MADIA DISCUSSION PAPER 5 F 8323 _~~- _ _f _ MANAGING= DEVELOPMENTT IN f, 7mi4~- .1 U _ I N _ _ _ _ _ _ ___ AFRCA ___ ,r_f^~f r,,nsfY ,, f.-; ',3 i-w . . t. - s .. FOREWORD The MADIA study and the papers comprising this MADIA Discussion Paper Series are important both for their content and the process of diagnosis and analysis that was used in the conduct of the study. The MADIA research project has been consultative, nonideological, and based on the collection and analysis of a substantial amount of concrete information on specific topics to draw policy lessons; it represents a unique blend of country-oriented analysis with a cross-country perspective. The conclusions of the studies emphasize the fundamental importance of a sound macroeconomic environment for ensuring the broad-based development of agriculture, and at the same time stress the need for achieving several difficult balances: among macroeconomic, sectoral, and location-specific factors that determine the growth of agricultural output; between the development of food and export crops; and between the immediate impact and long-run development of human and institutional capital. The papers also highlight the complementarity of and the need to maintain a balance between the private and public sectors; and further the need to recognize that both price and nonprice incentives are critical to achieving sustainable growth in output. The findings of the MADIA study presented in the papers were discussed at a symposium of senior African and donor policymakers and analysts funded by USAID in June 1989 at Annapolis, Maryland. The participants recommended that donors and African governments should move expeditiously to implement many of the study's valuable lessons. The symposium also concluded that the process used in carrying out the MADIA study must continue if a stronger, more effective consensus among donors and governments is to be achieved on the ways to proceed in resuming broad-based growth in African agriculture. The World Bank is committed to assisting African countries in developing long-term strategies of agricultural development and in translating the MADIA findings into the Bank's operational programs. Stanley Fischer Edward V. K. Jaycox Vice President Development Economics Vice President and Chief Economist Africa Reyional Office MADIA DISCUSSION PAPER 5 FERTILIZER POLICY IN AFRICA LESSONS FROM DEVELOPMENT PROGRAMS AND ADJUSTMENT LENDING, 1970-87 UMA LELE ROBERT E. CHRISTIANSEN KUNDHAVI KADIRESAN THE WORLD BANK______ - WASHINGTON, D.C.- Copyright ( 1989 All rights reserved The International Bank for Reconstruction Manufactured in the United States of America and Development/THE WORLD BANK First printing November 1989 1818 H Street, N.W. Washington, D.C. 20433, U.S.A. MADIA Discussion Papers are circulated to encourage discussion and ment. at the address shown in the copyright notice above. The World Bank comment and to communicate the results of the Bank's work quickly to the encourages dissemination of its work and will normally give permission development community: citation and the use of these papers should take promptly and, when the reproduction is for noncommercial purposes, with- account of their provisional character. Because of the informality and to out asking a fee. Permission to photocopy portions for classroom use is not present the results of research with the least possible delay, the manuscript required, though notification of such use having been made will be has not been prepared in accordance with the procedures appropriate to appreciated. formal printed texts, and the World Bank accepts no responsibility for The complete backlist of publications from the World Bank is shown in errors. The findings, interpretations, and conclusions expressed in this the annual Index of Publications, which contains an alphabetical title list and paper are entirely those of the author(s) and should not be attributed in any indexes of subjects, authors, and countries and regions. The latest edition is manner to the World Bank, to its affiliated organizations, or to members of available free of charge from the Publications Sales Unit. Department F, The its Board of Executive Directors or the countries they represent. World Bank, 1818 H Street, NW., Washington, D.C. 20433. U.S.A., or from The material in this publication is copyrighted. Requests for permission Publications, The World Bank, 66, avenue dlena, 75116 Paris, France. to reproduce portions of it should be sent to Director, Publications Depart- Uma Lele is the manager of Agricultural Policy in the Africa Technical Department at the World Bank. Robert E. Christiansen is an economist of Agricultural Policy in the Africa Technical Department at the World Bank. Kundhavi Kadiresan is a consultant of Agricultural Policy in the Africa Technical Department at the World Bank. Library of Congress Cataloging-in-Publication Data Lele, Uma J. Fertilizer policy in Africa: lessons from development programs and adjustment lending, 1970-87. (MADIA discussion paper; 5) Includes bibliographical references. 1. Agriculture-Economic aspects-Africa, Eastern. 2. Agriculture and state-Africa, Eastern. 3. Managing Agricultural Development in Africa (Organization) I. Title. II. Series. HD2126.L44 1989 338.1'8676 89-22728 ISBN 0-8213-1321-5 Contents Introduction .......................................... 5 Trends in Donor Policy Toward Fertilizer Use .......................... 7 Evolution of Donor Views Toward the Role of Fertilizer ...... ........... 7 Policy Assessment .......................................... 7 Fertilizer Use in the MADIA Countries ................................. 9 Overview ............................................ 9 IntraCountry Patterns of Fertilizer Use ............................... 11 Policy Regimes and Growth of Fertilizer Consumption .................. 19 Policy Reform in the Context of Past Performance ..................... 22 Kenya ............................................ 22 Tanzania ............. ............................. 24 Malawi ............................................ 25 Cameroon ............................................ 27 Senegal .......................................... 28 Nigeria .......................................... 30 Determinants of the Economic Benefits of Fertilizer Use ..... .......... 32 Fertilizer Price ........................................... 32 Producer Price and the Relative Cost of Fertilizer ..................... 33 Response Coefficients .......................................... 36 Benefit-Cost Ratios for Fertilizer Use ................................ 41 Summary, Conclusions, and Implications ............................. 47 Appendices .............. ............................ 50 Notes ....... : 66 Bibliography ....... 74 Acknowledgments This paper has benefited greatly from our interactions with individuals in USAID field offices in Kenya, Malawi, Cameroon, and Senegal who have provided valuable material on the adjustment experience and responded to many of our queries. In particular we would like to thank Messrs. A. Radi (Malawi), S. Sinding (Kenya), Tham Truong (Cameroon), and Wayne Nilsestuen (Senegal). We are grateful to M. Agarwal, R. All, V Bindlish, S. Carr, K. Cleaver, G. Desai, G. M. Higgins, S. Jammeh, R. Nolan, D. Pickering, A. Seth, R. Southworth, A. Spurling, D. Spurling, T. N. Srinivasan, D. Steeds, 1. Twyford, N. van de Walle, and M. Westlake for their valuable input into the paper. Kim Tran and Christina Dhanaraj provided secretarial support. We alone, however, are responsible for any deficiencies that remain. Illustrations 4. Growth Rate of Export and Food Crop Production in the MADIA Countries, 1970-85 .................................... 53 Tables , 5. Table 1: Trend in the ratio of producer prices for export crops to I. Actual and projected per capita arable land in MADIA countries .... 5 food crops in the East African MADIA countries ....... . ........ 53 2. Role of export and food crops in the balance of payments in MADIA 5. Table 2: Trend in the ratio of producer prices for export crops to countries, 1981-85 ......................................... 6 food crops in the West African MADIA countries ................ 54 3. Food self-sufficiency ratios for MADIA countries ................. 6 6. Fertilizer Use by Crop in the MADIA Countries .54 4. Fertilizer use per hectare of arable land ........................ 6 7. Malawi's Smallholder Fertilizer Revolving Fund .55 5. Factors affecting fertilizer use in MADIA countries, 8. Producer Prices for Primary Nutrient Types Used in the MADIA 1970-87 ....................... 9 Countries, 1971/72-1987/88 .................................. 57 6. Average fertilizer use in MADIA countries ......... ............. 11 9. Table 1: Producer prices for maize in MADIA countries, 1971-87 .... 57 7. Maize deficit and surplus areas by province and district in Kenya, 9. Table 2: Producer prices for seed cotton in MADIA countries, extent of high and medium potential land, and crop response to 1970-87 ... 58 fertilizer .... ..................................,14 9. Table 3: Producer prices for coffee in Kenya and Cameroon, 8. Population, rainfall, arable land, and cropping pattern by region in 1971-87 .................................... 58 Senegal .............................................. 17 10. Analysis of Nutrient Price/Crop Price Ratios .................... 58 9. Rate of explicit fertilizer subsidy in MADIA countries, 1970-87 .19 10. Table 1: Nutrient price/crop price ratios for selected crops in MADIA 10. Budgetary cost of fertilizer subsidies and their share in total countries .60 government expenditure in West Africa MADIA countries, 1974-87 19 IL Notes and Sources for Crop Response Data .61 11. Budgetary cost of fertilizer subsidies and their share in total Il. Table 1: Fertilizer recommendations of IFDC and ISRA for millet in government expenditure in East Africa MADIA countries, 1978-87 ..20 Senegal .... 62 12. Fertilizer policy reforms in the MADIA countries ..... .......... 22 11. Table 2: Fertilizer recommendations of IFDC and ISRA for 13. Prices for principal fertilizers in MADIA countries, groundnuts in Senegal . 63 1971/72-87/88 ......... 32 12. Sources and Notes for Tables 23 and 24 ....................... 63 14. Actual cost of transport and as a share of marketing cost and ci.f. 13. Distance, Number of Outlets, and Areas Served by Retail Outlets in fertilizer cost in selected countries .. 34 the MADIA Countries ............... 65 15. Nutrient price/crop price ratios for selected crops in East Africa, 1980-88.34......brev...at...ons..and...Acronyms. 3 16. Nutrient price/crop price ratios for selected crops in West Africa Abbreviations and Acronyms 1980-87 ................................................ 36 ADP Agricultural Development Program 17. Response coefficients for selected crops in East Africa ........... 37 AF Amelioration Fonciere Agricultural Research Program 18. Response coefficients for selected crops in West Africa .... ...... 38 AISCO Agricultural and Industrial Supplies Company 19. Coefficient of variation in international fertilizer prices, 1975-85 .... 41 A/S Ammonium Sulfate 20. Estimated range of benefit-cost ratios due to changes in ASA Annual Survey of Agriculture international fertilizer prices ....... ......................... 42 BCCC Bank of Credit and Commerce-Cameroon 21. Coefficient of variation in commodity prices ........ ............ 42 CASUCO Cameroon Sugar Company 22. Estimated range of benefit-cost ratios due to changes in commodity CAN Calcium Ammonium Nitrate prices .. 43 DAP Di-Ammonium Phosphate 23. Benefit-cost ratios for fertilizer use in the East Africa MADIA EEC European Economic Community countries .. ...~ , , .......... ...... ...... ... , , 44FAO Food and Agricultural Organization 24. Benefit-cost ratios for fertilizer use in the West Africa MADIA GOK Government of Kenya countries.~~~~~~~~~~~ GDP Gross Domestic Product countries, ..... .,.,,,,,,,,,,,,,,,,,,,,,..5........................... 45 rs oesi rdc 25. Benefit-cost ratios with high analysis fertilizer in Malawi .......... 46 GNP Gross National Product 26. Benefit-cost ratios for harvest prices and estimated range of benefit- IAR Institute of Agricultural Research cost ratios due to intrayear crop price variations ........ ....... 47 IDA International Development Association IFDA International Fund for Agricultural Development Figures IFDC International Fertilizer Development Center I. Gross food imports for the MADIA countries, 1971-86 ............. 5 IFPRI International Food Policy Research Institute 2. Trends in fertilizer consumption in the MADIA countries, 1970-87... 10 ISRA Institut Senegalais de Recherches Agricoles 3. Fertilizer use by crop in the MADIA countries .................. 18 K Potash 4. Prices for primary nutrient types used in the MADIA countries .... 33 KTDA Kenya Tea Development Agency 5. Producer price of maize in the MADIA countries, KPCU Kenya Planters Co-operative Union 1971-87 .. , 35 MADIA Managing Agricultural Development in Africa 6. Producer price of cotton in the MADIA countries, N Nitrogen 1970/71 to 1986/87 .......... .. , 35 NRDP National Rural Development Program 7. Producer price of coffee in Cameroon and Kenya, ONCAD Office National de la Cooperation et d'Assistance pour 1971-87 .. , 36 le D6veloppement P Phosphorus S Sulphur Appendices SAED Societe dAmenagement et d'Exploitation des Terres viip ~ ~~~~~~~~~~~~~~~~~~~~du Delta du Fleuve Se6negal 1. Gross Food Imports (CerealsI in the MADIA Countries, SEMRY Societe d'Expansion et de Modernisation de la 1971-86 ........ , .............,,,,,,,,,, 50 Riziculture de Yagoua 2. Table 1: Ratios of producer prices to international prices for major SFRF Smallholder Fertilizer Revolving Fund smaliholder crops in East African countries, 1970-85 (using nominal SODECOTON Soci4te de Developpement du Coton du Cameroun exchange rate) ........................,,,.50 SODEFITEX Societe de Ddveloppement des Fibres Textiles 2. Table 2: Ratios of producer prices to international prices for major SONACOS Societe Nationale de Commercialisation des smaliholder crops in West African countries, 1970-85 (using nominal Oleagineux du S6negal exchange rate) ......... , .,,,., .......... 51 SOSUCAM Societe Sucriere du Cameroun 2. Table 3: Ratios of producer prices to international prices for major TANU Tanganyikan African National Union smalihoider crops in East African countries, 1970-85 (using TFA Tanzania Farmers Association purchasing power parity exchange rate) ........................ 51 TFC Tanzania Fertilizer Company 2. Table 4: Ratios of producer prices to international prices for major TRDB Tanzanian Rural Development Bank smaliholder crops in West African countries, 1970-85 (using TSP Triple Superphosphate purchasing power parity exchange rate) .52 UCCAO Union Centrale des Cooperatives Agricoles de l'Ouest 3. Volume of and Growth Rates for Fertilizer Consumption in the UNCA Union Nationale des Cooperatives Agricoles MADIA Countries .52 USAID United States Agency for International Development Introduction In many parts of Africa over 80 percent of the value added Table 1 in smallholder agriculture comes from a production process Actual and projected per capita arable land in MADIA where the handhoe is frequently the only other major input countries besides labor. As a result, average labor productivity among Hectare per capita African smallholders is substantially lower than in Asia Hectare per capita forruralpopulation (Delgado and Ranade 1987). The problem of low labor Country 1985 2000 1985 2000 productivity is exacerbated by increasing population pres- sure on arable land.' Although Africa is commonly viewed Kenya 0.73 0.42 0.86 0.66 as a land-surplus continent, this view is no longer true of Malawi 0.73 0.45 0.84 0.60 many African countries where population pressure is Tanzania 2.30 1.44 2.59 1.68 causing a reduction in the traditional bush fallow system Nigeria 0.71 2409 5.23 4.86 and is increasing the movement of population to marginal Senegal 1.62 1.04 2.41 1.76 land (Eicher 1982; Lele and Stone 1989). This, in turn, has the potential to contribute to a decline in soil fertility and Source: Lele and Stone 1989. growing deforestation that have serious implications for the sustainability of African agriculture. Table I shows the actual and projected per capita arable land in the MADIA value of Senegal's exports for the 1981 to 1985 period. In countries.2 While the specific experiences of these coun- Nigeria's case, the once large share of agriculture in exports tries as regards the dynamics of population and land are had virtually vanished to be replaced by food and other explored in other MADIA papers (Lele and Stone 1989), the agricultural imports (i.e., cotton, edible oil) that constituted intensity of the growing population pressure is evident in one-fifth of the vastly expanded import bill (see Table 2). four of them (Kenya, Malawi, Nigeria, and Senegal). The impression of increasing dependence on imported At the macroeconomic level, the urgent need for rapid food is confirmed by the decline in self-sufficiency ratios growth in both food and export crop production is evident between the 1960s and 1980s in all the MADIA countries, from rising food imports and the inability of many countries except Malawi (see Table 31, where import dependence has to finance them because of stagnant export earnings. Figure increased since 1986 due to the growing refugee popula- I shows the growing food import dependence of the tion, together with declining per capita maize production.3 MADIA countries over the period of the 1970s and 1980s A crucial ingredient in the process of increasing agricul- (also see Appendix 11, and Table 2 shows the fundamental tural productivity is clearly the increased use of chemical importance of agriculture in employment, trade, and GDP in fertilizers, although it must be acknowledged that they these countries. Food imports constituted 43 percent of the alone cannot solve the complex problems of declining soil Figure 1 Gross food imports in the MADIA countries, 1971-86 East African countries West African countries Million tons of cereal Million tons of cereal 2.4 2.4 - 2.2 -2.2- 2 -2 - 1.8 -1.8 : 0" 1.6 -1.6 1.4 -1.4 - 1.2 -1.2- 1 1 /' 0.8 - 0.8 - 0.6 -/00. 0.4 - -0.4 - * 0.2 0 0.2- 1971 73 75 77 79 81 83 85 88 1971 73 75 77 79 81 83 85 86 . , .. Malawi Cameroon Kenya ............ Senegal - - - - Tanzania Nigeria Source: See Appendix 1. 5 Table 2 Role of export and food crops in the balance of payments in MADIA countries, 1981-85 Average share of Total Agricultural Gross food imports as Agriculture as a Export export export percentage of total share of Country crop earnings eamings imports exports employment GDP Percent East Africa Malawi Tobacco 50 56 15.5 14 83 38 Tea 20 23 Sugar 12 14 Kenya Coffee 24 42 8.5 14 81 31 Tea 16 28 Tanzania Coffee 34 32" 8.0 18 86 58 Cotton 12.5 15'* West Africa Senegal Groundnut 19 85 24.0 43 81 19 Cotton 2 15 Cameroon Cocoa 9.5 34 10.5 11 70 21 Coffee 10.5 37 Cotton 3 10 Nigeria Cocoa 2 4- 20.0 19 68 36 Notes: * Indicates an average for the years 1980-84. Indicates an average for 1981 and 1982. Indicates average for 1979-81. Source: Lele 1988; World Bank Database. Table 3 Table 4 Food self-suff iciency ratios for MADIA countries Fertilizer use per hectare of arable land 1960-69 1970-79 1980-86 Region 1970 1975 1980 1985 Kilograms of nutrient/hectare Cameroon Mean 0.94 0.89 0.87 Africa 10 13 18 20 Maximum 0.97 0.94 0.90 Latin America 20 29 44 41 Minimum 0.91 0.84 0.84 Oceanic 34 29 36 32 Coefficient of variation (percent) 2.05 3.10 2.30 Developing Countries 18 27 49 58 Senegal Asia 26 37 68 85 Mean 0.74 0.66 0.61 North America 70 87 99 85 Maximum 0.79 0.79 0.72 Western Europe 176 188 221 226 Minimum 0.69 0.57 0.49 World 49 63 80 87 Coefficient of variation (percent) 3.90 12.64 13.24 Source: FAO, Fertilizer Yearbook, 1986. Nigeria Mean 0.98 0.90 0.84 Maximum 0.99 0.96 0.89 Minimum 0.96 0.75 0.79 fertility and quality. A more complete solution requires that Coefficient of variation (percent) 1.02 8.48 4.63 chemical fertilizers be used in conjunction with a variety of Kenya policies that promote soil and farm management tech- Mean 1.03 1.02 0.91 niques.4 Given the potential for increased fertilizer use, it Maximum 1.14 1.07 1.01 is ironic that despite massive amounts of donor assistance Minimum 0.94 0.97 0.77 since the early 1970s,5 Africa's fertilizer use per hectare Coefficient of variation (percent) 6.51 2.87 9.03 remains the lowest in the world (see Table 41 and issues Tanzania related to protecting land quality have largely been Mean 0.95 0.93 0.92 ignored. Moreover, the share of the Sub-Saharan region in Maximum 0.99 1.02 0.94 the otherwise rapidly growing fertilizer consumption rate of Minimum 0.89 0.77 0.89 the developing world has declined since 1970-71 6 The Coefficient of variation (percent) 3.68 8.49 1.93 underutilization of fertilizer makes fertilizer pricing, sub- Malawi sidy, and distribution policy, together with the alleviation of Mean 1.02 0.99 1.03 other technological and institutional constraints one of the Maximum 1.07 1.02 1.14 most pressing issues in the modernization of African small- Minimum 0.98 0.90 0.96 holder agriculture. Coefficient of variation (percent) 2.80 3.58 5.99 Source: FAO database. 