, the meas- The Demand Side ures of supply and demand must be in the same units. Because domestic supply St is deflated by real While maintaining the same specification of the capital, we also deflate domestic demand Dt by real supply side, we now add a simple model of the capital. Thus, the demand side of the model is given demand side. Domestic consumers may consume by: either domestically produced goods or imported Dt Dtd l * Yd goods, which are available at an exogenously deter- (18-5) Kt - -T - + ddm ptnpm ) Kt mined world price. As was done on the supply side, we follow a dual approach and derive domestic con- Yt t sumers' demands for domestically produced goods + b (1- +bdt Kt + a (1-Yt) and imported goods by partially differentiating their Kt Yt- K &K Yt-i expenditure function with respect to domestic prices 1 .Yt Yt and imported prices respectively. + adK- - +d+(Yt At a very high level of aggregation, one may posit It Kt- t only two goods-domestically produced goods and Y imported goods. However, as we are working with + Yyd-(2t-1) + Ud sous-branche level data, there are many more goods, t each with its own price. In general, the expenditure Mt Mt11. Y 1 P*d 1 2 Yt function would be a function of the prices of all goods K, K Y - 2bdm ( ) - in the system. In order to simplify the model, we t_t 2 Kt assume strong separability among goods in different 1 Y Y t Y sous-branches and across other non-manufacturing + bm- (1 _ t) + bmt + amK (1 _ 7-) sectors. This implies that demand for domestically Kt Kt t- produced goods will be a function only of the prices I. Yt Y of domestically produced goods and imports within + am- Yt- + pm Kt (Yt-Yt-1) that sous-branche. Our implicit assumption is that t t- 1 t we are imposing a set of zero restrictions on parame- Yt ters in the expenditure function. We do not test for + Ym"K (2t-1) + um these restrictions, but, in principle, they could be relaxed at the cost of reducing degrees of freedom. where: Dt is the domestic demand for domestically This is obviously an assumption one would want to produced goods; Mt is the demand for imported go- relax at a later stage. ods;Yt is real GDP; P't is the price of imported goods. We adopt the same functional form for the expen- The b's, a's, p's, and y's are parameters to be esti- diture function as we did for the restricted profit mated, and all other variables are defined as before. function, namely the generalized McFadden form. The expenditure function is defined as: The Data (18-4)E (P,Pf,Y1 ) =ddm ()Yt + bddt The data used in the estimation are drawn from t two sources. The export, output, and import price + bmmPP2Yt + bdPF + bmP?' + bdPftYt + ad4Pt indices were obtained from the Ministere du Plan et de l'Industrie. These series were only available from + amPFt + PdPiWY + PmP?Y? + YdPNt2Yt 1980. The intermediate goods index was calculated by applying the output price indices to the input-out- + yPVpt2Yt put matrix for production in C6te d'Ivoire. It was available only at the two digit level. To ensure con- Partially differentiating the expenditure function formability with the other indices, we calculate a with respect to the domestic price of domestically labor cost index as the median nominal wage of all produced goods and imported goods yields the do- firTms in the sous-branche in year relative to the me- 187 Response to Relative Price Changes in C6te d'lvoire: Implications for Export Subsidies and Devaluations dian nominal wage in year t - 1. The median was * the theoretical curvature conditions, that the chosen to reduce the effect of outliers. profit function be convex with respect to Data on the value of imports within the sous- prices, are satisfied in all but one of the models branche classification are also obtained from the the estimated. For all cases except for the estima- Ministe re du Planet de l'Industrie. The imports refer tion of all firms with the demand side and a to imports of final goods within this classification, constant returns to scale technology, all the not the imported inputs used by domestic firms eigenvalues of the matrix of price coefficients within the sous-branche. were positive, indicating a positive semidefi- All other data on domestic sales, exports, expendi- nite matrix. ture on labor, and expenditure on intermediate in- puts was calculated from information provided to Because of the nonlinearities present in all the the Banque des Donne es Financie res by individual models, it is difficult to draw inferences from tables firms in the manufacturing sectors. The value of of the estimated coefficients of the structural model. domestic sales was measured by the value of gross Model simulations based on the estimated coeffi- sales minus the value of exports. Capital was meas- cients provide a clearer picture of the effect of price ured as net cumulated investment, deflated by an changes on output and exports.8 By simulating the aggregate capital goods price index. All differences model one hundred times, one can calculate not only are real differences, expressed in prices prevailing at the mean elasticity but also a standard error of the time t - 1 and deflated by the appropriate price index. simulation. For over 90 percent of the sous-branches This information was available goingback to 1976, and years, one can not reject the hypothesis that the but we could not consider a longer sample period distribution of the one hundred solutions of the en- owing to the limitation on the price data. Details on dogenous variables is normal. Thus, these standard the construction of the data and a brief description of errors can be used to form confidence intervals for the data set are available from the authors. the simulated elasticities. Table 18-2 presents arc elasticities based on si- Estimation and Results mulations of the model with supply equations only and non-constant returns to scale for the sample We estimated models with and without the de- aggregated over all firms. The reported elasticities mand side equations and with and without the re- capture the partial effects of a change in one exo- strictions of a constant returns to scale technology for genous variable, holding all other exogenous vari- two different samples.6 The models with only the ables constant. In particular, it is important to note output supply and input demand equations were that the elasticities with respect to changes in export estimated using iterative seemingly unrelated re- prices, labor costs, and intermediate goods costs gression. The models with the demand side were were obtained holding domestic prices constant. estimated using iterative three-stage least squares. These elasticities pertain to the short run, keeping The results are presented for the sample obtained capital fixed. Standard errors of the simulations are from an aggregation over all firms in the sous- presented in parentheses. There is very little disp- branche in table 18-1. ersion around the mean effect. The low standard We also estimated the model with only exporting errors reflect the fact that the price variables that are firms using a two-stage correction for possible selec- changed in the simulations were precisely estimat- tivity bias, but found little difference in the results ed in table 18-1. with the more complicated selectivity bias correc- Table 18-2 indicates a positive export supply re- tion.7 (The result of these regrssions are presented in sponse to increases in export prices. Given the esti- Lavy,Newman,anddeVreyerl990).Themainpoints mated elasticity of 0.34, a 40 percent increase in ex- to make with respect to these estimates are that: port prices (of the order of magnitude contemplated * the parameters are all of the expected signs; in the subsidy cum tariff program) would be ex- * on the whole, the parameters on the price vari- pected to lead to a 12 percent increase in exports. This ables are estimated quite precisely; is very close to the value of the short-run own price * the results with exporting firms only are simi- export supply elasticity obtained by Diewert and larto thosewith all firms, butwith a somewhat Morrison (1988) using aggregate U.S. data. Using a greater sensitivity to price variation; very different methodology and aggregate quarterly * the restrictions implied by a constant returns data for Greece, Balassa et al (1986) find a short-run to scale production process are rejected in at elasticity of around 0.6. Their estimated long run the one percent level in all cases; and elasticities are considerably higher. 188 Response to Relative Price Changes in C6te d'Ivoire: Implicationsfor Export Subsidies and Devaluations Table 18-1. Allfinns Supply side only Joint supply and demand Parameter CRTS Non-CRTS CRTS Non-CRTS bsx 4.38* 3.26* 39.87* 39.85* (1.35) (1.30) (1.48) (1.97) bsi -0.60* -0.47* -3.62* -3.12 (0.17) (0.15) (0.20) (0.23) bsg -2.92* -2.28* -27.37* -30.04 (0.97) (0.95) (1.08) (1.54) bix 0.42* 0.44* 0.67* 0.64* (0.04) (0.04) (0.05) (0.04) big 0.17 0.07 2.20* 2.08* (0.12) (0.10) (0.18) (0.21) bgx 2.52* 2.19* 19.01* 23.22* (0.74) (0.73) (0.98) (1.41) bxt 0.60* 0.64* 1.75* 1.15* (0.12) (0.15) (0.22) (0.25) bx -1.04 -27.3 (11.6) (19.8) ax 2.77 23.7* (5.19) (9.4) PX 1.15E-5 5.96E-5 (9.98E-5) (1.43E-4) -x -0.03 0.01 (0.02) (0.03) bit -0.01 -0.07 0.91* 0.49* (0.06) (0.05) (0.09) (0.10) bi -2.23 13.2** (1.84) (7.3) ai -2.56* -13.7 (0.83) (7.3) Pi 1.59E-5 4.4E-5 (1.57E-5) (5.12E-5) yi 0.0078 0.01 (0.0034) (0.01) bgt 0.13 0.06 7.28' 5.90* (0.33) (0.33) (0.64) (0.87) bg -16.3 1.38E+2* (12.3) (64.1) ag -2.79 -1.1OE+2* (5.50) (30.4) pg 8.69E-6 2.71E-4 (1.06E-4) (4.6E-4) yg -0.03 0.04 (0.02) (0.10) bst 1.95* 2.25* -10.59 -8.39* (0.47) (0.53) (0.97) (1.23) bs -74.6 -2.12E+2* (30.5) (93.8) as 54.74 1.75E+2* (13.64) (45.2) PS 3.62E-4 -4.91E-4 (2.65E-4) (6.3E-4) TS -0.06 -0.08 (0.06) (0.014) bdm 0.00016 -3.75E-4"* (0.0002) (0.0002) 189 Response to Relative Price Changes in C,te d'Ivoire: Implicationsfor Export Subsidies and Devaluations Table 18-1. (continued) bdt 0.0016 0.0021* (0.0010) (0.0009) bd -2.92E+3 -2.29E+4* (1.44E+4) (1.17E+4) -2.57E+2 -2.21E+2* (99.91) (84.8) f3d -6.01E-8 4.47E-7* (2.76E-7) (2.24E-7) 0.0012 0.0010* (0.0004) (0.004) bmt -0.11* -0.24* bmt ~~~~~ ~ ~ ~~~~~~~~~~~~~~(0.06) (0.06) bm 2.56E+6* 1.62E+6 (1.08E+6) (1.13E+6) 1.19E+4* 2.92E+4* am (6.20E+3) (6.89E+3) 4.93E-5* 3.12E-5 (2.06E-5) (2.16E-5) -0.05* -0.14* (0.03) (0.03) Objective*N 546.00 530.00 239.34 211.95 N 139.00 139.00 139.00 139.00 Note. Standard errors are in parentheses. *Significant at the 5 percent level. "Significant at the 10 percent level. The low output supply elasticity (0.03) and the is 1.12 for domestic sales and 0.51 for exports, based negative domestic supply elasticity (-0.11) with re- on a similar 10 percent increase in prices. spect to an increase in export prices, suggests that the The supply curve for domestic sales is clearly more increase in exports comes about primarily from a responsive to price that that of exports. In other decrease in sales to the domestic market. Holding words, it takes a higher increment in price to induce domestic prices constant, the increase in the relative an equal increase in output for export goods than for profitability of exporting due to a 40 percent increase domestic goods. Given our framework, we can only in export prices is estimated to lead to roughly a 4 speculate why this is the case. The differences may percent decrease in domestic sales. Since the output reflect higher costs of transportation, higher costs of effect is small, the effect of the increase in export achieving a standard quality for export markets, prices on input demands for labor and intermediate higher transaction costs, or increased risk. goods is also estimated to be small. Consistent with the notion of a fairly flat supply Increases in input prices have the expected effects. curve for domestic sales, the simulations suggest An increase in labor costs reduces labor use, export relatively high elasticities of output, use of labor, and supply, supply to the domestic market, and output. use of intermediate goods with respect to the domes- The own price elasticity for both labor and interme- tic price. As output would have to increase substan- diate goods is close to minus one. Output appears to tially before domestic prices would rise, higher do- respond more in the short run to changes in the cost mestic prices would be associated with a large reduc- of intermediate goods than to changes in the cost of tion in exports. We do not want to place undue labor. emphasis on the results of a change in domestic The third column is included to allow for a com- prices, since we do not believe that they should be parison of our results with other approaches that considered exogenous. assume that domestic prices are exogenous and do Table 18-3 presents quantity responses based on not respond to changes in export or import prices. simulations of the model with joint estimation of The simulations suggest that both the supply curve supply and demand, with non-constant returns to for domestic sales and for exports are responsive to scale and for the sample aggregated over all firms. changes in domestic prices. The own price elasticity Again, the simulations pertain to the short run, with 190 Response to Relatizv Price Changes in Cdte d'Ivoire: Implications for Export Subsidies and Devaluations Table 18-2. Exportingfirns only Supply side only Joint supply and demand Parameter CRTS Non-CRTS CRTS Non-CRTS bsx 7.32* 6.64* 54.29* 35.41* (1.93) (1.98) (1.68) (1.72) bst -1.75* -1.40* -2.39* -3.41 (0.35) (0.36) (0.47) (0.25) bsg -3.90* -3.72* -40.49* -25.93 (1.42) (1.43) (1.26) (1.33) bix 1.04* 0.84* 0.97* 0.66* (0.13) (0.13) (0.11) (0.04) big 0.16 0.08 0.63* 2.20* (0.26) (0.13) (0.41) (0.21) bgx 3.23* 3.27* 31.81* 19.35* (1.13) (1.14) (0.98) (1.20) bxt 0.65* 0.73* 2.16* 1.10* (0.16) (0.19) (0.27) (0.23) bx -14.81* -42.44 (5.56) (46.55) ax 2.07** 16.13* (1.17) (10.16) ,3x 2.14E-6 5.80E-5 (8.68E-5) (1.31E-4) Yx -0.01 0.014 (0.02) (0.029) bit -0.56* -0.44 0.85* 0.65* (0.14) (0.15) (0.15) (0.11) bi -8.80* 15.95 (2.91) (18.74) at -0.68 -1.11E+1* (0.61) (4.10) pi 1.70E-5 3.70E-5 (4.54E-5) (5.20E-5) yi 0.01 0.013 (0.01) (0.012) bgt 0.40 0.31 9.04* 5.39* (0.48) (0.50) (0.93) (0.75) bg -7.99 1.76E+2 (8.30) (1.39E+2) ag -0.77 -9.94E+1* (1.74) (30.12) [Sg 3.06E-5 2.30E-4 (1.30E-4) (3.9E-4) Yg -0.01 0.036 (0.04) (0.088) bst 1.66* 2.13* -12.86 -7.73* (0.47) (0.82) (1.30) (1.06) bs -0.86 -2.33E+2 (20.9) (2.OE+2) as 5.65 1.57E+2* (4.39) (80.29) A3s 2.62E-4 -9.80E-7 (3.27E-4) (3.81E-7) s ~-0.09 -0.0014 (0.09) (0.036) bdm 0.00012* -4.5E-4** (0.0003) (0.0002) 191 Response to Relative Price Chaanges in CMte d'lvoire: Inaplications for Export Subsidies and Devaluations Table 18-2. (continued) bdt 0.00185 0.0030* (0.0015) (0.0008) bd -6.83E+4 -5.12E+4* (4.36E+4) (1.98E+4) atd -5.50E+1 -3.02E+2* (1.38E+2) (80.29) Pd -1.31E-6 -9.80E-7* (8.42E-7) (3.81E-7) Yd 0.0014 0.0014* (0.00062) (0.00036) bmt -0.22* -0.80* (0.04) (0.046) b1m 7.58E+6* 3.23E+6* (2.7E+6) (1.49E+6) am 1.68E+4* 6.61E+3* (4.07E+3) (5.04E+3) Pm 1.46E-4* 6.17E-5* (5.23E-5) (2.86E-5) Ym -0.063* -0.27* (0.019) (0.22) Objective*N 530.00 510.00 276.42 202.60 N 134.00 134.00 134.00 134.00 Note: Standard errors are in parentheses. *Significant at the 5 percent level. 'Significant at the 10 percent level. capital fixed. In these simulations, domestic prices prices are not very responsive to changes in import are permitted to respond to changes in one of the prices, which would shift the domestic demand exogenous prices. In the previous simulations pre- curve. These results are consistent with the fairly flat sented in table 18-2, the increase in exports due to an domestic supply curve, suggested from the simula- increase in export prices resulted from a movement tions of the previous model, and with a fairly inelas- along the export supply curve. As the domestic tic domestic demand curve. prices were kept constant, no shift in the supply Allowing domestic prices to respond reduces the curve for exports took place. Thus, the previous estimated supply responses as theory would predict simulations provide evidence of elasticities of the and as discussed previously. An increase in export supply curve. In contrast, the simulations presented prices would lead to a movement along the firms' in table 18-4 do not yield elasticities of supply and export supplyfunction. This increase in relative prof- for that reason the term elasticities is in quotes. The itability of exports shifts back the domestic supply simulations do present percentage changes in quan- function and because the domestic demand is down- tities with respect to percentage changes in prices, ward sloping leads to an increase in the domestic but the changes are calculated by comparing two price. With a 40 percent increase in export prices, different equilibrium points after allowing domestic domestic prices are estimated to increase by roughly prices to adjust. As a change in domestic price shifts 8 percent. This increase in domestic prices reduces the export supply curve, the two equilibrium points the incentive for firms to substitute away from do- correspond to two different short-run export supply mestic sales and towards exports. While the sign curves. pattern is the same as with the supply side only, the Table 18-3 indicates that domestic prices, in fact, estimated changein exports and domesticoutputare do respond to changes in exogenous prices. Domestic one third of their previous values in table 18-1. The prices are responsive to changes in export prices, same 40 percent increase in export prices leads to a 4 labor costs, and costs of intermediate goods-all percent increase in exports, compared to the 12 per- changes that would shift the domestic supply curve. cent increase in the previous case when domestic With an elasticity of approximately 0.2, a 40 percent prices were not allowed to vary. increase in export prices would be expected to in- The fact that the export and output responses in crease the domestic price by 8 percent. Domestic this simulation are smaller is not an indication of a 192 Response to Relative Price Changes in C6te dIfvoire: Implications for Export Subsidies and Devaluations lack of responsiveness to price change in the manu- combination with an increase in export prices (as facturing sector. As the export supply and domestic would occur under a tariff cum subsidy program). supply functions have exactly the same form in the Given the apparently relatively flat domestic supply two models, one can directly compare the partial curve, increases in import prices that would shift out derivatives of the supply functions with respect to the domestic demand curve have a sizable effect on the output price. The results indicate that the esti- domestic sales and output. Holding all other prices mated slopes of the supply function are actually constant, an increase in import prices of 40 percent is larger in the model where demand and supply are estimated to decrease imports by 12 percent and estimated jointly. Moreover, the evidence has indi- increase domestic salesby 5.4 percent. As the increase cated that domestic prices do respond to changes in in domestic demand holding export prices constant export prices and that firms react in a theoretically makes sales to the domestic market relatively more consistent manner. attractive, exports would fall. The 40 percent increase As before, the net effects on output and the use of in import prices is estimated to lead to a 2.4 percent labor and intermediate goods are small. One would drop in exports. The net effect on output is to raise expect the estimated output response to be greater total output by 3.6 percent. when both export and domestic prices increase than As there is a significant effect on output with the when only export prices are allowed to increase. increase in import prices, the use of labor and inter- However, it appears that the net output response is mediate goods mustgo up to generate the additional too small to measure this effect. production. Expenditure on labor is estimated to go Accounting for a downward sloping domestic de- up by 3.2 percent and expenditure on intermediate mand curve also leads to a much smaller drop in goods by 6.4 percent with the 40 percent increase in output and input use with an increase in input pric- import prices. es. However, the same pattern prevails of output The simulated effect of the tariff cum subsidy pro- being more responsive to changes in prices of int- gram (a simultaneous 40 percent increase in import ermediate goods than to changes in labor costs. In and export prices) was estimated to increase output table 18-3, labor and intermediate goods are esti- by 4 percent. The percentage increase in domestic mated to be substitutes. In table 18-3, the cross-pri- sales is more than double that of exports (52 to 2 ce elasticities were negative, but imprecisely esti- percent). Thus, more of the increased output comes mated. about from an increase in domestic sales rather than Using the model with demand equations as the exports. All finns benefit from the increase in the basis for the simulations allows us to consider the import price of their competing products. Exporting effects of an increase in import prices, alone, or in firms benefit directly from an increase in export Table 18-3. Arc elasticities based on estimates of supply side model (with non-constant returns to scale, sample of all firmis) Change in price International Export Export Domestic Labor goods Change in real quantity (+10%)a (+40%) (+10%) (+10%) (+10%) Exports 0.51 0.34 -0.48 -0.06 -0.17 (0.01) (0.01) (0.04) (0.01) (0.03) Domestic sales -0.14 -0.11 1.12 -0.16 -0.80 (0.01) (0.01) (0.04) (0.01) (0.03) Total output 0.06 0.03 0.62 -0.13 -0.60 (0.01) (0.01) (0.03) (0.01) (0.02) Labor 0.08 0.07 0.87 -0.84 -0.11 (0.01) (0.01) (0.03) (0.01) (0.02) International goods 0.04 0.03 1.09 -0.03 -1.07 (0.01) (0.01) (0.04) (0.01) (0.03) a. The percentage changes in the prices that were used in calculating the arc elasticities are presented at the top of each column. Because the model is nonlinear, the point estimates of the elasticities calculated on the basis of a 10 percent change in price are different from those calculated on the basis of a 40 percent change. b. Standard errors of the simulations are presented in parentheses. 193 Response to Relative Price Changes in Cote d'Ivoire: Implications for Export Subsidies and Devaluations prices, while nonexporting firms benefit only to the nied by an equal decrease in import prices. This is extent that higher export prices increase domestic consistent with the finding of Nogues (1990) that prices. export subsidies failed (in the sense that they did not In drawing inferences about a subsidy program lead to increased exports) in Argentina where there from these results, it is important to note a critical was high import protection, but succeeded in Brazil distinctionbetween an export subsidyand a nominal when they were accompanied by import liberaliza- devaluation. Anominaldevaluationwillalwayslead tion. However, at least in the short run in Cote to an increase in export prices, while export subsidies d'Ivoire, increases in exports do not compensate for will lead to higher export prices only if firms choose the decreases in domestic sales that occur when do- to participate and take advantage of the subsidy mestic producers confront lower prices for compet- program. If, due to lack of credibility of the program ing imports. or high transaction costs, firms do not participate, their export prices will not rise. Only if enough firms The Program and its Stylized Facts participate in the subsidy program to affect the do- mestic price level, will firms that do not export bene- The behavioral relationship estimated and re- fit from the export subsidy program. ported in this paper could be used to predict output Previous evaluations of export subsidies have ap- and export supplyunderthe set of export and import praised their performance on the basis of the extent prices that prevailed with the Ivorian export subsidy to which they lead to an increase in exports. They cum import tariff program. These predictions should paid less attention to net effects on output and em- then be compared to, and evaluated against the ac- ployment. Our simulation results suggest that in the tual changes of exports, imports, and output. An short run a greater increase in exports (4 percent as estimate of the net effect of the program could be opposed to 2 percent), would occur if the export obtained by comparing the predicted supply under subsidy was not accompanied by an import tariff. the actual export and import prices with the pre- Indeed, we ran an additional simulation (not re- dicted supply under the export and import prices ported in table 18-3) that suggested that a still larger that would have prevailed in the absence of the pro- increase in exports (8 percent) would be obtained if gram. Unfortunately, neither complete price nor out- a 40 percent increase in export prices was accompa- put information is available to us at the moment. We Table 18-4. Arc 'elasticities" based on joint estimation of supply and demand (uith non-constant returns to scale, sample of all firms) Change in price Export & International Export Export Import import Import Lab or goods Change in real quantity (+10%)a (+40%) (+10%) (+10%) (+40%) (+10%) (+10%) Exports 0.15 0.10 -0.07 -0.06 0.05 -0.03 -0.05 (0.02) (0.01) (0.01) (0.01) (0.01) (0.01) ((0.02) Domestic sales -0.03 -0.03 0.20 0.16 0.13 -0.02 -0.16 (0.002) (0.01) (0.01) (0.01) (0.01) (0.002) (0.01) Total output 0.03 0.01 0.12 0.09 0.11 -0.02 -0.12 (0.01) (0.004) (0.004) (0.004) (0.01) (0.005) (0.01) Imports 0.01 0.01 -0.04 -0.03 -0.03 0.01 0.04 (0.001) (0.001) (0.002) (0.002) (0.002) (0.003) (0.002) Domestic price 0.19 0.19 0.02 0.01 0.21 0.10 0.72 (0.002) (0.002) (0.001) (0.001 (0.002) (0.004) (0.002) Labor 0.04 0.03 0.10 0.08 0.12 -0.64 0.48 (0.008) (0.006) (0.006) (0.005) (0.007) (0.008) (0.01) International goods 0.01 0.003 0.20 0.16 0.17 0.12 -0.34 (0.01) (0.01) (0.01) (0.01) (0.01) (0.002) (0.01) a. The percentage changes in the prices that were used in calculating the arc elasticities are presented at the top of each column. Because the model is nonlinear, the point estimates of the elasticities calculated on the basis of a 10 percent change in price are different from those calculated on the basis of a 40 percent change. b. Standard errors of the simulations are presented in parenthesis. 194 Response to Relative Price Changes in Cte d'Ivoire: Implications for Export Subsidies and Devaluations therefore do not attempt to evaluate the effectivness their plans for investment. Producers of import sub- of the program. Instead, we only sketch the main stitutes assert that the scheme has had very little elements of the subsidy-tarrif program, present what effect on the prices of imports and, therefore, has not is known about its implementation, and attempt to enhanced their competitiveness. draw a picture of the main changes during its first In the next sub-section we present the evidence two years. We also analyse the likely causes of the regarding the performance of the exporting sector observed pattern, relating it to the elasticities and and of the premium-receiving firms after the imple- simulations reported above. mentation of the program. Since the firm data from the Banque des Donne es Financieres is incomplete The Export Subsidy Cum Import Tax Scheme for 1987 and 1988, we base the inference of changes only on firms for which data is available for pairs of The tariff cum subsidyprogram wasannounced at adjacent years (t and t-1). The annual files are the end of 1985, with the first disbursements taking matched and merged for every two preceding years, place in the middle of 1986. Initially, an export sub- allowing a comparison using identical firms in both sidy at a rate equal to the tariff rate on a similar t and t+1. Using information on the date of creation, product was instituted for three sectors-wood surviving firms are distinguished from "true" entry products, textiles, and agro-processing. The program and also from "true" exiting firms. This distinction is was implemented selectively primarily because of required in order to detect any response of entry into the concern over its budgetary impacts. The gross more rewarded sectors, and exit from less rewarded subsidy rate varied from 10 to 40 percent even within sectors as a response to the change in the trade re- the same subsector. In January, 1988 the export sub- gime. All the relevant nominal values were deflated sidywas expanded to the chemical and rubberindus- by the appropriate price deflators. tries. Other industries such as machinery, electronics, etc. were still excluded. The payments were to be The Experience with the Program: 1986-1987 made within three months, but, in practice, have been paid out with a longer lag which reduces the During 1986-1987 the nominal (CFA) value of value of the subsidy. At the beginning of the pro- exports declined by 11 percent in 1986 and by 1.0 gram, import tariffs were adjusted and increased percent in 1987. However when exports are deflated with the objective of achieving enough revenue to by export prices the results are just the opposite: the finance the export subsidy. volume of exports increased sharplyboth in 1986 and From 1986 to August 1988,52 firms took advantage 1987 (15 and 22 percent respectively). To verify this ofthenewexportincentives.Wehaveidentified most result, we compared it with the volume of exports of these firms in the sample, but only 42 had continu- data obtained from the customs figures. This alterna- ous data for the years since their creation. Most of tive source provides an independent series on the these firms are "veteran' exporters. They are much quantity of exports and its confirms our earlier re- larger than other exporters: in 1987 their mean size sults, suggesting also a significant increase in the (1,020 workers) was more than twice theaverage size quantities exported. This growth took place while of all exporters (438). Their export level (in nominal export prices (in domestic currency) in most sectors CFA) is again almost twice that of other exporting declined, mainly as a result of the large appreciation firms, reflecting not only the size differential but also of the CFA versus the U.S. dollar. Thus the total a much higher propensity to export. The premium volume and the dollar value of exports increased, receiving firms exported on average in 1987 almost though exports values in terms of CFA or French 60 percent of their total output compared to only 28 francs have decreased. percent for non premium receiving firms. The output and export behaviour of the premium No thorough analysis of the impact, or even the firms resemble closely that of the export sector. Dur- potential impact of this scheme, has yet been con- ing the period 1981-1985, the export performance of ducted. (The study by Noel and Gilles (1984) studied the two groups of firms was similar, averaging a vol- the impact of uniform export subsidy and import ume growth of 13 percent annually. During 1986, ho- tariff within a general equilibrium SAM model.) Pre- wever, the premium firms experienced a lower gro- liminary evidence, in the form of firm surveys con- wth of exports compared to other firms (6 versus 15 ducted by the Govermment, indicates that to date the percent). In 1987 the export growth rate was almost scheme has not had its desired effect. Firms currently equal to that of the non premuim firms (23 percent). exporting state that the subsidy has helped them to Theoutputandexportperformance(expansion)in maintain their market share, but has not affected 1986-87 could bebased on three possible effects. First, 195 Response to Relative Price Changes in CMe d'Ivoire: Implicationsfor Export Subsidies and Devaluations the export subsidy could have contributed to this against lower priced imports. Our estimates indic- growth by offsetting to some extent the decline in ate that the introduction of an export subsidy alone export prices. Second, the appreciated CFA would would be insufficient to increase output of the trad- have led to lower imported input prices leading to a able goods sector. The combination of an export sub- shift to the right of the supply curve. Thirdly, the sidy with an import tariff, which comes closer to decline in domestic demand as well as the shift to the mimicking the effects of a devaluation, would serve right of the supply curve, could explain the export to counteract some of the short-run adverse output performance in 1986 and 1987. The model and the effectsofanovervaluedexchangerate.Wecannotsp- simulations presented in the previous section sug- eak to the longer run effects of such a program, to the gest that, in the past, all three of these effects have budgetary implications, or to implementation issues. contributed to growth in real exports. The elasticities The results presented in this paper do not in any reported in tables 18-3 and 18-4 suggest a large way constitute an evaluation of the actual tariff cum sensitivity of exports to domestic prices and a more subsidy program that was adopted in C6te d'Ivoire modest export response to export and input price in 1986. The subsidy program has not resulted in the changes. However, in order to evaluate the exact type of uniform increase in export prices within a relative importance of the three factors more infor- sous-branche that was analyzed in this paper. As of mation is needed about the actual changes of export August 1988, only 52 firms had taken advantage of and input prices as well aggregate demand. the new export incentives. These firms comprise a third of all eligible exporters, but a lower proportion Conclusions of all manufacturing firms within the sous-branches eligible for the subsidy Most of the firms which From a methodological standpoint, two results applied for the subsidy are large, veteran exporters, emerged from this paper. First, the exercise of esti- employing on average more than 1000 workers and mating firms' output supply and input demand exporting more than 60 percent of their output. functions using flexible functional forms was suc- The program had started with long delays in sub- cessful. The estimates satisfied theoretical curvature sidy payments, which may have discouraged firms properties and the price effects were precisely esti- from applying for the subsidy. Even though tariff mated. Second, joint estimation of supply and de- rates on imports were raised, no increases in tariff mand leads to considerably different estimates of revenues could be detected. By the fall of 1988, the export and output supply responses than estimates program's budgetary burden led the government to based on supply relations alone. Simply instrument- stop all payments and the program was de facto ing for endogenous domestic prices may take care of abandoned. However, this program remains a policy simultaneous equations bias, but is not an adequate option under discussion. substitute for modeling of the demand side. Without modelingthe demand side, it is notpossible to obtain Notes estimates of the net effect of changes in export prices on exports and output. Although our specification of 1. Properlyspeaking, errorsinthefirst-stageestimation the demand side was relatively simple, the results (the instrumenting equation) should be taken into account were consistent with theoretical expectations. Fur- in the calculation of the second stage standard errors (see ther work using less restrictive specifications of the for example, Duncan, 1987; Pagan, 1986). demand side would be useful, as would estimations 2. As defined by Diewert (1974) and as stated in Die- based on alternative flexible functional forms. wert and Wales (1987), a flexible functional form of cost Contrary to prior expectations, our results sugg- function is one that would provide a second order diff- est that firms in Cote d'Ivoire do sell more to the erential approximation to an arbitrary twice continuously foreign market when it is more profitable to do so. differentiable cost function that satisfies the linear ho- Exports respond positively to increases in export mogenity in prices property at any point in the admissable prices and negatively to increases in import prices. domain. Flexibility of the functional form of a profit func- However, the fact that exports would be lower if an tion is defined analogously. The expression "restricted" in export subsidy were combined with an import tariff the profit function refers to capital being fixed. is not an argument for introducing an export subsi- 3. We also specified producer supply and input de- dy alone. Firms producing in the tradable goods mand functions derived from the more restrictive sector suffer from an overvalued exchange rate not Cobb-Douglas production function, but didnot obtain rea- only because they would receive a lower price for sonable results. For example, the effect of increases in their exports, but also because they must compete export prices on exports was negative. 196 Response to Relative Price Changes in C6ted'Ivoire: Implications for Export Subsidies and Devaluations 4. See Diewert and Wales (1987). Diewart, W. E., and C. Morrison. (1988). "Export Supply 5. All efforts to estimate models using individual firms and Import Demand Functions: A Production Theory as the unit of observation and replacing the unobserved Approach." in Feenstra (ed.) Empirical Methods for In- firm prices by sous-branche level price indices were unsuc- ternational Trade, MIT Press, Cambridge, Mass. cesful. 'We suspect that measurement error in the price data Diewert, W. E. and T.J. Wales. (1987). "Flexible Functional was responsible for the poor resultas. Forms and Global Curvature Conditions", Econometrica 6. AllmodelswereestimatedusingSAS'sPROCSYSN- 55:43-68. LIN. Duncan, G. (1987). "A Simplified Approach to M-Estima- 7. In the first stage we predicted the probability of each tion with Applications to Two- Stage Estimators", Jour- firm in the sous-branche would export, constructed a Mills nal of Econometrics, 34: 373-389. ratio for each exporting firm, and estimated the models Faini, R., (1988). 'Export Supply, Capacity, and Relative using sous-branche level data with the sum of all individ- Prices", The World Bank, WPS Paper #123. ual firm's Mills ratios. This has theimplicit assumption that Goldstein, N., and M.S. Khan. (1978). "The Supply and the covariance matrix of the errors in the export probability Demand for Exports: A Simultaneous Approach." Re- equations and the second stage model equations is the view of Economics and Statistics, 60: 275-86. same for all the firms. Kohli, U. (1978). "A Gross National Product Function and 8. AllsimulationsweredoneusingSAS'sPROCSIMN- the Derived Demand for Imports and Supply of Ex- LIN. ports." The Canadian Journal of Economics 11:167-182. Lavy, V., J.Newman and P. de Vreyer, 1990. "Export And References Output Supply Functions With Endoggenous Domestic Prices", mimeo, The World Bank, October. Aspe, P., and F. Giavazzi. (1982). "The Short-Run Behavior Lorch, Klaus. (1989). "C6te d'lvoire: Industrial Competi- of Prices and Output in the Exportables Sector." Journal tiveness During Economic Crisis and Ajustment" of International Economics 12:83-93. Mimeo, The World Bank, Washington, D.C. Artus, J. and S. Rosa. (1978). "Relative Price Effects on Milanovic, B. (1986). "Export Incentives and Turkish Export Performance: The Case of Non-Electrical Machin- Manufactured Exports, 1980-1984." World Bank Staff ery," IMF Staff Papers 25, pp. 25-47. Working Paper, No. 660. Balassa, Bela. (1987). "Economic Incentives and Agricul- Ministere de l'Industrie, (1988). "Schema Directeur duDe- tural Exports in Developing Countries." World Bank, veloppement Industriel de la C6te d'lvoire" Republique DRD Discussion Paper DRD250, Washington, D.C. de C6te d'lvoire, (in four volumes). Balassa, B., E. Voloudahis, EP. Fylahtos and Sub Suhtai. Nogues, Julio. (1990). "Experience of Latin America with (1986). "Export Incentives and Export Growth in Devel- Export Subsidies." Weltwirtschaftliches Archiv, Federal oping Countries: An Econometric Investigation." World Republic of Germany, 126, No. 1:97-115. Bank, DRD Discussion Paper DRD159, Washington, Pagan, A. (1986). 'Two Stage and Related Estimators and D.C. (February) Their Applications", Review of Economic Studies, 53: Baumann, Renato, and Braga, Helson C. (1988). "Export 517-538. Financing in LDCs: the Role of Subsidies for Export Zilberfarb,B. (1980)."DomesticDemandPressure,Relative Performance in Brazil." World Development (UK). Prices and the Export Supply Equation: More Empirical 16.821-33. Evidence." Economica 47:443-450. 197 19 Do African Countries Pay More for Imports? Yes AlexanderJ. Yeats The debt crisis and decliningliving standards require the recent industrial organization literature on these careful husbanding of critically scarce foreign ex- issues, see Bresnahan 1989). change in most African countries. But economic the- Institutional factors may prevent developing ory suggests that smaller countries, which import countries from attaining the best terms for imports. from only a few intemational suppliers and cannot Helleiner (1978) argues that restrictive trade prac- support competitive markets and infrastructure, tices, national and international cartels, or some would be likely to pay more, rather than less for countries' lack of countervailing power may work imports. Analysis of import unit values for 1962-87 againsttheefficientfunctioningof intemationalmar- shows that the 20 African former French colonies kets. In addition, antitrust laws are often weak, paid a price premium of 20 to 30 percent on average nonexistent, or unenforceable at the international over other importers for iron and steel imports from level. Similarly, Edwards (1972) documents the ad- France. The losses associated with these adverse verse effects of restrictive inter-firm practices such as prices total approximately $2 billion by 1987. The agreements for the allocation of territorial markets; study also finds that similar price premia (of 20 to 30 pooling and allocation of patents, trademarks, and percent) were paid by former Belgian, Portuguese copyrights; fixing prices and discriminatory pricing; and British colonies in Africa for imports of these allocation of export business shares among firms; products from their former rulers. and establishment of reciprocal, exclusive, or prefer- Development of optimal trade and commercial poli- ential dealing. At the national level, inter-firm cies in developing countries depends crucially on agreements on exports extend not only to the alloca- factors such as whether transnational corporations tion of foreign markets, but even to individual extract excessive profits, or whether multiple sources foreign customers, allocation of specific goods to be of supply will produce lower import prices. This is exported, fixing of prices and levels of bidding on important since the poorest countries must pay the foreign contacts and the selection in advance of the lowest possible prices for imports of industrial firm that will submit the lowest bid. All of these equipment and production inputs required for eco- factors can lead to higher prices than those that nomic growth. However, if market imperfections ex- would prevail under more competitive conditions. ist, or if competition is less vigorous than it mightbe, To determine if such "excess" prices are being some developing countries may pay more than com- charged to African importers, this article first exam- petitive prices for imports, or receive less than com- ines the distribution of import prices paid by petitive prices for exports (for examples of studies developing countries which have highly concen- that have found evidence of this, see Hewett 1974, trated trade with a major exporting country (France) UNCTAD 1975, and Yeats 1978 and, for a review of using extensive time series data on unit values for 198 Do African Countries Pay Morefor Imports? Yes homrogenous goods. These prices are compared with prised 40 to 60 percent of French iron and steel ex- those paid to France by other countries whose imp- ports over the 25 year period. The products and the orts come from more diversified sources and, where 20 African French-associate countries are listed in evidence of "excess" prices are found, the level of eco- table 19-1. nomic costs is quantified. In addition, the analysis Next, the size of any overall price margins that employs correlation and regression tests to account French-associated countries may have paid relative for the influence on relative prices of other economic to other importers was estimated. Unit values, U, of and institutional factors such as the degree of market good i imported by French associates (p (or the com- concentration, or the size of the importing market. parator group, g), were calculated as the total f.o.b. Next, the article examines the pricing policies of oth- value (Vi) over the quantity (Qi). "Price" margins for er European countries (Belgium, Portugal and the each good were compiled as the ratios of unit values- United Kingdom) with theirformercolonies. Itcloses for the French associates to unit values for the with an overall assessment of the implications of comparators: these findings for developing countries' trade and commercial policies and suggest related research that Mi = Vf/Qi /Vig/Qig appears to have a high priority in addressing these issues. To derive an aggregate across goods, Mf,g, the in- dividual goods' margins were weighted by the share The Methodological Approach of the value of each good in the total value of the sampled iron and steel shipments imported by the By comparing various European countries' share associated French countries (VTf) and summed: in the trade of their former colonies with similar data fora control group of developingcountries, Kleiman (19-1) Mfg = (Vf/Qif X Qig/Vig) X Vij/ VTf (1976) develops an index of relative trade concentra- tion.The results suggest that former colonial coun- Equation( 19-1) computes the aggregate price dif- tries' trade with theUnited Kingdom was three times ferential that French associates pay (positive or the normal level for developing countries, for the negative) relative to other countries, weighted by the French associates about eight times as high and for value of shipments to the associates. The results are the Italian, Belgium and Portugese colonies the con- presented for two year time periods in an attempt to centration was even higher. As this might suggest, smooth out the effects of any unrepresentative trade whileFrance was selected as themain focus forstudy values that might influence annual figures. due to its very high trade intensity with its former A second measure of the costs (orbenefits) of these colonies, the analysis will also show that the findings price differentials (Ef) is derived: can be generalized to Belgium, Portugal, the United Kingdom, and, quite probably, other countries. (19-2) Ef = i(Uif - Uig) 'Qf The first step in the empirical analysis was to com- pile annual data on the quantity and value of French where Uif and Uig are unit values for the French exports (free on board -- f.ob.) on a joint product-by- associates and other developing countries (i.e., the country basis from United Nations Series D V/Q terms in equation (19-1) for the imported prod- Commodity Trade Tapes and to compute unit values uct. By multiplying the difference between the unit for every five-digit Standard International Trade values of the two country groups times the quantity Classification (SITC) iron and steel product exported of imports, this equation computes how much more by France from 1962 to 1987. Where more disaggre- (or less) the associated countries pay for their total gated data were not available, similar statistics were imports of the product. These calculations are then drawn for several higher level products (four-digit summed over all iron and steel imports, and are SITC). An effort was made to hold the four-digit expressed in current and present value terms. items to a minimum, however, since their unit values Several points about theexport unit value statistics can be affected by differences in their product mix. employed in this study should be noted. Analyses Several products had to be excluded from further based on unit values must generally be treated with analysis when tests showed they were exported to caution since differences among products grouped in too small a number of countries, or when full 1962-87 the same SITC category, or differences in quality value and quantity data were not available. This left among othenvise homogenous products may be re- :11 four- and five-digit SITC steel products that com- flected as price differences. The statistics on some 199 Do African Countries Pay More for Imports? Yes coated five-digit SlTC steel products, forexample, do with African data: there are major gaps in many of not differentiate between zinc-coated, tinplate and the African countries' data, many of the records are electric-sheet or other similar products. Several pre- United Nations "estimates", and some African coun- vious studies involving five-digit iron and steel tries only report trade data at the three-digit level - products, however, suggest that the overall effects of onlyfour report at the five-digit level employed here. such variations may be minor. In fact, this homoge- While these factors cause major discrepancies in the neity is such that studies by Stigler and Kindahl official import statistics of many African countries (1970), McAllister (1961), and others have used iron (Yeats 1990), the analysis in this study is based on and steel unit values to assess the accuracy of whole- French and other developed countries' offic ial exp- sale price quotafions employed by the United States ort statistics which are not subject to the same mag- Bureau of Labor Statistics. nitude of error as the African data. It is unlikely that An additional factor that could affect the quality of major incentives exist to induce the iron and steel ex- the trade statistics is the invoicing practices of imp- porters studied here to falsify their export vouchers. orters and exporters. Exporters and importers may over- orunderinvoice customs vouchers to reduce ta- The Emrpirical Findings riff liabilities, to evade restrictions on the use of for- eign exchange, or to illegally obtain subsidies (Bha- Initial comparisons of the relative prices paid by gwafi 1967, Sheikh 1974, JNESCO 1974). In some sit- the French associated and other developed and de- uationsitmaybepossibletouncoverevidenceoffal- veloping countries over the 1962-87 period are se invoicingby comparing reported exports with the reported in table 19-2. These statistics show that the partner country's reported imports. For several rea- French-associated countries are paying more for sons this procedure cannot be used satisfactorily their imports than other developed or developing Table 19-1. The value and destination of French iron and steel exports: 1962 to 1987 Share by importing country groups (prcentage)a French Iron and steel exports Developed countries Developing countries of which Sampled All Sampled products Total less products productsb share AU of which. French Latin French Socialist Year (S million) (S million) (percent) countries EEC c EFTAd associates America Asia associates' countries 1962 786.8 461.1 60.0 68.4 48.0 11.4 12.7 4.7 2.9 11.7 6.0 1965 966.4 556.6 57.6 74.7 46.4 11.2 13.3 4.0 3.5 7.8 3.6 1968 1,013.1 561.0 55.3 73.9 48.0 9.7 10.7 3.7 2.0 8.2 6.0 1971 1,532.1 814.0 53.1 77.6 48.8 9.0 10.5 3.3 2.2 7.0 4.4 1974 3,978.5 2,181.6 54.8 73.8 48.4 8.4 11.2 3.1 1.4 7.2 6.8 1977 4,279.3 1,938.3 45.2 68.8 46.4 5.7 12.3 3.4 1.5 8.8 9.3 1980 7,290.0 3,035.2 41.6 69.9 51.7 6.8 14.4 5.0 2.5 7.58 7.5 1983 4,854.1 1,33.9 39.8 69.0 46.4 6.3 15.2 3.3 4.7 6.4 7.4 1986 6,152.5 2,446.5 39.8 75.7 53.0 6.2 12.2 2.7 4.1 4.8 7.1 1987 6,642.7 2,619.0 39.4 76.9 53.8 6.3 11.7 2.4 3.9 3.6 6.6 Notes: a. The developed, developing and socialist country trade shares may not sum to 100 since the destination of some French exports is not shown in the database. b. (SITC numbers in parentheses): iron and steel simple wire excluding rod (677.01); iron and steel plates and sheets of other than high carbon or alloy steel (674.81); bars and rods of other than high carbon or alloy steel (673.21); tubes and pipes of iron other than cast iron (678.3); plates and sheets, less than 3 mm thick, of other than high carbon or alloysteel (674.31); iron and steel simple big sections (673.41); tube and pipe fittings of iron and steel (678.5); wire rod of other than high carbon oralloysteel (673.11); heavy plates and sheets of iron and steel other than high carbon or alloy steel (674.11); medium plates and sheets, 3.5 mm to 4.75 mm thick, of other than tinned plates and sheets (674.21); angles, shapes and sections, less than 80 mm thick, of other than high carbon or alloy steel (673.51). c. European Economic Community (10 member countries). d. European Free Trade Association. e. These countries consist of Algeria, Benin, Burkina Faso, Cameroon, Central African Republic, Chad, Congo, Cte d'lvoire, Gabon, Guinea, Madagascar, Mali, Mauritania, Mauritius, Morocco, Niger, Reunion, Senegal, Togo and Tunisia. The declining importance of these countries as a destination for France's iron and steel exports is due primarily to major reductions in France's share of the associate's total iron and steel imports. An additional factor was that the growth in total import demand in these countries generally lagged well below that of other regions. See appendix table I for statistics on France's share of the associated countries' iron and steel imports. Source Author's calculations based on United Nations Series D Commodity Trade Tapes. 200 Do African Countries Pay Morefor Imports? Yes countries. For the full 26 year period, the unit values of the 20 countries and exceeds the long-term debt of for French-associated countries always exceeded 12 of the associated countries in 1987. those of developed market economy countries (their The average premiums or discounts paid by indi- average premium for this period was approximately vidual associated countries for all imports of 24 percent), and in only one two-year period (1976- sampled iron and steel products are shown in table 77) did the associates' price fall below that for all 19-3. As before, these price comparisons are based other developing countries. The French associates solelyon French export unitvalues. For the full 1962- paid an average premium of 23 percent above the 87 period the individual country premiums average unit value for other developing countries over the close to 27 percent; some of the lowest values were full 1962-87 period. recorded for 1974-77 and 1982-83. The most striking The total cost of the price differential of French-as- point to emerge from table 19-3 is the extreme vari- sociates relative to all other developing countries, ance of average premiums paid both among coun- calculated using equation (19-2), was around $430 tries in any one period and by any one country over million, or, in present value terms, close to $900 mil- time. For example, over the 26 year period these pre- lion over the 1962-87 period. It should be miums averaged 3.1 percent forMorocco, but for Al- remembered, however, that these are losses only geria, Gabon and Mauritania, they were at least 15 ti- from imports of the sampled steel products which mes greater. The structure and size of markets in the- constituted only 39 to 60 percent of French steel and se countries are examined below to indicate if they iron exports. If the same pattern of price premia can suggest possible reasons for these price margins. holds for all iron and steel shipments the present value of the associated losses on imports would be An Analysis of Differences in Unit Values approximately $2 billion. To place this ($2 billion) figure in context, note that it is equivalent to 60 Why are there such major differences between the percent of the total gross international reserves of 18 f.o.b. export unit values for different countries of Table 19-2. Comparative unit values for France's exports of iron and steel products Frenchfo.b. exports to associated countries Premium or discount paid by French-associated countries compared to.developing countries Value Unit value Middle (S thousands) (S) Total a All developed All non-French Latin America East 1962-63 118,446 167.00 37.9 40.5 36.9 26.8 50.6 1964-65 98,593 151.50 27.5 29.8 21.8 20.4 23.5 1966-67 86,042 143.80 24.6 26.8 21.0 21.6 18.9 1968-69 101,180 150.00 28.5 31.3 23.9 32.7 14.2 1970-71 119,695 199.30 29.6 32.6 16.7 13.3 13.0 1972-73 187,362 234.80 23.0 26.9 18.6 22.0 16.6 1974-75 368,537 386.70 18.1 26.4 8.1 16.7 17.2 1976-77 341,378 375.80 13.1 20.4 -3.6 10.2 2.9 1978-79 465,702 496.60 19.5 19.8 26.1 24.1 12.1 1980-81 489,195 581.20 25.4 28.6 20.9 26.3 -11.2 1982-83 350,566 458.30 6.6 8.3 8.6 6.0 -13.8 1984-85 318,623 442.90 17.4 15.7 36.2 34.2 16.8 1986-87 269,537 668.00 40.1 37.0 66.5 54.7 10.9 Net revenue gains or losses ($ thousands) Actual dollar amoumt' 432,199 Present valueC 876,183 a. Excludes the French-associated countries in Africa. b, Calculated as: Adf- (Uf - UO) x Qf where Qf is the quantity of French associated country imports and UO, is the average unit value paid by all otherdevelopingcountries. These values are then summed over the 1962-87 period. c. The present value in 1987 of all annual gains or losses computed as above (Adf) discounted at 8 percent. Note: See Yeats (1989) for similar statistics for each of the four and five-digit SITC products included in these computations. The French-associated-country premium or discount Plbf is calculated as: PiSf - [(Uf- Ug)/Ug] x 100, where Uf is the unit value for the French associates and Ug is the unit value for the comparator group of countries. The premiums (or discounts) are averaged over all sampled iron and steel products. Source. Author's calculations based on United Nations Series D Commodity Trade Tapes. 201 Do African Countries Pay More for Imports? Yes destination? Since these items are generally homoge- tion is generally among the most concentrated of nous in nature, differences in product characteristics industries in developed countries so the potential should have a fairly limited influence on unit values number of exporting firms is limited. Firms head- and, because these are poorer countries, one would quartered in the same exporting nation may have a expect that they would purchase lower price, poorer greater opportunity and tendency to participate in quality goods if any product or quality differentia- cartel arrangements or collusive decisions on foreign tion does exist within the categories. prices. During the 1962-87 period there appear to The relationships of these unit value premia with have been only 3 or 4 French firms producing the seven possible explanatory factors are analyzed sampled steel products for export, and at various through simple pairwise correlations. The concentra- times the links between these companies were rein- tion of import supply among a small number of forced by nationalization. firms, for instance, could lead to oligopolistic pricing The size of the export market or export shipments practices. To determine the extent of this concentra- might also be expected to influence the pattem of tion the share of iron and steel supplies originating relative prices. In a study of United States machinery in the largest, and three largest exporting countries exports, for example, Hufbauer and O'Neill (1973) was measured. Because each country could have find evidence 'that a small importing country pays a many exporting firms, these measures alone do not much higher price for its machinery' (p. 272). Thus, provide direct evidence on the potential for collusive the quantity of each country's iron and steel imports practices and over pricing. In support of the country from France was computed to determine if larger ratios, however, is the fact that iron and steel produc- shipments were associated with lower import prices. Table 19-3. Average unit value differentials paid by French-associated countries relative to other countries: Selected iron and steel imports from France, 1962-87 (percentage) Importing country 2962-63 1964-65 1966-67 1968-69 1970-71 1972-73 1974-75 1976-77 1978-79 1980-81 1982-83 1984-85 1986-87 1962-87 Morocco 14.1 6.9 -2.1 4.7 3.2 3.5 0.3 0.9 10.9 0.9 -14.6 -5.6 17.0 3.1 Togo 17.2 21.0 11.7 2.7 -1.0 -4.2 -5.9 -14.3 -3.0 185 11.0 25.8 69.9 11.5 Burkina Faso 29.6 29.6 37.1 27.8 12.0 21.6 5.6 6.7 -1.0 2.8 -2.8 -1.0 10.7 13.7 Senegal 21.5 22.8 23.2 17.2 9.8 7.2 5.6 -3.2 6.9 12.4 2.3 21.8 52.3 15.4 Reunion 18.2 20.9 20.6 24.2 13.2 9.4 -1.5 -1.5 10.6 14.4 8.1 26.4 55.5 16.8 Cote d'lvoire 28.2 27.8 34.7 28.8 16.0 5.9 8.0 0.5 0.2 40.0 8.0 17.4 36.9 19.4 Madagascar 22.1 31.9 26.5 18.3 12.9 -0.3 8.5 4.2 15.3 35.5 18.4 19.5 40.2 19.5 CentralAfrican Republc 29.8 26.7 28.0 19.9 13.1 11.3 -0.8 1.6 26.8 4.4 5.7 29.0 60.4 19.7 Benin 36.1 22.4 33.4 11.7 1.9 2.8 3.2 -3.5 1.9 24.0 44.0 20.5 79.6 21.4 Mali 28.7 32.0 73.8 57.1 46.8 10.8 7.6 8.8 10.1 9.9 -2.5 -2.4 16.3 22.8 Mauritiusa n.a. -8.2 -17.7 -2.7 21.1 42.2 60.4 66.1 n.a. n.a. n.a. 24.9 n.a. 23.3 Chad 19.7 36.0 30.2 23.8 10.4 26.6 6.7 9.8 18.9 15.9 10.1 34.1 75.5 24.4 Cameroon 30.8 46.2 44.0 34.2 22.7 8.8 18.0 9.2 23.3 38.2 19.4 20.8 78.0 30.3 Congo 27.3 50.4 48.6 20.4 25.9 32.7 10.4 -0.8 22.0 20.6 40.7 46.8 97.0 34.0 Tunisia 15.1 31.3 45.8 48.5 68.4 46.0 42.4 22.1 31.9 18.3 -1.6 15.0 66.0 35.0 Niger 17.4 41.4 14.6 29.0 34.6 15.1 9.6 20.3 41.9 47.7 12.2 73.1 100.5 35.2 Guinea 43.8 55.0 59.2 38.6 45.5 66.0 51.0 29.2 30.8 45.7 49.4 36.6 34.4 45.0 Mauritania 28.3 60.0 49.0 36.3 35.3 35.7 35.0 20.4 62.9 30.8 27.1 48.4 32.6 46.3 Gabon 51.2 49.4 63.8 60.5 58.6 47.5 55.2 5.5 55.9 33.7 28.4 22.4 81.3 47.2 Algeria 77.9 41.0 43.6 50.5 70.8 60.7 33.2 135.9 65.2 58.3 18.1 27.0 22.9 54.2 Weighted average French- associate 37.9 27.5 24.6 28.5 29.6 23.0 18.1 13.1 19.5 25.4 6.6 17.4 40.1 26.9 n.a. - not available a. For some specific years low import volumes precluded computation of a relative unit value. 202 Table 19-4. Correlation between iron and steel relative importunitvalues and indicators of market sructure and effective size: 1968-69 and 1986-87 Market structure variables Market size variables Dummy variabes Share of 3 largest Relative quantity SizeandStructure Relative price No. of trade partners country suppliers of imports Total imports _ Associaed countries Developed countries variable 1968-69 1986-87 1968-69 1986-87 1968-69 1986-87 1968-69 1986-87 1968-69 1986-87 1968-69 1986-87 2968-69 1986-87 Number of trade partners in -0.448 -0.564* selected products Share of 3 largest country 0.384* 0.472* -0.762 -0.569 suppliers of selected products Relativequantity of imports -0.134 -0.708 0.355* 0.711* -0.150 -0510 of selected productsa Total imports (value) -0.17 -0.626* 0.474* 0.831* -0.219* -0.471 0.842* 0.817* Asociated country group 0.064* 0.447* -0.778 -0.671* 0.680* 0.407* -0.216* 0.507* -0.377* -0.753* Developed country group -0.200* -0.663 0.593* 0.5588 -0.441 -0.423* 0.454* -0.717* 0.598* 0.688* GNP per capita -0.287* 0.572* 0.560* 0.727* -0.385 -0.466* 0.575* 0.755* 0.716* 0.820* -0517* -0.745* Q0799* 0.748* Note: The selected products are noted in table 1. An asterisk () indicates statistical significance at the 99 percent confidence level. a Defined as the ratio of iron and steel shipments(in tons) to an individual countru to the average tonnage shipped by France to all trading partners. Source Authores caluculations based on United Nations Series D Commodity Trade Tapes. Do Afican Countries Pay More for Imports? Yes In addition, the absolute size of each nation's total number of trading contacts since this variable was imports from all sources was measured, to indicate significantly and positively correlated with GNP per whether there might be economies of scale associated capita, market size, relative quantities purchased and with larger shipments, or wheth- er French pricing the developed country dummy. policies are different for large export markets where Somewhat surprisingly, although a strong, nega- countervailing power may be influential. The corre- tive and significant association between relative lation of the number of altern- ative (country) prices and both the market size variables is evident suppliers with unit value differenc- es isestimated to in 1986-87, in 1968-69 these associations were lower, determine whether a large variety of contacts and and only significant at the 95 but not the 99 percent potentially greater sources of information on com- confidence level. The indirect effects of size appear petitive prices are related to import prices. Since important in both periods, however, since table 19-4 quality differences in imports of machinery have shows that both these variables are strongly and been found to be positively associated with real in- significantlycorrelated with market structurewhich, come (Hufbauer and O'Neill 1972), the correlation of in turn, influences market prices. each country's GNP per capita with unit value differ- While the correlations between relative prices and ences was estimated. Finally, a dummy variable was these variables are the primary focus of this analy-sis, used to designate transactions between France and some of the intercorrelations between the indepen- another developed country, while a second dummy dent variables are also of interest. For example, com- was used for shipments between France and a forner pared to the developed countries developing nations colonial country. had significantly fewer trade contacts, smaller mar- The results of the correlation analysis are summa- kets and higher concentration ratios, all of which rized for the 1968-69 and 1986-87 periods in table undoubtedly contribute to higher import prices. 19-4. Since 1968-69 France considerably broadened its trade contacts thus providing a larger base and Additional Evidence From Other African Countries range of country characteristics for the price com- parisons than was available earlier. To provide the widest possible interval for the intertemporal com- The previous analyses raise the question of parisonsof correlation results theperiod 1986-87was whether other industrial countries' exports show also selected. Although many of the 1986-87 correla- similar pricing patterns. Fora test of this proposition, tions appear stronger than those for the earlier f.o.b.unitvalueswerecomputedfortheUnitedKing- period, the variables that were significantly corre- dom's exports of major iron and steel products to lated with prices and market structure in 1968-69 had former African colonies (Gambia, Ghana, Kenya, Ni- a similar relationship in 1986-87. geria, Sierra Leone, Sudan, and Tanzania) as well as As shown in the first column of the table, five to all other developing countries. Next, similar com- variables had a significant relationship with relative putations were made for Belgium (for Burundi, French export prices in 1968-69 with all of the vari- Rwanda and Zaire) and Portugal (Angola and Mo- ables being significant in 1986-87. As is the case with zambique). These data were then used to compute industrial country market studies, variables relating the average premia or discount that the Belgian, Por- to market structure are strongly correlated with rela- tuguese or British former colonies paid over the tive prices. Forexample, a highlysignificant positive 1962-87 interval reported in table 19-5. relation (r = 0.384) exists between relative prices and Over the full 1962-87 period the average premia the share of imports controlled by the three largest paid by the former Belgian and French colonies are supplying countries in 1968-69 and the relation was remarkablyclose (23.7 and 23.2 percent, respectively) even stronger (r = 0.472) in 1986-87. A significant while the former British colonies paid a slightly inverseassociationalsoexistsbetweenrelativeprices lower premium of 20.0 percent. The same pricing and the number of trading partner (country) con- pattern emerges during 1962-75 for Portugal's ex- tacts. Thus, those importing countries maintaining ports to its former colonies, but from 1976 on the trade relationswithalargernumberof exporters, and premiums more than tripled and averaged over 120 theoretically benefiting from greater competition percent. It appears that the hostilities in Angola were and information on comparative prices, pay less for a major factor behind this dramatic rise as domestic their exports.Unfortunately, from the view of devel- firms may have employed excess pricing as a means opment policy, there is evidence that the smaller, of transferring resources out of the country. The sta- poor countries may not be able to sustain a larger tistics in table 19-5 are important, however, as they 204 Do African Countries Pay More for Imports? Yes show that the payment of price premia for imports is distribution, service and information systems. widespread among African countries. Smaller African markets and demand also may not be sufficient to support the required domestic distri- Alternative Hypotheses on the Causes of Price bution and service operations of more than one or a Differentials few foreign suppliers. And problems of language, finance or size or operations all could limit active The empirical approach employed in this paper competition.In addition, existing international trans- has close parallels to previous structure-performance port lines may limit African countries' ability to trade studies of industrial countries, which found that with some low cost producers. Most African coun- prices and profits were higher in markets where tries do not have direct access to North American or aggressive competition was absent. A series of other Far Eastern producers and imports from these geo- factors that could influence price margins include: (i) graphic areas may require costly offloading and institutional arrangements at the national level be- reloading in foreign ports in transit. tween governments; (ii) transnational firm linkages A final set of factors that could affect relative prices or special commercial arrangements between enter- derive from the extent, variability and enforcement prises; (iii) factors limiting access to international of government intervention in African markets. For markets due to information, transport, service, or example, taxes, government regulations and cur- marketing and distribution constraints; and (iv) fi- rency controls create incentives to falsify customs nancial risk differences. invoices to transfer foreign exchange abroad. A re- An example of institutional arrangements at the lated question is whether African countries are national level which may adversely influence Afri- somehow riskier than others, which is a function of can import prices is thepracticeof tyingbilateral aid, the reliability of market support systems, demand, so that recipients must use the funds to buy goods and government policies. Exports of durable capital produced in the donor country. Because they are in goods are often fully financed by the exporter with asense"captive importers" the African countries may the f.o.b. export prices reflecting these costs of fi- not be offered prices that would prevail in interna- nance and insurance. If the African countries were tional markets more generally. The bargaining generally considered riskier than most alternative power of Africancountries mayalsobe limitedby the destinations their price premiums may reflect these rules of origin under the Lome' Convention. In order higher finance charges. to qualify for preferential market access when they are exported to Europe, any assembled iron and steel Summary and Policy Implications or other fabricated products must use components produced in the European Community. Similarly, under "reverse preferences", imports from France or Using techniques which have been employed for Britain were admitted into their former colonies at analysis of domestic market performance in indus- tariffs considerably below those paid by other ex- trial countries, this study established that African porters until the early 1970s. Such arrangements countries paid higher import prices for iron and steel reduced competitive pressures on domestic Euro- shipments than did other developing or developed pean firms and allowed them to raise f.ob. export countries. The magnitudes of these excess prices in- prices above those of other (non-African) countries. dicated that they have been an important drain on A second set of factors that can influence relative foreign exchange: the overall premia for 1962 to 1987 prices relate to transnational operations or inter-firn had a present value of close to $1 billion in 1987. If ties which may result in collusive practices. For ex- the same pattem of excess prices applied to all iron ample, subsidiaries of foreign firms maybe required and steel imports (rather than just those products (formally or informally) to purchase from the parent selected) the magnitude of the costs would approxi- companyevenwhenotherinternationaltraderswere mately double. The differentials calculated here are offering goods of equal quality at lower prices (Kre- relative to average unit values for other developing inin 1988). This tie was found to be particularly country importers of French products. If the equiva- strong for Japanese enterprises. Such overpricing lent price differentials were calculated relative to the could, of course, be used as a means of transferring lower-cost suppliers such as Japan, the total losses profits and capital out of Africa. would be even greater. Whether collusive practices Price premia may also reflect weaker African infra- among importers, restrictive government polices or structure for domestic transport, marketing, higher unit costs in the smaller African markets pre- 205 Do African Count7ies Pay Morefor Imports7 Yes vented them from importing steel and iron at these period. As noted, the excess prices margins are fully lower prices, however, is not known. consistent with both economic theory on the func- These figures calculated here relate solely to iron tioning of markets and results from investigations of and steel shipments, and a key question is whether markets where monopoly elements exist. However, such excess price margins also apply to other capital it was not possible within the scope of the current goods imports. There is some tentative evidence in investigation to precisely identify the factors that support of this proposition. Yeats (1978, p. 178) com- were adversely affecting the African countries. pared four-digit SITC product unit values for all Among the possibile factors are: the relatively small French shipments to selected associated and non-as- size of their markets, which could be important given sociated African countries for 1962-69 and found that economies of scale in distribution, financing and in- unit-values for the ex-colonials averaged between 13 surance; the influence of factors that limit access to and 18 percent higher. If this excess price margin competitive suppliers such as tied aid, and estab- applied to all manufactured imports this would lished lines of intemational and domestic transport; mean that the associates could have been over- a lack of information on competitive suppliers; the charged by approximately $25 billion. The fact that use of overpricing to facilitate graft and corruption; trade intensity ratios are lower for most iron and steel or transfer pricing by subsidiaries of foreign firm s in products imported by the French-associated coun- the African countries. Definitive information on the tries than they are for other items (see appendix table relative importance of such factors will require a 19-2) also suggests that the price margins found here detailed analysis of the procurement practices and may actually be less than those of other products for problems of African importers. whichsupplyishandledbyfewercountries. Similar There are several lines that this related research analysis is needed, however, to estimate such mar- might take. First, it would be useful to extend the gins for other goods. procedures developed in this study to other types of While this study established that Africa countries homogenous products (i.e., glass, cement, nonfer- pay higher prices than other countries for iron and rous metals, etc.) to see if further evidence of steel products, a question of key importance is why discriminatory pricing exists for these items. Sec- they do and have done so over such an extended ond, trade intensity and other structural variables Table 19-5. Prenium or discount charged by selected European countries on iron and steel exports to associatedAfrican countries Average premnium or discount charged to associated countries Yazr Belgiuma Franceb Portugac United Kingdomd 1962-63 20.7 36.9 12.7 4.0 1964-65 21.2 21.8 37.3 8.8 1966-67 25.7 21.0 25.6 14.4 1968-69 19.1 23.9 29.9 12.4 1970-71 15.2 16.7 43.7 13.0 1972-73 18.0 18.6 18.7 15.5 1974-75 26.4 8.1 42.9 9.9 1976-77 35.3 -3.6 n.a. 22.5 1978-79 37.0 26.1 n.a. 15.1 1980-81 17.1 20.9 n.a. 19.2 1982-83 25.5 8.6 n.a. 36.5 1984-85 16.0 36.2 n.a. 37.9 1986-87 31.5 66.5 n.a. 53.0 a. Burundi, Rwanda and Zaire. b. See table 19-3 for French-associate countries. c. Angola and Mozambique. From 1976-77 to 1986-87 the premiums on Portugal's exports rose dramatically and averaged over 120 percent. It appears likely that the hostilitiesin Angola were a major factorcausingthelarge increase in premiumsoverthose which prevailed during 1962-63 to 1974-75. d. Gambia, Ghana, Kenya, Nigeria, Sierra Leone, Sudan,Tanzania and Uganda. Note' Based on the four and five digit SITC products listed in table 19-1. The average premium or discount has been calculated relative to the average unit value for each product paid by other developing countries. premiums on Portugal's exports rose dramatically and averaged over 120 percent. It appears likely that the hostilities in Angola were a major factor causing the large increase in Source Author's calculations based on United Nations Series D CommodityTradeTapes. 