Exchange controls and the parallel market economies in sub-Saharan Africa

SWP711 Exchange Controls and Parallel Market Economies in Sub-Saharan Africa Focus on Ghana Ernesto May WORLD BANK STAFF WORKING PAPERS Number 11   WORLD BANK STAFF WNORKING PAPERS Number 711 Exchange Controls and Parallel Market Economies in Sub-Sahairan Africa Focus on Ghana Ernesto May The World Bank Washington, D.C., U.S.A.  Copyright (C 1985 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 January 1985 This is a working document published informally by the World Bank. To present the results of research with the least possible delay, the typescript has not been prepared in accordance with the procedures appropriate to formal printed texts, and the World Bank accepts no responsibility for errors. The publication is supplied at a token charge to defray part of the cost of manufacture and distribution. The World Bank does not accept responsibility for the views expressed herein, which are those of the authors and should not be attributed to the World Bank or to its affiliated organizations. The findings, interpretations, and conclusions are the results of research supported by the Bank; they do not necessarily represent official policy of the Bank. The designations employed, the presentation of material, and any maps used in this document are solely for the convenience of the reader and do not imply the expression of any opinion whatsoever on the part of the World Bank or its affiliates concerning the legal status of any country, territory, city, area, or of its authorities, or concerning the delimitation of its boundaries, or national affiliation. The full range of World Bank publications, both free and for sale, is described in the Catalog of Publications; the continuing research program is outlined in Abstracts of Current Studies. Both booklets are updated annually; the most recent edition of each is available without charge from the Publications Sales Unit, Department T, The World Bank, 1818 H Street, N.W., Washington, D.C. 20433, U.S.A., or from the European Office of the Bank, 66 avenue d'lena, 75116 Paris, France. Ernesto May wrote this paper while in the Country Policy Department as a member of the World Bank's Young Professionals Program. He is now an economist in the East Asia and Pacific Regional Office of the Bank. Library of Congress Cataloging in Publication Data May, Ernesto, 1955- Exchange controls and parallel market economics in sub-Saharan Africa. (World Bank staff working papers ; no. 711) Bibliography: p. Includes index. 1. Foreign exchange administration--Africa, Sub-Saharan. 2. Black market--Africa, Sub-Saharan. 3. Foreign exchange admini.stration--Ghana. 4. Black market--Ghana. I. Title. II. Series. III. Title: Parallel market economics in sub-Saharan Africa. HG3982.M39 1985 332.4'564'09667 84-27012 ISBN 0-8213-0471-2  ABSTRACT This paper provides a theoreticael framework to understand the way in which exchange controls modify the behavior of the different agents in the economy, leading to the creation of a parallel economy. It gives the neces- sary theoretical elements to analyze this parallel economy and provides a simple methodology to obtain relevant quarLtitative information about it. Finally, it elaborates some of the policy implications of the existence of such an economy. The model developed shows that parallel market activities can be explained through the optimizing behavior of exporters and importers. Exporters will keep shifting their exports from official to parallel market channels until the expected marginal benefit in both activities is the same. Import traders will devote resources to rent-seeking activity until the moment in which an additional unit spent in this activity is equal to the savings derived from being able to use official channels instead of the parallel markets. Their combined behavior determines the amount of import and export smuggling, the level of the rent-seeking activity, and the black market exchange rate that is consistent with an equilibrium position where no one has any more incentives to move from their attained position. A methodology has been developed to detect the presence and assess the magnitude of the parallel market economy, as well as to explain its behavior in a quantitative fashion. This methoodology is applied to Ghana, a country in Sub-Saharan Africa, where parallel market activities seem to be widespread and relatively open in the economy. In the first part, a reduced form equation of the black market exchange rate that is derived from the context of our model is estimated for Ghana. There is a definite negative relationship between the real official exchange rate and the black market rate in this case. By letting the real official exchange rate appreciate, the Government of Ghana has been losing an important amount of foreign exchange related to the exports that are now being smuggled out of Ghana instead of going through the official channels. This has meant an important reduction in the amount of foreign exchange allocated to imports and, therefore, an increase in the demand for smuggled imports in the economy. In all, the real appreciation of the official exchange rate seems to be very much related to the increasing importance of parallel market activities in the economy. The sescond part of the methodology focuses on the estimation of the size of the parallel market economy and its evolution in time. The approach emphasizes the fact that there is a demand for currency related to the activities of the parallel market economy. It is based on the methodology developed by Tanzi (1982) to estimate the underground economy in the United States.  The yearly estimates of the parallel market economy seem to indicate a very clear trend: the government has been losing control over the economy as more and more transactions are being diverted to the parallel markets. The parallel market economy rose almost steadily to a 32.4 percent of the official GDP in 1982. ACKNOWLEDGEMENTS I am grateful to Takamasa Akiyama, Yaw Ansu, Kathie Krumm, Robert Myers, Jeffrey Hammer, Homi Kharas, Michael Michaely, Demetrios Papageorgiou, Marcelo Selowsky, and my colleagues in the Country Policy Department for their valuable comments, encouragement, and support during this project. Janet Entwistle and Marinela Dado provided excellent research assistance. I retain full responsibility for the remaining errors in the paper.  