6 Trends in Donor Policy Toward Fertilizer Use Evolution of Donor Views lbward the Role Evidence of these offsetting benefits is, however, contro- of Fertilizers versial, especially in the case of food crops. Food prices in During the 1970s, donors generally used subsidies on many West African countries tend to be determined largely fertilizer in the development projects that they funded as by domestic market forces, and there is little government a means of encouraging rapid growth in its use. Among the intervention in food markets with the exception of controls justifications for subsidies were that they (I) encourage on external trade and rice prices (see, for instance, Gelb learning by doing, (2) reduce the risk of using fertilizer and 1988). Reflecting excess demand for food, internal food help overcome credit constraints, (3) help poor farmers, (4) prices in these countries have tended to be above world contribute to maintaining soil fertility, and (5) offset market prices throughout the 1970s and early 1980s, even disincentives caused by taxation of output (World Bank when considered at purchasing power parity exchange rates 1986a, p. 95). In addition, Srinivasan (1986) has observed (see Figure 5 and Appendix 9). At best only a fifth of total that when planning and implementing capacity is limited, production went through official channels. The primary governments of developing countries tend to increase benefit of internal liberalization of food markets has spending on fertilizers as a means of achieving quick therefore been the budgetary gain to governments that had results. This is in contrast to investments in other areas, e.g., subsidized grain marketing interventions. Although privat- agricultural research and irrigation, that require a much ization has reduced the risks and costs encountered by longer gestation period (Srinivasan 1986). Srinivasan ob- food producers in clandestine marketing (e.g., in Kenya and serves that " . . in principle an argument for intervention Tanzania), consumers have lost the protection that they can be made in almost all cases," because enjoyed because of the intra- and interyear price and the efficiency and optimality of non-intervention is supply stabilization achieved through government interven- based on... "a complete set of contingent commodity tion, e.g., in Malawi.9 markets" (in particular, the existence of markets fo. Even in the case of export crops (to which the argument markets". (npriua, th exstnc of marketsfor favoring price reform mainly applies), although there was insuring against all kinds of risk) and.. .in no econ- scope at the beginning of the structural adjustment process likmy to be m or dt e for increasing output prices through exchange rate adjust- likely to be met.... ments and removal of taxes, the limits of these increases He concludes, however, that "In practice ... the costs of have been reached in many countries. Indeed, after the intervention must be weighed against its benefits" (Srini- initial increases reflecting the world market prices, down- vasan 1986, p. 49). Due to many of these same arguments ward adjustments were necessary in many countries as fertilizer subsidies, public sector monopoly of imports, and international prices declined.'0 Further, many of the earlier active public sector promotion of fertilizer use played an benefits of price correction have been eroded by the important role in fueling the green revolution in Asia, even continued increased cost of imported inputs especially though far more dramatic new technologies were available since the withdrawal of input subsidies. In addition, access for rice and wheat in Asia with high and assured returns to institutional credit has declined in some countries and, from their adoption than is the case in rainfed African where liberalization of credit has occurred, the private agriculture (Ahmad 1988). In the Asian context fertilizer sector's response has been mixed. In most cases the subsidies were considered to be more efficient than process of liberalization has demonstrated the complex support of product prices (Barker and Hayami 1976), reality of the task of developing distribution networks in the although in Asia guaranteed minimum prices for rice and agricultural sector. wheat have also been provided by governments as an The outcome of these and related developments, which important element of an integrated policy toward agricul- this paper examines in detail, has been stagnation in tural intensification (Siamwallah 1981). fertilizer consumption in some countries and the continued During the 1980s, increased budget deficits and doubts slow growth in others. The lesson of this experience is that about the effectiveness of public sector interventions (e.g., drastic changes in pricing and distribution policies in favor concern that subsidies result in high cost public sector of increased private sector involvement at the expense of monopoly of importation and internal distribution, that they a public sector presence, in contrast to attempts to do not reach their intended beneficiaries, and that they introduce pluralism in institutional arrangements at a more cause wastage and misallocation of resources) have led deliberate speed, often hinder input use. The implications donors to conclude that the costs of fertilizer interventions of this phenomenon for future donor assistance in the areas outweigh their benefits.' Many donor-supported liberaliza- of fertilizers and other inputs are explored in this paper. tion programs have therefore tended to remove subsidies and promote the role of the private sector in fertilizer Policy Assessment importation and internal distribution.8 Donors have also It is clear that fertilizer policy and its reform must be argued that the benefits of improved internal terms of assessed from a perspective that includes not only the trade for agriculture, resulting from measures such as budgetary issues that gave rise to the initial impetus devaluation, reduction of export taxes, and liberalizatiOn of toward reform, but the entire range of factors affecting agricultural commodity markets, would more than offset the fertilizer use in the broader developmental context. This increased costs to farmers of higher fertilizer prices. perspective and the set of factors it embraces have been 7 well articulated by Desai in his analysis of fertilizer There are five main sections in the paper. The first, consumption in India: Fertilizer Use in the MADIA Countries, provides an overview The pace of growth in total fertilizer consumption... of fertilizer use and the effects of price and nonprice lisl also ... .governed by the processes that convert the constraints in the six countries under discussion. It also potential into actual farmers' demand. This would examines three issues concerning fertilizer use that have a include development of an adequate and efficient special bearing on the formulation of overall agricultural distribution system, efforts to promote fertilizer use policy: regional concentration, size of farming operations on different crops, and increased availability of (i.e., large-scale/estate, small-scale commercial, and sub- fertilizer through domestic production and imports. sistence), and use on food vs. export crops. The second Thus viewed, incomplete diffusion of fertilizer use on section, Policy Regimes and Growth of Fertilizer Consump- all land where its use is potentially profitable should tion, begins by focusing on the impact of subsidies, not be considered as resulting only frorn time lags in budgetary costs, and foreign exchange shortages on fertil- farmers' demand, caused by changes in agroeconomic izer consumption. In connection with the foreign exchange variables. It is equally important to determine factor, it also discusses the drawbacks of reliance on ad hoc, whether the time series of total fertilizer consumption short-term, tied aid; the consequent need for long-term ... lisi influenced by the ways in which fertilizer import support to address financial and institutional distribution, promotion, and supply systems ... larel constraints on fertilizer use; and finally the important role developing (Desai 1982, p. 55). played by development projects in addressing institutional By employing such a perspective, it is possible to identify constraints and promoting the diffusion of fertilizer. the constraints hindering increased fertilizer use, and The third section, Policy Reform in the Context of Past thereby enabling the adoption of more effective fertilizer Performance, examines each MADIA country in detail for policies. Toward this end, this paper analyzes the patterns the extent to which the fertilizer reform programs have of fertilizer use during the 1970 to 1987 period in the six addressed the constraints that prevent small farmers from MADIA countries-Kenya, Malawi, Tanzania, Nigeria, Camer- increasing fertilizer use. To assess the impact of subsidy oon, and Senegal-which collectively account for 40 percent removal and currency devaluation on the economic returns of the population of Sub-Saharan Africa and 50 percent of to fertilizer use, the succeeding section, Determinants of its gross product. These countries represent almost all of the Economic Benefits of Fertilizer Use, presents an the ecological zones and grow most of the major crops of extensive analysis of the benefit-cost ratios obtaining in the Africa. Further, despite differences in their production MADIA countries, linking the results to the special issues of possibilities, technological sophistication, policy regimes, regional concentration, types of farming operation, and and institutional arrangements, there are enough similari- types of crops. The final section summarizes the paper's ties among these countries to allow lessons to be drawn findings, draws conclusions, and enunciates the main from their comparative experience. implications for the formulation of future policy. 8 Fertilizer Use in the MADIA Counties Overview Kenya has been a marginal exporter or importer of maize" The fundamental role of fertilizers in overall development (Cleaver and Westlake 1987). (See Appendix 2 for ratios of strategy, and the way the lack of such an overall strategy producer prices to international prices for smallholder affects fertilizer use can only be appreciated by a compre- crops in the MADIA countries.) hensive analysis of the various macroeconomic, sectoral, With regard to nonprice constraints, Kenya has had a institutional, and technological factors affecting fertilizer strong tradition of agricultural research for smallholder use, which can be divided into price and nonprice varia- crops, especially those for expott, and a well-developed bles. A comparative overview of these factors as constraints institutional setup to service the smallholder sector. is presented in Table 5. This typology shows that fewer Reflecting these policies, Kenya has had the best agricul- constraints relating to price policy have operated in Kenya tural performance among the MADIA countries, showing a than in the other five countries. Kenya has avoided currency strong growth in the production of food and export crops, overvaluation, significant taxation of its export crops, and growth in which the share of smallholders has increased fertilizer subsidies. It has also, by and large, linked the over time. domestic producer price of maize to an average of export Although Kenya's overall agricultural policy framework has and import parity prices-a reasonable policy given that therefore been sound, it nonetheless faces complex pricing, Table 5 Factors affecting fertilizer use in MADIA countries, 1970-87 Kenya Malawi Tanzania Cameroon Nigeria Senegal Price factors' 1. Currency overvaluation2 N N Y N3 Y4 N3 2. Explicit fertilizer subsidy N Y ys y y y 3. Explicit output taxation N Y Y y Y6 y Nonprice Factors' 1. Budgetary restrictions N, y Y7 N y7 Y7 2. Foreign exchange restrictions Y N8 Y Na N8 Y 3. Institutional instability N N Y N Y Y 4. Credit availability N Y Y Y Y Y 5. Input import licensing restrictions Y N Y Y Y Y 6. Crop response to fertilizer N Y9 Y N Y9 Y 7. Domestic marketing systems for fertilizer Y Y Y Y Y Y ' A "Y" in a cell indicates that for most of the period under consideration the factor acted as a constraint on fertilizer use, whereas an "N" indicates that the factor was not, generally speaking, a constraint. 2 Recall that currency overvaluation results in an implicit taxation of exports and an implied subsidy on imports, which consist mainly of food crops. 3 The issue of currency overvaluation in Cameroon and Senegal is complicated by their participation in the CFA zone and the resulting link to the French franc and inability to devalue. A purchasing power comparison suggests that there is overvaluation of about 10-20 percent in each country, although the entire issue of the extent of overvaluation of the CFA zone is highly controversial. I The Nigerian currency was overvalued prior to 1986, when the establishment of the second-tier foreign exchange market ISFEM) led to the devaluation of the naira by 400 percent, i.e., the exchange rate changed from N = $1, to N 4 = $1. The naira has depreciated further since then, with the exchange rate in May 1989 being N 8 = $1. Therefore, since 1986 currency valuation has not been a constraint for export crop pricing. In the case of food crops that were above world market prices, the level of internal prices have obviously not constituted a constraint, although their seasonal and year-to-year variability has. 5 The present subsidy on fertilizer is the result of not adjusting local prices for the devaluations of the Tanzanian shilling, and due to the fact that grant aid fertilizer is not priced at its economic cost. Therefore, while the subsidy does not result in a direct payment by the government, there is an opportunity cost of revenue forgone. The combined impact of the devaluation on the c.i.f. price, and the general price increase on internal costs not being reflected in the prices the farmers pay, has resulted in a subsidy estimated to be between 60-66 percent (Rioseco 1989; Carr 1989). 8 Until about 1979, the producer prices paid by the Nigerian marketing/commodity boards for export crops included an explicit tax. Thereafter, however, a subsidy was provided in order to partly compensate producers for the overvalued exchange rate. The commodity boards were abolished in December 1986 under the structural adjustment reforms, and export crop marketing was privatized. Markets in Nigeria for food crops, which account for most of the fertilizer used, have always been dominated by private trade, although unpredictable imports and import restrictions have caused uncertainty. While the government has had a policy of providing minimum support prices for grains, it has largely been ineffective because market prices have usually been much higher than the support prices (see Lele, Oyejide, et al. 1989). 7 Budgetary pressure is relaxed as the subsidy is gradually removed in Cameroon. In Senegal, the government does not bear a subsidy as of 1986. In Nigeria, the budgetary constraints became important after 1981, with the softening of the oil market and the decline in government revenues. 8 In Malawi, the foreign exchange constraint has been addressed directly by an IFAD/IDA-funded agency (Smallholder Fertilizer Revolving Fund) which has guaranteed supplies of foreign exchange. Nigeria and Cameroon did not face foreign exchange shortages until 1981-82 owing to oil exports; recently they have begun to experience foreign exchange shortages. 9 For local varieties of maize, low response coefficients are a serious impediment to increased fertilizer use. For hybrid varieties, response coefficients are not a problem, but the hybrid varieties are not popular for household consumption for several reasons. See Kydd (11989) for an analysis of smaliholders' preference for flint maize varieties. In Nigeria, fertilizer responsiveness under mixed cropping circumstances is questionable. It is not known how well hybrids and composite varieties grown in mixtures respond to fertilizers, e.g., for sorghum and millets. 9 technological, and institutional challenges to ensure equi- Senegal, Nigeria, and Tanzania have all shown poor table and sustained growth in fertilizer use on smallholder growth performance, with exports stagnating at best and agriculture and indeed, growth of fertilizer use in the small- food crop production not keeping pace with the growth of holder sector is much less impressive than smallholder population. In each of these cases, this is a result of agricultural performance as we will show below.'2 A much inappropriate macroeconomic and sectoral policies. Inter- more complex set of problems faces the other countries estingly, while fertilizer consumption has stagnated in where policy distortions have taxed exports either explicitly Senegal and Tanzania, it has soared in Nigeria which has (i.e., in Malawi, Cameroon, and Senegal) or implicitly had the highest subsidy and rate of growth in fertilizer use. through overvaluation of the currency (i.e., in Nigeria and This growth of fertilizer use may have averted Nigerian per Tanzania). As a result, producer prices have moved in favor capita production from falling more rapidly. of food crops, relative to export crops, although subsidies Figure 2 (and Appendix 3) shows the trends in fertilizer have led to the increased use of fertilizer in some consumption in MADIA countries between 1970 and 1987.13 countries. Table 6 shows the changes in the average levels of fertilizer After Kenya, Cameroon has the next best record of use per hectare of arable land for the years 1970, 1975, 1980, agricultural performance, and one that is related to the and 1985. (In Tanzania's case, the higher average growth rate growth of fertilizer use on smallholder export and food suggested by the increased per hectare use of fertilizer is crops. This is in contrast to the relatively small role that in contrast to that derived from the estimates based on increased fertilizer use has played in the growth of small- government import/production data in Figure 2. It illus- holder production in Kenya. trates the data problems surrounding an analysis of Although Malawi's overall growth in export crop produc- fertilizer use. The higher trend suggested by Table 6 is in tion and fertilizer use was strong in the 1970s, the conse- all likelihood due to a change in the estimate or definition quences of estate strategy for income distribution have of arable land.) The large differences in growth rates been a source of concern, as have the prospects for between countries and the size of the year-to-year varia- maintaining past rates of growth of agricultural production, tions in use are noteworthy although we need to be with growing subsistence orientation of rural households cautious because some countries start from a small base, and their related inability to undertake risk. Historically, e.g., Nigeria.'4 Nigeria and Cameroon, which are identified production growth has come largely from the estate sector in Table 5 as having numerous problems affecting fertilizer while the smallholder sector has stagnated, and in per use, experienced the most rapid growth (18.0 percent and capita terms production of most food crops has declined 11.7 percent annually). Malawi, although the poorest country over the period covered. We will demonstrate how the in the sample, ranked third (7.7 percent). Kenya, with the increased use of fertilizer in the smallholder sector, while smallest number of apparent pricing policy related con- crucial for ensuring broad-based growth, is hindered by straints, ranked fourth (6.7 percent). Tanzania, on the other biases that favor the estate sector, despite Malawi's hand, had a growth rate of only 2.9 percent between 1974 excellent record of implementation of NRDP which is and 1987. Nevertheless, its growth rate for the 1978-87 targeted on smallholders. period was higher and comparable to that for Malawi. Over Figure 2 Trends in fertilizer consumption in the MADIA countries, 1970-87 East African countries West African countries Thousands of tons of nutrients Thousands of tons of nutrients _D 300 aW- _ 280 26o _ ~~~~~~~~~~~~~~~~~280_ M04 240 (18.0%/) Zw- ~~~~~~~~~~~~~~~~~~220 2DO. ~~~~~~~~~~~~~~~~~200 180- 180/ . 88o 160 140 140 / 12D- ~~~~~~~~~~~~~~~~~~120 108 80 6 (0.8%) *0 - (.Y)___>_ fiB-,0¢ 40 - ...............;;/ 29 -(7.7%o) . . . . . . . . . . . .Ir. 20 ...............7%) 0 X . I . I . I I * I I * 8 I 8 I O I 8 170 72 74 78 78 80 #2 o 8 87 1970 72 74 76 78 80 82 84 86 87 () - Growth Rate . - - Malawi - Cameroon Kenya ........--- Senegal - --Tanzania .. - Nigeria Source: See Appendix 3. 10 Table 6 have been compounded by low, variable, and declining Average fertilizer use in MADIA countries rainfall and frequent droughts. In Kenya, while the share of grant aid fertilizer has also increased recently, inadequate Country Kilograms of nutrients used priority in allocating foreign exchange for the import of per hectare of arable land fertilizer constrained the growth in its use throughout the 1970 1975 1980 1985 1970s. East Africa Demonstrating the importance of a long history of Kenya 23.8 19.9 27.1 42.1 exposure to fertilizers, Kenya continued to rank first on the Malawi 5.2 6.6 14.3 11.4 basis of per hectare use (see Table 6) because of its high Tanzania 3.1 5.9 6.9 7.6 initial level of use and despite its slow recent growth in use. West Africa Malawi ranked second, owing to more recent high growth Nigeria 0.2 1.8 5.7 10.8 rates. Consumption levels in Nigeria and Cameroon rose Senegal 1.7 9.5 3.7 5.5 comparable to those in Malawi, while levels in Tanzania and Cameroon 3.4 2.0 5.1 8.1 Senegal lagged substantially behind. Indeed, Senegal, which had ranked second among MADIA countries in 1975, Source: FAO, Fertilizer Yearbook, 1986. had reached the lowest level in 1985. the 1970-87 period, Senegal showed no significant growth. Intracountry Patterns of Fertilizer Use The lack of a trend in Senegal's case is, however, mislead- Although comparisons of average per hectare use between ing; fertilizer consumption reached a peak of 116,000 tons in countries are valuable despite differences in agroclimatic 1977, after which fertilizer use declined to such an extent conditions and cropping patterns, they mask important that by 1987 it amounted to only about 25,000 tons. It has differences in internal patterns of consumption within each been argued by some (e.g., in USAID) that the peak levels country, comparisons that provide meaningful insights into of fertilizer use in Senegal were due largely to distortions the factors explaining fertilizer use and the implications for caused by a high subsidy on fertilizer. Others have, future policy. To the extent that data permit, there are three however, considered this to have been part of a conscious issues that this section explores: (I) regional concentration policy to promote growth of use (Pieri 1989). of fertilizer use within each country; (2) use by large farms/ The apparent discrepancy between the general policy estates, market-oriented small farmers, and farmers at sub- environment for agriculture and the growth of fertilizer use sistence or below; and (3) the pattern of use on food and is due to the fact that some of the constraints identified in export crops. Table 5 have been more binding than others on the growth There are major differences between countries in each of of use, so that their relative importance has varied greatly these respects, reflecting historical patterns of population among countries. For instance, readily available foreign settlement and land access as well as subsequent policy exchange enabled Nigeria and Cameroon-both oil export- responses. These differences point to the need for the ers-to import relatively more fertilizer than other countries implementation of country-specific policies toward fertiliz- who received a great deal of foreign aid (e.g., Senegal and ers that are integrally related to the overall agricultural Tanzania have been two of the highest recipients of policy, and that take into account the marginal productivity concessional assistance in the developing world), although of modern inputs and possibilities for growth potential on Nigeria and Cameroon have recently begun to experience the one hand, and objectives of regional and interclass severe foreign exchange shortages. Malawi ranked third in equity on the other. Precisely what weight to attach to these growth despite large current account deficits (as a share of objectives varies among countries depending on sociopolit- GDP) throughout much of the period. Since 1983, by which ical and economic objectives. The review of country time Malawi's foreign exchange problems had become experiences below illustrates the consequences of the compounded (see Lele 1989), a 5-year IFAD/IDA-funded weighting of these objectives by governments, in terms of Smallholder Fertilizer Revolving Fund (SFRF) aimed at the extent to which growth and equity objectives were guaranteeing supplies of fertilizer played an important role realized, and their implications for the future. in relieving the foreign exchange constraint (see also Note Whether fertilizers are being applied in each country in 8 to Table 5). However, the revolving fund was established regions where they show the highest physical response in in local currency with the result that several devaluations, contrast to the social return of fertilizer use across countries together with increasing transport costs, eroded the foreign is an issue of considerable interest in a continent where exchange value of the SFRE ADMARC also very nearly agriculture is preponderantly rainfed and few technologies diverted funds earmarked for fertilizers, but World Bank are available. Physical resource endowments, placement of supervision, a government rescue operation, and USAID populations, and transportation routes in each country all helped avert the crisis. Further, donor support for the SFRF have a bearing on this issue, which is of profound impor- ended in 1987 and this support had not yet been reestab- tance for production strategy, food security, export promo- lished at the time of writing this report. SFRF has had tion, and regional equity (see the attached map which several weaknesses in design and implementation detailed describes the agro-ecological zones in the MADIA countries in Annex 7; it nevertheless provides a useful model to and provides a useful reference for the discussions in the follow in other countries. rest of the paper). The weak growth of fertilizer consumption in Tanzania is Whereas areas with higher and more stable rainfall are a result of a complex set of macroeconomic, sectoral, and in typically most responsive to fertilizer use, population particular, institutional factors. As Tanzania has been by far densities in some countries have tended to be