206 Do African Countries Pay More for Imports? Yes (see appendix) could be computed for a large number countries to test for evidence of monopoly pricing. of bilateral trade flows and the results used to distin- Third, the procedures should be applied to homoge- guish outlier countries which may be subject to nous goods exported from developing countries to oligopoly or monopoly pricing. The procedures used determine if theymaybe receivingless than competi- in this study might then be applied to these specific tive prices for this trade. Appendix. Measures of Trade Concentration This appendix presents summary statistics relat- U.K., the Federal Republic of Gernany, and all devel- ing to market shares, trade intensity ratios and oped and developing countries. indices of import concentration in the French associ- Market structure indices like equations(19- 3) and ated countries' markets. Appendix table 19-1 shows (19-4) have been used extensively in structure-per- the share of France in the associates' total imports of formance studies of industrial countries where they iron and steel products (SITC 67) and all goods for are based on data on individual firms. There is a selected years from 1962-85. Because some of the potential problem in applying these measures to na- associated countries did not report their imports for tional trade data, because a high ratio at the national specific years, France's share and the trade intensity level may conceal a large number of (nafional) com- ratios could not be computed for these years. The peting firms. In OECD countries, however, there are table also gives a trade 'intensity" index (Iij) defined relatively few iron and steel firms (some of which are as the share of country i's (France) exports to associ- nationalized) so this should not be a major problem ate countryj (Xij/Xi) relative to theshare ofj's imports for the current study. (Mj) in world imports net of i's imports (Mw - Mi). Two major points clearly emerge from these indi- That is, ces. First, France has maintained a dominant position in almost all the associated countries' markets (Mau- (19-3) Iij = XijIXi /Mj/Mw - Mi ritius is an exception) although many of the ratios de- clined over the period. The fact that 14 of the 20 cou- The index can take values between zero and infin- ntries have higherbilateral trade ratios for all imports ity with values above one indicating a greater thanforironandsteelin l985suggeststhat"overpric- intensity of trade between two countries than can be ing" may extend beyond this one sector to all goods. accounted for by the countries' importance in world Yeats (1978, table 4, p. 178) shows that the average trade. That is, a value of two would indicate that the unit values for all four-digit SITC products imported intensity of trade between countries was twice as by selected associate countries from France are con- great as what would be expected on the basis of their sistently higher than those of other African countries. share in world trade. Second appendix tablel9- 2 shows that the mar- Appendix tablel9-2 provides statistics on thecon- kets of the ex-colonial countries for iron and steel centration of associate countries' iron and steel imports remain far more concentrated than those of imports from alternative major suppliers. A three- developed or developing countries although the country import concentration ratio (C3j) was market structure indices also are falling. Still, by 1985 computed: the three largest supplying countries controlled 70 percent or more (over 90 percent in the case of Chad (19-4) C3j = (M3j /MTj) x 100 and Reunion) of theassociates' imports. In industrial market studies such very high levels of concentration where M3j is the value of associate country j's iron have consistently been found to be associated with and steel imports from the three largest supplying higher seller prices and profits. countries and MTj is the total value of imports. The Hirschmann concentration index (Hj) was also com- Notes puted, This index may take values ranging from zero to one with the higher numbers indicating more The author is an economist in the International Econom- concentrated markets. Similar statistics have been ics Department of the World Bank. He would like to thank computed for the total imports of Brazil, the U.S., Azita Amjadi for assistance with much of the empirical 207 Do African Countries Pay More for Imports? Yes Table 19-Al. The shares of French exporters in associated countries imports and their bilateral trade intensity indices: 1962 to 1985 Share of France in associates' imports French-associate bilateral (percentage) trade intensity ratioa Country/product group 1962 1965 1970 1975 1980 1985 1962 1965 1970 1975 19801985b Algenia Iron & steel n.a. 60.1 28.0 20.8 12.3 17.4 n.a. 4.35 2.36 2.06 1.04 1.64 All items n.a. 70.4 42.4 33.5 23.2 26.0 na. 11.00 6.24 4.85 3.57 4.41 Benin Iron & steel 71.4 61.2 37.6 n.a. 33.0 25.7 4.43 4.43 3.17 n.a. 2.78 2.57 All items 59.3 54.8 42.2 n.a. 25.2 27.4 9.88 8.56 6.20 n.a. 3.87 4.13 Burkina Faso Iron & steel 83.1 89.1 49.2 64.0 72.5 50.5 5.16 6.46 4.15 6.33 6.12 5.05 All items 52.2 53.9 50.7 43.4 39.3 27.9 8.70 8.42 7.45 6.29 6.05 6.44 Cameroon Iron & steel 78.7 89.4 54.0 58.1 58.2 42.4 4.89 6.48 4.55 5.75 4.91 3.66 All items 54.5 58.1 50.5 46.3 44.7 42.1 9.08 9.08 7.43 6.71 6.88 6.19 Central African Republic Iron & steel 84.3 91.6 59.3 73.3 68.4 81.1 5.24 6.64 5.00 7.25 5 .77 7.01 All items 60.5 60.9 58.4 57.0 60.7 52.7 10.08 9.52 8.59 8.26 9.34 9.95 Chad Iron & steel 91.9 97.1 47.3 52.3 72.5 86.7 5.71 7.04 3.99 5.17 6.12 8.67 All items 53.2 46.4 39.8 40.8 31.0 33.3 8.87 7.25 5.85 5.91 4.77 5.08 Congo Iron & steel 89.8 79.9 55.4 76.4 76.7 44.6 5.58 5.79 4.67 7.56 6.47 4.22 All items 67.7 61.2 55.1 49.7 47.8 45.5 11.28 9.56 8.10 7.20 7.35 7.72 C6te d'lvoire Iron & steel 84.5 76.2 52.7 67.7 63.0 44.9 5.24 5.61 4.45 6.70 5.32 4.25 All items 66.7 62.4 46.2 39.1 40.8 32.1 11.12 9.75 6.79 5.67 6.27 5.44 Gabon Iron & steel 84.1 91.0 69.7 71.0 56.8 65.4 5.22 6.69 5.89 7.02 4.79 6.54 All items 61.9 58.5 56.6 66.9 58.4 54.2 10.32 9.14 8.32 9.70 8.98 9.57 Guinea Iron & steel n.a. n.a. n.a. n.a. 31.1 58.0 n.a. n.a. n.a. n.a. 2.62 5.80 All items n.a. n.a. n.a. n.a. 32.6 32.3 n.a. n.a. n.a. n.a. 5.01 5.34 Madagascar Iron & steel 93.1 88.9 59.1 67.1 45.5 78.2 5.78 6.44 4.98 6.63 3.83 7.40 All items 74.9 62.5 54.7 40.9 37.6 29.5 12.48 9.76 8.04 5.93 5.78 5.00 Mali Iron & steel 90.0 38.7 43.4 72.4 62.3 46.7 6.21 2.80 3.66 7.16 5.25 4.67 All items 39.2 24.1 38.4 34.1 36.3 25.3 6.53 3.77 5.65 4.94 5.58 4.15 Mauritania Iron & steel 97.2 90.5 57.6 78.0 81.1 41.4 6.05 6.56 4.86 7.72 6.84 4.14 All items 72.5 44.4 35.7 42.3 34.6 23.8 12.08 6.94 5.25 6.13 5.32 3.90 Mauritius Iron & steel 4.9 10.1 0.6 3.0 1.6 10.4 0.30 0.73 0.05 0.30 0.14 1.04 All items 4.8 5.7 6.9 8.6 10.7 11.8 0.80 0.89 1.01 1.25 1.65 1.93 Morocco Iron & steel 75.1 73.8 41.8 50.4 31.7 31.2 4.66 5.34 3.53 4.99 2.68 2.95 All items 42.7 38.0 31.0 30.4 24.8 22.8 7.12 5.94 4.56 4.41 3.82 3.86 Niger Iron & steel 95.0 84.6 73.4 73.5 64.6 30.1 5.90 6.13 6.19 7.27 5.45 3.01 ZOB Do African Countries Pay More for Imports? Yes Table 19-Al. (continued) Reunion Iron & steel 92.7 67.7 67.9 80.0 68.7 66.0 5.75 4.90 5.73 7.92 5.80 6.24 All items 68.8 67.6 62.1 62.6 65.3 65.0 11.4 10.5 9.13 9.07 10.05 11.02 Senegal Iron & steel 90.6 90.5 71.5 52.8 71.7 74.1 5.63 6.56 6.03 5.22 6.05 7.41 All items 65.0 53.1 51.2 41.5 34.1 43.2 10.83 8.30 7.52 6.01 5.25 7.08 Togo Iron & steel 51.0 52.2 32.4 30.7 54.8 30.1 3.17 3.78 2.73 3.04 4.62 3.01 All items 33.5 31.2 29.5 35.1 25.0 19.6 5.5 4.88 4.38 5.09 3.85 3.21 Tunisia Iron & steel 70.4 37.3 43.5 59.9 33.5 22.4 4.37 2.70 3.67 5.92 2.83 2.12 All items 52.2 39.0 34.7 34.4 25.2 27.6 8.70 6.09 5.10 4.99 3.88 4.68 a. The index represents the share of France in all exports to the associated country divided by the share of France in world trade (see equation 3). Avalue greaterthan unity indicates a greater intensity of trade than would be expected based on France's share in world trade. b. Because more recent information was not available for Benin, Burkina Faso, Central African Republic, Chad, Gabon, Guinea, Mali, Mauritania, Mauritius, Niger, Senegal and Togo the statistics shown in these columns are for 1983. Because 1985 data were not available for Cameroon the information shown relates to 1986 trade. Source Author's calculations based on United Nations Series D Commodity Trade Tapes. Table 19-A2. Supplier share and concentration indicesfor iron and steel imports: 1962 to 1985 Share of imports from three largest suppliers Import supply concentration indexa (prc*ntage) Country 1962 1965 1970 1975 1980 1985 1962 1965 1970 1975 1980 1985 Algeria 99.3 92.1 57.5 62.3 60.0 57.8 0.98 0.86 0.40 0.49 0.39 0.39 Benin 99.9 98.4 93.0 84.2 81.0 72.3 0.82 0.71 0.55 0.51 0.50 0.47 Burkina Faso 99.3 95.0 87.8 88.2 89.2 71.5 0.92 0.83 0.73 0.73 0.73 0.48 Cameroon 95.8 88.9 80.5 84.4 73.6 75.7 0.82 0.77 0.52 0.62 0.64 0.65 Central African Republic 98.7 97.4 92.6 93.0 93.7 88.9 0.87 0.90 0.69 0.70 0.68 0.69 Chad 99.0 98.0 86.9 97.0 88.4 96.2 0.96 0.92 0.6 0.66 0.74 0.68 Congo 97.7 91.4 77.7 84.6 93.2 70.8 0.90 0.79 0.56 0.69 0.79 0.45 C6te d'lvoire %.9 98.2 81.1 88.2 79.5 85.2 0.87 0.79 0.58 0.74 0.66 0.61 Gabon 97.1 %.2 84.6 87.6 86.8 54.1 0.90 0.90 0.66 0.71 0.65 0.67 Guinea 45.7 91.9 96.9 90.8 83.1 70.5 0.75 0.81 0.74 0.60 0.53 0.49 Madagascar 98.5 95.0 89.7 95.0 87.0 81.9 0.94 0.86 0.66 0.76 0.63 0.61 Mali 99.9 99.7 98.2 94.5 93.8 74.6 0.97 0.71 0.58 0.69 0.72 0.51 Mauritania 99.9 98.1 82.3 94.7 87.8 86.1 0.97 0.85 0.66 0.86 0.81 0.57 Mauritius 76.7 68.4 64.6 72.1 87.7 87.5 0.88 0.88 0.79 0.78 0.84 0.84 Morocco 97.8 94.6 71.5 76.6 81.3 82.7 0.84 0.84 0.53 0.55 0.53 0.49 Niger 99.3 91.0 95.7 88.1 63.4 64.9 0.98 0.82 0.81 0.74 0.58 0.44 Reunion 98.0 97.2 92.1 98.6 96.4 97.6 0.59 0.43 0.44 0.52 0.55 0.70 Senegal 99.0 95.9 92.5 71.0 89.6 83.8 0.92 0.82 0.71 0.54 0.76 0.60 Togo 91.9 90.9 83.0 79.1 88.6 75.3 0.58 0.82 0.54 0.52 0.62 0.46 Tunisia 93.3 73.1 69.4 82.8 78.8 72.3 0.79 0.49 0.52 0.68 0.50 0.45 Brazil 67.4 65.4 67.6 69.7 65.3 64.7 0.41 0.42 0.43 0.46 0.39 0.40 Germany, ed. Rep. 78.2 67.7 64.2 58.7 53.7 48.7 0.51 0.45 0.43 0.39 0.37 0.33 United Kingdom 38.2 43.3 41.1 44.0 49.5 52.6 0.26 0.31 0.34 0.34 0.35 0.38 United States 52.9 63.8 66.8 67.5 63.6 55.9 0.34 0.45 0.48 0.49 0.44 0.38 All developed countries 58.6 49.7 46.6 50.2 46.4 40.2 0.37 0.34 0.33 0.34 0.31 0.30 All developing countries 52.6 53.7 60.7 64.1 59.9 57.4 0.37 0.37 0.43 0.47 0.46 0.47 a. Hirschmann concentration index: Hj = (L )2. Source: Author's calculations based on United Nations Series D Commodity Trade Tapes. 209 Do African Countries Pay More for Imports? Yes analysis and Paul Meo for many helpful comments and Kleiman, Ephraim (1976). 'Trade and the Decline of Colo- suggestions. nialism' Economic Journal, vol. 86 (September), pp. 459- 1989 The International Bank for Reconstruction and De- 480. velopment/The World Bank Kreinin, Mordechai (1988). 'How Closed is the Japanese Market? Additional Evidence," The World Economy, (De- References cember), pp. 529-542. McAllister, Harry (1961). 'Statistical Factors Affecting the Stability of the Wholesale and Consumer Price Indexes," Avramovic, Dragaslov (1978). "Common Fund, Why and in U.S. Congress, Joint Economic Committee, Govern- What Kind," Journal of World Trade Law, 12 (October), pp. ment Price Statistics Hearing. Washington: U.S. Govern- 370-43. ment Printing Office. Bhagwati, Jagdish (1967). "Fiscal Policies, the Faking of Sheikh, Munir (1974). "Underinvoicing of Imports in Paki- Foreign Trade Declarations, and the Balance of Pay- stan," Oxford Bulletin of Economics and Statistics, (Novem- ments," O#ord Bulletin of Economics and Statistics, (Febru- ber). ary). Stigler and Kindahl (1970). The Behavior of Industrial Prices, Bresnahan, Tumothy F. 1989. "Empirical Studies of Indus- (New York. National Bureau of Economic Research). tries with Market Power," in Richard Schmalensee and United Nations Economic and Social Council (UNESCO) Robert D. Willi, eds., Handbook oflndustrial Organization, (1974). International Trade Reconciliation Study, Vol. 11, Amsterdam: North-Holland, pp. 1011-55. (E/CN.3 454), Geneva. Edwards, Corwin (1972). "Barriers to International Com- United Nations Conference on Trade and Development petition: Interfirm CompetitiveBehavior,"inR. Hawkins (UNCTAD) (1975). The Control of Transfer Pricing in and 1. Walter (eds), The United States and International Greece, Geneva: United Nations. Markets, Lexington: D.C. Heath. Yeats, Alexander (1978). "Monopoly Power, Barriers to Helleiner, G. (1978). World Market Imperfections and Devel- Competition, and the Pattern of Price Differentials in oping Countries, Washington: Overseas Development International Trade, Journal of Development Economics, 5 Council. (une), pp. 167-180. Hewett, E.A. (1974). Foreign Trade Prices in the Council for Yeats, Alexander (1989). "Do African Countries Pay More Mutual Economic Assistance, London: Cambridge Univer- for Imports? Yes," Policy Planning and Research Working sity Press. Paper No. 265, Washington: World Bank, September. Hufbauer, G.C. and J.P. O'Neill (1972). "Unit Values of U.S. Yeats, Alexander J. (1990). "On the Accuracy of Economic Machinery Exports," Journal of International Economics, Observations: Do Sub-Saharan Trade Statistics Mean vol. 2, pp. 265-276. Anything?' The World Bank Economic Review 4, no. 2. 210 20 Comments on Trade Policy Ademola Ariyo and C D. Jebuni Ademola Ariyo vice and support of investment projects in primary products, such as cocoa. These three papers address very interesting issues While the authors'efforts deserve commendation, which throw light on the relationship between trade a number of issues need to be addressed, especially and some macroeconomic indices of less developed as they may influence future studies on the subject countries (LDCs) of Africa. Given the potential policy matter. First, the paper attempts to throw light on a relevance of all the papers, my discussion is influ- policy issue affecting a set of producers of a primary enced by this policy orientation. product. Hence, as much as possible, the institution- al factors should be incorporated into the analysis. The Panagariya and Schiff Paper For example, even if a Nash tax framework were to prove beneficial, one is not sure if each (sovereign) The first paper, by Panagariya and Schiff, at- nation wouldbehave inamannerconsistentwith the tempted to provide an insight into the ongoing de- model given. For example, the short-run impact of bate regarding the potential effect of an expansion of changes in trade tax on government revenue may the production of a primary product (i.e. cocoa) on affect a particular governments decisions on cocoa the terms of trade, export revenues and real incomes pricing. of LDCs. In this regard, the simple and simultaneous Some major policy related developments have also effects of an alternative tax (Nash) arrangement, in- been recorded in some of the countries concerned. creased production efficiency and trade liberaliza- For example, Cote d'Ivoire recently put an embargo tion (devaluation) on real income, export and tax on the release of cocoa output into the world market revenues were simulated. While the authors indi- as a means of boosting prices. If this had succeeded, cated that additional studies need to be carried out, it probably would have encouraged others to do like- the main tentative conclusion was that each simula- wise. In addition, Nigeria has totally liberalized co- tion has a differential impact on the chosen macro- coa trade. Under this arrangement, in addition to the economic aggregates. The results also indicate that excise tax charged by the federal government, each there is a high degree of interdependency whereby cocoa producing state within the country is free to the policy actions of a country have significant effects impose additional levies on graded cocoa. Unless on the macroeconomic indices of other countries con- this practice exists in all producing countries, it might cerned. It was therefore the view of the authors that not be possible to accurately determine the trade tax the donor communities should coordinate policy ad- on cocoa in Nigeria to talk less of comparing same 211 Comments on Trade Policy with other countries. This problem is further com- and articulated policies and programs. An initial fo- pounded by the fact that the state-level tax rates are cus on less capital intensive, agro-allied industries not uniform. may ease the demands for the take-off. Following closely is the paper's implicit assump- Some additional observations mayalsobe made as tion that these cocoa producing countries would al- theymayimpact future studiesonthepaper's subject ways continue to export raw cocoa beans or that on- matter. First, the assumption of absence of smuggling ly an insignificant proportion of the total output is convenient for the theoretical analysis. However, could be processed locally. Again, recent develop- smuggling does exist in practice, and suggestions ments in Nigeria suggest the need to modify this aimed at eliminating or significantly reducing the assumption. For example, the Federal Government phenomenon would be of policy interest. Second, of Nigeriahasbanned the exportation of cocoabeans there is a renewed emphasis on intraregional econo- effective from the 1991/92 cocoa season. It is recog- mic cooperation. The possible impact of this on trade nised that representations have been made to the flows, terms of trade and other macroeconomic aggr- effect that existing local processing capacity could egates identified by the authors should be kept in accommodate only about one-third of total output. view. Third, tax rates for 1982 and 1983 were used for However, they are mainly concerned with a phased the analysis based on 1986 data. The authors need to implementation, rather than an abandonment, of the check or confirm that there has not been a significant proposal. It is desirable to anticipate the demonstra- change in the tax profile between 1982 and 1983 or tion effects of this development on other producers beyond, with a view to making appropriate adjust- and its implications on the issues addressed in the ments. paper. There were some conclusions contained in the pa- The Newman, Lavy, Salomon and de Vreyer paper per which may require further elaboration. For ex- ample, the authors indicated that a Nash strategy The paper by Newman, Lavy, Salomon and de could enhance Africa's share of the cocoa market by Vreyer attempts to evaluate the differential impact of reducing that of, say, Malaysia. One is tempted to two possible trade related policies: tariff-cum-sub- argue that Malaysia's emergence and prominence in sidy versus devaluation. The findings suggest that the cocoa trade reflects the positive impact of a relig- the latter would be more beneficial for the economy. iously implemented long term strategic plan. Hence, The reasons adduced for the plausibility of the ob- other countries need similar plans rather than mere served results were indicated by the authors. Among Nash strategy to alter Malaysia's position. Further- these are the inability of (manufacturing) producers more, the authors suggest that any country experi- in Africa to promptly adjust to changes in market encing significant distortions and disequilibria in its signals and the smallness of short run export supply economy should embark on a structural adjustment elasticity. The authors also believe that export subsi- program and then adjust its tax rate to the optimal dies ultimately leave unaffected the profitability of value. The possible non-feasibility of adjusting tax as exporting activities. They however admitted that the suggested had been referred to earlier. With regard to results are sensitive to the assumed endogeneity of the former, all the countries concerned have similar the domestic price and to the inclusion or exclusion macroeconomic problems and virtually all have em- of the demand factor. barked on structural adjustment programs, includ- Although theauthors stated that theessence of the ing devaluation. Hence, the authors could have dis- analysis was not an evaluation of the reasonableness cussed the implication of similar action by all the or otherwise of the tariff-cum-subsidy program in countries rather than a partial analysis that was done C6te d'Ivoire, it is difficult to appreciate the isolation for Ghana only. of the findings from its potential policy relevance. As the authors rightly observed, trade policy inter- This is important given that data based on C6te ventions should be implemented with a view to im- d'Ivoire were used. It is in this light that one would proving long term efficiency of resource allocation. raise some observations regarding the assumptions In addition, it is agreed that the price elasticity of underlying the model employed. In doing this, I demand for cocoa is very low. One then wonders if must state upfront that I had a problem interpreting devaluation and a Nash type tax regime are the best the word "price" in may instances in the paper re- methods for achieving this long term objective. To garding when the word refers to input or sales price my mind, the long run interest of primary producing of either exports or domestic sales. Hence, any mis- nations will be better served by accelerating their interpretation of some statements may be attribut- industrial development efforts through deliberate able to this limitation. 212 Comments on Trade Policy First, the authors should not be surprised that nants of the observed phenomenon, suggesting that manufacturers in C6te d'Ivoire are able to expand other equally important factors might have been left their exports in the short run in response to an in- out. crease in export prices. The behavior is expected of To enhance the ability of African countries to rational economic entities. Furthermore, the nature meaningfullyaddress thisissue, thefollowingobser- of production may be amenable to 'shock'. In addi- vations are noteworthy. First, it should be noted that tion to the diversion of output for domestic sales to prior to and at independence, most of the prominent the export market as the authors found, the possibil- trading houses were overseas affiliates of the home ity exists for some excess capacity which could be based companies and hence, naturally patronized utilized without affecting the quantity of output these companies abroad. Over time, this could have available to the domestic market. become a habit. An issue of interest therefore is an Second, the assumed endogeneity of domestic investigation of the mix of the states of origin of the prices is in order except that the validity rests upon trading houses in developing countries since then. empirical facts. For example, the assumption could Second, developed countries usually have eco- holdi in an open economy only when the import nomic (trade) agreements or treaties with former content forms an insignificant proportion of total colonies as a means of consolidating their market manufacturing inputs. It is probably because of the share. Hence, the nature of trade flows could persist contextual irrelevance of this assumption that the thereafter, especially for public sector transactions. authors found that domestic prices were sensitive to Hence, an assessment of the differential effect of the prices of intermediate inputs, as expected in an treaties on trade between former colonies and non- LDC environment. In fact, some of the results indi- colonies could provide some useful insight. In ad- cate that domestic prices do respond to changes in dition, the private sector in these former colonies is exogenous prices. expected to develop over time and engage mean- Third, the authors had no information on the out- ingfully in international trade. It would thereby be put or export prices of individual firms. Given the expected to make more contacts than the public sec- heterogeneity of the operations of the firms, the re- tor, which is already constrained by the provisions of sults based on aggregate data would have related to some treaties. Hence, a distinction between public no firm in particular, nor to any specific industiy. and private sector activities is recommended. In spite of the above observations, one needs to Third, it may be useful to compare purchases fu- commend the efforts of the authors for the significant nded through external loans or those which are cont- methodological improvements compared to pre- ractor financed with those purchased with internally vious studies. The findings provide useful empirical generated funds. Such analysis may provide some evidence bearing on the possible outcomes of alter- indication of the effect of tied aid or external loans, native policy options. However, as noted earlier, ad- as well as the source of the trading activities, on the ditional work is required to enhance the policy rele- observed results. Finally, the results did not show a vance of the study's findings. discernible trend, with 1982/83 appearing to be to- tally out of tune with other years. The author should The Yeats Paper endeavor to identify the causes of this phenomenon. Assuming that the analysis in the paper validly The paper by Yeats provides another valuable accounted for the observed results, a major policy revelation regarding the outcome of trade relations issue relates to an assessment of the prospects for between the developed economies and their former reversing this undesirable trend. The following sug- colonies. Although the study focussed primarily on gestions are offered. First, additional studies should France and its former colonies, there is reason to be embarked upon as recommended above and also believe that similar results could be expected with by the author in order to identify the major causes of respect to other developed countries and their form- the observed results. The findings should also be er colonies. The possible causes and the estimated widely disseminated, especially by being brought to costs of the observed scenario were well doc- the attention of both the developed countries and umented. In order to avoid a repetitive discussion of their former colonies. what the participants would have known from re- Second, with the current macroeconomic prob- ading the paper, no attempt is made here to discuss lems confrontingmanyLDCs, there isa likelihood of the reasons adduced for the results reported. How- greater reliance on foreign loans and/or aid for exe- ever, as indicated in the paper, the explanatory vari- cuting their recovery or development programs. ables accounted for about 56 percent of the determi- Hence, the problems associated with tied aid or its 213 Comments on Trade Policy equivalent may be intensified. The upward trend in links than with new ones, especially in those coun- the premium as reported on table 2 of the paper may tries that give little room for this practice. be instructive in this regard. It is suggested that loans Finally, many of these former colonies engage the or aid through multilateral sources should be pre- services of inspection agencies. Hence, the author ferred to bilateral arrangements, since the latter will should evaluate the usefulness of the services of most likely confine the recipient to a specified source these inspection agencies. This could be done by of purchase. analysing the differences in the premium paid by Third, there is need to address squarely the prob- those former colonies which engaged the services of lem of transfer pricing and deliberate over-invoicing inspection agents relative to those that did not. A of purchases (especially in the public sector). This similar test could be carried out between the former maybe easier to do between long-established trading colonies and other relevant countries. CD. Jebuni Trade policy is one of the key elements of the struc- The Panagariya and Schiff Paper tural adjustment policies adopted almost everywhere in Africa. The package usually pivots The paper by Panagariya and Schiff addresses this around an export-led growth strategy. The three pa- issue. Their results show that uncoordinated donor pers presented, by concentrating on the visible side policy advice and support of investments projects of the current account and the relevant policies leading to simultaneous expansion of exports of therein, can contribute positively to the process of commodities such as cocoa, could result in a fall in structural adjustment taking place in Africa. export and tax revenues. Balassa's (1979) work on changing comparative Our experience, at least in Ghana between 1985 advantage and some of the references cited in these and 1989, is consistent with this finding. As a result papers have dispelled part of the pessimism concern- of this experience, policymakers are shifting to a pos- ing less developed countries' manufactured exports ition where policywill aim at maximizingexpo- rt re- as a fallacy of composition. The more recent concern, venues from cocoa. This may involve output changes however, is that a simultaneous expansion of pri- or increased processing of cocoa for export. In this mary commodity exports resulting from the adop- connection, the next phase of the authors' paper, in tion of structural adjustment programs may lead to which they plan examining revenue, ma- ximizing declining export revenues and real incomes. strategies will be critical. This exercise would be The papers by Panagariya and Schiff, and New- more useful if it incorporated the possibility of proc- man, Lavy, Salomon and de Vreyer concentrated on essing part of the cocoa produced before exports. policies and/or strategies for increasing export vol- It is strange that all the exercises performed by the umes or revenues and real income from exports. But authors related to output increasing strategies and equally important in terms of the current account of policies. What will be the effect on export revenues the balance of payments are imports. In the face of and real incomes if output fell? In a meeting of the increasingcurrentaccountdeficits, theconcernabout Cocoa Producer Alliance in Accra in March 1990, imports relates to ways of reducing expenditure on consideration was being given to the possibility of imports without adversely affecting domestic pro- controlling supply with a view to increasing export duction and reducingconsumptionbelow levels that earnings. An exercise involving a decrease in cocoa may generate political and social unrest. The usual exports will therefore be useful. policy prescriptions have aimed at reducing imports In simulating the effect of a devaluation on cocoa expenditures through a cut in consumption of im- exports, the authors failed to take into consideration ports either as final goods or immediate imputs. No the institutional arrangements for the marketing of analyses, to my knowledge, have investigated the the product. Cocoa, in Ghana, is purchased by a possibility that considerable foreign exchange sav- marketing board which also fixes the price. Cocoa ings could be made on imports through the prices we production and exports respond to changes in cocoa payforourimports. In that context, Yeats' paperand prices, but their response to devaluation is doubtful investigations have potentiallyuseful policyandbal- (jebuni, Sowa and Tutu 1990). This is because the ance of payments implications. marketing board may not pass on the currency de- 214 Comments on Trade Policy valuation to producers. Failure to take this into con- Francophone countries in using devaluation. The sideration would bias the results for a devaluation. problem for these countries is that a devaluation of Even in cases where exports are not purchased by a the CFA franc will require the cooperation and agree- marketing board but by a middleman, this mecha- ment of the countries in the Zone and France. For an nism has to be taken into consideration. individual country, the alternative it would seem is A broader issue of increasing concern regarding to mimic a devaluation. The authors consider the agricultural exports and real incomes is the domestic possibility of using export subsides and import tar- economy repercussions of such export expansion. I iffs as a policy option. On the basis of the short run will illustrate this using cocoa in Ghana. The stylised responses of manufacturing firms in Cote d'Ivoire to facts to take into consideration include: relative price changes, the authors have sought to 1) cocoa is marketed by the Ghana Cocoa Market- dispel part of the skepticism surrounding this policy ing Board which may pass on part or all of policy option. Their results can make a major contribution changes to cocoa farmers through increased pro- to this debate. ducer prices. However their concentration on short run respo- 2) Prices in the other subsectors of agriculture are nses may still raise questions among the skeptics. As determined by supply and demand in weekly village their results show, in the short run, firms may re- markets and are not directly affected by government spond to increased profitability in exporting by div- policies. erting domesticsales to exports. In thelongrun, there 3) The connection between these village markets willbe a need for capacitybuilding. In the meantime, and the consuming urban markets is through mid- the diversion of domestic sales to exports could affect dlemen or middlewomen, who buy from the village domestic prices and inflation and have repercussions markets and sell in the urban markets and who may for the rest of the economy. Thus, the addition of a not behave as perfect competitors. long run perspective will be illuminating. 4) Cocoa producers' demand for products of the An important issue in the debate is the mode of fin- other agricultural subsectors is zero, and; ancing the export subsidy and how this impacts on 5) the Arthur Lewis unlimited supply of labor is no fiscal and monetary policy and in turn, influences the longer applicable. domestic price level and the ensuing effects. The ex- In these circumstances an increase in cocoa prices port subsidy must by financed and it is not clear that as a result of a devaluation or tax cuts, increases cocoa the additional revenue from the increased import production and exports. With limited labor supply, tariff can do this. This could affect the government cocoa farmers bid up the price of labor in the rural budgetandhavewiderimplicationsfortheeconomy areas. At the same time, because of the nature of price depending on the size of the budgetary effect. determination in other sectors in agriculture, pro- In examining policy options one might wish to ducer prices may not increase even though urban incorporate the reactions of neighbors, particularly prices are rising. The combination of increasing labor when they have similar economic structures. Cur- costs and probably declining terms of trade for the rently, considerable amounts of Ivorian Dinor Oil can other subsectors could cause a decline in their pro- be found in the Ghanaian market. Meanwhile Gha- duction or the rate of growth of output. Loxley (1988) naian producers cannot sell their oil. Even though the and IFAD (1988) have documented this process for Ghanaian government is committed to import liber- the economic recovery period in Ghana. As a result alization, it is coming under increasing pressure to of increasing cocoa prices, agricultural wages dou- do something about the situation. The argument by bled while the terms of trade for the food subsector, Ghanaian manufacturers that these products are sub- the largest subsector of agriculture in Ghana, de- sidized in Cote d'Ivoire is convincing. clined. Further complications could be added if we consider the prices of imported inputs in the other The Yeats Paper sectors and the effects of the polices on urban real inc- omes. Incorporating these considerations will requ- Using techniques of organization (market struc- ire a macroeconomic approach to modelling the rela- ture) on domestic prices, Yeats has demonstrated that tions between commodity exports and real incomes. French-associated African countries pay higher prices for their iron and steel imports. In the 26-year The Newman, Lavy, Salomon and de Vreyer Paper period from 1962 to 1987, French-associated African countries' unit value of iron and steel imports always The paper by Newman, Lavy, Salomon and de exceeded those of developed market economy coun- Vreyer addresses the policy dilemma faced by the tries by an annual average of 24 percent. They paid 215 Comments on Trade Policy an average premium of 23 percent above the unit cost of paying these premia must relate the cheapest value for other developing countries for the same source of these imports. The appropriate unit values period. The present value of the losses associated for this purpose must be based on the cheapest with these premiums could amount to $2 billion if all source unit values of the relevant imports. If this were iron and steel imports by these countries from France done, it would be seen that the present values of are considered. losses presented in the paper will be underestimates The paper shows that this phenomenon is not of the true losses. confined to the French-associated African countries, Earlier analysis had shown that less developed but may also apply to iron and steel imports by other countries pay more for "aid" imports. Thus, there African countries from their former colonial masters. might be two sets of forces operating to generate This fact, in combination with the association of these premia those related to 'aid" and those related these premia with size, number of contracts, etc. to monopolistic practices. It should be noted that the suggests that considerable foreign exchange savings analysis covers relations between some African could be made by bulk orders, diversification of the countries and their former colonial masters. These sources of purchase, etc. Another strategy for reduc- former colonial masters, it can be expected, also ex- ing these premia, not discussed by the author, is the tend the most bilateral aid to these countries. Over- use of international companies forvetting thequality pricing associated with "aid" imports need not be and prices of imports. related to market characteristics. It is essential to The analysis, as the author noted, has to be ex- make this distinction if we are to identify the deter- tended in several directions in order to realize its full minants of the prices paid by African countries for policy potential and implications for balance of pay- their imports. By ignoring this distinction, market ments arrangements. There is a need to extend the characteristics may be conveying influences which analysis to cover other imports and imports from are due to other forces. other countries and also exports by African coun- tries. Such an extension will tackle the questions of References whether this is a general phenomenon for all Africa's imports from all countries orwhether it is peculiar to Balassa, B. (1979), 'The Changing Pattern of Comparative their relations with their former colonial masters. It Advantage in Manufactured Goods", Review of Eco- will also address the question of whether African nomics and Statistics, Vol. 61. export marketing strategy results in losses from re- IFAD (1988), Report of the Special PRogramming Mission ceiving export unit values below what they could to Ghana", Report No. 0105-9H, Rome. have obtained. This is important for maximizing ex- Jebuni, C.D., Tutu, K.A., Sowa, N.K. (1990), 'Exchange Rate port revenues. Policy and Macroeconomic Performance in Ghana". Pa- On methodology, while it is enough for purposes perpresentedat AfricanEconomicResearchConsortium of calculafing the premium or discounts relative to Workshop. May 27-31, Nairobi. other countriesby usingFrench export unit values to Loxley,John, (1988). "Ghana:Economic Crisisand the Long other countries, this may not be appropriate for esti- Road to Recovery". The North-South Institute, Ottawa, mating the costs of such pricing practice. The true Canada. 216 21 Integration Efforts in Sub-Saharan Africa: Failures, Results and Prospects -A Suggested Strategy for Achieving Efficient Integration Ali Mansoor and Andras Inotai The recent economic performance of Sub-Saharan tive because of the inefficient resource allocation that Africa has been unsatisfactory. Over the period 1980 has resulted from the emphasis on regional import to 1985, real GDP declined by almost 1 percent annu- substitution and attempts at regional industrial plan- ally, while population increased by 3.3 percent a year ning. However, the progress with outward oriented on average. Since then, GDP has grown at an average adjustment in Sub-Saharan Africa sets the stage for annual rate of just under 4 percent, thus leaving per and will facilitate the implementation of a new ap- capita income substantially below the level at the proach to regional integration, that will emphasize start of the decade. the complementaritybetween external trade liberali- To restore significant per capita income growth, zation and regional liberalization of factor markets. the 'World Bank has supported reforms aimed at In turn, such regional integration will facilitate and making Sub-Saharan African economies more flex- extend ongoing adjustment efforts by increasing in- ible. By the end of 1989, 33 countries in Sub-Saharan ternal competition and providing greater flexibility Africa had adopted some form of structural adjust- for factors to adjust to changing economic condi- ment program with the support of the World Bank, tions. Selectivity and adaptability will be key con- and 17 countries had accepted the need to liberalize cepts in implementing the new approach. their trade regimes. However, despite these develop- Section two of this paper looks at the performance ments and some encouraging progress in countries of integration schemes in Sub-Saharan Africa and that have implemented strong reform programs,' far among other developing countries. Section three pre- more needs to be done to improve the outlook for sents the rationale for regional integration efforts in Sub-Saharan Africa. Emphasis needs to be given to Sub-Saharan Africa. Section four discusses the scope measures that will strengthen ongoing reform efforts for change and the main elements of a new approach to achieve more rapid and extensive improvements to regional integration inSub-Saharan Africa. Section in efficiency. In practice this requires initiatives in five contains concluding remarks. many directions, including strengthening and deep- ening of the adjustment process through regional Regional Integration in Sub-Saharan Africa and integration and cooperation.2 Among Other Developing Countries It is with these considerations in mind that this paper reviews the performance of various regional The eight major economic integration arrange- groupings formed to promote regional economic in- ments in Sub-Saharan Africa aim at industrial coop- tegration. This experience has been generally nega- eration and trade liberalization in one form or an- 217 Integration Efforts in Sub-Saharan Africa: Failures, Results and Prospects -A SuggestedSbategyforAchievingEffidentIntegration other.3 Their objectives are to promote collective self- The comparisons in table 21-2 suggest that unions reliance and to establish a common market. These in Africa have been less successful than those else- objectives are generally consistent with the Lagos where, except for the Andean Pact, which failed Plan of Action, which was adopted in April 1980. through overemphasis on industrial planning. Sur- Except for the West African Economic Community prisingly, African unions have witnessed more exter- (CEAO), the regional organizations have not nal than internal trade generation and have moved achieved any significant increase in intraregional from being relatively closed in the 1960s to about the economic activity. same degree of openness as unions in other parts of In general the results of integration schemes in the world.6 Sub-Saharan Africa have been more disappointing than elsewhere in terms of generating regional Regional Integration Anong Developing Countries trade.4 Table 21-1 summarizes some basic indicators for the Central American Common Market (CACM), Economic integration among developing coun- the Andean Pact,the Association of South East Asian tries was a major policy issue in the 1960s and the Nations (ASEAN), the Latin American Free Trade early 1970s. Subsequently it appeared to have has lost Area (LAFTA), the European Free Trade Area (EFITA), its relevance. However, despite adverse economic and the European Community. and political conditions, regional integration The groupings in Sub-Saharan Africa have experi- schemes did not dissolve. Their policy relevance is enced the greatest degree of opening up of all unions. indicated by the continuing membership of most This is all the more remarkable given the rhetoric of developing countries (except in Asia) to at least one self-reliance and de-linking from the world trading regional integration scheme. More importantly, in system that underlies much of the official justifica- the last few years new initiatives have appeared. tion for such groupings.5 However, the share of re- Thus, while some countries are considering closer gional trade and intraunion trade creation in Sub-Sa- ties to neighboring economies for strategic rather haran Africa are generally lower than in other re- than economic reasons (Maghreb and other Arab gions. It would appear that the opening up of the regional groupings), many established integration African groupings is more a reflection of the failure schemes are redefining policy priorities and instru- to achieve extensive and efficient import substituting ments. This activity has been partly spurred by re- industrialization than of explicit policy decisions. gionalization trends in the developed market econo- The CACM achieved large increases in regional mies (European Community 1992, US-Canada free trade. It has been the most successful union in the trade pact, Pacific Rim initiatives). developing world, in terms of intraunion trade crea- tion. However, its trade gains were based on an Original motivations. At the end of the 1950s, eco- import substitution strategy in which consumer nomic integration among developing countries was goods were produced at prices exceeding those on a response to the limits of national import substitu- the world market. The inputs for this production tion and the emergence of the European Economic were financed by buoyant commodity export re- Community (EEC). It aimed for import substitution ceipts. When commodity prices collapsed in the late on the regional level through the use of trade policy 1970s and foreign financing became unavailable in instruments such as those being applied by the EEC. the early 1980s, the earlier gains were reversed (see However, the attempt to emulate the EEC did not World Bank 1989c). sufficiently take into account the characteristics of Outside Sub-Saharan Africa, all organizations exc- the members of the EEC. These included the size and ept for ASEAN and the Andean pact generated more level of development of the market, the high share of internal than external trade. In Sub-Saharan Africa intraregional trade before integration, historically the reverse is true. In the case of ASEAN this result is developed production and financial linkages, and not surprising given the negligible emphasis placed the high income and industrialization levels which on intraunion trade liberalization. Despite this, the provided considerable financial resources for com- sh-are of regional trade within ASEAN is relatively pensation payments. More importantly, the EEC high, while ASEAN also displays the greatest degree aimed at achievinggreater regional integration while of openness. Similarly, the CEAO is both the most pursuing, in parallel, external trade liberalization. open group in Sub-Saharan Africa and the one with The commitment of all the members of the EEC to an the largest share of intraregional trade. This suggests increasingly liberal world trading regime, under the that regional exchange maybe favored by openness. auspices of GAIT, ensured that the initial movement 218 Table 21-1. Basic indicators for selected economic and trade groups (percntage) Trade GDP Distribution of GDP in % Share of Net trade diversion regional Degree of gains I1 (-) or gains Area trade 1/ openness (Change in Crude (+) with Growth (thousands (Irade=Exp+ (Tradel degree of trade non-partners US$ bils Avg. anual rates POP (mils) GDP p.c. Manufa of square Group Imp) GNP) opennes) 2/ creation 3/ 3/ 1986 1965-80 1980-86 mid-1986 (USS) Agr 1986 Ind 1986 1986 kms) Outside SSA CACM 2 43 8 8 - 22 4.6 0.0 24 914 23 27 19 423 EC9 53 48 13 7 6 3067 3.7 1.5 264 11629 5 37 20 1527 EC6 n.a. 31 3 5 -2 2508 3.8 1.2 198 12645 3 36 24 1169 EFrA 24 36 -1 2 -3 475 35 2.3 32 15047 5 38 23 1236 Asean 23 58 13 1 12 192 7.8 3.5 295 652 18 33 22 3064 AndeanPact 4 41 9 1 8 120 5.6 0.0 83 1458 16 33 19 4719 Lafta 11 15 -6 - -6 550 5.3 0.1 351 1566 15 35 22 19311 In SSA CEAO 13 57 31 3 28 18 4.1 0.7 47 379 37 22 13 4444 ECOWAS 4 33 30 2 28 78 6.3 -1.8 179 436 40 26 9 6093 PrA 7 41 .18 1 17 33 3.9 2.1 146 227 40 22 14 4973 UDEAC 2 42 35 1 34 18 5.9 6.2 23 805 19 36 5 1708 MRU 1 4 35 1 34 4 3.3 0.2 12 346 41 23 3 429 CEPGL - 47 11 - 11 9 2.4 1.3 43 208 35 31 14 2399 IOC 3 42 n.a n.a. n.a. 7 n.a. 2.5 13 517 n.a. n.a. n.a. 562 Note: I/Data is for 1983.2/For SSA, calculationsare based on differencebetween 1%5 and 1983. For othergroupings, various dates are used corresponding to a period before and after the union became effective. 3/Here trade creation/diversion is measured as the cahnge in rade relative to GDP. A more precise measure should take account of production and demand in a well specified model (see Corado and De Melo (1986). Note that Net Trade Gains (column 4) = Crude Trade Creation (column 5) + Trade Diversion (-) or Gains with non-partners (column 6). Crude Trade Creation is the change in the share of regional trade in GDP following the creation of the Union. Source: Inotai 1986 (rable 1, p. 44), Robson 1987, OECD Various years, and International Monetary Fund 1988. Integration Efforts in Sub-SaharanAfrik. Failures, Results and Prospects -A Suggested StategyforAchievngEffident Integration Table 21-2. Characteristics of trade amongmembers of selected economic integration and cooperation schemes Major Export trade- among related members in Intrascheme exports as a percentage of total exports charact- 1987 (USS Group eristicsa millions) 1970 1975 1980 1983 1985 1987 Central American Common Market 1,2,7 492 27 23 22 22 15 12 Andean Group 1,2,7 683 3 5 3b 4 3 3 Caribbean Community 1 323 7 7 6 9 8 6 UDEAC in Central Africa 1,2,4 38 3 4 4 2 8 6 W. Afiican Economic Comm. (CEAO) 2,3,4 383 9 7 7 13 9 688 East African Common Market 5 142 17 13 8 7 7 7 Economic Community of West Afrcian States (ECOWAS) 885 2 3 4 4 4 6 RCD (Iran, Pakistan, Turkey) 1,305 1 1 5b 9 10 5 Latin American Integration Association 6 8,103 10 14 14 11 9 11 Association of Southeast Asian Nations 6,7 14,529 15 16 18 23 17 218 Memorandum item: European Economic Community 555,616 49 49 53 53 55 59 Note: 1 =Free trade among members; 1 =common extemal tariff; 3=redistribution of proceeds from tariff to settle payments imbalances among members; 4=common currencv 5=now defunct; 6=some preferential trade treatment among members; 7=joint positions international trade negotiations.b 1981. chis total and the shares do not include Singapore's very large exports to Indonesia, which are not reported by mutual agreement. Source: Inotai 1986 (rable 1, p. 44), Robson 1987, OECD Various years, and International Monetary Fund 1988. to a common external tariff involved, in general, a Additional constraints emerged from the imple- lowering of tariffs. mentation of trade policy instruments. Originally scheduled tariff reductions did not take place or re- Performance. As a consequence of the "fallacy of quiredtime-consumingprocedures. Asliberalization transposition" (see Laghammer and Hiemenz 1989), approached the highly protected domestically pro- overambitious goals were formulated. Regional inte- duced goods, the liberalization process was stopped gration was expected to become an engine of growth and in some cases reversed. The integration schemes which would increase international competitiveness failed to generate the benefits that could have been by making use of the advantages of specialization, expected from exposing the highly inefficient na- larger economies of scale, enhanced regional compe- tional import substitution industriesto regional com- tition, and a regional training ground. It appeared to petition. offer collective self-reliance and decreased depend- Larger and more developed countries became the ency on the developed world. main beneficiaries of the regional market while Instead, the generally low level and modest share smaller and less developed economies registered of intraregional trade, accompanied by rather mod- growing trade deficits. Payments arrangements and est export-to-GDP ratios in various countries, failed preferred treatment status granted to deficit econo- to alter the determining role of extraregional trade in mies provided at best a temporary remedy. Particu- growth (see tables 21-1 and 21-2). The regional mar- larly in regional integration schemes involving coun- ket reamined too small to secure significant econo- tries with inconvertible national currencies, this mies of scale. This was compounded by the tendency problem often proved disruptive and contributed to to consider integration of markets by product the reduction of intraregional trade volumes. Trade groups. Progress was further handicapped by weak liberalization was not accompanied by factor market (if any) industrial structures, thelackof intraindustry liberalization, and thus impeded the flow of capital linkages, and usually nonexistent or underdevel- from surplus countries to deficit countries to offset oped infrastructure. trade imbalances. 220 Integration Efforts in Sub-SaharanAfica Failures, Results and Prospects -A SuggestedStratgyforAchivlngEffidentIntegration Tabl,e 21-3. Gross domestic product, 1987 (US$ billions) Afria region Asia region EMENA region LAC region Lesotho 0.3 Bhutan 0.3 Yemen PDR 0-8 Haiti 2.3 Mauritania 0.8 Laos, PDR 0.7 Yemen Arab Republic 4.3 jamacia 2.9 Sierra Leone 0.9 Nepal 12.6 Jordan 4-3 Nicaragua 32 Chad 1.0 Papau New Guinea 3.0 Oman 8.2 Honduras Liberia 1.0 Sri Lanka 6.0 Tunisia 8.5 Trinidad and Tobago 4-3 Central African Rep. 1-0 Bangledesh 17.6 Morocco 16.8 Costa Rica 43 Malawi 1-1 Malaysia 312 Kuwait 17.9 Bolivia 4.5 Burundi 1.2 philippines 34.6 United Arab Emirates 23,7 Paraguay 4.6 Togo 1.2 Thailand 48-2 Syrian Arab Republic 24-0 El Salvador 4.8 Mauritius 1.5 Indonesia 69.7 Hungar 26.1 Dominican Republic 4.9 Mozambique 1.5 Korea, Republic 121.3 Pakistan 31.7 Panama 5.5 Botswana 1.5 India 220.8 Portugal 34.3 Uruguay 6.4 Benin 1.6 China 293.4 Egypt, Arab Republic 34.5 Guatemala 7.0 Burkina Faso 1.7 Yugoslavia 60,0 Ecuador 10.6 Somalia 1.9 Turkey 60.8 Chile 19.0 Mali 2.0 Algeria 64.6 Columbia 31.9 Zambia 2.0 Saudi Arabia 71.5 Peru 45.2 Madagascar 2.1 Venezuela 49.6 Rwanda 2.1 Argentina 71.5 Congo, People's Rep. 2.2 Mexico 141.9 Niger 2.2 Brazil 299.2 Tanzania 3.1 Gabon 3.5 Uganda 3.6 Senegal 4-7 Ethiopia 4.8 Ghana 5.1 Zimbabwe 5.2 Zaire 5.8 Kenya 6.9 C6te d'lvoire 7.7 Sudan 8.2 Cameroon 12.7 Nigeria 24.4 Subtotal 126.1 849.4 491.7 727.0 Percentage to grand 6.0 39.0 22.0 33.0 total Grand total 2,194.2 Source: World Bank 1989b, Table 3. 221 Integration Efforts in Sub-Saharan Afica Failures, Results and Prospects -A Suggested Strategy forAchueving Efficnt Integration Some groups made an attempt at joint industriali- ures (especially in Sub-Saharan Africa), it is even zation. However, those sectors selected as integration harder to find evidence of gains. industries were capital and technology intensive, A less homogeneous picture is provided by the sh- which further distorted the production pattern of the are of member countries in intraregional trade. Sm- member countries and resulted in inefficient alloca- aller, landlocked economies generally show figures tion of scarce imports and investment resources. This above average, with intraregional trade shares of be- policy could not prevent redistributional disputes; it tween 30 and 60 percent of total trade. There is a mar- merely transferred the problem to another, politically ked difference between intraregional trade and extra- highly sensitive area. The distribution of industries regional trade, with manufactured goods and nontr- was based on political viability, thus contribufing to aditional products accounting for a larger share in in- the inefficient location of investments. traregional trade. Different manufactured goods are In the EEC, the high share of intraregional trade sold in the regional and international markets. In mo- required the introduction of a common external tariff st countries, goods based on domestic resources and to prevent differences in competitive position arising labor intensive consumer products are oriented to from different national tariffs towards third coun- the international market, while capital intensive go- tries. In economic integration among developing ods and products based on imported inputs have a countries, given the modest intraregional share of relatively higher intraregional share in total expor ts. total trade, the common extemal tariff was mainly required to support the common industrialization Response to crisis. By the mid-1970s, fundamental policybyprovidingprotectionto the new integration changes in the international economic environment industries. Thus, a generalized decrease in tariffs was had exerted a decisive impact on economic integra- precluded and instead the common external tariff of- tion among developing countries. First, the com- ten resulted in a substantial tariff increase for the rel- modity price boom and subsequent terms of trade atively most open and developed economies of the changes had a differential impact on the member group. In this way, it increased the level of protection countries of the same integration area according to in the member countries that were relatively most their export and import pattern. Second, adjustment integrated into the world economy. Consequently, to the new realities required an export oriented de- the common external tariff became an instrument ag- velopment strategy as opposed to regional integra- ainst later trade liberalization (e.g. the CACM). An tion based on regional import substitution. additional shortcoming of a high common external Large trade deficits and growing debt burdens tariff is that, within the integration area, it protects pressured countries to increase exports to earn con- multinational companies that settled down in the vertible currencies, and curtail domestic investment period of national import substitution. These become and the ability to import inputs for processing into the main beneficiaries of enhanced protection and regional exports. The result was a relative decline in can be expected to oppose later most favored nation the share of regional trade. Exports were diverted tariff reductions. from intraregional to extraregional markets8 re- duced investment had an adverse impact on prod- Evaluation of trade performance. According to cus- uction capacities and infrastructural projects; and toms union theory, the success of regional integration reduced import capacity led to concentration on 'es- schemes can be measured by the share of intrare- sential items" generally only available from outside gional trade in total trade. However, a growing in- the region. Large and competing devaluations add- traregional trade share can only be seen as a positive ed to the shift of exports to extemal markets. The development if it is accompanied by increasing rela- previously achieved level of regional trade policy tive weight of the region in world trade. The Latin coordination was diluted by "temporary" reintro- American Integration Association (LAIA) and duction of previously lifted quantitative restrictions, ASEAN are the only regions with intraregional trade slowdown or stopping of originally agreed trade flows of some international importance-above 8 liberalization schemes, implementation of "adminis- billion dollars (see table 21-2). Intraregional trade tered trade', and increased tariffs in intraregional reached more than 10 per cent of total trade only in trade. ASEAN,7 the Central American Common Market, Member countries with substantial intraregional and LAIA. However, intraregional trade shares show trade surpluses became unwilling (or unable) to fi- a stagnating or decreasing trend, between 1975 and nance the deficits of other members. At the same 1987. If one takes account of unrecorded trade, which time, regional imbalances were considered as secon- in some cases surpasses the officially registered fig- dary issues compared to the servicing of intema- 222 Integration Efforts in Sub-Saharan Africa: Failures, Results and Prospects -A Suggested Strategy forAchieving Efficent Integration tional debts. Economic integration among develop- national direct investment Japanese investment in ing countries has not provided the capital, technol- ASEAN). Another reason is the generally higher eco- ogy, and human skill that has become increasingly nomic growth characteristic of world market ori- important in the international division of labor. It has ented economies, which results in higher demand for not taken accountof the necessityof connecting trade domestic and regional goods. Additionally, export liberalization with factor market liberalization. And oriented economies generally have a lower debt ratio it has not provided the rapid and flexible response and more liquidity.Therefore tradingwith them may required due to institutional bottlenecks, protracted be easier and will not require the use of different decision-making processes and constraints imposed compensation mechanisms, strict bilateralism or by national interests in the search for a harmonized other trade minimizing instruments. Finally, adjust- regional response. ment to the international economy usually fosters Countries following an export oriented path of policy coordination among participating countries. development were able to carry out a more rapid Once their national currency becomes convertible, as adjustmentwith fewerlosses than were inward look- a result of the adjustment policy, a major barrier to ing economies. In regional integration schemes with intraregional trade can be eliminated. less emphasis on regional import substitution (ASEAN), member countries could adjust more eas- Results From Ongoing Integration Efforts in ily to changing world economic conditions than in Sub-Saharan Africa integration schemes based on the philosophy and practice of import substitution (Latin America). The performance and results of each of the main New, orientations. As a result of the past failures and ongoing integration efforts currently underway in the newly emerging consensus on the importance of Sub-Saharan Africa are reviewed in the World Bank an outward orientation, the pattern of regional integ- report, "Intra Regional Trade in Sub-Saharan Africa" rationhasbeguntochange.Regional integration sch- (World Bank 1989d). The major themes which emes are becoming more flexible and selective. They emerge from this review are similar to those for allow increasingly for optional measures to be imple- economic integration among developing countries in mentedbyasubgroupwithgreatercomplementarity other regions. The only relatively successful union, of interests.9 Cautious member countries, which pre- the CEAO, is characterized by currency convert- iously blocked development, may be bypassed. ibilityand capital mobility. These features are shared Member countries that are more prepared to further only by IJDEAC, which has been one of the least integrate into the international economy may under- successful unions. The CEAO is also the only union take more substantial steps than others. In this spirit, with significant labor mobility. Thus it would appear the green light has been given to establishing indivi- that currency convertibilityand capital mobility may dual relations with nonmember countries. Similarly, be necessary, but not sufficient, for success. And it some countries have lowered their tariff level below would appear that it may be necessary to implement the common external tariff imposed by the integra- extensive labor mobility to achieve successful inte- tion scheme (Costa Rica in CACM). gration and enhanced intraunion trade. The earlier trade policy approach of regional inte- The erection of a common external tariff around gration has given way to other approaches'0 and the any of the major unions is likely to lead to some trade policy instruments which are still applied have diversion of recorded African trade, in addition to experienced fundamental modifications. In overall diverting trade from the rest of the world. Except for trade strategy, regional protection and "training the CEAO, none of the Unions has been particularly ground" arguments favoring gradual global com- successful in creating intraunion trade or achieving petitiveness through the regional market have been a large overall share of intraunion trade. However, replaced by the viewpoint that regional competitive- recorded intra-African trade is not necessarily low." ness is the outcome and consequence of global com- Of the top 10 largest economies in Sub-Saharan Af- petitiveness. Liberalization of national trade policies rica, only C6te d'Ivoire and Kenya have shares of has enabled industries (enterprises) exporting to the intra-Sub-Saharan Africa trade above 6U percent. extraregional market to increase their intraregional But 17 out of 39 Sub-Saharan African countries for exports as well (ASEAN). One reason for this is the which data was available had more than 10 percent comparative advantage provided by global speciali- of their recorded trade with Sub-Saharan African zation. World market orientation offers large room partners. Seven of these countries achieved a share of for making use of intraindustry linkages which can at least 20 percent. The smallestunions, theMRUand be supported by the increasing participation of inter- CEPGL, have little relevance for the promotion of 223 Integration Efforts in Sub-Saharan Africa= Failures, Results and Prospects -A Suggested StrategyforAchieving Efficient Integration intra-African trade. Unless a compelling case can be Benefits from integration would, therefore, arise made for other functions, it is unlikely that their from increasing competition within Sub-Saharan Af- continued existence could be justified. rica by liberalizing factor and goods and services There may be a positive relationship between de- flows across African boundaries, rather than activat- gree of openness and the share of intraregional trade. ing customs unions that would grant high external This would fit in with the observations in the Carib- protection against third parties. Without economic bean and ASEAN. However, while this appears to be integration, Sub-Saharan Africa may improve per- true in comparing the performance of the different formance by moving to an outward orientation, but unions in Sub-Saharan Africa, the opposite is ob- it will fail to seize the benefits of horizontal and served at the country level. This suggests an inverse vertical integration available to larger economies. relationship between the degree of openness and the Thus, for example, by pursuing outward oriented share of Sub-Saharan African trade. adjustment during the 1980s, Mauritius achieved Most of the countries with substantial shares of high rates of growth and eliminated unemployment. intra-Sub-Saharan African trade are located in West However, future growth is handicapped by lack of Africa. None of the Southern African countries has a access to the labor and raw materials readily avail- share of intra-African trade exceeding 10 percent. In able in neighboring countries. East and Central Africa, except for Rwanda and The proposed approach to regional integration as- Uganda, all the countries have less than 15 percent of sumes, as argued by Balassa (1979), that intraregional their trade accounted for by regional partners. It is trade is not in itself more valuable than other inter- possible that currency convertibility explains part of national trade. It requires an outward oriented strat- these variations, since seven of the ten West African egy that "provides incentives which are neutral be- nations with more than a ten percent share of intra- tween production for the domestic market and expo- African trade are members of the Franc Zone. rts' (World Bank 1987, page 8). Experience shows The landlocked countries (which tend to be more that "the important lesson is that the strongly inward closed) are those with the largest shares of intrare- oriented economies did badly" (World Bank 1987, gional trade. Of the 11 landlocked countries for page 8); the validity of this lesson is not affected by which data are available, all but three have at least a the size of the economy (Lachler 1989). Table 21-3 12 percent regional trade share. It is possible that presents data on gross domestic product by country re-exports by neighbors might be inappropriately and by region for developing countries in Sub-Saha- classified as exports from these countries. This would ran Africa; Asia; Europe, the Middle East and North artificially inflate the recorded share of regional trade Africa; and Latin America. This could be considered for these countries. Zimbabwe, Zambia, and the Cen- a crude indicator of domestic and regional market tral African Republic have unusually low shares of size. regional trade for landlocked countries, which may The emergence of a single European market in suggest the presence of large unrecorded trade flows. 