CONDENSE Cette 6tude propose un cadre th6orique destine a faire mieux comprendre la facon dont le contr6le des changes modifie le comportement des differents agents de l'economie, aboutissant a la creation d'une economie parallele. Elle fournit les 6lments th6oriques necessaires pour analyser cettes economie parallele en meme temps qu'une methode simple pour obtenir les donn6es quantitatives pertinentes a son sujet. Enfin, elle examine certaines des implications de l'existence d'une telle 6conomie sur le plan des politiques. Le modele 6labore montre que les activites du marche parallele peuvent s'expliquer par la volonte des exportateurs et des importateurs d'optimiser leur situation. Les exportateurs s'efforceront de faire passer leurs exportations du march6 officiel au marche parallele tant que les avantages marginaux attendus des deux marches ne seront pas les memes. Les importateurs consacreront leurs ressources a des activites susceptibles de produire un revenu jusqu'a ce que l'unite supplementaire depensee a de telles activites soit egale a l'6conomie qui resulterait de la possibilit6 d'utiliser les voies officielles au lieu des march6s paralleles. Ensemble, par leur comportement, les uns et les autres determinent Le volume d'importations et d'exportations pass6es en contrebande, le niveau d'activit6 a but lucratif et le taux de change du marche noir compatible avec une position d'6quilibre oui personne n'a plus de raison de vouloir bouger de sa position. Une methode a ete mise au point pour deceler la presence et 6valuer l'ampleur de l1'conomie du marche parallele, ainsi que pour expliquer son comportement de facon quantitative. Cette m6thode est appliqu6e au Ghana, pays de l'Afrique au sud du Sahara oii les activites du march6 parallele semblent florissantes et semblent se pratiquer d'une maniere relativement ouverte dans l'6conomie. Dans la premiere partie est 6tablie une estimation pour le Ghana du taux de change sur le marche noir a partir d'une 6quation sous forme r6duite decoulant de notre modele. I1 existe dans ce cas une relation negative entre le taux de change officiel reel et le taux du marche noir. En laissant le taux de change officiel r6el remonter, le Gouvernement du Ghana a perdu un volume appr6ciable de devises du fait de la sortie en contrebande d'exportations qui auraient da normalement emprunter les voies officielles. Cela a entrain6 une forte reduction du volume de devises allou6 aux importations et, par consequent, un accroissement de la demande de produits import6s en contrebande. En somme, la remontee en valeur reelle du taux de change officiel semble etroitement li6e a l'importance accrue des activit6s du march6 parallele pour l'economie.  Le second volet de cette methode porte sur l'estimation de I'ampleur de 1'economie du marche parallele et sur son 6volution dans le temps. L'etude souligne le fait qu'il existe une demande de devises li6e aux activites de cette 6conomie. Elle se fonde sur la methode mise au point par Tanzi (1982) pour estimer l'6conomie clandestine aux Etats-Unis. Les estimations annuelles de l'economie du marche parallele semblent indiquer une tendance tres claire : l'Etat perd le contr6le de l'6conomie a mesure qu'un nombre croissant de transactions se trouvent detournees vers les marches paralleles. En 1982, 1'economie du marche parallele, en hausse presque constante, a atteint 32,4 % du PIB officiel.  EXTRACTO Este estudio proporciona un marco te6rico para comprender la manera en la que los controles cambiarios modifican el comportamiento de los diverso,s agentes que actuan en la economia y conducen a la creaci6n de una economia paralela. Brinda los elementos te6ricos necesarios para ana- lizar esta economia paralela, asi como una metodologia simple para obtener la informaci6n cuantitativa pertinente acerca de la misma. Finalmente, explica en detalle algunos de los efectos en materia de politicas deriva- dos de la existencia de tal economia paralela. El modelo elaborado muestra que las actividades del mercado paralelo pueden explicarse mediante el. comportamiento de los exportadores e importadores tendiente a la optimizacion. Los exportadores persistiran en trasladar sus exportaciones de los cauces del mercado oficial a los del mercado paralelo hasta que el benefici4o marginal esperado en ambas activi- dades sea el mismo. Los importadores dedicarAn recursos a actividades de busqueda de renta hasta el momento en que una unidad adicional gastada en tales actividades sea igual al ahorro derivado de la posibilidad de usar cauces oficiales en vez de los del mercado paralelo. Su comportamiento combinado determina la cantidad de contrabando de importaciones y exporta- ciones, el nivel de actividades de bdsqueda de renta y el tipo de cambio de mercado negro que guarde armonia con una posici6n de equilibrio en la que nadie tenga mas incentivos para salir de la situaci6n alcanzada. Se ha elaborado una metodologia para detectar la existencia de la economia del mercado paralelo y evaluar su magnitud, asi como para explicar su comportamiento de una manera cuantitativa. Esta metodologia se ha aplicado a Ghana, pais de Africa al Sur del Sahara donde las activi- dades del mercado paralelo parecen muy difundidas y relativamente abiertas en la economia. En la primera parte del estudio se calcula con respecto a Ghana una ecuaci6n de formato reducido del tipo de cambio del mercado negro que se desprende del contexto de nuestro modelo. Hay en este caso una rela- ci6n negativa definida entre el tipo de cambio oficial real y el del mer- cado negro. Al dejar que se valorice el primero de los dos, el Gobierno de Ghana ha venido perdiendo una suma apreciable de divisas en relaci6n con las exportaciones que ahora se realizan clandestinamente, en vez de a tra- v6s de los cauces oficiales. Esto ha significado una reducci6n considera- ble de la cantidad de divisas asignadas a las importaciones y, por lo tanto, un aumento de la demanda de los productos importados clandestina- mente. En conjunto, la valorizaci6n real del tipo de cambio oficial parece estar muy vinculada a la importancia creciente de las actividades del mercado paralelo dentro de la economia.  