greater in the largest recipient of grant aid fertilizers, in addition to areas of lower and more variable rainfall because of a lesser financial aid, it may be inferred that foreign exchange was incidence of human and animal diseases.'5 In areas of low not the main constraint. In Senegal, institutional problems population densities and a high incidence of animal II disease, a shortage of labor and animal traction poses Table 7). problems for intensification as increased use of fertilizers The sources and causes of deficits and their implications increases the demand for labor which requires concomitant for fertilizer policy in a broader context of overall economic policies toward agromechanization, migration, and human and agricultural policy are complex. For instance, the settlements. Also large production surpluses, when they do Central province which has high responses to fertilizer occur, make immense demands on the internal transporta- (average responses, from between 20 kilograms to 25 tion network to distribute surpluses to rural and urban kilograms of maize per kilogram of nutrient) is food deficit areas of high population concentration. Movement of export in maize in both moderate and good agricultural years, as crops to ports similarly makes demands on the transporta- cash crops predominate and farming households choose to tion systems. The problem is especially difficult in the case depend on the market. In the semiarid areas, however, of staple food crops, as the value of food crops tends to be where rainfall is low and variable, and fertilizer use is risky low in relation to their bulk compared to many export and crop responses are low, a large proportion of the crops, leading to high transport costs in relation to their households depend on food markets out of necessity.'8 unit value.16 In areas of low agricultural potential and high Therefore, if a policy of fertilizer use in the areas of high population concentrations, if the transport networks are response is pursued to maximize growth of output-as weak, household food security is particularly threatened.'7 might be essential-then this needs to be combined with Given the high risks in rainfed agriculture and the low price a policy of support for areas of low potential-such as elasticities of demand for food, increasing production stable food supplies and prices, fertilizer subsidies tar- through the use of modern inputs may require a price geted on food deficit households, and investment in support program to maintain producer incentives. Thus, a education and tranportation to facilitate outmigration. (This policy of concentrating fertilizer use in areas of highest assumes that both the average and marginal product of physical response-which is essential for growth-needs to fertilizer in high potential areas is greater than that in low be examined in conjunction with equity concerns, food potential areas. Given the low levels of application even in security, the role of transportation, markets, price support the high potential areas, this is a reasonable assumption.) programs, and risks resulting from low and variable fertilizer Only rapid growth in agricultural production will enable responses across all regions. Complex interactions are financial sustainability of such a two-pronged strategy. rarely taken into account simultaneously in the formulation From this viewpoint, in 1982/83 almost two-thirds of of fertilizer policies as the section on country-specific Kenya's fertilizer use was in the highest response areas of issues illustrates. It reviews the important differences Central and Rift Valley provinces and in parts of Eastern between each of the countries from these various view- province. These areas account for most of the tea, coffee, points in order to explore their implications for the and maize production in the country and contain approxi- formulation of future policy. mately 55 percent of the population. Crop responses to Among the MADIA countries Kenya offers the greatest fertilizers are, however, medium (between 12 kilograms to scope for exploiting regional comparative advantage in 17 kilograms of maize per kilogram of nutrient) in most production due to the high degree of population concentra- other parts of the Rift Valley (Baringo, Kericho, Nakuru) and tion in the areas of good growth potential. As much as 70 in the Western (Bungoma, Busia, Kakamega) and Nyanza percent of Kenya's population is concentrated on 13 provinces (Kisumu, Siaya, and South Nyanza), and they are percent of the land of mostly high and medium potential. lower still in Elgeyo Marakwet, Kajiado, Laikipia, Nandi, and However, the rate of population growth in the marginal Narok in the Rift Valley and Machakos in the Eastern semiarid areas has been faster than the average for the province. Although response data are not available for the country, reflecting a combination of a rapid population Coast and North Eastern provinces, the fertilizer responses growth rate and an increasing constraint on land access in are believed to be very low because most of the land is of better areas. Among the provinces, Western, Nyanza, and low potential. The predominantly maize growing districts in Central have a substantial amount of high and medium the Rift Valley (Elgeyo Marakwet, Kericho, Nandi, Trans potential land. But the relatively high population densities Nzoia, Uasin Gishu), Eastern province (Embu, Machakos) in these provinces result in very low per capita arable land. and in the Western province (Bungoma and Kakamega) The per capita arable land available now in Kiambu is 0.21 comprise the very few that are projected to be maize hectare, 0.22 hectare in Kisii, and 0.25 in Kakamega. By the surplus in the year 2000. year 2000, this is expected to decline to a minisule 0.09 Growth in fertilizer use in Kenya between large-scale and hectare, 0.10 hectare, and 0.11 hectare, respectively. In the smallholder agriculture over time is of immense interest other provinces, such as Eastern and Coastal, the land but such data are not available suggesting that this available per person is much higher (see Table 7), but the important information may not have been used in policy land quality is poorer. (Although maize is grown throughout formulation toward intensification of smallholder agricul- Kenya, over 50 percent of the crop is cultivated in the Rift ture. What data exist suggest, however, that smaliholder Valley and Western provinces.) fertilizer consumption rose from 15 percent of total use A recent study on grain marketing in Kenya indicates that during the early 1970s to about one-third (36 percent) of with a 2.5 percent overall growth in the yield of maize, a I total use in 1985/86. Reflecting this trend in fertilizer use, percent area growth in Central, Nyanza, and Western smaliholder production in Kenya has made a long-term and provinces, and a 2 percent area growth in other provinces, diversified contribution to domestic food and export crop even some of the currently maize surplus districts such as production. Nevertheless, productivity of the estate/planta- Kirinyaga, Siaya, and Busia will become maize deficit by the tion sector (defined in terms of yield per hectare) has year 2000 (these projections are based on the assumption increased more rapidly than in the smaliholder sector, due of a moderate year). Meanwhile, in most of the other in part to greater use of all inputs (Lele and Agarwal 1989). districts that are already in deficit, the extent of the deficit The predominance of estate use of fertilizer has been will on average increase more than three to four times (see facilitated by the recent liberalization of fertilizer imports in 12 X a S. e . ° V-ne I~~~~~~~ I~~~~~~~ o M 96LLt asiu Table 7 Maize deficit and surplus areas by province and district in Kenya, extent of high and medium potential land, and crop response to fertilizer Maize Balance' Per capita Maize response ('000 MT) Total percent Arable land to fertilizer Moderate Year Good year high and medium Hectares/person Kg/per kg Province/District 1980 2000 1980 2000 potential land 1979 2000 of nutrient2 Nairobi -79.82 -224.65 -79.82 -224.65 Central Kiambu -46.73 -169.25 -34.34 -151.79 51% 0.21 0.09 23.6 Kirinyaga 5.84 -7.17 19.10 18.36 66% 0.33 0.14 20.2 Muranga -35.49 -125.27 -21.29 -97.69 73% 0.28 0.12 24.6 Nyandarua -8.38 -32.24 0.18 -16.94 56% 0.89 0.39 21.1 Nyeri -33.93 -97.93 -24.22 -80.39 42% 0.33 0.14 NA Subtotal -118.69 -431.26 -60.57 -328.45 56% 0.34 0.15 Coast Kilifi -21.44 -71.57 -5.91 -41.47 10% 1.65 0.72 NA Kwale -29.43 -60.10 -26.50 -69.80 25% 2.54 1.11 NA Lamu -3.02 -10.22 -2.02 -7.40 60% 13.04 5.72 NA Mombasa -34.32 -79.85 -33.99 -79.20 - 0.00 NA Taita/Taveta -6.67 -24.13 -1.11 -14.7 34% 3.96 1.74 12.0 Tana River -8.18 -29.74 -6.59 -26.6 41% 9.25 4.06 NA Subtotal -103.06 -275.61 -76.22 -239.24 12% 2.56 1.12 Eastern Embu -11.37 -11.37 -3.99 14.84 29% 0.76 0.34 24.6 Isiolo -2.18 -3.03 -0.86 1.49 - 0.00 NA Kitui -36.98 -38.89 -29.04 0.65 10% 4.32 1.90 NA Machakos -22.91 -3.73 53.51 234.70 26% 1.10 0.48 9.0 Marsabit -9.01 -28.33 -7.80 -24.20 - 0.00 NA Meru -34.17 -20.63 -16.38 40.19 29% 0.64 0.28 16.9 Subtotal -116.62 -105.98 -4.57 267.47 7% 1.42 0.62 North Eastern Garissas -2.37 -8.77 -2.37 8.76 NA Mandera -1.96 -3.74 -1.97 -3.75 NA Wajir -2.62 -7.88 -2.62 -7.88 NA Subtotal -6.95 -20.39 -6.95 -20.40 Nyanza Kisii -13.35 -65.20 -0.12 -38.52 88% 0.22 0.10 14.5 Kisumu -35.77 -87.48 -33.12 -82.61 76% 0.33 0.15 17.2 Siaya 3.81 -26.68 23.15 6.64 81% 0.43 0.19 15.6 South Nyanza -1.43 -35.30 18.00 2.78 72% 0.55 0.241 16.5 Subtotal -46.74 -214.66 7.92 -111.71 77% 0.38 0.17 Rift Valley Baringo -18.76 -43.21 -16.42 -40.29 20% 3.53 1.55 16.0 Elgeyo Marakwet 21.32 35.83 33.91 51.55 48% 0.98 0.43 12.1 Kajiado -8.31 -40.49 -5.50 -34.33 2% 2.23 0.98 12.0 Kericho 44.77 81.85 72.94 144.90 85% 0.53 0.23 17.0 Laikipia -5.10 -28.98 0.63 -19.19 14% 6.01 2.64 10.0 Nakuru -8.49 -24.38 0.57 4.39 46% 0.72 0.31 14.0 Nandi 99.13 177.40 127.27 70% 0.64 0.28 12.9 Narok -10.71 -52.23 -6.49 -44.00 34%- 5.65 2.48 10.0 Samburu -9.71 -17.09 -9.37 -16.39 - 0.00 NA Trans Nzoia 98.21 183.09 121.01 236.88 75% 0.60 0.26 NA Turkana -20.80 -20.62 20.77 -20.68 - 0.00 NA Uasin Gishu 43.72 74.70 52.10 93.83 82% 0.92 0.41 NA West Pokot -2.54 -50.14 1.78 -43.54 15% 3.06 1.34 22.1 Subtotal 222.73 274.73 350.38 545.09 15% 1.55 0.68 Western Bungoma 28.63 53.53 43.80 88.72 65% 0.40 0.17 15.4 Busia 0.08 -26.05 8.69 -9.63 83% 0.45 0.20 18.2 Kakamega 43.58 101.39 85.37 198.23 73% 0.25 0.11 16.4 Subtotal 72.29 128.87 137.86 279.33 72% 0.32 0.14 Total -176.86 -868.95 268.03 167.431 2% 0.96 0.42 '15 percent deducted for fodder and losses. Assumes 2.5 percent overall yield growth distributed in accordance with districts' growth potential. Area growth 1 percent in Central, Nyanza, and Western provinces, otherwise 2 percent. Minor errors in line. Maize response to fertilizers are average responses to Nitrogen derived from FAO Trials conducted between 1 969-73. Some have expressed doubts about the districtwise maize balance results in this table. For instance, G. Stern observes, ". . . Machakos production fluctuates between feast and famine depending on the weather, but it is hard to believe that in a favorable year, by 2000 its surplus would be second in the country and very close to first.... Kakamega data [are also] surprising. At one time, the district (called North Nyanza) included Busia and Bungoma, and it was Bungoma that generated major surpluses.... [it is] hard to believe that Kakamega with some of the most densely populated areas could generate sizeable surpluses. One can divide the district into the heavily populated South that will be as or more food deficient than Kiambu district; a reasonably self-sufficient, fairly heavily populated center and a potential surplus, less densely populated North. The surpluses in the North could not do more than meet the deficit of the South" (Personal communication with the authors). 2 Average response to 40:40:0 or 60:60:0. Source: IFDC 1986; Maize Balance and Population Data: Githongo & Associates 1983. Agricultural Land Statistics: Farm Management Handbook of Kenya, Vol. ll, as reported in ISNAR. 14 Kenya, as estates are better able to organize and finance the marketing board ADMARC at prices that are one-third imports. Smaliholders, particularly those in the already to one-half of those received by estates for those crops that organized institutional network for tea and coffee have also they are allowed to grow. Smaliholders are charged lower benefited from liberalization policies. On the other hand, prices for fertilizers, and while the fertilizer is more easily for a large share of households outside this network of accessible in the customary areas through an impressive commercial agricultural services, owing to their lack of network of ADMARC distribution points, quite ironically this access to credit for smallholders, effective agricultural ease of access enables the estates to purchase fertilizer extension, and especially due to poor technologies, these from ADMARC distribution points. Within both the estate farmers are unlikely to benefit greatly from liberalization if and the smaliholder sectors, there is dualism, i.e., large and the early reports detailing the liberalization experience are small estates, more commercially-oriented, and below of any guide. subsistence smaliholders (Lele and Agarwal 19891. Contrary In Malawi, as in Kenya, population pressure is greatest in to the usual belief, the private sector network catering to the Southern region. Fertilizer use among small farmers has, estates is poorly developed and fertilizers processed by however, been the highest in the Central region where Optichem for estates through a South African monopoly farms are relatively large and the better-off smallholders invariably cost more than the well-run SFRF for small- have been able to organize in groups to gain access to holders.20 While fertilizers distributed for maize are differ- institutional credit. In a separate paper Lele and Stone ent from those used on tobacco, estates do not seem to (1989) have argued that the Boserup hypothesis, which mind the application of the "wrong" fertilizers. There is no suggests that intensification will proceed in the areas of monitoring of the sale of fertilizer intended to benefit high population densities, does not hold in Malawi, where smaliholders, and in all likelihood it is the smaller estates only regressive intensification occurs. In 1986/87, 72 percent that benefit from the fertilizers given to the relatively more of the fertilizer consumed by smallholders in Malawi was commercial small farmers on credit. Clearly much can be used in only 2 agricultural development districts (ADDs), done to ensure that only smallholders qualify to purchase Lilongwe and Kasungu, by about 25 percent of the total fertilizers although its resale to estates cannot be avoided, number of smallholders. In the Southern region, where half unless the relative incentives between the two sectors of the population lives (including most of the 55 percent of increase returns in the smallholder sector. the households with less than I hectare of land), only 11 Given this pattern of sales in Malawi together with slow percent of the nation's total fertilizer use occurred in 1987/ growth of hybrid maize adoption, it is not surprising that 88. average maize yields have been stagnant in the smallholder In Malawi, estates produce 95 percent of exports, but sector and nearly 70 percent of the smallholder area is cover less than 20 percent of the area under cultivation. devoted to the production of maize alone. Only 5 percent Estates mostly in the Central region, where much of the of smallholder area is under hybrid maize in Malawi burley and flue-cured tobacco estates are located, reported compared to 60 percent in Kenya, where unfortunately the nearly 40 percent of fertilizer use in 1987/88, with the production of fertilizer use has not progressed among small remaining 60 percent used by smallholders.'9 Between farmers at the rate it should. Yields per hectare on 1972/73 and 1987/88 reported sales of fertilizer on estates Malawian large estates, on the other hand, have tended to increased at an annual rate of only 4.5 percent, while sales be considerably higher than on small farms as a result of in the smaliholder sector increased at a rate of 11.7 percent the greater use per hectare on estates of other complemen- annually. These estimates, however, in all likelihood under- tary inputs, and reflecting the higher marginal value product state estate use as there are leakages of subsidized of fertilizer on export crops (Lele and Agarwal 1989). A fertilizer from the smallholder sector to estates. Estate further dualism within the smallholder sector is creating a production of tobacco during 1972/73 to 1987/88 increased major future challenge for supplying with fertilizer the from 15,677 metric tons to 51,220 metric tons and that of tea nearly 75 percent of smallholders who sell labor and and sugar increased by 5 percent and 12 percent annually depend on the market for food. Policy debate between the whereas (as stated earlier) smallholder production stag- government and donors is already shifting from a general- nated. In the context of structural adjustment lending ized targeted subsidy to the smaliholder sector as a whole (discussed below), which has involved removal of fertilizer to one targeted explicitly to the poorest of the small- subsidies, there has been a major dispute between the holders. Clearly complex fiscal, political, administrative, government and donors about the extent of leakage to management, welfare, and growth issues are raised by this estates. Estimates in studies funded by the World Bank and stratification in which lack of dogmatism and emphasis on USAID have put the leakage to estates at approximately 25 learning by doing will be critical for donors and the percent of total sales to smaliholders (Nathan 1987). government, as the experience with the reform program has Government has argued that leakages are much lower but well illustrated. owing to lack of survey data there is no reliable information In land-abundant Tanzania-where pockets of population available. The central importance of this issue in the presssure nevertheless are significant (Lele and Stone formation of a fertilizer policy will become evident from the 1989)-government policy has been to encourage produc- discussion of the reform measures, especially the question tion in the less populated high potential areas of the of justifying a subsidy for the smallholder sector including Southern Highlands. This is in part to relieve pressure on its budgetary implications. the more densely populated traditional areas of food and Several factors explain the phenomenon of leakage. The export crop production in the Northeast Highlands and the most fundamental is that policies have affected relative Lake Victoria Basin and in part due to the politics of returns to the application of fertilizers differently between regional balance, which has caused substantial discrimina- the estate and the smallholder sectors. Estates are allowed tion against the traditional cash cropping areas in a general to sell their produce at near world market prices in context of no or slow growth in fertilizer use. Although auctions, whereas smallholders are required to sell theirs to macroeconomic policy distortions that discriminate against 15 export crops also hurt export crop producers as a group tions, market-oriented commercial small farmers who (especially until 1987 when the Tanzanian shilling was produce coffee, cocoa, cotton, and vegetables, and subsis- devalued), had fertilizer consumption been growing rapidly tence farmers. The smallholder enclave schemes have been across the board, regional diversification in use would have the primary focus of fertilizer promotion through donor- been welcome although transport poses problems. funded development programs. Subsistence food crop In 1986/87, about 70-75 percent of total fertilizer use was producers have received little fertilizer. The data on in four regions in the Southern Highlands (Iringa, Mbeya, fertilizer use on plantation agriculture are not available for Ruvuma, and Rukwa), where use increased from 35,000 tons estimating the trend in fertilizer use on crops grown there. of product in 1975 to 91,500 in 1986/87, even though only 18 In Senegal and Nigeria, there is divergence between the percent of the population lives in these four southern areas of high population densities, and the areas where regions.2 In contrast, in the five northern regions (Arusha, fertilizer use is potentially most profitable. In Nigeria, the Mara, Mwanza, Shinyanga, and Kigoma), where the majority more urbanized Southern rain forest zone, with almost 50 of food and export crops were traditionally grown and percent of the population but less than one-fourth of the where nearly a third of the population resides, fertilizer use arable land area, has the highest population densities.22 had declined from 22 percent (20,300 tons of product) in Nevertheless, because of its heavy rainfall which militates 1975 to less than 10 percent (10,700 tons of product) in against fertilizer use, and large areas of soils that respond 1986/87. Also reflecting the effects of a discriminatory poorly to fertilizer, it accounts for only about 10 percent of macroeconomic policy environment toward export crops, all fertilizer use in Nigeria (Lele, Oyejide, et al. 1989). The this shift in the pattern of regional consumption reflected politically influential but resource poor semiarid North, the growing importance of fertilizer use on food crops and where agriculture is the mainstay of the economy and per the declining use on cash crops, for example, coffee and capita income is lowest among the three regions, accounts cotton (FAO and World Bank 1987). Accordingly, and in for about two-thirds of the total cultivated area and combination with a policy effort of pan-territorial pricing of fertilizer use in Nigeria. Fertilizer responses are, however, maize and more recently of giving premiums rather than lower in the North than in the Middle Belt, which has higher discounts to certain distant areas, there has been a shift in rainfall and moderately better soils but has the lowest the areas of marketed maize production away from the population densities among the three regions. Consistent densely populated regions in the North to the less densely with this, the per hectare use of fertilizer in the Middle Belt populated Southern Highlands. Given the proportion of (54 kilograms of product), is one and three-quarters higher population concentrated in the Northeast Highlands and than in the North (31 kilograms), and about two and one- the Victoria Basin, the growing population pressure, and half times higher than in the South (23 kilograms), although good agricultural potential for food and export crops in the total use there is low. these areas, there is a clear need for increased intensifica- In Senegal there is currently no evidence of a systematic tion of agriculture in these regions, the growth of which effort to intensify production in the context of an overall could finance development of other regions in much the agricultural strategy. The problem lies in part in the same way as has occurred in Kenya. The World Bank extremely limited production possibilities and their degra- Agricultural Reports in 1973 and 1983 stressed this point, dation over time. The assumptions made in the 1970s about but as has been documented elsewhere, this knowledge the loss of export market prospects for Senegal's groundnut had relatively little effect on the projects the World Bank or production has become a self-fulfilling prophecy. Without a the other donors financed in the 1970s, as the idea of clear production policy toward groundnuts and emphasis promoting areas of high productive potential (congruent on their extraction, Senegal's groundnut (and sorghum/ with population densities) conflicted with government millet) production has stagnated and it has lost shares in policy of achieving