1992, the US-Canada Free Trade Pact, and sugges- This is especially true for Zambia, which has a very tionsforaFreeTradeAgreementbetweentheUSand open economy. Pacific rim countries reflect concern that generalized concessions within the GATT may be harder to ach- The Rationale for Integration Efforts in ieve in the futurecompared to thepast. The emergen- Sub-Saharan Africa ce of large blocs may help mitigate the free rider pro- blem and thus facilitate the movement to ever freer Does Regional Integration in Sub-Saharan Africa Make international trade. Sub-Saharan Africa as a whole Economic Sense? has the economic size of a small industrialized coun- try (such as the Netherlands) and so the relevance of The above review suggests that the failure of past a Sub-Saharan Africa bloc should notbe overempha- models of economic integration was based on import sized. However, to the extent that internal liberaliza- substitution behind regional barriers. This is consis- tion within Sub-Saharan Africa facilitates a lowering tent with Lachler's (1989) evaluation of growth per- of extemal barriers, this reduction in extemal protec- formance, which determined that for a given level of tion could be tumed to credit from the GAIT. external protection, achieving a greater degree of intraregional openness will yield significant eco- Regional Integration and Trade Liberalization nomic benefits. Lachler's findings can be explained by the efficiency gains that arise from intraregional After a decade of structural adjustment efforts in factor mobility and competition. Sub-Saharan Africa, progress with trade liberaliza- Z24 Integration Efforts in Sub-Saharan Africa Failures, Results and Prospects -A Suggested Strategy forAchieving Effident Integration tion has been slow. In general, trade regimes in Sub- forcing competition first with firms that are of com- Saharan Africa remain restrictive. Less than one third parable levels of (in)efficiency. This would allow a of the countries have eliminated all quantitative re- reduction of costs, through mergers, acquisitions and strictions (import licensing and/or foreign exchange takeovers, that maybe significant enough to facilitate allocation systems).12 African governments seem to survival in the world market. Such an opening of be aiming at average protection of 30 to 40 percent factor and product markets within a program of compared with less than 10 percent in industrial general liberalization may induce investment that countries (World Bank 1987, page 136). would not be forthcoming in a purely national con- A regional approach may offera new dimension to text. Under efficient management, possibly from a supplement the unilateral and uncoordinated na- neighboringcountty, the restructured firm will have tionai efforts of the sort currently being engaged in a large enough internal demand to eliminate "waste- with World Bank and IMF support. Binding reforms ful overcapacity'.13 through an international treaty will make it harder A similar approach has been advocated for the for opponents to challenge than national legislation. Central American Common Market (CACM). Until According to Nellis (1986) and Steel and Evans 1980 the CACM was the most successful regional (1984), African countries have attempted to industri- grouping, in terms of generating intraunion trade, in alize by relying on the public sector to set up opera- the developing world (see table 21-2). Thus, it is tions geared at small national markets. As a result, argued that "the removal of intraregional trade bar- many public enterprises in Africa are inefficient, riers in this context would enable efficiency gains badly managed, and overdimensioned. "The persist- Moreover, the reduction in intraregional trade barn- ence of capacity underutilization... suggests that a ers, as recommended here, would be accompanied temporary shortage of foreign exchange is not the by lower extra-regional trade barriers. This would fundamental problem. The underlying problems are also reduce the danger of trade diversion... These the dependence of production on imported rather arguments present a case in favor of achieving re- than domestically produced inputs, and excessive gional reintegration in the context of an overall free growth of production relative to the growth of im- trade environment, but not as an alternative to exter- port capacity-and in many cases relative to the size nal trade liberalization since Central America, even of the market" (Steel and Evans 1984, page 55). if fully integrated, would still be better off by liberal- Steel and Evans (1984) also note that high effective izing trade with the rest of the world." (World Bank protection encourages new investment, even with 1989c, pages iv & v). unused capacity. "High cost operation is an inevita- Thus, the proposed strategyemphasizes extension ble consequence of these problemsof capacity under- of the size of all product and factor markets as a utilization, inadequate infrastructure, declining pro- means of increasing competition and efficiency, the- ductivity, and excessive capital intensity. At the same reby facilitating a lowering of external barriers. This time, incentives to reduce costs have been blunted by contrasts with traditional approaches that empha- high effective protection, import prohibitions, re- size the extension of the market for selected products stricited competition, and administered pricing sys- to exploit the economies of scale perceived to be tems ... As a result, a large share of industrial produc- necessary for import substitution behind high barri- tion in Africa takes place at costs that are not com- ers. Economies of scale tend to be more relevant for petitive in termns of world market prices, and it is not heavy industry. They are irrelevant for agricultural uncommon to find some firms that actually use more goods and for services such as marketing and trans- foreign exchange than they save" (Steel and Evans port where significant increases in economic activity 1984, pages 59-60). This supports Nellis' (1986) view could be expected in Sub-Saharan Africa. Freeing that making firms competitive requires competition. agricultural trade is particularly important, both to However, he notes that in "most African countries support national reforms to shift the terms of trade internal markets are so small that at least large manu- in favorof the agricultural sector and to enhance food facturing firms frequently acquire automatically a security. As noted by the World Development Report monopolistic or oligopolistic position." (Nellis 1986, 1987 (World Bank 1987), "even large economies, if cut page 44). A regional approach may be politically off from international trade, would lack stimuli for more acceptable, not only by building on the pan-Af- efficient industrial development." (page 3). This ex- rican sentiment of the continent, but also by reducing plains why, despite occasional temptation to rely on the cost of adjustment. scale economies for achieving import substituting in- A program of rapid liberalization with regional dustrialization, Brazil has tended to have an outward partners may diminish the costs of adjustment by orientation (World Bank 1987, Figure 5.1, page 83). 225 Integration Efforts in Sub-Saharan Africa. Failures, Results and Prospects -A Suggested Strategy for Achieving Effident Integration None of the existing subregional groupings in Sub- amounted to little more than declarations of intent to Saharan Africa are large enough to rely on internal reduce tariffs. Existing organizations should allow trade. The case of Nigeria is relevant in emphasizing and indeed encourage subgroups of states to imple- that significant economic benefits cannot arise sim- ment more rapid and extensive elimination of trade ply from extensive economic integration. Noneof the barriers and obstacles to labor and capital mobility economic groupings (including ECOWAS without between them, while recognizing that some mem- Nigeria) has as large and unified a market, total bers would not want to proceed further than already factor mobility, harmonized investment and tax agreed. codes and common external tariff, to the same extent as Nigeria (viewed as an economic unit). The poor Main Elements of a New Approach economic performance of Nigeria reinforces earlier arguments conceming the importanceof generalized Most favored nation trade liberalization. Any new liberalization to secure major economic benefits. In- approach must start with the recognition that the tegration is only a useful means to that end. regional market can be increased thomugh most fa- Sub-Saharan African countries depend for their vored nation trade liberalization pursued by each growth "upon their ability to trade relatively freely member country. At present nontariff barriers and with the rest of the world." (World Bank 1987, page tariffs in intraregional trade tend to be higher than 3). Notwithstanding any dynamic or static gains barriers imposed by extraregional partners. In other from regional trade, the most important benefits re- words, intraregional trade is in a dispreferred situ- side in expanding extemal trade. Regional integra- ation. A number of products now directed to the tion should be supported as a means of achieving external market would find their way to regional greater eventual outward orientation. This iswhy the markets if trade conditions were identical. Particular focus should be on creating competition in regional importance bears on the rapid lifting of nontariff factor and product markets as a step to external barriers. Conceming tariff reduction, a compulsory liberalization. time schedule for reduction is required, including unilateral measures. The Way Ahead Consistency of national and regional liberalization. Re- Current efforts at regional integration have not gional integration should provide an additional proceeded as rapidly as hoped for in Sub-Saharan means of moving groups of African countries to- Africa. Thus the Organization of African Unity states wards overall economic liberalization by improving that, "Five years after the adoption of the Lagos Plan conditions for a more active role by private agents of Action and the Final Act of Lagos, very little pro- across the frontiers of Africa. The approach needs to gress has been achieved in the implementation of the be consistent with and promote liberalization efforts Plan and the Act" (Organization of African Unity at the national level and not set back programs that 1985). This is why the regional organizations are any individual country maywant to undertake on its increasingly tuming to the World Bank and other own. At the same time, the prospects for efficient donors for support.14 There is thus a need for African economic integration, based on a regional liberaliza- govemments, regional organizations, and donors to tion of factor flows combined with external trade develop a framework for responding to such re- liberalization, are enhanced by the increasing recog- quests in a manner supportive of the objectives of nition by African governments of the need to imple- efficient integration and of national adjustment ef- ment outward oriented adjustment strategies. forts. Specific details and problems have to be resolved Existing subregional organizations. Closer economic in a more concrete context. Nevertheless a general links should be based on existing and potential com- framework is required to ensure that integration con- plementarities and trade flows. To the extent possi- tributes to improved growth prospects in Sub-Saha- ble, new initiatives should work with and through ran Africa by strengthening and deepening ongoing the existing subregional organizations, in particular national adjustment efforts. The following sugges- ECOWAS, PTA, UDEAC and SADCC. New initia- tions should be seen as a first step towards achieving tives should promote factor mobility and the free a consensus on the critical elements. To reconcile movement of goods and services within the group as these elements, it is proposed that the starting point well as generalized trade liberalization with the rest should be the existing undertakings and agreements of the world; they should not aim at reversing the by the member states. To date these have generally failure to activate customs unions with high barriers 226 Integration Efforts in Sub-SaharanAfrica: Failures, Results and Prospects -A Suggested Strategy for Achieing Efficient Integration against nonmembers including those on the conti- objectives to concentrate initial efforts on eliminating nent.. Specific provision will have to be made for non-tariff obstacles to trade. interurtion liberalization and to avoid excluding non- Similarly, the proposed emphasis on factor mobil- member African trading partners that may have ity is fully consistent with the broad objectives of the more liberal trading arrangements and/or have ex- major economic groupings in Sub-Saharan Africa. tensive trade relations with some group members, They see the current agreements "as a first step to- such as Botswana and Zimbabwe inthecontextof the wards the establishment of a Common Market and PTA and Zaire and Congo in the case of UDEAC. eventuallyof an Economic Community ... to promote co-operation and development in all fields of eco- Donors' role. To support these efforts, donors could nomic activity particularly in the fields of trade, cus- assist those individual member States that are pre- toms, industry, transport, communications, agricul- pared to relax controls on cross-border flows of fac- ture, natural resources and monetary affairs with the tors of production and goods and services. The main aim of raising the standard of living of its peoples, of justification for donors to support African requests fostering closer relations among its Member States, would be to provide national authorities with the and to contribute to the progress and development necessary leverage (arising from technical expertise of the African continent." (PTA Treaty, Chapter 2). and financial resources) to overcome political obsta- The major area of divergence concerns the grant- cles that would not otherwise be removed and which ing of regional preferences. The objective of existing hinder economic progress. unions is to pursue import substituting industriali- zation behind high barriers against third parties, Consistency With the Aims of the Existing Regional while the World Bank would argue for a lowering of Groupings extemal barriers. In practice regional preferences to date have been few and generally insignificant. Fur- In 1980 African Heads of State adopted the Lagos ther, there is an increasing commitment of African Plan of Action that advocated "a far-reaching re- governments to outward oriented adjustment. Thus, gional approachbased primarilyon collective self-re- the stage is set for the Common Extemal Protection liance" (Organization of African Unity 1980, Pream- to be interpreted as an objective of sustaining lower ble, Article 1). In the current climate of adjustment, external barriers than would be feasible in individual the objective- of the Lagos Plan need to be inter- countries, thanks to the efficiency gains from factor preted even more broadly. "To achieve the goals of mobility and increased internal (regional) competi- rapid self-reliance and self-sustaining development tion. Such an approach has been informally accepted and economic growth" (Organization of African by the Secretariats of the PTA and ECOWAS. Unity 1980, Preamble, Article 3) requires the emer- gence of strong African firms able to compete in the Consistency With Ongoing World Bank Operations world market. Thus, the spirit of the Lagos Plan is fully consistent with the promotion of extensive re- The World Bank is supporting the efforts of Sub- gional liberalization as the first step to generalized Saharan African countries geared to the eliminafion liberalization. of economic distortions and greater integration into Economic integration may offer a powerful means the World Trading System. This involves greater re- for achieving an outward orientation15 in some Sub- liance on market clearing prices for the allocation of Saharan African countries. Regional cooperation resources, including foreign exchange; efforts geared should result in increased openness (and thus a re- at rationalizing and reducing effective protection duction in protection vis-a-vis the rest of the world) through tariff reform coupled with the elimination of on the part of the member countries of the regional quantitative restrictions; and measures designed to union. This should lead to a reduction in barriers to increase private sector economic activity. Bank sup- trade in services and factors of production in addi- ported regional operations will need to reinforce tion to trade in goods. these actions. The groupings usually acknowledge, as in the case of the PTA, that "unless simultaneously with the Greater reliance on market clearingprices. Securing tr- reduction and eventual elimination of tariffs, nontar- ade and investment flows that contribute to lasting iff barriers are also eliminated, the effort of eliminat- growth will involve establishing some market clear- ing tariffs can be easily nullified" (PTA 88-017, page ing price mechanism for regional activity. It will be 4). Therefore, it would be consistent with the agreed an important prerequisite to allow free pricing of all 227 Integration Efforts in Sub-Saharan Africa Failurs, Rcsults and Prospects -A Suggested Strategy forAchieving Efficient Integration transactions by the private sector, including trade of tially dynamic actors from contributing to increased regional currencies. This may lead to distortions aris- regional economic activity. ing from multiple exchange rates. To minimize these, mechanisms will be required to ensure (1) that there Compensation is rapid movement towards exchange rate unificat- ion at a level consistent with an open external current The purpose of integration should be to improve account, and (2) that arbitrage through existing par- resource allocation, which will raise absolute in- allel markets is fully exploited to limit deviations of comes. The relative incomes of some partners may regional cross-rates from those on the parallel market diverge, and the emergence of a few poles of indus- for each regional currency against hard currencies. trialization should be expected. The primary objective of compensation should be Tariff rejform Customs unions, with their potential to equalizebenefits to privateeconomicagents rather for trade diversion behind a high common external than to national governments. This can best be tariff, should not be the primary objective of integra- achieved by extending employment and investment tion efforts. Inevitably, a regional strategy will give opportunifies and opening up goods and services some preference to regional partners, but avoiding a markets in the more advanced countries to those common external tariff will limit potential trade di- from the economically weaker ones. Nevertheless, version and emnphasize trade creation.16 A common political sensitivitywillhavetobe takenintoaccount external tariff requires joint agreement by several in addition to purely economic arguments. Thus, countries for changes and this will needlessly com- there will have to be workable means of transferring plicate trade liberalizafion. compensation between governments. In implement- The short term aim would be to complete the ing any such scheme, it will be essential to avoid dismantling of intraregional nontariff barriers and negafing the benefits that should accrue to the pri- implement most favored nation trade liberalization vate sector. according to an accelerated time-table. Over the me- Compensafion schemes should move away from dium term, extemal tariffs should be reduced to- the existing pattem in Sub-Saharan Africa, which wards the levels in the more open developing coun- emphasizes the financing of supposedly regional tries. Once tariffs are relatively low, it would prob- projects in the poorer countries. The danger is that ablybe welfare enhancing to dismantle intraregional funds specifically set aside for regional projects could tariffs (to secure the full benefits of regional compe- be spent for political reasons on schemes with a low tition) provided this was accompanied by a further or even negative rate of return. There is also a danger reduction of external tariffs. In no case should in- that such funds would finance plant and equipment traregional tariff preferences be excessive (say more that would be incompatible with a rational deploy- than 20 percent) to avoid significant trade diversion ment of resources from a regional perspective. and inefficient investment. In the long run, there A more attractive option would be a direct transfer must be an opening up to the discipline of the inter- to the budget of the weaker economies, on the model national market and therefore any regional trade of the Southem Africa Customs Union (SACU). preferences should not be excessive and should be SACU shares trade taxes with deliberate overcom- reduced after a clearly specified adaptation period pensation to the weaker members. To avoid conflicts (say 5 to 10 years). between external trade liberalizafion and fiscal reve- nue objectives, it may be preferable to share total Increased private sector economic activity. Ideally, a revenue or sales taxes (on the model of the EC, which regional initiative would involve complete nontariff shares value added tax receipts). This is especially liberalization of trade and capital flows, and full relevant in the sort of framework advocated here labor mobility among regional partners. There which emphasizes increased competition and factor would be free flow of inputs and dividends across mobility instead of tariff concessions to partners, national boundaries within the group. Freer market whilemaintaininghighbarriers against third parties. access and greaterfactormobilityis needed to attract African governments should avoid reliance on do- the foreign private investment that is sorely lacking nors to finance compensation schemes. This would in the region. Discrimination according to ownership discourage countries from participating mainly to of firms and product must end, at least for those benefit from foreign aid and to ensure that the countries to benefit from World Bank financial assis- mechanism is sustainable. This should not preclude tance.Uncompetitivebehaviormustbedealtwithby donors from financing regional projects and infra- appropriate regulation without excluding poten- structure, but such financing should be provided on 228 Integration Efforts in Sub-Saharan Afria. Failurms, Results and Prospects -A Suggested StrategyforAchkving Efficent Integration the merit of the projects and not as a means of pro- ult to envisage how the World Bank could support viding compensation. With budgets remaining se- African governments in implementing tariff prefer- verely constrained for the foreseeable future, the ences. temptation to renege on regional responsibilities From a purely theoretical perspective that ignores maybe difficult to resist. This suggests some urgency social and political costs, rapid and far-reaching uni- in devising a mechanism with some automacity in lateral most favored nation trade liberalization with the transfers to weaker states. the commensurate exchange rate action should ac- company the regional integration of factor markets. Surnurnay Recommendations and Conclusion This would lead to rapid and extensive restructuring driven by the competition from imports that would The major risks of this approach are lack of interest channel factors of production to newly profitable by Sub-Saharan African governments and regional activity. organizations, and an implementation that results in Where possible, African governments should diverting trade from efficient suppliers because the adopt such an approach. In practice, however, we are expected generalized liberalization does not materi- far from a first best situation and a wide gap persists alize. The failure of existing groups largely reflects between the objectives and achievements of Bank the unrealistic expectation that it is feasible for many supported trade liberalization efforts in Sub-Saharan countries to simultaneously agree on major and rap- Africa (see World Bank 1989d). It is noteworthy that id liberalization. It is essential to accept that different even the Bank's current objectives do not aim at countries are unlikely to be able and willing to libera- anything close to the level of protection in industrial lize at the same rate. Provided a few countries might countries, let alone full trade liberalization. This is be willing to accelerate adjustment through a re- why the regional approach may have practical rele- gional apProach, it maybe worthwhile to explore this vance for advancing trade liberalization, quite apart option. from its emphasis on regional factor mobility. The integration effort must be placed within the fr- Nevertheless, the recent study of liberalization ex- amework of a general process of liberalization to av- periences summarized by Papageorgiou, Choksi and oid diverting efficient trade away from industrializ- Michaely (1 987) and Michaely (1988) argues that one- ed countries and other developing countries (includ- shot liberalization may also be preferable for pr- ing those in Sub-Saharan Africa not in the arrangem- actical reasons. The study finds that expanding sec- ent). However, because of policy and market failures tors grow fast enough to offset unemployment in the in Sub-Saharan Africa, there may be potential for ef- adjusting sectors when liberalization is immediate, ficient trade switching to less expensive African exp- while phasing the process provides opportunities for orters from costlier traditional suppliers. This potent- contracting sectors to overturn the reform by provid- ial could be as high as between four and five billion ing opportunities to generate political resistance. dollars (see World Bank 1989d). There is also an un- The Papageorgiou review includes no Sub-Saha- realized trade potential in agricultural commodities ran African countries, and Sub-Saharan Africa may that is hindered by national regulations on both the not be comparable to other regions concerning the export and import side. Harmonization of policies to impact of trade liberalization. This is because in Sub- eliminate unofficial trade flows will also have to be Saharan Africa the economic base is so narrow, the an important element of any integration strategy. distortions so extensive, and the supply of invest- While economists have always agreed on the bene- ment so constrained that there may be few sectors or fits of free trade, in practice no economic develop- firms that can provide an offset to the general con- ment experiment, including that of the Newly Indus- traction induced by the liberalization. In this regard, trialized Countries, has everinvolved systematic and even in Nigeria, which is one of the most diversified extensive free trade. The issue of the rate and scope economies and the largest economy in Sub-Saharan of external trade liberalization should be seen in a Africa, the experience with liberalization has been practical context of movement towards a system that that unemployment rises in the short run. This is is more open and less biased against exports. because theexpandingsectorsgrowfrom alowbase, Efforts should be redirected from the activation of while the contracting sectors involve larger firms customs unions or the granting of tariff preferences. (Zanini 1987, van Eeghen 1988). Instead, emphasis should be put on the removal of Thus, thePapageorgiou findings notwithstanding, restrictions on cross-border factor movements and it is unrealistic to expect most Sub-Saharan Africa on the dismantling of nontariff barriers to the free re- governments to adopt immediate and extensive gional movement of goods and services. It is diffic- trade libealization. It is this view together with the 229 Integration Efforts in Sub-Saharan Africa: Failures, Results and Prospects -A Sugested StrategyforAchievzing Efficient Integration substantial progress still possible that leads to a prag- more lasting and we-i±hw.iile results in some cases matic view that regional liberalization may yield where extensive unilateral liberalization is resisted. Appendix Active Regional Integration Organizations in Africa SADCC - Southern African Development Coor- dination Conference (9) CEAO - Communaute Economique de l'Afrique Founded: 1980 de l'Ouest (7) Angola, Botswana, Lesotho, Malawi, Mozam- Founded: 1972. Established: 1974 bique, Swaziland, Tanzania, Zambia and Zimbabwe. Benin, Burkina, C6te d'Ivoire, Mali, Mauritania, Niger and Senegal. UDEAC - Union Douaniere et Economique de l'Afrique Centrale (4) CEPGL - Communaute Economique des Pays Founded: 1964. Established: 1966 des Grands Lacs (3). Cameroon, Central African Republic, Congo, Ga- Founded: 1976 bon, Chad and Equatorial Guinea Burundi, Rwanda and Zaire Notes ECOWAS - Economic Community of West Afri- can States (16 members -15 initial but Cape Verde This paper was prepared for the African Economic Is- split from Guinea-Bissau). (Also known as CEDEAO sues Conference to be held in Nairobi, Kenya June - Communaute Economique des Etats de l'Afrique mports.5-7, 1990. We are grateful to Michael Sarris, Paul de l'Ouest). Meo, Ajay Chhibber, Charles Humphreys and Jorge Founded: 1975 Culagovski for comments and help with this paper. The Benin, Burkina, Cape Verde, Cote d'Ivoire, Gam- paper does not necessarily reflect the views of the World bia, Ghana, Guinea, Guinea-Bissau, Liberia, Mali, Bank or its affiliated organizations. Mauritania, Niger, Nigeria, Senegal, Sierra-Leone 1. See World Bank and United Nations Development Pro- and Togo. (Integrates the CEAO members and MRU gramme 1989. members and Cape Verde, Gambia, Ghana, Guinea- 2. See World Bank 1989a for a detailed discussion. Bissau, Nigeria, Togo). 3. See appendix for a list of organizations, membership and date of formation. IOC - Indian Ocean Commission (5) 4. Notethatinthissectionapositiveevaluationismade Founded: 1982. Established: 1989 in terms of the objectives of the unions to increase intrare- Comoros, France (Reunion), Madagascar, Mauri- gional activity. No normative judgement should be tius and Seychelles. inferred from references to success or failure in achieving increased intraregional trade. MRU - Mano River Union (also known as Union 5. It should be noted that the heavy dependence on ex- du Fleuve Mano) (3). port commodityproduction, the increases of import prices, Founded: 1973. Established 1974 especially of fuels, and the increasedreliance on food impo- Guinea, liberia and Sierra-Leone rts haveimportantlycontributed to this result. It is also true Guinea only joined in 1980 that in some cases import substitution requires imported inputs to produce output with negative value added at wo- PTA - Preferential Trade ARea (17). rld prices. Nevertheless, regardless of such explanations Formed: 1982. Started 1984 (uly 1) this increased dependence on the world economy serves to Angola, Burundi, Comoros, Djibouti, Ethiopia, emphasize the folly of autarchic policies for Africa. Kenya, Lesotho, Malawi, Mauritius, Mozambique, 6. This is because of the failure of the unions to achieve Rwanda, Somalia, Swaziland, Tanzania, Uganda, their stated objectives of regional import substitution. Zambia and Zimbabwe. Sudan and Zaire are actively 7. It is significant that ASEAN has achieved the highest considering membership. sustained level of intraregional trade while being the group with the most outward orientation and the least emphasis on preferential trade arrangements. 230 Integration Efforts in Sub-SaharanAfrica. Failures, Results and Prospects -A Suggested Strategy forAchieving Effident Integration 8. In the PTA, for example, recorded intraregional trade References was almost halved from about $1 billion in 1984 to about $500,000 in 1988. Balassa, Bela. (1979). "Intra-IndustryTradeand the Integra- 9. For example, bilateral agreements between Argent- tion of Developing Countries in the World Economy", In ina and Brazil give special preferences to the capital goods Herbert Giersch, ed., On the Economics of Intra-Industry sector, create a special compensation mechanism in case of Rade. large bilateral trade imbalances by including capital flows Berg, Elliot. (1988). "Regionalism and Economic Develop- into the deficit country, and considerbilateral economic rel- ment in Sub-Saharan Africa." A study prepared for the ations as a component of planning and development pro- United States Agency for International Development. grams. October. 10. The modified role of regional integration schemes Inotai, Andras. (1986). Regional Integmtions in the New World is perhaps best reflected in the increasing attention paid to Economic Environment. Budapest: Akademia Kiado. traditional cooperation areas, without economic and trade InternationalMonetary Fund. (1988). Direction of Itade Year- policy coordination. Joint activities in developing physical book. Washington, D.C. and human infrastructure may substantially increase the Lachler, Ulrich. (1989). "Regional Integration and Eco- "integration capabilities' of a region. In Latin America, the nomic Development." Industry Series Working Paper ASEAN and the Arab region, there are ambitious plans to Number 14. Washington, D.C.: World Bank Processed. interconnect national electrical systems and develop the region's~~~~~~~~~~~~~ enrysco.Alotedvlpmn ,fifra Laghammer, Rolf J. and U. Hiemenz. 1989. Regional Inte- region~~~~~~~~~~~~~~~s~~ enrg setr loth eeomn o nongration amowng Developing Countries -Survey of Past Per- tion technologies requires close cooperation. Additional forance angeng fout ur vey of PastP- possibilities may be opened in the exploitation of geologi- formance and Agenda for Future Policy Actin. Kiel: Insti- cal resources, joint prote ction of the environment and the utilization of hydrological resources. In this paper, how- Michaely, M. (1988). "Trade Liberalization Policies: Lessons ever, we are more concerned with the benefits of factor and of Experience." World Bank. Paper presented at the con- goods market integration. See Berg 1988 for a discussion of ference in Sao Paulo for a New Policy Towards Foreign these issues in the Sub-Saharan African context. Trade in Brazil. Washington, D.C.: World Bank. April. 11. Thisis not to deny that in general intra-African trade Nellis, John R. (1986). "Public Enterprises in Sub-Saharan is lcwer than in Asia or Latin America. However, notwith- Africa." World Bank Discussion Paper 1. Washington, standing the possible distortions from re-exports to D.C.: World Bank. Processed. landlocked countries, one third of Sub-Saharan African Organization of African Unity. (1985). Africa's Priority Pro- countries achieve recorded trade with their neighbors that grammefor Economic Recovery 1986-1990. is riot out of line with experience in other parts of the Organization of African Unity. (1980). Lagos Plan of Action developing world. for the Implementation of the Monrovia Strategyfor the Eco- 12. See World Bank 1989d, table 2. nomic Development of Africa. Addis Ababa. 13. This is not an argument in favor of import substitu- Organization for Economic Co-operation and Develop- tiorn in a large internal market. Given existing widespread ment. Various years. Foreign Trade Statistics. Paris. underutilization of capacity, a policy of regional liberaliza- Papageorgiou, D., M. Michaely and A.M. Choksi. 1987. tion will lead to fewer plants at greater levels of capacity "The Phasing of a Trade Liberalization Policy: Prelimi- utilization, but this should provide the basis for external naryEvidence."WorldBankDiscussion Paper. Washing- trade liberalization rather than be used as a justification for ton, D.C.: World Bank Processed. continued third party protection. Preferential Trade Area. Report of the Twelfth Meeting of 14. The World Bank has received specific requests from the PTA Coundl of Ministers. PTA, ECOWAS, CEAO, UDEAC, and SADCC. Robson, P. (1987). The Economics of International Integration. :L5. Outward orientation simply means not creating London: Allen and Unwin. anti-export bias. Many countries have achieved this by LodnAleadUwm offsettingxpot s . o anti-export bas ofhimportbiesb Steel, William F. and Jonathan W. Evans. (1984). 'Industri- offsetting by otal anting such barriers alization in Sub-Saharan Africa - Strategies and Perform- ratmtter than -uy totally dismantlig such barriers (see Weorld Wsigo,D.. oluak Bank 1987, page 81). ance." Washington. D.C.: World Bank. 16. Tradediversionoccurswhenmoreexpensiveunion van Eeghen, Willem. 1988. "Nigeria: Survey of 33 Manu- partners substitute for less expensive external trade part- facturing Enterprises." Washington, D.C.: World Bank. ners as a result of preferences. Trade creation is the Wonnacott, Paul and Mark Lutz. 1988. "More Free Trade substitution of expensive domestic production by less ex- Associations?" Unpublished paper prepared for a con- pensive partner imports. ference of the Institute for International Economics. 231 Integration Efforts in Sub-SaharanAfrica Failures, Results and Prospects -A Suggested Strategy for Achieving Efficient Integration World Banlk (1987). World Development Report 1987. New (1989d). "Intra-Regional Trade in Sub-Saharan Africa.' York. Oxford University Press. Washington, D.C.: World Bank (December). (1989a). Sub-Saharan Africa: From Crisis to Sustainable World Bank and United Nations Development Pro- Growth.ALong-TermPerspectiveStudy.Washington,D.C.: gramme. (1989). Africa's Adjustment and Growth in the World Bank 1980s. Washington, D.C.: World Bank (1989b). World Development Report 1989. New York: Ox- Zanini, G. (1987). Note on SFEM Impact on 32 Industrial ford University Press. Companies in Nigeria Surveyed in February and May (1989c). "Trade Liberalization and Economic Integration 1987, unpublished report prepared by the West Africa in Central America." Report Number 7625-CAM. Wash- Projects Department (Industry and Finance Division). ington, D.C.: World Bank. Processed. Washington, D.C.: World Bank 232 22 The Record and Prospects of the Preferential Trade Area for Eastern and Southern African States Nguyuru H.I. Lipumba and Louis Kasekende The quest for the political and economic integration Each region was supposed to pass through three of Africa preceded the political independence of Af- stages before the formation of the African Common rican states in the late 1950s and early 1960s. The Pan Market. First, each region would gradually establish Africanist movement was largely founded by Afri- a free trade area in which tariff and nontariff barriers can Americans and Caribbeans. It was joined by would be removed. Second, each region would be- leading African nationalists, including Osagefyo come a customs union by introducing a common KwameNkrumahandMzeeJomoKenyatta,afterthe external tariff. Third, a full fledged economic com- Second World War. It called not only for the formal munity with a monetary union, preferably using a independence of African states, but also for political common currency, would be formed. It would allow and economic unity. After Ghana's independence in freedom of movement of goods and services and 1957, Nkrumah envisaged that all independent Afri- labor, capital, and entrepreneurs. Under the eco- can states would immediately form a unitary nomic union, fiscal and monetary policies and in- government and a military high command to liberate vestment codes would be harmonized. After the at- those countries that were still under colonial rule. tainment of an economic union in each region, inter- The political realities of African states after inde- regional barriers would be removed and an African pendence have prevented an all African political and Common Market would be formed. economic union. However, despite the poor per- The Preferential Trade Area for Eastern and South- fornance of most of the African regional economic em Africa (PTA) is the brain child of the Economic arrangements, interest in the formation of regional Commission for Africa and the Lagos Plan of Action. trade areas and common markets has persisted.! It was formed to incorporate 20 countries of Eastern Most African states have a GDP that is less than the and Southern Africa.2 The PrA was formally estab- value of output of goods and services produced in a lished in 1981 when 12 member states ratified the medium sized town in Europe and North America. Treaty. Operations were begun in July 1982. At the In order to address the problem of economic balkani- time of writing (May 1990) only Botswana, Madagas- zation and to promote African economic integration, car, and Seychelles have not formally ratified the the African heads of state adopted the Lagos Plan of Treaty and joined the PTA. Sudan, which according Action. The Lagos Plan envisaged the formation of to the Lagos Plan of Action was to be part of the an African Common Market by the year 2000. For Central Africa region, has joined the PTA. purposes of implementing the Lagos Plan, Sub-Saha- The objective of this paper is to analyze the per- ran Africa was divided into three regions: West Af- formance and prospects of the PTA in promoting rica, Central Africa, and Eastern and Southern Africa. intraregional trade and economic integration. After 233 The Record and Prospects of the Prefential Trade Ara for Eastern and Southern African States this introduction, the second section brieflyanalyzes Problems of Economic Integration the advantages and problems of economic integra- tion and sets out a conceptual framework for analyz- One of the problems of economic integration is that ing the performance of the PTA and its future pros- tariff reduction may not necessarilyconfertrade pref- pects. The third section briefly analyzes the econo- erences to partner states. The reduction of tariffs mic characteristics of member states and discusses among members of a regional economic arrange- the objectives of the PTA and policy instruments to ment while maintaining higher tariffs on third coun- be used to attain these objectives as provided in the try imports will produce trade preferences if tariff Treaty. The fourth section looks at the PTA's perf- inclusive import prices are the effective determinant ormance in trade promotion, cooperation in mone- of import sourcing. However, in many countries do- tary affairs, and cooperation in transport and com- mestic currencies are not convertible and sourcing of munications, industry, and agriculture. Section five imports depends more on securing import licenses discusses problems and prospects facing the PTA. and foreign exchange allocation. Another problem is that the formation of a customs Advantages and Problems of Econornic Integration union with a high common external tariff may be used to support import substituting industrialization According to international trade theory, the first at the regional level. The Central American Common best choice is not regional economic grouping,3 but Market (CACM) succeeded in promotingindustriali- free world trade. Free trade areas and customs unions zation in some of its member countries;6 however, the are considered as second best choices. Preferential increased level of trade within the CACM was largely trade arrangements would be welfare increasing if trade diversion. Establishing internationally com- trade creation (i.e. the replacement of expensive do- petitive industries is a necessary condition for sus- mestic production by cheaper imports from other tained growth of manufactured output even under a members) would be larger than trade diversion (i.e. regional common market. the replacement of cheaper imports from the rest of There are problems related to the unequal distribu- the world by more expensive imports from partner tion of costs and benefits among member countries. states).4 Article24 of the GATTpermits the formation The customs union will confer more benefits to the of customs unions. This runs against the nondis- more developed member countries, which will at- criminatory principle of the most favored nation tract increased investment in manufacturing. The status, but it is believed that trade creation will domi- less developed member countries may suffer from a nate trade diversion in a customs union. This section low level of industrial development because most of looks at some of the advantages and problems of the manufactured consumer goods will be supplied economic integration, and presents a conceptual by the advanced member countries. Also, in many framework for a developmental approach to regional small less developed countries, particularly in Africa, economic integration. customs duties account for a large percentage of total govemment revenues. Large trade diversion under a Advantages of Economic Integration preferential trade arrangement may drastically re- duce government revenues which may not be easily The perceived advantages of economic integration recouped by other tax measures. are related to increased productivity and efficiency. There are problems related to regional industries. Increased market size would facilitate the expans-ion More advanced member countries, such as Kenya of production based on economies of scale, speciali- within the East African Community, oppose plann- zation, and comparative advantage. It would enable ed distribution of industries. They see the allocation the setting up of core industries for industrialization system as an unnecessary constraint to the expansion and development. And it would increase the viability of their manufacturing sector. Planned regional loca- of projects, which would improve countries' abilit- tion and market guarantees for regional industries ies to attract capital for project funding. Regional may conflict with the objective of fostering competi- payment and clearing arrangements and increased tion to improve efficiency. And it maybe difficult to intraregional trade would enable countries to save on plan and run regional industries that involve many hard currencies. Greater factor mobility within the governments. Regional executives are usually not customs union would foster increased productivity. given adequate powers to effectively complement And successful economic integration would require regional projects, and some member countries may and at the same time enhance political stability. not give the necessary priority to regional projects. 234 The Record and Prospects of the PrfeentiaJ Trade Ara for Eastern and Southern African States A Developmental Approach to Regional Economic The Economic Characteristics of the PTA Countries and Integration the Objectives of the PTA A developmental approach to regional economic The establishment of the PTA was intended to integration would protect the regional market from address the region's economic development prob- outside competition. For the less developed member lems. This section provides data on the economic states, a policy of unilateral tariff reduction is supe- characteristics of the PTA countries and discusses the rior to joining a custums union.7 A regional economic main objectives of the PFA. arrangement should emphasize the broadening of the production base by planning and implementing Economic Characteristics of the PTA Countries joint projects in infrastructure and manufacturing industries. Particularly important are the large-scale When the PTA started operations in 1982, only 12 heavy industries such as iron and steel, fertilizer and states ratified the Treaty. By 1990 the organization other chemicals, transport equipment, etc. had 17 member states. The total population of the A developmental approach to regional integration PTA countries in 1987 was estimated tobe 200 million demands more political commitment than does mar- (see table 22-1). In 1987 the total GDP of the PTA ket oriented integration. It requires the delegation of region was about 50 billion dollars; average GDP per powers to a regional institution. Governments capita was $257. Only in Mauritius was per capita would be closely involved in planning, implement- GNP greater than $1,000. Fourteen of the countries ing, and mobilizing finance for development pro- had a per capita GNP of less than $400. jects. In the planning of the location of industries, The overall economic performance of the PTA governments would be expected to address the prob- countries during the 1980s was dismal. Per capita lem of equitable distribution of benefits and costs as GNP in all the PTA countries except Malawi, Mauri- well as the efficiency criteria. tius, and Swaziland decreased during the period (see Table 22-1. Basic economic indicators of the PTA member states Average Average Area GDP GNP annual annual Population (thousands (miUions (per capifa growth rate growth rate (millions) of square of dollars) dollars) (percentage) (percentage) Country mid-1987 kilometers) 1987 1987 1980-1987 1982-1987 Angola 9.2 1,247 7,740 - - - Burundi 5.0 28 1,150 250 -2.0 2.6 Comoros 0.4 2 199 370 - - Djibouti 0.4 22 206* 572* - - Ethiopia 44.8 1,222 4,800 130 -1.6 0.9 Kenya 22.1 583 6,930 330 -0.9 3.8 Lesotho 1.6 30 270 370 -0.9 2.3 Malawi 7.9 118 1,110 160 0.0 2.6 Mauritius 1.0 2 1,480 1,490 4.4 5.5 Mozambique 14.6 802 1,490 170 -8.2 -2.6 Rwanda 6.4 26 2,100 300 -1.0 2.4 Somalia 5.7 638 995 290 -2.5 2.2 Sudan 23.1 2,506 8,210 330 -4.3 -0.1 Swvaziland 0.7 17 369 700 1.2 3.3 Tanzania 23.9 945 3,080 180 -1.7 1.7 Uganda 15.7 236 3,560 260 -2.4 0.4 Zambia 7.2 753 2,030 250 -5.6 -0.1 Zimbabwe 9.0 391 5,240 580 -1.3 2.4 Total PTA 198.7 9,568 50,959 256 - NVote. 1985. 235 The Record and Prospects of the Preferential Trade Area for Easter and Southern African States table 22-1). Foreign exchange shortages and unsus- to the eventual establishment of a common market tainable balance of payments deficits characterized among member states. most of the PTA economies. Inadequate foreign ex- The Treaty allowed for a period of 10 years for the change led to the underutilization of the import de- elimination of all tariffs and nontariff barriers. Tariffs pendent manufacturing industries. The manufactur- of commonly agreed upon commodities were to be ing sector accounted for at least 20 percent of GDP in reduced at different rates, depending on the classifi- only Mauritius, Swaziland, Zambia, and Zimbabwe cation of the goods. After the initial tariff reduction, (see table 22-2). member states were to reduce tariffs by 10 percent every two years and remove the remaining tariffs in The main objectives of the PTA. The PTA has two 1992. The initial common list of products whose tar- main objectives. The first is to promote growth in iffs had tobe reduced contained 212 items; by the end intraregional trade by removing tariff and nontariff of 1989 it contained over 700 commodities. Member barriers, gradually moving towards a common mar- states also agreed not to increase tariffs on any im- ket, and eventually forming an Economic Commu- ports from the region above the levels prevailing in nity for Eastern and Southern Africa. The PTA's sec- September 1982. ond main objective is to promote regional economic Lesotho, Swaziland, and Botswana (a nonmem- development through planning and mobilizing ber) are members of the South Africa dominated funds for financing regional projects. Southern African Customs Union (SACU). They Promoting Intraregional Trade. According to Arti- have a common external tariff and cannot reduce clel2oftheTreaty,theobjectiveofpromotingintrare- tariffs to PTA member states without violating the gional trade was to be achieved in two ways. First SACU Treaty. Swaziland and Lesotho were granted customs duties and nontariff trade barriers among temporary exemption from implementing the tariff member states would be gradually reduced and reduction measures of the PTA Treaty. Comoros and eventually eliminated. Second would be the gradual Djibouti argued that they were very small and de- evolution of a common external tariff with respect of pendent on customs duties for government revenue. all goods imported from third countries, with a view The Treaty recognized the special economic condi- Table 22-2. Other economic indicators of the PTA member states, 1987 Current Current account Debt Annual account GDP balance service rate of balance as a Share in (percentage) (millions of Total as a percentage inflation percentage of Country Agriculture Manufacturing dollars) debt of exports (percentage) GDP Angola 46 3 - - - - - Burundi 59 9 -185 755 38.9 7.3 -16.1 Comoros 37 4 -61 203 31.2 - -30.7 Djibouti - - - 181 - - - Ethiopia 42 12 -475 2,590 39.9 -2.4 -9.9 Kenya 31 11 -639 5,950 30.2 5.2 -9.2 Lesotho 21 .15 -16 241 4.01 5.2 -5.9 Malawi 37 - -53 1,363 32.0 25.2 -4.8 Mauritius 15 24 47 775 6.1 0.5 3.2 Mozambique 50 - -676 - - - -45.4 Rwanda 37 16 -250 583 13.9 4.1 -11.9 Somalia 65 5 -59 1,965 - -28.3 -5.9 Sudan 31 8 -702 11,388 93.4 19.5 -8.6 Swaziland 24 20 -2 293 5.9 12.5 -0.5 Tanzania 61 5 -605 4,335 65.7 29.9 -19.6 Uganda 76 5 -200 1,405 31.22 38.0 -15.6 Zambia 12 23 -12 6,400 58.7 43.0 -0.6 Zimbabwe 11 31 -22 2,512 25.41 12.5 -0.4 Source: World Bank, Sub-Saharan Africa: From Crisis to Sustainable Growth, 1989. 236 The Record and Prospects of the Prefential Trade Area for Eastern and Southern African Statos tion of these countries and granted temporary ex- in the issuing of import and foreign exchange li- emptions from the full application of certain provi- censes during the 10 year interim period before the sions of the Treaty. They were required to reduce complete removal of the nontariff barriers. However, tariffs for commodities in the common list by only 25 the Treaty does not explicitly recognize that the non- percent of the prescribed percentages. convertible currencies of most member states make To be granted PTA preferences by other member tariff preferences ineffective instruments for promot- states, goods had to pass stringent rules of origin. The ing intraregional trade. products must have been produced in the member To promote and facilitate trade in goods and serv- states by enterprises which are subject to manage- ices, Article 3(4)(iii) of the Treaty aims to establish mentbya majority of nafionalsand atleast 51 percent appropriate payments and clearing arrangements equity holding by nationals of the member states or among themselves that would facilitate trade in a Government or Governments of the member states goods and service. The main objectives of the PTA or institutions, agencies, enterprises or Corporafions Clearing and Payments Arrangements as stipulated of such Government or Governments, annex III Rule in Annex VI of the Treaty include: 2(a). * promoting the use of national currencies ex- For goods to qualify for PTA tariff preferences, the pressed in the Unit of Account for the Prefer- value added by PTA countries could not be less than ential Trade Area (UAPTA) in the settlement 45 percent. Certain goods that are designated by the of eligible transactions among member states, PTA council of ministers as goods of particular im- * providing machinery for the multilateral set- portance to the economic development of member tlement of payments among the member states could qualify for tariff preference as long as states, and they contain not less than 25 percent of value added * undertaking regular consultations among from the member states. themselves on The ownership rules of origin were included in * monetary and financial matters. order to prevent foreign companies in one country from takingadvantageof thelargermarket. This rule The PTA Clearing House was established and is illegal under GATT principles and has led to started operations in February 1984. Member states heated discussion because foreign ownership in the agreed that the Clearing House should be housed in manufacturing industry is predominant among the the Reserve Bank of Zimbabwe until an independent more industrialized countries including Zimbabwe, Clearing House could be established. The PTA secre- Mauritius, Kenya, and Swaziland. Some countries, tariat has urged the member states to direct all their particularly Zimbabwe, requested temporary ex- intraregional trade through the Clearing House us- emptions from the ownership rules of origin. This ing PTA currencies. It was agreed that multilateral rule was temporarily relaxed by adopting a three tier clearing of payments would take place every 75 days. sliding scale preferential formula. The level of tariff Net debtor countries would pay the net creditor reduction would be based on the ownership shares countries in convertible currencies. The maximum of PT1 A nationals in the producing enterprises. This limits of net debit and net credit for each member complicates the administration of tariff preferences state would depend on the level of its intra-PTA within the PTA. The classification of potential im- trade. A country that exceeds its net debit position portables and exportables within the PTA region into would have to pay the excess amount in hard cur- 10 commodity groups with different initial tariff re- rency within three business days. duction levels adds to the administrative complexity. To facilitate travel within the region so as to pro. The value added requirement that is not strictly uni- mote trade in goods and services, the PTA, through fonr also complicates the administration of tariff its Eastern and Southem Africa Trade and Develop- preferences. ment Bank, introduced its own UAPIA travellers The Treaty recognizes that nontariff barriers are cheques. The PTA travellers cheques are cleared in the major impediments to the promotion of intra- the PTA Clearing House. Net creditors are paid in PTA trade. These include quantitative restrictions, dollars at the end of the transaction period. export and import licensing, foreign exchange licens- ing, stipulation of import sources, advance import, Promoting regional economic development. The sec- deposits, conditional permission for imports, and ond main role of the PTA is to promote overall re- special charges for acquiring foreign exchange li- gional economic development through cooperation censes. The Treaty calls for preferential treatment of by member states. The Treaty recognized the impor- thePTAmemberstatesintheallocationofquotasand tance of efficient transport and communication 237 The Record and Prospects of the Preferential Trade Area for Eastern and Southern African States links ... for the development of the Preferential Trade The Treaty called for the establishment of the Eastern Area (Article 23). Transport and communication and Southern African Trade and Development Bank, links established during the colonial period mainly with the objective of providing financial and techni. linked the export enclaves to the colonial power. The cal assistance to promote economic and social devel- lack of adequate intra-African transport and commu- opment in the member states. It would promote the nication links has continued even after three decades development of trade among member states by fi.- of independence. nancing trading activities and projects designed to The PTA Treaty also calls for member states to make the economies of member states more comple.. cooperate in the field of industrial development to mentary to each other. It would also supplement the promote collective self reliance, complementary in- activities of the member states' national develop.. dustrial development, the expansion of trade in in- ment agencies by jointly financing operations and dustrial products and provisions of related training using such agencies as channels for financing specific facilities within the Preferential Trade Area (Article projects. The PTA Trade and Development Bank was 24). Priority areas for cooperation include basic and established in 1985 with its headquarters in Bujum.. heavy industries (metallurgical, chemical, and petro- bura, Burundi. chemical industries, and mechanical, engineering, electrical, and electronics industries), manufacturing Performance of the PTA Since Launching its and processing industries for consumer goods pro- Operational Phase duction, and facilities in relation to raw material and related infrastructure such as programs for the devel- The PTA integration program is aimed at promot- opment of power and energy (Annex VII Article 3 ing cooperation among member states in all sectors (c)). of the economy. A program of implementation is The Treaty recognizes that individual member currently in place to eventually establish a common states' markets are too small for the establishment of market. Many programs have been initiated by the some of the large-scale industries. Member states are PTA Secretariat. There has been increased contact: encouraged to establish multinational industrial en- between policy makers, business people, and politi- terprises where combined markets of more than one cians of the different member states of the PTA re- member state are required to make an enterprise gion. However, trade creation and diversion are yet profitable and require huge amounts of capital in- to respond significantly to the signals in place. Per- vestment. Guidelines on the formation, location, and formance of the PTA should be judged in terms of regulation regarding ownership and management of trade promotion, cooperation in monetary affairs, multinational industrial enterpriseswere tobedevel- and cooperation in transport and communications, oped by the committee on industrial cooperation. industry, and agriculture. In the field of agricultural development, the Treaty calls for cooperation in the following Trade Promotion * development research, extension, and the ex- change of technical information and experi- The PTA program for the promotion of intra-PTA ence; trade includes: * production and supply of foodstuffs that will * The establishment of a common list of selected ensure surpluses of food, especially grains, commodities of import and export interest to and the establishment of adequate storage fa- member states (Article 3, annex 1, pg. 44), cilities and strategic grain reserves; * The relaxation and eventual elimination of tar- * coordination of the export of agricultural com- iff and nontariff barriers, and modities; * The setting up of a PTA clearing house to save * harmonization of programs in agricultural on foreign currency (Annex VI, pg. 115). and livestock production; * development of land and water resources; The PTA common list of commodities and the tariff * sharing of agricultural service and technology; reduction program are in place. However, delays in * and the marketing and stabilizing prices of implementing treaty provisions have resulted in agricultural commodities (Annex IX article 2). postponement of the deadline for trade liberaliza- tion. The tariff and nontariff barriers to intra-PTA The objective of cooperation in the industrial and trade were originally to be eliminated by 1992; this agricultural sectors makes PTA not only a Preferen- has since been revised to the year 2000. The program tial Trade Area, but also a development institution. is currently running behind the revised schedule. 238 The Record andProspects of the Pref*rvntial Trade Areafor Easter and Southern African States Statistics on trade promotion are presented in table * to conduct systematic supply and demand 22-3.Thetabledoesnotshowevidenceofan increase market surveys on products identified by in intra-PTA trade arising out of the implementation member states as being of priority import and of the PTA program. The share of intra-PTA trade in export interest, total trade is less than 10 percent. A review of the * to convene buyer and seller meetings as a trade patterns at a disaggregated five digit commod- complementary activity of the supply and de- ity level reveals that some 25 percent of total trade is mand surveys. in comLmodities in which PTA members are both By November 1988, the PTA Trade Development exporters and importers.8 Centre had created a computer data base with infor- The 1binding constraints to trade are nontariff bar- mation on over 2,000 enterprises in the region and riers, particularly the use of foreign exchange con- trade data on 1,000 products for the period from 1980 trols and import licensing. Even if a country reduced to 1985. The PTA has organized a series of buyer and tariffs, there would be no prior grounds for expecting seller meetings and trade fairs. These have increased increased volumes of imports from the PTA or other- contacts among business people and increased wise slnce imports would still be subject to import knowledge of potential sources of imports. It is esti- controls and foreign exchange allocations. It is not mated that at the second trade fair, held in Zambia in surprising that PTA trade has not yet responded 1988, close to 100 million dollars in business was significantly to the tariff reduction program. Few initiated. Nontariff trade barriers, in most cases, hin- member states accord preferences to importers from der further negotiations and the conclusion of such the region when issuing import licenses and allocat- contracts. ing foreign exchange. Only Zimbabwe is on record as having earmarked a proportion of foreign ex- Use of the clearing house. To encourage the use of change reserves for financing intra-PITA trade. How- national currencies in transactions between member ever, the amount set aside in 1989 was less than the states and the use of the Clearing House, the PTA value of imports from other PTA states in 1988. authority decided that all transactions in goods and Information gathering and dissemination. There is a services between member states would be settled lack of readily available trade information on export through the Clearing House. This is in line with the and import sources in the PTA region. The PTA sec- broad objectives of promoting and facilitating intra- retariat, with the assistance of ITC, UNCTAD, and PTA trade. GATT, has implemented a project in information The Clearing House was established in February gathering and dissemination with the following ob- 1984. In the period up to end-December 1984 (see jectives: table 22-4) only six member states participated in the to establish a regional trade and production Clearing House. By its third year of operation, all information network with a central data base member states (with the exception of Comoros and at the PTA secretariat inter-linked with desig- Djibouti) had participated in the Clearing House. nated national trade promotion organizations Between 1984 and 1988, Clearing House payments within each member state, and the ratio of payments to total intra-PTA imports Table 22-3. Trends in total and intra-PTA trade: 1982 to 1988 (in miVions of dollars and percentages) 1986 1983 1984 1985 1986 1987 1988 Total exports 6,505 5,489 6,060 5,968 6,560 7,630 8,392 Total imports 5,344 5,015 5,349 5,203 6,035 6,511 7,105 Total trade 11,849 10,504 11,409 11,171 2,595 14,141 15,497 Intra-PTA exports 402 377 374 339 359 538 575 Intra-177A imports 542 430 434 355 391 584 643 Total intra-PTA trade 945 807 809 695 750 1,122 1,218 Intra-P7A exports share 6.2 6.9 6.2 5.7 5.5 7.1 6.9 Intra-lTA imports share 10.2 8.6 8.1 6.8 6.5 9.0 9.1 Intra-:P A trade share 8.0 7.7 7.1 6.2 6.0 7.9 7.9 Note: These figures exclude Swaziland and Lesotho. Total exports are less than total imports largely because of underreporting of imports to evade customs duties. 239 The Record and Prospects of the Preferential Trade Area for Eastern and Southern African States increased significantly, and the net settlement in for- Bank (refer to document PTA/TC/CP/XIIUI1o). eign exchange as a percentage of gross trade financed During August to December of 1988, total sales of the decreased. According to a report given by the Secre- cheques by member states amounted to UAPTA 1.5 tary General to the 14th meeting of the Clearing and million. In the first four months of 1989, total sales Payments committee (November 1989), improved amounted to UAPTA 1.1 million. performance of the Clearing House continued in 1989; its volume of business amounted to UAPTA Establishment of a monetary union. In a further en- 334.8 million and only 41.3 percent of payments deavourto attain the objectives of trade liberalization through the Clearing House were settled in foreign and economic integrafion, the PTA is moving in the exchange currency. These figures indicate success in direction of establishing a monetary union. At the the use of the Clearing House as a multilateral insti- 13th meeting of the Council of Ministers held in tution for settlement of payments. Arusha (November 1988), it was agreed to undertake a study to examine the feasibility of setting up a Cooperation in Monetary Affairs monetary union and how the PTA could attain this goal on a gradual basis. The study is currently in Included in the PrA's cooperation in monetary progress. affairs are the establishment of the PTA Trade and Development Bank, promotion of UAPTA travellers Cooperation in Transport and Communications, cheques, and establishment of a monetary union. Industry, and Agriculture The PTA trade and development bank. The PTA Trade The medium and long term objective of the PTA is and Development Bank was established in Novem- to restructure the economies of the PTA member ber 1985 and became operational in January 1986. In states through cooperation to establish industrial, May1989 the Bank introduced financing schemes in agricultural, transport, and other projects. This is respect of letters of credit confirming preshipment intended to increase capacity utilization in existing finance and providing credit to export oriented in- industries and create new capacities. A charter on dustrial projects. By October 1989, the Bank had con- multinational industrial enterprises between two or firmed letters of credit for import of critical inputs more member states is already in place. Table 22-5 from third countries worth 4.2 million dollars. The presents a list of projects and probable locations Bank is currently concentrating on providing short which have been identified by the ECA and PTA termn trade financing to facilitate intra-PTA trade. Secretariats. The list is not exhaustive. Also under consideration is the setting up of an Export Development Finance Department (EDF) in Problems and Prospects of the PTA the Bank. Tlhe PTA is a group of poor countries with weak UAPTA travellers cheques. UAPTA travellers economies. At the formal poitical level there is ap- cheques were launched in August 1988 with the main pearance of a firm commitment to cooperation. How- objective of facilitating intra-PTA travel. By April ever, many countries in Africa are too small to be 1989, a total face value of UAPTA 37.6 million had viable economic units. The landlocked countries are been distributed to 14 member states by the PTA unlikely to sustain economic development if their Table 22-4. Gross transactionsfinanced through the ClearingHouse Net settlement Intra-PTA Payments in foreign exchange Payments Payments imports as a percentage as a percentage (millions of Annual (millions of of intra- of gross trade Period UAPTA) percentage rate UAPTA) PTA imports financed 1984 37.2 - 423.8 8.8 70.4 1985 48.9 31.5 350.0 14.0 86.9 1986 59.4 22.7 333.0 17.8 51.7 1987 87.9 47.9 451.9 19.5 54.8 1988 142.0 61.5 478.8 29.7 50.1 Source PTA Clearing House Executive Secretary's Report, June 1989. 240 The Record and Prospects of the Prefeential Trade Area for Eastern and Southern African States neighboring countries are in economic turmoil. For crisis. Many are currently implementing World Bank these countries, economic cooperation with neigh- and International Monetary Fund designed struc- boring states is necessary, though not sufficient, for tural adjustment programs that require, among other sustained economic development. things, trade liberalization. Overall trade liberaliza- A major problem facing the PTA is the diversity of tion has the effect of undermining tariff preferences its membercountries. Five of thememberstateshave to member states. At present, this is an academic populations of less than one million. Lesotho and point because tariff reduction does not confer effec- Swaziland cannot effectively participate because of tive preferences. Effective liberalization of import their membership in the Southern African Customs licensing and payment regimes may actually benefit Union. Comoros and Djibouti consider themselves intra-PTA trade by removing the foreign exchange too small and dependent on customs duties as a control impediments to intraregional trade. source of government revenue to be able to effec- The fosteringof economic cooperationwill require tively implement the PTA tariff reduction program. coordination of macroeconomic and trade policies. Angola and Mozambique are facing civil wars that The PTA is currently looking at the modalities of make tariff reduction and monetary union look like forming a monetary union which will require a com- academic exercises. A few countries other than those mon currency or fixed exchange rate regime and free that have specific derogations have published the movement of capital. It would be a pleasant surprise reduced tariff rates applicable to goods imported if the PTA countries were to adopt one currency. The from the PTA. However, the complex commodity monetaryunionwhichthePTAcentralbankersprob- classification and the three tier system of the rules of ably have in mind is coordination of monetary and origin make it difficult for business people to under- exchange rate management to attain exchange rate stand clearly which goods have what preferences stabilityamongmembercountries. Similarrmonetary under the PTA program. and fiscal policies will be necessary, though not suf- Another major problem the PTA faces is the elimi- ficient, for achieving rates of inflation that do not nation of nontariff barriers. These are imposed in dif- differ excessively among member states. ferent formsbyvirtuallyall member states. Most cur- The PTA has not confined itself to promoting and rencies are inconvertible. Foreign exchange allocati- facilitating trade. It is also a development institution on and import licenses are required before goods can that will identify and promote development projects, be imported. Unlessgoods in thecommon list are put in particular in industry and agriculture. The selec- inanopengenerallicensebymemberstates, tariff pr- tion, planning allocation, and implernentation of eferences will not be effective. Many member states, multinational projects to serve the whole region is however, have expressed the fear that opening up in bound with political and technical problems. The use this manner will result in large foreign exchange out- of the efficiency criteria is likely to lead to concentra- flows in settlement of the net indebted position aris- tion of industries in countries with better infrastruc- ing from intra-PTA trade. They consider foreign ex- ture and existing industrial bases. Equitable distribu- change allocation, for example, as a necessary tool for tion may lead to establishment of high cost indus- management of scarce foreign exchange resources. tries. There is an inherent conflict between efficiency One of the key objectives of the PTA is the use of and growth of the region as a whole, and equity. The national currencies in intra-PTA trade and the multi- smaller and more coherent East African Commu- lateral clearing facility. To date there has been little nity's experience in regional industrial allocation use of the clearing house and many exporters have does not offer optimism that the larger and more continued to invoice in hard currencies. This is ex- diversified PTA region is likely to succeed in this plained largely by the fact that many of the PTA area. countries maintain overvalued and rapidly appreci- The PTA has a small secretariat and an ambitious ating currencies. The export retention schemes in program. It attempts to initiate cooperation on many operation in Uganda and Tanzania also discourage fronts at the same time. It so far has not consolidated invoicing in local currencies as exporters stand to achievements in the specific area of promoting and benefit if they earn hard currencies. A large propor- liberalizing trade. There is a danger of its efforts tion of trade among some member states goes being dissipated in too many areas without showing through unofficial channels to avoid official foreign concrete results in any particular area. Moreover, exchange and import control and to evade customs Angola, Botswana (non-member), Lesotho, Malawi, and sales taxes. Mozambique, Swaziland, Tanzania, Zambia, and Most of the member states have gone through a Zimbabwe are members of the Southern African De- period of drastic economic decline and persistent velopment Coordination Conference (SADCC). Z41 The Record and Prospects of the Prefemntial Trade Area for Eastern and Southern Afrcan States Table 22-5. Sub-Regionalprojects identifiedforestablishmentin thePTA Projects and projet idcas Location Other participants 1. Steel, steel products, and Zimbabwe related activities Uganda Kenya Mauritius Tanzania Lesotho Malawi Swaziland 2. Manufacture of engines for Zimbabwe Must identify other states to supply parts and farm equipment and road Kenya components transport components 3. Manufacture of diesel Ethiopia As above engine-mounted chassis Tanzania 4. Low cost multi-purpose Uganda As above vehicles Zambia 5. Agricultural machinery Zimbabwe As above 6. Irrigation equipment Zambia 7. Sections and bars for high Zimbabwe tension electricity wires 8. Energy equipment Zambia 9. Manufacture of transformers Zambia Participation encouraged 10. Fertilizers Kenya Other states invited to participate Lesotho Malawi Ethiopia Burundi Tanzania Uganda Zimbabwe Mauritius Swaziland Zambia 11. Caustic soda production Kenya 12. Building materials Mauritius 242 The Record and Prospects of the Preferential Trade Area for Eastern and Southern African States Table 22-5. (continued) Projwts and project ideas Location Other participants 13. Processing of fish Djibouti and other seafood Comoros (exploratory study proposed) Somalia 14. Improvement of cement (Idea being promoted industry in SADCC) 15. Utilization of steel Zimbabwe plant waste 16. Production of (Action plan pharmaceuticals formulated) 17. Production of pesticides (Action plan formulated) 18. Manufacture, maintenance, (Being promoted in and repair of railway rolling SADCC and involving stock Tanzania, Zambia, and Zimbabwe) 19. Production of hospital (Being promoted in equipment and spare parts SADCC) 20. Aircraft maintenance Ethiopia Tanzania Kenya Zambia Zimbabwe 21. Aircraft training Ethiopia Kenya Uganda Zambia 22. Development of inland Ethiopia ports and strengthening Mauritius national shipping companies Kenya Malawi Tanzania Uganda SADCC has promoted trade among its member oped and should be granted special and differential countries through exchange of trade information and treatment in the implementation of the Treaty Provi- possibly counter trade. It has, however, rejected the sion and, if possible, be assisted financially by other use of tariff reduction as a means of trade promotion. member states. Using the standard UN criterion of Thus, within the PTA region, there is duplicity of classifying the least developed countries based on institutions with somewhat similar objectives. There income per capita, share of manufactured output in is a need for coordination to avoid conflict and du- total GDP, and life expectancy, all PTA countries are plication of activities. least developed with the exception of Kenya, Mauri- Small countries such as the Comoros, Djibouti, tius, Swaziland, Zambia, and Zimbabwe. Mauritius Burundi and Rwanda feel that they are least devel- and Swaziland are small countries with populations 243 The Record and Prospects of the Preferential Trade Area for Eastern and Southern African States of less than one million each. Zambia's economy has 8. PTA document presented to the Inter-government collapsed. Among the large countries, only Kenya Committee of Experts in Nairobi, November 1989. and Zimbabwe can be considered as more devel- oped. It is not feasible for the two countries to offer References economic assistance to the rest of the member states. Discussion of the problems facing the PTA may Atsain, A. (1990): Intra AfricanTrade of the West African make the reader pessimistic about successful eco- Monetary Union, mimeo. nomic integration in Eastern and Southern Africa. Cline, W.R., E. Delgado (eds) (1978): Economic Integration in However, the potential for expanding intraregional Latin America (Washington, D.C.: Brookings Institute). trade does exist and should be exploited. Cooper, C.A. and B. F. Massel (1965): Towards a general The PTA has succeeded in bringing business peo- theory of customs unions in developing countries, Jour- ple and traders from the region together and they nal of Political Economy, Vol. 73. have formed a PTA Chamber of Commerce. Informa- El-Agraa, E.M. (1988): international Economic Integration tion on products produced by each member country (London: MacMillan). is increasingly available. What is needed is a more Hazlewood (1975): Economic Integration: the East African liberal payment regime at least for intra-PTA trade Experience (London: Heinemann) and easy movement of goods and people. The PTA Hazlewood (1979): The End of the East African Commu- should and can concentrate and solidify measures of nity: What are the Lessons for Regional Integration liberalization and facilitation of trade before attempt- Schemes?, Journal of Common Market Studies. ing to move forward in the difficult area of selecting Kiggundu, S.l.: APlannedApproach toa Common Market in planning, locating, and implementing multinational Developing Countries (Nairobi: Coign Publications). projects. PTA: A Brief to the Sixth Meeting of the PTA Authority on Problem-Areas that are Delaying the Implementation Notes of Activities of the PTA, PTA/AUTH/VI/3. PTA/CM/XI5: Report of the Eleventh Meeting of the 1. The success of the West African Community, CEAO, Council of Ministers. November 1987. and the West AfricanMonetary Unioninpromotingintrare- PTA/CM/XII /5: Report of the Thirteenth Meeting of the gional trade is documented by Atsain (1990). Council of Ministers. November 1988. 