La segunda parte de la metodologia se centra en el calculo de la magnitud de la economia del mercado paralelo y su evoluci6n en el curso del tiempo. El metodo destaca el hecho de que existe una demanda de moneda vinculada a las actividades de la economia del mercado paralelo. Se basa en la metodologia elaborada por Tanzi (1982) para calcular la eco- nomia sumergida en los Estados Unidos. Las estimaciones anuales de la economia del mercado paralelo parecen indicar una tendencia claramente definida: el Gobierno ha venido perdiendo el control sobre la economia a medida que se ha desviado hacia el mercado paralelo un nulmero cada vez mayor de transacciones. La econo- mia del mercado paralelo se elev6 en forma casi sostenida a un 32,4% del PIB oficial en 1982.  TABLE OF CONTENTS Page I. Introduct:ion ............................................ * .............. I II. A Historical Perspective of Exchange Controls in Sub-Saharan Africa ................... ......... .. ................ 9 III. Exchange Controls: The Theoretical Framework*................... 19 A. Exchange Controls as a Balance of Payments Policy .......... . 19 B. Analysis of Exchange-Control Regimes ....................... . 26 1. The Model ..................................................... 31 a. The Importer's OptimizaLtion Problem ................. 32 b. The Exporter's Optimization Problem ... 39 c. Black Market for Foreign Exchange . .......... 42 2. Comparative Statics ................... .. . . ..... .. 44 a. Depreciation of the Official Exchange Rate .......... . ....... 44 b. Increase in the Level of Government Policing Activity ................. ................. . 51 c. Exogenous Increase in Exports ...................... 55 d. Exogenous Decrease in Imports ... ................... 58 IV. Exchange Controls and the Parallel Market Economy: The Case of Ghaa ..60 A. Black Market Exchange Rate Determination .............ra ......t 62 B. The Size of the Parallel Market Economy .. 65 1. Cocoa Smuggling and the Amount of Domestic Currency Traded in the Black Market ..................... 67 a. The Econometric Model,, ........................ . ..... . 70 b. Estimation of the Cocoa Supply Equation ............. 75 2. The Size of the Parallel Market Economy ................. 81 V. Conclusions: Some Policy Implications of the Existence of a Parallel Market Economy.* .. .... o .......................... 92 Appendix A: Exchange Control Policy in Sub-Saharan Africa .96 Appendix B: Data ....*........................ ....................... 127  LIST OF TABLES Page Table 1: Exchange Rate Regimes in Sub-Saharan Africa: Three Analytical Groups. ................................ 10 Table 2: Real Effective Exchange Rate Indices, 1972-1982 . . 13 Table 3: Nominal Effective Exchange Rate Indices, 1972-1982 .. 14 Table 4: The Black Market Exchange Rate, Ghana: 1972-1982 . . 64 Table 5: Cocoa Smuggling and Production, Ghana: 1960-1979 ......... 69 Table 6: Cocoa Supply Equation for Brong-Ahafo Region: 1960-1981 ................. 77 Table 7: Cocoa Supply Equation for Volta Region: 1960-1981 .. . 77 Table 8: Estimated Ghanaian Cocoa Smuggled to Ivory Coast: 1960-1982 ........79 Table 9: Demand for Currency Relative to Total Money, Ghana: 1965-1982 . ....... .... 85 Table 10: Actual and Predicted Values of Currency Holdings, Ghana: 1965-1982 ........... 87 Table 11: Estimates of the Parallel Market Economy, Ghana: 1965-1982 ......... 89 Table Al: Exchange Control Policy in Sub-Saharan Africa, 1982 ...... 105 Table Bi: Regression Data for the Black Market Rate Equation, 1972-1982 ........................................................ 127 Table B2: Regression Data for the Black Market Rate Equation, Instruments. 1972-1982 ...... . .... .... ................. . . 128 Table B3: Black Market Exchange Rates and Cocoa Prices: Ghana, Ivory Coast, and Togo. 1960-1981 .................. 129 Table B4: Regression Data for Cocoa Supply Equations, 1960-1981 .... 130 Table B5: Regression Data for Currency-Demand Equation, 1965-1982 ..................................... .. .... ..... 131  LIST OF FIGURES Page Figure 1: Real Effective Exchange Rates, 1972-1982 ................. 15 Figure 2: Nominal Effective Exchange Rates, 1972-1982 .............. 16 Figure 3: Foreign-Exchange Market .................................. 20 Figure 4: Imposition of an Import Quota ............................ 27 Figure 5: Depreciation of the Official Exchange Rate ............... 48 Figure 6: Depreciation of the Offici.al Exchange Rate with Reserves Target ... . .. ............................... 50 Figure 7: Increase in Policing Activity, Case 1 .................... 53 Figure 8: Increase in Policing Activity, Case 2 .................... 54 Figure 9: Exogenous Increase in Exports ............................ 56 Figure 10: Exogenous Increase in Exports Having a Reserves' Target ........ 57 Figure 11: Exogenous Decrease in Imports ............................ 59 Figure 12: The Size of the Parallel Market Economy, Ghana ........... 91   I. INTRODUCTION The fact that economists have for so long regarded exchange controls as "obviously nonoptimal from the resource allocation viewpoint" 1/ has resulted in a very small amount of effort, in terms of economic theory, to explain the behavior of economies where exchange controls are a "fact of life." An increasing number of economists have recognized that the standard models developed by the economic theory in the field of open macroeconomics, which has been mainly concerned with cornvertible currencies, may not be applicable to an economy in which there is rationing in the official foreign exchange market. 