regional equity (Lele and Meyers 1987). export markets for groundnuts and related products (Lele An interesting feature of Cameroon, which also applies to 1988). The growing rice imports, on the other hand, resulted Tanzania, is that despite its large physical area, population in a policy of economic diversification, i.e., import substitu- is concentrated in small areas of the country (Lele and tion of rice through irrigation. This strategy has had an Stone 1989). In Cameroon, 73 percent of the population important regional dimension in which the role of the resides in 33 percent of the area. According to a govern- Groundnut Basin has declined and that of Fleuve in the ment estimate, about 80 percent of the cultivable land is North has increased, while Casamance and Eastern Senegal still underutilized. Population densities are the highest in which are both low in population densities have been the Far North, Littoral, Northwest, and Western provinces, economically and politically marginal. These developments constituting nearly 57 percent of the total population in 18 have been reflected in the shift in the pattern of fertilizer percent of the total arable land. Unlike the other regions use in the country. where rainfall is high, ranging between 1,500 and 3,000 Almost 50 percent of the total population (and nearly 65 millimeters, in the Far North rainfall is only 400-800 percent of the total rural population) lives in the Groundnut millimeters. But nearly 59 percent of the total rice produc- Basin, only 9 percent in the Fleuve, 14 percent in Casa- tion and 41 percent of the cotton production in the country mance, and 6 percent in Eastern Senegal (see Table 8). The is in the Far North. In the Littoral, Northwest, and Western growing population pressure in the Groundnut Basin, provinces coffee and maize predominate. In 1985, 55 together with declining rainfall has weakened the ecological percent of the fertilizer sold in Cameroon was in the base of the Basin, especially in the absence of successful Western and Littoral provinces, the principal coffee growing intensification efforts. The policies that are pursued for the areas, which constituted 75 percent of the subsidized development of agriculture and regional equity, and the fertilizer. About 26 percent was sold in the Northern role fertilizer will play in this process is of fundamental province, where cotton is grown; these sales included only importance for the future in a country with by far the 8 percent of the subsidized fertilizer. poorest set of agricultural resources and the greatest risks Cameroonian agriculture is divided into industrial planta- in production-even by the generally low standards of 16 Table 8 Population, rainfall, arable land, and cropping pattern by region in Senegal Region Rainfall Population Population Arable land Unused but Regional share of density (as % of as % of potentially area cropped (%) per/sq km total) total cultivable' Groundnut MilletSorghum Rice Maize Thies (350-600 mm) 130 13 4 1 Diourbel (350-600 mm) 115 8 3 0 782 772 - 252 Louga (Up to 350 mm) 17 8 25 33 Kaolack/Fatick (Sine-Saloum) (600-800 mm) 54 20 18 10 St. Louis (Fleuve) (200-300 mm) 14 9 17 2 - 23 Dakar/Cap Vert (350-600 mm) 2673 23 0 11 - - - Ziguinchor/Kolda (Casamance) (More than 800 mm) 31 14 14 16 13 10 70 42 Tambacounda (Eastern Senegal) (More than 800 mm) 6 6 18 2 8 8 29 14 I Potentially Cultivable defined as "Terres inutilisees et susceptibles d'utilization agricole ou foresti6re." 2 Share of area cropped for the Groundnut Basin as a whole. Source: Situation Economique du Senegal 1982, Direction Statistique, et Rapport Annuel Direction Eaux, Forets et Chasses, 1978. rainfed agriculture in the MADIA sample. This makes the Tanzania, this regional diversification would not be a pursuit of a systematic long-run agricultural strategy that problem-indeed, given the higher productive potential of clearly takes into account the implications of the high risks the latter regions referred to above-such diversification in agriculture fundamental for achieving any growth at ali in would be welcome-except that it has been associated the future. Donors have, however, not taken into account with the serious decline in Senegal's overall fertilizer use the importance of risk as a factor of policy formulation in since 1977 and its collapse in the Groundnut Basin. Pieri Senegal, nor have they helped provide stability to the observes, for instance, that fertilizer use on groundnuts in complex and otherwise ecologically unstable production Sine-Saloum went down from 30 kilograms per hectare per environment-a situation caused in part by the govern- year during 1960-69 to 23 kilograms in 1970-79. Over the ment's own haphazard policy toward agriculture. same period, fertilizer use on millet/sorghum went up from About 80 percent of the groundnut and millet/sorghum 13 kilograms per hectare to 23 kilograms per hectare and a quarter of the maize production in the country takes because of farmers' concern for food security resulting from place in the Groundnut Basin. The climatic and soil the growing pressure on land. Per capita cultivated area conditions are most favorable in Kaolack/Fatick (the former declined from 1.35 hectares in 1960 to 0.85 hectares in 1979 Sine-Saloum) region lying in the southern Sudanian zone. (Pieri 1989, pp. 91-96). Collapse of fertilizer use in the Nearly one-fifth of the total population resides in this Groundnut Basin is of serious concern. But the strategic region, and as much as half of the total groundnut and issues of long-term agricultural development, including the sorghum/millet production of the Groundnut Basin occurs critical role of public policy in ensuring agricultural intensi- here. Rainfall levels are lower in the northern part of the fication has not been adequately explored in Senegal. Basin, where 60 percent of the Basin's population resides, Fertilizer use in the Fleuve has clearly increased in support ranging between 350 and 600 millimeters. Rainfall in this of irrigated rice and vegetables. There are, however, no area has been declining and becoming more variable in the reliable data on the Fleuve to judge its potential as an last two decades. economic user of fertilizers. Sine-Saloum, Casamance, and The Casamance, located in the Sudano-Guinea zone with Eastern Senegal have assured rainfall, and land in these relatively higher and assured rainfall (more than 800 regions is believed to be of higher potential, with a millimeters) is believed to have by far the greatest potential combined population among the three regions of 40 among the rainfed areas, and accounts for 70 percent of the percent. Obviously, priority should be given to all these area under rice in the country, and about 40 percent of the regions in fertilizer consumption.24 However, as in the area under maize. In Eastern Senegal (Tambacoundal, also Middle Belt of Nigeria, intensification in Ziguinchor/Kolda a well-watered region, cotton is predominantly grown along (Casamance) poses problems stemming from labor short- with other food crops (rice, maize, and millet/sorghum). ages; the region also has the disadvantage of being the Political and economic action has, however, led to the most isolated area in the country. Thus without a strategy development of irrigated rice and horticulture crops in the devised simultaneously to address the problems related to Fleuve at high costs. fertilizer use including mechanization and investment in During the late 1970s, nearly 75 percent of the fertilizer infrastructure, priority for fertilizer use will clearly not work. was used in the Groundnut Basin. Fertilizer use has The lack of widespread use of fertilizer among a large become regionally more diversified over time, with a number of small farmers is also of concern. In Nigeria, much movement away from the Groundnut Basin toward Casa- of the new technology available is for sole cropping and, mance, Eastern Senegal, and the Fleuve. Due to population since small farmers tend to intercrop to spread risks, the pressure there has also been a shift in fertilizer application new crop varieties are adopted mainly by the larger, more from export to food crops.23 Again, as in the case of progressive farmers toward whom agricultural extension and 17 fertilizers have been directed (Lele, Oyeiide, et al. 1989), through their macroeconomic and sector policies by shifting although fertilizer application on traditional mixtures have prices and other more subtle incentives in their favor become quite common. should be expected to use increased fertilizer over time on The divergence of population densities and fertilizer use these crops.28 Information on the actual on-farm use of among various regions and sizes of farmers' holdings raises fertilizers is spotty. But insofar as growth of use has been the question of fertilizer use on small and large farms in largely associated with certain regions and, through devel- general and because of its effect on achieving broad-based opment projects, tied to particular crops, that information, growth in agricultural production, in turn, its potential combined with IFDC surveys carried out in specific years, impact on the growth linkages of agricultural development information on relative prices, and the authors' field with the rest of the economy; within the smallholder sector investigations, allows strong inferences on end use. The itself, it raises the issue of the more rapid adoption of general proposition above is supported by the available fertilizer among the relatively more commercially-oriented evidence on use. Figure 3, which depicts the cropwise small producers compared to their counterparts who are pattern of fertilizer use shows that, in Nigeria, the food crop close to or below subsistence.25 sector accounted for over 80 percent of fertilizer use. In From this viewpoint the per hectare use of fertilizer Malawi, as much as 83 percent of the fertilizer used in the among smallholders may be lower in Kenya than in Malawi, smallholder sector was applied to maize alone (see Figure provided that reported use by smallholders in Malawi is 3).29 In Tanzania, maize accounted for nearly half of the total accurate and reflects actual use and not leakages to the fertilizer use. In Senegal, as stated earlier, owing to the estate sector.26 Although in Kenya, unlike Malawi, superior collapse of distribution in the Groundnut Basin, fertilizer price incentives for export crops and the right of small use has shifted from groundnuts and sorghum/millet to rice farmers to grow all crops lead to a significant portion of the and cotton. acreage and fertilizer use being devoted to these crops In contrast, in Kenya two-thirds of the total fertilizer use with farmers relying on the market for nearly 50 percent of (by estates and smallholders) was on three export crops their food.27 Nevertheless, even among these more com- (coffee, tea, and sugar) in 1985/86.30 In Cameroon, similarly, mercial small farmers in Kenya, there is greater scope for nearly two-thirds of the fertilizer use was directed toward further intensification. Smallholder yields per hectare of export crops, chiefly coffee and cotton, with coffee account- export crops, especially tea and coffee, are lower (half to ing for as much as half of the total reported fertilizer use.31 32 two-thirds) than those of large estates in Kenya, and yield In practice, however, field investigations suggest that, differences by size class are greater still in Malawi (Lele and reflecting the high producer prices of food crops in Agarwal 1989). To provide further insights into the precise Cameroon, farmers have tended to apply some of it to food policies that should be pursued it is, however, necessary to crops, especially in the coffee areas of the Western have additional data and information. Highlands. This phenomenon has been at work in Kenya, Countries that have favored the production of food crops too. Figure 3 Fertilizer use by crop in the MADIA countries Malawi 1984/85 Kenya 1982/83 Tanzania 1985 Sugar ~~~~~~~~~~Others 1)6u.8a% 53.0% Others hr 16.5% Tea Maize 83.5% e~~~~~~~~~~az 47.0% Nigeria 1979/80 Senegal 1985 Cameroon 1985 Groundnuis Others Food Crops 2.3% ~~~35.20/o Groundnuis MitleVlSorghum 150 RIC ^ 32 Cotton 3 8.0% Rice 2--- Cotton 11.2%200 Mai'ze orhm Rice% Miiei 3.8 15.8%o Coffee 18 Source: See Appendix 6. Policy Regimes and Growth of Ferdlizer Consumptdon As a prelude to the discussion of the nature of policy Table 9 reform in each country and the likely financial and eco- Rate of explicit fertilizer subsidy in MADIA countries, 1970- nomic profitability of fertilizer use at the farm level, we first 87 discuss the rates of fertilizer subsidies, budgetary costs, and foreign exchange availabilityP3 It is not easy to Year Malawi Tanzania Cameroon Senegal Nigeria Kenya determine rates of subsidies and budgetary impact over (Percent) time because of a number of factors: changes in exchange 1970 NA NA NA 36 NA NA rates and international prices of fertilizer; changes in the 1971 NA NA NA 52 NA NA importance of "aid" fertilizers and the way they are treated 1972 NA NA NA 54 NA NA in the setting of internal prices; changes in the rates of 1973 NA NA NA 52 NA NA explicit subsidies themselves; and finally the informal 1974 NA NA NA 54 NA NA market in fertilizers within each country and even across 1975 NA 75 NA 76 NA NA national borders. Nevertheless, some judgments about the 1977 NA N6A NA 73 85 NA rates of subsidies and their budgetary impact can be made. 1978 NA 50 48 52 85 0 The rates of explicit fertilizer subsidy are shown in Table 1979 NA NA 54 48 85 0 9 and their estimated budgetary costs for the East and 1980 NA NA 53 54 85 0 West African group of the MADIA countries are presented in 1 981 NA 60 58 61 85 0 Tables 10 and I. The extreme case is Nigeria, which has 1982 NA 60 48 46 85 0 had the largest explicit subsidy (85 percent in the early 1983 25.0 60 NA NA 83 0 1980s) as well as a substantial implicit subsidy owing to the 1984 28.6 60 NA NA 50 0 overvaluation of the currency converting subsidies from an 1985 23.4 NA NA NA 34 0 incentive to a liability. In 1985, Nigeria's explicit subsidy 1986 22.6 NA NA 0 28/82 0 constituted 32 percent of the total agricultural expenditures 1987 22 NA 65 0 82 0 by the federal government, and 3.7 percent of total Note: Subsidy in Tanzania was abolished in 1984; however, there is an government expenditures (Tables 10 and 11). By 1987 as a explicit and implicit subsidy due to exchange rate overvaluation and result of the devaluation and the increase in the naira cost the fact that fertilizer is not priced at full cost to farmers. The combined explicit and implicit subsidy is estimated to range between of imports, the expenditure on the subsidy had soared to 60-66 percent in 1988/89. nearly I billion naira. Clearly in a country with substantial Since 1986, the Government of Senegal has stopped subsidizing competing demands on the use of resources-especially in fertilizers. But a limited subsidy is provided by USAID on fertilizers circumstances of declining budgetary revenues following bought for cash. The subsidy from USAID was as follows: In 1986/ the fall in oil prices-the resources allocated to the FA8,0007per ton(FPRI/IFDC 1988)1 fertilizer subsidy must be considered in terms of alternative With the institution of a two-tier exchange rate in Nigeria in October opportunities to create other more permanent assets. For 1986, the actual rate of subsidy was 82 percent. instance small-scale irrigation has had a powerful impact on Source: Malawi: World Bank 1988a; Tanzania: Spurling 1982; diversification of Nigerian agriculture, until very recently. Cameroon: BergaAssciatndes 1987 Nigeria: LeleandBindlish Elsewhere we have documented the importance of feeder 1988. Table 10 Budgetary cost of fertilizer subsidies and their share in total government expenditure in West African MADIA countries, 1974-87 Cameroon Senegal Nigeria %of %of %of %of %of %of Millions Total Agricultural Millions Total Agricultural Millions Total Agricultural Year of US$ budget budget of US$ budget budget of USS budget budget 1974 1.2 NA NA NA NA NA NA NA NA 1975 2.1 NA NA 21.6 NA NA NA NA NA 1976 2.3 NA NA 17.0 NA NA NA NA NA 1977 2.4 NA NA 8.3 NA NA NA NA NA 1978 7.5 NA NA 11.4 2.4 NA NA NA NA 1979 8.0 NA NA 8.0 1.2 NA NA NA NA 1980 9.0 NA NA 18.8 2.6 NA NA NA NA 1981 13.6 NA NA 7.8 1.3 NA 171.1 1.1 24.8 1982 7.6 0.5 NA NA NA NA 196.0 1.7 24.6 1983 15.7 1.0 NA NA NA NA 103.6 1.0 16.1 1984 16.0 NA NA NA NA NA 213.0 3.8 43.1 1985 NA NA NA NA NA NA 240.9 3.7 32.1 1986 NA NA NA - - - NA NA - 1987 23.0 0.9 NA - - - NA NA - 19 Table 11 foreign exchange availability, foreign exchange shortages Budgetary cost of fertilizer subsidies and their share in have destabilized supplies and undermined growth by total government expenditure in East African MADIA encouraging reliance on ad hoc, short-term, tied aid. espe- countries, 1978-87 cially as donor support for fertilizer imports has grown Malawi Tanzania substantially in most countries. This year-to-year support by % of % of % of % of donors, with all of its uncertainties, has not allowed Millions Total Agricultural Millions Total Agricultural governments to plan the expansion of fertilizer use or to Year of US$ budget budget of US$ budget budget develop an administrative capacity for formulating a sound, long-term fertilizer policy, such as that begun by the SFRF 1978 NA NA NA 4.43 2.9 4.1 in Malawi. Ad hoc donor assistance has also caused prob- 1979 NA NA NA 5.0 3.3 4.5 lems of mismatches between fertilizer supplies provided 1980 NA NA NA 7.2 2.3 3.8 by donors and those needed by recipients. This, combined 1981 NA NA NA 8.3 2.0 54 with delays in delivery, shortfalls in quantity, and the 1982 NA NA NA 6.6 3.2 434 added strain on the administrative capacity of governments 1983 NA NA NA 4.7 3.4 3.6 that are already overstretched administratively has made 1985 2.5 1.3 6.3 - - - grant aid fertilizer a mixed blessing. Further, until recently, 1986 1.6 1.0 NA - - - grant aid fertilizer has incurred relatively high transport 1987 1.2 0.7 NA - - - costs because in the absence of location-specific fine-tuned packages, donors have tended to promote low analysis Note: The U.S. dollar cost of subsidies, in terms of actual expenditures by fertilizer which is more expensive to transport. government, was computed using official exchange rates. In The aid effectiveness studies carried out for MADIA Tanzania, the fertilizer subsidy was eliminated after 1984, and in suggest that increased import support results in increased Senegal the government has not directly subsidized fertilizers since possibilities for tied aid and this may have increased the 1986. Estimate. attractiveness of import support among donors (in addition Source: World Bank database and World Bank files. to their interest in macroeconomic policy reform). Under the recently established Special Program of Assistance (SPA) for debt distressed countries in Sub-Saharan Africa, there has roads for integration of agricultural markets and their been some progress in getting donors to untie aid provided present dearth in Nigeria. At the same time, we have also in support of structural adjustment.34 This untying provision pointed out that the past investments in feeder roads have does not, however, apply to project assistance, even if not resulted in the creation of permanent assets because of cofinanced, or to any bilateral assistance given by individual serious maintenance problems (Lele, Oyeiide, et al. 1989). donors. The possibility of untying urgently needs to be Thus, while the choices on allocation are quite clear in applied to a program of long-term import support for principle, in practice without the strengthening of the fertilizers, built around a sound fertilizer policy on a institutional structures at the state, district, and local levels country-by-country basis. Such long-term support is to ensure the creation and maintenance of public goods uneeded all the more as institutional instability, credit needed to improve productivity, fertilizers have seemed to unavailability, import licensing restrictions, and input be the most attractive instrument to the government to marketing systems have also acted as constraints to the transfer the oil bonanza to the North. These softer, more growth of use. All of these problems need to be addressed strategic aspects of smallholder agricultural development on a long-term basis, as the discussion below of fertilizer policy, the sequencing and phasing of investments, and the reform programs will illustrate. capacity to plan and implement must be given more Before turning to that discussion it is important to stress attention in donor assistance. For instance, the World Bank that although long-term fertilizer import or distribution alone has committed $1.7 billion to Nigeria's agriculture but system support perse has not been offered by donors to any the establishment of policy planning capacity has not of the viADIA countries, the institutional constraints that received the attention it needs. The government has in prevailed in the 1970s have been partially offset by the area addition devoted II billion naira to agriculture since the oil and/or commodity-specific development projects that the boom. Expenditures on agriculture increased by 63 times in donors financed. Indeed, the regional and cropwise pat- less than a decade, a large part of which was accounted for terns of use discussed earlier for each country are closely by large-scale irrigation and fertilizer subsidies (Lele, linked to the dominance of projects, e.g., ADPs in Nigeria, Oyejide, et al. 1989). NRDP in Malawi, SODECOTON in Cameroon, and KTDA and Malawi, on the other hand, had the smallest rate of Coffee Development Programs in Kenya. The ability of explicit subsidy among those with subsidies (less than 30 these programs to offer more reliable supplies of inputs, percent), while Senegal, Cameroon, and Tanzania had credit, and integrated services, together with extension explicit subsidies of intermediate rates typically amounting advice, has made a critical difference to the uptake of to about half of procurement costs (see Tables 9, 10, and I1). fertilizer, although subsidies have also made a difference Tanzania also provided a substantial implicit subsidy by influencing the price of fertilizer relative to that of through currency overvaluation. Kenya has not had a outputs-an issue discussed later. This important role subsidy on fertilizers since 1977 and therefore there has played by development projects in fertilizer promotion is been no budgetary cost of fertilizer use to the government. an unexpected conclusion to emerge from the MADIA Budgetary considerations have caused supplies to be study, especially given the generally negative perception of rationed in each country, except Kenya, leading to a major the sustainability and the low rates of return of integrated debate within the context of structural adjustment. rural development projects, which have led to the virtual With the exception of Malawi, where the relatively long- discrediting of the project approach (Lele 1988; World term and stabilizing effects of the SFRF have ensured Bank/OED 1985). 