2. The twenty countries are Angola, Botswana, Bu- PTA/TC/CP/XIV/10: ReportoftheFourteenthMeeting rundi, Comoros, Ethipoia, Kenya, Lesotho, Madagascar, of the Clearing and Payments Committee. November Malawi, Mauritius, Mozambique, Rwanda, Seychelles, So- 1989. malia, Swaziland, Tanzania, Uganda, Zimbabwe, and PTA/TC/CT/VII/3: Study on the Feasibility of Elimi- Zambia. nating Customs Duties and Non-Tariff Barriers to Intra- 3. For a discussion of advantages and experiences of PTA Trade by September 1992. April 1987. economic integration in different parts of the world, see PTA Report of the Study Team on the Equitable Distribu- El-Agraa (1988). tion of Costs and Benefits in the PTA. November 1989. 4. The conepts of trade creation and trade division were PTA Treaty for the Establishment of the Preferential Trade introduced by Viner (1950). Area for Eastern and Southern African States. June 1982. 5. For a critical review of the failure of regional eco- PTA/TC/CP/XIII/10: Report on UAPTA Travellers nomic integration among developing countries, see Vaitsos Cheque Operations. June 1989. (1978). Vaitsos, C.V. (1978): Crisis in Regional Economic Coop- 6. For a detailed analysis of the benefits and costs of eration (integration) among Developing Countries: A economic integration in Central America, see Cline and Survey, World Development, Vol. 6. Delgado (1978). Viner, J. (1950): The Customs Union Issue (New York; 7. See Cooper and Masell (1965). Carnegie Endowment for International Peace). 244 23 Enhancing Trade Flows Within the ECOWAS Sub-Region: an Appraisal and Some Recommendations Ademola Ariyo and Mufutau I. Raheem The Economic Community of West African States riers, especially along the Anglophone-Francophone (ECOWAS) was formed in 1975. It was seen as one of dimension. Third, export performance is better than the surest means of enhancing the accelerated devel- import performance, mainly because the required opment of the subregion. Although several imports are not being produced within the region. appraisals of the performance of this organization This finding forms the basis for the suggestion that have been donebefore, another look at this issue was the success of ECOWAS requires focus not only on necessary for the following reasons. distribution, but also on production. It requires a First, the consolidation period for the operation of deliberate production arrangement to facilitate the ECOWAS was expected to last 15 years. Hence, 1990 availability of some of the required goods hitherto marks an important milestone for evaluating its per- not available within the subregion. formance. Second, most of the previous studies were The remainder of the paper is organized as follows. undertaken before the introduction of stabilization The second section provides a brief discussion of the and structural adjustment programs by some mem- rationale for the establishment of ECOWAS. The ber countries of the ECOWAS subregion. Hence, the third section provides an in-depth analysis of the possible incremental impact of this development as nature and growth of intra-ECOWAS trade. Some of reflected in the structure and volume of intra- the major causes of the unimpressive level of trade ECOWAS trade is yet to be documented. Third, little flows within the subregion are highlighted in the a ttention has been paid to the trade matrix amongst fourth section. Some suggestions for enhancing trade members of the community. Finally, attention had flows within the subregion are offered in the fifth earlier been more focussed on distribution, and less section. Conclu ding remarks are presented in the last on the production of tradeable goods within the section. region. The aim of this paper is to address some of these Fonnation of ECOWAS: The Rationale issues. The major findings of the study are as follows. First, there has been a slight improvement in the Regional economic integration is an issue that has volume of trade within the region. This may be due been extensively discussed in the economic literature to the benefits of stabilization and structural adjust- with respect to both developed and the less devel- ment programs which encourage intraregional eco- oped countries (LDCs). Recently emphasis has been nomic cooperation. Second, the structure of trade on the need for LDCs to form economic groupings as flows has broken the language and geopolitical bar- a means of enhancing their growth and development 245 Enhancing Trade Flows Within the ECOWAS Sub-Region: an Appraisal and Some Recommendations Table 23-1. ECOWAS merchandise trade: degree of opennessfor 1987 MiUions of dollars Degree of opennas Country Total exports Total imports Total trade GDP (percentage) Benin 114 607 721 1,570 46 Burkina Faso 127 416 543 1,650 33 Cape Verde 9 124 133 158 84 C6te d'lvoire 3,166 2,153 5,319 7,650 70 Gambia 100 182 282 172 164 Ghana 1,017 1,019 2,036 5,080 40 Guinea 453 336 789 2,166 36 Guinea-Bissau 4 70 74 135 55 Liberia 912 1,482 2,394 990 242 Mali 109 437 546 1,960 28 Mauritius 370 364 734 840 87 Niger 413 355 768 2,160 36 Nigeria 8,300 5,390 13,690 24,390 56 Senegal 645 1,174 1,819 4,720 39 Sierra Leone 212 189 401 900 45 Togo 233 561 794 1,230 65 Source: World Bank 1989; IMF International Trade Statistics prospects. This suggestion was based on the poten- which are characterized by slow growth in industrial tial benefits of regional economic interaction. countries, declining terms of trade in LDCs, mount- Several attempts have been made in the past three ing protectionism within the industrial countries decades to initiate economic grouping schemes in against developing countries' exports, and huge ex- Latin America, Asia, and Africa. Informing this effort ternal debt repayment which constitutes a serious has been the usual general argument for trade expan- burden on the LDCs. Cicin-Sain and Marshall (1983) sion among developing countries through trade lib- have emphasized the need to conceive and establish eralization. It is believed that, by opening up a large preferential trading and payment arrangements and growing market among developing countries, among the developing countries to allow, inter alia, economies of scale and more broad based specializa- for a better achievement of individual countries' bal- tion can be efficiently achieved than would otherwise ance of payments objectives. They conclude that this be the case (UNCTAD 1979). Consequently, some can be facilitated through a faster expansion of in- incentives will emerge that will stimulate production traregional trade. in the tradable sectors and ultimately lead to an It was against this background that ECOWAS was expansion of trade and economic activities. formed. The Agreement (or the Treaty) establishing Economic integration among LDCs may occur in the ECOWAS was signed in Lagos on May 28, 1975, response to the present world economic conditions, and became effective in June 1975 when the required Table 23-2. Total and intra-ECOWAS trade (millions of dollars) Intra- Intra- Intra- Total Total Total ECOWAS ECOWAS ECOWAS (4) as a (5) as a (6) as a exports imports trade exports imports trade percentage percentage percentage Year (1) (2) (3) (4) (5) (6) of (1) of (2) of (3) 1975 11,747 9,713 21,460 502 392 894 4.3 4.0 4.2 1980 34,289 24,977 59,266 1,472 1,441 2,913 4.3 5.8 4.9 1984 19,031 13,150 32,181 1,083 1,207 2,290 5.7 9.2 7.1 1987 18,285 14,859 33,144 1,098 1,291 2,389 6.0 8.7 7.2 Source: IMF International Trade Statistics (various issues) 246 Enhancng Trade Flows Within the ECOWAS Sub-Region: an Appraisal and Some Recommendations Table 23-3. ECOWAS members' trade statistics: 1975 (millions of dollars) Exports to Imports fom ECOWAS ECOWAS Exports Imports countrie countries as a as a as a as a percentage percentage Intra- Intra- percentage percentage of total of total Total Total regional regional of tofal of total ECOWAS ECOWAS Country exports imports exports imports exports imports exports imports Benin 57 197 4 16 7.0 8.1 0.5 1.9 Burkina Faso 44 151 24 35 54.5 23.2 0.4 1.5 Cape Verde - - - - - - - - C6te d'Ivoire 1,188 1,126 181 57 15.2 5.1 10.0 10.8 Gambia 149 59 2 4 1.3 6.8 1.3 0.6 Ghana 807 791 11 19 1.4 2.4 6.8 7.6 Guinea- 143 164 5 19 3.5 11.6 1.2 1.6 Guinea-Bissau 6 38 1 1 16.7 2.6 0.1 0.4 Liberia 395 331 5 4 1.3 1.2 3.3 3.2 Mali 50 241 9 77 18.0 32.0 0.4 2.3 Mauritius 199 227 3 32 1.5 14.1 1.7 2.2 Niger 91 102 6 21 6.6 20.6 0.8 1.0 Nigeria 7,995 6,032 163 55 2.0 0.9 67.5 57.9 Senegal 462 581 78 38 16.9 6.5 3.9 5.6 Sierra Leone 136 196 4 7 2.9 3. 1.1 1.9 Togo 126 174 6 8 4.8 4.6 1.1 1.7 Total 11,848 10,410 502 393 4.2 3.8 100.0 100.0 Source: World Bank 1989; IMF International Trade Statistics number of seven countries ratified it. The Commu- were not obliged to eliminate existing duties. From nity encompasses 16 countries with the membership the third year, however, member states were to com- comprising nine French, five English and two Portu- mence a reduction in the level of tariffs and to pro- guese speaking countries. The broad objectives of gressively eliminate quotas and other quantitative ECOWAS as contained in Article Two of the Treaty restrictions on trade. Harmonization of external tar- envisage the promotion of: iffs was expected to have been fully implemented by ''cooperation and development in virtually all the tenth year. Finally, the Treaty specifically indi- fields of economic activity, particularly in the fields cates that a customs union would be established of industty, transport, telecommunications, energy, within the first 15 years of operation of the scheme. agriculture, natural resources, commerce, monetary The establishment of the ECOWAS was the culmi- and financial questions, and on social and cultural nation of several past efforts and initiatives to bring matters, for the purpose of raising the standard of together these countries of diverse political and so- living of its people, of increasing and maintaining cioeconomic backgrounds. This historical antecedent economic stability, of fostering closer relations hasbeenextensivelydiscussedbyHazelwood (1967) among its members and contributing to the progress Sohn (1972), Onwuka (1982), Ezenwe (1984) and Die- and development of the African continent." jomaoh (1985), among others. Prior to the formation Promotion and expansion of intra-Community of ECOWAS, economic cooperation was organized trade is therefore the principal motive for the estab- along two separate blocks-Anglophone and Fran- lishment of ECOWAS. To enhance the realisation of cophone groupings. The speed with which the ECO- tlhis objecfive, manyprovisionswere inserted into the WAS provisions could be implemented was prob- Treaty. A timetable for the implementation of these ably affected by the existence of these subgroups. provisions was agreed upon as follows. It was stipu- ECOWAS is now about 15 years old. Many of its lated that two years after the coming into effect of the decisions, protocols, and Treaties remain unimple- Treaty, member states would be required not to im- mented. As the Community embarks on the second pose new duties on intra-Community trade. They phase of its existence, it is desirable to provide evi- 247 Enhanang Trade Flows Within the ECOWAS Sub-Region: an Appraisal and Some Recommendations Table 23-4. ECOWAS members' trade statistics: 1980 (millions of dollars) Exports to Importsfrom ECOWAS ECOWAS Exports Imports countries countries as a as a as a as a percentage percentage Intra- Intra- percentage percentage of total of total Total Total regional regional of total of total ECOWAS ECOWAS Country exports imports exports imports exports imports exports imports Benin 63 331 9 24 14.3 7.3 0.2 1.3 Burkina Faso 90 358 39 84 43.3 23.5 0.3 1.4 Cape Verde 4 68 1 6 25.0 8.8 0.0 0.3 C6te d'lvoire 3,142 2,187 482 185 15.3 8.5 9.2 8.8 Gambia 30 167 9 14 30.0 8.4 0.1 0.7 Ghana 1,206 1,130 9 261 0 23.1 3.5 4.5 Guinea- 358 324 19 14 5.3 4.3 1.0 1.3 Guinea-Bissau 11 55 2 4 18.2 7.3 0.0 0.2 Liberia 600 535 11 10 1.8 1.9 1.7 2.1 Mali 148 458 15 168 10.1 36.7 0.4 1.8 Mauritius 235 316 1 38 0.4 12.0 0.7 1.3 Niger 566 594 71 111 12.5 18.7 1.7 2.4 Nigeria 26,802 16,517 549 227 2.0 1.4 78.2 66.1 Senegal 477 11,052 147 153 30.8 14.5 1.4 4.2 Sierra Leone 219 336 22 35 10.0 10.4 0.6 1.3 Togo 337 551 89 108 26.4 19.6 1.0 2.2 Total 34,288 24,979 1,475 1,442 4.3 5.8 100.0 100.0 Source: World Bank 1989; IMF International Trade Statistics dence on the extent of its achievements relatingto the termined by their external sector. They are highly growth, structure, and direction of trade flows, and susceptible to the vagaries of the world economy. to examine some of the major solutions for overcom- Table 23-2 presents data on the total volume of ing existing constraints. These will hopefully be use- foreign trade by ECOWAS member countries for the ful, especially for the political leaders and policy- years 1975, 1980, 1984, and 1987. The proportion of makers in redesigning the strategy for enhancing the intra-ECOWAS trade in total trade is also given. The achievement of the laudable objectives of ECOWAS. data indicate that the region's total trade increased from $21 billion in 1975 to $59 billion in 1980. This The Nature, Growth, and Relative Importance of represents a 176 percent increase over the period. Intra-ECOWAS Trade Between 1980 and 1984, the value decreased by 46 percent to $32 billion. The downward trend is attrib- Foreign trade isaveryimportant economicactivity utable to the global recession which adversely af- for all the member states of ECOWAS. They depend fected the performance of the member countries' on trade to meet their growing needs for foreign economies. No significant improvement was exchange, essential imports, and government reve- achieved since then as the trade volume increased nue through import and export duties. The extent of marginally to $33 billion in 1987. this dependence, normally referred to as the degree The trend of intra-ECOWAS trade follows a similar of openness, is measured by the ratio of foreign trade pattern. The volume of intraregional trade increased to the gross domestic product. The higher the value by 226 percent, from $894 million in 1975 to $2,913 of this indicator, the higher the degree of dependency million in 1980. By 1984, it had decreased to $2,291 on trade. Table 23-1 provides data bearing on this million, after which it experienced a marginal in- issue for member countries for 1987. crease to $2,390 million in 1987. The figures suggest that most countries are open In spite of its increase in absolute terms, the pro- to external economies. This suggests that the eco- portion of intra-ECOWAS trade to the total foreign nomic performance of these countries is largely de- trade remains unimpressive. However, there are con- 248 Enhandng Trade Flows Within the ECOWAS Sub-Region: an Appraisal and Some Recommendations Table 23-5. ECOWAS members' trade statistics: 1984 (millions of dollars) Member Member Exports to Imports from country's country's ECOWAS ECOWAS exports imports countries countries relative to relative to Intra- Intra- relative relative total total Total Total regional regional to total to total ECOWAS ECOWAS Country exports imports exports imports exports imports exports imports Benin 114 396 9 43 7.9 10.9 0.6 3.0 Burkina Faso 80 255 16 76 20.0 29.8 0.4 1.9 Cape 'Verde 3 71 1 5 33.3 7.0 0.0 0.5 C6te d'lvoire 2,700 1,507 368 240 13.6 15.9 14.2 11.5 Gambia 76 112 38 7 50.0 6.3 0.4 0.9 Ghana 530 621 5 118 0.9 19.0 2.8 4.7 Guinea 409 251 33 13 8.1 5.2 2.1 1.9 Guinea-Bissau 18 51 0 5 0.0 9.8 0.1 0.4 Liberia 451 363 12 49 2.7 13.5 2.4 2.8 Mali 136 341 42 102 30.9 29.9 0.7 2.6 Mauritius 270 331 17 19 6.3 5.7 1.4 2.5 Niger 228 298 39 88 17.1 29.5 1.2 2.3 Nigeria 13,142 7,116 359 89 2.7 1.3 69.1 54.1 Senegal 535 1,000 118 259 22.1 25.9 2.8 7.6 Sierra Leone 148 166 3 63 2.0 38.0 0.8 1.3 Togo 191 271 24 31 12.6 11.4 1.0 2.1 Total 19,031 13,150 1,084 1,207 5.7 9.2 - - Source: World Bank 1989; IMF International Trade Statistics siderable variations in the value and share of intra- other countries (Benin, Burkina Faso, Gambia, ECOWAS trade of individual countries compared to Guinea, Mali, Mauritania, Niger, and Nigeria) which their global trade, as Tables 23-3 to 23-6 indicate for maintain dealings with between seven and nine 1975, 1980, 1984, and 1987. The picture that emerges other member nations. Ghana, Liberia, Sierra Leone, from these tables is that the largest exporters to the and Guinea Bissau deal with between three and six, subregion include Nigeria, C6te d'Ivoire, and Sene- while Cape Verdi appears to export to none of the gal. The largest importers from the subregion are countries within the subregion. Senegal and Togo C6te d'Ivoire, Ghana, Mali, Mauritania, Senegal, have the most diversified import sources. Cape Verde Burkina Faso, and Nigeria. is the least diversified, as it imports from only two With respect to exports, Nigeria's position is un- member countries. derstandable because of increased petroleum exports There is a wider spread in the countries' imports to m,ost ECOWAS member countries. However, Ni- than in exports. Trade concentration in the subregion geria's performance is not impressive when related isinfluencedbymonetaryzoneandlanguagegroup- to its total volume of foreign trade. In relative terms, ings. Countries in the same (table 23-7) monetary the landlocked countries-Burkina Faso, Mali, and zone or language group tend to trade more with Niger-conduct a substantial proportion of their for- themselves than with other countries. For instance, eign trade with other members of the subregion. there is a stronger and more intensive trade link They collectively account for the highest proportion among the Francophone countries than among the of intra-ECOWAS trade for the years considered in Anglophone countries. This may be due to the com- this paper. mon historical and economic background within Tables 23-7 and 23-8 show the flow of each coun- each group. try's exports and imports to and from other mem- The commodity composition of the member states' bers. Senegal exports to all other member countries, exports and imports is also important. Tables 9 and while Togo and Cote d'Ivoire trade with 10 or more 10 present data on the structure of merchandise ex- countries in the group. Following closely are eight ports and imports for ECOWAS member countries. 249 Enhancing Trade Flows Within the ECOWAS Sub-Region: an Appraisal and Some Recommendations Table 23-6. ECOWAS members' trade statistics: 1987 (millions of dollars) Member Member Exports to Imports from country's country's ECOWAS ECOWAS exports imports countries countrics relative to relative to Intra- Intra- relative relative total total Total Total regional regional to total to total ECOWAS ECOWAS Country exports imports exports imports exports imports exports imports Benin 114 607 10 46 8.8 7.6 0.7 4.2 Burkina Faso 127 416 11 129 8.7 31.0 0.8 2.9 Cape Verde 9 124 3 5 33.3 4.0 0.1 0.9 C6te d'lvoire 3,166 2,153 467 335 14.8 15.6 20.1 14.8 Gambia 100 182 44 4 44.0 2.2 0.6 1.3 Ghana 1,017 1,020 5 120 0.5 11.8 6.5 7.0 Guinea - - - - - - - Guinea-Bissau 4 70 0 3 0.0 4.3 0.0 0.5 Liberia 912 1,482 9 34 1.0 2.3 5.8 10.2 Mali 109 437 31 137 28.4 31.4 0.7 3.0 Mauritius 369 364 38 73 10.3 20.1 2.3 2.5 Niger 413 355 30 79 7.3 22.3 2.6 2.4 Nigeria 8,300 5,390 289 53 3.5 1.0 52.8 37.1 Senegal 645 1,174 111 212 17.2 18.1 4.1 8.1 Sierra Leone 212 189 2 23 0.9 12.2 1.3 1.3 Togo 233 561 9 20 3.9 3.6 1.5 3.9 Total 15,730 14,524 1,059 1,273 6.7 8.8 - - Source: World Bank 1989; IMF Intemational Trade Statistics Similar tables cannot be computed for the intrare- ever, there seems to be room for optimism. For exam- gional trade flows due to nonavailability of data. ple, previous studies asserted that the level of intrare- However, itisassumed that the tables presented here gional trade was between three and four percent are representative of the intragroup trade. Theexport during the 1970s and early 1980s. This study has profile is dominated by fuels, minerals and metals, shown that it had increased to about seven percent and other primary commodities, which includes in the late 1980s. This may be a result of several effo- food items and agricultural raw materials. The ex- rts to induce intraregional trade. Most countries are port of manufactures is very negligible with the ex- now trading with a relatively larger number of oth- ception of Mali, Senegal, and Sierra Leone, where er member countries than in the early 1980s, thus manufactures accounted for more than 25 percent of cutting across the monetary and language barriers total exports in 1987. witnessed in the past. Even so, the present level of Food, fuels, and manufactured goods have domi- about seven percent is very low. Some of the major nated the region's import profile. The share of food factors militating against a significant improvement in merchandise imports is greater than 15 percent in in intraregional trade flows, after 15 years of the about half of the countries in 1987. The share of existence of ECOWAS, are highlighted in the next machinery and transport equipment in total imports section. is also relatively high. It accounts for more than 30 However, a substantial volume of intraregional percent in Nigeria, Senegal, Ghana, Guinea Bissau, trade is unrecorded. This may be due to the activities Burkina Faso, Mali, and Mauritania. It appears that of smugglers and petty traders along the interna- ecological differences, tional boundaries where unofficial markets are lo- The level of intraregional trade in West Africa is cated. Even though there is no certainty as to the still very low. All the ECOWAS member countries magnitude and intensity of unrecorded trade, there carry out a large proportion of their trade transac- is ample evidence which tends to suggest that the tions with countries in Europe and America. How- size may not be so small. 250 Table 23-7. Intra-ECOWAS trade matrix-exports (USS millions) Burkina Cape CMte Guinea- Maurit- Sierra Country Benin Faso Verde d'lvoire Gambia Ghana Guinea Bissau Liberia Mali ania Niger Nigeria Senegal Leone Togo Benin - 0.12 _ 0.45 _ 1.77 _ _ 0.01 0.25 3.24 1.46 0.06 _ 0.60 Burkina Faso 0.26 _ _ 1.74 _ 0.67 _ _ 0.49 - 0.75 0.22 - - 0.34 Cape Verde - - - - - C6te d'lvoire 16.6 98.0 - - __ 28.4 - - 14.0 47.6 6.10 27.04 12.4 36.7 _ 27.7 Gambia - - - - - 39.13 3.88 0.07 0.03 - 0.05 _ 0.47 0.03 Ghana - 2.80 - - 0.10 - - _ 1.10 - - 0.10 020 _ _ 0.10 Guinea - - - - 0.23 0.05 _ _ 0.48 1.04 _ _ _ 0.02 0.08 0.23 Guinea-Bissau - - - - 0.014 - 0.019 - _ - - - - 0.15 _ _ Liberia - - - 0.10 _ 0.30 0.70 _ - - - - 5.90 - 1.20 Mali - 1.32 _ 3.87 _ - - - 1.55 - 0.85 0.29 4.45 0.16 0.10 Mauditania 053 - _ 12.92 _ 0.70 - - 1.66 - - _ 3.53 2.29 _ 0.70 Niger 2.84 2.16 _ 1.61 _ 0.02 0.08 - _ 0.76 0.01 _ 21.58 - - 0.08 Nigeria 4.00 0.10 _ 1.40 _ 36.0 - - 6.00 - - 6.00 _ 74.0 1.90 1.40 Senegal 4.30 4.77 1.48 27.23 3.20 0.15 6.84 2.94 0.15 18.74 7.49 1.91 1.98 _ 0.44 4.00 Sierra Leone - - - - 0.18 - 0.04 - 1.19 - - - 0.01 0.02 _ _ Togo 0.43 0.98 - 0.82 - 0.35 0.08 - 0.03 0.42 - 1.65 0.18 0.02 - - Source Computed from IMF: International Trade Statistics. Table 23-8. Intra-ECOWAS trade matrix-imports (USS millions) Burkina Cape C6te Guinea- Maurit- Sierra Country Benin Faso Verde d'lvoire Gambia Ghana Guinea Bissau Liberia Mali ania Niger Nigeria Senegal Leone Togo Benin - 0.28 - 18.27 _ 0.04 - - 0.05 - 0.58 0.87 6.44 4.77 _ 0.70 Burkina Faso 0.24 _ _ 107.81 _ 3.08 _ _ _ 1.45 _ 2.38 1.25 5.24 - 5.00 Cape Verde - - - 0.90 - - - - - - - - - 1.60 - - C6te d'lvoire 0.50 1.90 _ _ _ 0.10 _ _ - 4.30 14.2 1.80 154.4 30.0 _ 0.07 Gambia - - - - - 0.11 0.25 0.02 0.02 - - - - 3.52 0.20 - Ghana 1.90 0.70 - 31.2 43.0 _ 0.10 _ 0.30 - 0.80 - 39.0 0.2 - 0.40 Guinea - - - - 4.27 - - 0.02 0.72 0.09 - 7.53 0.04 0.06 Guinea-Bissau - - - - 0.078 0.038 - - - 0.02 3.23 _ _ Liberia - - - 15.40 _ 1.20 0.50 - - 1.70 1.80 - 6.00 0.20 - Mali 0.34 0.54 _ 107.39 - - - - - - 0.98 0.64 20.6 - 0.47 Mauritania - - - 6.68 0.06 - - - - 0.93 0.01 0.06 8.24 _ - Niger 2.38 0.82 _ 29.75 _ 0.14 1.60 -_ 38.73 2.10 - 1.82 Nigeria 1.00 - _ 14.00 1.00 - __ 7.00 - 4.00 1.00 _ 2.00 - _ Senegal 0.06 - 0.01 40.32 0.04 0.44 0.02 0.02 0.01 0.17 2.52 - 81.53 _ 0.30 0.02 Sierra Leone - - - 0.23 - - 0.09 - 0.36 _ - _ 21.26 0.49 _ Togo 0.29 0.58 - 8.17 - 0.58 0.13 - 0.18 0.11 0.72 0.45 1.0 4.89 - - Source: Computed from IMF: InternationalTrade Statistics. Enhancing Trade Flows Within the ECOWAS Sub-Region: an Appraisal and Some Recommendations Table! 23-9. Structure of ECOWAS merchandise exports (percentage of share) Fuels, minerals Other prinary Machinery and transport Other and metals commodities equipment manufacturers Country 1980 1987 1980 1987 1980 1987 1980 1987 Benin 51 42 45 38 1 6 3 15 Burkina Faso 0 0 89 98 2 1 8 1 Cape Verde - - - - - - - - C6te d'lvoire 6 4 84 86 2 2 7 7 Gambia 1 1 96 92 0 0 4 7 Ghana 81 37 67 60 0 0 2 2 Guinea - - - - - - Guinea-Bissau Liberia 59 57 38 41 1 0 2 1 Mali 0 0 83 71 0 1 16 28 Mauritius 76 31 23 66 0 0 2 2 Niger 86 86 12 13 1 0 2 1 Nigeria 96 91 3 8 0 0 0 1 Senegal 22 17 49 43 2 3 26 37 Sierra Leone 13 22 31 19 0 1 56 58 Togo 76 66 16 26 1 1 6 7 Source: World Bank data Table 23-10. Structure of ECOWAS merchandise imports (percentage of share) Other primary Machinery and Other Food Fuels commoditie equipment manufacturers Country 1980 1987 1980 1987 1980 1987 1980 1987 1980 1987 Benin 19 11 4 34 8 2 21 49 49 37 Burkina Faso 19 16 13 3 4 5 29 34 34 42 Cape Verde - - - - - - - - - - C6te d'lvoire 17 19 17 15 3 4 28 35 35 35 Gambia 11 43 7 4 3 14 19 60 60 30 Ghana 9 6 27 17 4 3 30 31 31 37 Guinea - - - - - _ _ _ _ _ Guinea-Bissau 7 5 15 9 3 3 34 41 41 49 Liberia 18 19 28 21 3 3 28 23 23 29 Mali 16 12 17 16 2 2 39 26 26 27 Mauritius 29 26 9 10 3 2 36 25 25 27 Niger 13 18 26 6 4 11 27 29 29 20 Nigeria 17 8 2 3 3 3 33 45 45 50 Senegal 15 10 17 8 3 3 30 35 35 43 Sierra Leone 19 17 14 9 5 4 18 44 44 49 Togo 19 20 20 6 3 6 20 28 38 40 Source: World Bank (1989) 253 Enhancing Trade flows Within the ECOWAS Sub-Region: an Appraisal and Some Recommendations Factors Accounting for Low Volume of Intraregion- pendence, membercountries continue to have strong al Trade in West Africa economic links with the former metropolitan coun.* tries in Europe. This linkage is manifested in trade Several factors have been identified as the causes treaties and monetary arrangements between the of the low level of trade among ECOWAS member two sets of countries. As Pobbi-Asaman (1980) countries. One of these is the noncomplementary pointed out, most of the former colonial powers have nature of the production structure arising from simi- given preferential access to their former colonies" larities in the characteristics of the national econo- produce. These preferences and arrangements in ef. mies of member states. Virtually all the countries in fect have discriminated against countries which are the region rely to a large extent on the exportation of not part of such a grouping. primary (agricultural and mineral) products. The im- Political factors, especially political instability in port structure favors manufactured goods which are member countries, have also been a hindrance to the not produced domestically. Furthermore, some in- growth of trade within the subregion. There have dustrial outfits produce light manufactures such as been cases where political leaders of member coun-- consumables, textiles, and soft drinks, which are des- tries have accused one another of direct or indirect tined only for the local market. Hence, it may be interference in their domestic affairs. This usually difficult to satisfy the import requirements of a coun- provides an excuse for the closure of intemational try by the exports of other member countries. borders with the attendant negative effect on trade Another major factor is that some of the member flows within the ECOWAS subregion. Related to this countries have policies which are not conducive to is the problem of nonimplementation of the legions the growth and development of intraregional trade. of protocols and decisions reached by the ECOWAS An important aspect of this is the nonliberalization member countries. The ECOWAS Secretariat (1985) of trade (especially with respect to unprocessed has lamented the low level of compliance with Coin- goods, traditional handicrafts, and industrial goods), munitypolicies and the nonimplementation of Com-i which was considered necessary for the creation of a munity decisions and programs which in effect have free trade zone within the region. The achievement been the major constraint to the implementation of of this goal has continued to be stalled by the failure trade liberalization arrangements within the subre- of member nations to comply with the ECOWAS gion. Available information suggests that this high Resolution on Liberalization of Trade. While the tar- default rate arises not necessarily because these deci- iff is still being retained in several instances, it has sions have been found unacceptable by the authori- become fashionable among the countries to impose ties of the member states, but rather due to the lack different kinds of nontariff barriers on trade flows. In of political will to implement them. particular, the retention of the existing tariff and the The poor state of infrastructural facilities within introduction of new tariffs are encouraged by the fact the region is another factor militating against en- that indirect taxes (mostly in the form of import and hanced trade flows within the ECOWAS subregion. export duties) account for a large proportion of gov- This relates particularly to transportation (air, rail, emnment revenue in all the countries in the region. and road) and communication networks. For exam- Tariff removal, therefore, does not appear attractive ple, national railway systems were built according to to these governments without an alternative ar- different technical specifications. The implied gaps rangement that would provide adequate and app- in the routes and differences in gauges, rolling sto- ropriate compensation for those member countries ck, and facilities have made the cost of transport whose import duties were to be lowered or remov- prohibitive. This is especially true in the case of the ed through trade liberalization. The operation of the landlocked countries (UNCTAD 1978). The absence ECOWAS Fund for Compensation, Cooperation and of regular shipping facilities and lack of suitable air Development (based in Lome, Togo) was meant to connections aggravate an already bad situation. Per- alleviate the problem of loss of revenue. Unfortu- haps with the exception of one or two member co- nately, it is bedeviled with several operational pr- untries, there are no modern harbors conducive to oblems. Furthermore, different kinds of nontariff efficient handling of import and export activities. barriers to trade flows have also been introduced. Intracommunity communications are very expen- Another factor contributing to the low level of sive, irregular, and unreliable. In many instances, intraregional trade is the traditional attachment of telephone, telegraphic, and telex networks often member countries to former colonial master coun- have to be channelled through European outposts. tries, especially the United Kingdom, France, and The special problem of landlocked countries (Mali, Portugal. Even after attaining political inde- Nigeria, and Burkina Faso), which have transit trade 254 Enhandng Trade Flows Within the ECOWAS Su-.Region: an Appraisal and Some Recommendations arrangements with the coastal states, deserves men- difficulties, and some other weaknesses that have tion. These arrangements are usuallybedeviled with been pointed out in other studies (Cicin-Sain 1980; problerns relating to transportation mode, summary Frimpong-Ansah 1987). The cash dominated trading declaration of goods, harmonization and simplifica- activities also encourage smuggling and unrecorded tion of customs documents, and imposition of cus- trading activities which thereby distort the true pic- toms duties and other charges on transit traffic by ture of trade flows within the ECOWAS subregion. coastal countries. As Alabere (1987) rightly pointed The existence of other economic subgroupings that out, the inherent transportation difficulties and par- run parallel with ECOWAS may also jeopardise the ticularly high transportation costs make it difficult growth of intraregional trade within ECOWAS. Arti- for exports of these inland countries to compete fa- cle 59 of the ECOWAS Treaty is considered to be very vorably with those of coastal countries. This problem flexible and tolerant with respect to the estab- has two main negative effects. First, these landlocked lishment of other micro-integrative subsystems countries would have to lower the prices on exports within the Community. The Article allows any mem- to stay competitive, thereby lowering the income of ber country to be a member of other regional or rural producers. On the other hand, the higher prices subregional associations with other member and which these countries must pay for their imports nonmembercountries. This has given rise to conflicts compared with coastal prices tend to fuel inflation- of interest between ECOWAS and other subgroups, ary pressures thereby raising the real cost of living and divided loyalty on the part of countries belong- (IMF 1970). ing to more than one group. Examples of such ar- Several studies (Olofin 1977, McLanaghan et al rangements include the Economic Community of 1982, Cicin-Sain and Marshall 1983, Raheem 1983, West Africa (CEAO) under the Treaty of Abidjan ECA 1982 and 1984) have identified the widespread (1973), the Mano River Union, and the Free Trade controls and restrictions on exchange transactions as Area between Cape Verde and Guinea Bissau. The a major factor impeding the development of intra- Anglophone countries have no subregional group- West African trade. These exchange controls and re- ing arrangements. However, there are a number of strictions contribute to a large extent in making in- otherbilateral trade arrangements between Commu- convertible most of the currencies of the member nity members. nations. The region is characterized by differences in Contradictions in the treaties of these different payments restrictions applied by member countries. groups are bound to arise, and may retard the grow- This is partly in response to the balance of payments th of trade within ECOWAS. For instance, there are problems they face, and partly influenced by diver- glaring divergencies in terms of the 'rule of origin' gent policy approaches. The proliferation of these issue (World Bank 1989). Also, the customs union mostly nonconvertible currencies constitutes a seri- proposed by ECOWAS, which involves the abolit- ous problem for intragroup payments on commercial ion of all tariffs, is incompatible with the CEAO's transactions. preferential system. The complications arising from The dominance of cash transactions in intra- the conflicting demands of the various subgroup ECOWAS trade (Oke 1989) reduces the pace of devel- arrangements have probably been a major cause of opment of intraregional banking and financial coop- the nonimplementation of the phased program of eration. For example, the extent of use of some con- trade liberalization envisaged in the ECOWAS ventional instruments like letters of credit, bills for Treaty. collection, documentary credits, sight drafts, and Other factors militating against trade growth bank financing of trade would be hindered by reli- within ECOWAS include inadequacy of relevant in- ance on cash transactions. The negative effect of the formation and lack of uniform and standard meas- problem has also been highlighted by the World urement and grading systems. There is a dearth of Bank (1989), which emphasized that even if African functional centers to provide information on the suppliers could be competitive in price and quality, availability of products and related opportunities inadequate financial arrangements tend to place amongmembercountries.Thereisnoefficientmeans them at a disadvantage with non-African competi- of contact between prospective business partners. tors. The establishment of the West African Clearing The existence of the West African Chamber of Com- House (WACH) has facilitated the removal of some merce has not helped much regarding information of the identified obstacles that emerge as a result of dissemination. Thebankingsector (international and the lack of direct convertibility of the ECOWAS cur- local) in these countries has not been very helpful in rencies. However, the operations of the WACH are this regard. The absence of uniform grading and being constrained by structural problems, regulatory measurement standards has made impossible mean- 255 Enhancing Trade Flows Within the ECOWAS Sub-Region: an Appraisal and Some Recommendations ingful price and quality comparisons across coun- and reconciliation of tariff structures shouldbe given tries. The absence or inadequacy of efficient modern priority. However, this should be complemented storage facilities fortradeableproducts such as meat, with a frank assessment of the viability of the fresh fish, fruits, and vegetables constitutes another ECOWAS Fund to enable each member state to real- serious problem. istically assess the net effects of tariff review on its revenue. Enhancing Intra-ECOWAS Trade Flows: Some Third, all impediments to the free flow of trade Suggestions should be identified and the extent of amendments attainable should be quantified. Prominent among The preceding sections have highlighted the ra- these is the existence of competing groupings whose tionale for the establishment of ECOWAS, the struc- terms of reference conflict with those of ECOWAS. htre of trade relationships among member countries, Attempts should therefore be made to identify and and the problems militating against the achievement reconcile the conflicts and appropriate adjustments of the objectives underlying the establishment of should be made. The activities of these subgroups ECOWAS. The findings indicate that, in spite of the should beconsistentwith theobjectives of ECOWAS. current low level of intraregional trade, ECOWAS Fourth, ECOWAS member states should prioritize member countries are striving to ensure the growth trade profile targets. For example, our analysis of trade flows within the region. The relative volume showed that food items account for a sizeable pro- of intraregional trade flows has increased from about portion of total imports within the region. Hence, the 4 percent in 1984 to over 7 percent in 1987. desire for self-sufficiency in, say, food items within a Some of the previous studies commissioned by specified number of years may be a priority. It was ECOWAS have emphasized the scope for increasing also noted earlier that there exists surplus exportable intraregional trade ingoodsbeingproducedbyexist- goods within the region. The hindrances to their ing industries in the region. Attention was expected exportation should be identified and appropriate in- to be focussed on semiprocessed and light industrial centives and proposals should be provided. Finally, products. These include beverages, textiles, confec- there must be improved information flows and infra- tioneries, nails, timber and wood products, cement, structural facilities to encourage interaction among and candles, among others. Our analysis indicates member states and their citizens. that, given the high level of import demand for manufactured goods in each of the member states, New Areas of Focus manufacturing holds the key to a meaningful and appredable growth in intra-ECOWAS trade. The suggesfions presented here so far have related The potential effects of the industrial policies of to eliminating some lapses in the exisfing decisions. member countries should not be overlooked. The It is also appropriate to suggest some new areas of countries have similar industrialization programs focus. The ECOWAS Treaty is essentially concerned based on the import substitution strategy. Hence, for with the movement of tradeable goods (i.e. distribu- most industries the domestic market is the target. tion) only. The Treaty has not paid particular atten- This leads to excess capacity problems, and massive tion to production. However, the analysis indicates importation of spare parts, which affect member that the volume of imports from outside the region states'balance of payments positions. still dominates intra-ECOWAS imports. Efforts should therefore be made to identify those items still Recommendations being imported from outside the region and a joint production and investment program should be es- The fairly substantial level of smugglingwithin the tablished. It should be possible to identify and estab- region would have led to distortions and most prob- lish some production units specifically meant to ably an understatement of the actual level and direc- serve the ECOWAS market. The prioritization of the tion of trade flows within the region. Hence, the first implementation of such projects may be guided by recommendation offered here is that efforts should the multipliereffects, potential market size, technical be made to minimize this practice. This observation feasibility, comparative advantage, geographical leads to the second suggestion that there should be a spread, and cost of implementation (UNCTAD 1986). renewed determination to eliminate tariffs and other The existing ECOWAS Treaty does not seem to hindrances to trade within the region. Existence of emphasize trade in raw materials and intermediate and differences in tariff rates are major causes of inputs. Goods in these categories could be purchased smuggling within the ECOWAS subregion. Review from outside only when not available from within 256 Enhanang Trade Flows Within the ECOWAS Sub-Region: an Appraisal and Some Recommendations the region. In addition, an overhaul of the financial leaders to ensure meaningful implementation of system should be given priority to facilitate account these proposals. settlements within the region. In this light we com- The study also identified new areas of focus. These mend the existence of the West African Clearing relate to an expansion of the scope of activities cov- House, although the scope and efficiency of its op- ered by the existing ECOWAS Treaty, to include raw erations need to be strengthened. Finally, each mem- materials and intermediate inputs as tradeable items. ber state should endeavor to formally incorporate the The harmonization of production and investment pr- growth of intra-ECOWAS trade into its economic oposals should be part of the package, with a view to policy decisions. A good example is the construction producing items currently being imported from out- policy recently launched in Nigeria, which made the side the region. The public sector may focus on the ECOWAS region its primary point of call for required provision of infrastructural and communication fa- imports. cilities, as part of the incentives to mobilize and While the implementation of these suggestions enhance the meaningful participation of the private would go a long way in enhancing intra-ECOWAS sector. trade, it is pertinent to register two additional obser- Finally, two recent developments within and out- vations. The first relates to the fact that the absence side the subregion will significantly impact on future of the political will to implement decisions reached trade flows within ECOWAS. These are the adoption by member countries constitutes a serious problem. of stabilization programs by ECOWAS member Political leaders need to be reminded of the enor- countries and the formation of Europe 1992. The mous advantages of the success of the ECOWAS possible opportunities and threats to intra-ECOWAS arrangement. Issues of concern to these leaders trade arising from these developments should be should be made clear and appropriate remedial ac- studied in depth, with a view to designing appropri- tions should be taken. ate strategies for capitalizing on the opportunities The second issue relates to the expected role of the and averting any threats to enhanced intra-ECOWAS private sector. In the expansion of intra-ECOWAS trade. trade, efforts should be made to mobilize the private sector. This might allow the public sector to focus on References the improvement of infrastructural facilities. The in- volvement of the private sector may also augur well Albere, B.R. (1987), 'Economic Integration in Less Devel- for the free flow of resources. Entrepreneurs should oped Countries: The Case of ECOWAS". Unpublished be free to locate anywhere within the region, with M.Sc. Economics Project Report,Departmentof Econom- appropriate incentives and protective measures, es- ics, University of Ibadan, Nigeria. pecially against imports from outside the ECOWAS Allen, R.L. (1961), 'Integration in Less Developed Area'. subregion. The establishment of a limited convert- Kyklos, pp. 315-335. ibility arrangement and subsequently a monetary Aitken, N.D. (1973), 'The Effects of the EEC and EFTA on zone in the region should be pursued vigorously. European Trade: A Temporal Cross Section Analysis'. American Economic Review, Vol. 63, pp. 881-892. Concluding Remarks Anusionwu, E.C. (1984), "West African Lessons From East African Failure in a Common Market" in Second Collo. Fifteen years after the formation of ECOWAS, most qium on the Economic Integration of West Africa. of the objectives of its founding fathers have yet to be Austin, D. (1967), 'Economic Cooperation and Integration achieved. Paramount among these is the inability to in Africa. Report of the Seminar Programme of Congress achieve an appreciable volume of intra-ECOWAS for Cultural Freedom, No.11. trade. For various reasons, member states still ac- Balassa, B. and A. Stoutjesdijk (1975), "Economic Integra- tively engage in trade with the developed economies tion Among Developing Countries". Journal of Common of Europe and America, even in some commodities Market Studies, XIV 1, September. that could be supplied from within the subregion. Cicin-Sain, A. (1980), "Strengthening the West African Thus, theproportionoftradeamongECOWAScoun- Clearing House and Adopting Limited Convertibility tries is still less than eight percent of the total. Arrangements or Steps Towards an ECOWAS Monetary The study has identified some of the major causes Zone", Project RAF/77/032, December ECOWAS Secre- for this, and has proffered some solutions. On one tariat. hand, promises and decisions are yet to be imple- Cicin-Sain, A. andJ. Marshall (1983),"Study on theLimited mented. These relate to tariff elimination and trade Currency Convertibility Among ECOWAS Countries", liberalization proposals, and to the will of political UNCrRAD/ECDG/142 GE 83-55469. 257 Enhandng Trade Flows Within the ECOWAS Sub.Region: an Appraisal and Some Recommendations Diejomaoh, V.P. (1985), 'Economic Integration in the Union Arrangements',Journal of Common Market Studies, ECOWAS States: Problems, Progress and Prospects' in Vol. XVI, September. Second Colloqium on the Economic Integration of West Onwuka, R.I. (1982), Development and Integration in West Africa. Africa: The Case of ECOWAS (Ile-Ife: University of Ife Diejomaoh, V.P. and M.A. Oyoha (eds.) (1980), Industriali- Press). sation in ECOWAS (Lagos: Heineman Educational Books Pazos, F. (1973), 'RegionalIntegration of Trade Among Less Nigerian Ltd.) Developed Countries' World Development 1 (7), July. Economic Commission of Africa, United Nations (1982), pp.1-12. 'Analysis of Bottlenecks to the Development and Expan- Pobbi-Azaman, K.O. (1980), 'Prospects for Increased Eco- sion of Intra-Sub-Regional Trade and Remedial Meas- nomic Cooperation and Integration in Africa' in Second uresinWest Africa'. Report submitted to the5thMeeting Colloquium on the Economic Integration of West Africa. of the Committee of Officials of the Niarney MULPOC in Pournakoris, M. (1969), "Economic Integration and De- Banjul, The Gambia, February. veloping Economies with Similar and Different Eco- Economic Commission of Africa, United Nations (1984), nomic Systems", Economic Internationale, Vol. 32 (1) pp. 'Monetary and Payments Implications of Border Trade 112-125. in Africa". E/ECA/TRADE/15 August 1984. Raheem, M.I. (1983), 'The Impact of Economic Integration Economic Community of West African States (1985), Ten on Intra-Regional Trade Flows: The ECOWAS Experi- Years of ECOWAS: 1975-1985. Lagos: ECOWAS. ence". Unpublished M.Sc. Economics Prcject, University Economic Community of West African States (1987), of Ibadan, Nigeria. "ECOWAS Economic Recovery". Lagos: ECOWAS. Robson, P. (1980), The Economies of International Integration. Edozien, E.C. (1975), 'Intra-African Trade as a Step To- (London: George Allen & Unwin Ltd.). wards an Inward-Outward Oriented Looking Develop- Robson, P. (1968), Economic Integration in West Af- ing Policy". The Nigerian Journal for Economics and Social rica. (London: George Allen & Unwin Ltd.). Studies, Vol. 17,65-76. Sohn (1972), Basic Documents of African Regional Organiza- Ezenwe, U. (1984), ECOWAS and the Economic Integration of tion. (New York. Oceana Publishing Inc.) Vol. 3. West Africa (Ibadan: West Books Publisher). Truman, E.M. (1969), 'The European Economic Commu- Frimpong-Ansah, J.H. (1987), Analysis of Suggested Meas- nity: Trade Creation and Trade Diversion"t. Yale Economic ures to Supervise the Functionary of the West African Essays, vol. 9, pp. 201-257. Clearing House UNCTAD/ECD/TA/20, April. United Nations Conference on Trade and Development Hazlewood, A. (1967), African Integration and Disintegration (1978), 'Study of Recorded Trade Flows'. Report pre- (London: Oxford University Press). pared for ECOWAS Trade, Customs and Monetary International Monetary Fund (1970), Surveys of African Study Project, March. Economies, Vol. 3, Washington, D.C.: IMF p. 43. United Nations Conference on Trade and Development McLenaghan, J.B. et. al. (1982), 'Currency Convertibility (1979), 'Preliminary Reports on Trade Liberalization Op- the ECOWAS", IMF Occasional Paper No. 13. tions and Issues for the ECOWAS'. Geneva, June. Oke, B.A. (1989), 'Legislative and Regulatory Obstacles to United Nations Conference on Trade and Development the Promotion of Intra-Regional Cooperation in Trade, (1986), 'Current Problems of Economic Integration: The Banking and Finance in Africa". Economic and Financial Problems of Promoting and Financing Integration Pro- Review, Vol.26, No. 4, December. jects'. TD/B/6/ .7/79. Olofin, S. (1977), 'ECOWAS and the Lome Convention: An World Bank (1989), Sub-Saharan Africa: From Crisis to Sus- Experiment in Complimentary or Conflicting Customs tainable Growth (Washington, D.C.: World Bank). 258 24 Comments on Regional Integration Achi Atsain and Sylviane Guillaumont Jeanneney AchiAtsain tection against third parties have not succeeded in increasing intraregional trade. If protection means The papers presented by Mansoor and Inotai, U- high tariffs against third parties, Lipumba and pumba and Kasekende, and Ariyo and Raheem have Kasekende have shown in their study that PTA has raised major issues confronting the economic group- reduced tariffs over the years, but intraregional trade ings of the developing countries. They present has not increased within this economic grouping. additional evidence of the overall poor perfornance The countries of the WAMU-CEAO, which have had of the economic groupings of the developing regions. quite similar practices, have, on the contrary, experi- Although Mansoor and Inotai tend to imply that enced increases in their intraregional exports and ASEAN and CACM seemed to have performed rela- appear to be the most telatively successful integra- tively better than similar economic groupings in tion scheme in Sub-Saharan Africa. This is to say that Sub-Saharan Africa, my own calculations show that trade liberalization would not be a necessary and the intraregional trade of these two groups has con- sufficient condition to increase intraregional trade. stantly declined between 1960 and 1987 as a Trade liberalization must be accompanied by com- percentage of their total exports. plementary policies such as factors mobility, which This comment raises a methodological problem as has been correctly pointed out by the three groups of data provided also by Ariyo and Raheem do not authors. support my own findings. They find that the intrare- While Mansoor and Inotai have paid little atten- gional trade of ECOWAS has steadily increased be- tion to monetary factors, Lipumba and Kasekende, tween 1975 and 1987. My own data show declining on the one hand, and Ariyo and Raheem, on the values of exports from ECOWAS and that its share of other, have linked the promotion of intraregional intraregional exports with respect to its total exports trade to the use of a convertible currency. Again one was estimated at roughly 3.1 percent between 1975 canarguethattheuseofacommoncurrencymaybut and 1987. partly explain the greater gains in intraregional trade Besides the problem of data source, the papers for UDEAC, which uses CFA francs, compared to the appear to have identified, although incompletely, the CEAO, which has not experienced increases in in- causes of the poor performance. Mansoor and Inotai traregional trade over a comparable period. In- tend to be more provocative by overemphasizing the creased intraregional trade, as argued by Landell- industrial developmentstrategypursued by the Sub- Mills and others, depends also on the availability of Saharan African countries which favored, in their finance and financial instruments such as exports, words, inefficient import substitution policy as the credits, export insurance and other sources to traders major cause of the poor perfornance of the Sub-Sa- and firms. It depends on the solidity of the financial haran African economic groupings. They argue that institutions in Sub-Saharan Africa. The liquidity cri- customs unions that have granted high internal pro- sis that is experienced by many African commercial 259 Comments on Regional Integratiox banks represents an impediment against increased prospects of increased intraregional trade by assist- intraregional trade. Unfortunately, sufficient atten- ing economic groupings in Sub-Saharan Africa in tion was not paid to this factor in any of the three removing these constraints. After all, no country in papers. the world has removed all forms of barriers to trade A further point that I would like to make is to stress with the rest of the world. Issues raised by free trade that trade liberalization, financial instruments, avail- associafions are complex and slippery and the most ability of finance and labor and capital mobility will valid generalization is that it is very difficult to gen- no longer be sufficient to expand intraregional trade eralize. World free trade is an ulfimate goal to achieve without reforms of the general administration and and the world is far away from free trade. the specialized institutions that have been estab- The last point of my comments has also received lished to promote intraregional trade. As argued by attention in the papers, although it is not sufficiently Landell-Mills and others, economic agents-firms articulated. Political tempers have done a great deal and traders-need a trading system which is free in slowing progress towards economic cooperation from any regulatory and procedural barriers to in- in Sub-Saharan Africa. Political confidence and trust traregional trade. Although I do not firmly believe which nourishes political will must be sought that unrecorded trade far exceeds officially recorded through all forms of interaction if integration trade, it could be that the occurrence of the former schemes in Sub-Saharan Africa are to bear more fruit must be the result of too many administrative regu- for the African people. Political trust is a prerequisite lations. for any form of development policy harmonization A further point that I would like to make and and for safeguarding income growth and the distri- which has been discussed in the three papers relates bution of benefits from integration. If political trust to the direct relationship between improved infra- is secured, viable regional projects could be found structure and increased intraregional trade. Infra- and funded by donors such as the World Bank or the structure here includes information, as well as com- African Development Bank. If not, I tend to agree munication and telecommunications. Contrary to with Mansoor and Inotai that no viable regional pro- Mansoor and Inotai, who tend to make the World ject could be found and those who continue to advo- Bank financial assistance conditional to the removal cate integration by production rather than by market of all forns of discrimination against third parties, I would find little empirical evidence to support their would argue that donors would greatly improve the case. Sylviane GuillaumontJeanneney The three papers I have to discuss are very inter- A Methodological Issue Raised by Smuggling esting. The Mansoor and Inotai paper gives us an exhaustive survey of the benefits that developing The overall importance of unrecorded trade, as countries can derive from economic integration and well as its unequal weight from one country to an- of the difficulties they meet. The Lipumba and other, isemphasizedbythethreepapers.Unrecorded Kasekende and Ariyo and Raheem papers enlighten trade not only leads to an underestimation of the us about two significant African experiences of inte- effective integration, but it also makes it difficult to gration: the Preferential Trade Area for Eastern and evaluate the progress of intraregional trade. Indeed, Southern Africa (PTA) and the Economic Commu- an increase in recorded trade can be due to the fact nity of West African States (ECOWAS). that some informal trade has become an official flow, The three papers reveal how difficult it is to evalu- and the converse is also true. ate the effectiveness of integration. So I first want to However, one of the purposes of regional integra- make some comments of method. ThenI shall under- tion is perhaps just to reduce the part of informal line that two issues would have called for some trade to the benefit of the formal one, in order to greater attention because of their significance for a widen the fiscal basis and so to increase public re- policy aimed at improving integration. ceipts as it is suggested in Ariyo and Raheem's paper. 260 Comments on Regional Intkgration Indeed it can be considered that a lowering of cus- exchange outside the official market. But when there toms duties and nontariff barriers within a region is no foreign exchange shortage, traders have an aims at reducing the informal part of trade since it incentive to underinvoice imports more than exports decreases the benefits of smuggling. Therefore, even because customs and fiscal duties are generally if an increase in recorded trade results from reduced higher on the former than on the latter. smuggling or from a lower underinvoicing of com- mercial flows, and so does not reflect a real growth Two New Significant Ways for Improving of intraregional trade, it can be considered as a suc- Integration cess of the customs union. From the same point of view, we may discuss There are two possible policies to improve African Mansror and Inotai's conclusion that "to ensure that regional integration, that have been overlooked until Most Favored Nation trade liberalization proceeds in recently. The three papers have presented a very parallel with the intraregional liberalization of labor interesting study of the impediments to African inte- and capital flows, a Common External Tariff should gration. All three have stressed the role of inconverti- be avoided until tariffs are generally low." If it is true bility of African currencies. CEAO, which is mostly that there is a risk that a Common External Tariff between UMOA countries, is presented as a counter would be excessively protective, it is also true that example. without a Common External Tariff intraregional But I think that another major impediment is the smuggling is stimulated. instability of real exchange rates between African The PTA and ECOWAS papers have collected sta- countries. This factor is all the more important since, tistical data on official intraregional trade. The study in view of integration, the three papers say that pro- of their evolution is a little more optimistic for duction has to be coordinated and private invest- ECOWAS than for PTA since in the first area "the ment has tobe stimulated. So the firstquestion is how findings suggest a slight improvement in volume of to stabilize real exchange rates between members of trade within the region" whereas in the second, there ECOWAS, PTA or other regional arrangements. is 'no evidence of an increase in intra PTA trade It is not realistic to foresee a monetary union be- arising out of the implementation for PTA pro- tween so manyand diverse countries as the members grams." However, we must notice that the share of of ECOWAS or PTA. As written by Lipumba and official intraregional trade in total trade was by 1987 Kasekende, "it will be a pleasant surprise of the slightly higher in PTA than in ECOWAS, i.e. 7.93 century if PTA countries adopt one currency". percent vis-A-vis 7.21 percent, which both remain Let us consider the case of West Africa: the main unimpressive. So we may consider that the perform- problem is to stabilize the real rate of exchange be- ances of these two regional arrangements are rather tween the naira and the other currencies of the re- similar. gion. Is the solution to peg the CFA franc, the CEDI, These statistics on official trade arebuilt from sepa- the Guinean franc, etc. to the naira? Is the solution to rate data on imports and exports. These records show promote, through adjustment programs, a more sta- a difference between total intraregional imports and ble real effective exchange rate of the different coun- exports, two flows which should be equal by defini- tries, including firstly Nigeria? tion. The gap is obviously due to a discrepancy in Until now, structural adjustment programs have customs registration. It can not be interpreted as a led to very different real depreciations of African specific performance of intraregional imports rela- currencies. As the same international institutions tive to intraregional exports (or conversely, accord- (IMF, World Bank and now EEC) assist African gov- ing to the direction of the difference), as it is sug- ernments in the implementation of their macroeco- gested in the ECOWAS study. Another explanation nomic policies, should they not be more concerned is necessary. about the implications that an adjustment program In many African countries, exchange restrictions in a specific country has for its neighbors? have been made more severe since 1980. We can Moreover, part of the international financial assis- perhaps find here a reason why during the 1980s tance could be directed toward the stabilization of recorded intraregional imports are greater than re- real exchange rates, while simultaneously aiming at corded intraregional exports, whereas the opposite greater convertibility. The means could be monetary was the case during the 1970s. When the currency is cooperation between African governments and one not convertible, import registration is needed to ob- of these intemational institutions. The experience of tain Joreign exchange on the official market whereas the Franc Zone provides a good example of what can unrecognized exports can be used to obtain foreign be done in this field. Certainly the maximum degree 261 Comments on Regional Integration of assistance, corresponding in the Franc Zone sys- The three papers s-. -io agree that a high com- tem to the possibility of unlimited negative balance mon tariff may lead to an inefficient import-substitu- in the operations account, is extremely unlikely to tion industrialization, and that African integration meet with much support in the international frame- has to be seen as a step toward multilateral trade work. But a range of alternative systems, implying a liberalization, which is a more or less transitory step smaller degree of exposure, is possible. It authorizes according to the authors. However, a means of com- a negative balance only for a specific limited term pensation for the poorest countries can be a special and/or places a fixed limit on the size of the allow- Common External Tariff for basic agricultural pro- able deficit. On the other hand, African monetary duction such as cattle or grain. authorities should cooperate with international insti- There is no evidence that trade liberalization must tutions to define their monetary policy; but this co- go at the same pace in agricultural and industrial operation could be less tight than it is in the case of fields. Indeed, theworld conditions of marketingand Franc Zone countries with French authorities. Surely price deternination of agricultural and manufac- this way by which intemational cooperation can pro- tured products are not the same, since the protection mote African integration calls for further investiga- in industrialized countries is higher and cases of tion. intemational dumping prices more frequent for the The second question I want to ask is how to con- former than for the latter. tribute to equity among the members of regional Another connected problem which requires fur- organizations. The main target of regional integra- ther attention is whether financial compensation in tion in Africa is generally to promote faster industri- favor of the poorest countries is better suited than alization. But it is likely that the poorest countries commercial protection. And how can international will benefit less from industrialization. What kind of institutions contribute to assess and finance these coompensation has to be given to them? compensations? 262 25 Human Resources, Technology and Industrial Development in Sub-Saharan Africa Sanjaya Lall Sub-Saharan Africa remains one of the least industri- fared worse in Africa than in many other developing alized regions in the world. Industrial production regionswith similarhandicapsand policies. We have has stagnated or declined in many Sub-Saharan Af- to look, therefore, to some structural features of Afri- rican countries over the past decade. Inefficiencies can industry that account for its past record and and external shocks, exacerbated by poor policies, impose limitations on its future development. These have caused many industries to become a drag on features (described in Lall 1989 and 1990a) have more their economies. This is the more worrying because to do with the human resource base available for Africa is still predominantly specialized in relatively industrialization than with physical conditions or simple, low technology industries. Its long-term de- trade and otherpolicies; however, thelatter influence velopment would entail entry into more complex the development of appropriate skills and capabili- activities, where technologies are changing rapidly. ties and so affect industrialization directly as well as However, the prospects for growth in these highly through human resources. This framework, broader demandingareas are not encouraging. The industrial and more comprehensive than the received ap- lag of Africa is set to increase rather than diminish in proaches to industrialization, leads to the considera- the foreseeable future. tion of much longer term policy requirements than is Two broad sets of solutions have been suggested currently fashionable. for the problems of African industrialization, based The paper is laid out as follows. The second section on differing analyses of the causes of its poor per- describes the recent African experience of industrial formance. One has focused on adverse external cir- development, evaluates its success more broadly cumstances, the other on the management of macro- than only in terrns of output growth, and comments economic policies, trade regimes or internal indus- on some of its most striking features. The third sec- trial strategies. tion sets out a simple but general framework for Both approaches clearly have merit. Exogenous analyzing the determinants of industrial develop- shocks have hit many African countries particularly ment. The fourth section applies the framework to hard, and economic management, both in policy for- Africa. The fifth section concludes the paper with mulation and its implementation, has suffered from some recommendations on the promotion of healthy many faults. The basic physical conditions for indus- industrialization. trialization have not been very propitious in many (but not all) African countries: small, fragmented Background to Industrial Development in markets, poor infrastructure, remoteness and often a Sub-Saharan Africa scarcity of relevant natural resources. Even when all these factors are taken into account, however, there In 1965, manufacturing contributed nine percent remainsalargeunexplained"residual".Industryhas of GDP in Sub-Saharan Africa, compared to 20 per- 263 Human Resources, Technology and Industrial Development in Sub-Saharan Africa cent in middle-and low-income countries, 14 percent Cameroon (8.5 percent) and Mauritius (10.9 percent) in South Asia, 26 percent in East Asia, and 23 percent was fairly impressive by any standards. In the group in Latin America (World Bank 1989a, Table 3, and of low-income countries, Burundi, Benin and Leso- World Bank 1989b). By 1987, this figure had risen to tho turned out average annual growth rates of 5 10 percent in Africa, compared to 18 percent in South percent or more in this period. Asia and 28 percent in Brazil, 30 percent in Korea, 25 This is not to deny that a large number of other percent in Mexico, 24 percent in Thailand, and some African countries suffered from low or negative 40 percent in Taiwan. A few Sub-Saharan African growth of manufacturing in the recession following countries had much higher shares for manufactur- the second oil crisis (some had stagnated for much ing; of the 45 countries in the region, in four(Zambia, longer). The worst affected were Zaire, Tanzania, Zimbabwe, Swaziland and Mauritius) manufactur- Zambia, Uganda, Somalia, Sierra Leone, Central Af- ing contributed 20 percent or more to GEDP, while in rican Republic, Nigeria, Ghana and Liberia . The another 10 it contributed 10 percent or more (World recession and its aftermath led to severe underutili- Bank 1989b, Table 3). At the same time, at least 10 zation of capacity as foreign exchange for imported African countries suffered declining or stagnant inputs and equipment was more severely rationed shares of manufacturing in GDP over the 22 years and domestic demand fell. The squeeze on modern from 1965 to 1987 (World Bank 1989b), and some industry led to considerable unemployment, and registered a larger share for manufacturing only be- some of the unemployed entered into informal sector cause GDP itself was declining rapidly (as in Zam- activity. bia). With the exceptions noted, therefore, the level of In Ghana and Nigeria, for instance, the expansion industrialization by this measure remained fairly of informal sector activity was particularly notice- low and stagnant in much of Africa. able in this adjustment period. Much of such activity The total value of manufacturing value added was in low productivity, low technology activities (MVA) in Africa came to $16.3 billion in 1986, which that needed few imported inputs. It is important, was 46 percent of that of India, 66 percent of that of however, not to "glorify" the resilience and capabili- Korea, and 28 percent of that of Brazil (World Bank ties of the informal sector. It fulfilled a valuable func- 1989b, Table 8, and World Bank 1988a). Only four tion of permitting survival, but it remained on the countries-Nigeria, C6te d'Ivoire, Cameroon and margins of subsistence. It may have provided the Zimbabwe-had an MVA of over $1 billion each, seedbed for some entrepreneurship and skills, but as while 15 had an MVA of under $10G million each. Of such the existing informal sector does not have the the group of low-income countries, Nigeria had the largest MVA; of the middle-income countries, Zim- babwe, Cote d'Ivoire and Cameroon had the largest. Table 25-1. Sub-Saharan Africa: Growth rates of In terms of growth rates (table 25-1), African in- GDP and manufactuing dustry did fairly well in the 1965 to 1980 period (8.8 (average annual percentage) percent average annually). Though rates of growth GDP Manufacturing appear overstated due to the small initial base, it is 1965- 1973- 1980- 1965- 1973- 1980- clear that the first flush of import substitution, built Africa 1973 1980 1987 1973 1980 1987 on aid and revenues from generally booming pri- All Africa 5.9 2.5 0.5 10.1 8.2 0.6 mary product exports, was vigorous. During 1980 to Nigeria 4.0 1.7 2.3 n.a. 1.9 3.4 1987, the African growth rate for manufacturing fell ig4 .n 9 to an annual average of 0.6 percent, while the South Low income 6.0 2.8 -0.4 10.7 10.2 -1.0 Asian rate rose to 8.0 percent. East Asia kept up a Nigeria 3.3 1.9 1.4 n.a. 1.5 1.4 healthy 10.4 percent while Latin America, beset by Mid-income debt problems, fell to the same rate as Africa, 0.6 economies 5.2 1.4 3.8 n.a. 2.5 6.1 percent.1 six most populous The figures in table 25-1 suggest that the slow- economies 7.0 3.0 -0.8 12.0 12.9 -1.3 down observed for Africa as a whole was in fact Sahelian confined to low-income economies (which also did economies 1.0 3.5 2.5 n.a. n.a. n.a. poorly in the 1970s), and was sharply affected by the Oil exporters 7.5 2.8 -0.5 13.5 15.0 -1.0 Nigerian performance. Middle-income countries ac- All low-income celerated their manufacturing growth in the 1980s, economies 6.0 4.6 6.1 9.1 8.1 10.3 and the performance of C6te d'lvoire (8.2 percent per China, India 5.9 4.3 1.7 8.3 10.7 3.9 annum during 1980 to 1987), Congo (9.7 percent), Source: World Bank 264 Human Resources, Technology and Industnal Development in Sub-Saharan Africa dynamism to provide the basis of sustained growth. factured exports of Hong Kong. Taiwan or Korea in Once it is "modernized", with competitive industrial that year (which were $44.6 billion, $47.3 billion and technologies and skills (in, say, the Italian mold), it $43.9 billion respectively), and only 55 percent of that may well be a source of growth, but this involves of Thailand. It contributed less than I percent of the problems of capability and efficiency which also af- $371.5 billion of the total manufactured exports of the fect the formal sector. Third World as a whole; in 1973 this share had been African industrial growth (see Steel and Evans nearly four times higher. 