2/ To a great extent, this has to do with the emergence of a parallel marlket economy that is related to the imposition of exchange con- trols. It is the purpose of this paper to take the first steps in developing a unified analytical framework to study the behavior of economies with exchange controls, by providing a framework to analyze parallel market activities (i.e., smuggling of imports and exports, rent-seeking, and trading of currencies in the black market). In a first approximation, there is no difference in the economic analysis of a direct quantitative restriction on imports, or an import restriction through the use of exchange controls. The use of the more elaborate administrative procedure of exchange controls can only be justified to the extent to which governments find that foreign trade is more easily and effectively controlled through the foreign exchange market. Traders need to change from foreign to domestic money (or vice versa). 1/ A. Krueger (1983), p. 175. 2/ See, for example, Krueger (1983).  -2- This similarity falls apart the moment we consider illegal transac- tions. A direct import quota may create incentives to smuggle imports as long as the import-related premium outweighs the risks ensuing from the illegal transactions. If the quota is imposed through exchange controls, then the foreign exchange needed to smuggle goods into the country needs to be purchased through the black market. In this case there is an incentive to smuggle imports as long as the import premium outweighs the black market premium together with the risks involved in illegal transactions. As has been pointed out by the literature, direct import quotas can only create incentives for smuggling but do not give incentives for the creation of a black market in foreign exchange. 3/ In the case of exchange controls it is not very meaningful to pose the question of import smuggling and its effects on welfare without exploring the means of its finance. This possible black market aspect of smuggling is not captured by the existing literature on smuggling in international trade. The reason is that the analysis of smuggling has always been carried out in a framework of trade restrictions (quotas, tariffs) and not of exchange controls. But in the former, black markets of foreign exchange are never an issue. There is another phenomenon related to quantitative restrictions on trade that has not been related either to smuggling or black markets: rent- seeking. 4/ Anne Krueger (1974) shows how quantitative restrictions, which 3/ M. Sheikh (1976), pp. 9-10. 4/ Defined as the activity undertaken to benefit from rents or revenues created by trade restrictions. This may involve the use of resources such as manpower to carry out all the administrative procedures to obtain an import license, bribes, increase of installed capacity if this is used as an indicator by the authorities to allocate licenses, etc.  -3- carry premiums, attract the use of resources in order to earn them. This use of real resources entails a greater loss of welfare than a Bhagwati equivalent tariff would. Krueger avoids in her analysis the direct relation between the activities of smuggling and rent-seeking. With trade restrictions and no exchange controls we can think of a "constrained importer" as having two options to acquire its desired imports: (i) use resources to smuggle the goods into the country and (ii) use resources to get hold of the import license. An increase in the amount of smuggled goods decreases the import premium of a given license and therefore diminishes the incentives for rent- seeking. To understand rent-seeking activities it is important to relate them to smuggling activities. Furthermore, the moment we enter into a world of exchange controls we have to face the interrelation between smuggling, rent- seeking and black markets. It seems then that the theory of international trade has yet to develop a single unifying analytical framework to study parallel market activities. There has been an extensive amount of related work done in the literature, but there are still some important loose ends. We now try to tie up some of these loose ends by developing a model of exchange controls. It brings together, through an analytical framework, the phenomena of smuggling, rent-seeking, and black markets which have been so far analyzed separately in the literature. By relating the official exchange rate, the black market rate, and the import premium through the: activities of smuggling and rent- seeking, the model helps to understand the way in which exchange controls affect the behavior of the different agents in the economy, leading to the creation of a parallel market economy.  -4- The remainder of this introductory section serves as a reader's guide by presenting the outline of the paper and reviewing its major findings with an indication where these are located in subsequent sections. The next section gives an historical perspective of exchange con- trols in Sub-Saharan Africa, the region that is emphasized throughout the paper. It analyzes the different exchange rate regimes in Sub-Saharan Africa trying to establish their common characteristics as well as their marked differences. For countries where the common pattern of response to foreign exchange scarcity has been to rely on exchange controls, a detailed summary of their exchange control measures is presented in Appendix A. Section III presents the theoretical framework to analyze exchange controls. The first part describes the range of policy instruments available to the authorities for dealing with balance-of-payments problems. Exchange control is primarily a balance-of-payments policy and this part relates it to all the other policy alternatives. It also presents a brief discussion on the possible rationale behind the prevalent use of exchange controls in LDCs, particularly in Africa. The second part outlines the ways in which exchange controls have been analyzed in the literature and develops a model of exchange controls. The model developed shows that parallel market activities can be explained through the optimizing behavior of exporters and importers. Exporters will keep shifting their exports from official to parallel