20 This conclusion does not mean, of course, that the for both cash and credit, for use on both food and export project approach has been perfect. It needs substantial crops, meanwhile increasing considerably the information improvement as a complementary two-pronged approach flow available to farmers on the specifics of fertilizer use. to the development of African agriculture which provides The use of the media, including television, needs to be project assistance to ensure the implementation of con- seriously considered for this purpose as an alternative to cerns expressed through adjustment lending. A major the recurrent cost-intensive and ineffective extension problem with development projects, for instance, has been services, which continue to proliferate even in a period of their monopoly on distribution, and even worse their severe recurrent resource crunch. restricting the availability of fertilizer to certain crops, areas, It is clear that making the right types of fertilizers or farmers. Indeed, as pointed out earlier, frequently the available for the right crop at the right time requires a only way for farmers to obtain fertilizer for food crops in substantial additional effort; more imaginative approaches areas where export crop projects have operated has been in Africa, where more refined, location-specific knowledge to acquire membership in such projects and to transfer of the soils and farming systems of small farmers, including fertilizer to other areas and crops. For instance, not only the constraints the more risk-averse subsistence producers have the coffee and cotton producing farmers diverted face, are crucial. Addressing concerns expressed through fertilizers to food crops in Cameroon, but even in Kenya, adjustment lending for this purpose is of course also where tea and coffee prices have been highly favorable to fundamental, as is the increased availability of credit. Here, smallholder producers, a significant portion of the fertilizer suffice it to say that the process of diffusion of fertilizer distributed by KTDA and coffee cooperatives is reported to from the specific crops and areas to which it was originally have been diverted to maize. This diversion has of course targeted to other crops and areas is well along in Africa, led to problems with the use of fertilizers. For instance, in more systematically in some countries and areas than in Kenya and Cameroon the application of the nitrogenous others. It is a pattern that has occurred in Europe, North fertilizers (ammonium sulfate and 20:10:10 as a basal America, and more recently in Asia (Desai 1982). It is in this dressing) supplied for tea and coffee to maize and other context that we now turn to the reform programs for a more food crops that require fertilizers high in phosphates tends detailed examination of the extent to which they address to result in wastage.35 36 This could be averted by making the real constraints facing small farmers in increasing their the projects the vehicle for delivering appropriate fertilizers fertilizer use in each of the MADIA countries. 21 Policy Reform in the Context of Past Performance Kenya including control of supplies and price fixing. Still other Table 12 provides a comparative overview of the types of factors constraining the use of fertilizers include the lack of reforms that individual countries have implemented from credit and institutional support to food crop producers. among the long set of links that constitute the fertilizer These were being addressed by donors as part of the chain in each country. Integrated Agricultural Development Program (IADP), but In Kenya, quantitative restrictions (ORs) on fertilizer were abandoned as unworkable (Lele and Meyers 1987). imports through an import licensing system introduced in Nonetheless these issues need to be addressed and were the 1970s, have been one of the most important constraints not covered by the recent reform program (discussed to the growth of use despite a much more developed below) that has stressed fertilizer importation and distribu- institutional setup for small farmers. These were accompa- tion procedures. Donors, including the World Bank, USAID, nied by "jawboning" by government through the establish- and the Dutch government have already contributed much ment of fixed prices and distributor's margins announced to the understanding of the farm use of fertilizers through together with the Maximum Retail Price (MRP).37 Prior to the important studies38 and have provided financing for fertil- reform measures introduced in 1983, importers used to find izer imports, albeit on an ad hoc basis. The existence of the it unprofitable to import fertilizer because of uncertainty knowledge generated through these studies and past about the availability of licenses, difficulty and delays in lessons of implementation of IADP now need to be obtaining letters of credit, and uncertainty about the MRP, converted into action programs for fertilizer distribution on although farmers were frequently unaware of its existence. a long-term basis through predictable and reliable supplies These regulations led to insufficient levels of fertilizer of foreign exchange with a built-in provision of growth in imports and shortages, especially in areas outside major fertilizer use at 10 to 15 percent per year annually. district capitals where government-determined margins USAID has taken the lead in the reform measures in were insufficient to cover the extra transport costs. Indeed, Kenya, as elsewhere. These reforms have concentrated on our earlier comparisons between per hectare use in Kenya (i) liberalizing import procedures to ensure an increased and Malawi and the evidence of yield differences between number of importers and timeliness of imports based on a smallholders and estates suggest that because of institu- more correct assessment of demand,39 (ii) increased private tional problems, even in high and medium potential areas, participation in the fertilizer distribution networks, (iii) the fertilizer use is much lower than its economic potential. setting of distribution margins to more accurately reflect Much of the good performance in Kenya is explained by a marketing costs (especially in remote areas), and (iv) shift from low to high value crops by small farmers owing bagging fertilizers in smaller quantities to make fertilizer mainly to the lack of restriction on their production and more accessible to small farmers. Since the introduction of sale. This shift now appears constrained by the need of these measures, fertilizer consumption reached 100,000 small farmers to produce more food, as reflected in the tons of nutrient in 1986 from the low level of 75,000 tons in diversion of fertilizers provided for export crops to food 1984-a year of drought-although the increase is less crops (World Bank 1986b). impressive when compared to a peak of 91,300 tons There have been other criticisms of Kenya's fertilizer reached in 1983 (see Figure 2). Also the extent to which the policy, such as allegations of impropriety in allocation of increase is a result simply of increased aid fertilizers licenses and oligopolistic behavior on the part of licensees, provided because of "liberalization measures," as distinct Table 12 Fertilizer policy reforms in the MADIA countries Type of reform Malawi Tanzania Kenya Cameroon Senegal Nigeria Improvement of import licensing procedures Y Y Improvement or privatization of imports Y Y Y Credit to importers y y Privatization of wholesale distribution Y Y Y Privatization of retail distribution Coops, Y Y Y Y Change in retail margins Y Subsidy removal y y y y y Introduction of HAF2 Y y y y Fertilizer in small bags Y y Targeted subsidy Y Notes: A "Y" in a cell indicates that the preferred policy reform is being undertaken or under consideration. I In Tanzania since 1984, the primary cooperatives have been reintroduced and they play a major role in the retail distribution of fertilizers. 2 Introduction of high analysis fertilizer. 22 from greater demand being reflected in more imports of imports by donors to improve import procedures in the private investors is difficult to say. Nevertheless, fertilizer long-term, there are major hindrances to achieving this stocks also increased, thereby improving farmers' confi- objective through donor policies themselves. dence in timely availability of supplies. Perhaps the most difficult and long-standing issue facing A report by Agriconsult (1988) argued that the reform policymakers in Kenya, but one which is acquiring increas- program has addressed many of the relevant institutional ing urgency, is the regional focus of future intensification constraints on increased fertilizer use, by: efforts and the government role in this process. Among the a. eliminating sole distribution rights of grant aid fertil- MADIA countries, Kenya presents by far the largest differ- izers by the Kenya Grain Growers' Co-operative Union ences in terms of fertilizer responses between the high (KGGCU, formerly the Kenya Farmers Association, a potential and semiarid areas Isee section on Response large farmer cooperative); Coefficients below). Given Kenya's rapidly growing popula- b. increasing availability of fertilizer among cooperatives; tion, increasing pressure on land, and decelerating growth c. increasing awareness among small farmers about in smaliholder production, there is a critical need to fertilizers of different types, especially in small bags to increase production in areas of high (and medium) poten- increase its access; and tial by prioritizing them in fertilizer distribution. The later d. increasing the number of fertilizer distributors in discussion on returns to fertilizer use will demonstrate that major market centers. fertilizer application is highly profitable in these areas, given the response coefficients and without subsidies. Also, The same source also points out that for large-scale the fact that smalholder yields are only half to two-thirds farmers the prnvatization and leliza tion zeasures of those of estates and large farms means that there is ment system by granting allocations directly to end users. and ium potentalthrough amon othe high Cooperative cosstn of lag famr (eg,KGU ;n and medium potential areas through, among other things, Cooperatives consisting of large farmers (e.g., KGGCU) and the increased use of fertilizer (see Lele and Agarwal 1989). cash crop producing smallholders (e.g., those in coffee) as Prmtn fetlie us,hwvr.ass raesuso well as the Kenya Tea Development Authority (KTDA), Promoting fertcltzer use, however, raises broader issues of which has large bulk needs, have been able to obtain the public sector role in agricultural development. Increas- licenses to import fertilizer direCtly.40 On the other hand, it ing maize surpluses in the areas of traditional comparative is questionable whether effective competition in imports advantage will continue to require the government to play has increased because of the small quantity of fertilizer that an active role, albeit as a complement to the private sector, Kenya imports. The allocation of import licenses to end chiefly as a buyer and seller of last resort for maize. The Ksenyhas imrts. the alocrationbusines of im licestutoend government, through the maize board, would also have to users has taken th e lucrative busines-f dstributin supplement the movement of surpluses to remote food fertilizer to the large-farm or bulk small-farm users away dfctaes uha iu,Klj,Kae ais.adWlr from the importers/distributors. They have been left with deficit areas, such as Kitui, Kiliji, Kwale, Garissa, and Wayir, the distribution to small-scale agricultural operations-a to ensure food security. segment of the market that is characterized by low and Fertilizer policy can complement a food security policy in variable demand and high unit costs of..distribution, an important way, since the costs of maize price support variable demands ands high unit costs of dsistribution, an.taiizto poice hvbenlreiKnyad preventing possibilities of cross subsidization in operations. donors have sought to reduce government's role in maize Further, the lack of credit for stockists has hindered their mark (aver an restae 1987).mAst a s imple distrib- performance, so that fertilizer is often unavailable in rural areas, resulting in little price competition. The elimination utive measure to improve household and regional food self- of the oligopoly in fertilizer distribution will encourage sufficiency in marginal areas42 and as a way to minimize the private retailers to stock fertilizers. Nevertheless, without cost of the government's role in maize distribution, a an active and meaurblgalfnceaincomplementary policy might actively encourage increased an active and measurable goal Of increasing fertilizer use friie s nmiei hs ra hr,a niae among small farmers as part of a larger agricultural policy, fertilizer use on maize in these areas where, as indicated and without expanding their access to public or cooperative earlier, population is growing more rapidly than average. andswithouti expanding th eirpaccess t oepbli o roer ative The demand for fertilizers in these semiarid areas is still institutions taprmely the use of .fertil prov ide nascent.43 As households have relatively limited access to growth of fertilizer usel credit, effective agricultural extension does not exist, and Closely related to the role of the public sector in Kenya turnover and profits from fertilizer sales are too low to is the growing imprtacefdnorsupliencourage private retailers to become involved in cash a grant basis (which now account for over 60 percent of sales. Even more fundamentally, technological packages fertilizer supplies). Donor presence necessitates that the have not been developed for these semiarid areas. This fenyanlgovernmuien. binvol r presenencesicingst dhei- suggests that total reliance on the private sector may not be Kenyan government be involved in fert r p g dthe most effective way of increasing fertilizer application or sions. Despite its very limited administrative capacity, the food consumption in these areas, even though removal of government must also cope with the complexities of distributive margins and monopolies in fertilizer and food procuring supplies tied to donor sources.4' This has led to distribution might help matters. a continued constraint, as observed by the Agriconsult One way to address the problem while avoiding monop- report: oly may be to subsidize transportation costs of fertilizers In the last two years.. . although the allocations have (by the government delivering fertilizer to important depot been released in time, the performance of importers centers) to encourage private sector fertilizer sales. has deteriorated largely because the approved selling Increased agricultural research and credit availability for prices were announced late, and by the time they subsistence farmers will be crucial ingredients. This were announced, the prices were not commensurate approach might minimize the need for an active public with the world market price (Agriconsult 1988, p. 45). sector food distribution program, except in the years of Later, in the discussion of Cameroon, it will be demon- severe droughts. strated that despite the urgent need for untying fertilizer In sum, Kenya's circumstances require a two-pronged 23 agricultural policy that actively promotes self-sustaining spare parts and railroad wagons has also severely hurt the growth in the high potential areas. In addition agricultural reliability of deliveries. Similarly, the secondary distribution policy must address the need for fertilizers in low potential network, from railheads to local depots, has suffered from areas until other policies are devised to encourage migra- a shortage of vehicles, poor roads, and institutional weak- tion of population out of resource-poor regions or to nesses. Compounding the transport difficulty is the severe promote nonagricultural employment and relieve popula- institutional instability and the lack of coordination that has tion pressure on the environment. characterized agricultural pricing and marketing policy in Tanzania for nearly a decade and a half. These problems Tanzania seem to be deeply entrenched in Tanzania's political Fertilizer policy has not been a central issue in the policy system through the active role the political party plays in dialogue between donors and government in Tanzania policy, almost as an alternative government, and where because serious macroeconomic distortions and institu- responses to reforms are not very practicable. Between the tional instability combined with apparent land abundance late 1960s and 1974, the cooperative unions and the have overshadowed the need for intensification (Lele and Tanzanian Rural Development Bank (TRDB) had primary Meyers 1987; Lele 1989). Nevertheless, the need for inten- responsibility for farm input supply. The rapid expansion sification is increasingly evident. For example, in the and growing economic and political importance of the traditional export crop areas of Arusha, Kilimanjaro, cooperative movement, however, were attended by a Mwanza, Shinyanga, and Tabora regions, the per capita growing inefficiency. Increasingly perceived by the ruling arable land available is already less than one hectare (Lele party (TANU and later CCM) as an alternative center of and Stone 1989). However, with improvement in macroeco- political power, the cooperative unions were dismantled by nomic policy, cooperative and private sector institutions the government in 1974.46 The policy of uiamaa villages have begun to make a comeback which reflects the basic implemented at around the same time was developed for strength of the enterprising Tanzanians. It has important reasons of facilitating the provision of services, but it led to lessons for future fertilizer policy (Lele 1988). excessive concentration of population, accelerating soil The general weaknesses of the Tanzanian public sector degradation by reducing bush fallow, and increasing defor- have pervaded all areas of fertilizer supply, i.e., importa- estation for fuel. Crop authorities established around the tion, domestic production, and distribution, and these will mid-1970s, which purchased output with the newly estab- need to be addressed if the strong private sector response lished ujamaa village cooperatives, were initially given to increased price incentives is to be sustained. responsibility for some input distribution, while TRDB had As part of the country's Basic Industrialization Strategy, responsibility for input distribution in other areas together Tanzania established a fertilizer factory-the Tanzania with the provision of credit. Inadequate coordination and Fertilizer Company (TFC) in Tanga-in 1968, well before lack of clear lines of responsibility between the crop other MADIA countries emphasized such basic import authorities (parastatals), the TRDB, and the village cooper- substitution.44 However, design problems and foreign atives in credit recovery through crop purchases resulted in exchange shortages plagued the factory, affecting the a large accumulation of farmer debt to TRDB. The resulting availability of feedstock and spare parts and leading to disputes among these institutions caused the credit and supply shortages.45 fertilizer distribution system to become completely mori- In addition, the domestic distribution system has suf- bund. By the end of the 1970s, many farmers were willing fered from a weak and deteriorating transport infrastruc- and able to pay cash for fertilizer (as, with the growing ture, which increases the cost and decreases the reliability money supply and shortage of consumer goods, cash of the entire agricultural marketing system. This weakness balances had begun to accumulate in rural households), but is an especially critical one in Tanzania because of its large by 1982 there were only 13 retail outlets in the entire size and the fact that, in contrast to Kenya, the portion of country, most operated by the Tanzania Farmers Associa- the population engaged in crop production resides on the tion (TFA). (The TFA is a large farmer cooperative based in perimeters of the country, while the natural markets and Arusha; see World Bank 1983.) As a growing parallel access to ports are in neighboring countries. Opening up economy controlled much of the rural production, the new areas of production in the South has tended to alter government's frustration with the institutional chaos caused this pattern and make heavy demands on the transporta- by the proliferation of parastatals led to the reestablish- tion network. Yet the share of transport in public invest- ment of cooperative unions in 1984 and to a much greater ment in Tanzania dropped radically, as revenues were acceptance of the role of the private sector.4' In 1986/87, directed to the Basic Industrialization Strategy, to the the 23 reestablished cooperative unions operating at the expansion of social services, and since 1978 to defense regional level accounted for about 73 percent of fertilizer expenditures (Lele and Meyers 1987). At the same time, it distribution. must also be stressed that Tanzania's requests during the Finally, the farmers' access to fertilizer has been hin- 1970s for donors to support increased investment in the dered by uncertainty about prices, availability of inputs, transportation network did not receive adequate attention. and reliability of output markets. For example, in the early Western donors instead diverted one of the largest aid 1980s, the TRDB (now known as CRDB) was responsible for programs in Africa to other industrial and social pursuits, advancing fertilizer to creditworthy villages. In theory, an leaving the transport network and its role in Tanzania's eligible village was one that had repaid last season's loan. economic development to China, who invested in railroads. In practice, however, most villages were given credit for two Unfortunately even this investment has faced major main- seasons before being completely cut off from TRDB sup- tenance problems. plies. The effect of this scheme was to restrict, rather than The TFC, which has had a monopoly on imports and expand, the use of inputs (Spurling 1982, p. 26). distribution in major regional centers, has had difficulty The pricing of fertilizer also underwent several changes.08 transporting fertilizer between the factory, the port of entry Beginning in 1973/74, as part of the attempt to boost food (Dar es Salaam), and the major railheads. The shortage of production, USAID and the World Bank financed the 24 National Maize Program. The government then extended seems well underway and a substantially improved trans- the program to many marginal semiarid areas of the country port infrastructure. Donors have now begun to invest more where cassava and sorghum/millets are the predominant in transportation, but recent data on donor financing of crops. According to one senior Tanzanian official, the transport still show a woeful neglect of the traditional government tended to promote maize "where it grew production areas of Arusha, Kilimanjaro, and the Lake politically rather than agronomically." Under this program, it Victoria Basin, and recurrent financing is severely con- was also decided to issue free fertilizer to maize growers strained (Gaviria 1989), suggesting that the government throughout the country, a policy that lasted for two seasons, policy of regional redistribution from regions where well 1973/74 and 1974/75, but caused a severe wastage of over one-third of Tanzania's population is concentrated may fertilizer through distribution to the areas of least potential. well be continuing. In 1975/76, the free fertilizer program was replaced by a 75 Reform measures in Tanzania as elsewhere have tended percent subsidy, which was subsequently reduced to 50 to concentrate on short-term measures. Thus, given the percent on domestically produced fertilizer, and a 100 macroeconomic crisis, the export rehabilitation program percent subsidy on the cost of transporting fertilizer from that the World Bank approved in 1981 concentrated on the factory to district distribution points (Spurling 1982, p. increasing producer prices, ensuring adequate foreign 4). By the early 1980s, the level of subsidy on both fertilizer exchange to meet the basic import requirements of the and transport had declined and finally was eliminated in agricultural and transport sectors, and restructuring the 1984 with the full cost passed to producers. The National parastatal crop authorities, as the government was then Maize Program had very little impact on maize production, unwilling to entertain any privatization. although it did play an important role in introducing In 1986, the Bank funded another Multisector Rehabilita- farmers to the concept of fertilizers, an objective that could tion Credit in which it offered several specific recommenda- have been achieved at a fraction of the cost. tions with regard to input supply and