1984 and Gulhati and Sekhar 1982) has been highly Part of the poor export performance of African protected and overwhelmingly inward oriented. manufacturing is explained by the inward orienta- Launched primarily by foreign companies (or resi- tion of its trade regime. With the sole exception of dent non-Africans) to serve local markets or process Mauritius, there is no African economy that is raw materials for export, it was led later by state strongly export oriented in the East Asian mode. enterprises. Yet the weakness in indigenous indus- However, the Mauritian manufactured export boom, trial entrepreneurship, which shows up most clearly which was almost entirely in knitwear and other in the paucity of our modem African small-scale garments, is based on factors which differentiate it industrial (as opposed to informal) activity, could not from the rest of Sub-Saharan Africa. It has a strong be remedied by setting up parastatals. Some activi- indigenous entrepreneurial class, a well-educated (if ties, especially those which had simple technologies, not technically advanced) labor force, and a large had been in existence for a long time and had good influx (stimulated by the trade regime and the qual- (usually foreign) management, did achieve effi- ity of labor) of direct investment from Hong Kong. ciency. However, a very large proportion of industry The specific interaction of incentives and skills did not, especially in countries without ready access sparked off the Mauritian success. The only other to a plentiful supply of foreign managerial, entrepre- African country that is "moderately" export oriented neurial and technical skills. The countries that have is Cote d'Ivoire (World Bank 1987). Its export success the best industrialization records (Zimbabwe, C6te has been limited and directed mainly to neighboring d'Ivoire, Kenya, Gabon, and Mauritius) are the ones Francophone countries. Moreover, its own skills which could draw on expatriates or resident non-Af- have been strongly supplemented by expatriates at ricans,2 or could continue to attract sufficient foreign all higherlevels of management, but thishas notbeen investments. sufficient to launch it on the course to becoming a The degree of inefficiency in African industry new Newly Industrialized Country (NIC). seems to rise with the degree of capital and skill The export record of African industry thus reflects intensity of the facilities set up. Many traditional its inherent competitive weakness as well as the in- industries also display considerable inefficiency in centive structures provided by the trade regime. De- comparison to standards of other developing re- tailed micro level studies of technical efficiencyshow gions. Parastatal industries have tended tobe among very low levels of capability to operate or improve the mrost inefficient (Nellis 1986). Most African indus- on imported technologies (Mlawa 1983 andWangwe try is highly import dependent and has remained so forthcoming on Tanzania, Page 1980 on Ghana, and overtime. Local linkages have been mainly confined Mytelka forthcoming on Cote d'Ivoire), though to primary inputs, while manufactured components where foreign skills are brought in to provide a mini- or intermediates, technical and consultancy services, mum base and a "teaching" role the situation is much and technological inputs have continued to be im- better (Pack 1987 on Kenyan textiles). ported. The degree of import dependence has thus been much higher than in most other developing A Framework for Analyzing Industrial Success countries. Import dependence per se would not matter if the Neither of the two conventional explanations of transformation of imported inputs took place with the problems of African industry comes to grips sufficient efficiency to permit growth, diversifica- properly with its structural weaknesses. The conven- tion, and most importantly, penetration of foreign tional explanation based on exogenous shocks ex- markets. It is in the sphere of manufactured exports, plains the fall in production because of lack of im- however, that the weakness of African industry ported inputs and domestic recession, but does not shows itself most clearly. As data collected by the deal with the reasons for its lack of competitiveness, Worlld Bank (1989b, Appendix Table 17), show, the dynamism or linkages, and its extreme dualism. The total manufactured exports of Africa came to $3.5 conventional explanation based on incentives (trade billion in 1987. This was less than a tenth of manu- and industrial policies) explains part of the reasons 265 Human Resources, Technology and Industrial Development in Sub-Saharan Africa forpoorexport performance and lackof competitive- ing from competition. Competition provides the ba.. ness arising from protected inward orientation. sic spur to investment in capability development, buit However, it places the whole burden of the explana- is a double-edged weapon. Too little competition can tion on incentives rather than structural factors, and lead to inadequate or misguided capability acquisi.. thus tends to be partial and incomplete, but not tion; too much can wipe out firms that cannot finance entirely wrong. the costs of capability acquisition. The ideal set of' It is necessary to adopt a more comprehensive incentives thus combines some competition (of the approach to the determinants of industrial perform- right sort, ideally from world markets) with protec. ance. Recent research on the acquisition of techno- tion for the period of learning when costs are high logical capabilities at the firm level in developing and quality low. Entry into more complex activities countries provides a convenient starting point (Dahl- necessarily requires greater protection, but as the man et al 1987, Lall 1987, Katz 1987, Fransman 1986). experience of Korea and Taiwan shows, can be ac- The antecedents of such work go back to economic complished if it is combined with incentives to enter historians like Rosenberg, originators of the 'evolu- export markets or face domestic competition in the tionary theories" of growth like Nelson and Winter, near future (Lall forthcoming). analysts of technical change like Freeman, and theo- Fourth, incentives are only part of the story. The reticians dealing with problems of innovation and ability of firms to respond to incentives depends on informafion like Arrow. The application of these di- their initial base of capabilities and their access to verse approaches to problems of developing country skills within the economy. It also depends on the, enterprises has yielded many interesting new in- development of infrastructure, a supplier network sights. While we cannot review these insights here at and an institutional structure that allows informat- any length, a few relevant points may be usefully ion and technology to flow, adequate standards to be highlighted. set, research to be conducted, finance to be provid- First, the process of becoming efficient even in a ed and labor to be trained. In-house capability acqui- static sense (i.e. acquiring mastery of a given technol- sition has to be complemented by the external edu- ogy), can be slow and difficult. It goes well beyond cation and training system, the development of sup- simply setting up a physical facility, becoming famil- pliers and service firms, and an institutional fra- iar with its working in some passive sense, and real- mework that allows markets to function efficiently. izing technical economies of scale. It requires firms In developed countries these requirements are tak- to develop new skills, acquire new informafion, de- en for granted; in developing countries they cannot velop new supplier networks, and set up new organ- be. izational structures. In developing countries, where These firm level factors can be protected at the "ready made' skills, information, suppliers and in- national level. Taking for granted that macroeco- stitutions do not exist in the external environment, nomic conditions and physical infrastructure are ap- firms necessarily have to invest heavily in building propriate to industrial development, its progress de- up their capabilities. Thus, there is thus no predict- pends on the complex interplay of three sets of fac- able learning curve which all firms traverse over tors: incentives, capabilities and institutions (Lall time. The process of becoming efficient is usually 1990 and forthcoming). Incentives compriseboth the uncertain, risky and often incomplete. This is true of neoclassical prescriptions of export orientation developed and developing countries, but the latter (strictly speaking, neutral incentives to sell in foreign suffer greater lags, and larger dispersions round the and domestic markets) and the internal competition norm, because of the weaknesses of institutions and espoused by Balassa, Bhagwati and Krueger, and support systems. sufficient selective protection to allow diversification Second, efficiency cannot remain static when in- and deepening to take place. The standard neoclassi- puts, products and technological conditions are al- cal recipe makes some allowance for infant industry ways changing. Dynamic efficiency requires more protection to coexist with export orientation. How- developed capabilities than simple mastery, to adapt, ever it recommends only low, uniform and short- improve and innovate on existing technologies. No lived protection across industries to minimize the firm can be self-sufficient, since specialization is es- "distorting" effects of selective protection. There is sential for competitive success, but in certain critical neither theoretical nor empirical support for such a areas of its operation it must deepen and broaden its recipe. Once the possibility of market failures due to capabilities if it is to remain fully competitive. dynamic and unpredictable learning, extemalities or Third, the development of capabilities is extremely complementarities are admitted, protection has to sensitive to market incenfives, especially those aris- varyby activity according to its complexity and link- 266 Human Resources, Technology and Idustrial Development in Sub-Sa haran Africa ages to other activities. If judiciously administered, lowed the industrialization path set by interventions selective protection is a necessary element of indus- on incentives. trial deepening (Pack and Westphal 1986, Lall forth- Table 25-2 (taken from Lall 1990) sets out some of coming). the relevant data on performance and capabilities in Capabilities include three things: the launching of the NICs and near-NICs. It shows, among other physical investment, the provision of human capital, things, how the two medium-sized NICs invested and the undertaking of technological effort. The abil- heavily in education and training to back up their ity to set up physical facilities is such an obvious drive into advanced industry, and mounted substan- determinant of industrial growth that it needs little tial research and development efforts. Korea led the comment, except to point out that investment skills technological effort because of its unique combina- differ greatly between countries. Human capital for tion (in the developing world) of nationalism, heavy industrialization arises from the formal education industry push, and aggressive export orientation. system and employee training undertakenbyindus- The capabilities drive also involved interventions, try l'King1984).Therelevanceof education and train- both selective and functional in nature, since the ing for industrial development is obvious. However, education and technology 'markets" could not pro- the need for specialized technical training tends tobe duce the requisite output of skills or information overlooked in many discussions of industrialization. needed without direction and assistance (Lall forth- This is fairly low at initial levels of industrial devel- coming). opment (though some high level skills and training are always needed), when general education and Incentives, Capabilities and Institutions in Africa flexibility are more important. Skills and training become productive only when We now apply this general framework to Africa combined with technological effort in industry to and identify its comparable weaknesses. absorb new knowledge, adapt and improve on it. While advanced innovation is accepted as the life- Incentives blood of industrial success in developed countries, the need for more mundane technological effort there The framework sketched out here does not dismiss and (especially) in developing countries tends to be the role of correct incentives in the industrialization overlooked. Yet it is this which determines how suc- process. On the contrary, an export oriented trade cessful newcomers to the industrial scene are in pro- regime is taken to be superior to an import substitut- ducing efficiently and building up their dynamic ing one. However, "correct" incentives are defined to comparative advantage. . include selective interventions to provide protection Finally, institutions here refer to those entities set to infant industries. This kind of intervention, with up to facilitate the workings of markets. In the indus- its emphasis on selectivity and rapid gains in com- trial context, these can provide finance, information, petitiveness, should be sharply distinguished from services, standards, export assistance, and the like, the by-now "classic" type of indiscriminate and per- which comprise the whole network of external link- manent protection given in most developing coun- ages that allow individual firms to operate efficiently. tries. The pattern of interventions in the incentive This network may not be thrown up automatically regime in Africa have, sadly, been overwhelmingly by the free market. of this sort (Steel and Evans 1984). The results have The experience of the NICs illustrates the rele- been worse than in India ( Lall, 1987) because of vance of this framework nicely. As far as incentives smaller, more fragmented markets; lower levels of are concerned, three of the four East Asian NICs (the technical skill and technological effort; and a greater exception being Hong Kong) intervened both selec- shortage of entrepreneurial abilities (exacerbated by tively and functionally to promote a chosen strategy the leading role assigned to parastatals). However, in of industrial deepening. Of the three, Singapore African countries with more developed capabilities chose to rely most heavily on foreign investors for or access to foreign capabilities, the results have been entrepreneurial leadership and technology. At the distinctly better than in others. Thus, Zimbabwe set other extreme, Korea chose to rely primarily on na- up an impressive and reasonably functional struc- tional enterprises and to push massively into heavy ture of capital and intermediate goods production in industry. This entailed high degrees of intervention the years of isolation under UDI. Its integrated iron to promote selected activities and to build up the and steel mill (ZISCO) was the onlyone in Sub-Saha- skills and technological capabilities needed in those ran Africa with acceptable levels of efficiency, and it particular activities. The capabilities developed fol- was state owned and highly protected. Similarly, 267 Human Resources, Technology and Industrial Development in Sub-Saharan Africa Table 252. Indicators of national technological capability in selected NICs Hong S. Korea Taiwan Kong Singapore India Brazil Mexico Thailand Structure and Performance Value added $b. (1985) 24.5 22.2 6.7 4.3 35.6 58.1 43.6 7.7 Growth 1965-80/1980-86 18.7/9.8 16.4/12.9 17.0/7.0 13.3/2.2 4.3/8.2 9.6/8.2 7.4/0.0 10.9/5.2 Exports (1986) $b (1986) 31.9 35.9 32.6 14.7 7.2 9.1 4.9 3.9 Growth of merchandise-exports 1965-80/1980-86 27.3/13.1 19.0/12.7 95/10.7 4.7/6.1 3.7/3.8 9.4/4.3 7.7/7.7 8.5/9.2 Gross Domestic Invt. as % GDP (1986) 29.0 19.0 23.0 40.0 23.0 21.0 21.0 21.0 Capitol Goods Prod. as % of Total Manufacturing (1985) 23.0 24.0 21.0 49.0 26.0 24.0 14.0 13.0 Capitol Goods Imports $b (1985) 10.6 5.6 7.1 8.1 3.7 2.2 6.1 2.7 (as % MVA) (43.3) (25.2) (106.0) (188.4) (10.4) (3.8) (14.0) (35.1) Stock of Foreign Direct Investment 2.8 8.5 6.0-8.0 9.4 1.5 28.8 19.3 4.0/5.0 $b (1984-6) Stock as % GDP 2.8 8.1 20-26 53.8 0.7 9.6 13.6 10.5-13.1 Education A) Education Expenditure as % household consumption (1980-5) 6.0 - 5.0 12.0 4.0 5.0 5.0 6.0 ,) Public Expenditure %GNP 4.9 5.1 2.7 2.9 3.7 2.9 2.6 3.9 Year) (1985) (1986) (1978) (1980) (1985) (1984) (1985) (1984) Central Government Expenditure on Total Government Expendure (1986) 18.1 20.4 - 21.6 2.1 3.0 11.5 195 Age Group Enrolled (1985) Primary 96 100 105 115 92 104 115 97 Secondary 94 91 69 71 35 35 55 30 Tertiary Education 32 13 13 12 9 11 16 20 Vocational Ed. Enrol. (1984) ('000) 815 405 32 9 398 1,481 854 288.0 as population working age 3.06 3.24 0.86 0.5 0.07 1.83 2.0 0.96 No. of tertiary level students in S/E fields (000) 585 207 36 22 1,443 535 563 360 % population 1.39 (Year) (1987) (1984) (1984) (1984) (1980) (1983) (1986) (1985) in engineering (000) 228 129 21 15 397 165 282 n.a. % population 0.54 0.68 0.41 0.61 0.06 0.13 0.35 - Science And Technology Patents Granted: Total (1986) 3,741 10,615 n.a. 598 2,500 3,843 2,005 n.a. of which % local 69 56 n.a. 8 20 9 9 n.a. RD % GP 2.3 1.1 n.a. 0.5 0.9 0.7 .06 0.3 (Year) (1987) (1986) (1984) (1984) (1984) (1984) (1985) RD in Productive Sector % GNP 1.5 0.7 n.a. 0.2 0.2 0.2 0.2 n.a. Scientists/ Engineers in RD Per million population 1,283 1,426 n.a. 960 132 256 217 150 All scientists/engineers (a) Total nos. (1000) 361.3 n.a. 145.5 38.3 1000- 1,362.25 565.6 20.3 2000 (b) Per million population 8,706 n.a. 26,459 15,304 1282- 11,475 10,720 472 2564 (Year) (1986) - (1986) (1980) (1985) (1980) (1970) (1975) Sources: Asion Development Bank, Foreign Direct Investment in Asia and the Pacific, 1988; World Bank, World Development Report, various; World Bank, various country reports; U.N., Statistical YearbookforAsia and the Pacific1986-1987, Bangkok; UNESCO, Statistacal Yearbook 1988, Paris, 1989; Republic of China, Statistical Yearbook of the Republic of China 1988, Taiwan; Republic of China, Education Statistics of Republic of China 1984, Taiwan; Republic of China, Science and Teehnoly Data Book,Taiwan, 1987; Republic of Korea, Introductionof Science and Technolty, Seoul, 1988; Evenson, R.E. (1989), 'Intellectual Property Rights, R & D, Inventions, Technology Purchase and Piracy in Exonomic Development", (mimeo). 268 Human Rcsourses Technology and Industrial Development in Sub-Saharan Africa Maunitian enterprises, supplemented by investors technological learning and spillovers are lost in com- from East Asia, successfully set up garment exp- parison with the NICs, where investment costs are orting facilities to take advantage of its export orien- much lower and many externalities are captured. tation. The technology involved was simple to ma- ster, but the organization of production and export Humn capital. A significant part of the entrepre- marketing required skills and enterprise lacking in neurial, technical and managerial skills required for many other African countries. The textile industry in industrialization arises from the previous experience Kenya has also achieved a respectable degree of te- of commerce and industry. Africa has been, in com- chnological mastery, with substantial and prolong- parison with other developing regions, particularly ed infusions of skills from experienced technicians unfortunate in this respect. With the exception of a from India (Pack 1987). In general, however, the in- few trading communities, the indigenous popula- centive structure in Africa has not been conducive to tions of most of Africa have little traditional experi- the healthy development of industrial capabilities ence in modem commerce or manufacturing. How- (see M.eier and Steel 1989, Lall 1989, World Bank ever, there is no shortage of entrepreneurial drive: the 1989b). informal sector in Africa is as active and vibrant there But what would the response of Africa be to more as in other developing regions (Page 1979), and the liberalized incentives? In the sphere of manufactur- recent recession has swelled the ranks of infornal ing industry, apart from areas of obvious compara- entrepreneurs. However, the drive to profit from op- tive advantage based on processing local mineral portunityisnotthesameperseastheentrepreneurial and agricultural resources, do sufficient capabilities capability required to organize, set up and run mod- exist to permit a spurt of export oriented growth? em industry (Kilby 1971). And, in the longer term, what would determine its There tends to be a progression in most developing dynamic comparative advantage? The answers to societies from commerce and infornal industry to these questions depend on African capabilities and formal small- and medium-scale manufacturing, and institutions. within manufacturing from smaller or simpler to larger, more complex activities. There is, in other Capabilities words, a process of learningentrepreneurial capabili- ties just as there is a process of learning technological Physical investments. The efficient implementation or managerial capabilities. The inherited structures of industrial investments requires a broad spread of of African economies, shaped by colonial rule, have technical, design, organizational and construction placed them far down on the learning curves for all skills. Many such skills can be imported from spe- three sorts of capabilities. Attempts to force the pace cialized engineering or consulting companies or by "Africanization" of industry via public enter- capital goods manufacturers in developed countries. prises or small-scale industry promotion schemes However, the cost of relying heavily on foreign con- have not been able to bypass the learning process. tractors can be heavy, and certain critical investment However, some learninghascertainlyoccurred. West functions (initial project preparation; negotiating Africa in general has a better developed entrepre- technology design and transfer; participation in en- neurial class than Eastern or Southern Africa because gineering, monitoring, and equipment selection) of a smaller direct colonial presence in the past, and should be handled at least partlyby the project spon- some communities have more advanced trading sor. Otherwise the country risks biases in project skills than others. Butby and large, the inheritance of design, technology, and location, and much higher industrial capabilities is too small to pennit a dy- investment costs than are necessary. The problems of namic African industrial class to have emerged so far. inappropriate techno]ogy choice (Stewart 1977) and Thesmall-scale formal sector remainsveryweakand "white elephants" in Africa are well known, as are underdeveloped (Page 1979). the very high costs of setting up projects (a steel mill Past experience is only one source of capability in Nigeria costs three to four times as much as the acquisition; education and training are more impor- same investment in other developing countries). tant. The relationshipsbetween education and indus- What is perhaps less widely appreciated is that the trialization are important and binding, but the pre- lack of local participation in design leads to sub- cise links between particular types of education and sequent failure to master, adapt and improve upon specific levels or forms of industrialization are not imported technologies, and to the absence of link- always easy to trace (King 1984). The operation of ages with potential local suppliers of investment easy low technology, activities with which industri- goods (Lall 1990a). Thus, valuable opportunities for alization generally starts, requires literacy and 269 Human Resources, Technology and Industrial Development in Sub-Saharan Africa schooling a range of basic technical skills and some that the quality of vocational training is lower than high level technological and managerial skills. How in other developing countries (Middleton and Dem- do the structure and achievements of African educa- sky 1989). Secondary education has a low technical tion compare with other countries? content and poor quality of instruction and equip- Table 25-3 below sets out aggregate data on gross ment (Zymelman 1990). educational enrollments in 1965 and 1985 for Africa The available data on technical training at tertiary and other developing regions. (A more detailed levels shows that the total number of tertiary stu- breakdown for a large sample of African countries dents enrolled in scientific and technical fields in and others for comparison is available from the Africa came to 175,000 in 1983, which was below half author.) that of Thailand (in 1985), below 30 percent that of The data show that Africa started with education Korea (in 1987), and only 5 times that of the tiny levels far behind those of other regions and, while it island economy of Hong Kong. As a percentage of made considerable progress in enrolling pupils at all the population, Africa had 0.04 percent in technical levels, still lags far behind other developing regions education, compared to 1.5 percent for Korea, 1.1 (Vorld Bank 1989b, Zymelman 1990). The most criti- percent for Hong Kong and 0.7 percent for Thailand. cal input for industrial development-secondary No African country had a proportion higher than education-is particularly backward, while at the 0.17 percent (Madagascar). Of the more industrial- tertiary level the lag is even greater. In relation to ized countries, Nigeria and Zimbabwe had 0.02 per- enrollments in the "model NIC", Korea, Africa as a cent, Kenya and Cote d'Ivoire 0.06 percent; and whole had 45 percent basic literacy, 17 percent secon- Cameroon, Mauritius and Zambia 0.03 percent. It is dary enrollment, and 6 percent tertiary enrollment in interesting that Mauritius was fairly low on this the 1980s. Nearly half of the African labor force had scale, in contrast to its attainments in literacy and noschoolingwhatsoever, compared to 20 percent for secondary schooling. The expansion of garments Korea. University enrollments per 100,000 popula- manufacturing does not require specialized technical tion in Korea came to 3,606 in 1986, compared to 330 skills. However, if Mauritius were to diversify and for Zimbabwe (the highest in Africa), or only 10 in upgrade its export base in emulation of, say, Hong Mozambique. In terms of literacy and secondary Kong, a massive push would be needed in its techni- level enrollments, the best education by far was pro- cal skills base. videdbyMauritius, which provided the skillbase for Data on narrower technical fields are even more its export success. directly relevant to industrial skills. Total enroll- Figures for gross enrollments are misleading be- ments in engineering for Africa come to 48,000, cause they do not show the proportions of students which was slightly more than double that of Hong that stay on to complete the course. Dropout rates in Kong, and only 21 percent of that of Korea. As a Africa tend to be particularly high (World Bank proportion of the population, Africa has 0.01 percent, 1988b), especially in comparison with East Asia. The Hong Kong 0.41 percent and Korea 0.54 percent. The figures also do not show the qualityof training given, highest for any African country, again Madagascar, nor the technical orientation of the courses. There are is 0.02 percent, followed by Kenya and Gabon, with reasons to believe that educational quality in Africa 0.016 percent each.4 has been declining recently (World Bank 1988b), and UNESCO collects data on the total number of "po- tential scientists and engineers" in each region. Rough as the data are, they again indicate the small Table 25-3. Gross enrollment ratios in education. size of Africa's human capital base for industrializa- (pcrcentage of group) tion. In 1985, Africa had 1,376 potential scientists and Primary Secondary Tertiary engineers per million population, compared to Region 1965 1986 1965 1986 1965 1986 11,730 for Asia (including Japan), 11,759 for Latin Sub-Saharan America, and 8,263 for all developing countries. Africa 41 66 4 16 0 2 Total enrollment in vocational training for the re- East Asia 88 123 23 45 1 5 gion was 667.1 thousand in 1983 (World Bank 1988b). South Asia 68 123 24 32 4 5 This was 80 percent of enrollments (in 1984) in Korea. Louth Aserica 68 108419 24 32 20 As a proportion of the population, Africa has 0.16 Latin Ameinca 98 108 19 48 4 20 percent and Korea 2.07 percent, Hong Kong and countries 88 106 21 52 5 18 Thailand about 0.6 percent and Brazil 1.1 percent. Source. WDR, 1989, Table 29 The highest enrollments in Africa are registered by Zaire, Cameroon, Malawi and Gabon, with most of 270 Human Resources, Technology and Industrial Development in Sub-Saharan Africa the large industrializers lagging in training their edge (Lall 1987 1990b). It is very difficult to measure workers in basic technical skills. this kind of effort applied to routine production or Whi'le some of these data may not be accurate or investment activity. The best proxy is the incidence strictly comparable, their broad implications are of engineers and technicians in the workforce, but it clear. "The educational structure of Sub-Saharan Af- is a crude proxy. The evidence adduced earlier sug- rica is unsuitable for industrialization" (McMahon gests that relatively little of such effort is undertaken 1987, p. 19). If the data were adjusted for quality, and in most African countries. The largest repositories of if firm-level training were taken into account, it is technological skills are probably long established likely that Africa would fall further behind the NICs firms with strong technical links abroad, e.g. affiliates of East Asia and all the indications are that the gap is of multinationals or local firms with good foreign growing larger over time. management and technicians. The low level of human capital relevant to industry The component of technological effort which is suggests why the region presents a general picture of most easily measurable is fornal research and dev- poor technological mastery and dynamism in indus- elopment. This is likely to be a small part of the tot- try. The exceptions that exist are explained precisely al effort needed in most developing countries to mas- by efforts made at the firm-level to create skills and ter imported technologies. However, it is increas- train workers. However, such efforts have not been ingly a critical input. As more complex technologies widespread or intensive enough to make up for the are imported, and as older technologies get fully general scarcity of skills. Firms are always reluctant mastered, local research and development becomes to invest heavily in training when there is a risk of essential for assimilating, adapting and improving thebeinefits of theirinvestment "leaking out" to other on these technologies (see Cohen and Levinthal firms if workers leave. This is a classic case of market 1989). failure for which economists recommend subsidies Data on research and development in Africa are for training or govemment sponsored training. In patchy. The data available here cover research in all Africa. such interventions do exist, but they are not its forms (in agriculture and other nonindustrial sec- sufficient to ensure firm level training comparable to tors), and it is likely that in the region much of it is that undertaken, say, in Korea (where 5 to 6 percent not devoted to manufacturing objectives.6 Moreover, of firm turnover is required to be spent on training, the data do not distinguish between research and McMahon 1987).5 development conducted by manufacturing enter- Firn level training is not, of course, a substitute for prises as opposed to government laboratories sepa- the education system but a complement to it. If the rated from industrial activity. system does not produce the base of literacy or for- The data show that the levels of total research and mal training needed for industrial capabilities, firns development in Africa are generally low (some of the cannot create such abase. African industry must, in figures, as for Togo, are dubious). The numbers of consequence, operate with a small layer of relevant scientists and engineers involved in research and capabilities which have to be spread very thinly. It is development are also low, though Ghana, Mauritius someithing of a paradox, then, that this situation (of and Sudan stand out as exceptions when these are much of industry substituting with inadequately deflated by total population (they exceed Thailand trained employees) coexists in many countries with and approach the levels reached by Brazil). The av- unemployed engineers and technically qualified per- erage for the region, 49 scientists and engineers in sonnel in nontechnical occupations (Bennell 1984). research and development per million population, is This inability to exploit the potential capabilities that well below the norm for developing countries as a existisitselfareflectionofthelowlevelofmanagerial whole (127) and is below 4 percent of the figures competence, and of the lack of competitive pressures achieved by Korea and Taiwan. However, as far as in mcst African economies and some others. To some manufacturing is concerned, these greatly overstate extent it is also a reflection of the scarcity of technical the technological effort in Africa (see note 4). personnel-they tend to give up "dirty" production The low intensity of formal technological effort in jobsand move intoeasyadministrative ones (Bennell Africa is entirely expected and given its early stage 1984). of industrial development, is notentirely inappropri- ate. A booming export oriented economy like Thai- Technological effort. No enterprise can achieve effi- land has managed well until now by relying pas- ciency, even if it is well endowed with skilled em- sively on imported technology for most of its indus- ployees, if it does not undertake conscious, directed trial needs. It is in the longer term, with growing effort to collect and assimilate new technical knowl- deepening of the industrial structure, that research 271 Human Resources, Technology and Indutrial Development in Sub-Saharan Africa and development in industry becomes a significant and starved of funds. The inward orientation of, and determinantof competitiveness. At the stagereached other interventions in, the economies reduce the in- by most African industiy, what is more relevant is centive for private agents to find institutional solu- production related technological effort and capabil- tions of their own. ity development. This is where countries like Thai- land and the East Asian NICs have established a Conclusions and Policy Implications distinct lead. Zymelman (1990) surveys the African record of Successful industrial development requires a con- scientific effort as revealed by scientific publications. ducive macroeconomic setting, an adequate supply He finds that Africa's share of scientific publications of foreign exchange, and adequate infrastructure. in the world's total (0.4 percent) is much lower than Policy makers must address the entire spectrum of its share of population (8.5 percent), and has lower- determinants of industrial performance if they wish than-average citations per article. Publications are to broaden, deepen and improve their industrial highly concentrated by country (Nigeria, Kenya and base. The strategies open to different countries are Sudan account for 70 percent), and byfield, with only bound to differ, but each country can act on the three 3 percent of publications in engineering and technol- main factors described above in some fashion. ogy (medicine and biology account for 82 percent). It is generally agreed that trade and industrial In this field, the most relevant for industry, Africa policies in Africa have been too inward looking, in- accounts for only 0.15 percent of relevant world pub- terventionist and oblivious to needs of efficiency. lications in 1986. They have permitted too many "white elephants" to come into existence and have held back potentially Institutions competitive activities from realizing their potential. Ownership patterns have been skewed to promote The development of industry related institutions Africanization faster than is economically efficient. (defined narrowly here to include those which facili- Loss making enterprises have been kept in operat- tate the functioning of markets and the development ion for too long. Domestic impediments placed on of capabilities) is central to abroad based industriali- competition and growth have added to the con- zation process. Firms cannot function efficiently as straints imposed by high levels of indiscriminate isolated units; they have to establish a variety of protection. strong linkages with the rest of the economy. The While the correct policy response in most African economy, in tum, has to provide a variety of inputs, countries would be to liberalize on domestic and services, information and infrastructure, and stand- foreign competition, allow freer entry to potential ards and rules, to enable firms to produce, invest and investors and ensure that incentives become more grow. The primitive market structures with which neutral between domestic and foreign markets, this developingcountries start on industrialization suffer may not amount to a case for free trade or even for widespread structural deficiencies in furnishing all low, uniform rates of infant industry protection. these linkages and services. Some deficiencies are There is a good theoretical and empirical case for remedied by private agents in response to market intervening in market driven incentives to build up signals. Others require direct government interven- competitiveness in progressively more difficult in- tion. Still others call for the setting up of permanent dustries. Without such intervention, comparative ad- institutions to play facilitating roles, with autono- vantage may well stay quite static. mous status, specialized skills, and market account- The arguments against such selective intervention ability. are well known. In essence they come down to three: Some relevant institutions exist in several African (a) governments maybe nobetter, and in practice can countries, but the general scarcity of trained labor be much worse, than markets in "picking winners"; seems to have held back proper institutional devel- (b) the costs of entering new activities should be opment. Limited support is provided to manufactur- borne by capital markets (which should be tackled ers in terms of technical services, training, informa- directly if they funcfion imperfectly), while other tion or standards. Interlinkages between large and arguments for protection dynamic and unpre- small firms are minimal, and are not facilitated by dictable learning externalities, complementarities) appropriate institutional assistance to potential are trivial; and (c) if a selective intervention goes small suppliers or subcontractors (World Bank wrong, it is difficult to rectify because vested interests 1 989b). Where institufions exist, they are often poorly build up in preserving "losers". There is clearly some staffed and managed, given conflicting objectives, merit in these arguments. 272 Human Resources, Technology and Industrial Development in Sub-Saharan Africa The arguments against selectivity can, neverthe- in the African context. But perhaps this is too pessi- less, be carried too far. The problem with past inter- mistic. There are countries with considerable indus- vention has been non-selectivity and lack of strategy trial potential and past policy achievements. Instead rather than selection and strategy based on experi- of urging them to withdraw from selectivity alto- ence, analysis and economic evaluation. Wholesale gether (which may be unrealistic in most cases), it import substitution did not pay much attention to the may be most productive to guide them to a level of process of reaching international competitiveness. selectivity which is least prone to "hijacking' and Selective intervention in Korea, by contrast, concen- creates the least long term distortions. trated on a few activities at a time, enforced a degree It would appear that the area of capabilities is the of market discipline by forcing export orientation in area with the greatest need for government action in combination with domestic protection, and moni- Africa. It is beyond the scope of this paper to analyze tored the process of maturation closely (closing the means by which general educational quantity, down or restructuring emerging losers) (Pack and quality and relevance should be improved in Africa Westphal 1986). More important, since protection (see World Bank 1988b, Zymelman 1990, Middleton itself is not a remedy for high costs or poor quality and Demsky 1989). It is relevant, however, to men- arising from poor training, institutional weakness, tion some measures which apply more narrowly to deficient infrastructure or lack of extemal support, industrial skills and technology (Freeman 1989 has a the process of selective maturation requires a broader brief and succinct review of the problems of latecom- strategy which includes capability building and ers to industrialization which has relevance to these other supply-side measures as well as protection. issues). Where the supply side elements are missing, protec- It is essential that capability development in Africa tion only generates social cost. It sustains and en- be based on the present endowment of skills and courages inefficiency, since firms cannot overcome social structures (Hyden 1983), and that it aim at a handicaps arising in the external environment, and healthy and competitive industrial structure. Formal slowly slip into slothful routines. education and training are clearly needed for all The preconditions for the efficient deployment of levels of African industry. Though in a lot of infornal selective intervenfion are quite demanding. It re- and small-scale industry there is not much formal quires a strong and competent government which is need for technical training, literacy and some famili- driven by economic objectives. This government arity with modern technologies are clearly impor- should be able to analyze technological information tant. Modern small-scale industry can, on the other and select only activities which, given resource, skill hand, involve highly sophisticated operations, and and technological resources (and the "state of na- sustained expansion and upgrading of this sector ture") can become efficient in a reasonable period. It will inevitably call for the expansion of vocational (or should be able to monitor progress, ensure entry to similar) training and high level technical education. foreign markets and a degree of foreign competition Formnal large-scale industry needs far higher inputs at home, act to remedy errors or close down losers. of well trained workers, technicians, managers and Most importantly, it should be able to progressively engineers. It is imperative for governments to pro- boost the capabilities and institutions that determine vide such training at the right level and cost, and of industrial efficiency. the right quality (Bennell 1984, Dougherty 1989, Many of these conditions may not be satisfied in World Bank 1988b and 1989b). the majority of African countries. Governments may For instance, firms need help, at least in the initial not be strong or objective enough to manage such stages, in acquiring information on technology, strategies, and the existing structure of political econ- equipment, materials, consultants, and so on. Certain omy may "hijack" sensible economic strategies technological functions cannot be undertaken in- (Biggs and Levy 1990). They may lack the human house because they are too "lumpy' or have too resources to mount the analytical, surveillance and many externalities (public goods). Thus, only the promotion efforts needed. The given endowments govemment can set up standards or testing institu- with which they have to work may be so small that tions, provide technical extension or conduct basic much of modern industry maybe out of reach for the research. Only the government can establish a sci- foreseeable future. Under these conditions of high ence and technology infrastructure linking industry risk of "government failure" or insuperable struc- with laboratories, consultants, universities and for- tural constraints, it may be best to "leave it to the eign entities. In many African countries, where even market". The costs of past government failures are so basic quality control or preventive maintenance high that many analysts veer to this view, certainly skills are lacking, functions that are normally per- 273 Human Resources, TechnoloSy and Industrial Development in Sub-Saharan Africa formed in-house may be "externalized' and pro- considerble ingenuity and dynamism in some met- vided by common units. This would conserve skills alworking jobs (utensils, tools) and repair (automo- and provide "tutors" to in-house operators. biles). If the level of skills and know-how can be One important way to stimulate local technologi- raised to that required for more modem (small-scale) cal activity and diffusion is the promotion of consult- industrial engineering activity, this would provide a ancyorganizations. Consultantsarethemostimpor- dynamic growth sector as well as facilitate the tant means of transferring technology in several in- growth of other sectors. dustries (mainly with continuous processes). They Subcontracting and interfirm linkages in general are also repositories of technical knowledge gathered have been slow to develop in African industry. This from experience in a variety of enterprises. Fostering reflects the lack of capabilities on the part of both the growth of consultants in more advanced fields large firms (whichhavetoexpend considerableeffort often requires protection in the forrn of restricting the and transfer a lot of skills and technology in setting entry of foreign consultants or ensuring a twinning up subcontractors) and potential suppliers (which arrangement between foreign and local consultants. lack the entrepreneurial or basic technical skills to Like all protection, this is a tool which needs to be appear viable as reliable sources). It also reflects poor used with great care and discretion. In Africa, a com- infrastructural facilities for small firms, biases in pol- parative advantage in consultancy services can only icy and credit markets and the lack of a technical be developed in the simple end of the spectrum of extension network. Governments should try to rem- activities, concentrating on technologies which are in edy these deficiencies and correct existing biases, common use (construction, food processing, textiles, while encouraging large firms to establish local sub- cement). While most consultants should be spon- contracting systems. Needless to say, this potent tool sored in the private sector (a common source of for diffusing technology and skills and promoting consultancy organizations is the specialized project small-scale industry should not be used too bluntly. divisions of large manufacturers), some may have to Forcing the growth of inefficient local suppliers is not be set up in the public sector to provide technical conducive to healthy development. assistance to small enterprises on a subsidized basis. The broader economic setting should not be for- Again, local consultants maybenefit from "tutoring" gotten (Gulhati and Sekhar 1982). "The process of by experienced foreign experts recruited for limited developing industrial capabilities is self-reinforcing, periods. with different elements interacting to support each A promising path to promote institutional devel- other. The general industrial environment affects its opment in the technology and training area is to help content and direction: a competitive, outward-look- industry associations set up facilities for their mem- ing regime is likely to call forth an appropriate set of bers. In Korea, industry associations provided many technological responses. If macro policies and incen- valuable technical functions in this way, and also tive structures are improved, and sufficient foreign acted as interlocutors for the implementation of gov- resources provided to enable Africa to resume a strat- emnment policy, export targeting and feedback of egy of long term growth, the ultimate determinants informationr Well staffed and funded associations of how successfully it industrializes will be its edu- can be a substitute for several types of institutions cation and training systems and increasingly its sup- that would take much longer to evolve. ply of technical labor and firm level investments in Two areas of industry call for special strategies. technological effort. Africa as a whole lags well be- The first is the engineering industry, the second is hind the rest of the developing world in these areas. subcontracting between large firms and small sup- Even the "best' countries in Africa are very far from pliers. Basic metalworking skills are generic to a the best elsewhere, i.e. the NICs of East Asia. whole host of other industries, and the ability to The wide dispersion of industrial development build, copy, repair or improve capital goods (even of within Africa will mean that future paths may di- a simple sort) is widely regarded as the seedbed and verge even further. The cumulative nature of capabil- but of technological progress (Fransman 1986). Such ity development dictates that the better-off countries skills are scarce in most African countries, which will, ceteris parabus, continue to industrialize more adds to thecountries'inabilityto utilize theirexisting efficiently than others. If they plough back part of capital stock efficiently or to raise its productivity theirgrowth in revenues into developinghumanand over time. The inability to make spares locally or technological resources, their lead will increase fur- carry out troubleshooting services adds greatly to ther. Some countries with greater access to foreign operating costs and downtime (when equipment lies exchange will be able to compress the learning curve idle). At the same time, the informal sector shows by attracting foreign "tutors" and better technology. 274 Human Resources, Tednology and Industrial Development in Sub-Saharan Africa Those with significant local supplies of non-African Cohen, W.M. and Levinthal, D.A. (1989), -Innovation and capabilities may similarly improve their position if Learning The Two Faces of R & D." Economicjournal 99, they give full rein to those capabilities. Other coun- pp. 569-96. tries, with neither foreign exchange nor large rep- Dahlman, Cj., Ross-Larsen, B. and Westphal, L.E. (1987), ositories of non-indigenous skills, may continue to "Managing Technological Development: Lessons from lag industrially, whatever incentive policies they Newly Industrializing Countries." World Development adopt. TL hey will go up the learning curve only at the 15:6, p. 759-75. speed their own human capital development per- Dougherty, C. (1989), 'The Cost Effectiveness of National mits. Training Systems in Developing Countries.' World Bank, PPR Working Papers, WPS 171. Notes Fransman, M. (ed) (1982), Industry and Accumulation in Africa. London: Heinemann. 1. All these data are from World Bank 1989b. Fransman, M. (1986), Technology and Economic Development. 2. For instance, some 70 percent of top managerial and Brighton: Wheatsheaf Press. technical positions in industry in Cote d'lvoire and Zim- Freeman, C. (1989), 'New Technology and Catching Up" babwe are held by Europeans (World Bank country in C. Cooper and R. Kaplinsky (eds.), Technology and reports). Deuelopment in the Third Industrial Revolution. London: 3. This subsection draws heavily on Lall (1990b and Cass, p. 85-99. forthcoming). Gulhati, R. and Sekhar, U. (1982), "Industrial Strategy for 4. Zymelman (1990) Notes that Africa graduates only 1.2 Late Starters: The Experience of Kenya, Tanzania, and people per 100,000 population in science and. engineering Zambia." World Development. 10:11, p. 949-72. per annum, compared to 65 in industrialized countries. Hayden, G. (1983), No Shortcuts to Progress: African Devel- Engineering by itself is even weaker: "developed countries opment Management in Perspective. London: Heinemann. graduate 166 times more engineers per capital than do Katz, J. (ed.) (1987), Technology Generation in Latin American Sub-Saharan African countries" (p. 25). Manufacturing. London: Macmillan. 5. As BAS (1989) Notes, on-the-job training is very weak Kilby, P. (ed.) (1971), Entrepreneurship and Economic Devel- in much of modern African industry, while traditional opment. New York: Free Press. forms of apprenticeship training are inadequate for impart- King, K. (1984), "Science, Technology and Education in the ing modern technical skills. Development of Indigenous Technological Capability", 6. UNESCO (1989) data on four African countries in M. Fransman and K. King (eds.), Technology Capability (Congoz, Malawi, Mauritius and Zambia) show that of R&D in the Third World. London: Macmillan, p.31-64. perfonrmed in the "productive sector", agriculture, mining, Lall, S. (1987), Learning to Industrialize: The Acquisition of utilities or construction, accounted for all spending or em- Technological Capabiity by India. London: Macmillan. ployment of scientists and engineers. Manufacturing did not Lall, S. (1989), "Human Resource Development and Indus- register any R &D at all in these countries. No other African trialization, with Special Reference to Sub-Saharan Af- countries were included in these tabulations. rica." Journal of Development Planning 19, p.129-58. Lall, S. (1990a), "Structural Problems of Industry in Sub-Sa- References haran Africa." in Background Papers: The Long-Term Per- spective Study of Sub-Saharan Africa. (Vol. 2), World Bank. Bas, D. (1989), "On-the-Job Training in Africa." Interna- Lall, S. (1990b), Building Industrial Competitiveness: New tional Labour Review 128:4. P. 485-496. Technologies and Capabilities in Developing Countries Paris: Bennell, P. (1984), "The Utilization of Professional Engi- OECD. neering Skills in Kenya", in M. Fransman and K. King Lall, S. (forthcoming), "Explaining Industrial Success in the (eds.) Technological Capability in the Third World. London: Developing World." in VN. Balasubramanyam and S. Macmillan, p. 335-54. Lall (eds.) Current Issues in Development London: Macmil- Biggs, T. and Levy, B. (1990), "Strategic Interventions and lan. the rolitical Economy of Industrial Policy in Developing McMahon, W.W. (1987), "Education and Industrializa- Countries." in D. Perkins and M. Roemer (eds.), Economic tion," World Bank, Background Paper for WDR 1987. Systems Reform in Developing Countries.Boston: Harvard Meier, G.M. and Steel, W.F. (1989), Industrial Adjustment in University Press. Sub-Saharan Africa, New York. Oxford University Press, Chenery, H.B., Robinson, S. and Syrquin, M. (1986), Indus- EDI Series in Economic Development. trialization and Growth: A Comparative Study. New York: Middleton, J. and Demsky, T. (1989), "Vocational Educati- Oxford University Press, for the World Bank. onand Training." World BankDiscussion Papers, No.51. 275 Human Resources, Technology and Industrial Development in Sub-Saharan Africa Mlawa, H.M. (1983), "The Acquisition of Technology, Tech- Stewart, F. (1977), Technology and Underdevelopment Lon- nological Capability and Technical Change: A Study of don: Macmillan. the Textile Industry in Tanzania." University of Sussex, Stewart, F., Lall, S. and Wangwe, S. (eds.) (forthcoming), Dissertation. Alternative Development Strategies in Africa. London: Mac- Mytelka, L. (forthcoming), "Ivorian Industry at the Cruss- millan. roads." in Stewart et al. UNESCO (1989), Statistical Yearbook 1988, Paris. Navaretti, G.B. (forthcoming), "Joint Ventures and Autono- Wangwe, S. (forthcoming), "Building Indigenous Techno- mous Industrial Development: The Magic Medicine?" in logical Capacity." in Stewart et al. Stewart et al. World Bank (various years), World Development Reports. Nellis, J. (1986), "Public Enterprise in Sub-Saharan Africa", New York Oxford University Press. World Bank Discussion Paper, No. 1. World Bank (1981), Accelerated Development in Sub-Saharan Pack, H. (1987),Productivity, Technology and Industrial Devel- Africa. Washington, D.C. opment. Oxford University Press for the World Bank. World Bank. 1987. World Development Report 1987. New Pack, H. and Westphal, L.E. (1986), "Industrial Strategy York: Oxford University Press. and Technological Change: Theoryversus Reality." Jour- World Bark. 1988a. World Development Report 1988. New nal of Development Economics2l, p.87-128. York Oxford University Press. Page, J.M. (1979), "Small Enterprise Development: Eco- World Bank. 1988b. Education in Sub-Saharan Africa: Policies nomic Issues from African Experience." World Bank for Adjustment, Revitalization and Expansion. Wash., D.C. Staff Working Paper, No. 363. World Bank. 1989a. World Development Re,ort 1989. New Page, J.M. (1980), 'Technical Efficiency and Economic Per- York: Oxford University Press. formance: Some Evidence from Ghana," Oxford Economic World Bank; 1989b. Sub-Saharan Africa From Cisis to Sus- Papers 23:2, p. 319-39. tanable Growth. Washington, D.C. Steel, W.F. and Evans, JW. (1984), "Industrialization in Zymelman, M. (1990), Science, Education and Development in Sub-Saharan Africa: Strategies and Performance', World Sub-Saharan Afric World Bank Technical Paper No. 124, Bank, Technical Paper No. 25. Africa Technical Department Series. 276 26 Entrepreneurship and Growth in Sub-Saharan Africa: Evidence and Policy Implications T Ademola Oyejide Against the background of poor economic performn- This paper is aimed at making a modest contribu- ance during the 1970s and 1980s, many African tion to this broad research agenda. Its primary pur- countries have implemented ongoing structural ad- pose is to survey the available literature as a means justment programs. In most parts of Africa there is of identifying major elements of the dynamics of increasing emphasis on private initiative and market private enterprise development in Africa. In the sec- forces as a means of reviving and restoring sustain- ond section the main characteristics of African entre- able economic growth. It is presumed that this new preneurship are analyzed. The third section exam- orientation in economic management will improve ines the patterns of upgrading, graduation, and the flexibility with which African economies can re- growth in African entrepreneurship. Section four spond to rapid and unexpected changes in the contains a discussion of the constraints impeding intemational economic environment. The new orien- enterprise development and growth. Emerging pol- tation also implies an expanded role for the private icy issues are identified and concluding remarks are sector in the development process. offered in the fifth section. There have been some fundamental changes in both the macroeconomic environment and the regu- Characteristics of African Entrepreneurship latory framework in many African countries in recent times. However, there is increasing concern that the Typically, the private sector in an African economy supplyresponse of private sectorenterprisesappears contains a wide variety of enterprises including in- to be largely muted and uninspiring. This raises the formal enterprises; microenterprises; and small-, me- question of whether there are enough entrepreneurs dium-, and large-scale modem firms. This section is of the appropriate types to be able to respond ef- concemed with the ownership and size distribution fectively to the newly created opportunities. It also of these firms, their capabilities for development, and calls for study of the determinants of African entre- their overall economic growth enhancing capabili- preneurship and its capability for response, and ties. analysis of the factors which constrain the develop- ment of entrepreneurship and the private sector in Ownership and Size Distribution Africa. Such study and analysis would provide im- proved knowledge of the origins, characteristics, The size distribution of private sector enterprises constra- ints, and growth potentials of the private in Africa features several distinct characteristics. enterprise sector. There isanoverwhelmingpredominanceof informal 277 Entrepreneurship and Growth in Sub-Saharan Africa: Evidence and Policy Implications enterprises and microenterprises, which generally marketing. As a result . .dence abounds that many produce simple consumer goods. These enterprises of the important functions associated with the effi- are not integrated with the formal sector of modem cient management of modem business enterprises small-, medium-, and large-scale finns. These enter- are either not performed at all orbadly carried outby prises account for the largest share of employment in the typical African craftsman-entrepreneur who the private sector. Surveys indicate that firns with owns and runs an informal or small-scale enterprise. less than 10 employees provide 59 percent of total Decision making is often excessively concentrated private sector employment in Kenya, 75 percent in in the hands of the owner-manager, whose authority Nigeria, 83 percent in Zambia, and 90 percent in Sie- is rarely delegated. In addition, there is limited pro- rra Leone (Page 1979, Liedholm and Mead 1986, Kil- pensity to form partnerships or take advantage of the by 1988). The share of total employment accounted limited liability public corporate form. Rather, the for by firms with between 10 and 50 workers aver- overwhelming majority of small African enterprises ages less than 10 percent in many African countries. is organized as sole proprietorships (Marris and Indigenous capital accounts for a small share of Somerset 1971, Akeredolu-Ale 1975, Kennedy 1980). total private sector capital. Local capital is heavily This organizational form of ownership and control concentrated in the informal and small enterprises. generates growth retarding enterprise succession Foreigners, resident aliens, and the state control the problems. medium- and large-scale firms (Kilby 1988, Levy Surveys show that the majority of African enter- 1990). These studies also show that African capital prises do not keep adequate financial records and accounts for 20 percent of manufacturing capital in that financial controls do not feature prominently in Nigeria, 80percentof outputand20percentof capital the management of the enterprises (Kilby 1962). in Ghana, and 9 percent of value-added in Kenya. There is a lack of experience, formal education, and Various attempts have been made in the past to training in record keeping, compilation, and use of modify this ownership pattern. An important exam- numerical data. Even when rudimentary accounting ple is the Nigerian case in which an indigenization records are maintained, they are often deficient and program carried out in the 1970s transferredbetween not frequently used as a management tool (Kilby 40 and 100 percent of the equity capital of many of 1962). This, in tum, leads to overcapitalization, main- the medium- and large-scale modern enterprises to tenance of excessive stocks of raw materials, and local shareholders. The subsequent privatization incomplete separation of business from household program is in the process of sharply reducing the and personal accounts. The managerial functions re- state's share in the transferred equity. The proportion lating to marketing and promotional activities re- of the total capital of Nigerian enterprises owned by ceive very little attention; market coverage tends to private citizens and associations should increase sub- be limited, thus hampering enterprise expansion and stantially. A recent study laments the lack of adequate growth. research on the growth of larger indigenous enter- prises in Nigeria. It argues "that in terms of the scale Technological Capability of individual enterprise, the degree of corporate or- ganization and the size and diversity of investment, Owners of African informal, micro, and small- Nigerian private capital has advanced well beyond scale enterprises tend to be entrepreneurs who have African enterprise in Kenya, the Ivory Coast, Zim- emerged from the ranks of craftsmen. Their skills babwe and other Sub-Saharan African countries" have usually been acquired through a process of (Forrest 1989, p.1.). apprenticeship, the length and formality of which reflect the extent of technical barriers to entry into the Managerial Capability particular product group or industry. Surveys reveal that between 60 and 90 percent of such entrepreneurs Various studies of African private sector enter- had received training as apprentices before setting prises have identified a low level of managerial skills up their own enterprises (Gerry 1974, Aryee 1977). as one of the major deficiencies in the private sector. The system of apprenticeship is reported to be more Owners of informal, micro and small-scale enter- formalized and older in West Africa than in the East- prises tend to be craftsmen-entrepreneurs who have em, Central, and Southern parts of Africa. probably acquired a level of technical proficiency in The heavy reliance on this traditional system of their production processes. However, they are apprenticeship for the acquisition and accumulation largely without any extensive or formal training in of technical skills implies that these skills are not business organization, financial management, and easily transferable to subsequent activities in the for- 278 Entrepreneurs hip and Growth in Sub-Saharan Africa. Evidence and Policy Implications mal medium- and large-scale enterprises. Similarly, simultaneously, enterprises may go thioughagradu- skills learned in the formal sector do not appear to ation process as they are transformed from micr- forrn. a basis on which small-scale enterprises are oenterprises and small-scale enterprises into me- built. Thus, a kind of technological dualism seems to dium- and large-scale firms. Both of these constitute exist in which entrepreneurs with training and expe- important features of entrepreneurship develop- rience in the modem sector are concentrated in small ment in Africa. forrn.al sector finns, while less than 15 percent of It has oftenbeen suggested in the literature that the infornal and microenterprise owners claim prior ex- informal sectorand microenterprises provide a valu- perience in the modem medium- and large-scale able training ground for advanced entrepreneurial industry. and managerial capabilities upon which more mod- The level of formal educational attainment among em medium- and large-scale firms are established informal and microenterprise owners and managers and sustained (Page 1979). To the extent that informal is astonishingly low. Surveys indicate that estimates and small-scale enterprises play a significant mle in of the proportion of African entrepreneurs lacking the building of human capital, they would form an formal education range from a low of 13 percent in important bridge for the development of entrepre- Nigeria to a high of 77 percent in Sierra Leone. Fur- neurship in Africa. thermore, it is estimated that an average of about 50 In practice, however, the role of the informal sector percent of the entrepreneurs surveyed had not in augmenting the supply of entrepreneurial and achieved functional literacy. Less than 15 percent of managerial skills for the overall economy is probably these entrepreneurs had attended technical or voca- limited by several considerations. In so far as infor- tional schools prior to the establishment of their en- mal sector enterprises and microenterprises are pre- terprises. The low level of functional literacy and dominantly owned and managed by craftsmen-en- technical training probably accounts, to a significant trepreneurs, the types of skills they embody may not extent, for the rudimentary nature of the technology be readily transferable to the management of modern employed in the production processes of many Afri- medium- and large-scale firms. This limitation could can informal sector and microenterprises. arise from the fact that enterprise growth in the mod- em sector requires skills that are quite different from Enterprise Upgrading, Graduation, and Growth those possessed by craftsmen-entrepreneurs whose primary domain is the informal sector. In particular, From the point of view of enterprise development, their lack of formal education and failure to acquire the primary importance of the informal microenter- adequate training in financial management and prises and small-scale enterprises derives from their other organizational skills would limit the size and actual and potential growth dynamics. These enter- complexity of firms which they could efficiently run. prises contribute to the overall growth of the econ- In any case, there apparently exists a strong ten- orny through increased efficiency and productivity, dency on the part of informal sector entrepreneurs to income generation, and employment creation. The diversify their holdings of micro and small-scale en- number of enterprise start-ups and closures and the terprises rather than to concentrate resources to- growth of enterprises are indicators of the develop- wards expanding a single enterprise in the direction ment of entrepreneurship. of a larger and more efficient size (Kilby 1962, Marris Enterprise growth has several dimensions. Firms and Somerset 1971). Such growth by diversification may growinnumber,in size, and by diversifying into could represent a rational response on the part of different product lines. Each of these changes may small enterprises to a structure of incentives which is occur with or without increased linkages between hostile to growth in firn size. However, diversifica- firms. Probably growth in firm size would be most tion tends to detract the entrepreneur's attention conducive to increased efficiency, productivity, and from efficiency concemns. It might be hypothesized, income generation. Therefore an analysis of growth therefore, that modern small-scale firms owned by oriented entrepreneurship development should fo- individuals who are more familiar with market op- cus largely on this aspect of enterprise growth. But portunities and who have acquired greater educa- even when the growth of enterprises is so narrowly tional, commercial, and managerial experience are conceived, it is important to recognize that this likely to constitute a more effective bridge for enter- growth can be accomplished through two interre- prise development than the craftsmen-entrepreneurs lated but separate routes. Firms can experience a in the inforsnal sector. transition from an informnal to a formal status in the The proportion of modern small- and medium- process of expansion and growth. Subsequently or scale enterprises in Africa which came from the ex- 279 E"treprcnrship and Growth in Sub-Saharan Africa: Evidencs and Policy Implications pansion of microenterprises in the informal sector is * ability to market the goods and services pro- not entirely clear. Aggregate data on the size distri- duced by the enterprise, bution of enterprises are not sufficiently revealingon * ability to progressively improve the technical the issue of the pattern of upgrading and graduation. capabilities of the enterprise, and But a broad interpretation of the evidence from four * ability to manage relations. African countries indicates that in Nigeria, Rwanda, Botswana, and Sierra Leone, modem small- and me- Within this framework, three sets of obstacles to dium-scale firms did not graduate from micro begin- enterprise development can be identified. These are: nings, but rather started life autonomously. Appar- a shortfalls in the supplyof individualswith the ently some regional differences can be discerned attributes of entrepreneurs, from the available data. In Western Africa over 30 * weaknesses in the external economic environ- percent of the modem small- and medium-scale ment within which enterprises operate and firms trace their origins to microenterprises. Twenty associated high costs of gaining access to re- percent or fewer of such firms graduated from micro sources and markets, and ranks in Eastem and Central Africa. Compared with * regulatory environments which make enor- Asia and Latin America, the rate of upgrading of mously burdensome demands on enterprises smaller firms (those with between one and ten em- in their relations with government bureauc- ployees) to larier firms is much lower in Africa. racy. Enterprise Development Constraints The factors which relate to the internal capabilities of the entrepreneur (i.e. entrepreneurial, managerial, Constraints against enterprise development pre- and technological) can be separated from the factors vent informal enterprises from acquiring formal which relate to external constraints (i.e. the macro- status and hamper the ability of microenterprises economic environment and regulatory framework). and small-scale firms to graduate into medium- and Many studies have concluded (e.g. Levy 1990) that large-scale units. The identification and classification there does not appear to be a serious shortfall in the of these constraints hasbeen approached in a number supply of indigenous entrepreneurs. Hence, the fac- of different ways in the literature. For instance, Lall tors responsible for the underdevelopment of Af- (1989) offers a comprehensive approach to the deter- rica's private sector will probably be found in the minants of enterprise growth and nerformance pri- macroeconomic and regulatory environments. marilyintheformofacomplexinterplayofthreesets These factors form a barrier which hinders the of factors, i.e. incentives, capabilities, and institu- transition of enterprises from informal to formal tions. In this approach, incentives cover macroeco- status, hampers the graduation of microenterprises nomic conditions, physical infrastructure, and int- and small-scale firms into medium- and large-scale ernal competition. Capabilities include the ability of units, and slows down the rate of start-ups of me- entrepreneurs to launch physical investments, the dium-scale enterprises. It would be useful to identify provision of human capital (in terms of skill, training, the policy induced distortions and institutional fac- and ability to absorb and applynew knowledge), and tors which constitute critical elements of this barrier. the undertaking of technological effort. Institutions Analytically, the barrier can be represented as a include entities established to facilitate the working hump in the typical firm's unit cost curve in relation of markets and the network of extemal linkages to its size. The transition (from informal to formal) which allow individual firms to operate efficiently. and graduation (from small- to medium-scale) issues Another set of identification and classification cri- are captured, in varying degrees, by the need to teria emanates from the functional approach to en- overcome the difficulties created by the hump. The terprise development which focuses on thebundle of existence of a hump in the firm's unit cost curve may activities that must be performed in a successful also explain the reaction of entrepreneurs whose en- enterprise (Levy 1990). These include: terprises grow primarily by diversification into * initiative to identify and respond toopportu- manyproductlines.Thismaybeduetotheperceived nities for profit, difficulties of going over the hump either by becom- * discipline and skill to manage an enterprise so ing formal or by expanding the scale of operations in as to maximize revenues and minimize costs, the same product line. * ability to gain access to finance and other in- An analysis of the firm's enabling environment puts (e.g. labor, raw materials, equipment, (i.e. macroeconomic, institutional, and regulatory etc.) needed for production, framework) should provide some pointers which 280 Entreprnclnship and Growth in Sub-SaharanAfric Evidence and Policy Implications would endow the hump hypothesis with more (and straint to dynamism on the part of indigenous small testable) specificity. Several elements of the enabling and medium enterprises than do weaknesses in the environment may be presumed to contribute to the external economic environment (especially in the fi- locationand magnitude of the hump in relation to the nancial sector). Other studies (e.g. Schatz 1965) re- size and type of the enterprise. For instance, financial gard capital shortage as an illusionary constraint to market distortions may playa role. Credit subsidiza- enterprise development. tion and non-price rationing policies generally favor An important component of the economic envi- large-scale enterprises and discriminate against (thus ronment is the degree and kind of internal competi- crowding out) smaller firms. Gaining access to insti- tion in the field of entrepreneurship in African econo- tutionalized and increasinglycheaper credit may re- mies. Lall (1989) makes a persuasive case for some quire making the transition from informal to formal protection in this area. His plea finds strong support status. This may, in turn, involve a relafively high in previous studies, particularly in Nigeria, in which up-front cost that, at the beginning, may outweigh survey results have consistently cited the retarding the eventual benefits of improved access to and re- effect of the activities of expatriate entrepreneurs on duced costs of credit. the development of indigenous enterprises through Similarly, labor market distortions emanating crowding out effects (Akeredolu-Ale 1975). It seems from minimum wage legislations or standards may that the infant industry argument for protection may play a role to the extent that the firm's unit labor cost also find an application in the more general area of increases sharply as it moves from an informal status enterprise development. to a formal status and/or graduates from a small- to medium-scale operation. This involves a movement Policy Issues from relatively heavy reliance on family cum appren- tice labor to greater reliance on wage labor. What is The accumulated evidence on Africa's informal involved is graduation into a firm size (usually 10 or sector and microenterprises suggests that few of more employees) for which minimum wage regula- these firms grow into modern firms of any size. It tions apply and are enforced. remains less clear, however, which constraints hind- Other regulatory constraints often constitute entry er the growth of these enterprises and the extent to barriers which may contribute to the sharp rise in the which specific constraints are binding. Deeper un- firm's unit cost. These include the need to obtain derstanding of the inhibiting factors and their in- various licenses and satisfy other registration for- cidence would constitute an important prerequisite malities that become mandatory as an enterprise for the formulation and implementation of policies attempts to make the transition from informal to aimed at relaxing the binding constraints. formal status and/or to graduate from small- to me- This significant knowledge gap has not prevented dium-scale. Such formalities are often a requirement African governments and the donor community for becoming eligible for certain cost-reducing bene- from mounting various types of special assistance fits enjoyed by larger firms. The costs of satisfying the programs for eliminating perceived deficiencies in requirements are normally incurred before the corre- the development of African entrepreneurship. Policy sponding benefits accrue; and on a per unit basis, actions can be taken to remove particular policy in- such costs constitute a greater burden for small than duced distortions or build up desired institutions to for large enterprises. facilitate enterprise development. But special assis- Surveys and studies of African entrepreneurship tance programs have usuallybeen targeted at comen- development have found deficiencies with respect to sating for the negative impact of given constraints. several aspects of the internal capabilities of enter- Specific assistance programs can be justified as com- prises, and key elements of the macroeconomic envi- pensatory measures in cases where the removal of ronment and institutional and regulatory frame- existing distortions is, for some important reason, not work. While entrepreneurs appear to exist in suffi- immediately feasible. Assistance programs can be cient quantity in many African countries, their mana- used as a means of off-setting extemalities emanating gerial and technological capabilities are either inade- from various types of market failures and as a means quate or inappropriate. Institutions which would en- of strengthening desirable multiplier effects. These able indigenous enterprises to perfonn efficiently are prgrams can be used to promote the development in short supply (Lall, 1989). In terms of ranking the of entrepreneurship in the face of major constraining three major sets of constraints, Levy (1990) finds that elements in the enabling environment. in Tanzania regulatory controls are extremely restric- However, the results of many previous attempts at tive. However, they constitute less of a binding con- using special assistance programs to promote entre- 281 Entrepreneurskip and Growth in Sub-Saharan Africa: Evidence and Policy Implications preneurship development in Africa have been less References than desirable. Special credit assistance programs have not been effective in many countries for a van- Akeredolu-Ale, E.O. (1975). The Underdevelopment of In., ety of reasons, including inappropriate design and digenous Entrepreneurship in Nigeria. Ibadan Univer. implementation, and inadequate proportion of en- sity Press. terprises covered. Special training programs have Aryee, G. (1977). Small Scale Manufacturing Activities: A also mostly fallen short of their targets. Study of the Interrelationships between Formal and In- Aspects of the ongoing structural adjustment pro- formal Sectors in Kumasi, Ghana. WEP Working Paper. grams in many African countries generally address Geneva, ILO. various deficiencies in the macroeconomic environ- Gerry, C. (1974). Petty Producers and the Urban Economy ment as well as the institutional and regulatory WEP Research Paper 8, Geneva, ILO. framework. But putting the right incentives in place Kennedy, Paul. 1980. GhanaianBusinessmen: From Artisan will not elicit an appropriate response until the abil- to Capitalist Entrepreneurs in a Dependent Economy. ity of indigenous entrepreneurs to respond to the Munich: Weltfoum Verlar. new opportunities is enhanced. There can be a br - Kilby, Peter. (1%2)_. The Development of Small Industries oad general approach for reforming the macroeco- in East Nigeria. Enugu: Ministry of Commerce and U.S. nomic environment and regulatory framework Agency for International Development. p. 6. Table 1. which may apply across most African countries. 