market channels until the expected marginal benefit in both activities is the same. Import traders will devote resources to the rent-seeking activity until the moment in which an additional unit spent in this activity is equal to the savings derived from being able to use official channels instead of the parallel markets. Their combined behavior determines the amount of import and  -5- export smugglLng, the level of the rent-seeking activity, and the black market exchange rate that is consistent with an equilibrium position where no one has any more incentives to move from their attained position. In this context, rent-seeking, smuggling, and black markets are phenomena tha,t coexist in an economy with exchange controls. Indeed some very interesting Interrelationships between these variables are determined in the model. On one side, the rent-seeking activity seems to be inversely related to the official exchange rate, and directly related to the black market rate and the level of government policing activity. On the other side, the fraction of exports that is channelled through parallel markets is directly related to the black market rate, and inversely related to the official exchange rate and the level of government policing activity. There are nevertheless some relationships that cannot be determined a priori by the model. The most important relationship is the one between the black market rate and the official exchange rate. For the case of a country with severe foreign exchange constraints one can expect this relationship to be negative. On the one hand, the productivity of the rent-seeking activity is very limited as a result of the overall scarcity of foreign exchange, and on the other hand, we can expect that any increase in the government foreign exchange earnings will be reverted to the economy by relaxing the exchange controls. But as a general conclusion, the way in which a depreciation of the official exchange rate affects the black market rate seems to be an empirical question that can only be determined for each particular country under its specific circumstances. Section IV develops a simple methodology to detect the presence and assess the magnitude of the parallel market economy, as well as to explain its  -6- behavior in a quantitative fashion. This methodology is presented through a case study for Ghana, a country in Sub-Saharan Africa where parallel market activities seem to be widespread and relatively open in the economy. But, in principle, the methodology outlined in this section could be applied, by the country economist, to each particular country where exchange controls are imposed. In the first part, a reduced form equation of the black market exchange rate that is derived from the context of our model is estimated for the case of Ghana. There is a definite negative relationship between the real official exchange rate and the black market rate in this case. By letting the real official exchange rate appreciate, the government of Ghana has been losing an important amount of foreign exchange related to the exports that are now being smuggled out of Ghana instead of going through the official channels. This has meant an important reduction in the amount of foreign exchange allocated to imports and therefore an increase in the demand for smuggled imports in the economy. In all, the real appreciation of the official exchange rate seems to be very much related to the increasing importance of parallel market activities in the economy. The estimation results are encouraging and give ample support to the key determinants of the black market exchange rate derived from the context of our model. They also seem to indicate that the exchange rate policies pursued by the government of Ghana have led to the emergence and growing importance of the parallel market economy in this country. The second part of the section focuses on the estimation of the size of the parallel market economy and its evolution in time. We use the general methodology developed by Tanzi (1982), with which he estimates the underground economy and tax evasion in the United States. In our case, the approach  -7- emphasizes the fact that there is a demand for currency related to the activities of the parallel market economy. In particular, we postulate that the demand for currency in Ghana is directly related to the amount of domestic currency that is being traded in the black market. In this way, the method of estimating the size of the parallel market economy involves a two-stage procesdure. First, the amount of domestic currency traded in the black market has to be determined. This is, of course, an unobservabLe variable. We can, therefore, only estimate a proxy variable for the amount of domestic currency tradied in the black market which we do by estimating the quantity of cocoa smuggled out of Ghana. Second, we use this proxy variable in the estimation of the demand-for-currency equation. This gives an estimate of the amount of "illegal money" in the economy, which can then be transformed into a GDP estimate of the parallel market economy by assuming that the velocity of money is t'he same in the parallel market economy as it is in the "official" economy. The yearly estimates of the parallel market economy show the increasing importance of parallel market activities in the Ghanaian economy. From a situation where parallel market activities were nearly non-existent in 1965, the parallel market economy rose almost steadily to a 32.4 percent of official GDP Ln 1982. This discouraging picture of t'he Ghanaian economy pushed its government to undertake the policy measure it tried to avoid for so many years: by April 1984 the official exchange rate had been depreciated to 33 Cedis per U.S. Dollar, from the original 2.75 Cedis per U.S. Dollar, after having followted a transitional arrangement of a multiple exchange rate system introduced in April 1983.  -8- The historical lesson that Ghana embodies should be clear, its policy implications understood: Governments that impose exchange controls based on a hostile attitude toward exchange-rate changes, thinking that they are gaining control of the economy by having a "direct" allocative mechanism for foreign exchange, as distinct from an amorphous and unpredictable price mechanism, may find themselves losers in their own battle. With higher domestic inflation rates than world inflation rates these governments are progressively losing control of the economy as more and more transactions are diverted to the parallel markets. Furthermore, this loss of official control over the economy comes with an additional cost given by the inefficiencies related to the discriminatory process in which import licenses are assigned, as well as the increasing amount of resources that the economy loses in the rent-seeking activity. Finally, Section V elaborates on some of the policy implications of our analysis. It shows that the existence of a parallel market economy has a significant effect on the repercussions that follow exchange controls. For example, the imposition of exchange controls as a substitute for a formal devaluation does not avoid the adverse repercussions on prices or real wages of a devaluation. The emergence of a parallel market economy in response to such controls and the depreciation of the black market exchange rate have similar consequences to those of an official devaluation. Thus, the overall results of this paper indicate that the use of a more integrated framework to analyze exchange controls has important policy implications. Hopefully the paper will inspire further work along these lines.  -9- II. A HISTORICAL PERSPECTIVE OF EXCHANGE CONTROLS IN SUB-SAHARAN AFRICA During the last few years there has been a growing recognition that part of the roots of economic stagnation in Sub-Saharan Africa lie on domestic policy deficiencies. Although the external environment has been an aggra- vating factor, the record of poor growth in most Sub-Saharan African countries suggests that policymakers have given ina,lequate attention to increase the efficiency of resource allocation. This was the conclusion of the Sub-Saharan Africa report (1981) and is now being conEirmed by the analysis done for the new Sub-Saharan report (1984). A generalized impression is that trade and exchange-rate policy is at the heart of the failure to correct deep distortions in relative prices in these economies. To analyze this further we can, as a first approximation, divide Sub-Saharan Africa into three analytical groups of countries 5/, as shown in Table 1. 5/ This division is made on the basis of traditional balance-of-payments analysis that asserts: there are essentially three means by which a country's external accounts may be kept in balance: monetary and fiscal policies under fixed exchange rates responding to the imperatives of the balance-of-payments; flexible exchange rates; and exchange control, under which quantitative restrictions on international transactions are adjusted to restrict payments to foreigners to a level commensurate with foreign exchange availability from earnings and borrowing (Krueger, 1983, p. 173).  -10- Table 1: EXCHANGE RATE REGIMES IN SUB-SAHARAN AFRICA: THREE ANALYTICAL GROUPS Group Country Currency Pegged to I. Restrictions on 1. Botswana Currency Composite payments for current 2. Burundi U.S. Dollar transactions 3. Ethiopia U.S. Dollar 4. Gambia, The Pound Sterling 5. Ghana Currency Composite 6. Guinea SDR 7. Guinea-Bissau SDR 8. Kenya SDR 9. Madagascar Currency Composite 10. Malawi SDR 11. Mauritania Currency Composite 12. Mauritius SDR 13. Nigeria a/ 14. Rwanda U.S. Dollar 15. Sierra Leone a/ 16. Somalia SDR 17. Sudan U.S. Dollar 18. Tanzania Currency Composite 19. Uganda a/ 20. Zaire SDR 21. Zambia SDR 22. Zimbabwe Currency Composite II. No restrictions on 23. Liberia U.S. Dollar payments for current 24. Lesotho South African Rand transactions 25. Swaziland South African Rand III. CFA Area - The Central African 26. Cameroon French Franc Currency Union 27. Central African Republic French Franc 28. Chad French Franc 29. Congo, People's Republic French Franc 30. Gabon French Franc - The West African 31. Benin French Franc Currency Union 32. Ivory Coast French Franc 33. Niger French Franc 34. Senegal French Franc 35. Togo French Franc 36. Upper Volta French Franc - Other 37. Mali French Franc a/ Country follows a managed floating policy. Source: IMF, Exchange Arrangements and Exchange Restrictions, Annual Report 1983.  -11- As expected, no country in Sub-Saharan Africa follows a flexible exchange rate policy. In 1983 there were three countries: Nigeria, Sierra Leone, and Uganda that followed a managed floating system. The rest of the countries determined their exchange rates on the basis of a peg to a given currency or a composite of currencies, including the SDR. In the first group of countries, which is the largest one, the common pattern of response to foreign exchange scarcity has been to rely on exchange controls. 6/ Governments have imposed import restrictions rather than resorting to devaluation or restrictLve monetary and fiscal policies to conserve foreign exchange. The second group of countries have basically no restrictions on payments for current transactions and therefore rely on exchange rate adjustments and corresponding fiscal and monetary polices for balance-of-payments adjustments. Finally, the third group is represented by the CFA Franc Area. In this case the members of the two monetary unions, the Central African Currency Union and the West African Currency Union, signed an agreement of monetary cooperation with France whereby the exchange rate between the French franc and the franc of African Financial Cooperation (CFA) was fixed, foreign exchange reserves were pooled, exchange controls were common to the whole zone, and an "operation account" at the French Treasury guaranteed the convertibility of the CFA franc. R. Mundell explains the outcome of this arrangement concisely: Balance of payments equilibrium is maintained by rigid financial discipline of the member countries, the governments of which are prevented from using the Central Bank as a source of inflationary finance. Government accounts must be balanced except insofar as 6/ In general, when we talk about exchange controls in this paper, we focus on the related restrictions on payments for current tran- sactions. Most countries in Sub-Saharan Africa have restrictions on payments for capital transactions.  -12- the government has cash reserves at its disposal or can acquire extra credits from one of the common banks or from abroad. 7/ This does not mean that these countries have not employed quanti- tative restrictions, but rather that any such restriction has been imposed for motives other than balance-of-payments constraints, and have been relatively invariant with respect to the availability of foreign exchange. The marked differences in the exchange rate regimes followed by the countries in the analytical groups can be appreciated by observing the behavior of their nominal and real effective exchange rates. 8/ This is presented in Tables 2 and 3, and Figures 1 and 2. 9/ 7/ R. Mundell (1972), p. 33. The nominal effective exchange rate is here defined as the import-weighted geometric average of the relevant individual bilateral exchange rates. The actual formula is given by: Eio where Eit represents the price of domestic currency in terms of the ith partner country at time t (Eio is for the base period), and wi is the import weight for the ith trading partner. The real effective exchange rate is an import-weighted geometric average of the bilateral exchange rates adjusted by the ratio of domestic consumer price index to the corresponding trade partner consumer price index. They are calculated according to the formula: E Pt w REER - i (g t I) io it where Pit represents the price level of the ith country at time t relative to the base period, and Pt represents the price level of the home country, also relative to the base period. For a good review study of the major conceptual and methodological problems involved in the use of real effective exchange rate indices see Maciejewski (1983). .-I The indices were only computed for countries in which data were readily available. Central African Republic, Chad, Benin, Guinea, Guinea-Bissau, Mali, Uganda, Seychelles, Botswana, Zimbabwe, Lesotho, and Swaziland were not included.  -13- Table 2: REAL EFFECTIVE EXCHANGE RATE INDICES, 1972-1982 (Indices, 1972 - 100) 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 GROUP 1 Burundi 100.00 97.74 102.20 101.61 97.05 86.47 89.66 109.73 108.39 124.99 143.48 Ethiopia 100.00 101.87 100.41 94.04 120.05 123.99 119.79 123.66 113.90 124.78 135.62 Gambia 100.00 109.27 105.40 108.32 103.06 100.62 101.70 103.29 103.75 99.13 99.45 Ghana 100.00 115.64 124.11 139.76 212.67 415.53 467.49 347.68 454.88 1005.70 1252.21 Kenya 100.00 96.80 99.81 101.16 98.87 102.84 108.73 105.05 102.91 99.15 102.40 Madagascar 100.00 99.47 104.62 105.31 99.37 91.75 89.92 94.48 99.83 112.35 122.11 Malawi 100.00 89.17 90.13 90.30 90.09 84.65 87.15 86.00 87.02 85.50 82.36 Mauritania 100.00 101.92 103.70 99.75 110.18 110.80 100.03 94.09 91.40 113.63 124.57 Mauritius 100.00 93.29 101.46 99.03 99.97 99.84 101.07 100.53 99.96 100.03 93.84 Nigeria 100.00 91.07 96.4b 114.26 139.51 144.75 146.71 148.45 157.08 176.59 180.16 Rwanda 100.00 101.63 107.24 121.44 125.47 125.32 117.68 122.55 120.27 134.87 157.35 Sierra Leone 100.00 89.97 87.63 85.84 80.28 74.97 76.64 80.25 77.21 88.55 113.20 Somalia 100.00 103.27 109.11 114.68 132.82 130.64 124.51 137.21 188.89 277.57 245.82 Sudan 100.00 103.24 113.94 122.88 122.43 129.50 124.70 131.87 121.34 140.03 104.13 Tanzania 100.00 97.53 102.18 109.27 102.30 102.13 103.67 97.31 109.81 140.98 166.86 Zaire 100.00 96.78 110.98 123.77 145.66 192.85 250.12 226.09 178.23 174.14 177.62 Zambia 100.00 100.64 98.38 96.41 104.69 101.62 100.40 96.31 91.63 95.13 104.87 Group Avg. 2-j 100.00 96.93 101.58 113.59 136.15 166.09 178.93 158.71 176.11 316.07 397.28 GROUP 2 Liberia 100.00 102.55 108.68 108.26 109.15 104.88 98.34 98.07 100.29 109.04 116.90 GROUP 3 Cameroon 100.00 105.08 106.06 111.56 110.05 113.31 118.26 115.35 112.40 106.54 103.69 Congo 100.00 97.35 88.99 96.26 93.21 96.27 98.09 97.28 92.99 94.42 93.25 Gabon 100.00 101.26 97.86 117.63 127.01 130.39 134.74 133.48 133.37 123.71 125.26 Ivory Coast 100.00 105.37 106.01 109.31 109.50 124.93 130.61 140.00 142.78 132.97 124.23 Niger 100.00 106.03 94.38 95.27 105.27 115.29 117.55 115.01 112.40 118.52 115.32 Senegal 100.00 105.18 105.15 128.42 116.90 116.86 111.98 111.83 107.54 98.01 101.17 Togo 100.00 99.11 94.87 105.57 103.98 111.78 104.04 103.59 102.49 101.35 96.86 Upper Volta 100.00 102.14 91.91 105.23 85.63 99.80 105.71 110.57 110.59 100.47 96.21 Group Avg. - 100.00 103.93 101.74 111.96 111.52 119.18 121.89 127.07 124.08 115.07 110.74 Sub-Sahar a7 Africa - 100.00 98.19 101.87 113.58 132.21 158.59 169.74 154.67 168.35 287.25 355.53 1/ Individual country indices are aggregated for each group by using GNP weights.  -14- Table 3: NOMINAL EFFECTIVE EXCHANGE RATE INDICES, 1972-1982 (Indices, 1972 = 100) 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 GROUP I Burundi 100.00 100.18 103.51 100.45 99.82 92.65 83.73 82.65 83.83 99.89 115.26 Ethiopia 100.00 102.49 108.10 107.80 118.33 117.16 107.04 104.79 106.21 123.80 140.76 Gambia 100.00 112.22 114.95 110.45 100.19 96.23 94.84 99.56 105.46 103.89 101.72 Ghana 100.00 106.22 109.15 107.65 115.17 113.78 79.30 41.71 40.76 46.06 50.45 Kenya 100.00 96.67 97.85 96.23 93.85 94.42 91.53 90.06 88.34 84.46 78.68 Madagascar 100.00 100.86 98.49 101.69 99.60 97.22 96.43 97.13 97.58 94.17 85.85 Malawi 100.00 91.78 90.86 90.52 96.88 98.43 100.27 99.80 95.50 96.06 92.92 Mauritania 100.00 102.86 106.11 104.00 109.29 109.30 99.35 94.37 93.34 108.45 116.69 Mauritius 100.00 90.78 88.31 86.36 85.64 86.89 86.28 81.96 64.58 62.60 57.52 Nigeria 100.00 93.82 100.99 102.53 111.29 106.74 97.28 96.90 103.90 106.40 108.81 Rwanda 100.00 100.48 92.15 90.15 96.87 94.10 85.17 85.16 86.79 100.70 115.00 Sierra Leone 100.00 92.73 90.62 85.58 75.83 72.24 71.48 67.44 65.76 67.69 70.89 Somalia 100.00 107.58 113.37 114.46 131.26 134.01 129.51 129.84 133.95 160.20 131.01 Sudan 100.00 99.56 101.84 103.21 112.63 113.79 99.45 89.14 75.88 79.74 52.21 Tanzania 100.00 96.30 97.49 95.00 92.52 93.16 92.14 82.92 81.58 92.93 94.29 Zaire 100.00 90.60 92.76 89.74 63.86 54.50 50.67 23.74 14.58 11.61 9.37 Zambia 100.00 103.31 107.38 109.84 111.67 100.67 92.01 89.24 86.41 87.77 93.96 Group Avg.!