credit. These The prices farmers now pay for fertilizer do not reflect included the need to reassess the implications of increased the subsequent devaluation and the increase in internal reliance on unpredictable grant aid fertilizers and to transport costs because of an explicit and implicit subsidy reexamine the feasibility of domestic fertilizer production. estimated to be in the range of 60-66 percent in 1988/89 Given scale economies in fertilizer production and the (Rioseco 1989; Carr 1989). questions about the design of the factory, the latter policy Related to the question of input pricing policy is that of should be pursued with the utmost caution if the past pan-territorial producer prices, as the latter made it experience of donor support of premature industrialization relatively more attractive to produce high bulk, low value in Tanzania is to be avoided. Therefore, a choice between foodgrains such as maize in the more remote regions. small-scale fertilizer production and grant aid fertilizers Although Kenya and Malawi have also pursued such a actually contains a trade-off between low returns from high policy, for which they have been criticized by donors due to average cost and high risk from uncertainty in fertilizer the small size and population concentrations, the financial supply. In addition, the Bank recommended that coopera- implications of such a policy in terms of additional trans- tives encourage cash sales through special incentives, that port costs are far less significant in these two countries planning and procurement procedures for fertilizers be compared to Tanzania. A majority of maize purchases by the improved, and that the rural infrastructure be diversified maize board in Kenya (75 percent in 1985/861 and ADMARC (World Bank 1986c; FAO and WJorld Bank 1987). Avoidance of in Malawi (59 percent on an average between 1981-86) are nationwide campaigns (instead of less ambitious concentrated in the Rift Valley and Central regions, respec- approaches), institutional effectiveness, and weaknesses of tively, whereas they are scattered throughout the country in the transport network will continue to be Tanzania's most Tanzania owing to the location of its producing and important developmental challenges, along with the long- consuming populations. In addition to distorting the term maintenance of a conducive macroeconomic policy location of production of foodgrains, such relative prices environment. This means that the issues of long-term probably contributed to the substitution of maize for export development and rehabilitation of the physical and admin- crops in the more remote regions, for example, maize for istrative infrastructure in the country are yet to be tackled. tobacco in Ruvuma during the 1970s, when their compara- tive advantage would suggest the reverse (World Bank Malawi 1983a). To meet the consumption needs of Dar es Salaam it Both its fertilizer policy and the circumstances of its was reasonable to encourage maize production in remote agricultural sector distinguish Malawi from the other MADIA areas; however, this strategy made large demands on the countries, although fertilizer use has increased at a moder- transport system to purchase and transport even the ate rate (7.7 percent between 1972/73 and 1987/88). First, smallest quantities of maize sold in the distant areas. This the need to increase productivity in agriculture in Malawi is policy was later reinforced by certain regions receiving great-perhaps greater than in the other MADIA coun- premium prices regardless of their location and the tries-because it is among the poorest countries in the transportation costs. As in Kenya, the policy issue Tanzania world and has one of the highest population densities in now faces is how to ensure the growing food consumption Africa (World Bank 1988b). Second, the agricultural sector needs at reasonable prices both of Dar es Salaam and of dominates the economy to a greater degree than in the some of the remote drought-prone rural areas. The needs other MADIA countries, accounting for 36.4 percent of GDP of the former have been met through increasing food in 1988 and about 80 percent of employment (Government imports, including food aid, while the latter have been of Malawi 1988). Third, Malawi suffers from dualisms within addressed by the government in an unstable and costly agriculture (Lele 1989). Not only has the share of the estate manner. At the same time, how to encourage private sector in land use been growing but the smallholder sector production and consumption in rural areas in the context of itself is dualistic with 55 percent of the 1.3 million small- a legal economy has been an important issue. This requires holder households cultivating traditional varieties of maize a continued improvement in macroeconomic policy that on holdings of less than one hectare. Further, these 25 households do not produce adequate food and depend on level of 7.5 percent in 1982/83 to 3 percent in 1986/87 the market for purchases of maize to meet their domestic (Nathan 1987). food needs. Fertilizer use tends to be concentrated on In addition, the liberalization of grain markets in 1987 larger land holdings-those in excess of 1.5 hectares- reduced the government's food stocks and contributed to because of the risks associated with fertilizer use and concerns about food insecurity. The influx of refugees from problems with credit availability (Carr 1988, p.9).49 50 On Mozambique, which exceeded 500,000 by mid-1988, had average during the 1980s, only 25-30 percent of the small- also added to domestic food demand. As a result of these holder population who had access to credit applied concerns, the government withdrew from the subsidy fertilizer and over 80 percent of the fertilizer used by small- removal agreement in mid-1987 and is presently continuing holders is applied to maize.5 Nevertheless, not all farmers to subsidize smallholder fertilizer prices by about 25 using fertilizers are using improved seed that responds to percent. As this level of subsidy has been constant over the fertilizer. The land area under improved varieties of maize past several years, fertilizer prices for the smallholder is consistently less than 8 percent as compared to 60 sector have continued to increase in the face of devalua- percent in Kenya. Overall maize yields in Malawi have tions. In order to maintain economic incentives for fertilizer stagnated at about I metric ton per hectare, as there has use, the producer price for maize, and in turn the consumer been little progress in adoption of improved maize varie- price, has been raised between 1987/88 and 1988/89 by 44 ties by farmers despite 20 years of agricultural development percent (16.67 tambala per kilogram to 24 tambala per projects. The low maize productivity requires that as much kilogram). Increasing output prices have deleterious conse- as 70 percent of total cultivated area be devoted to subsis- quences for the welfare of urban and rural food-deficit tence maize production, thereby allowing little opportunity households as they devote a third of their income to food for cash crop production. purchases. Indeed the situation in Malawi now resembles The focus of the fertilizer reform program in Malawi many Asian countries where food and fertilizer subsidies to (sponsored mainly by the Bank and USAID) has been the increase internal food self-sufficiency were considered elimination of the fertilizer subsidy. Increasing efficiency by essential due to growing landlessness. replacing conventional fertilizers with high analysis fertiliz- USAID and the World Bank, which had adopted a staunch ers is perhaps more advanced in Malawi and Kenya than in antisubsidy stance during the first three SALs, have recently the other MADIA countries, in part due to the progress become more receptive to maintaining the present subsidy made during reform measures. The primary motivation for until improvement of maize technology and increased the subsidy removal program was to reduce the govern- access to credit by small farmers make the subsidy ment's budget deficit, which exceeded 12 percent of GDP unnecessary. Two outstanding issues remain, however: (i) in 1980. Import liberalization has not been an issue in how to increase fertilizer use among cash-short and food- Malawi because of the combination of difficult external deficit poor rural households in light of the leakages of transport problems"2 and the success of the IFAD/IDA subsidized fertilizers to the estate sector and the relatively funded Smallholder Fertilizer Revolving Fund (SFRF) in larger small farmers; and (ii) how to ensure that the food procuring and importing fertilizers (see Appendix 7).53 The deficit farmers have the means to pay for fertilizer. With issue of subsidies is vexing in Malawi, because the fertilizer respect to the first issue, the two alternative programs price/maize price ratios faced by Malawian smallholders under consideration are a targeted subsidy on fertilizer for have tended to be three to four times those in Kenya and food-deficit households versus continuation of a general- as much as ten times those in Nigeria.54 At the same time, ized subsidy for the smallholder sector as a whole. Most as mentioned earlier, leakage of fertilizers made available donors lean toward a targeted subsidy, whereas the to the smallholder sector to estates that are favored in government prefers a general subsidy which, quite surpris- policies toward land access, prices, and markets has also ingly, is less demanding of budgetary resources. There are been a problem. also differences with respect to choices of an instrument for The Government of Malawi initially agreed to the subsidy a targeted subsidy, with some preferring a food-and- removal program, but a combination of devaluations and fertilizer-for-work program, while others favor a program increased external transport costs following the distur- targeted more directly on land-scarce (food deficit) house- bances in Mozambique caused domestic fertilizer prices to holds. Clearly, so little experience exists about the targeting rise and undermined the government's resolve to persist in of subsidies under Malawi's particular political and admin- eliminating the subsidy. This may have been because the istrative circumstances, that any program devised will have subsidy contributed relatively little to the large overall to explicitly recognize this fact and be experimental in budget deficit but helped prevent prices of inputs from nature-indeed more than one approach may be attempt- rising further in a situation where food shortages were likely ed in different areas-with clear monitoring systems to ensue from a combination of the influx of refugees and established to learn by doing. external transport bottlenecks. The government argued that The other important issue affecting Malawi's use of continuation of the subsidy was necessary partially to offset fertilizer is access to institutional credit. Even though food- changes in producer prices, which had shifted relative deficit households cannot sell the direct proceeds of prices away from maize to export crops. The perceived fertilizer use (i.e., maize) without jeopardizing their food consequences of the subsidy removal program for food security, this does not mean they cannot repay credit. The security were aggravated by declines in the volume of problem becomes one of how to develop mechanisms to marketed maize between 1984 and 1987. These declines ensure that these households are able to repay credit as were a result of the constant nominal producer price of repayment is typically linked to the sale of output. Failure maize over the preceding four years together with rising to repay would impair Malawi's excellent record on credit fertilizer prices, and prices of competing crops undertaken recovery, reinforcing the conviction held by some that only as a measure of agricultural diversification. The area larger farmers are creditworthy. This raises complex issues planted to improved maize dropped from the already low about the design and viability of the credit system, given 26 that many poor households are currently reluctant to Cameroon borrow because of the fear of not being able to meet the The issues surrounding increased fertilizer use in Camer- repayment record which is 96 percent. Less stringent oon are less urgent than in Malawi, Kenya, or Tanzania, repayment criteria, by which 85 percent of the total chiefly because land pressure is less severe but also repayment will be tolerated from the groups who borrow, because oil wealth and the stability of past policy allow are being devised. Another solution is to allow many Cameroon a healthier cushion. Nonetheless, while Camer- already market-dependent, food-deficit households to oon's agricultural performance has been relatively strong increase their earnings by growing burley tobacco, which is compared to other MADIA countries (Lele 1988), structural currently grown almost exclusively on estates, and thereby weaknesses have hindered the design and implementation increase access to income and purchased inputs. Small- of agricultural policy. in particular, the fragmentation of holders are more efficient producers of burley, although agricultural programs through development companies and yields on estates are substantially higher. (The DRCs for a weak national ministerial structure (weaker than in either smallholder burley production are lower than for estate Kenya or Malawi) have served as a limitation on agricultural burley production; see Lele and Agarwal 1989.) By enforcing policy. Consequently, much of the growth of fertilizer current quota restrictions on the production of burley, consumption has taken place through the efforts of devel- however, estate owners have realized substantial rents and opment parastatals, many of which have been managed a cheaper supply of wage labor to intensify production, and with varying degrees of expatriate input, for example, they tend to resist attempts to liberalize production. SODECOTON and SEMRY (Lele, van de Walle, and Gbeti- A third and important issue relates to the relative roles bouo 1989). of hybrid (dent) and local (flint) varieties of maize. The The salient issues in the fertilizer subsector were identi- differences in the fertilizer responsiveness of each, espe- fied by an IFDC study (1985) in preparation for dialogue on cially at low levels of fertilizer application, are not known, reforms between donors (chiefly USAID) and the govern- despite the fact that Malawi has had one of the best ment. That study identified several weaknesses in the organized and well-functioning agriculture research systems fertilizer subsector and made several recommendations among the MADIA countries, showing how little emphasis including: (i) an accelerated crop production and fertilizer there has been in agricultural research to make it farmer- research program, and (ii) a reduced role for the govern- oriented. Agronomists argue that enough is known to make ment in the procurement and distribution of fertilizer, recommendations. First, even the highest recommendations complemented by a greater role for the private sector (IFDC are below the point of diminishing returns. Second, the 1985, pp. XiV-XX).55 These recommendations correctly reflect response curve in all likelihood is a straight line in the problems with fertilizer importation and with recommenda- relevant range, and the much maligned blanket recommen- tions for fertilizer application in light of the varied and dations are not so bad. However, small farmers are fragile soils in Cameroon.56 At the same time, there are unwilling to adopt high risk, input-intensive hybrid maize other dimensions to the problems of intensification in which does not meet their consumer preferences. At the Cameroon that the IFDC report and consequently USAID same time, the national need to maximize the supply of had not originally anticipated. USAID together with the calories from the least amount of land is increasing in order government appears to have begun to address these with to ensure that enough land is available for increasing export considerable sensitivity and understanding. Easier ele- volumes to maintain much needed imports. The dualism ments of privatization appear to have progressed without created by Malawi's land policy has not only led to hindering fertilizer availability although several problems increased land in the estate sector but consequently the remain in accelerating fertilizer use. They are, most notably, number of households in the customary sector is increasing the weak domestic private trade and transport network and rapidly, contributing to poverty and the number of risk the absence of an effective financing mechanism for working averse producers. Thus, while donors and the government capital-either for importation, wholesale, and retail distri- debate the best means of promoting fertilizer use on the bution, or for small farms for purchasing agricultural inputs. flint maizes preferred by smallholders, the macroeconomic As a consequence, the results of liberalization in Cameroon implications of that strategy compared to concurrent are not dissimilar from those in Kenya-efficiency gains promotion of the use of higher-yielding dent maizes which accrue to large-scale agricultural enterprises, and some are sold to ADMARC and milled for resale must be notable improvements have occurred for the more commer- considered as the need for sales by ADMARC will in all cially-oriented small-scale farmers with as yet limited effect likelihood increase. This explains the governmental on the large majority of subsistence-oriented farmers, research system's focus on dent maizes, albeit at poor rates whose productivity must be improved and who face of adoption. Equal priority must also be given to research numerous interlinking constraints. While this should not be on the flint maizes that small farmers prefer. surprising given the short duration of the privatization Although donors and the government recognize the program, how far and how fast the private sector responds importance of many of these individual issues, the highly to the challange of creating new demand remains to be interactive nature of the issues requires that a cohesive seen. long-term agricultural strategy, addressing a number of A five-year, two-phase fertilizer reform program that was constraints simultaneously, be formulated urgently. Donors agreed to by USAID and the Government of Cameroon can then provide long-term financial support in pursuit of (GOC) contained four main components:57 (i) the liberaliza- more equitable growth as an essential complement to tion and privatization of fertilizer imports and distribution, delivering the needed macroeconomic balance. As in the (ii) the continued expansion of the private sector in case of Kenya and Tanzania, however, such a long-term fertilizer and other input distribution services, (iii) the agricultural development strategy with its many interacting phased elimination of the fertilizer subsidy,58 and (iv) an parts is yet to be developed for Malawi. annual review of smallholder crop prices to "avoid negative income effects on farmers and prevent a decrease in 27 fertilizer demand/use..." (USAID/Cameroon 1987, pp.43- rice producers, although they along with others have come 44, 51-54). To this end, USAID/Cameroon pledged that GOC under heavy criticism for their high cost of operations.63 will "review coffee price policy on an annual basis with the Nevertheless, to the extent that the success of many of objective of raising the producer price the estimated 10-12 the parastatals in credit recovery and production has been percent required to offset the increased cost of fertilizer' due to the effective integration of input supply with applied to coffee." monopoly crop purchases, privatization throws into ques- The agreement also called for the "creation of a credit tion the long-term development of future credit supply and fund within selected, well functioning commercial banks," marketing arrangements for export crops. and "reduction of uncertainties related to subsidy pay- Donors traditionally engaged in support of export crops, ments by recommending that the GOC deposit the subsidy especially the EEC and France, have not been entirely payment in commercial banks for management and dis- enthusiastic about the speed of privatization encouraged bursement to fertilizer distributors" (USAID/Cameroon by USAID. USAID's past experience in Cameroon has been 1987, pp.46-48). focused mainly on food crops, whose active markets and Several issues are raised by the privatization of importa- high prices have provided a strong incentive to use tion and distribution. For instance, the effort to extend fertilizers; however, there is mixed evidence on how credit to private importers posed initial problems, as the competitive or integrated the food markets are. With commercial banks were reluctant to advance credit on respect to export crops, France and the EEC have imported fertilizer used as collateral. They required import- expressed concerns about the robustness of the internal ers to provide additional collateral in excess of the value of distribution channels and the time that should be allowed the loan, given the absence of a lending history for fertilizer to develop effective cooperative and private sector chan- in the case of any private firms (with the exception of a few nels. This applies especially in remote areas where cooper- sales for horticultural crops) and the banks' lack of familiar- ative activity has been weak and unofficial access to ity with the fertilizer sector. USAID moved quickly to ensure markets in neighboring countries for cotton and rice that commercial bankers would receive the necessary requires a major realignment of markets and approaches. guarantee by establishing a USAID-funded loan facility for European aid has also often been tied to fertilizers mixed importers at local commercial banks.59 in their own (or member) countries, with financing provided Whether the absence of institutional credit to small through their own offshore banks rather than through farmers is a constraint to the growth of fertilizer use is a Cameroonian banks. This partly explains the reluctance to much debated issue, especially for the development of move to privatization quickly. traditional food and export crops. However, small farmers Senegal near major urban centers in the southern part of Cameroon have shown willingness to pay cash for unsubsidized In Senegal, where the fertilizer distribution system in the fertilizers for use on high value vegetables for urban Groundnut Basin has collapsed, the same issues of the markets. Indeed, a recent study of rural finance points to appropriate roles of the public, private, and cooperative the buoyancy of informal consumer credit systems such as sectors, institutional and policy stability, and the speed of the tontine, while at the same time pointing out that credit liberalization arise, but given the much lower productive demand for cash crop production has been weakened by potential and higher risks in farming, these issues have a the low returns (with the exception of cotton). Cotton much greater significance than in other countries. producer prices in Cameroon are considered too high by Kelly, who has done farm surveys of households in the international standards and are being lowered to maintain Groundnut Basin, observes: financial viability of the industry. The issue of economic ... agricultural policy in general, and input distribution benefits of fertilizers and the ability of small farmers to policy in particular, has been in a state of relative finance purchases is thus important in Cameroon, where chaos since 1980. The GOS has a stated policy, the role of the private sector in financing production credit particularly with respect to fertilizer, but to date it has is not known.60 It is not clear, for instance, whether tontines, been unable to implement many facets of the policy. which mainly give consumption credit and do not experi- The rules for