1988. "Breaking the Entrepreneurial Bottleneck in Late However, policies to enhance the internal capabilities Developing Cuntres: Is there a Useful Role for Gover- of enterprises (i.e. entrepreneurial, managerial, and ment?" Journal of Development Planning, No. 18, Special technological capabilities) are more likely to be suc- issue on Entrepreneurship and Economic Development. cessful if they are specifically designed to fit the Lssue on Etpe ursipan Economi opment. pmarticular circumstances of each country. A similar Lall, Sanjaya. (1989). "Structural Problems of African In- particular circumstances of each country. A similar dustry." Presented at the workshop on Alternative De- approach is probably necessary for building up spe- velopment Strategies in Africa. Oxford: Queen Elizabeth cial institutions designed to facilitate the working of House. (1S1-13 December). markets. The formulation, design, and implementation of Levy, Brian. (1990). "Private Sector Development inTanza- such country specific policy interventions will re- nia: Obstacles and Opportunities." Washington, D.C.: quire an improved knowledge base about the ch- World Bank. Mimeo. aracteristics of African entrepreneurs and the en- Liedholm, Carl and Donald C. Mead. (1986). "Small Scale vironment in which they operate. It will also require Industry in Sub-Saharan Africa: Empirical Evidence and an approach which does not insist that owners and Strategic Implications." In Bos Bero and Jennifer Whi- managers of modern medium- and large-scale ent- taker, eds., Strategies for African Development. Berkeley: erprises should evolve from the traditional oper- University of California Press. Abstract in Meier and ators whose domain has been the informal sectorand Steel, eds. 1989. microenterprises. It should not be surprising if the- Marris, P. and A. Somerset. (1971). "African Businessman: se are eventually supplanted by a new breed of en- A Study of Entrepreneurship and Development in trepreneurs who are more literate and knowledge- Kenya." Nairobi. able in relation to both managerial and technolo- Page, John M. (1979). "Small Enterprise in African Devel- gical capabilities, and much less apologetic about the opment, a Survey." World Bank Staff Working Paper No. contributions which the private sector can and 363. Washington, D.C.: World Bank. (October) Processed. should make to the economic development of Afr- Schatz, S.P. (1965). "The Capital Shortage Illusion: Govern- ica. ment Lending in Nigeria." Oxford Economic Papers, Ox- ford, XVII, No. 2. (uly). 282 27 Comments on Human Capital and Entrepreneurship William F Steel and Samuel M. Wangue William E Steel dence of what Oyejide referr-ed to as a 'new breed of entrepreneurs" who are better educated and trained I would like to review the papers on human resources than their predecessors and who are applying their from the point of view of the prospects for industrial skills to exploit new opportunities, not simply fol- development in Africa. Although I am normally rea- lowing their par-ents' trades.2 sonably optimistic about the way industrial structure The question, however, is not whether entrepre- is adjusting to policy reforms in Africa, Imust confess neurial and labor force abilities have been improv- that these papers are rather discour-aging about the ing, but whether they are expanding fast enough. prospects for sustained growth of productivity in Several responses in Lall's paper give cause for con- industry. cern. First, he points out that "the drive to profit from T'hese papers suggest three important conclusions opportunity (in small and informnal activities) is not for- industry about human resource development: per se the same as the entrepreneurial capability r-e- substantial investment in human capital is a quired to organize, set up and run modem industry. necessary condition for industrial developme- Second, he shows that Africa still lags far behind the nt to be indigenous rather than just a transpla- rest of the world in education and training, which ntation of foreign machinery and managers; countries such as Korea have used successfully to lay *Africa's resource gap with the rest of the world the basis for technological and industrial growth. may be increasing rather than closing; and This does not mean that no progress has been made: *other competing claims on scarce public re- enrollment and literacy have improved. But they sources may drain away the investments that have continued to improve just as fast in other devel- are needed to raise pr-oductivity, oping regions as well, so the gap remains. (It would be useful if the paper could also include statistics to Before discussing these further, I should first note show how Africa today compares to Korea, Hong that there are some positive signs of progress. I agree Kong, Brazil and Thailand before they began indus- with Oyejide's conclusion that entrepreneurs exist in trializing rapidly.) Third, he argues that the techno- substantial numbers. There certainly is a long entre- logical effort is particularly weak in Africa, making it preneurial tradition in the commercial sector, espe- difficult to follow the adaptive approach that was in cially in West Africa.1East Africa also has come a long large part responsible for the success of the East wiay in entrepreneurial development since the colo- Asian newly industrializing countries. nial days when indigenous commercial activity was Even more discouraging are the indications that explicitly discouraged. Although few existing entre- the situation may be getting worse. Educational ex- preneurs may have the skills and capital needed to penditures havebeen squeezedbybudgetary proble- lead industrial tr-ansformnation, there is growing evi- ms and the quality of education has probably fallen 283 Comments on Human Capital and Enbteprenewwhip in many countries. Technical education continues to With limited resources, a strategic approach is be underemphasized and lack scientific equipment. needed, that is, some first principles must be empha- Perhaps it should count as a positive sign that we sized. These papers do not give a clear sense of where are much more aware today of the importance of to begin, although Lall argues for selective interven- early development of laborand entrepreneurial skills tion to promote certain strategic industries. He quite than policymakerswere at the timeof independence. rightly notes that this approach can impose high In fact, many leaders promoted an import-substitu- costs if the supply conditions are not right. I am tion industrialization strategy on the assumptions concerned that the selective protection he argues for that this would generate missing labor skills and that would foster those high costs in precisely the wrong the state could substitute for the apparent lack of industries: those with high linkages to the rest of the entrepreneurs.3 But in the 1980s, we woke up to the economy.Iwouldrathersuggestthateffortstostimu- fact that many state-owned industries were being late such industries concentrate on making condi- designed and run by bureaucrats, not enterprizing tions more conducive for them through direct incen- managers, and that labor productivity was falling tives, reduction of regulatory and entry barriers, instead of rising. training programs and a high level commitment to The emphasis shifted to liberalizing markets so solving their supply problems. that increased competition would force firms to be I would draw two strategic conclusions from the more efficient. In many countries, however, while papers: inefficient firms are indeed feeling the pinch, invest- * Investment in education, and especially tech- ment in more efficient firms seems slow in coming. nical training, is critical if the process of indus- Thereare entrepreneurs whowould like toadapt,but trialization is to be a flexible one of adapting most are lacking the experience, finance, marketing to change and applying new techniques. Rais- information, or technological know-how to do so. ing the skills and access to infornation of the Today, industrialization is seen as a continuous general labor force is the only effective way to process that requires substantial input of human and achieve the objective of growth through rising technological capabilities and that cannot be sus- productivitywhile pursuingthegoal of equity tained simply through protection that stimulates ca- through greater participation in the economy. pacity creation regardless of cost, on the one hand, or * It may be a strategic error to focus too much unsupported exposure to international competition, on large-scale industry as the primary source on the other. However, this more sophisticated view of industrial development in Africa over the makes it difficult to design a practical strategy. For next decade. The conditions for large indus- instance, the guidelines for the Second Industrial tries to be competitive in world markets have Development Decade for Africa mention an incred- not been adequately established and African ible range of things that need to be developed as a ba- consumers can ill afford the heavy costs of sis for industrialization. These include the following maintaining inefficient large industries for human and technological development: technical through high protection or subsidies to state schools and universities, research and development enterprises. To develop a self-reliant, sustain- bodies, consultancy services, industry associations able model of industrialization, as put forward and entrepreneurship training, among others. For in the Lagos Plan of Action and the Industrial physical infrastructure they include roads, bridges, Development Decade for Africa, much more ports, railways, telecommunications, energy, market attention is needed to building up the needed sites, storage, etc. All these are important, but the experience, skills and institutions. Much of the tremendous cost and effort involved are beyond the capability-building is likely to occur in small- means of low income, less developed countries. and medium-scale activities. S.M. Wangwe The Tall Paper haran Africa examines the background to industrial development in Sub-Saharan Africa and argues that The paper by Sanjaya Lall on Human Resources, the two conventional explanations of the problems Technology and Industrial Development in Sub-Sa- of African industry (i.e. that based on exogenous 284 Comments on Human Capital and Enfrprneurs hip shocks and that based on incentives-trade and in- Zimbabwe is only 0.005 percent, that for C6te dustrial policies) do not come to grips with the struc- d'Ivoire is 0.004 percent, and that for Nigeria is 0.005 tural factors of industrial perfornance. The author percent. It isunlikely thatthese countries havebelow develops an alternative framework based on an in- average levels of technological development in spite teraction of three sets of factors, i.e incentives, capa- of their below average shares of engineers in total bilities and institutions. population. In terms of training in science and in voc- The analysis of the interaction of the three sets of ational training, Nigeria and Zimbabwe are well be- factors permits recognition of the important role of low the average for Sub-Saharan Africa. While this the process of learning and capability development unexpected relation maypartlybe a reflection of data as a determinant of industrial success. In this sense problems, levels and pattems of utilization of scien- the framework adopted in this paper departs from tific and technological personnel could be more im- the static approaches to industrial development. The portant. process of learning and capability development has While effort is made to raise the low level of the been a central characteristic of the current industrial stock of scientific and technological personnel in revolution (since 1945) and is likely to continue to be Sub-Saharan Africa, the question of effective utiliza- important well into the 21st century. tion deserves attention. More specifically, the follow- The framework of analysis adopted in this paper ingareexamplesof deterninantsof utilizationwhich has a merit of permiting a balance between market should be accorded attention at the same time. and administrative failures and permits a very im- making adequate supporting staff available at portant but changing role of the state in bringing all levels. For example, engineers require an about industrial development in Africa. In addition, adequate number of technicians and artisans a balanced analysis of protection based on the prin- to be effective as their roles are complemen- ciples of objectivity and selectivity is permissible. tary. Appropriate ratios of different levels and This paper brings out the important role of human types of technological skills will have to be capital in industrial development. This is a signifi- made available. cant departure from the more common emphasis on * scientific and technological personnel cannot physical capital based on the Harrod-Domar type of work effectively if supporting tools and facili- models. While a shift towards empasis on human ties are not put in place. Technological institu- capital is being suggested one lesson ought to be tions and scientific laboratories, for instance, learned from the experience of the past two to three will need to be equiped with the necessary decades in Africa. The adoption of the development facilities for personnel to perform their work model which emphasized the shortage of capital and effectively. the need to mobilize investment capital also tended * motivation and reward for achievement. to take utilization of capital for granted. By the 1970s a upgrading, retraining and systematic on-the- the problem of capacity underutilization began to job training programs will need to be put in emerge and has worsened in the 1980s. It is now place. recognized that investment in new capacity has to be * foreign personnel will continue to be em- balanced with adequate attention to the detenni- ployed but attention should be paid to their nants of capacity utilization. This lesson is important utilization in training and in complementing in the proposed shift towards investment in human not substituting local personnel. capital so that utilization of human resources is ac- corded at least as much attention as investing in the As regards institutions, the role of research and creation of new stocks of human capital. This would development willbe central to the process of learning ensure that the existing stock of human resources and acquisition of technological capabilities. At least translates into capability development. On this point three aspects of research and development deserve the data which the paper presents already suggests attention. First, during the economic crisis of the that in Africa there seems to be a weak relationship 1980s many research and development activities between indicators of the stockof scientificpersonnel have been starved of funds and facilities following and technological development. For instance, the av- an attempt to reduce government expenditure. This erage share of engineers in the population of Sub-Sa- needs to be reversed to ensure that research and haran Africa is only 0.01 percent, but the correspond- development activities are adequately funded and ing figure for some of the African countries which are equipped. Second, the link between research and expected to have acquired above average technologi- development institutions and industry needs to be cal capabilities are below this average. The share for established and/or strengthened. Third, the priority 285 Commcnts on Human Capital and Entreprenewuwhip setting and policymaking government ministries adaptingimported technologyto suit thescalewhich should work closely with research and development is affordable by local entrepreneurs in terms of capi- institutions. tal and skill requirements. Third, the pattern and The framework adopted in this paper raises the structure of organization of production may or may question of the relatioship between local technology not be supportive of transition and/or graduation. and foreign technology and the role of the latter. The These factors will have to be tackled in a framework relationship has often been more competitive than which captures these interrelated factors. complementary in Africa. The role of imported tech- nology in augmenting local technological efforts in a Notes complementary way will receive high priority in the suggested framework. 1. The paper on rural-urban linkages which was re- moved from the program at the last minute contains the The Oyejide Paper historical example of the state-owned monopoly for coco- nut oil processing, which proved no match for The paper by Oyejide on Entrepreneurship and "enterprising urban women [who] descended on Nzima- Growth in Sub-Saharan Africa surveys available lit- land to participate in a restructuring of commodity erature on the determinants of entrepreneurship production and attendant social relations." By connecting with a view to identifying major elements of the rural coconut producers with urban consumers and by dynamics of private enterprise development in Af- introducingconsumergoodsintherural areas, theystimu- rica. The paper examines various aspects of enter- lated a pattern of growth with strong spillover effects into prise upgrading, graduation and growth. The issues transport, housing and marketing. The state-owned proc- raised in this paper draw attention to the barrier that essing factory had few such linkages because it tended to can be represented as a hump in the typical firm's meet all its own needs from within. Despite its supposed unit cost curve in its relation to size. This hump technical superiority, it simply could not compete in eco- inhibits transition from informal to formal activities nomic terms with informal producers. and inhibits graduation from small- to medium-scale 2. Recent studies that provide corroborating evidence operations. include Jonathan Dawson, "Small-Scale Industry Devel- An important policy issue which arises from this opment in Ghana: A Case Study of Kumasi", (London:" survey is how best to disaggregate the contents and Overseas Development Administration, ESCOR, 1988, structure of this hump. There are at least three rele- processed) and William F. Steel and Leila M. Webster, vant but interrelated factors here. First, there are "Ghana's Small Enterprise Sector: Survey of Adjustment capital requirements to facilitate jumping over the Response and Constraints", (World Bank, Industry Dev- hump. Second, there are technologyrequirements for elopment Division, draft working paper, 1990, process- the transitionand /orgraduation to take place, which ed). often call for replacement of indigenous technology 3. A state-centered approach suited some leaders' de- by imported technology. Graduation is inhibited sire to avoid building up an indigenous business class that largely because two sides of the same coin are usually might serve as an opposition power base. The notion that not given adequate attention-i.e upgrading local indigenous entrepreneurs didn't exist conveniently justi- technology or combining it with relevant aspects of fied this approach. In no case did leaders adopt strategies foreign technology in a coplementary way and to build up the entrepreneurial class. 286 28 Growth and Adjustment in Sub-Saharan Africa BennoJ. Ndulu Deceleration of growth in Sub-Saharan Africa set in process of adjustment. Longer term issues related to during the second half of the 1970s. It was sustained structural transformation, institutional develop- during the 1980s mainly as a result of exogenous ment, and the development of human capacity are shocks and policy misdirection. Key among the ex- definitely very fundamental to the development ogenous shocks were a collapse of primary process in general and in African economies' capac- commodity prices, an increase in interest rates on ity to respond (adapt) to changing conditions in par- foreign debt, and a steep cut in net foreign resource ticular. These sets of issues and the political economy inflows during the first half of the 1980s. Policy mis- of the adjustment process merit complete treatment direct,ion was of two types. In the first, which has on theirown. Killick (1990) and Helleiner (1990) have been referred to as 'capital habit' (Wheeler 1984), excellent reviews on issues concerning the relation- investment was driven beyond sustainability. For- ship between the structure of the economy and its eign resources played a significant role in easing capacity to adapt to changing conditions, and the constraints for financing capacity expansion; how- relationship between structural adjustment and long ever, utilization of this capacity relied dominantlyon term development. own resources whose rate of expansion fell short of There are definite links between medium term what was required. In the second, the anti-export objectives and long term development. Sustained policy bias of the 1 970s contributed to the decline in growth is fundamental to sustainable redistributive export growth. Structural maladaptation was also policy and poverty alleviation. Stable and healthy related to a wrong expectation that the exogenous export growth is as important for medium term shocks of the late 1970s were short term and could be growth as it is for long term growth. The same study ridden out like the first oil crisis of 1973. concluded that for low-income countries investment Most countries, even before adopting Interna- in human capital is a very significant determinant of tional Monetary Fund (IMF) programs, undertook long term growth. The last decade has shown that stabilization measures. Programs adopted with pres- economic stagnation in the medium term constrains sure from the IMF, the World Bank, and bilateral such investment to the detriment of longer term donors continued with this perspective of adjust- growth. Many of the features of growth with adjust- ment. The perspective subsequently shifted to ad- ment are thus implicitly and in some cases explicitly justment with growth. This change was premised on consistent with longer tenm growth and develop- broad acceptance of the contention that sustained ment. economic growth is central to an adjustment strategy This paper attempts to highlight the key trade-offs aimed at achieving sustainable internal and external between growth and macroeconomic balance macroeconomic balance. through discussion of resource gaps for growth, vari- This paper emphasizes a medium term perspec- ous channels for closure of such gaps, and implica- tive. It does not take up distributional issues in the tions for macroeconomic balance. In the second sec- 287 Growth and Adjustment in Sub-Saharan Africa tion, the peculiar features of the growth process in Import Compression, Capacity Utilization, and Growth Sub-Saharan Africa are presented. These require analysis beyond conventional growth models. In the A large part of the literature on growth has fo- third section, an attempt at modelling growth incor- cussed on the relationship between capacity growth porating these peculiar features is undertaken. Em- and investment. Growth of capital stock has been the pirical testing of the major conclusions for the model focus of a majority of the models that deal with is provided and adjustment experience is discussed labor-surplus economies, such as those in Sub-Saha- in light of the growth model. The fourth section ran Africa. These models have ranged from fixed- provides concluding remarks. proportions in production emphasizing incremental capital-output relationship (Chenery and Strout The Peculiar Features of the Growth Process in 1966) to perfect factor substitutability production Sub-Saharan Africa (Solow 1956, Robinson 1971, Fischer 1987). Over the last fifteen years, there has been a decel- The process of growth entails not onlythetrade-off eration of real growth in spite of reasonably high between current consumption and capacity growth rates of investment in Sub-Saharan Africa (see tables (via investment), but also between capacity growth 28-1 and 28-2). Therefore it has become necessary to and capacity utilization in an import compressed emphasize the distinction between capacity growth situation. Thus actual growth targets and estimates and actual growth. of resource gaps for growth should take into account Real growth in a situation where capacity underu- the possibility of much higher actual growth poten- tilization obtains is the sum of capacity growth gp, tial than that implied by investment projections which is driven by the investment rate, and growth alone. While net foreign resources are currently criti- in the rate of capacity utilization gu:1 cal for the recovery of export performance, policy measures and resource allocation in favor of a diver- (28-1) g 8 gp + g sified export sector remain crucial for sustained growth performance in the longer run. Table 28-1. Growth of production, investment and external trade: Sub-SaharanAfrica (average annual percentage) indicator 1965-73 1973-80 1980-87 Gross domestic product Sub-Saharan Africa 5.9 2.5 0.5 Low-income economies 6.0 2.8 -0.4 Agriculture Sub-Saharan Africa 2.2 -0.3 1.3 Low-income economies 2.2 0.0 1.0 Manufacturing Sub-Saharan Africa 10.1 8.2 0.6 Low-incomne economies 10.7 10.2 -1.0 Gross domestic investment Sub-Saharan 9.8 4.0 -8.2 Low-income economies 10.3 4.5 -9.9 Export volume Sub-Saharan Africa 15.1 0.2 -1.3 Low-income economies 16.9 -0.6 -3.6 Import volume Sub-Saharan Africa 3.7 7.6 -5.8 Low-income economies 3.4 8.1 -7.8 Implicit GDP deflator Sub-Saharan Africa 7.5 6.8 15.2 Low-income economies 8.2 17.4 17.0 Source- World Bank 1989a Sub-Saharan Africa: From Crisis to Sustainable Growth. 288 Growth and Adjustment in Sub-Saharan Afrka High investment rates could coexist with declin- capacity utilization, especially in the "presumed" ing actual growth as long as the decline in capacity dynamic sector, industry (Gulhati and Datta 1983, utilization, gu, exceeds the increase in capacity Wangwe 1983, Ndulu 1986, Rattso 1987). Import in- growth,gp. Fora given resourceconstraint, theremay stability hence translates into growth instability be a trade-off between capacity growth and capacity (Helleiner 1986). utilization. A number of African economies experi- Import compression occurs under two conditions. enced this during 1975 to 1981, as capacity underu- First, if substitution between imported and home tilization accelerated while investment rates in- goods is imperfect, a cut in capital and intermediate creased. This was reflected in the study by Gulhati goods imports will result in a cut in real activity and Datta (1983), which observed rapid increase in growth (Khan 1974, Khan and Knight 1988, Balassa ICORs and decline in investment productivity from et al 1986). Unlike in the extreme case of the fixed 83.8 percent during 1961 to 1973, to 6.2 percent dur- coefficients relationship, the reduction in real activity ing 1980 to 1987 (World Bank 1989, p.26). The process growth will be proportionately less than the rate of of growth does not only depend on factors that influ- decline in imports, since some substitution away ence capacity growth but also those affecting the rate from imports may take place. of capacity utilization. This contention is noteworthy The remarkable stability of the ratio between real given the fact that savings and hence investment imports of capital goods and real investment over the ultimately depend on actual income growth. Re- last decade suggests a close to fixed proportional source allocation between capacity expansion and relationship. For Sub-Saharan Africa as a whole, the capacity utilization becomes as important as re- ratio of real capital goods imports (machinery and source mobilization for investment to enhance ca- transport equipment) to real investment2 increased pacity growth. very slightly from 35. 4 percent in 1980 to 36. 8 The critical role of import capacity in the realiza- percent in 1987. In spite of a steep deceleration in tion of real growth can be looked at from the perspec- import volume growth, the share of capital goods in tive of these two components of actual growth. In- total imports increased slightly between 1980 and vestment requires capital goods that are not domes- 1987, from 31 percent to 32 percent (World Bank tically produced and therefore domestic savings can- 1989a). However, the economywide ratios of real not be transformed into investment goods unless intermediate imports to value added seem to have foreign exchange is available (Ram 1985). On the shown a compression during the last decade of for- other hand, it has been shown that intennediate eign exchange supply bottlenecks. While the annual imports are the single most important constraint to growth rate of import volumes decelerated from 7.6 Table 28-2. Basic indicators of resourcegapsforgrowth. Sub-Saharan Afica, selected years Indicator l965 1980 1985 1987 Percentage od GDP Gross domestic investment Sub-Saharan Africa 14.0 20.2 12.1 16.0 Excluding Nigeria 15.0 20.0 15.7 16.0 Gross domestic savings Sub-Saharan Africa 14.0 22.0 12.6 13.0 Excluding Nigeria 15.0 14.3 13.8 11.0 Resource balance Sub-Saharan Africa 1.0 1.4 0.4 -3.3 Excluding Nigeria -5.0 -5.7 -2.0 -4.9 Fiscal defecit (excluding grants) Sub-Saharan Africa - -9.6 -10.6 -11.4 Excluding Nigeria - -9.9 -10.9 -11.8 Ratio Import surplus* Sub-Saharan Africa 1.06 0.89 0.78 1.11 Excluding Nigeria 1.06 1.15 0.90 1.12 Imports/ exports. 289 Growth and Adjustment in Sub-Saharan Africa percent during 1973 to 1980 to 5. 8 percent during Whereforeignsavingsconstituteadominantpropor- 1980 to 1987, the rate of growth of production de- tionof import capacity, such rigidityhasled toasteep clined from 2.5 percent to 0. 5 percent. This suggests decline in capacity utilization, and growth has imperfect substitution. The capacity utilization rate largely been exogenous. declined proportionately less than the reduction in real intermediate imports. This was achieved Exogeneity of Growth in Sub-Saharan Africa through change in the relative sectoral contribution to GDP in favor of the less import-dependent sectors The bulk of savings, and hence investment, is de- in production, particularly agriculture. However, pendent on income growth itself. Where growth is substitution in the individual production processes constrained by foreign exchange, the level of exports, is more rigid, as evidenced by the steeper decline in which is an important part of income, plays a crucial growth observed in the more import-dependent in- role in enabling the savings-investment process. In- dustrial sector (for agriculture and manufacturing creased exogeneity of growth reduces the efficacy of see table 28-1). policy efforts and has strong implications for me- Import compression can also occur when import dium term options. There are a number of exogenous volumes in low-income countries are predominantly factors that influence growth. Key among these are constrained by import capacity rather than deter- the external terms of trade, net foreign resource in- mined by relative prices and real income as tradi- flows, weather vagaries in economies dominated by tional models postulate (Hemphill 1974, Moran rain-fed agriculture, and discovery and exhaustion 1990). In a model that takes into account import of mineral resources. The extent of exogeneityof the capacity directly, Moran (1990, p.288) estimates short growth process depends on the relative dominance run income elasticity for import demand of 0.2 and of these factors over endogenous factors affecting price elasticity of -0.1. The argument here is that growth. when foreign exchange constraints are explicitlycon- Two main sets of factors explain the performance sidered, import restrictions dampen the transmission of real exports (income terms of trade):export volume mechanism of relative prices. changes, which are policy related, and changes in the As long as imports are imperfect substitutes for barter terms of trade, which are exogenously deter- domestic goods, a squeeze on the capacity to import mined. In a study covering 33 countries in Sub-Saha- will result in a decline in investment and capacity ran Africa during 1970 to 1985, the explanatory con- utilization. Import compression thus ceteris paribus tributions of these two factors were decomposed leads to reduced capacity growth and rate of capacity (Svedberg, 1990). The changes in export volume fac- utilization. The severity of import compression ef- tor dominated in 19 countries, the barter terms of fects varies from country to country depending on trade factor dominated in 11 countries, and three the extent of rigid dependence of investment and countries had a mixture of the two. Among those production on imports. most affected by barter terms of trade were oil ex- The two main sources of import capacity are ex- porters and monocultural economies. A similar de- port earnings and foreign savings. Export perform- composition analysis for the period from 1980 to 1988 ance and the rate of net foreign resource flows are reveals the more significant role of the barter terms central to alleviation of import compression and of trade (see table 28-3). growth. However there are other measures condu- The period of economic stagnation was charac- cive to enhanced efficiency in the utilization of im- terized by increased exogeneity of the growth proc- port capacity, especially in the longer run. ess both directly (in terms of contribution to value The above suggests that the problem of import added) and indirectly via import compression. This compression in relation to growth is not merely that is particularly significant for monocultural econo- of the absolute availabilityof real import capacity,but mies where exports contribute a large proportion of also of its allocation between capacity epansion and GDP and import capacity. Typical cases are Zambia, utilization. However, in some cases selective comple- Nigeria, Gabon, Congo, Mauritius, and Ethiopia. mentary investment-particularly in infrastructure Dependence on foreign resources ties growth per- to remove bottlenecks for utilization of previously formance to the economic performance and condi- installed idle capacity-may have a higher pay off. tions of donor countries. As Rattso (1990) shows Export performance and the rate of net foreign re- conceptually and as trends in resource inflows show, source inflows are central to the alleviation of import the pattern of flows has tended to be procyclical. compression and growth. However, net foreign in- World economic recessions weaken demand for Sub- flows have often been tied to investment projects. Saharan Africa exports and reduce resource trans- 290 Growth and Adjustment in Sub-Saharan Africa Table 2-3. Sources of variation in the purchasingpower of exports: Sub-SaharanAfrica, 1980 to 1988 Growth rate of Barter terms Variation of Variation of purchasing Quantum of trade purchasingpower purchasing power power growth rate growth explained by quantum explained by terms of Period (perentage) (percentage) (percenfage) (percentage) trade (percentage) 1980-87 -5.8 -1.6 -4.2 27.6 72.4 1980-85 -5.6 -3.3 -2.3 58.9 41.1 1985-88* -7.3 0.7 -7.7 -9.0 109.0 Data Jror 1988 are projections. Sources: Data from World Bank 1989a World Deelopment Report; World Bank 1989b African Economic and Financial Data fers. This occurred during the late 1970s and the first Khan and Reinhard (1990). More important are the half of the 1980s. In addition, this period was charac- qualifications to the results offered by these two terized by increased real interest on extemal debt, authors regarding the indirect effects of public in- which further reduced net inflows to finance growth. vestment on growth via raising profitability of pri- vate investment and absorption capacity of the econ- Public Sector and Growth omy more broadly. The second issue relates to the link between fiscal The public sector's significant role in the growth deficitfinancingand the monetaryprocess and hence process in Sub-Saharan Africa stems from its partici- inflation. Where public sector savings fall short of pation in investment and overall resource allocation. financing requirements for public sector investment, In addition to providing economic and social infra- the gap will have to be financed either by noninfla- structure, governments in Africa have participated in tionary borrowing, at least in the short run (net new directly productive activities through statal and foreign resources to govemment and non-bank do- parastatal entities. Through direct controls of alloca- mestic borrowing), or through monetization. The tion of financial resources and foreign exchange, they gap between public investment and public sector have sought to control the pattern of resource alloca- savings determines the public sector financing re- tion beyond budgetary operations. quirement and the proportion of it monetized will Two issues arise with respect to the impact of high depend on the availability of the other two sources. government participation in resource allocation for Depending on how the fiscal deficit is financed, a growth. The first relates to the impact of public sector trade-off between growth and inflation will ensue. If investment programs on private investment. On the other sources are not available, an inflation tax be- expenditure side, public sector investment, espe- comes a major instrument for financing growth via cially in infrastructure, raises the profitability of pri- monetization of the fiscal deficit, but with the cost of vate investment. On the financing side, however, if price instability. Such a strategy is nonsustainable, as public sector borrowing exceeds new deposits with has been well documented in the Olivera-Tanzi ef- the banking system, the public sector investment fect, due to fiscal drags as a result of collection lags programs crowd out private investment. The net and declining revenue yields beyond a certain level impact of public sector investment programs will of inflation rate [Olivera (1967), Tanzi (1977, 1988), thus depend empirically on the relative magnitudes Aghevli and Khan (1977), Roe (1989)]. Increases in of the two effects. Evidence from several studies on net foreign resource inflows and reductions in for- this has shown a net crowding in effect of public eign debt servicing largely borne by governments sector expenditure on investment.3 The variation in have important implications for reducing monetiza- the size of the crowding in effect of public sector tion of the fiscal deficit and hence inflation. expenditure on investment is most likely explained Public sector deficits are not all related to invest- by the extent of infrastructural expenditure in total. ment programs, since large proportions have re- Blejer and Khan (1984) show for developing coun- cently been associated with current dissaving. Subsi- tries as a group that longer term infrastructural ex- dies and transfers via absorption of losses of distri- pendi tures, rather than short term public investment, bution and marketing parastatals have significantly positively induce private investment. The issueisnot widened the financing gap. Part of this is directly that of comparative direct effects on growth meas- absorbed into the fiscal budget through subventions ured by relative productivity of private and public to parastatals and part of it shows up in a build up investment as discussed and empirically tested by on non-serviced overdrafts in the official banking 291 Growth and Adjustmcnt in Sub-Saharan Africa system. The latter part of absorption has recently variation across countries. The response to shocks threatened viability of the banking system in some and closure of resource gaps are discussed in the countries (e. g. Tanzania). This is a result of a large context of the model for the 1980 to 1985 and post- mandatory accumulation of non-performing assets 1985 periods in Sub-Saharan Africa. in the banks against continued honoring of its obli- gations on the liabilities side, albeit at increasing Growth Model interest rates, as efforts towards achieving positive real rates are being implemented. Here the model is used to demonstrate the major factors which enhance and constrain real output Growth and Resource Gaps growth in Sub-Saharan Africa. A prototype growth model which incorporates the Growth and investment. In summary form, the peculiar conditions obtaining in Sub-Saharan Africa growth-investment relationship is presented in is presented in this section. The model is a variant of equations (28-2) through (286): the one developed by Taylor (1988,1989) and is based on a consistent macroeconomic accounting frame- (28-2) gp ' & + ki work. It could also be considered an extension of the World Bank's Revised Minimum Standard Model (28-3) i- ip + ig (RMSM) and is even more closely similar in spirit to the integrated model developed by Khan and Haque (28-4) ip, iiU + i2 ig (1990). Its distinctiveness from the latter two models lies in two areas. First, it explicitly incorporates ca- (28-5) gp - go + k(io + ilu) + k(i + i2) ig pacity utilization in the determination of resource gaps for capacity growth. This enables consideration (28-) ig - [(V( + i2)1 [(gp - go) k - (i + ilu)] of trade-offs between capacity growth and capacity utilization, and therefore of the relationship between Capacity output growth gp is assumed to be deter- capacity growth and actual growth for economies mined by the level of investment i and its productiv- starting with low capacity utilization rates. Second, ity k. Behind the growth-investment relationship is by including the consolidated fiscal gap and its fi- an imperfect substitution production function with nancing, it provides an explicit but approximate link capital being the most constraining factor. of growth and financing to the inflationary process Total investment is the sum of private ip and public via monetization. investment ig. Public investment is determined ex- In order to allow consideration for alleviation of ogenously by government decision and constrained the foreign exchange constraint through improved by financing sources. Private investment is deter- export performance, the version of the model pre- mined by availability of profitable investment op- sented here includes the real exchange rate as a rela- portunities and constrained by availability of funds. tive price (Solimano 1990, Ndulu 1990). In this sense We shall assume for now that public investment's it reduces the deterministic nature of the other mod- crowding in effect on private investment exceeds its els. Unlike in the models by Hemphill (1974) and crowding out effect. This seems to be plausible cur- Moran (1990), import capacity is not considered ex- rently in view of the deterioration of economic and ogenous in entirety. Export eamings, a component of social infrastructure over the last decade. The net the capacity to import, is itself influenced by policy public expenditure crowding in effect on private in- stance. vestment is represented by i2ig. The pervasive finan- The discussion first relates the model to the pecu- cial repression in Sub-Saharan Africa makes inves- liar features of the growth process in Sub-Saharan tible funds quantity rather than price rationed. The Africa. An empirical calibration of the model is done impact of the interest rate on private investment is fora single country (ranzania), to bring out the range thus assumed to be negligible. of strategic choices for closure of growth gaps and All variables are normalized by capacity output yp their implications for sustainability. Two empirical and hence 3yp is capacity utilization u. The simple tests are carried out to demonstrate the impact of acceleratoris ilu and equation (28-6) is the consistent import compression on growth and the relationship decision rule for public investment. between financing gaps and inflation. These tests are performed using cross-sectional data for Sub-Saha- Growth constraints. Three potential constraints to ran Africa, and are largely indicative in view of wide capacity growth related to investment finance are 292 Growth and Adjustment in Sub-Saharan Afric private savings, public sector savings, and overall substitutes for imported inter.mediate goods (Khn foreign savings. Overall accumulation, which is total and Knight 1988). Competitive imports mc are as- savings, is represented by s in the following: sumed to be exogenous and dependent on the quan- tity rationing behaviour of the authorities. (28-7) sp _ + b1 (1- w)u Exports e are assumed to be supply-constrained and responsive to the internal terms of trade and S8 - zg + t - foreign competitiveness, which is represented by the real exchange rate rer. It is assumed here that exports (28-8) Z8 =4+ ZIU respond positively to the real exchange rate meas- ured in domestic currency units. The capacity utili- sf=mc + mZ + + ji - e - t zation rate u is considered a scale factor of total exports. mz - a, + alu, mk - (1-c)i Equilibrium growth conditions consistent with financ- c = Domestic Content Of Investment ing constraints. The equilibrium condition consistent with u being a scale factor, with theoverall financing (saving) constraint is given as: (28-9) e - eo + eirer (28-11) i .sp + s+ Sf (28-10) s = sp + sg + Sf Substitution yields the growth rate consistent with All variables are again expressed in normalized the overall savings constraint gs: terms. Private savings sp constitute the private sector's (28-12) g c k changes in commodity stock hoarding, enterprise c [do + dju - d2rc provision for depreciation, and net additions to fi- d go nancial assets held with the banking system. Aggre- k + b0 + z,+ mc + a, +j + t - eO gate private savings depend on disposable income flows [(l-w)u]. di - bi (1-w) + zi + al > 0 Public sector saving sg in gross terms is defined here to include domestic and foreign components. d2 - el > 0 Their sum is the public sector surplus including grants. The domestic component Zg is dependent on The stability condition for the growth equation is income flows, given the tax rate and domestic recur- that the domestic component of investment c is less rent expenditure. The foreign component includes than one. Savings-constrained growth will increase net official foreign transfers t and interest on foreign with capacity utilization through increased domestic debt j*. savings sp + sg, which increase with capacity utiliza- Overall foreign savings sf is equal to the negative tion u. In economies with excess capacity, the in- of the current account deficit plus unrequited net crease in net exports will not crowd out investment transfers t. For simplicity it is assumed that only since domestic saving increases with u. The effects of government borrows abroad and is a recipient of changes inoffsettingexogenous net foreign transfers, transfers. We separate the current account balance domestic saving that rises with u, and real deprecia- into two components. The exogenous component is tion impacts via exports have been separated in interest on foreign debtj*. Thebehavioral component equation (28-12). is based on trade account determinants. Foreign sav- Based on a foreign exchange-constrained invest- ings increase with various components of imports ment equation with a fixed import content of invest- and debt servicing, given exports and exogenous net ment, the growth rate consistent with the balance of foreiign transfers. resources and uses of foreign exchange & is Imports m are assumed to be quantity rationed. Imports of capital goods mk are demanded in fixed (28-13) proportion to investment. Intermediate imports mz k are determined by a derived demand of the capacity gf (1-c) [0+ T* e,- mc-j' -a -alu + e2rer] +g0- utilization rate, since domestic inputs are imperfect 293 Growth and Adjustment in Sub-Saharan Africa The stability condition is that 1-c < 1. tax revenue and net officai foreign transfers; it de- Capacity output growth will fall as the capacity creases with increases in foreign debt servicing. utilization rate increases. This is because the increase in intermediate imports required for a given import Growth and inflation. The fiscal constraint provides capacity would imply a reduction in imports of capi- a link to the inflationary process which depends on tal goods for investment. This is the trade-off be- the monetized proportion of the borrowing require- tween capacity growth and capacity utilizafion in an ment. Using a simple monetarist approach, the im- import compressed economy. In the case where plication for inflation consistent with a given target al < 1-c, actual growth will increase. More specifi- growth rate and financing gapscan be approximated caily, starting with a situation of capacity underutili- (cost-push factors are assumed to be incorporated zation, as long as (1 - c)/alu > k, a unit of foreign into the growth rate of output on the supply side). exchange switched from capital goods imports to Demand for money is assumed to be stable and intermediate imports will raise actual growth. Since changes only with real income in the medium term. domestic savings depend on actual output, an in- In the short run, velocity changes and portfolio sub- crease in u will eventually lead to increased invest- stitution may induce instability. In economies with ment. This would further relax the overall savings undeveloped financial infrastructure, the dominance constraint. of transacfions demand makes this assumption plau- Real depreciation of the exchange rate would lead sible. to higher growth through increased exports. Assuming constant velocity in the medium term, Through increased foreign exchange earnings, real the equation of exchange is expressed in growth depreciation would allow increased importation of terms by equation (28-16). The change in money intermediates and capital goods. Increases in net of- supply has three sources, which are represented in ficial transfers would strengthen the real exchange equation (28-17). Credit expansion targets consistent rate effect; increases in foreign debt servicing would with inflation and growth targets can be set using reduce it. Under foreign exchange-constrained equafion (28-18), assuming sterilization of NFA. growth, the impact of real exchange rate depreciation on capacity growth is unambiguously positive. (28-16) M - P + X The fiscal constraint to growth is derived based on the public sector borrowing requirement psbr. The 1; = broad money supply growth psbr is determined for a given exogenously deter- = price level rate of change (inflation) mined public sector investment rate, public sector X = real output growth savings which is dependent on the capacity utiliza- tion rate, and exogenous recurrent expenditure lev- (28-17) AM - h;ruPyp + AL + ANFA els. The normalized public sectorborrowing require- ment is then given as: AL = change in new bank credit to private sector ANFA = change in net foreign assets (28-14) pu = ig-zg+j - t = ig-z1u +fj-t h = proportion of psbr monetized hruPyp = new bank credit to public sector p = psbr/y and hence pu = (psbr/y)(y/yp) uPyp = px actual output The psbr increases with public investment and for- (28-18) p , v[AL/px + ha] - (ga + gyp) eign debt servicing; it declines with increases in do- mestic public sector savings and net official foreign v = costant velocity transfers to the government. Substituting for ig from (28-6) and solving for the Given the public sector borrowing requirement, growth rate consistent with the fiscal constraint the proportion of it monetized will depend partly on yields: the magnitude of net official transfers. A higher t (28-15) gp k(1 + i2) [(n + Z1) + ill u implies a lower h. The public sector borrowing re- + k(l + i2) (zo - j* + t) + quirement also depends on the extent of foreign debt go + ki0. servicing. Hence, reduction in inflationary pressure can be achieved via increased t and reduced j*. In the Fiscal constrained growth, given a target psbr, in- longer term sustainable inflationary control will de- creases with capacity utilization through increases in pend on reducing the domestic sources of high psbr. 294 Growth and Adjustment in Sub-Sahara Africa Increased overall domestic savings not only helps Figure 28-1. Constraints to Capacity Growth reduce psbr directly via zg, but also creates an oppor- tunity for public sector financingfrom domesticnon- i F inflationazy sources. gp s Figure 28-1 depicts the three keyconstraints-sav- ings, fiscal, and foreign exchange constraints-in the G growth-capacity utilization space. It also depicts the trade-off between capacity utilization and inflation via the fiscal constraint. Ex post these three schedules must intersect at a given level of capacity utilization. 0 In calibration of the model for a given year, this is gp- - - achieved via appropriate adjustments in the inter- cepts if historically-estimated values of slope coeffi- cients are to be preserved. G F Implications for Growth and Adjustment Based on the Calibration of the Model and Simulations Using Tarzania as a Case A. variation of this model (holding the real ex- P change rate constant) has been calibrated for six Sub- Saharan African economies-Ghana, Nigeria, Uganda, Zambia, Zimbabwe, and Tanzania (Green 1989, Oyejide and Raheem 1989, Sepheri and Loxley 1989, Mkandawire 1989, Rattso and Davies 1989, and Ndulu 1989). A full version (including a variable real exchange rate) was also calibrated for Tanzania (Ndulu 1990). All six studies underscore the domi- nance of the foreign exchange constraint in the growth process. The trade-off between capacity grcwth and capacity utilization is emphasized under ity utilization can be obtained to achieve a target this constraint. Increased export performance to re- growth rate. This, however, can produce values of duce import compression and increased capacity foreign resource inflows that are not sustainable and utilization to raise domestic savings are seen as the which could be symptomatic of growth without ad- key measures for growth and stabilization. The mag- justment. nitude of foreign debt service is also underscored by A comparison of the two strategies and the esti- all six studies. mated parameters for the Tanzania economy are Behavioral paraameters are obtained using histori- given in table 28-4. The target growth rate for poten- cal data and least squares estimation. There are two tial output is set at 4 percent. The simulation entails approaches for achievinga consistent solution for the the two strategies above compared to the actual base model. First, given a target growth rate, the saving year (1986) performance and that of 1976, which and foreign exchange gaps are solved simultane- historically reflected reasonable sustainability in ously for required changes in capacity utilization and terms of growth and macroeconomic balance. the real exchange rate. The solution for capacity utili- Relative to the base year, the strategy that entails zation is then used to obtain a fiscal balance consis- growth and adjustment (via the real exchange rate) tent with the desired growth rate. This enables us to comes out superior to the one that relies on foreign obtain a consistent public sector borrowing require- resource inflows to support growth. In the growth ment. Using obtained values of the borrowing re- with adjustment strategy, the extent of self-financing quirement, potential growth rate, growth in capacity for growth is much higher. The foreign savings re- utilization and given exogenous non-inflationary quirement to achieve the target growth rate of poten- sources of finance, the extent of monetization and tial output (nt) with adjustment is 13.2 percent com- implications for inflation can then be drawn. pared to 17. 5 percent with the other strategy. Since Alternatively, for a given real exchange rate, the increased export performance entails greater capac- solution for required transfers and increased capac- ity utilization (85 percent compared to 76 percent 295 Growth and Adjustment in Sub-Saharan Africa with the other strategy), domestic savingsare signifi- yield of the adjustment process reflects a shift in cantly higher (13. 3 percent compared to 10. 7 per- relative prices. The foreign debt servicing ratio in- cent). The external current account deficit also im- creased from 0. 8 percent in 1976 to 2. 1 percent of proves slightly from 6. 0 percent of potential output potential output in 1986. Higher levels are projected without exchange rate adjustment to 5. 4 percent if debt relief is not forthcoming. For the same level of with adjustment. Actual output growth is signifi- exports and imports, a higher current account deficit cantly higher with the adjustment scenario due to is implied. Since external debt is largely serviced by much higher growth of capacity utilization. In view the government, its impact on the fiscal deficit is of the lower public sector borrowing requirement in more pronounced. The domestic savings rate is sig- the growth with adjustment scenario, pressure for nificantlylowerin the 1986 simulated scenarios than monetization and inflation decline. In brief, the strat- in 1976. This is a result of a lower savings effort, and egy incorporati'-g jarsment demonstrates much particularly the public sectoi's ,educed tax effort. more sustainability. Deterioration in supportive infrastructure for pro- Compared with actual performance in 1976 the duction and export performance are reflected in re- above scenarios indicate a deterioration in the sus- duced public sector spending. This and deterioration tainability of growth. This significant drop in the in the terns of trade have reduced the yield of rela- Table 28-4. Comparative simulationfor Tanzania withfourpercent targetgrowth rate a) Parameters b) Growth equations with intercepts (estimated using OLS) adjustedfor base year (1986) values g = -0.16 + 0.25i (i) gs = -0.02 + 0.07u (without rer) i = 0.053 - 9.89u - 1.148ig g = -0.064 + 0.214u - 0.082 rer sp = -0.04 + 0.20u (ii) gt = 0.165 - 0.193u (without rer) Zg = -0.068 + 0.083u gf = 0.077- 0.234u + 0.166 rer Sf = 0.068 + 0.31u - 0.22 rer + 0.33i (iii) gp = -0.007 + 0.053u mk = 0.33i mz = -0.156 + 0.31u e = -0.234 + 0.22 rer c) Comparative simulations for 4% growth target (all variables normalized by yp) Base year Projection with change Projection with change Actual 1976 Variable (1986) in u and t in u and rer (good year) g 0.030 0.040 0.040 0.033 i 0.178 0.220 0.220 0.193 u 0.700 0.760 0.846 0.907 rer(1972=1) 0.703 0.703 0.970 t 0.068 0.096 0.068 0.014 0.026 0.021 0.012 0.057 u 0.018 0.016 0.010 0.052 Sp 0.100 0.112 0.130 0.150 sS -0.010 -0.005 0.003 0.052 +t 0.126 0.155 0.122 0.034 +nt 0.146 0.175 0.132 0.045 is 0.055 0.086 0.080 0.110 ip 0.123 0.134 0.140 0.083 e 0.061 0.061 0.120 0.200 mz 0.038 0.054 0.083 0.098 0.021 0.021 0.021 0.008 t = official transfers; nt = net transfers (including private) 296 Growth andAdjusment in Sub-SaharanAfrica tive price measures. The implications of the above Sudan, and Ethiopia were excluded due to large changes are that, although a strategy of growth and fluctuations in import volumes related to food im- adjustment issinequanonforsustainability, itentails ports.6 The estimated equation for the impact of im- a long haul. External resource inflows are an impor- port compression on growth is the following tant component of the adjustment process in the medium term. However, for these to be effective in (28-20) g - -2.378 + 0.309 + 0.210gm + 0.104gtot promoting sustainability, continued adjustment is (-1.803)(4. 872) (2. 918) (0. 397) critical. R2 =0.647 F=13.85 DW=1.99 Empirical Testing of the Major Conclusions from the t - statistics are in brackets Mode! Using Sub-Saharan Africa Data These results show that during 1980 to 1985, capac- Two simple empirical tests are carried out using ity continued to be an important source of growth cross-sectional data for Sub-Saharan Africa for the across the countries included. Import compression period 1980 to 1985. The first demonstrates the im- impacted on real growth via capacity utilization. pact of import compression on growth. The second Since gm was negative in most of the countries, de- demonstrates the relationship between financing cline in capacity utilization rate must have partially gaps and inflation. offset capacity growth gp, generated via continued high investment rate. The tenns of trade had an The impact of import compression on growth. Follow- insignificant impact on growth apart from import ing the above model, real actual growth is postulated compression. to depend on the rate of capacity creation and capac- The instrumental variable method was also used ity utilization. The domestic investment rate determ- to estimate equation (28-20) to further check for pos- ines the rate of capacity creation and hence capacity sible interdependence between i and g. The results of growth. Domestic investment makes the first claim the two-stage least squares method support our pos- to available import capacity. This contention is plau- tulation of exogeneity. sible at least for the period up to 1985, when the use of foreign savings was tied to investment projects. The relationship between financing gaps and inflation. The residual of the import capacity is allocated for The relationship between the inflation rate and current use, mainly for imported intermediates. The growth of domestic credit and growth of real output proportion of import capacity to support capacity is estimated in equation (28-21). Again, the analysis utilization is thus more sensitive to the severity of uses the average data for the 1980 to 1985 period. This import compression measured here as the growth hypothesis is tested to provide some insight on the rate of import volume. A third independent variable, impact of monetization of the fiscal deficit on infla- growth in the terms of trade, is included to capture tion in an attempt to close resource gaps for growth. the effect of exogenous shocks. To the extent that this The period 1980 to 1985 involved a large expansion is already captured via import compression, this of the fiscal deficit as a proportion of GDP which was variable may not have a separate significant statisti- monetized as evidenced by high domestic credit cal impact on growth. Possible multi-collinearity be- growth (table 28-5). This occurred in spite of decel- tween these two variables is dampened by the strong eration in actual real growth. presence of exogenous components in the total im- This test was done cross-sectionally for 21 coun- port capacity. tries for which complete data was available.7 The grwth rate of the consumer price index for 1980 to (28-19) g = go + gi + gZgm + g3gtt 1985 was used as a measure of the inflation rate. Inflationary pressure was postulated to increase with i= investment rate the rate of expansion of domestic credit and to de- gm = growth rate of import volume crease with increases in real growth. The following gtot= growth rate of terms of trade results were obtained. This equation was estimated cross-sectionally for (28-21) gcpi = 0.157 + 0.942 g&c, - 1.007g 20 countries4 using average values for 1980 to 1985, (0.048)(7.748) (-1.980) which is a period long enough to accommodate lags and cyclical changes.5 Although complete data was R2 = 0.76 F = 32.72 DW = 1.823 available for 28 countries, the Sahelian economies, gPr = growth rate of domestic credit 297 Growth and Adjustment in Sub-Saharan Africa These results support the simple version of the the economy will settle at a new equilibrium with inflationary process linked to growth and its financ- lower capacity utilization, lower capacity growth, inggaps as discussed in the model. Dependingon the and high inflation. Due to the diversity of the coun- extent of monetization of borrowing requirements tries involved, thediscussionbelowisanapproxima- for growth, resultant growth in domestic credit will tion of the dominant features of the responses in the be partly transmitted into higher inflation. The coun- region. teracting effect of increased real growth is inversely related to inflation (see equation 28-18)). These two The 1980 to 1985 period. During 1980 to 1985, resp- factors explain the largest proportion of variation in onses to resource gaps for growth were predomina- inflation rates across the countries. The near-unity of ntly passive and recessionary. Pressure from these the sizes of the coefficients strongly support the sim- gaps was accommodated through import compress- ple specification with its corresponding assump- ion effected by quantitative restrictions and hence, tions. deceleration in real growth. Lower export growth However, the results obtained above are only in- (table 28-3) and reduced net inflows of new foreign dicative of broad explanatory relationships across resources resulted in reductions in real imports. Ca- the continent. These may not be directly transferable pacity utilization fell drastically as imports of inter- to country specific analysis using time series data mediates were reduced relatively more than capital (see London, 1989). Inflation expectations, for exam- goods. ple, are deemed important and can best be included Gross domestic investment rates were maintain- using fime series data. Diversity of cost structures, ed at approximately 20 percent of GDP for Sub-Sa- structural bottlenecks, and pricing systems across haran Africa as a whole during this period. The gr- countries and their variations over time are impor- oss domestic savings rate declined sharply from 21. tant considerations for analysis of inflation (Chhib- 6 percent in 1980 to 11. 