/ 100.00 96.95 101.19 101.94 107.45 103.95 92.07 86.76 89.45 88.49 87.52 GROUP 2 Liberia 100.00 93.07 93.76 92.03 96.15 94.27 87.49 84.70 84.08 92.92 100.32 GROUP 3 Cameroon 100.00 102.37 99.51 103.60 101.98 100.09 100.52 101.07 101.14 97.00 92.88 Congo 100.00 101.19 99.08 101.59 100.08 98.47 98.52 98.96 99.11 96.25 93.18 Gabon 100.00 102.18 99.24 103.17 100.78 98.76 99.35 99.95 99.98 95.35 90.97 Ivory Coast 100.00 102.13 99.05 103.06 100.88 98.73 98.83 99.52 99.47 95.36 91.50 Niger 100.00 106.03 94.38 95.27 105.27 115.29 117.55 115.01 112.40 118.52 115.32 Senegal 100.00 102.41 99.88 104.38 104.70 104.48 106.20 107.80 109.04 106.33 104.62 Togo 100.00 103.21 98.87 104.99 101.80 98.67 98.96 100.71 99.99 92.82 87.11 Upper Volta 100.00 104.17 101.43 114.28 115.77 119.16 135.04 149.77 172.70 186.04 200.79 Group Avg.!' 100.00 102.62 99.15 103.56 102.48 101.45 103.11 105.33 105.33 102.85 99.73 Sub-Sahar an Africa - 100.00 97.84 100.79 102.15 106.57 103.58 93.85 89.34 91.96 90.60 89.39 1/ Individual country indices are aggregated for each group by using GNP weights.  Figure 1: REAL EFFECTIVE EXCHANGE RATES, 1972 - 1982 (1972 = 100) INDEX 489 350 uROUr A= / 250 / GROUP I 11/O GHANA 200 159 100 GROUP 3 50 I l l I I | YEAR 72 73 74 75 76 77 78 79 80 81 82  Figure 2: NOMINAL EFFECTIVE EXCHANGE RATES, 1972 - 1982 (1972 = 100) INDEX as~~~~~~~~~~~~~~~~~IO P 3 19 g0 + , ~~~~~~~~~GROUP 1 195 GROUP 1 85 W/O GHANA 89 75 70 65- I I I I I I I I I I YEAR 72 73 74 75 76 77 78 79 89 81 82  -17- The GNP weighted average of real effective exchange rates in Sub- Saharan Africa appreciated by 54 percent 10/ between 1972 and 1982. But the difference between analytical groups is enormous, and can be seen clearly in Figures 1 and 2. At one extreme, most CFA countries experienced relatively stable nominal and real effective exchange rates over the period. On average, the rates of these countries appreciated in real effective terms by only 11 percent from 1972 to 1982. By contrast, real effective exchange rates in Group I appreciated by 77 percent over the same period, even after having a 31 percent depreciation of their average nominal eff'ective exchange rates. These figures are heavily influenced by the experience of some of the countries in the group where the need to offset relatively high domestic inflation rates by exchange rate changes is precludled by the existence or imposition of exchange controls. This result tends to confirm the general expectation that "relatively appreciating" countries must eventually intensify import restrictions to maintain overvalued exchange rates and avoid running out of reserves. The characteristics of these analytical groups reflect the marked differences irn historical experience between Francophone and Anglophone countries. As Mundell (1972) puts it, it: is not surprising that the differences in cultural patterns arising from the different methods and 10/ The percentage changes in effective exchange rates referred to in the text were computed disregarding Ghana, which is a very extreme case.  -18- techniques of colonization reveal themselves in the currency and banking systems inherited by the African countries, on the approaches to economic policy, and the settings within which policies are carried out. Overall Better monetary policy has been the beneficiary in the French countries; greater monetary experience in the English. Some of the former British colonies gained financial experience and institutional development in the process, but the experience has not been acquired without cost and as a consequence more rigid systems of exchange control, with its consequenMe for illict trading and distortions, have been imposed. It is the purpose of this paper to analyze some of these consequences of exchange controls for the countries that have relied on them. -L In particular, the paper will focus on the way in which exchange controls affect the behavior of the different agents in the economy, leading to the creation of a parallel market economy. 11/ Mundell, op. cit., p. 27. 12/ A detailed summary of the exchange control measures adopted by the different countries in Group I is presented in Appendix A.  -19- III. EXCHANGE CONTROLS: THE THEORETICAL FRAMEWORK A. Exchange Controls as a Balance of Payments Policy Exchange control is primarily a balance of payments policy, but it is just a particular option within a wide range of balance-of-payments policy instruments available to achieve external balance. We will now make a quick review of these policy alternatives. We will do so by analyzing their impact on the foreign exchange market, depicted in Figure 3. As we know, the exchange market is a conceptual device that allows us conveniently to summarize the forces determining equilibrium in exchange between countries. Lurking behind the demand and supply curves of Figure 3 are people's desires to hold curcencies as financial assets as well as to use them as means of payment. We must remember that the demand and supply for foreign exchange shift wit'h any of innumerable real and monetary economic conditions. To begin with, let us assume that the exchange rate is overvalued at eo, and, therefore, we have an excess of ex-ante payments over receipts. The theory of international economics has shown that there are essentially two automatic adjustment mechanisms to an imbalance of international transactions. If the exchange rate is flexible, the price of foreign exchange will immediately be bid up and domestic money will depreciate. The gap between autonomous demand and supply of exchange will be closed by movements along the existing schedules to the equilibrium exchange rate, e*. Alternatively, if the exchange rate is pegged, the deficit will gradually reduce the net foreign liquidity of the country.  -20- Figure 3: FOREIGN-EXCHANGE MARKET Price of Foreign Exchange D S e e __ _ o- __ - X _ __ _ ___--- X Q5Q* Qd Quantity of