input distribution, shaped to a large ence the high degree of seasonality in the demand for extent by GOS economic constraints, have changed credit for inputs, can provide working capital to relatively radically from year to year. Farmers and distributors small farmers on a large enough scale to make a difference have received little advanced warning of such in macroeconomic terms. (Despite the more active money changes. At the farm level, the end result has been lending enterprise in Asia-in India for instance 70 percent sharply reduced fertilizer consumption and the devel- of the credit was once provided by the informal sector-the opment of new strategies for acquiring fertilizer and share of their lending in financing modern inputs has been compensating for diminished access (Kelly 1988, p. insignificant.) In Africa, it is unlikely that informal credit 72). could meet the credit needs of farmers. The IFDC report The situation has been made more complicated than in seems to share these concerns, observing that a "lack of Cameroon by a weak and variable demand for fertilizer purchasing power at the farm level was a major constraint because of the increasing variability and declining amounts to fertilizer use" (IFDC 1986a, p.208). of precipitation throughout the country between 1960 and The role of formal public or cooperative financial institu- 1983 at an average annual rate of 2.2 percent (lammeh and tions in providing small farmer credit raises another set of Lele 1988; Kelly 1988). The fundamental importance of issues for the future. Since the major financial institutions increasing groundnut and sorghum/millet production in in Cameroon have not made significant headway in lending Senegal from the point of view of export earnings and food to small-scale agriculture, credit in kind through parastatal security has been greatly complicated by this weakness of development agencies has been the main source of small demand and by the simultaneous withdrawal of credit, seed farmer credit.61 SODECOTON and SEMRY62 are, for instance, distribution services, and public sector distribution of generally acknowledged to have worked well for cotton and fertilizer following the dissolution of the state's marketing 28 apparatus, ONCAD, in 1980. This was part of a larger of costly storage. transportation, and distribution attempt by donors to reduce the public sector deficit and given the uncertainty about future fertilizer price and employment levels through one of the earliest structural credit policies. Furthermore, the absence of a farmer adjustment loans, which included efforts to eliminate credit program made it virtually impossible to accu- fertilizer subsidies.64 65 How successful the attempts to rately estimate effective demand; and the few inde- reduce overall deficits and employment levels have been is pendent traders who expressed interest claimed a not clear, although the effect of measures on fertilizer use major constraint was lack of commercial credit (Kelly is indisputable. Fertilizer consumption declined by 88 1988, p. 701. percent from 1979/80 to 1985/86 and shifted to rice and A similar view was taken by participants during a seminar cotton outside the Basin.66 However, public sector employ- held in Dakar in mid-1986 on the privatization program: ment, which was 56,888 in 1979/80, rose to 67,519 in 1984/ (a) private traders are hesitant to participate in the 85. Over the same period, the average public sector distribution of fertilizer because the business is monthly wage bill rose by 60 percent, this in a country that considered financially risky; lb) the lack of a credit already had the largest share of public administration in system for both farmers and private traders keeps the GDP among the MADIA countries and a decline in real per private market of fertilizer very limited; (c) farmers are capita GDP of over 1.0 percent annually during the period still skeptical about the profitability of fertilizer use 1960 to 1987. It is difficult to disentangle the effects of (World Bank, 17 luly 1986d, p.l, office memo). institutional and climatic factors and, within institutional issues, to assess the relative roles of the lack of fertilizer Interestingly, outside of the Groundnut Basin, withdrawal supply, the absence of credit, and the unwillingness of the of the state's marketing apparatus (ONCAD) has left behind private sector to supply fertilizers on either cash or credit a complicated system of input and output marketing based in the absence of stable effective demand-the latter on a few remaining development companies, cooperatives, reflecting the poverty of Senegalese farmers and the high and a patronage arrangement centered on the marabouts risks in cropping due to environmental factors. The effect of (Waterbury 1989). Ironically, it is SAED and SODEFITEX, two the decline in rainfall, more than 3 percent annually in two parastatal institutions, together with a few cooperatives, that provinces in the Groundnut Basin that historically have maintained fertilizer distribution in each of the regions been an important source of fertilizer use67 68 is described outside the Basin-Fleuve, Upper Casamance, and Eastern by Kelly, as follows: Senegal, although in the last two years, reportedly SAED In general, reliance on organic fertilizer is not consid- has gradually disengaged from its production-related ered adequate. In recent years, however, threat of activities-which included input distribution-in the Fleuve drought has caused farmers to rely on organic fertil- region, and the private sector has been promoted. Fertilizer izer rather than assume the dual risk of crop loss and distribution in the Fleuve is now reported to be under- cash investments associated with chemical fertilizers taken entirely by the private traders. But unlike the Kelly (Kelly 1988, p. 12). study, there are no data to show what the real fertilizer Kelly goes on to argue that situation appears to be from the field level. Progress in getting private traders involved in retail distribution in ... low fertilizer demand in recent years has been due other parts have been admittedly slow, although private to (1) low farm incomes, (2) low and uncertain fertilizer traders are handling wholesale transactions with coopera- response, (3) farmers' belief that fertilizer is not an tives. Interestingly, in these other regions repayment rates essential input, (4) farmers' preference for alternative have been high in both SAED and SODEFITEX. As in investments considered more profitable and less Cameroon, however, both have come under severe criticism risky, (5) the lack of credit (viewed by many farmers as for their high costs of operations. a form of insurance), and (6) an unresponsive distribu- While diversification attempts justify proceeding else- tion system (Kelly 1988, p. 254). where, there still remains the question of the future of Ability to pay cash is also an important concern. agriculture in the Groundnut Basin. We have stressed the Kelly's logit analysis indicates that 53 percent of those fundamental importance of increasing fertilizer use in farmers with less than a 10 percent chance of buying groundnut and sorghum/millet production, because fertilizers have farms of less than 6 hectares. Field surveys exports of groundnuts are stagnating, land pressure is indicate that 100 percent of these cash strapped farmers increasing, and soil degradation is worsening (lammeh and understandably give priority to buying peanut seeds over Lele 1988). There is also the question of the future roles of fertilizer, as being the most crucial input to realizing any the private, cooperative, and public sectors and the speed production. Only 18 percent of farmers with less than 6 with which the past balance can be changed. hectares paid cash for fertilizer sometime during Senegal's As in Cameroon, SAED's success in input distribution and Programme Agricole (as compared to 100 percent of farmers credit recovery comes from its ability to integrate credit, who cultivate more than 6 hectares), and none currently have inputs, and rice marketing. In addition, SAED can reclaim access to a reliable source of noncrop revenues.69 The irrigated parcels from debtors. Similarly, because of its absence of noncrop revenues and lack of access to credit monopsony control of cotton marketing, SODEFITEX is in a means that the probability of fertilizer use is low.70 The fact good position. Further, it has the advantage of working with that much of the fertilizer is now being consumed in the small, cohesive, self-managed producer groups with trained areas of relatively high rainfall is considered by donors officers, which can eventually be turned into grassroots (World Bank, USAID) as reassuring. Even in these areas, cooperatives, although started in a paternalistic manner. however, private sector sales have made little headway: Donors have tended to treat cooperatives as synonymous ... USAID offered to finance a subsidy on all cash with privatization, and they have not adequately focused on sales made by the private sector (SONACOS, Cooper- the politics of cooperatives. It is not clear, for instance, atives, and commercial outlets). The private sector, when and whether a genuinely grassroots cooperative however, was generally unwilling to assume the risks movement will develop in Senegal, which means that such 29 collective nurturing of cooperatives may be essential, as To summarize, although privatization of fertilizer distribu- well as developing some role for public sectorally-led tion is in many ways an attractive option, it is likely to commercial institutions. After independence in 1960, the progress slowly in Senegal, due to high risks in adoption of government abolished colonial marketing systems based on modern technology and the poverty of Senegalese farmers. Lebanese traders to increase its political control on Further, reliance on the private sector for fertilizer sales is economic activities. It then led Prime Minister Dia to unlikely to work without a comprehensive marketing strat- expand the state's role in agricultural marketing through egy for inputs, credit, and output purchases. This strategy animation rurale. However, he encountered resistance from poses a dilemma, however, because farmers have preferred the marabout traders, whose spiritual, economic, and to use credit and inputs, but have frequently avoided political power is undeniable in Senegal (Schumacher 1975). repayment by selling output in the informal sector. This has With Mr. Dia's implication in a coup attempt in late 1962, occurred in large part because for political reasons the the grassroots orientation of the cooperative movement government and the marabouts have tended to retain that he attemped to steer had ended. The movement political control with farmers having been allowed little continued, however, with the Senegalese state attempting stake in the institutions that determine their livelihood. to ensure more compatibility of cooperative activities with Changing this state of affairs will require a radical rethinking marabout interests, whose support Mr. Senghor needed to about the role of grassroot-based farmer cooperatives consolidate political power (Waterbury 1989). The top down which can become a political and economic force on par bureaucratic nature of cooperatives which has since evolved with the marabouts and the Senegalese state. For such a appears to have become ingrained. change to occur, donors would need to support coopera- The risk-induced financial problems of public sector tives, while recognizing that cooperatives face tremendous commercial organizations also need to be separated from risk in terms of financial problems that must be separated political patronage and management inefficiency problems, from their management failures. although the former have received relatively little attention from donors. In Senegal, both problems seem to have been Nigeria at work and in policy disclosure their relative roles have As in the other MADIA countries, the emphasis of policy never been classified. For instance, when farmers were reform in Nigeria has been on removal of the massive unable to repay debts because of droughts in 1977/78, fertilizer subsidy and privatization of fertilizer procurement 1980/81, and 1981/82, BNDS and ONCAD encountered and distribution.7" The two issues are intertwined in Nigeria financial difficulties. Forced by recurring droughts, and because the rapid growth in fertilizer use (from less than motivated by political concerns to broaden peasant sup- 20,000 metric tons of nutrient in 1972 to 263,000 metric tons port, the government intervened four times between 1977 of nutrient in 1987) has been coincident with a uniform and and 1981 by ordering ONCAD to forgive farmers' debts-an high rate of subsidy on fertilizer prices (85 percent subsidy action for which ONCAD was not reimbursed. Agriculture over most of the period, combined with a highly overvalued credit was essentially seen by the government as a form of exchange rate-meaning that fertilizer has been virtually agricultural insurance. Although the concept of agricultural distributed free of charge) and the centralization of the insurance is frequently popular in donor circles, they have fertilizer import and distribution system. The overwhelming not considered credit forgiveness as a form of insurance for use of fertilizer in the politically powerful North where it small farmers. On the other hand, massive problems of has been used by successive military governments as a overdue payments by farmers having political support has means of redistributing the oil wealth to the North makes not dissuaded them from approving repeated loans to the problem of subsidy removal sensitive. The relative credit agencies, e.g., Kenya's AFC (Lele and Meyers 1986). In influence of these various factors in the growth of fertilizer Senegal, mismanagement of funds and overexpansion of use is difficult to disentangle. For this reason and given the ONCAD's staff (because of political pressures to increase complexity of these issues a thorough review of these employment) also contributed to ONCAD's financial prob- topics is crucial to understand the consequences of the lems but the relative roles of the two factors have not been policy reform program. analyzed. This led to the abolition of ONCAD in 1980. The Four phases are discernible in the Nigerian government's continued public sector growth in the context of a declining fertilizer subsidy policy since the centralization of fertilizer economy appears to be a serious problem in Senegal, procurement and distribution in 1976 following the first oil precisely because an alternative employment-oriented price increase. In the period 1976-79, the federal govern- strategy that will generate employment in the private sector ment subsidized the cost of fertilizer imports, port clear- has yet to emerge. ance, and transportation to state warehouses. Officially the USAID and the Caisse Central de Cooperation Economique subsidy amounted to 75 percent of the landed cost of (CCCE) recognized the likely adverse impact of climatic and fertilizer in state capitals, with farmers expected to pay the institutional problems on the demand for fertilizer. To remaining budgetary 25 percent. The northern state govern- address budgetary concerns, they jointly funded a limited ments which used 70 percent of the fertilizer, however, subsidy on fertilizer through the New Agricultural Policy, were rarely able to recover the intrastate transportation which the government launched in 1984, but unfortunately costs from farmers, with the result that the actual subsidy to they restricted it only for those farmers able to pay cash for farmers was close to 85 percent (Idachaba 1987). fertilizer.7 This did not address the problem of the majority During the second period (1980-83), the subsidy was to of small farmers who have insufficient cash to buy fertilizers. be the joint responsibility of the federal and state govern- Further, such a subsidy tends to be regressive since it only ments with each paying half of the 75 percent budgetary benefits large farmers who can pay cash, as Kelly has subsidy on the landed cost of fertilizers in the state demonstrated." This subsidy is now scheduled to end in capitals. As before, however, the subsidy actually accruing 1989.73 As in Malawi the issue of whether a targeted or a to farmers was closer to 85 percent because of the inability generalized subsidy should be provided for fertilizer is of the state governments to recover the intrastate transpor- clearly an issue that will need to be faced in Senegal. tation costs. Also, most of the cost of the subsidy fell on the 30 federal government as the states frequently failed to pay expected to play a larger role at the retail level, although their share (idachaba 1987). government wants to retain its monopoly on fertilizer Prior to 1975 in its initial involvement with ADPs, the imports. Nigeria's cooperatives, however, have some of the World Bank had taken the view that a fertilizer subsidy was same weaknesses as those of other countries. Although the essential to promote use. After 1975, the Bank raised ADPs have played an important role in ensuring fertilizer increasing opposition to input subsidies, but especially on supply in the states, they have played a small role in fertilizer because of the financial cost, its wastage, and sales developing local institutions whether commercial, coopera- across the border to neighboring countries where prices tive, or governmental. This means institutional weaknesses were considerably higher. By 1983, which marked the will pose a problem unless a gradualistic approach is beginning of the third phase, the Bank was sufficiently adopted to transferring responsibility for fertilizer distribu- opposed to subsidies, partly as a result of its field tion and other inputs to cooperatives.75 76 experience with the ADPs, to make the phased removal of The issue of how much subsidy will still be needed on subsidies a condition of its 1983 fertilizer import loan to fertilizers, albeit at a lower level, is important as there is Nigeria. Partly in response to the loan conditions, but considerable divergence between the crops on which mainly because the decline in oil revenues had increased fertilizer use is potentially most profitable, and those on budgetary pressure on the Nigerian government, it reduced which it is actually used. More than half of all fertilizer use the fertilizer subsidy from 85 percent in 1982 to 28 percent in Nigeria takes place on sorghum, millet, and maize grown in 1986, i.e., before the due date for abolition of the in the North through the ADPs which have only recently subsidy. The implicit subsidy due to the overvaluation of spread to the Middle Belt and the South. Fertilizer use the naira remained large (see Tables 6 and 7). appears to be the least profitable on these crops when the The beginning of the fourth and current phase is ratios of subsidized fertilizer prices and the market prices associated with the introduction of the second tier foreign that have typically prevailed in Nigeria in recent years are exchange market of the naira in October 1986. Despite the considered. Fertilizer use would not be profitable on those fourfold increase in the cost of imports as a result of the crops at market prices if the subsidy were removed 400 percent effective devaluation, the price of fertilizer was altogether, given the physical responses to application. maintained by the government at its predevaluation level. Rice, constituting more than 10 percent of total fertilizer use Thus, the subsidy on fertilizer again amounted to more than but under 2 percent of the total cultivated area, accounts for 80 percent. Meanwhile, because of good weather and a disproportionately high share of total fertilizer use in surpluses, the prices of domestically produced food also Nigeria, but little fertilizer use takes place on cassava, declined in 1986/87, causing concern about subsidy cowpeas, and yams. The prices and responses for these removal; because of a drought prices increased sharply crops are good and use needs to be increased. again in 1987/88 and have continued to spiral upward in Greater efficiency in fertilizer use can be attained by 1988/89, in spite of good harvests (Lele, Oyejide, et al. encouraging its greater use in the Middle Belt, and on 1989). At the same time, the naira has continued to foods such as rice, cassava, cowpeas, and yams, but this depreciate precipitously, with the exchange rate in May 1981 poses a dilemma. While the quality of the land in the being N 8 = $1, as compared to the rate of N 4 = $1 that Middle Belt is moderately better than in the other two prevailed immediately following the devaluation in October regions, and there is room for expanding the cultivated 1986. As the fertilizer price has remained more or less area, labor shortages resulting from the low population unchanged, this means that the subsidy in the most recent densities impose a constraint on intensification (Lele, period has amounted to about 90 percent. While the Bank Oyejide, et al. 1989). A similar dilemma results in terms of has taken a more cautious approach to the issue of subsidy encouraging the greater use of fertilizer on crops like removal in part because there is an increased recognition cassava, cowpeas, and yams. Because the market demand in the Bank now of the importance of chemical fertilizers in for these crops tends to be fairly inelastic with respect to increasing agricultural production, and in maintaining soil prices, the production increases resulting from the fertility, the budgetary implication of the subsidies in increased use of fertilizer could lead to precipitous Nigeria, their lopsided distribution to the North where declines in prices, thus creating a disincentive for produc- responses to fertilizer are perhaps less favorable, and the ers, but the introduction of support prices-which have opportunity cost of these resources for the creation of more been used extensively in Asia-poses problems that are permanent assets remain major issues. In 1987, the fertilizer discussed elsewhere in the MADIA papers. Finally, as in subsidies are likely to have accounted for three-fourths of East Africa, efficiency of use can be increased by shifting to the federal government's total agricultural budget (Lele, high analysis fertilizers instead of the 15:15:15. However, a Oyejide, et al. 1989). severe lack of location-specific technologies in Nigeria's As the oil revenues and the ability of the government to complex mixed cropping system makes transition more transfer resources have declined, privatization of wholesale difficult than in East Africa. This now leads us to the issues and retail business is being considered, with cooperatives related to physical responses. 