5 percent in 1982 (The World beret al, 1989). Bank, 1989b). Although the continued high rate of investment partially cushioned the deceleration in Response to Shocks and Closure of Resource Gaps in the the rate of growth, the decline in capacity utilization Context of the Model: Sub-Saharan Africa rate more than offset capacity growth. Average an- nual actual growth during 1980 to 1983 was -0. 7 The closure of resource gaps for growth can be percent. During this period, policy actions to rever- analyzed in the context of the model. We focus here se the decline in export growth were not imple- on the case of a foreign exchange-constrained sce- mented and decline in the terms of trade reinforced nario. As a result of involuntary import compression, the trend. Table 28-5. Indicatorsfor closure of resource gaps: Sub-Saharan Africa excluding Nigeria, 1980-1987 Percentage of GDP Current Net official account transfers to the deficit less Real government as Domestic Current interest Gross exchange a percentage of credit account payments on Fiscal Real domestic rate indexZ the overall growth Period deficit foreign debt deficit imports investment (1978=100) fiscal deficit3 (percentage) 1980 9.5 7.4 -9.9 26.8 20.0 105.0 28.9 21.1 1981 11.6 8.9 -10.8 24.7 20.0 112.0 26.8 21.9 1982 11.5 8.8 -11.6 23.3 19.2 114.0 23.5 20.5 1983 9.1 6.5 -10.4 20.9 16.5 118.0 31.1 17.0 i984 6.8 3.8 -10.2 19.8 16.0 132.0 33.3 12.0 1985 6.6 3.7 -10.9 19.6 15.7 124.0 31.0 16.2 1986 9.0 6.2 -11.4 19.8 15.9 90.0 32.3 15.5 1987 - - -11.4 19.3 15.9 70.0 33.9 _ Note:: I Includes unrequited net transfers. 2 Foreign currency/domestic curency. 3 Overall fiscal deficit excluding grants. Sources: IBRD 1989b African Economic and Financial Data; IBRD 1989c World Devlopment Report 298 Growth and Adjustment in Sub-Saharan Africa As public revenue declined during 1980 to 1982 in drive to increase exports through real exchange rate the face of sustained high expenditure, the fiscal def- depreciation and increase in the share of foreign icit widened. In a number of cases pressure from fis- prices received by export producers. The annual cal populism resulted in increased monetization of growth rate for export volume increased to 0. 7 per- the fiscal deficit. Domestic creditgrowthaveraged 22 cent during 1986 to 1988. Of the 33 countries for percent annually during 1980 to 1982. After 1982 inf- which data is available, 21 registered significant ex- lows of new foreign resources declined. The domes- port quantum increases during 1987 and 1988. This tic investment rate declined from 18.2percentinl982 is in large contrast to the average annual quantum to 12. 1 percent in 1985 (World Bank, 1989b). Domes- growth rate of -3. 3 percent during 1980 to 1985 and tic credit growth decreased to an average of 15 per- 0.2 percent during 1973 to 1980. The real purchasing centas cuts in investment were effected and the fiscal power of exports during 1986 to 1988 registered an deficit declined from 11. 5 percent of GDP in 1982 to annual decline of 6. 3 percent due to a steep deterio- 10 percent in 1984 (see table 28-5 for magnitudes ration of the barter terms of trade. This exogenous excluding Nigeria). Correspondingly, the inflation shock of the steep fall in world prices for primary rate (GDP deflator) increased to an annual average commoditiessignificantlyreduced theyieldof policy of 18.5 percent during 1980 to 1985 compared to 6. 8 effort. percent during 1973 to 1980. In addition to increased Second, as a result of compliance to conditionality, monetization, inflationary pressure was exacerbated a resurgence of net new foreign resources was re- by negative growth in real output. corded (see table 28-5). The bulk of this was in net The resultant appreciation of the exchange rate official transfers, although net private transfers also (see table 28-5) had two additional impacts apart increased substantially. However, during the samne from discouraging exports. First, the real domestic period, foreign debt service payment reduced the value of foreign savings available in foreign currency amount of foreign savings available to support cur- units declined. Second, the increased relative attrac- rent expenditure (table 28-5). The foreign savings tiveness of holding foreign assets induced capital leakage via debt service payments in 1986 was 31. 1 flight (World Bank, 1989b). In some countries, the percent compared to 23. 3 percent in 1981. decreased tax effort during this period was a result Import compression was quite diverse across of the relative growth of the informal sector. countries depending on when adjustment programs The 1980 to 1985 period can thus be characterized were adopted and the extent of efforts to increase as one of passive accommodation to resource gaps exports. There seems to have been a shift in the for growth. The attempt to defend growth via in- allocation of import capacity between capacity ex- creased monetization of govemment deficit, external pansion and capacity utilization. There was more borrowing, and, in many cases, build up of payment concem for raising capacity utilization, often pre- arrears, could not be sustained beyond 1983. Import sented as 'consolidation of previous investment'. compression took its course and adjustment in the Foreign resources have become more fungible in savings-investment process followed suit with the terms of use since most of it now comes in the form end result being reduced real growth. The absence of of import support. active policy measures, especially on the side of rela- Investment for Sub-Saharan Africa as a whole in- tive prices, was particularly conspicuous. Continued creased modestly from 11. 5 percent of GDP in 1984 decline in per capita incomes and living standards to 16. 4 percent in 1987 (excluding Nigeria which was politically unsustainable. Decline in real public almost remained stagnant). This increase was pre- expenditures led to intensified deterioration of eco- dominantly supported by an increase in foreign sav- nomic and social infrastructure, which threatened to ings. The reversal in the downward trend in domestic reverse earlier achievements. savings has not matched the increase in investment. The post-1985 period. In the post-1985 period a The overall fiscal deficit (excluding grants) in- much more active policy stance was adopted, albeit creased from 10. 6 percent of GDP in 1985 to 11. 8 with some pressure from international financial in- percent in 1987. The gap between increased public stitutions and bilateral donors. Attempts were made spending and stagnant public saving was filled pre- to reduce import compression, raise domestic sav- dominantly via an increase in net official transfers. ings and investment, achieve a sustainable balance Government overall deficit, including grants, de- of payments, and reduce inflation. creased slightly from -7. 3 percent in 1985 to -7. 2 Attempts to tackle the import compression prob- percent in 1987 in spite of an increase in spending lems were two-pronged. First, there was an earnest which exceeded the increase in domestic revenue. Of 299 Growth and Adjustment tin Sub-SaharanAfrica the 40 countries for which data is available, 24 re- rate (relative to 1980 to 1985) owes its realization in duced their rates of domestic credit growth during many countries partly to a decrease in monetization 1986 to 1987 primarily by substituting net foreign of fiscal deficits as a result of foreign financing. transfers to the government for borrowing from the The apparent fragility of the recent adjustment banking system. A combination of reduced moneti- with growth is thus a major concem in the context of zation and an increase in the real growth rate (aver- sustainability. Similar growth rates in the 1960s and aging 1. 4 percent during 1986 to 1987 compared to first half of the 1970s had a much more sustainable -0. 3 percent during 1980 to 1985 for Sub-Saharan financing structure. While noting the current hostile Africa as a whole) led to a reduction in the rate of growth environment due to exogenous constraints, inflation (CPI) in 26 of 32 countries for which data is efforts to encourage domestic savings and a diversi- available. fied export performance have to be sustained. In- creased net inflows of foreign resources should be Concluding Remarks seen as a means for buying time to realize results from the adjustment measures. The process of growth in Sub-Saharan Africa en- tails peculiar features that have to be taken into ac- Notes count for revival of sustained growth in the future. There is need to emphasize not only the trade-off 1. Actualoutputyisobtainedbymultiplyingcapacity between current consumption and capacity expan- utilization u, with capacity output, yp: sion, but also between capacty growth and capacity y = uyp; dy - u dyp + ypdu . utilization in investment decisions. This is particu- Substitution and simplification yields: larly important given the current low rates of capac- dy/y gy dy6j + du/u - gp + gu . ity utilization and continuing import compression. 2. Importsofmachineryandtransportequipmentwere Rapid growth of capacity utilization will require deflatedbytheimportunitpriceindex(WorldBankl989b). credit expansion beyond that consistent with pro- Investment was deflated by a weighted average of the jected growth. At the same time, investment to reha- import price index and the GDP deflator with assumed bilitate supportive infrastructure is critical. weights of 30 percent and 70 percent, respectively. Based The continued need for additional net foreign re- on the six studies covering Nigeria, Tanzania, Zimbabwe, source inflows stems from two considerations. First, uganda, Zambia, and Ghana (op cit) the direct import rapidly deteriorating terms of trade have signifi- content of investment averaged around 35 percent during cantly reduced the yield from adjustment measures. 1985-86. Second, the debilitating foreign debt service burden 3. See Ndulu 1990 for Tanzania, Rattso and Davies 1990 constitutes a very large leakage from current re- for Zimbabwe, Oyejide and Raheem 1990 for Nigeria, sources to support growth. Green 1990 for Ghana, Sepheri and Loxley 1990 for Central to sustained and stable growth in the Uganda, Mkandawire 1990 for Zambia, and Boye 1990 for longer run, however, is the need to raise the extent of Senegal. own-financing for growth. This entails not only 4. CountriesinthesampleareBurundi,CentralAfrican measures for raising savings, but particularly more Republic, Ghana, Kenya, Liberia, Madagascar, Senegal, measures for improved and diversified export per- Sierra Leone, Tanzania, Togo, Zaire, Zambia, Botswana, formance. While reduction in import dependence Cote d'lvoire, Mauritius, Zimbabwe, Cameroon, Congo, can be achieved through structural transformation, Gabon, and Nigeria. this indeed will require a very long time frame. There 5. Lavy and Pedroni (1990) have an insightful discus- is still scope for raising efficiency in the use of the sion on the various ways to deal with this problem, scarce import capacity. This requires notonlyrealistic including the use of an explicitly dynamic framework to pricing of this scarce resource, but also fungibility of deal with cycles. use especially for external resources. 6. Data is from World Bank 1989b. More often than not, the revival of growth during 7. Countries in the sample are Burundi, Ethiopia, Gam- the last five years has occurred without stabilization. bia, Ghana, Kenya, Madagascar, Malawi, Niger, Senegal, The apparent reductions in the extemal account and Sierra Leone, Sudan, Togo, Zaire, Zambia, Cote d'lvoire, fiscal deficits are predicated on an increase in net Mauritius, Zimbabwe, Cameroon, Congo, Gabon, and Ni- foreign transfers for balance of payments and fiscal geria. support. Moreover, the current reduction in inflation 300 Growth andAdjustment in Sub-Saharan Africa References Khan, M. and C. M. Reinhard (1990): Pnvate Investment and Economic Growth in Developing Countries", World Aghevli, B. and M. Khan (1977): "Inflationary Finance and Development, vol. 18, no. 1. the Dynamics of Inflation: Indonesia 1951-72", American Khan, M. and N. Ul Haque (1990):"Adjustment with Economic Review, June 1977. Growth:Relating the Analytical Approaches of the IMF Balassa, B., E. Voloudakis, P. Fylaktos and S. T. Suh (1986): and the World Bank," Journal of Development Economics. "Export Incentives and Economic Growth inDeveloping Killick, T. (1990): "Structure, Development and Adapta- Countries: An Econometric Investigation", World Bank tion", African Economic Research Consortium Special Paper, DRD Discussion Paper no. DRD 159, Washington, D.C. no. 2, Nairobi. Blejer, M. I. and M. S. Khan (1984): "Government Policy and Lavy, V. and P. Pedroni (1990): "Are Returns to Investment Private Investment in Developing Countries', Intemna- Really Lower in Sub-Saharan Africa?", Mimeo, World tional Monetary Fund Staff Papers, vol. 31., no 2. Bank, Washington, D.C. Chhibber, A., J. Cottani, R. Firuzabadi and M. Walton London, A. (1989): "Money, Inflation and Adjustment Pol- (1989): "Inflation, Price Controls and Fiscal Adjustment icy in Africa: Some Further Evidence", African Develop- in Zimbabwe", IBRD WorkingPapers, WPS 192, Washing- ment Review, vol 1., no. 1. tor, D.C. Lipumba, N., B. Ndulu, S. Horton and A. Plourde (1988): Chenery H. and A. Strout (1966): "Foreign Assistance and "A Supply Constrained Macroeconometric Model of Economic Development", American Economic Review, Tanzania", Economic Modelling, vol.5, no. 4. VoL 56, No. 4. Mkandawire, T. (1990): "Growth Exercises on Zambia", a Fischer, S. (1987): "Economic Growth and Economic Pol- paper for WIDER Project on Medium Term Adjustment icy" inV. Corbo, M. Goldsteinand M. Khan(eds.) Growth Strategy, Helsinki. Oriented Adjustment Programs (Washington, D. C : IMF Moran, C (1990): "Imports undera Foreign Exchange Con- and World Bank). straint", The World Bank Economic Review, vol. 3, no. 2. Green, R. H. (1989): "Medium Term Constraints to Growth: Ndulu, B. (1986): "Investment, Output Growth and Capac- Ghana", a paper for WIDER Project on Medium Con- ity Utilization in an African Economy: The Case of straints to Growth, Helsinki. Manufacturing Sector in Tanzania", East African Eco- Gulhati, R. and G. Datta (1983): "Capital Accumulation in nomic Review, vol. 2. Eastern and Southern Africa: A Decade of Setbacks", Ndulu, B. (1990): "Macroeconomic Constraints to Growth: World Bank Staff Working Papers, No. 562, Washington, An Empirical Model for Tanzaniae, a paper for the D.C. WIDER Project on Medium Term Adjustment Strategy, Hel- Helleiner, G. (1986): "Outward Orientation, Import Stabil- sinki. ity and African Economic Growth: An Empirical Inves- Ndulu, B. and N. Lipumba (1988): "International Trade and tigation" in Sanjaya Lall and F. Stewart (eds) Theory and Economic Development in Tanzania" in R. Kanbur, J. F. Reality in Development (MacMillan). Ansah and P. Svedberg (eds. ) International Trade and Helleiner, G. K. (1990): "Structural Adjustment and Long- Economic Development in Sub-Saharan Africa (forthcom- term Development in Sub-Saharan Africa". Paper pre- ing). sented to the workshop on Alternative Development Olivera, J. (1%7): "Money, Prices and Fiscal Lags: A Note Strategies in Africa. Queen Elizabeth House, Oxford. on the Dynamics of Inflation", Banca Nazionale del Lavoro December 11-13. Quarterly Review. Hemphill, W. (1974): "The Effects of Foreign Exchange Otani, I. andD. Villanueva (1989): "Determinants of Long- Receipts on Imports of Less Developed Countries", lMF Term Growth Performance in Developing Countries", Staff Papers, 21:637-77. IMF Staff Working Paper (WP/ 88/ 97). World Bank (1989a): Sub-Saharan Africa: From Crisis to Sus- Oyejide, A. and M. Raheem (1990): "Macroeconomic Con- tainable Growth, (Washington, D. C.). straints and Growth Programming:Empirical Evidence World Bank (1989b): African Economic and Financial Data from Nigeria", a paper for WIDER Project on Medium (Washington, D. C.). Term Adjustment Strategy, Helsinki. World Bank (1989c): World Development Report. Ram, R. (1985): "Exports and Economic Growth: Some Khan, M. (1974): "Import and Export Demand inDevelop- Additional Evidence", Economic Development and Cul- ing Countries", lMF Staff Papers, 678-693. tural Change, vol. 33, no. 1. Khan, M. and M. D. Knight (1988): "Import Compression Rattso, J. (1988): "Import Compression Macro-dynamics: and Export Performance in Developing Countries", The Macroeconomic Analysis for Sub-Saharan Africa", (Uni- Review of Economics and Statistics, vol. LXX, No. 2. versity of Trondheim). 301 Growth and Adjustment in Sub-Saharan Africa Rattso, J. (1989): "The Asymmetric Relation Between Sub- Svedberg, P. (1990): "The Export Performance of Sub-Saha- Saharan Africa and the Rest of the World: A Theoretical ran Africa", Economic Development and Cultural Change, Analysis of the Role of Import Compression", (Univer- vol. 38. sity of Trondheim). Tanzi, V. (1977): "Inflation, Lags in Collection and the Real Rattso, J. and R. Davies (1990): "Growth Programming for Value of Revenues', IMF Staff Paper, vol. 24, 154-67. Zimbabwe", a paper for WIDER Project on Medium Termn Tanzi, V. (1988): "The Impact of Macroeconomic Policies on Adjustment Strategy, Helsinki. the Level of Taxation (and on the Fiscal Balance) in Robinson, S. (1971): "Sources of Growth in Less Developed Developing Countries", IMF Working Paper WP/88/95. Countries: A Cross Section Study", Quarterly Journal of Taylor, L. (1988): 'Medium Term Development Strategy". Economics, vol. 85, no. 3. Mimeo, Massachusetts Institte of Technology, Boston.. Roe, A. (1990): "Internal Debt Management in Africa", Taylor, L (1989): "Gap Disequilibria: Inflation, Investment, African Economic Research Consortium Special Papers Saving and Foreign Exchange", Mimeo, Massachusetts no. 4. 77, Nairobi. Institute of Technologr, Boston.. Sepheri, A. and J. Loxley (1990): "Medium Term Con- Tutu, K. A., N. K Sowa and C. D. Jebuni (1989): "Real straints to Growth: Uganda", a paper for WIDER Project Exchange Rate and Macroeconomic Performance in on Medium Term Adjustment Strategy, Helsinki. Ghana", paper presented at African Economic Research Solimano, A. (1990): "Macroeconomic Constraints for Me- Consortium Research Workshop (December, Harare). dium Term Growth: A Model for Chile", World Bank Wangwe, S. M. (1983): "Industrialization and Resource Al- Working Papers, WPS 400, Washington, D.C. location inaDeveloping Country: The Case of Recent Ex- Solow, R. (1956): "A Contribution to the Theory of Eco- periences in Tanzania", World Development, vol. 11, no. 6. nomic Growth", Quarterly Journal of Economics, vol. 70, Wheeler, D. (1984): "Sources of Stagnation in Sub-Saharan no. 1. Africae, World Development, vol. 12, no. 1. 302 29 The Liberalization of Price Controls: Theory and an Application to Tanzania Paul Collier and Jan Willem Gunning A subsidy raises the price to producers at the same how the economy behaves under shortages and the time as it lowers it to consumers. Thus it stimulates consequent difficulties liable to be encountered in the extra production and extra consumption and main- transition. The shortage economy turns out to have tains market clearing. By contrast, price control alarming properties. It implodes into subsistence. works by lowering the price to both producers and Response to producer price changes is perverse. The consumers. It therefore depresses production at the demand for money first rises and then falls. The same time as it induces extra consumption. Hence, if longer transition is delayed, the larger the price price controls succeed in altering prices, they give changes will need to be to achieve market clearing. rise to excess demand. Although desired consump- In the third section we show how this theory applies tion rises, achieved consumption falls. Many African to the collapse of the Tanzanian economy between governments have been tempted into using price 1978 and 1984 and, more especially, to its rapid recov- controls because they are free, whereas subsidies are ery between 1984 and 1989. The fourth section pro- a claim on thebudget. Usually price controls are only vides a brief conclusion. applied to a few items. The frustrated excess demand for the price controlled items can then spill over into The Theory of Rationing and the Transition to (satisfied) demand for other goods. However, some- Market Clearing times governments have applied the controls over such a wide range of goods that the excess demand Supply Response Under Rationing becomes generalized. An example of this occurred in Tanzania from the In order to understand the transition to market late 1970s until 1984. Consumer prices for nearly all clearing, it is necessary to investigate how the econ- goods were held at a level well below that at which omy behaves under conditions of shortage. markets would have cleared. There was intense ex- cess demand at these official prices. For the goods The volume ofcrop sales. Underconditions of market that people actually wanted a black market devel- clearing, the household is characterized as maximiz- oped. In 1984 the price controls were abandoned. ing utility subject to relative prices, endowments, Helped by extra imports of consumer goods, the and technology. If the price of output (crops) changes economy adjusted over the next two years back to relative to the price of consumer goods, the optimal market clearing. volume of crop sales will change. Although the direc- This paper focuses on the transition back to market tion of change is a priori ambiguous, empirical stud- clearing. In the second section we set out a theory of ies generally find supply response to be positively 303 The Lsberalization of Price Controls: Theory and an Application to Tanzania Figure 29-1. Influences on the Volume of Crop Sales money and abstract from any asset demand. If the household's transactions demand for money in- (6) creases, it must sell more crops to augment its cash Inflation Demand for balances. This is shown as relationship (4). In condi- money balances tions of market clearing, the usual reason fora house- hold to increase its cash balances is inflation. This is (5) (4) shown as relationship (5). In conditions of shortage, the demand for money may be radically altered depending upon the nature Availability of (2) Volume of of rationing. In most African contexts the more rele- consumer goods crop sales vant form is stochastic rationing. Here the household faces only a probability distribution of possible avail- ( t3) /t abilities of consumer goods which it might encoun- ter. Within limits the household can, depending on ^ / (1) the money balances which it chooses to hold, alter Real producer the expected value of its expenditure. Generally, as prices the household increases its moneybalances expected consumption increases and is an incentive for the household to increase its money balances. This be- havior is a form of rent seeking. Under market clearing conditions, peak money related to crop prices. In figure 29-1, which sets out balances equal planned expenditure; under stochas- a taxonomy of influences on crop sales, this is shown tic rationing, peak money balances exceed expected as relationship (1). consumption for all nonzero values. The offsetting If the household faces generalized shortages of effect reduces the demand for money and the net consumer goods, expenditure may not be a choice effect is a priori ambiguous. If goods are sufficiently variable. This depends on how goods are rationed scarce, then thedemand formoneywillbelowerthan and how the black market operates. For the present under market clearing. However, for moderate de- we will abstract from the black market. If rationing grees of shortage it is possible for money demand to is deterministic, then the household cannot signifi- be higher, even though expected expenditure is cantly influence the amount of goods it is permitted lower. Monte Carlo experiments reported in Bevan et to purchase, which, by assumption, is less than it al (1990) suggest that this is indeed likely to be the would wish to buy. Knowing its entitlement, the case. household therefore equates its income to this con- The change in the demand for money consequent strained level of expenditure. The household's sale upon the regime switch from market clearing to sto- of crops will be proportional to the availability of chastic rationing is referred to as the 'honeymoon consumer goods. In figure 29-1 this is shown as effect' because of its implications for crop sales relationship (2). Thus, in an economy characterised (Bevan et al 1987 and 1987a). If there is a once and for by shortages of consumer goods, a sure way of in- all increase in the demand for money, then peasants creasing crop sales is by increasing the availability of will choose to sell crops in order to build up their consumer goods. money balances. This is shown as relationship (6) in A surprising consequence of shortages of con- figure 29-1. Superficially this effect is like the infla- sumer goods is that, for a given entitlement, peasant tion tax; however, there is an important difference. supplyresponsetocrop pricesbecomes negative unit Inflation can be regarded as taxation because the elastic. To maintain constant income, crop sales must govemment does not increase its liabilities in real be reduced proportionately to the increase in crop terms. Understochasticrationingwithpricecontrols, prices. Thus by increasing crop prices under condi- households increase their real and nominal money tions of shortages of consumer goods, a government balances. This constitutes a liability for the govern- will inadvertently intensify the problem. This change ment, which turns out to be central in the transition in the supply response to price is shown as relation- back to market clearing. ship (3) in figure 29-1. Assuming that the predominant reason for African The market for goods and crops. So far we have fo- peasants to hold moneybalances is to finance expen- cused only on the peasant household. To complete diture, we consider only the transactions demand for the framework for the analysis of the transition back 304 The Liberalization of Price Controls: Theony and an Application to Tanzania to market clearing, we must specify the other agents Figure 29-2a. National Equilibrium Loci in the Goods in the economy. We assume that households trade and Crops Markets only with a government trading agency, which buys crops from households in exchange for domestic Nominal price of crops currency. It sells these crops abroad for consumer p G imports which it sells to households for domestic currency. Consumer goods are not produced domes- ES G tically. We assume that the trading agency is a price G ES C C taker on world markets and that it is unable to bor- EDC row either abroad or domestically. The government sets the prices of goods and crops. It requires the tradingagency tobuycropswhenever it has the cash 'c \ESG to do so and to sell goods whenever it has the foreign \ ESG exchange to import them. ES C The markets for goods and crops are depicted in G figure 29-2, which is a variant of the Malinvaud G model (1977). The space is defined on the two nomi- Nommal price of goods Pg nal prices, Pc, for the cash crop, and P& for goods. The schedules CC and GG denote the loci of notional equilibrium in these two markets for a single period. ES: Excess supply Notional equilibria are defined on the assumption of equilibrium in all markets other than that of the lo- cus being derived. Notional equilibrium in the goods market is achieved by variations in demand, supp- Figure 29-2b. Effective Equilibrium Loci in the Goods ly being treated as exogenous within the period. and Crops Markets Hence, an increase in the goods price leads to excess supply of goods, unless it is offset by the higher Pc G incomes generated by an increase in the price of cash crops. Proportionate increases in both prices are not neutral because of a real cash balance effect, which causes the GG locus to be steeper than a ray through \\\ \ the origin. Notional equilibrium in the cash crops \\ F(C) market is defined by the trading agency holding just \"\\\\\\\ enough domestic currency to buy the cash crops supplied by households. Above the CC locus the c \ trading agency holds insufficient currency to buy all & " '\\\\ of the offered crops. Below the locus there is quasi \ excess demand in the sense that the trading agency = is willing to purchase more supplies than are forth- \""""\\\\ F(G) coming. An increase in the price of crops leads to excess Pg supply unless it is offset by an increase in the price of goods. Again proportionate increases in prices are not neutral because as prices rise the real cash bal- ance effect induces an increase incrop supply. Hence, demand for crops, shifting the crops equilibrium the CC locus is flatter than a ray through the origin. locus to C'C' in figure 29-2. The two loci divide the space into four zones of This analysis can be used to distinguish different disequilibrium, inwhichnotallplanscanbefulfilled, ways in which crop and goods prices may be set so that agents change their behavior. These changes wrongly. For example, in an economy characterized in behavior alter the loci of market equilibria from by repressed inflation, the combination of excess notional to effective. The change implied by our demand for goods and crops is shown in the shaded previous theory is that, in response to an excess zone in figure 29-3. Policy advice would generally demand for goods, peasants revise downwards their infer that the excess demand for crops implies that marketing of crops. This expands the zone of excess their price is set too low. Prices Pc* and Pg* secure full 305 The Liberalization of Price Controls: Theory and an Application to Tanzania market clearing. From figure 29-3 it is evident that Figure 29-3. Regions of Crop Mispricing repressed inflation does not imply that crop prices are below P,*. Three regions of repressed inflation can be distinguished. In the region bordered by c c G C',E,Pc* the nominal price of crops is too high even though there is an excess demand for them. In the region PC* E,O the nominal price of crops is too low, but the relative price of crops to goods is too high. Only in the third region O,E,P'g are both the nominal and the relative prices of crops too low. Hence, it is c not possible to infer how crop prices differ from their equilibrium levels merely from the symptom of ex- cess demand for crops once the goods market is also in disequilibrium. The comparison between the ac- tual price of crops and their optimum price, Pc*, is relevant only if the goods price is simultaneously /// raised to Pg*. If instead the goods price is left unal- tered at a level below Pg*, then quite different consid- erations determine the appropriate change in the price of crops. If the economy is characterized by repressed infla- tion, there is invariably a need to raise the price of goods. However, even if this is done, the nominal P g price of crops should be lowered if the economy starts in region C',E,Pc*f and the relative price should be lowered if it is in region PA"EUO. If the goods price is not raised, then a sustainable position can only be reached if the economy is in region P'pk,E. The sus- tainable position is reached by reducing the nominal price of crops until the goods market equilibrium Figure 294. The Dynamics of Shortages locus is reached. p C Dynamic analysis. If the economy enters the re- pressed inflation zone and prices are not altered so as to restore equilibrium, then marketed supply will contract as peasants reduce planned income to ex- pected expenditure. For example, assume that the c economy is in equilibrium at E (in figure 29-3). The government purchases from the trading agency a c quantity of goods for its own consumption, which it finances by printing extra domestic currency. It m freezes nominal prices at their initial levels and k makes no further interventions. Although in sub- sequent periods goods supply to peasants reverts to P c --, normal levels, the peasant sector is in disequilibrium. In aggregate peasants can reduce their excess moneybalances only by selling less, which will cause a reduction in the supply of goods in the next period. If the government has set Pc and Pg at world prices so that there is no taxation, the reduction in goods I supply has the same value as the reduction in crop sales. The absolute magnitude of excess demand for Gl C" goods is unaltered. Peasants continue attempting to 0 . adjust by reducing their income below expenditure. Pg Pg Pg Pg 306 The Liberalization of Price Controls: Thteory and an Application to Tanzania This response in aggregate causes an implosion into nent increase in the price level. The honeymoon ef- subsistence. fect is analogous to an special inflation tax arrange- If, as is more likely, the government imposes some ment in which the government acquires the re- crop taxation so that Pa/Pg there is evidence that these supplies were also Volume of crop sold skewed towards urban centres. For example, of the 309 The Liberalization of Price Controls: neory and an Application to Tanzania supply of soap to towns other than Dar es Salaam derive an estimate of real changes in crop prices around 40% was probably via the parallel market, requires the construction of a price index of goods compared with only around 30% in villages. Such a bought by peasants from urban areas. However, in bias in the parallel market is consistentwith evidence Tanzania there is no cost of living index specifically that it was easier to operate as a black marketeer in for peasants. There is only the national Consumer urban than in rural areas (see Bevan et al 1989a). Price Index (CPI), which forour purposes is inappro- The Board of Internal Trade data suggest that the priate for two reasons. First, although the typical official supplies diverted to the black market3 re- peasant household has a consumption pattern simi- mained largely in urban areas. Bevan et al (1990) lar to the national average, it has a radically different found that in 1983 some three quarters of rural house- expenditure pattern because of subsistence con- holds in Tanzania did buy in black markets. How- sumption. Further, the peasant economy in aggre- ever, the majority of these responded that they were gate has an even more different pattern of net pur- unable to buy in the black market whenever they chases because most of the food which the typical wanted to do so (at the black market price). Thus, peasant household buys is purchased directly or in- these households were rationed in the official market directly from other peasant households. Since food and in the black market. has a weight of 64 percent in the CPI, these differ- Prior to the shortages in the consumer goods mar- ences are important. The CPI is dominated by ket, peasant sales of crops were based upon the ex- changes in food prices, yet the peasant economy in pectation that the resulting income could all be spent aggregate does not purchase food. on desired purchases. As a result of shortages of This is particularly important when measuring the consumer goods the supply curve has become nega- peasant terms of trade with the rest of the economy. tively sloped (Si) and, as shortages have intensified, The price of goods which peasants sell to the rest of it has shifted to the left (S2, S3). the economy should be deflated by the price of the Between 1977 and 1984 the supply of export crops goods which theybuy from it, ratherthanby the CPI. per capita fell by 30 percent and the real producer We have therefore constructed a peasant specific CPI price of export crops fell by 26 percent. According to based on the prices of those goods which peasants our theory, without the fall in crop prices, the decline purchase from the rest of the economy.4 The resulting in crop sales would have been even larger. Economet- series is shown in the second column of table 29-2. ric tests of supply response in 17 regions during the We use the series to deflate the nominal price series, period from 1978 to 1984 have found the volume of assuming that crop income earned in 1982/83 is crop sales to be positively and significantly related to spent in calendaryear 1983, etc. The result is that crop the supply of consumer goods to rural areas and prices did not increase in real terms during 1982/83 negatively (though insignificantly) related to real to 1987/88 (see table 29-1). In fact they declined crop prices. This supports the above account (see substantially: the index fell from 125 in 1984 to 87 in Bevan et al 1989, Chapter 10). 1988. Hence, the increase in the volume of crop sales is not a conventional supply response. Prices de- The 1984 to 1988 period. As shown in table 29-1, clined by 30 percent in real terms in the period of the since the 1984 reforms there appears to have been a refonns. Based on our analysis of the pre-reform substantial increase in the volume of crops sold. Be- decline, we argue that the decline was caused by a tween 1977 and 1985, the volume of official sales of deterioration in theavailabilityof consumergoods in export crops per capita of the rural population de- rural areas. clined by an annual average of five percent; between 1985 and 1988 it expanded by an annual average of Changes in the Denmandfor Money three percent. The recent growth in the volume of food sales has been far more dramatic, particularly In addition to financing expenditure, peasants through nonofficial channels. Total official and non- generate income to add to their cash balances. Cash official food sales per capita of the rural population balances are both the medium of exchange and the increased by 79 percent between 1983/84 and principal financial asset available to peasants. 1987/88. Overall, the volume of crop sales per capita Should desired cash balances fall, the peasant can of the rural population expanded at an annual rate of reduce income (i.e. crop sales) and maintain expen- 10 percent during this four year period. diture while money balances are being run down. Table 29-1 shows that nominal crop prices in- Here we distinguish between real and nominal creased dramatically during 1982/83 to 1987/88. To changes in money balances. 310 The Liberalization of Price Controls: Theory and an Application to Tanzania Table! 29-1. Crop sales: prices in quantities, 1982/83 to 1987/88 (indkces, 1982/1983 = 100) Quantities 1982183 1983/84 1984185 1985/86 1986/87 1987/88 Grains 100 88 107 160 160 183 Export crops 100 101 102 90 119 121 All cnmps 100 94 105 15 139 151 Prices Grains (nominal) 100 193 314 249 252 326 Expoirt crops (nominal) 100 132 174 254 325 416 All crops (nominal) 100 163 208 738 283 376 All crops (real) 100 125 110 92 83 87 Source: Bevan et al (1990a) Real cash balances. The demand for real money holdings of money balances. Although there is no balances is normally related to real expenditure and direct information on peasant holdings of cash bal- hence to real income. Between 1977 and 1983 peasant ances, the mean cash holding of the Tanzanian real cash income per capita fell substantially. Using a household is a reasonable proxy. The latter is known comparison of budget surveys, Bevan et al (1988, from the total cash in circulation (from the Central table 2-14) find a fall of 39 percent. The expectation Bank) and the number of households (from the Cen- is therefore that real cash balances per capita would sus). Since peasants make up 80 percent of the popu- also have fallen. This would have enabled peasants lation, it is likely that their holdings are closely re- temporarily to reduce their cash income by more lated to the Tanzanian average, Although peasants than the fall in their expenditure. The magnitude of have lowercash incomes than theaverage household such an effect depends upon the size of peasant cash (and so would tend to hold less money), they receive holdings relative to income. Peasant cash holding income much less frequently and less predictably was around 75 percent of cash income.5 Hence, had than wage earners, the other major group in the peasants reduced their money holdings so as to re- economy. This implies that, per shilling of income, store this ratio of money to income, the 38.7 percent theywould tend to hold considerablymorecash than fall in income would have permitted the equivalent wage eamers. These two differences from the aver- of a one-year reduction in cash income of 29 percent. age Tanzanian household are therefore qualitatively Had this reduction in money holdings been com- offsetting. The series on real cash balances per capita pleted by 1983, it would imply that over the six year is shown in table 29-3. period from 1977 to 1983, cash income would have Between 1977 and 1981 there was a 21.2 percent been reduced by 5 percent. Hence, crop sales, as an increase in mean real cash balances. If cash income important part of cash income, would also probably declined by around 39 percent during this period, as have been reduced by around this amount. implied by the survey evidence, then the velocity of in fact, however, despite the large fall in real cash circulation halved in the peasant economy.6 Our the- income, peasants appear to have increased their real ory explains why, with the onset of rationing, it is possible for money balances to rise even though Table 29-2. Peasant real crop income expenditure falls. We have also seen that as shortages (indices, 1983 = 100) intensify money holdings must eventually decline. Rural Read The evidence presented in table 29-3 is consistent Nominal purchases incomefrom with this theory. Starting from market clearing in Capfta crop sales from urban areas crop sales per 1977, real balances increased as expenditure declined 1983 100 100 100 until 1981. Thereafter, there was a substantial decline. 1984 152 131 114 By 1983 real cash balances were slightly below their 1985 257 189 130 1977 level. Hence, until 1981 crop supply was 1986 314 159 114 boosted by the change in real money balances. After 1987 398 340 107 1981 real cash balances were reduced; this helps to 1988 547 432 114 account for the severe decline in crop sales during the Source: World Bank data early 1980s. The analysis implies that this effect 311 The Liberalization of Price Controls: Theoiy and an Application to Tanzania would have continued as goods availability deterio- close to its previous peak. These large swings must rated. have contributed to the changes in crop supply. The We have noted that the effect of shortages on real sharp fall in the tax rate during 1983 to 1984 helps to money demand poses a potential problem for the explain the severe economic collapse of those years; transition back to market clearing. By 1986 goods the restorafion of tax rates by means of inflation were much more available in rural areas so that explains why the economy was able to overcome the peasants could reduce money holdings per shilling otherwise major transition problem posed by the of expenditure. As table29-3 shows, real balances fell decrease in real balances. by around a third between 1983 and 1986, and have During the phase of decline in crop sales, 1977 to sincebeen stable. The adjustmentbackonto the 1977 1984, the volume of crop sales per capita fell by ratio of money to expenditure has been achieved around 30 percent. We have argued that it is incorrect partly by increased expenditure, but mainly by re- to interpret this decline as being caused by the fall in ducing real money holdings. This rapid reduction in the real producer price of crops. Abetter explanation real money holdings might have implied that cash is provided by the special circumstances of shortages income, and hence crop supplies, would temporarily of consumer goods which prevailed during this pe- bereduced.Thiswas,however,avoidedbyoffsetting riod. During the four year period from 1983/84 to changes in the price level. 1987/88, there was a spectacular recovery of 47 per- cent in the volume of crop sales per capita of the rural The price level. The demand for money can rise for population. Yet the average crop price fell in real two reasons: either the desired holdings of real bal- terms by 30 percent during the same period. Clearly, ances can increase (as it did during 1977 to 1981), or the remarkable growth in crop sales which reversed the price level can rise. The latter would require a long period of fairly continuous decline was not a households to add to their money balances in order movement up a normal supply curve. Between to keep them constant in real terms. InTanzania the 1983/84 and 1987/88, the rural economy moved inflation tax was initially modest (see table 29-3). backtomarketclearingconditions.Hence,itreverted However, during 1980 to 1982 it became very power- to something approaching the original supply curve ful. This was because the rate of inflation increased (S-S in figure 29-1). As shown in table 29-4, this was and interacted with the honeymoon effect. Since achieved partly by larger imports of consumer peasants had increased their real mroney balances, a goods, financed mainlyby own-funded imports. The given rate of inflafion was eroding a larger stock of abandonment of price controls mainly raised the money. Thereafter, as peasants started to reduce real price level. The rise in the price level eroded real balances, the yield from the inflation tax fell despite money balances, and lowered real crop prices (being a generally rising inflation rate. only partly offset by nominal increases in producer The net effect of the inflation tax and the change in prices). the dem and for real balances is shown in table 29-3. It is expressed as the addition of the two effects and Table 29-3. Improvements in consumer choice, as a percentage of cash income in each year. The latter 1983 to 1988 is of interest because the income peasants need to Availabdity of Availablity devote to adding to money balances instead of to resources (1) of goods (2) expenditure is analogous to a tax on cash income. 1983 0.563 4.8 Therefore it is an implicit income tax, which the 1984 - 4.7 government was levying on peasants and achieved 1985 - 6.7 by the changes in money holdings. 1986 2.544 6.1 This analysis reveals substantial swings in the rate 1987 - 7.3 of implicit taxation. At the start of the period the rate 1988 3.825 7.4 was modest; it was 5 percent in 1977 and 8 percent in Note 1. Refers to how many varities of soap were available 1978. It then rose sharply to an average of 16 percent on a typical shopping trip. Data for soap varities is available during 1979 to 1982. During this phase there was a only for 1983,1986, and 1988. 2. Refers to eight goods: cooking change in the composition of the tax from the 'hon- oil, margerine, toilet soap, laundry soap, matches, khanga, effect' In 983 ad 1984therewas alarge kerosene, and cigarrettes. Figure refers to how many of these eymoon effect. In 1983 and 1984 there was a lage eight goods were found to be available on the average fall to an average of 7 percent. During 1986 to 1988 'shopping trip' by an enumerator of the Bureau of Statistics. The the rate increased to an average of 12 percent. Despite number is the unweighted average of 80 shopping trips per the low level of real balances, the inflation rate was yearS one per quater for eacpubished datagathered by the sufficiently high to restore the implicit tax rate to Bureau of Statistics for the construction of the CPI. 312 T/he Laberalization of Price Contols: Tlwory and an Application to Tanzania Hence our explanation for the paradoxical in- For example, if the number of varieties doubles crease in output is that it reflects the return to market and the elasticity of substitution is 2.5 (the sort of clearing. The abolition of price control made part of magnitude that might be expected), then the implied the money balances held by peasants unnecessary. increase in welfare is 32 percent.7 This is equivalent This monetary overhang would have led to a reduc- to a reduction in the cost of living of 24 percent if the tion in output, as peasants would run down their number of varieties is constant (1/(1-0.24) =1.32). excess money balances by producing less. However, To apply this theory to Tanzania we need data on their incentive to do this was offset by the rise in the the range of product choice available to the typical inflation tax. The combination of the return to market consumer. For this we have adapted the underlying clearing and price increases explains the paradox data gathered by the Bureau of Statistics for the Con- that output increased while producer prices fell in sumer Price Index. Our analysis is confined to the real terms. range of choice among different varieties of soap, but this is the single most important differentiated prod- Liberalization and Consumer Choice uct in the expenditure pattern of peasants according to the 1977 Household Budget Survey. The Bureau We now turn to a final aspect of this liberalization: samples 16 varieties of soap in twenty urban centres its effect on the range of choice of available goods. each quarter. Hence, each year there are 80 shopping visits. When a particular variety is not available, this The number of varities of a good. Price deflators are is recorded by the statistical officer. Using this infor- likely to mis-state the true change in the cost of living mation, we have calculated for 1983, 1986 and 1988 because they do not take into account changes in the theaverage number of varieties of soap found during rangeof choice available forthe typical differentiated these 80 visits. The results are shown in table 29-4. product. For example, even soap, a relatively stand- This example suggests that wider choice has substan- ardized product, is differentiated by scent, quality tially raised living standards. If the soap data are and function (laundry or toilet). Between 1978 and representative (and while for other goods the in- 1984 the range of choice in Tanzania narrowed as part crease in the number of varieties may not have been of the general contraction in supplies of consumer quite as dramatic, for many goods other than soap goods; since 1984 it has widened again. The range of one would expect the elasticity of substitution be- choice affects the cost of living in two ways. First, tweenvarietiestobelower) thechangebetween 1983 since consumers have different requirements, the and 1988 would have raised welfare by more than 20 wider the range of choice the closer are the types of percent even if one adopted as high a value for the good available to what different people ideally want substitution elasticity as 10. Were the substitution to buy. If people who want laundry soap have to elasticity 5 (still a rather high value) the improve- make do with toilet soap, the true cost of washing ment would be 47 percent. This would be equivalent clothes is increased. Second, for many differentiated to a 32 percent fall in the cost of living. products, the typical consumer wishes to buy more than one variety. People will wish to buy both laun- The number of different goods available. Choice can dry soap and toilet soap. also be increased by increasing the number of goods. We now attempt to quantify the effects on the cost In many ways the number of different goods avail- of living of the post-1984 widening of product cho- able is more important than the number of varieties ice. We use the theoretical framework developed by of a single good. This is because the scope for substi- Dixit and Stiglitz (1977). They show how for a given tution, say, between soap and matches is much lower total expenditure a consumer will get more welfare than that between different varieties of soap, and so as variety is increased. In order to achieve a given nonavailability has more deleterious effects upon level of welfare, a consumer will need to spend less living standards. as variety is increased. Thus, increased choice low- We attempt to quantify the widening in the choice ers the cost of living. The magnitude of this effect of goods using the same Bureau of Statistics data on depends upon how closely one variety of a product shopping visits. We consider eight important con- can substitute for another, which is measured by the sumer goods (each usually represented by several elasticityof substitution. When theconsumeris com- varieties). A good is available if at least one of its pletely indifferent between varieties, the elasticity of varieties is available. The results in table 29-4 reveal substitution is infinite and greater variety would not a marked improvement in the availability of the lower the cost of living. range of goods. Between 1983 and 1988 the number 313 The Liberalization of Price Controls: Theory and an Application to Tanzania of different goods available increased by 54 percent. ducer prices may be higher (in nominal or in real The quarterly data reveal that availability deterio- terms) than in a market clearing equilibrium, in spite rated sharply between the first and third quarters of of excess demand. Supply response is likely to be 1983 (from 6.1 to 3.8) and continued at this low level perverse as longas agents expect shortages to persist. until the second quarter of 1984. From then on, i.e. Relying on changes in controlled prices in the transi- cotemporaneous with the reforms which could be tion to market clearing is problematic not only be- expected to have reduced shortages, there was a cause of the informational requirements of such a rapid and fairly continuous improvement. In the case policy, but also because the overshooting which will of the increase in the number of varieties we were be involved maywell confuse agents. In addition the able to indicate the order of magnitude of the change transition can be frustrated by a monetary overhang. in living standards and in the cost of living which it We have suggested that in these circumstances there implied. To do the same thing for goods rather than is a case for aid. varieties is more problematic since there are more Ouranalysis of Tanzania suggests that thesubstan- likely to be large differences between goods in their tial increase in agricultural output since 1984 was not substitutability. Ideally, we would need to estimate a due to producer price increases, but to the return to demand system. Since this is not feasible on the data market clearing. Indeed, without taking into account we can only follow the same procedure as that used that peasants were rationed prior to 1984, it is diffi- for varieties. Again the magnitude of the effect is cult to explain why output increased in the period dependent upon the substitution elasticity chosen. from 1984 to 1988 since crop prices fell dramatically Here we choose a value of 2.5. The improvement in in real terms. availability between 1983 and 1988 (54 percent) would then imply a welfare improvement for given Notes expenditure of 19 percent, or equivalently stated, a 16 percent reduction in the cost of living. 1. The policy presupposes that the government is cred- The effects which we have discussed above are ible, a condition which may not be satisfied in all countries cumulative. The wider choice of varieties may (using for which the analysis is relevant. an elasticity of 5) have lowered the cost of living by 2. There is a proviso to this. It is not safe to conclude 32 percent. The wider choice of goods may have that the economy can move back up the original supply further lowered the cost of living by 16 percent. curve simply by restoring prices. During the intervening Hence, together, they imply a fall in the cost of living years agricultural infrastructure has deteriorated so that it of 43 percent (1 - (1-.32)(1-.16)). For a given level of would probably be unable to handle sales at their 1977 cash expenditure, this implies an increase in the wel- level. This has, in effect, steepened the supply curve. fare derived from it of 75 percent. This is in addition 3. InTanzaniapeople usually speakof parallel markets. to the more conventional changes in living standards That term however is often used in other countries to which we will be analysing below, namely increases denote legal transactions. In Tanzania parallel market ac- in cash income and increases in subsistence con- tivites were illegal and price control was enforced with sumption. The figure of 75 percent is little more than considerable sanctions. a numerical illustration. However, it does suggest 4. From the 1976/77 Household Budget Survey the that there have been substantial gains from wider Bureau of Statistics provided unpublished data on the choice, which are ignored in the usual statistical consumption pattem of rural households. From this we standard of living analysis. Once we take into ac- identified goods supplied from outside the peasant econ- count improved availability producer prices would omy (such as soap and bicycles). Virtually all of these have fallen less than the 30 percent we estimated goods were included in the price gathering process for the earlier. Indeed, if the 43 percent estimate in our ex- construction of the CPI. Hence, by going back to the un- ample is correct, then producer prices would have derlyingpricedataitwaspossibletoconstructapriceindex remained constant rather than fallen in real terms for rural purchases from the rest of the economy on 1977 (1.43 x 0.7 = 1.001). weights. This was undertaken for the period 1983-88. 5. Mean cash income in 1980 per peasant household is Conclusion known from a large sample rural survey (Collier, Radwan and Wangwe (1986)). Mean cash holdings from Central Bank and Census data. Our theoretical analysis of the case of generalized 6. Velocity of circulation (0.613/1.212) = 0.506). price controls has disturbing policy implications. 7. In the Dixit-Stiglitz approach different varieties are Conventional static analysis ceases to apply. Pro- modelled as constant elasticity of substitution substitutes 314 The Liberalization of Price Controls: Theory and an Application to Tanzania in demand. If different varieties are treated symmet- -Bevan, D.L., Collier, P. and Gunning, J.W., (1989(, Peas- rically, the consumer will choose to buy the same ants and Governments: An Economic Analysis, Oxford: quantity of all available varieties. If availability im- Oxford University Press. proves in the sense that n, the number of available ---(1989a), 'Black Markets: Information, Illegality and varieties, increases to n' but that aggregate availabil- Rents", World Development, Vol. 17, no. 12, pp. 1955-63. ity is unchanged (the consumer buys less of each --(1990), Controlled Open Economies: a Neoclassical Ap- variety, but his total expenditure remains constant proach to Structuralism, Oxford: Oxford University when is the substitution elasticity), then the welfare Press. change is given by --(1990a), Price Controls and the Transition to Market u'/u = (n'/n)4 Clearing: Theory and an Application to Tanzania, I.E.S. Hence welfare is increasing in the number of varieties. Applied Economics Discussion Paper No. 94 Collier, P., Radwan, S. and Wangwe, S. (1986); Labour and References Poverty in Rural Tanzania; Oxford: Oxford University Press. Bank of Tanzania (1988), Economic Bulletin, Vol. XVII I no. Dixit, A. and Stiglitz, J.E. (1977); "Monopolistic Competi- 3. tion and Optimum Product Diversity"; American Eco- Berthelemy, J.C. and Morrisson, C. ( 1987), "Manufactured nomic Review, Vol. 67, pp. 297-308. Goods Supply and Cash Crops in Sub-Saharan Africa", Malinvaud, E. (1977); The Theory of Unemployment Re- World Development, Vol. 15, no. 10/ 11, pp. 353-1367. considered; Blackwell; Oxford. ---(1987a), East African Lessons on Economic Liberaliza- Neary, J.P. and Stiglitz, J.E. (1983); 'Towards a Reconstruc- ton, Thames Essay, No. 48, Aldershot: Cower, for the tion of Keynesian Economics: Expectations and Con- Trade Policy Research Centre. strained Equilibriaw, Quarterly Joumal of Economics, ---(1988), 'Incomes in the United Republic of Tanzania Vol. 98; Supplement; pp. 199-228. during the 'Nyerere Experiment'", in W. van Ginneken (ed) Trends in Employment and Labour Incomes, l.L.O. Geneva. 315 30 The Prospects for an Outward Looking Industrialization Strategy Under Adjustment in Sub-Saharan Africa S. Olofin Sub-Saharan Africa continues to be in need of a viable and adopted by several African countries, no explicit and feasible long term industrialization strategy.1 reference is made to export oriented industrialization The inward looking industrialization strategywhich strategy. However, since the 1970s the World Bank many African countries have hitherto pursued has and the International Monetary Fund have favored been a dismal failure. Its inherent inefficiencies have this approached over inward looking industrializa- been well documented in Little et al (1970), Bhagwati tion strategy and have urged its emulation by other and Krueger (1973), Krueger (1978), and Balassa less developed countries (LDCs) (World Bank 1979, (1978), among others. This paper considers Sub-Sa- Hughes 1980, Benerji and Reidel 1980, Fei et al 1979). haran African countries' prospects in looking Perhaps what has come to be known as the Berg forward to an Asian type of export oriented industri- report (World Bank 1981) represents the World alization strategy. Bank's definitive statement regarding desirable eco- Export oriented industrialization strategy is some- nomic policies for Sub-Saharan Africa. times referred to as the G-4 model; its foremost prac- The question as to whether the G-4 model can be titioners are the so-called East Asian 'gang of four'- replicated in other LDCs is not a new one. Numerous Hong Kong, South Korea, Singapore, and Taiwan. studies havebeen undertaken to assess the feasibility Most countries in Sub-Saharan Africa have begun to of its adoption by other LDCs. The fallacy of compo- adopt some type of export oriented industrialization sition argument by Cline (1982) states that if a large strategy in the context of World Bank and Interna- number of LDCs were to simultaneously pursue an tional Monetary Fund assisted structural adjustment export oriented industrialization strategy, it would programs. As of 1988 at least 30 countries in Sub-Sa- lead to an outpouring of manufactured exports. This haran Africa had embarked upon structural adjust- in turn would generate a protectionist response from ment programs (ECA 1989). the industrialized countries. Hughes and Waelbro- Few studies have been undertaken to analyze the eck (1981) believe that the degree of accessibility to implications of structural adjustment programs for OECD markets is still such that generalization of the industrialization efforts in Sub-Saharan Africa. The G-4 model by LDCs is feasible, even if progress Asian newly industrialized countries, particularly would be slow. Studies by Schmitz (1984) and Evans the G-4 countries, have had great success with the and Alizadeh (1984) represent the dependency the- export oriented industrialization strategy. In the oryperspective.Theyargue thatgeneralization of the structural adjustment programs being proposed to G4 model under neoclassical prescriptions may not 316 The Prospects for an Outward Looking Industrialization Strategy UnderAdjustment in Sub-Saharan Africa be feasible, especially given the nonreproducible na- term problems. The third major issue addressed is an ture of the external circumstances which favored the altemative strategy that would focus on minimizing G-4 countries during the 1970s. and possibly eliminating the effects of these internal The first major issue addressed in this paper is the constraints. extent to which the export oriented industrialization The paper is divided into six sections. Following strategy-the G-4 model-may be expected to work this introduction, some theoretical issues are exam- in individual African countries. The fallacy of com- ined in the second section. There is a brief review of position factor is discounted; focus is placed on the past inward looking industrialization strategies in applicability of the model within the context of indi- the third section. In the fourth section, a quick review vidual countries rather than all African countries as and evaluation of export oriented industrialization a group. It is assumed that some of the internal strategy is undertaken and related to the African deficiencies and constraints facing individual Afri- context. This is followed in the fifth section by a can countries can be overcome. If several countries broad outline of a rural oriented smallholder indus- would embark on the strategy simultaneously, as tralization strategy. Possibly an interim alternative was the case in the erstwhile import substitution industrialization strategy, it would aim at dealing strategy, the countries would be unlikely to achieve with some of the existing internal constraints. And it the same level of intensity of exports as is assumed would possibly pave the way for adopting a sustain- in a typical G-4 model. However, the rate of progress able long term industrialization strategy. A short con- would be likely to vary across countries. This wou- clusion in the final section rounds out the paper. ld rule out the fallacy of composition factor, which relies strongly on the assumption of a uniform G-4 Sone Theoretical Considerations regime across countries (Cline 1982). It is assumed that no individual country would be able to influence It would seem that the theoretical arguments of the the supply of manufactured goods so as to instigate orthodox general equilibrium school and the unor- protective barriers against its exports. For individ- thodox disequilibrium approach are not directly rele- ual African countries, the constraints on intema- vant to the issues addressed in this paper. The ap- tional market demand would not be a major limiting proach being adopted by the World Bank and the factor. International Monetary Fund in designing ongoing The viability of an export oriented industrializa- adjustment programs seems to have little to do with tion strategy in individual African countries is ana- extreme doctrinaire considerations. It would appear lyzed. It is assumed that international demand con- that the underlying considerations are more prag- straints and absorptive capacity in industrialized matic. However, the pragmatic approach is linked to countries do not constitute major barriers.2 Supply some underlying theory. The neoclassical arguments and demand potentials for successful export-led in- surrounding free markets, trade liberalization, price dustrialization strategy are distinguished. Greater incentives, and other issues provide the springboard emphasis in this paper is placed on the formner. It is for the Asian newly industrialized countries' type of assumed that, for the Sub-Saharan Africa region, export oriented industrialization strategy. These and ceteris paribus, intraindustry, intraregional, and the dependency theorists' counter arguments should South-South trade in manufactured exports could be used in the search foran appropriate development substitute for interindustry North-South trade. In an strategy for Sub-Saharan Africa. intraregional and South-South trade context, market One of the central issues which needs to be ad- access would not be likely to constitute a major prob- dressed from a theoretical point of view relates to the lem. It is hypothesized that Sub-Saharan African role of the state vis-a-vis the invisible hand in the countries' collective terms of trade would improve as pursuit of standard neoclassical results. Some of the intraregional and South-South trade grows. existing empirical studies, for example Chow (1987), The second major issue the paper seeks to address tend to tilt the balance overwhelmingly in favor of is the serious nature of the limiting internal con- standard neoclassical prescriptions. They attribute straints which have contributed to the ineffective- positive results to assigning a greater role to the ness of inward looking industrialization strategy in invisible hand. The arguments and counterargu- Sub-Saharan Africa. The doctrinaire emphasis of a ments regarding the role of the state are ideological. modified version of export oriented industrialization They tend to make little or no room for examining strategywould appeartobeneglectingthese internal the realities in each country in relation to the mini- constraints. This is partly due to ideological biases mum necessary and sufficient conditions for obtain- and partly due to preoccupation with pressing short ing highly desirable standard neoclassical results. 317 The Prospects for an Outward Looking Industrialization Strategy UnderAdiusbtcnt in Sub-SaharauAfrica However, such necessary and sufficient conditions ported. Those countries with a mineral resource base do not exist in the majority of Sub-Saharan African -such as Zaire, Zambia, Liberia, and Nigeria-have countries. Too much focus on ideological biases tended to neglect agriculture, to take advantage of clouds the real issues. The choice between market readily accessible rents and foreign exchange earn- forces and state intervention may be purely secon- ings, and to pursue similar inward looking, import dary in light of the empirical experience of theregion. substituting strategy. With the onset of declining C6te d'Ivoire, Malawi, and Senegal have followed commodity prices and terms of trade in general, this market oriented strategies as opposed to the state inward looking strategy has run into serious difficul- interventionist strategies followed in Guinea, Ghana, ties in the agricultural primary product exporting Sudan, Zambia, and Ethiopia during the 1%0s and countries and in the rent earning countries. It has 1970s. Both groups of countries experienced similar resulted in the unprecedented debt burden that is results-little or no industrialization, heavy indebt- threatening to wipe out whatever modest growth edness, and other characteristics of the underdevel- gains may have been realized in the two or more opment trap. The evidence tends to suggest that decades of pursuing inward looking industrializa- across ideological lines the problems in the 1%Os and tion strategy. 1970s had more to do with the wrong strategy of In essence, inward looking industrialization devel- inward looking industrialization (Acharya 1981). opment strategy used trade restrictions in the form This paper focuses on two primary theoretical and of tariffs, quotas, and multiple exchange rates to empirical questions in relation to the challenges of promote industrialization and correct balance of underdevelopment in Sub-Saharan Africa. The first payments difficulties. The model and the results de- is how the necessary and sufficient internal condi- rived from its implementation have been subject to tions for deriving standard neoclassical results (as much scrutiny. The overwhelming conclusion has evidenced in the Asian newly industrialized coun- been one of disenchantment, even by some of its tries)maybecreatedinSub-SaharanAfrica.Thismay early advocates (Prebisch 1964, Hirschman 1968, require some measure of strong state intervention. Briton 1970, Baer 1972, Sutcliffe 1971, and Diaz-Ale- The second is how such considerations would go jandro 1975). beyond the static efficiency concerns of neoclassical The main arguments against inward looking in- prescriptions, to dynamic efficiency considerations. dustrialization strategy focus on the damaging ef- To achieve this, social costs and profitability criteria fects of too much protectionism. This has invariably would sometimes have to override private costs and led to inefficient allocation of resources due to distor- benefit criteria. tions in factor and product markets. Criticism is also made of this strategy's bias against exports and agri- Past Inward Looking Industrialization Strategy culture and its import intensity, which has often ex- acerbated the balance of payments pressure it was From the late 1950s to the early 1980s, the industri- intended to relieve. There has been widespread dis- alization programs that were adopted in Sub-Saha- illusionment with inward looking industrializafion ran Africa canallbecharacterized asinward looking. policy across ideological and analytical divides, es- As it was in other LDCs, this approach to develop- pecially in its static efficiency aspects. ment was inspired by the mainstream theories of the Most of the Sub-Saharan African countries, along 1950s and 1960s-such as in UN (1951), Singer (1950), with other debt ridden LDCs, have had no choice but Nurkse (1953, 1959), Scitovsky (1954), Leibenstein to adopt World Bank and International Monetary (1957),andRostow(1960),amongothers.Thesetheo- Fund assisted structural adjustment programs. ries advocated strategies which included the notion Many of these countries have balance of payments of the need for a big push, a productive role for the problems partly caused by the negative effects of state, and rapid industrializafion through inward inward looking industrialization strategy, and partly looking import subsfitution. by domestic mismanagement of resources and unfa- Most of the Sub-Saharan Africa countries, regard- vorable exogenous factors, such as declining terms of less of their ideological leanings, have continued trade. their preindependence pattern of relative neglect of While the empirical findings on inward looking the supply oriented productive sectors of the econ- industrialization strategy are generally accepted, omy. They favor a weak export policy oriented to- their orthodox interpretafions, policy conclusions, wards the exports of a few agricultural cash crops. and recommendations are not necessarily widely Finished consumer manufactures and raw materials shared. Proponents of the dependency theory school for import substituting light manufactures are im- argue for increased state intervention (Merhav 1969, 318 The Prospects for an Outward Looking Industrialization Sbtategy Under Adjustment in Sub-Saharan Africa Sunkel 1973, Vaitos 1974, and Thomas 1974). How- trialized countries. An attempt is made in this paper ever, despite the validity of some of these counterar- to address the second issue. For reasons of space, the guments and the political unpopularity of some of over-involvement issue and related concerns per- the conditionalities, structural adjustment programs taining to the political economy of adjustment are not have come to stay in most of these countries for a addressed. According to Gulhati (1988), these should number of reasons. be of great concern not only to economists and poli- First, the programs have inevitably become part of cymakers in these countries, but also to the World the policy packages or remedies most indebted LDCs Bank and the International Monetary Fund. must accept to obtain relief from a crushing debt The inward looking industrialization strategy has burden. Second, and perhaps more importantly, the produced modest growth in manufacturing output strong support from the advocates of the new policy, in some countries-such as C6te d'lvoire and Kenya. especially the Intemational Monetary Fund and the However, over an extended period this growth has World Bank, partly stems from the empirically ob- created very limited employment opportunities in servable success stories of those countries that have relation to the magnitudeof the investment involved. adopted the policy. In particular, the Asian newly Emphasis hasbeen on light manufacturingwith little industrialized countries, of which the G-4 is most scope for growth in complementary exports to pay prominent, have been successful. Inward looking for raw material inputs, and little explored scope for industrialization strategy is shown to be ineffective sourcingoutput inlocal raw materials. The prospects and is discredited, while the record of those countries for reviving growth based on relocation of produc- which switched to an outward looking, free market tion by industrialized countries is not bright, given oriented policy is there for everyone to see. Although the rising costs of imported energy and skills which this success may be difficult to replicate in other have rendered such ventures uneconomical as rela- developing countries, Cline (1982), the International tive prices have changed or foreign exchange Monetary Fund, and the World Bank continue to squeeze has rendered them inoperable. spread this "good news" through official and nonof- As table 30-1 shows, the manufacturing sector ficial channels-including Zulu and Nsouli (1989), now contributes over 10 percent of GDP in countries World Bank (1989), Landell-Mills et al (1989), Mc- such as C6te d'Ivoire, Senegal, and Kenya. However, Cleary (1989), Thomas and Chhibber (1989), and its contribution to employment remains between 2 Hum phreys and Jaeger (1989). and 3 percent of the labor force. Invariably, 80 percent There are at least two bothersome aspects about or more of industrial value added is in import substi- this new gospel which call for careful examination. tuting activifies, principally in early stage lines such The lirst has to do with the political economy of as food and beverage processing, textiles, garment adjustment. The World Bank and the Internafional manufacture, wood products, paper and printing, Monetary Fund's role appears to be gradually mov- and a few exceptionally capital intensive projects ing away from that of a disinterested lender of the such as cement manufacture, fertilizers, metal proc- last resort to debt ridden nations to that of deep essing, and petroleum refining. There is also a heavy involvement in country policy. The level of involve- bias against exports (Acharya 1981). ment in day-to-day policy formulation and imple- Thus inward looking industrialization strategy as mentation is now such that the role and status of the an industrialization and overall development strat- resident representative is beginning to approximate egy appears to have reached its limits. One doubts if that of a preindependence colonial governorin some an outward looking industrialization strategy or any countries. The second bothersome aspect relates to variantofitwouldfairbetter,aslongasthenecessary doubts as to which variant of neoclassical compara- basic preconditions for successful industrialization tive cost advantage theory inspires the policy pre- are not present. scriptions under structural adjustment programs. Some of the fears expressed for example in Mkan- Export Oriented Industrialization Strategy and the dawire (1988) imply that the seemingly deliberate African Context deindustrialization stance of structural adjustment programs in many instances maybe coming straight The characterization of the success in the Asian from Hla Myint (1958). This in effect would require newly industrialized countries varies between the that these countries go back to preindependence pro- supportive neoclassical interpretations and the criti- duction and trade structures. It would not enable cal alternative dependency theory viewpoints. The them. to look forward to the type of outward looking literature varies on which countries are included in industrialization achieved by the Asian newlyindus- the newly industrialized countries. However, there 319 The Prospects for an Outward Looking Industrialization Strategy UnderAdjustment in Sub-Saharan Africa Table 30-1. Agriculture amd manufactuirng as a percentage of Gross Domestic Product in African countries 1960-1975- 1980-1985' Country Agriculture Manufacturing Agriculturc Manufacturing Algeria 10 14 6 12 Benin 39 5 36 7* Botswana 34 12* 7 8* Burkina Faso 40 13* 43 14* Burundi 63 6 52 8 Cameroon 30 11 22 11 Central African Rep. 31 13* na na Chad 41 11* 41 16* Congo 13 8 7 4 C6te d'lvoire 27 13 28 12 Djibouti 3 6 4 3 Egypt 28 20 18 1 Equatorial Guinea 64 1 59 1 Ethiopia 60 7 45 10 Gabon 26 10 5 7 Gambia 38 5 35 5 Ghana 47 11 41 11 Kenya 33 10 27 11 Lesotho 62 1 21 6 Liberia 24 2 17 7 Libya 5 2 2 3 Madagascar 37 3 15* Malawi 61 6 52 9 Mali 53 7* 53 7' Mauritania 27 2* 19 9* Mauritius 20 12 13 17 Morocco 20 16* 18 17' Niger 61 6 43 4 Nigeria 59 4 22 8 Reunion 11 8* 1 1* Rwanda 52 13 43 16 Senega; 25 15* 18 21* Seychelles 10 5 7 9 Sierra Leone 30 6 39 7 Somalia 48 43 5 South Africa 12 21 5 20 Sudan 52 4 34 73 Swaziland 24 12* 20 16* Tanzania 42 7 53 5 Togo 46 5 27 6 Tunisia 25 8 15 12 Uganda 52 9 49 8 Zaire 17 8 rna na Zambia 14 7 14 22* Zimbabwe 18 14 12 26 Note: aThe year varies from one country to another with the year for which data is available being selected within the range for each country. na indicates that data is not available. (.) indicates a value of less than one half of one percent. 'Manufacturing combined with electricity, gas and water or mining and quarrying. Source U.N. Yearbook of National Accounts (various years). 320 The Prospects for an Outward Looking Industrialization Strategy UnderAdjustment in Sub-SaharanAfrica has been particular focus on the so-called Asian Ti- and subsidised infrastructure in the newly industri- gers or G-4 countries-South Korea, Taiwan, Hong alized countries. This was at a time when declining Kong, and Singapore. While the G-4 model specifi- profits, increasing labor costs, and taxes made com- cally refers to these four countries, to some extent petition in the industrialized markets of the West their development strategy has been followed in a more difficult. As special value added taxes were numberof othercountries, includingBrazil and Mex- introduced itbecame more profitableto farm out part ico. Reference to the G-4 model in this paper refers to of the production processes to low wage countries. the broad experience of this group of countries, The result was an increase in the multinational cor- which accounted for 62 percent of LDC manufac- porations' share in newly industralized countries' tured exports in 1975 (Lall 1980).Duringl965 to 1978, exports-31 percent for South Korea in 1974,84 per- it achieved annual real GDP growth rates of 11 per- cent for Singapore in 1975, 51 percent for Brazil in cent, yearly increases in manufactured exports of 1973, and 34 percent for Mexico in 1974 (Lall 1980). between 20 and 40 percent, and annual increases in The degree and manner in which the newly indus- manufacturing employment of between four and trialized countries became locations for multina- eight percent (Schmitz 1984). tional production varied from one country to an- The dominant explanation for this success is other. But in general their becoming such locations hinged on the Heckscher-Ohlin-Samuelson factor forned part of the context for the expansion of their proportions theory of comparative cost advantage exports. This relocational policywasenhancedby the (Balassa 1981, Bhagwati 1978, Krueger 1978, West- practice of permittingmanufactured imports to enter phal 1978, Little 1981, and Ranis 1981). The argument national markets partially free of duties whenever is that these countries adopted the right policies by raw material originated from the country of impor- liberalizing imports, adopting realistic exchange tation. It is estimated that United States imports from rates, and providingincentives for exports. Above all LDCs under these tariffs increased by 295 percent they managed to get factor prices right so that their annually between 1966 and 1979, with the largest economies, especially the manufacturing sector, shares coming from Mexico, Taiwan, Singapore, and could expand in line with their comparative cost Hong Kong (Schmitz 1984). advantage. It is further argued that their reliance on Coinciding with this was the emergence of the market forces and integration into the world econ- Eurodollar market, whose supply of funds was fu- omy yielded results superior to those under policies eled mainly by balance of payments deficits in the of protection and dissociation from the world econ- United States. These deficits resulted from massive omy (Schmitz 1984). military expenditure in Vietnam and, in the late Others do not support this embodiment of the 1970s, the recycling of surplus petrodollars from the neoclassical parable which ascribes their successes OPEC countries. It is estimated that credit from pri- primarily to export oriented policy changes. Instead vate multinational banks, which became the main they attribute the rise of the newly industrialized vehicle for recycling these funds, expanded by more countries to cyclical and historical national and inter- than 50 times between 1966 and 1978. Over 50 per- national factors. These have produced favorable ac- cent of the loans went to LDCs. Brazil, Mexico, and cess to advanced countries' markets, increased access South Korea accounted for about 50 percent of the to international finance, and increased relocation of total accumulated debt to multinational banks by production to the periphery by transnational corpo- 1980 (Griffith-Jones 1980). It is argued that access to rations (Bienfeld 1982). A number of countries were this private capital market allowed these countries to able to take advantage of these favorable conditions, avoid the Intemational Monetary Fund's condition- it is argued, due to their location and geopolitical ality on economic policy. This is not the current ex- significance, the existence of strong (repressive) in- perience of Sub-Saharan Africa countries (Schmitz temationally reliable regimes (Evans and Alizadeh 1984). The newly industrialized countries were able 1984), and the existence of technological infrastruc- to maintain levels of imports which were not sustain- ture resulting from earlier inward looking industri- ablefrom exports in the early periods of theirgrowth. alization policy and state control over industrial de- This enabled them to maintain high growth rates in velopment (Schmitz 1984). investment and output, despite their increasing The mid-1960s witnessed a tremendous surge in debts, because they were able to finance theirbalance locational decisions of multinational corporations of payments deficits from external private capital dealing in labor intensive manufactures. The multi- flows. Studies have shown that liberal trade policies nationals tookadvantageof thecheap labor force and were accompanied by significant capital market con- generous incentives in the fonn of tax exemptions trols, with governments controlling as much as 60 321 The Prospects for an Outward Looking Industriaization Strategy Under Adjtstment in Sub-Saharan Africa percent or more of investable funds as a means of gration to enhance productivity may not be sustain- state direction of economic activity (Wade 1982, able in the long run. The four internal constraints Datta-Chaudri 1981). would constitute major obstacles to effective absorp- There is also the viewpoint that inward looking tion of surplus rural labor into a modernized urban industrialization strategy and outward looking ex- sector. In such circumstances, a reverse urban-rural port oriented industrialization strategy may not nec- migration would relieve the pressure in urban pov- essarily be substitutes. They could be complements, erty enclaves. with the former achieving the accumulation of indus- A double pronged interim industrialization strat- trial experience, technical skills, and entrepreneur- egy could be an alternative to inward looking and ship, as was the case in Korea. The two strategies export oriented industrialization strategies. It would could be implemented together (Krueger 1981). first emphasize increasing labor productivity A study by Chow (1987), based on eight newly through improvements in skills, technology, and en- industrialized countries, concluded that there was trepreneurial ability in the rural setting, where 80 overwhelming empirical evidence supportive of the percent or more of the labor force resides in Sub-Sa- a proiri argument in favor of export expansion as a haran Africa countries. Second, it would take advan- development strategy for LDCs. However, there are tage of the relative abundance of labor and land to some basic characteristics that are common to Sub- pursue labor intensive production techniques in Saharan African economies that would be inimical to smallholder units in rural areas. The rural oriented an export oriented industrialization strategy.3 Some smallholder industrialization strategy advocated in of these characteristics predate or have been exacer- this paper is referred to as the ROSH strategy. The bated by colonial experience and would need to be ROSH strategy maybe viewed as an interim measure tackled if any industrialization program or overall that would pave the way for a longer term strategy. development strategy is to be sustainable over the It could possibly make the adoption of an export long run. The most limiting characteristics include: oriented industrialization strategy feasible. * absence of a viable entrepreneurial class, ROSH Development Strategy and State Intervention * poor work ethic and the resulting low produc- tivity of labor, The idea of a rural oriented smallholder industri- * acrimonious ethnic cultural plurality within alization strategy considered here is by no means artificial geographical boundaries that have entirely new.4 However, it remains unpopular in Af- made it difficult for cohesive nation states to rican LDC policy circles for a number of reasons. evolve, and First, there is a fascination with big projects and * technological backwardness. large-scale industries. There is misplaced confidence in their ability to enhance rapid economic growth. These four fundamental constraints have given This attitude is related to the big push theories of the rise to several others. 1950s. Second, colonial and post-colonial pattems of investment have always favored large-scale invest- A Case for Rural Oriented Srnallholder ment in capital intensive enterprises, infrastructure, Industrialization Strategy mining industry, and agriculture. Third, thegrowing incidence of heavy foreign debt burden has left pol- The four internal constraints on successful long icy makers with little or no choicebut to adopt strate- term industrialization listed above explain why the gies favored by their creditor countries and support- erstwhile inward looking industrialization develop- ing multilateral agencies, particularly the World ment strategy has not been sustainable in the Sub-Sa- Bank and the International Monetary Fund. haran Africa economies. The same would be true for The rationale for and content of ROSH strategy an export oriented industrialization strategy or any would in several respects be similar to its charac- urban oriented strategy based on capital intensive terization in Acharya (1981); Daniel, Green, and Lip- production. Given these constraints, a neoclassical- ton (1985); and Johnston and Kilby (1975). The advo- inspired liberalization policy would not cure market cates of an export oriented industrialization para- distortions; it would compound them in the absence digm have drawn inspiration from the successful of the prerequisites for taking advantage of market experience of the Asian newly industrialized coun- efficiency in a modem economy. Similarly, an urban tries. The ideas behind ROSH strategy are drawn oriented modernization effort aimed at realizing a heavily from the experience of the Peoples Republic Lewis-Fei-Ranis neoclassical type of rural-urban mi- of China, India, and to some extent Pakistan. In these 322 T7he Prospects for an Outwvard Looking Industralization Strategy Under Adjustment in Sub-Saharan Africa countries the smallholder approach to industrial de- state intervenfion, especially at the early stages when velopment has recorded some measure of success, such intervention maybe crucial to successful imple- although not on the unprecedented scale associated mentation. with the G-4 countries. This argument for direct state intervention differs Given that the basic elements of the small holder from the justification for such intervention by the approach to industrialization are sufficiently famil- dependency theorists. They argue for increased state iar, they will not be repeated here.5 What may be intervention on purely ideological grounds. Instead novel in this paper's characterization of ROSH strat- in this paper initial state intervention is seen as inevi- egy has more to do with the institutional framework table to promote the necessary initial conditions for for the implementation of a smallholder approach to effectiveness of private sector initiative. development. Earlier advocates of a ROSH type of strategy (Oshima 1962 and World Bank 1989) have ROSH Strategy and Small-scale Manufacturing tended to favor an implementation strategy which assigns a major role to the private sector. However, Industrialization efforts in Sub-Saharan Africa stronggovernmentintervention maybe crucial to the have been faced with two major constraints in the successful implementation of a ROSH strategy, espe- past. First, there has been little or no attempt to link cially within the context of African LDCs. manufacturing to domestic sources of raw material. Although private sector initiative should be en- Second, the inward looking industrialization strat- couraged as much as possible, the standard neoclas- egy form of manufacturing has not promoted devel- sical prescription of reliance on the profit motive is opment of domestic skills, technological know-how, not likely to succeed. First, the shortage of manage- and the effective adaptation of imported technology. rial and technical skills and the absence of an entre- This has followed from the import dependent nature preneurial class would require that these cadres be of manufacturing. created before one could expect market forces to For any long term industrializafion strategy to be operate efficiently. The responsibility for creating the sustainable, the parallel development of agriculture conditions for technological innovation and acquisi- and manufacturing must be broken. The two sectors tion of skills cannot be entrusted to the private sector. should be made to complement each otherby way of The provision of these minimum requirementsby the resource inputs and domestic acquisition of simple state is akin to its providing necessary physical infra- technical skills and domestic technological innova- structure and incentives to private investment capi- tions. These are prerequisites to the effective master- tal. A strong case could be made for initial direct state ing and adaptation of more sophisticated imported participation. The state would transfer its holdings technology. Consequently the focus of policy would to private investment capital as the original deficien- be on encouraging the growth of local small-scale cies would be overcome. This would make such ven- manufacturing in rural areas. This would make use tures sufficiently attractive to foreign and domestic of direct inputs from agriculture and supply direct private sector capital. inputs into improving the technology in agriculture. Second, the smallholder artisan units would likely The small-scale manufacturing of a wide range of initially produce low quality goods with lack of hand tools and animal powerfarmingimplements- standardization. Government intervention would be such as ploughs, cultivators, seed fertilizer drills, and required to ensure protection from higher quality, planters-could provide the beginnings for an in- better standardized products from large- and me- dustrialization process that would have adeqaute dium-scale plants (Little and Mazumdar 1977). inputs and foreign exchange. It would also increase Third, the observed failure of inefficient parastat- the potential for rural nonfarm employment genera- als in these economies over the last two or three tion. The acquisition of technical skills could serve as decades may not be entirely due to the inferiority of a basis for a wider form of industrialization at a later state intervention in private sector initiative. It may stage. The historical experience of India and the Peo- have had more to do with the deficiencies of inward ples' Republic of China shows that this form of rural looking industrialization as a development strategy, industrialization could provide fertile ground for the and its emphasis on large-scale capital intensive en- development of efficient and increasingly sophisti- terprise, which the overburdened bureaucracy was cated small-scale manufacturing enterprises (ohn- ill-equipped to operate efficiently. Carrying out nec- ston 1978). essary parastatal reforms and scaling down their The scope for this form of small-scale industry in activities to smallholder enterprises as envisaged un- Africa is largely uncharted. It could be significant, der a ROSH strategy may enhance the efficiency of especially if small businesses can be induced to take 323 The Prospects for an Outward Looking Industrialization Strategy Under Adjustment in Sub-Saharan Africa advantage of the mutual supporting links between Sub-Saharan Africa to replicate the experience of agriculture and rural industry to promote rural in- China and India with little or no external interfer- come growth (Acharya 1981). Exploiting this linkage ence. China and India have pursued inward looking, would foster quick transformation of agriculture to self-reliant smallholder industrialization strategies more scientific land intensive methods of cultivation. as a prelude to opening up to the rest of the world. Manufacturing could quickly progress from the This preparatory phase may be far from complete in manufacture of relatively simple implements such as the cases of China and India, which remain essen- ploughs and carts to that of seed drills, electric mo- tially closed economies. tors, diesel engines, stationary threshers, and power Another problem is the elite ruling class, which is fillers. In addition, the manufacture of simple farm a colonial legacy in Africa and has a track record of equipment has often led to the manufacture of con- inept resource management. It remains doubtful that sumer goods such as electric fans, bicycles, and sew- the commission agents who pass for entrepreneurs ing machines, and producer goods such as oil seed in these countries would support the emergence of expellers, lathes, and hand drill presses (johnston true entrepreneurs who have to eam their profits. 1978). Because of the pervasive importance of metal- There would also be the problem of weaning the working skills, the technical and managerial capa- population of their taste for high quality and sophis- bilities bred in rural workshops could greatly en- ticated imported manufactures. Low quality manu- hance the development of a light engineering indus- factures would be produced by small-scale units in try, which would be crucial for the assimilation of the initial stages of their development. Chronic po- more complex and appropriate technologies. litical and institutional instability are additional problems in these countries. Some Advantages of ROSH Strategy Over Earlier Perhaps the greatest obstacle that may militate Strategies against the successful adoption of ROSH strategy as a development strategy is the debt crisis in Sub-Sa- While primarily making for acquisition of skills, haran Africa. This obstacle will remain as long as the technological know-how, and entrepreneurial skills current debt burden relief measures continue to em- that have been terribly deficient in these countries, a phasize debt rescheduling and other short term pal- ROSH strategy would have a number of added ad- liatives rather than debt forgiveness, repudiation, or vantages over inward looking and export oriented cancellation. As long as there is a debt problem there industrialization strategies. First, it would minimize may be no room for any new industrialization or the need for a massive infusion of net foreign capital overall development strategy initiative. It may be inflows and free the development process from the that an approach like the ROSH strategy would not familiar savings and foreign exchange constraints in be seen as consistent with the short term interests of two-gap models of development. Second, it would creditor countries. However, it is hoped that it wou- conform with the self-reliance philosophy which is Id be obvious that whatever can be done to pull these dominant in these countries. Nationalist philosophy fragile economies out of their chronic dependency is inevitable in the face of increasing North-North and equip them tobe able to respond more meaning- trade links, with greater emphasis on high tech fully to the demands of a changing global economy manufactures and reduced dependence on raw ma- would in the longer run be in everybody's interest. terials from LDCs.6 Third, the reduction in manufac- turing dependence on imported raw material and the Conclusion corresponding self-sufficiency in food and basic manufactured goods could reduce balance of pay- The Sub-Saharan Africa region continues to search ments pressures. Fourth, successful implementation for a viable, long term industrialization strategy. of a ROSH strategy would prepare these economies There may be very little to borrow from Asian newly for fuller involvement and integration into the global industrialized countries-especially the G-4 coun- economy on a more equitable basis. tries-which would be of direct relevance to the re- gion's peculiar characteristics. More appropriate les- Some Envisaged Implementation Problems sons may come from India and the People's Republic of China, which have had a less spectacular, more As the industrialized world moves into the age of gradual approach to industrial growth and develop- the information revolution, it is questionable how ment compared with the G-4 countries. It is impor- feasible a deliberate isolationist strategy could be. It tant that the region try to tackle first its self-limiting is doubtful that the rest of the world would allow internal constraints. Otherwise it may continue to 324 The Prospects for an Outward Looking lndsstrialization Stfrategy Under Adjstment in Sub-Saharan Africa find it difficult to take advantage of any favorable M.Godrey (eds.) The Struggle for Development: National external conditions. A well implemented ROSH Strategies in an International Context. (Chichester: John strategy would enhance the prospects for a sustain- Wiley). able long term industrialization strategy regardless Briton, H.J. (1970): 'The Import Substitution Strategy of of external conditions. Economic Development: A Survey', Pakistan Develop- ment Review, Vol. 10, No. 2. Notes Chow, P.C. (1987): 'Causality between export growth and industrial development: Empirical evidence from the 1. This presupposes that rapid industrialization is still NICs', Journal of Development Economics, Vol. 26, No. 1. regarded as a desirable primary component in the drive Cline, W.R. (1982): "Can the East Asian Model of Develop- towards achieving self-sustaining economic growth and ment be Generalized?", World Development, Vol. 10, No. overall economic development. It would be safe to assume 2, pp. 81-90. that this still holds in most economic development minis- Daniel, P., R.H. Green, and M. Lipton (1985): 'A Strategy for tries in Africa, and probably all developing countries. the Rural Poor', Journal of Development Planning, No. 15 2. This assumption would have to be relaxed in an (United Nations), pp. 113-136. n-country case, where n > 1. Datta-Chaudri, M.K. (1981): 'Industrialisation and Foreign 3. A fairly detailed examination of these characteristics Trade: The Development Experiences of South Korea has been undertaken by the author elsewhere; see Olofin and the Philippines' in E. Lee (ed.) Export-led Industriali- (1989). sation and Development, Geneva: Asian Employment Pro- 4. See for example Oshima (1962), UNDP (1974), and gramme, International Labour Office. Kilby (1975). Althoughthisideaisnotnew,ithasnotbeen Diaz-Alejandro, C. (1975): 'Trade Policies and Economic incorporated into a rigorous and formal paradigm. From a Development' in Peter Kenen (ed.), International Trade policy point of view, it needs to be made more appealing and Finance: Frontiers for Research, (Cambridge: Cam- than the current emphasis on large-scale projects. bridge University Press). 5. For a lucid discussion of the main features of a small- Evans, D. and P. Alizadeh (1984): 'Trade, Industrialisation holder approach, see Acharya (1981) and Daniel et al and the Visible Hand', Journal of Development Studies, (1985,). Special Issue on Industrialisation. 6. There are ominous prospects for further decline in Fei, J.C., G. Ranis, and W.Y. Kuo (1979): Growth with Equity. North-South trade, due to the planned economic integra- The Taiwan case, (New York. Oxford University Press). tion of Europe in 1992 and the greater attention being Griffith-Jones, S. (1980): 'The growth of Multinational focused on Eastern Europe. Banking, the Euro-currency Market and the Effects on Developing Countries', Journal of Development Studies, References Vol. 16, No. 2. Gulhati, Ravi (1988): The Political Economy of Reform in Acharya, S.N. (1981): 'Perspectives and Problems of Devel- Sub-Saharan Africa, EDI Policy Seminar Report No. 8, opment in Sub-Saharan Africa', World Development, Vol. (Washington, D.C.: World Bank). 9, pp. 109-147. Hirschman, A.O. (1968): 'The Political Economy of Import Baer, W. (1972): Industrialisation and Economic Development Substituting Industrialisation in Latin America', Quar- in Brazil, (Homewood, Illinois). terly Journal of Economics, Vol. 82, No. 1, Februaiy. Balassa, B. (1981): The Newly Industrialising Countries in the Hughes, H. (1980): 'Achievements and objectives of indus- WVorld Economy, (Oxford: Pergamon Press). trialisation' in J. Cody, H. Hughes and D. Wall (eds.) Banerji, Randadev and James Riedel (1980): 'Industrial Policies for Industrial Progress in Developing Countries, employment expansion under alternative trade strate- (New York: Oxford University Press), pp. 11-37. gies: case of India and Taiwan: 1950-70', Journal of Devel- Hughes, H. and J. Waelbroeck (1981): 'Can Developing opment Economics, vol. 7, No. 4, pp. 567-577. Country Exports Keep Growing in the 1980s?', The World Bhagwati, J.N. and Anne Q. Krueger (1973): 'Exchange Economy, Vol. 4, No. 2. June. pp. 127-148. control, liberalization and economic development', Humphreys, C. and W. Jaeger (1989): 'Africa's Adjustment American Economic Review, May. and Growth', Finance & Development, June. Bhagwati, J.N. (1978): Foreign Trade Regimes and Economic Johnston, B.F. (1978): 'Agricultural production potentials Development: Anatomy and Consequences of Exchange Con- and small farmer strategies in Sub-Saharan Africa', in trol Regimes, (Cambridge, MA.: Ballinger Press). S.N. Acharya and BF. Johnston TVO Studies of Develop- Bienfeld, M.A. (1982): 'The Intemational Context for Na- ment in Sub-Saharan Africa, World Bank Staff Working tional Development Strategies: Constraints and Opp- Paper No. 300 (October). ortunities in a Changing World', in M. Bienfeld and 325 The Prospects for an Outward Looking Industrialization Strategy UnderAdjistment in Sub-Saharan Africa Kilby, P. (1975): 'Manufacturing in Colonial Africa' in P. Ranis, G. (1981): 'Challenges and Opportunities Posed by Duiguan and L.H. Gann (eds.) Colonialism in Africa: 1978- Asia's Super Exporters: Implications for Manufactured 1960, Vol. 4 (London: Cambridge University Press). Exports from Latin America', in W. Baer and M. Gillis Krueger, A.O. (1978): Foreign Trade Regimes and Economic Export Diversification and the New Protectionism - The Ex- Development: Liberalization Attempts and Consequences, perience of Latin America, National Bureau of Economic (Cambridge, Mass: Ballinger Press). Research and the Bureau of Economic Research, Univer- Krueger, A.O. (1981):'Export-led Industrial Growth Recon- sity of Illinois. sidered' in W. Hong and L.B. Krause (eds.) Trade and Rostow, W.W. (1960): The Stages of Economic Growth (Lon- Growth of the Advanced Developed Countries in the Pacific don: Cambridge). Basin, (Seoul, Korea: KDI). Schmitz, H. (1984): 'Industrialisation Strategies in Less De- Lall, S., (1980): 'Exports of Manufactures by New Industri- veloped Countries: Some Lessons of Historical Experi- alising Countries: A Survey of Recent Trends', Economic ence', Journal of Development Studies. Special issue on and Political Weekly, December 6 and 13. Industrialisation. Landell-Mills, P., R. Agarwala, and Stanley Please (1989): Scitovsky, T. (1954): 'Two concepts of external economies', 'From Crisis to Sustainable Growth in Sub-Saharan Afri- Journal of Political Economy. April. ca'. Finance & Development, December. Singer, H.W. (1950): 'The Distribution of Gains Between Leibenstein, H. (1957): Economic Backwardness and Economic Investing and Borrowing Countries', American Economic Growth, (New York: John Wiley). Review, papers and Proceedings, May. Little, I., I. Scitovsky, and M. Scott (1970): Industry and Trade Sunkel, 0. (1973): 'Transnational Capitalism and National in Some Developing Countries: A Comparative Study, (Paris: Disintegration in Latin America', Social and Economic OECD). Studies, Vol. 22, No. 1. March. Little, I.M.D. and D. Mazumdar (1977): Research Proposal Sutdiffe, R.B. (1971): Industry and Underdevelopment (Lon- on Small scale enterprise development. mimeo, (The don: Addison-Wesley). World Bank). Thomas, C.Y. (1974): Dependence and Transformation (New Little, I.M.D. (1981): 'The Experience and Causes of Rapid York: Monthly Review Press). Labour-Intensive Development in Korea, Taiwan Prov- Thomas, Vinod and Ajay Chhibber(1989):'Experience with ince, Hong Kong and the Possibilities of Emulation' in E. Policy Reforms Under Adjustment Lending, Finance & Lee (ed.) Export-led Industrialisation and Development, Ge- Development, December. neva, Asian Employment Programme, International La- UNDP (1974): 'Sharing in Development: A Programme of bour Office. Employment, Equity and Growth for the Philippines', McCleary, William A. (1989): 'Policy Implementation un- Report of an intra-agency team financed by UNDP and der Adjustment Lending', Finance & Development. De- organized by the ILO. (Geneva: ILO). cember. United Nations (1951): Dept. of Economic Affairs, Meas- Mkandawire, T. (1988): 'The Road to Crisis, Adjustment ures of Development of Underdeveloped Countries. and De-Industrialisation: The African Case', African De- United Nations (Various years) Yearbook of National Ac- velopment, Vol. XIII, No. 1. counts. Merhav, M. (1969): Technological Dependence, Monopoly and Vaitos, C.V. (1974): Intercountry Income Distribution and Growth, (Oxford: Pergamon Press). TRansnational Enterprises (Oxford University Press). Myint, Hla (1958): 'The Classical Theory of International Wade, R. (1982): Irrigation and Agricultural Politics in South Trade and the Underdeveloped Countries', Economic Korea (Boulder, Colorado: Westview Press). Journal, June. pp. 317-337. Westphal, L. (1978): 'The Republic of Korea's Experience Nurkse, R. (1953): Problems of Capital Formation in Underde- with Export-led Industrial Development', World Devel- yelped Countries, (Oxford Blackwell). opment, Vol. 6, No. 3. March. Nurkse, R. (1959): Patterns of Trade and Development, (Wick- World Bank (1979) World Development Report 1979. (Wash- sell Lectures). ington, D.C.). Olofin, S. (1989): 'The Asian NICs Growth Model as an World Bank(1981): AcceleratedDevelopment inSub-Saha- Alternative Development Strategy in Africa'. Paper pre- ran Africa: An Agenda for Action (Washington, D.C.). sented at Project LINK conference held in Paris. August. World Bank and UNDP (1989): Africa's Adjustment and Oshima, H.T. (1962): 'A Strategy for Asian Development', Growth in the 1980s (IBRD: Washington, D.C.). Economic Development and Cultural Change, 10, no. 3. World Bank (1989): Sub-Saharan Africa: From Crisis to Sus- Prebisch, R. (1964): 'Towards a New Trade Policy of Devel- tainable Growth, A Long-Term Perspective Study (Washing- opment', in Proceedings of the United Nations Confer- ton, D.C.). ence on Trade and Development, vol. Il (New York: Zulu, J. and S.M. Nsouli (1989): 'Adjustment Programs in United Nations). Africa', Finance & Development, March. 326 31 Comments on Growth-Oriented Adjustments Hafez Ghanem and C Obidegwu Ha fez Ghanem The second sub-question concerns the private in- vestment equation in the model and has to do with The Ndulu Paper the lack of a profit variable. I understand the reduced form equation would include the real exchange rate The Ndulu paper argues that development econo- since the structural equation includes capacity utili- mists should pay more attention to capacity zation. But shouldn't we explicitly include in the utilization rather than focusing exclusively on the behavioral model some profit variable as affecting rate of investment. It emphasizes the importance of investment behavior, maybe the real exchange rate the foreign exchange constraint and the necessity of or the interest rate? maintaining a realistic real exchange rate. My second set of comments relates to the relation- The first comment I have on this paper concerns ship between investment and imports where the the difference between changing the level of output author postulates that the relationship can be ex- and changing the growth rate. We normally think of plained in terms of fixed coefficients. This could be a changing capacity utilization as a once-and-for-all very realistic assumption in the short term but change in the level of output and not necessarily as shouldn't we expect that over the long run the rela- changing the growth rate of the economy. In most tionship between imports and investments, or the models, we talk about investment as affecting the import content of investment, would be a function of growth rate of the economy. I think this is an issue relative prices? that is not sufficiently discussed in the paper. Athird commentconcernsthebehaviorof savings. The second comment concems the specification of I was quite puzzled by the author's result where he the private investment equation. I have two sub- says that savings constrained growth declines with a questions conceming that. The first is on the issue of real depreciation, since the improved export per- crowding-in and crowding-out or the relationship formance will reduce foreign savings, i.e. the current between govemment and private investment. In the account deficit will be smaller. The author does try to model specification, the author makes private invest- nuance his conclusion with talk of higher foreign ment a function of capacity utilization and a positive transfers as a result of a real depreciation and by function of government investment. Shouldn't this talking about an increase in the domestic value of really be an empirical issue? I assume, as is men- foreign savings. I'm puzzled because most models tioned in the paper, that the crowding-in or crowd- I've seen have an accounting identity which says that ing-out effect would differ from one country to an- domestic savings are the difference between exports other and would also differ according to the alloca- of goods and non-factor services and imports of tion of public investment. goods and non-factor services. So I don't see how any 327 Comments on Growth-Or'ented Adjustments increase in exports can lead to a decrease in savings. due to wider consumer choice and wider product I would have thought that foreign savings would fall availability can be substantial. We should not ignore as the deficit becomes smaller but this fall is fully them, however, it is rather disappointing that the compensated for by higher domestic savings. authors could not actually estimate those gains for Concerning the structure of the model, looking at the Tanzania case because of data constraints. the equations and diagram, I wondered about the There is an important link that needs to be stressed conditions needed to ensure that the equilibrium between price deregulation and the trade regime. In reached in this model which is dynamic is a stable the case of Tanzania (and I know nothing about equilibrium. I feel that a discussion of the stability or Tanzania except from reading this paper) price de- instability of the equilibrium reached and what con- regulation was accompanied by relaxing import con- ditions are needed to ensure stability would be war- trols through allowing for own imports. I think this ranted. relationship needs to be stressed since price deregu- Concerning the estimation techniques, my first lation, while maintaining quantitative restrictions question is why did the author take the average on imports, would probably have a much smaller values of thedifferentvariables foranumberof years welfare enhancing effect than price deregulation for each country so that he has one observation per with import liberalization. In fact, many govern- country rather than pool time-series and cross-sec- ments would argue that they can not liberalize prices tion data. I would have thought that by pooling he because of import controls which imply that certain would have had more efficientestimates. Aftergoing producers have monopoly power on the local mar- through the model and the different equations, I was ket. Once we link price deregulation and the trade struck that the author's estimating equations are not regime, we immediately link price deregulation with derived from the model. This raises issues of simul- overall consistent macro policies. I think this is an taneous equations bias in some instances and miss- issue that needs to be stressed. If you have quantita- ing variables bias in others. The author mentions that tive restrictions, exchange controls and other restric- in the paper, and also states that he used two stage tions to imports, then you need to get rid of those least squares rather than OLS and that the results prior to or simultaneous with price liberalization. In were not different. But the question in the reader's order to do that, you need "correct macro policies" mind is, since we have gone through this model, why and in particular, the real exchange rate has to be in don't we try to estimate the structure of the equations equilibrium or the trade liberalization will not be using some maximum likelihood estimator, or at sustainable. To get maximum benefits from price least use reduced form equations? I found the infla- deregulation, you must have import liberalization tion equation especially problematic since it's de- which will only be sustainable in the long run if rived from a money demand function where there's macro polices lead to a real exchange rate which is an no interest rates and no expectation variables for equilibrium exchange rate. future inflation. The Olofin Paper The Collier and Gunning Paper The Olofin paper states that an outward looking Concerning the Collier and Gunning paper, the industrialization strategy would not bevery success- reader's first reaction is to try to explain the preverse ful in Africa and would not do much better than an result they get from the theoretical model concerning inward looking industrialization strategy because of the relationship between producer prices and output several constraints. Therefore the author argues for a and the reaction of supply to changes in producer ROSH strateg-a rural oriented smallholder indus- prices. In many countries in Africa, when there are trialization strategy-with initial direct state partici- price controls there are shortages, but there is also a pation. The author argues that industrialization in black market. You can usually buy the goods, but at Africa, whether inward or outward oriented, faces prices much higher than the administratively deter- four major problems which are: mined fixed prices. My feeling is that if we include * absence of an entrepreneurial class; this phenomena in the theoretical model we will not * poor work ethic; get the preverse supply response that the authors are political unrest; and getting. * technological backwardness. I thought the most important point made by the paper concerned price deregulation. The paper Once he posed the problem that way, my response shows that the welfare gains from price liberalization was to wonder why the ROSH strategy would work. 328 Comments on Growth-Orented Adjustments Given those problems, it would seem that the appro- continue to exist under the new strategy? Namely, priate response would be to tackle them directly, wouldn't the new projects be managed with the goal mayibe throughbettereducation,betterincentives for of obtaining some political and social objectives at people to work, political reform and other incentives the expense of economic efficiency? Hence, would- for technological transfer. n't it be more useful to try pushing the private sector I don't fully understand why government as an from the beginning? Given the various problems entrepreneur would be more successful under the African governments face when trying to privatize proposed ROSH strategy than under the failed in- public enterprises, is it realistic to go ahead with a ward looking industrialization strategy. It is true, as strategy based on public ownership with the hope the author states, that maybe the types of projects that the projects would be eventually privatized? thatwouldbefinancedbythepublicsectorunderthe Wouldn't it be more realistic to go ahead from the ROSH strategy would be smaller and better adapted beginning with a strategy based on private owner- to the African environment, etc. But wouldn't the ship and private participation while the government problems that plague public enterprises in Africa only plays a supportive role? C. Obidegwu It is my pleasure to comment on these three interest- ants to produce. This improves supply response and ing and excellent papers. Two of them-the Ndulu the pace of adjustment. and the Collier and Gunning papers-have much in The third paper, by Olofin, is concerned with common. Both use a disequilibrium framework in strategies for development and industrialization for their analysis, and conclude that external financial Sub-Saharan African countries. It discusses the fail- assistance is an important element in the transition ure of the post-independence import substitution to growth and equilibrium. industrialization strategies and the prospects for, and The Collier and Gunning paper is a micro-eco- impact of, adopting the export oriented industriali- nomic analysis of peasant behavior in an economy zation strategy by Sub-Saharan African countries. faced with shortages of consumption goods as a Olofin rejects both the import substitution and the result of the lack of foreign exchange for imports of export oriented industrialization strategies for Sub- prod-uction inputs and consumer goods. The paper Saharan Africa. According to him, import substitu- analyzes the impact of this scarcity on peasant pro- tion industrialization strategyhasalreadyreached its duction decisions and examines the difficulties of limits in providing growth in industrial output and stimulating production in such circumstances, as employment and export oriented industrialization well as the transition to market clearing. The Ndulu strategy would not be appropriate since these coun- paper is a macroeconomic analysis of a prototypical tries do not possess the basic pre-conditions for in- Sub-Saharan Africa economy which faces con- dustrialization. Olofin therefore advocates an alter- strained imports and import-substitution possibili- native strategy -- the Rural Oriented Smallholder ties. External assistance plays a direct role in stimu- (ROSH) industrialization strategy. I shall now dis- lating production activity by raising the availability cuss each paper in turn. of imLported inputs and consequently, domestic ca- pacity utilization. The Collier and Gunning Paper The two papers highlight the critical role that ex- ternal finance can play in reviving an economy with As the authors readily admit, this paper is really fundamentally incorrect relative prices and few sub- two papers put together. One part is a theory of stitution possibilities for imported goods. In that peasant behavior under sustained disequilibrium sense, these papers lend support to the practice of (excess demand for goods) and repressed inflation. multilateral agencies and bilateral donors in packag- The other is an application of this theory to the Tan- ing policy advice and financial assistance. In the zanian economy during 1977 to 1988. microeconomic framework of Collier and Gunning, The theory makes some very restrictive assump- the financial support increases theavailability of con- tions about the supply and prices of market goods. sumer goods, which provides the incentive to peas- There are two goods-a cash crop, produced by peas- 329 Comments on Growth-OrientedAdjustments ants and purchased by a government agent, and a ning model, a decline in crop prices (equivalent to a consumption good, which is supplied by the state. fall in real wages) leads to a rise in labor supply; this The government fixes the price of both goods; there perverse result is due the use of a one-period model. is a shortage of the consumption good at the prevail- In this formulation, the only reason to work is to buy ing relative price, controls are effective and therefore whatever goods are expected to be available; the parallel markets do not develop. In effect, in this peasants adjust their labor supply in order to acquire economy, the government or its agent can be re- the goods. A decline of crops prices has no impact on garded as the producer of crops and the peasants as supply of goods or its probability distribution. In the the suppliers of labor, receiving a wage rate Pc (the multi-period formulation such as that elaborated by price of crops). The real wage would therefore be Barro and Grossman, households expect that excess Pc/Pg (where Pg is the price of goods). However, at demand for goods will prevail for a finite period and this wage rate and the initial money holding there is therefore have an incentive to save. Thus, the effec- excess demand for both crops (labor) and goods, tive labor supply curve remains positively sloping thus, the economy is operating in the zone of re- and is steeper than the notional labor supply curve. pressed inflation (see Cuddington et al 1984, page In the second part of the paper, the authors apply 26). their model to analyze the behavior of the Tanzanian With these assumptions, and using a variant of the economy during 1977 to 1988. This exercise makes disequilibrium model of Malinvaud, the authors de- heavy demands on data, and the authors had to make rive a number of policy-relevant conclusions for this heroic assumptions to produce the required data. For economy. instance, the computation of urban demand for grain In the presence of excess demand for goods and assumes constant urban per capita consumption of labor, the supply of labor is lower than it would be grain during the period. An elaborate technique is under market clearing conditions. Households with- used to calculate the "import" CPI for peasants; this hold their labor because they are unable to get the CPI is used to deflate nominal income which is used goods they want. forexpenditures for goods including those produced 2. Given the excess demand for goods, the supply in the peasant economy, for example, purchases of response of crops to price changes is perverse. Thus, grain by coffee farmers who receive their income lowering crop prices would raise the supply of crops mostly from outside the peasant economy. Neverthe- and reduce the national demand for goods. There- less, the authors should be congratulated for their fore, raising the price of crops is counterproductive patience in assembling all the required data. unless coordinated with the improved availability of The authors find that the decline in production of consumption goods. The authors regard this as of export crops between 1977 and 1984 was due to a very practical interest "since countries with predomi- shortage of consumer goods. They conclude that in nantly agricultural exports are often advised to raise times of shortage, crops sales are determined by the the producer prices of export crops." availability of consumer goods instead of the normal 3. In an economy characterized by repressed infla- relationship to prices. However, it seems to me that tion it is not possible to infer how crop prices differ the normal relationship to prices should not be ex- from their equilibrium merely from the symptom of pected when the system is out of equilibrium and excess demand for crops once the goods market is with 'prices' that are really fictitious. The goods' also in disequilibrium. It is very difficult to achieve prices do not reflect the transactions costs of acquir- equilibrium by official manipulation of the these ing these goods. For instance, consumers endure the prices, thus, sequenced price reform is heavily con- frustrations of not being able to obtain desired goods strained. and incur search costs which raise the price of the 4. The existence of monetary overhang under- goods beyond the store price. As shortages become mines the adjustment to the abolition of price con- more acute, these costs rise, the implicit real price of trols as it delays production response. cash crops (or the real wage) falls, and, ceteris paribus, 5. External aid facilitates the transition to market output would be reduced. The authors' estimate that clearing prices by providing for an injection of goods the real producer price of the export crops declined into the market economy. by 26 percent between 1977 and 1984 therefore con- These results are consistent, except in one respect, siderably understates the real decline when the frus- with the analysis of the repressed inflation zone (ex- tration factor is considered. Unfortunately this factor cess demand for labor and goods) elaborated by is not easily quantified. Barro and Grossman (1976) and Cuddington et al The authors found that during 1984 to 1988 (the (1984). The exception is that in the Collier and Gun- reform period), there was a large increase in crop 330 Comments on Growth-Oriented Adjustments sales despite a decline of real prices. They assert that has implications for the pace of price reformn. It would a combination of the return to market clearing and indicate that rapid decontrol is preferable to an alter- nominal price increases explain the paradox that native pace in which the govemment attempts, over output rose while producer prices fell in real terms. a period of time, to provide price incentives while In my view, there is not necessarily a paradox. Rising decontrolling prices and distribution at a slow pace. output could have been due to good weather, im- This has implications for exchange rate adjustments proved availability of inputs and transport, and where the emphasis has been on upward adjust- lower costs of doing business due to some liberaliza- ments of the nominal exchange rate bya government tion of prices and distribution; these factors would agency, leaving the process of determination of the tend to shift the supply curve to the right. The agri- rate unchanged. cultural supply elasticity with respect to non-price factors has been found to be higher than the supply The Ndulu Paper elasticity with respect to prices, especially in poor countries with poor facilities and markets (see Ch- The Ndulu paper is concemed with adjustment hibber 1988). While realized pricesare generallyused and growth of Sub-Saharan African economies in time series econometric tests, the appropriate price which face constrained imports, and low capacity variable in production decisions is expected prices; utilization and growth. The paper notes that import farmers decisions are likely to be based on long-term compression affects not only current production but price trends. There is no evidence to suggest that the also the growth of investment, making the allocation decline in real prices constituted a long-term trend. of imports between capacity expansion and utiliza- In any case, the decline in real prices of 20 percent tion a key policy issue. The author suggests that the over four years is probably exaggerated by the use of centrality of improvement in export performance the peasants "import' price index, which excludes and net foreign resource inflows in the growth recov- food prices. ery process cannot be overemphasized. He further The discussion in the paper of the evolution of real notes that growth in Sub-Saharan Africa is substan- monetary balances seems quite complex. The tiallyinfluencedbyanumberof exogenous factors- authorsattributetheholdingofmoneytotheprecau- the changes in the barter terms of trade, external tionary motive - the need for peasants, faced with assistance, and the weather. For instance, in agricul- random shortages of goods, to be prepared to avail ture, growth fluctuations are mainly due to the vaga- themselves of those occasions when availability was ries of the weather. In another section of the paper, atypically good. However, involuntary holding of the author discusses the role of government in the money, as a result of the excess demand for goods economy, particularly in public investments and di- and the fluctuations in the supply, is also a plausible rect controls of financial resources, especially foreign explanation for the evolution of moneybalances. exchange. He discusses evidence indicating that This paper has a number of important lessons for public investments, particularly in infrastructure, development policy. One lesson is that widespread have a net crowding-in effect in several countries. use of price and distribution controls could easily The central adjustment and growth issues raised lead to, severe policy distortions, undermining the by the paper relate to the growth implications of use of prices as policy instruments. It is clear from import compression and the related problem of ra- this analysis that quantities rather than prices be- tioning of imports, export growth and diversifica- come more potent instruments, but in general, the tion, the role of govemment in the economy, and the manipulation of quantities is a far more difficult vulnerability of the economies to exogenous shocks. undertaking than prices. Another lesson is that con- The author specifies a macroeconomic accounting trols could easily get the economy into a quagmire framework, presumably to analyze these issues. from which it would be very difficult to extricate While I find the model interesting, the links between itself. According to the authors, external aid can play the policy issues and the accounting framework do a key role in the transition to market clearing condi- not seem to be particularly strong. The author should tions. In the Tanzanian situation, the abandonment make clear at the onset the kinds of policy issues the of widespread price controls and confinement, and model is designed to elaborate. the legalization of own-funded imports, as well as I have some problems with the specification of increased flows of official external assistance to fi- some of the relationships in the model. For instance, nance general imports, have been important ele- since an importconstrained economyis in disequilib- ments in the economic recovery. The authors' finding rium, the effective imports should not be determined that sequenced price reform is fraught with mistakes, by desired investment and capacity utilization. 331 Comments on Growth-OrientedAdjustments Rather, imports should be determined by the finan- lates to some characteristics common to Sub-Saharan cial constraints; investments and capacity utilization African economies, which the author regards as be- would depend on imports, using some explicit im- ing inimical to an export oriented industrialization. port allocation criteria. In the approach used in the The paper outlines two main interpretations of the model, the author sidesteps the issue of the allocation G4 industrialization miracle. One is that the coun- of imports which he deemed important. The model tries adopted the right policies to get factor prices has a simple supply function for exports which does right, relied on market forces and were integrated not provide much guidance on the issues of export into the global economy through liberal trade poli- growth and diversification. One key adjustment pol- cies. The other strand of argument is that the success icy issue relates to the factors that influence the re- of the G4 countries was due to cyclical and historical sponse parameter. In an import constrained econ- national and international factors which produced omy, import capacity would also tend to constrain favorable access to the markets of the advanced export production and diversification. For instance, countries, increased their access to international fi- the lack of imported intermediate inputs to work the nance, and encouraged the relocation of production mines reduces the exports of mineral dependent within their boarders by transnational corporations. economies and fertilizer shortages undermine agri- The interpretations of the success of the G4 coun- cultural production and diversification which di- tries are numerous. One participant in the debate rectly or indirectly reduce agricultural exports. The (Amsden 1989) has even suggested that Korea suc- paper implies that there is a behavioral specification ceeded by violating the cannons of economic wis- for foreign saving, when indeed the equation is sim- dom-by getting prices wrong. Some others, such as ply an identity. The impact of depreciation of the (Nam 1988) and Rhee et al (1984) have attributed the currency on foreign savings is not unambiguous as success to, among other factors, the Confucian heri- the author claims. tage, a belief system which places high value on hard According to the author, the full model has been work, thrift, loyalty, discipline and respect for calibrated for Tanzania and partial models have been authority. But Rhee et al have pointed out that the calibrated for a number of other countries. It would same Confucian heritage could be blamed for inhib- have been useful to present some simulation results iting economic development in Korea before the from the Tanzanian model. The rest of the paper is 1%0s. However, under compatible policies, these devoted to econometric estimates of two of the rela- human traits became a positive force for develop- tionships postulated in the model, and an evaluation ment. On the role of market forces Bradford (1986) of the progress of Sub-Saharan African countries in points out that what "seems to distinguish the East structural adjustment. The policy implications of the Asian development experience is not the dominance regression results are not clear, as they do not seem of market forces, free enterprise and internal liberali- to throw any light on the major policy issues raised zation, but effective, highly interactive relationships at the beginning of the paper. Finally, I feel that the between the public and private sectors characterized paper is rather ambifious in trying to tackle a wide by shared goals and commitments embodied in the variety of issues simultaneously. It could benefit development strategy and economic policy of the from a sharp focus on a few issues for which it could government." provide in-depth analysis. In spite of the plurality of interpretations, the fac- tors to which the success is usually attributed are, in The Olofin Paper general, not mutually exclusive. A favorable external environment helped the G-4 countries as well as The Olofin paper argues that the export oriented Brazil, Mexico, Turkey, etc. to achieve rapid industri- industrialization strategy, which implicitly under- alization. But when the global economy soured in the pins the current adjustment programs in Sub-Saha- 1970s, the G-4 countries were able to sustain rapid ran Africa, is inappropriate for these countries. The growth while the others faltered. The one distin- author therefore recommends instead, the Rural Ori- guishing factor between the G4 countries and the ented Smallholder (ROSH) industrialization strat- others is the relentless pursuit of export growth pro- egy. moted by credible export incentives, which were The author has two main lines of argument for provided by a mixture of proven, effective measures. rejecting export oriented industricalization strategy. Export promotion is, of course, the central feature of The first relates to his interpretation of the success of outward orientation. If the success of the G-4 coun- Taiwan, Korea, Singapore and Hong-Kong (Asian tries is regarded as exceptional, how about the recent G4 countries) in industrialization. The second re- successes of Mauritius, and the three ASEAN coun- 332 Comments on Growth-Oriented Adjustments tries-Thailand, Malaysia and Indonesia? These are ity as long as they are not part of a superstructure industrializing rapidly using export oriented indus- such as thegovernment. There are manyother details trialization strategy - combining export promotion, of the ROSH strategy that I have problems with, liberal trade policies with reliance on the private however. The paper advocates participation of the sector, and markets. Their success is perhaps more state in the direct development of the small-scale instructive for African countries than that of the industries, at least at the initial stages. It is unlikely Asian G-4 since these ASEAN countries share many that the state, which has proved incapable of manag- more characteristics with Sub-Saharan African coun- ing a relatively limited number of firms, will be suc- tries. cessful at coordinating the activities and managing The author outlines four characteristics which, in its investment in a multitude of diverse small firms. his view, limit the potential for industrialization in Besides, a transition from state ownership to the Sub-Saharan Africa-absence of an entrepreneurial private sector may prove very difficult to effect as the class, poor work ethic and low productivity of labor, current experience with privatization indicates. For acrimonious ethnic rivalry and lack of national cohe- successful industrialization in Sub-Saharan Africa, sion and identity, and technological backwardness. It production, whether in small or large firms, will have is doubtful that these factors are so endemic that they to bein the domain of theprivate sector, with thestate cannot respond to policy change, investment and providing social and physical infrastructure. The de- economic, institutional and political reform. It is not velopment and operation of small firms in particular clear also that these constraints will be less debilitat- entail risks that demand flexibility and quick re- ing to development under ROSH, or that ROSH will sponse; these are not the hallmarks of the bureau- be more successful in ameliorating them. It is not cratic environment in government. The paper advo- obvious to me that reliance on the profit motive is cates protection of small firms from competition "not likely to succeed' in spurring industrialization. from large- and medium-scale plants and, I assume, Irrespective of the development strategy, incentives from imports. However, such differential protection, are needed for individuals and firms tobeproductive taxation and regulation are difficult to manage effi- and these agents will only undertake economic ac- ciently, and protection will remove an incentive for tivities if there is an expectation of profit. the firms to be efficient. It is not clear that the ROSH There are many attractive features of ROSH. The strategy will necessarily minimize foreign exchange proposals to emphasize growth in labor productivity, constraints on growth. Small firms also make de- technology, entrepreneurial ability, labor intensive mands for imported capital and intermediate inputs; production techniques in smallholder units, and indeed some use of imported capital and inputs will government investments in social and physical infra- be necessary for the technological advances that structure are all excellent. These proposals, however, stimulate rapid industrialization. Experience shows are not incompatible with structural adjustment pro- that the only way to minimize foreign exchange crisis grams or export oriented development strategies. issustainedexportgrowth,notavoidanceofimports. Many structural adjustment and export develop- The paper alludes to the need to develop agricul- ment programs emphasize the development of ture and industry in such a way that they comple- small-scale industries, the use of labor intensive pro- ment each other. But it does not, in my view, vigor- duction and capacity building. In fact, the develop- ously emphasize the point that sustained rapid ment of small-scale industries underpinned the rapid growth of agriculture is a prerequisite to industriali- industrialization of Japan and the other successful zation, particularly a rural-based one. Structural ad- Asian countries. In contrast, India, which is sup- justment measures programs in Sub-Saharan Africa posed to have had an explicit smallholder strategy, have focussed on reviving agricultural growth by was less successful because the policies and regula- providing incentives for production and improving tory environment under the ruling import substitu- the access of farmers to the inputs that enhance their tion strategy inhibited the growth of small and large productivity. A rapid growingand profitable agricul- firms alike. tural sector will lay a strong foundation for industri- I agree with the author about the need to end the alization by meeting the basic food needs of the fascination with big projects, large-scale industry population, providing rural dwellers the resources and capital intensive industrialization. I am in sup- they need to invest in rural industry, and generating port of the emphasis on small-scale industry, al- the labor to work in such activities. though I do not see why it should be limited to rural It seems to me that the isolationist ROSH strategy areas. It is reasonable to expect that small-scale firms advocated by the author is a version of the import will make less demands on scarce managerial capac- substitution strategy, but with a bias against large- 333 Comments on Growh-Orented Adjustments scale urban production. The author realizes that Colinl.Bradford,Jr.(1986)."EastAsianModels: Mythsand there are severe obstacles to instituting and imple- Lessons", Development Strategies Considered, John P. Le- menting this isolationist strategy. Even if it is the case wis, and Valeriana Kallab (eds.). Transaction Books. that Africa is being increasingly marginalized in the Chhibber, Ajay (1988). "Aggregate Supply Response in Ag- world as some writers suggest, it still will be very riculture: A Survey", Structural Adjustment in Agricul- difficult for any African country to isolate itself from ture Theory and Practice, S. Commander (ed.). James the rest of the world. There are strong vested inter- Curry Publishers. ests, both inside and outside Africa, to maintain and Cuddington, John T., Per-Olov Johansson, Karl-Gustaf strengthen the links that exist. The best strategy for L?fgren (1984). Disequilibrium Macroeconomics in Open African countries is one that brings them the maxi- Economics, Basil Blackwell. mum benefits from being part of the global family. Sang-Woo Nam (1988). 'Altemative Growth and Adjust- ment Strategies in Newly Industrializing Countries in References Southeast Asia7 Beyond Adjustment: The Asian Experi- ence, Paul Streeten (ed.). International Monetary Fund. Alice H. Amsden (1989). Asia's Next Giant: South Korea and Yung Whee Rhee, Bruce Ross-Larson, and Gary Pursell Late Industrialization: Oxford University Press. (1984). Korea's Competitive Edge: Managing the Entry into Robert Barro and Herschel Grossman (1976). Money, Em- World Markets, John Hopkins University Press. ployment and Inflation, Cambridge University Press. 334 The World Bank Most countries of Sub-Saharan Africa adopted programs of economic reform in the 1980s. The basic principles and methods of implementation were well understood for reforms intended to reduce budget and balance of payments deficits. But other crucial aspects of the reform pro- grams-particularly measures to improve efficiency or restore growth-were less clearly known and agreed upon. When most of these programs were being implemented, little research had been done on how to apply general economic principles to specific reforms workable in Africa. This collection, by both African and non-African academics, policymakers, and development practitioners, reports on research that has been undertaken in the past decade to evaluate the effectiveness of these reform efforts. It draws together current findings and policy recommenda- tions on eight broad topics affecting African economic reform: exchange rate policy, parallel markets, fiscal deficits and expenditure policy, financial policy, trade policy, regional integration, human capital and entrepreneurship, and growth-oriented adjustment. An introduction highlights the principal observations of the twenty-two papers and eight sets of commentary, which were originally presented at a conference in Nairobi in June 1990. At the time this book was in preparation, Ajay Chhibber was principal economist and Stanley Fischer was vice president, development economics, in the World Bank's Development Economics Department. WORLD BANK PUBLICATIONS OF RELATED INTEREST African Development Indicators Africa's Public Enterprise Sector and Evidence of Reforms Daniel Swanson and Teferra Wolde-Semait Financial Systems and Development in Africa Edited by Philippe Callier Financing Adjustment with Growth in Sub-Saharan Africa, 1986-90 Long-Term Perspective Study of Sub-Saharan Africa The Making of Economic Policy in Africa Ravi Gulhati Poverty-Conscious Restructuring of Public Expenditure Marco Ferroni and Ravi Kanbur Social Dimensions of Adjustment in Africa: A Policy Agenda Structural Adjustment in Sub-Saharan Africa Cadman Atta Mills ISBN 0-8213-2063-7