31 Determinants of the Economic Benefits of Ferdlizer Use The demand for fertilizer is determined by its economic than in West Africa, a pattern that is true at both official value at the farm level, most commonly measured by the exchange rates and rates adjusted for purchasing power benefit-cost ratio. This, in turn, is determined by the parity (see Appendix 8 for the cost of fertilizer at official interaction of fertilizer prices, output prices, and the exchange rates). Even within East Africa, however, the cost responsiveness of crops to fertilizer application, in other of nutrients varies considerably, with prices generally being words, the response coefficient. As is evident from the higher in Malawi-the poorest country in the MADIA preceding discussion, the certainty with which the benefit- sample by per capita GDP-than in either Kenya or cost ratio is realized depends on the stability of input and Tanzania. Nigeria, which had the highest per capita income output prices and yields, which in turn influences the rate until recently, had the lowest nutrient price of the six at which farmers discount the use of fertilizers. Each of countries because of the high subsidy.77 Similarly, in these components is examined in this section. Senegal and Cameroon, nutrient prices even at purchasing power parity rates tend to be lower than in East Africa, Fertilizer Price although not as low as in Nigeria. Although farmers' decisions are determined by prices Although the variation in the fertilizer prices among encountered at the farm level, intercountry comparisons of countries is partly attributable to the types of fertilizers user prices, i.e., the price the farmer pays, that governments (i.e., straight nitrogenous, complex, and phosphatic) or to set for fertilizers (by converting them into U.S. dollar prices low and high analysis (see Table 13), price differentials also at official and purchasing power exchange rates) provide reflect a number of other factors including local and some useful insights, given the differences in per capita international transport costs, the level of subsidy, and the incomes among countries. Fertilizer prices used here are quantity and pricing treatment of grant aid fertilizers, as approximations, as they do not include the transport and well as the countries' ability to obtain the best deals on handling costs from the points of retail sales to farmers' prices. For example, in the case of ammonium sulfate (A/ fields where they are applied, nor do they indicate the S), the price variation between Malawi, Cameroon, and informal sales in fertilizers that we have indicated are Tanzania for 1985/86 was significant-$1,020 per metric ton significant. The producer level prices of nutrients (see Table in Malawi compared to $548 per metric ton in Tanzania and 13 and Figure 4) have been generally higher in East Africa $398 in Cameroon. Internal transport costs, which are Table 13 Prices for principal fertilizers in MADIA countries, 1971/72-87/88 Cameroon Senegal Nigeria Malawi Kenya Tanzania Year NPK A/S NPK NPK NPK CAN A/S DAP CAN A/S TSP (20:10:10) (6:20:10) (15:15:15) (20:20:0) US$/ton of nutrient adjusted using purchasing power parity exchange rates 1971/72 NA NA 122 NA NA NA NA 176 247 NA NA 1972/73 NA NA 132 NA 215 317 327 245 364 NA NA 1973/74 258 NA 190 NA 233 343 361 367 489 NA NA 1974/75 228 377 175 NA 587 852 833 348 464 NA NA 1975/76 477 727 203 NA 543 762 669 648 955 NA NA 1976177 432 548 249 91 510 714 628 588 1018 NA NA 1977/78 326 518 243 90 547 766 673 457 824 641 409 1978/79 307 474 281 86 585 820 720 498 952 680 435 1979/80 321 498 299 87 617 867 761 531 954 684 438 1980/81 328 546 327 90 608 1156 885 833 1195 612 392 1981/82 275 458 276 146 525 997 1057 819 1227 677 347 1982183 256 487 216 132 697 1106 1106 660 1085 519 266 1983/84 245 468 370 115 697 1036 1099 690 776 419 215 1984/85 208 396 514 186 648 884 953 603 912 291 149 1985/86 209 398 385 180 627 894 1020 NA NA 548 344 1986/87 262 498 488 277 645 856 1054 539 831 765 472 1987/88 NA NA NA 247 822 1066 1335 NA NA 625 386 Notes: Data are for official fertilizer prices adjusted using purchasing power parity exchange rates. The fertilizer type(s) listed for each country reflect what is predominantly used. Certain costs incurred within the country, such as handling cost and rebagging cost are not affected by currency overvaluation were not available separately for all years. Ideally these costs should not be adjusted for currency overvaluation; however, a lack of data and the fact that these costs comprise a small part of total fertilizer cost (less than 15 percent) mean the adjustment can be ignored. Fertilizer prices for Kenya are f.o.r. Nakuru. The prices in Tanzania after 1984 (when subsidy was abolished) refers to TFC's exstore prices at regional levels. Though retail prices are uniform throughout the country in both Kenya and Tanzania, the end-user prices vary according to location and depending on services rendered by retailers. Sources: For fertilizer prices, the following sources were used: Cameroon: Berg Report 1983; IFDC 1986b; Kenya: World Bank 1 986b; Personal Communication with USAID Office, Nairobi, Kenya; Senegal: Kelly 1988; Malawi: Nathan Report 1987; Nigeria: Lele and Bindlish 1988; Tanzania: Mhella 1985; FAO/World Bank 1987. Purchasing power parity exchange rates computed by Seka and Fishstein, MADIA Worksheet 1988. 32 Figure 4 Prices for primary nutrient types used in the MADIA countries (converted to US$/ton using purchasing power parity exchange rates) East African countries West African countries S/ton (thousands) S/ton (thousands) 1.3 1.3 12 CAN) 12 1.1 ~~~~~~~~~~~~~~~~~~~~~~~~~~1.1 0.9 0.- 0.8 ~~~~~~~~~~~~~~~~~~~~~0.8 0.7 ~~~~~~~~~~~~~~~~~~~~~~0.7 0.6 0.6 - ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~(S/A) 0.5~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~. 0.5 ~ ~ ~ ~ ~ ~ ~ ~~~/ 0.5 0.4 / \ /00.4 (6:20:10) 0.3 0.3 . -" 02 0.2 ,,,,,,,,..'-.,... 0.1 0.1 - - * * (15:15:15) 1971/72 73(74 75/76 77/78 79/80 81182 83/84 85/86 8687 1971/72 73174 75(76 77178 79/80 81/82 83/84 85/86 86187 ........ Malawi - Cameroon Kenya ''......Senegal - - - - Tanzania - Nigeria relatively high in the East African countries compared to output. Some governments (e.g., Nigeria) are already those in the West African countries, partly explain this financing the transport of fertilizer. Of course, greater difference (see Table 14). In Malawi, owing to its landlocked regional integration between countries allowing across- position and the civil war in Mozambique, the cost of border trade will be one additional way of addressing the transporting fertilizer from port to farm gates was relatively problem of market integration. Such sales are already high-$137 at the official exchange rate in 1987.78 extensive though illegal, as for instance between Tanzania In Kenya and Tanzania transport costs from port to farm and its neighbors, Cameroon and its neighbors, between gate were $33 and $124, respectively, at official exchange Senegal and Gambia, and between Nigeria and Chad. rates.79 Large internal distances and poor development of However, every government in the MADIA countries att- infrastructure explain some of the high costs in Tanzania, as aches high priority to food security and has tended to indicated earlier. Currency overvaluation also overstates the control the vital food trade routes so as not to be costs in dollar terms at official exchange rates, although embarrassed by an inability to feed its own population in since four-fifths of transport expenses consist of foreign periods of drought-the frequency of which is increasing. exchange costs, this is not a significant factor. In Nigeria, Herein lies the dilemma of the theory and practice of Cameroon, and Senegal the port to farm gate costs were interregional integration. Donors could help by financing considerably lower, $17, $22, and $9, respectively, at official regional and national stocks of food and fertilizer in exchange rates perhaps because of the lower costs of sufficient quantities over a long enough period to increase petrol.80 the reliability of inputs and food supplies and thereby These high internal transport costs can also be seen encourage African governments to permit interregional when expressed as a share of total marketing costs and trade (e.g., like the EEC) by increasing their confidence in compared to non-African countries (see Table 14). A more their own ability to address the politically explosive issue accurate comparison is between the transport cost as a of food shortages. This will expand markets, thereby share of the c.i.f. price of fertilizers among countries, since reducing risks, and in turn will promote intensification of some countries include taxes as part of total marketing agriculture. costs (Table 14). As can be seen, two MADIA countries- Malawi and Tanzania-have transport shares in excess of Producer Price and the Relative Cost of Ferilizer half of the landed price of fertilizer, and a third country, The high fertilizer prices in East Africa must be considered Nigeria, has a cost share amounting to nearly a third of the against the background of official prices of maize in Malawi cost of fertilizer. These are shares that are much higher than and Tanzania (at which between 10 and 20 percent of the in non-African countries. The issue of distances raises the maize production is traded) which have tended to be about more complex question as to whether it is not more half those in West Africa when using purchasing power efficient to promote the use of less bulky fertilizer through parity exchange rates (see Figure 5 and Appendix 9, Table subsidies on its transport rather than on food crop trans- 1)81 portation between food surplus and food deficit regions- Comparisons between export crops are more limited at least until transportation infrastructure and improved than for food crops because only a few countries produce production efficiency offsets the higher transport costs of crops in common, cotton and coffee being the most 33 Table 14 Notes: I For Malawi, internal transport cost is comparatively low; because and as a sare of mareting costMalawi is landlocked, it is necessary to include the transport costs Actual cost of transport and as a share of marketing cost between the ocean port and the domestic point of entry. As a result and c.i.t. fertilizer cost in selected countries of the war in Mozambique, the nearest port available is Durban, As a percentage of whereas earlier Beira or Nacala was a more economical alternative. internal total c.i.f. The transport cost from the port to the Malawian border is $122 per transport marketing fertilizer metric ton for a total transport cost of $137 per metric ton. Figures in Country (year) costs costs cost parentheses represent the total transport cost The transport cost in early 1983 through Beira/Nacala by rail to Lilongwe for containerized fertilizer was $30 per metric ton. $/metric ton Percentage 2 Transport cost for Kenya is from Mombasa to farms in Nakuru district. Malawi (1 987)' 15 (137) 22 (73) 6 (59) The cost of transporting fertilizers to the farms as far as Kisii district Kenya (1 984)2 33 33 15 is $51 per metric ton at the nominal exchange rate. Tanzania (1 985/86)3 124 50 52 3 Transport cost for Tanzania is from the FAO/FADINAP study (1987). It Nigeria (1985)4 67 27 30 is not clear whether the cost refers to transport of large truckloads of fertilizer from ports/factory to farm gate or if transport cost is a Cameroon (1 985)5 22 11 10 combination of rail and road cost. A more recent study (FAO and Senegal (1 984)6 9 NA NA World Bank 1987) shows the weighted average transport cost of Philippines (1985/86) 7 10 4 fertilizers for 1986/87 to be much lower-$72 per metric ton. Thailand (1985/86) 7 12 4 4 Transport cost for Nigeria is for the year 1985. After the devaluation in Argentina (1985/86) 12 29 117 1986,the transport cost in U.S. dollars would be about $15-17. India (1 985/86) 22 39 NA I Cameroon's weighted transport cost of fertilizers to all destinations in the country is for 1 984/85. To serve the Extreme North and North provinces in Cameroon a combination of rail and road transport would cost between US$70 and US$110 per metric ton. Source: Malawi: World Bank 1 987a; Kenya: MOA 1987; World Bank 6 Data on transport cost in Senegal were not available. The figures in the 1986; Nigeria: IFDC 1985a; Cameroon: IFDC 1986a; Senegal: table refer to the cost for transporting rice from the port to the farm Jammeh 1 987b; Tanzania, Philippines, Thailand, Argentina, India: gate in the Groundnut Basin. In 1984 the transport costs for the East FAO*FADINAP 1987. Senegal and Casamance regions were as high as $22.9 per metric ton at official exchange rates. I Transport costs as a percentage of exfactory prices. notable. In the case of cotton in 1980 (see Figure 6 and Table 15 Appendix 9, Table 2), three countries (Cameroon, Tanzania, Nutrient price/crop price ratios for selected crops in East and Malawi) set producer prices at $0.33 per kilogram at Africa 1980-88 purchasing power parity rates, two countries (Nigeria and Arabica Kenya) set prices 36 percent higher (at $0.45 per kilogram), Country Maize Rice Tobacco Coffee Cotton Tea and one country (Senegal) paid about 25 percent less ($0.26 per kilogram). In 1985, the price paid in Senegal was less Malawi than half that paid in Nigeria (See Lele, van de Walle, and 1980/81 8.8 1.0 Gbetibouo 1989). 1981/82 7.8 1.4 For coffee, the comparison between Cameroon and 1982/83 9.1 1.7 1983/84 9.0 1.1 Kenya, both of which produce significant quantities is of 1984/85 99 1.1 interest (see Figure 7 and Appendix 9, Table 3). Producer 1985/86 122 1.0 prices of arabica coffee in Cameroon are only about 50 1986/87 12.5 1.0 percent of those earned by producers in Kenya. This is due 1987/88 10.3 2.3 to a combination of the lower prices earned by Cameroon in the international market, because of poor quality relative Kenya to Kenya's premia, and because of a high rate of taxation. 1981/82 7.2 0.5 3.4 4.1 Together with the much higher maize prices in Cameroon 1982/83 4.5 0.4 3.2 3.0 compared to Kenya, this explains why the ratio of producer 1983/84 5.0 0.3 3.1 1.4 prices of coffee and maize in Cameroon have been one- 1984/85 5.2 0.4 2.9 2.5 fourth of those in Kenya. This also explains the reported 1985/86 NA NA NA NA shift in Cameroon of fertilizer meant for coffee to the 1986/87 3.4 0.2 3.2 2.5 production of maize and other horticultural production 1987/88 4.5 NA NA NA referred to earlier, and explains why the coffee yields in Tanzania Cameroon are only a quarter of those in Kenya. 1980/81 5.6 3.2 0.6 . 1.9 3.1 The nutrient price/crop price ratios are presented in 1981/82 5.4 3.5 0.8 2.5 3.8 Tables 15 and 16. Maize is the only crop for which data are 1982/83 5.1 3.0 0.7 2.4 4.1 available for all the countries and the ratios confirm the 1983/84 4.1 2.2 0.5 1.9 3.1 earlier impression that the relative cost of nutrients to 1984/85 2.2 1.5 0.4 1.5 2.2 output prices is high in East Africa when compared to West 1985/86 4.2 2.8 0.6 2.6 3.9 Africa. Malawi has the most expensive nutrients in terms of 1986/87 5.0 3.3 0.6 2.4 NA the price of maize, while Nigeria has the lowest.82 The 1987/88 5.0 2.9 NA NA NA variation in the nutrient price/crop price ratios is consid- Notes: These ratios are computed using official fertilizer prices that ered later in a discussion of the variability in benefit-cost reflect subsidies, the effect of grant aid fertilizer on cost, and the ratios. official exchange rate. The ratio does not reflect internal transport costs. For more details and sources, see Appendix 10. The nutrient- crop price ratios for maize and rice in Tanzania have been computed for producer prices in the premium areas. For the other areas, the ratios are bound to be still higher. 34 Figure 5 Producer price of maize in the MADIA countries, 1971-87 (converted to US$/ton using purchasing power parity exchange rates) East African countries West Alrican countries S/ton S/ton 340 ._ S340 300 300 260 , . . . 18260 . . ,\.. 220 220 180 ~~~~~180o 140~~~~~~~~~~~~~~~~~~~1 1971 73 75 77 79 8t 83 85 87 1971 73 75 77 79 81 83 85 87 ...... .. Malawi - - Cameroon Kenya ------- Senegal - -f- - Tanzania - - Nigeria Source: See Appendix 9, Table 1. Figure 6 Producer price of cotton in the MADIA countries, 1970/71 to 1986/87 (converted to US$/kg using purchasing power parity exchange rates) East African countries West African countries S/kg $/kg 0.44 0.60 * ~~~~~~0.55 0.4 . Sng Source: See Appendix 9, TableS2. * 0.50 0.36 *0.45 0.32 'i* 0.40. S \ 0.28 *. 0.35 /-a.- '5 024 /0.30 02 **** / ~~~~ ~ ~~~~ ~~0.25 D * j' - as.. S~~~~~~~~~~~~~~/00 o-000 0.1~~~~~~~~~~~~~~~~602 0.16 * ~~~~~~~~~~~~~~~~~~0.15 */-S 0.12 ** * , ,, SS I1 0.10 1 970/71 72373 74/75 76/7 7817 80/81 82/83 84/85 86/17 1970171 72373 74/75 76f77 7879 80/8 MM(8 64/8 817 . ... ..Malawi- - Cameroon Kenya .......Senegal - - -- Tanzania - -Nigeria Source: See Appendix 9, Table 2. 35 Figure 7 Response Coefficients Producer price of coffee in Cameroon and Kenya, 1971-87 We now turn to an analysis of the third and final factor (converted to US$/kg using purchasing power parity affecting the profitability of fertilizer use-response coeffi- exchange rates) cients, a factor that is difficult to analyze because of S/kg considerable variation in the agroclimatic circumstances 4 2 and because of weak and often inconsistent data on . A response coefficients. Often this reflects a lack of appreci- 3.S _ f \ ation among governments and of the fundamental need for 3.6 - I \ highly location-specific and well articulated recommenda- 3.4 - A } tions for fertilizer application, especially for micronutrients. 3.2 - The response data also mask the immense problems of zs I \ / variability of responses around means resulting from the Z6 _ climatic factors referred to earlier.83 Z4 _ / \ { \ { In order to present as complete an analysis of response Z2 _ / \ I \l coefficients as possible, data were collected from several 2 sources, specifically from trials by FAO, the International .I Fertilizer Development Center (IFDC), research and crop 16_ associations in each country, and the World Bank Staff ..4 - J Appraisal Reports (see Tables 17 and 18). Unfortunately, 1.2 _ - - \ o these sources frequently fail to specify a production . > ' - function, so it is difficult to ascertain whether a coefficient 0.8 is to be interpreted as a marginal or average value. Further, 1971 73 .9 31 e13 65 the sources used (i) do not provide a probability distribu- tion of the benefits of fertilizer use in an environment of K _ Cameroon high intra- and interyear rainfall variability; (ii) rarely specify Source:eneeyappedix 9, Table 3.the variety of seed(s) used or specify soil types and Source: See Appendix 9, Table 3. Table 16 Nutrient price/crop price ratios for selected crops in West Africa, 1980-87 Coffee Maize Groundnuts Rice Millet Arabica Robusta Cotton Cocoa Cameroon 1980 2.2 0.3 0.4 1.7 0.5 1981 2.1 0.4 0.4 1.5 0.4 1982 2.4 0.5 0.5 1.6 0.5 1983 NA 0.5 0.5 1.7 0.5 1984 NA 0.4 0.4 1.5 0.5 1985 NA 0.4 0.4 1.4 0.5 1986 NA 0.4 0.4 1.4 0.5 1987 NA 0.4 NA 1.2 0.5 Senegal 1980 1.9 1.5 1.7 1.7 1.3 1981 1.5 1.2 1.3 1.4 1.2 1982 1.5 1.2 1.3 1.4 1.0 1983 2.9 2.8 2.7 2.8 1.8 1984 3.2 3.2 3.2 3.5 2.5 (4.2) (4.2) (4.2) (4.5) (3.2) 1985 2.4 1.9 2.5 2.8 2.1 (4.2) (3.2) (4.4) (4.9) (3.7) 1986 2.6 2.0 2.1 2.6 1.8 (3.6) (2.8) (2.9) (3.6) (2.5) 1987 2.9 2.3 2.3 2.9 2.1 (3.6) (2.8) (2.4) (3.6) (2.5) Nigeria 1980 0.3 0.2 0.2 0.3 0.2 0.1 1981 0.5 0.4 0.3 0.4 0.3 0.1 1982 0.6 0.4 0.3 0.4 0.3 0.1 1983 0.6 0.4 0.3 0.5 0.3 0.1 1984 0.6 0.7 0.4 0.5 0.6 0.3 1985 0.7 0.6 0.4 0.5 0.5 0.3 1986 1.4 0.8 0.5 1.5 0.7 0.4 1987 1.4 NA 0.3 1.3 0.5 0.1 Notes: These ratios are computed using official fertilizer prices that reflect subsidies and the effect of grant aid fertilizer on cost. The ratio does not reflect internal transport costs. For more details and sources, see Appendix 1 0. Fiqures in parentheses are ratios for the unsubsidized price of fertilizer in Senegal. 36 conditions, or the impact of each recommendation; and (iii) but actual field conditions may not be representative as do not consider the gap between on-station and on-farm these fields are under the supervision of research staff conditions, such as the practice of sole versus mixed where the level of crop production management is likely to cropping; the quality of land preparation; the extent of be higher than that used by a typical farmer. So, while using weeding; type, mix, rate, and timing of fertilizer applica- the trial data for any analysis, it is necessary to understand tions; or the timeliness of planting. The reported response that these do not typify those on farmer plots. The coefficients also ignore the influence of crop rotation convention, according to Falusi (1987) is to assume that the patterns on fertilizer yields and do not test the conse- results that the farmers achieve are at best 60 percent of quences of suboptimal application rates on yields. There- those for trial plots. fore, it is often difficult to interpret and compare the There remains the question of crop response to total and limited data that are available. individual nutrients. Normally, there are two ways of looking The conditions and methods by which trials were con- at the response to nutrient use: (i) relate the response to ducted by these sources differed, causing the results to nitrogen, and (ii) relate the response to total nutrients. In vary. The data are from actual field conditions or from countries where the fertilizer types commonly used have N, experimental station plots, are expressed in terms of total P. and K in equal proportion, and where fertilizer recom- nutrients or in terms of individual nutrients (N, P, and K), mendations for farmers are broad-based rather than based and can be either average or marginal. The response data on individual soil analysis, estimates of crop responses to from FAO trials are more consistent because they are based total nutrients are considered more appropriate than on actual field conditions for several years. The trials and responses to individual nutrient (Falusi 1987). If, however, demonstrations in a particular plot, however, usually lasted the analysis is for countries that predominantly use straight for only one year (FAO 1974). Consequently, the results are nitrogenous fertilizers (like Malawi, Tanzania, and Camer- affected by different factors determining nutrient efficiency, oon), it is better to study crop responses in terms of such as potential residual and cumulative effects, particu- nitrogen alone. If the objective, as in the case of IFDC, is to larly with regard to phosphorus and potassium. On the define the optimum rates of application of N, P, and K for other hand, most of the IFDC trials and the National different regions, then crop responses must relate to Research Institution trials were in experimental stations, individual nutrients. Given the fundamental differences in lasting between a single season and 3 years, on mono-crop the nature of the data, comparisons of response coefficients farms under ideal crop management. Research trials of crop are limited to instances where they are wholly comparable. associations are often carried out in actual farmers' fields, Finally, there are differences between FAO data that are Table 17 Response coefficients for selected crops in East Africa East Africa Maize Sorghum Tea Coffee Rice Wheat Country Local Hybrid Green Arabica Kilograms of Output Per Kilogram of Nutrient' Malawi (ASA) 16.6 29 (FAO) - 20-37 (WB) 14 30 Kenya 10.4 East of Rift Valley 30-35 West of Rift Valley 15-20 (GOK) HPD 15-26 18-21 MPD 10-21 5-19 LPD 9-14 4 (FAO) (WNP) 15-17 (RVP) 12-22 (C&EP) 16-25 Tanzania (FAO)2 Existing Practices 13.5 10 13.2 11.9 Improved Management 11.53 12.8 11.53 4.83 (World Bank) 6 16 Notes: ASA - Annual Survey of Agriculture. GOK - Government of Kenya. *LPD - Low Potential Districts. FAO - Food and Agriculture Organization. HPD - High Potential Districts. WNP - Western and Nyanza provinces. WB - World Bank. *MPD - Medium Potential Districts. RVP - Rift Valley province. See Appendix 11 for classification of high, medium, and low potential districts. C&EP - Central and Eastern provinces. 1 See Appendix 1 1 for an explanation of the actual nutrients. 2 FAO figures are for medium to high potential districts. 3The yields under improved practices are higher. However, data suggest that crop responses to fertilizers under improved practices are lower than under existing practices. Thus, the extent to which agricultural extension is a substitute for fertilizer use rather than a complement to it needs serious further analysis based on strong empirical research. For detailed Notes and Sources: See Appendix 11. 37 Table 18 Response coefficients for selected crops in West Africa Nigeria Cameroon Senegal FAO Falusi World Bank Crops/region LP IP 1976-80 1984/85 Crops/region IFDC FAO Crops/region IFDC FAO Kilograms of Output Per Kilogram of Nutrient' Maize Maize 7.3 Maize NA Derived Savannah 4-11 Coastal Lowlands 32.1 S. Guinea Savannah 5-12 6-7 Guinea Forest (Ntui) 3.8 Forest 7-18 6-14 Maize after cotton 30.8 Sudan Savannah 6 8 5.6 Maize after groundnut 20.9 N. Guinea Savannah 6-10 8 Sorghum Sorghum 3.9 Sorghum 3.8 Sudan Savannah 3-8 2.5-7 8.5 2.5-3 Northern Plain (HYV) 7-30 North Sine-Saloum 4.3 N. Guinea Savannah 4-8 2-7 5-7 8 South Sine-Saloum 5.8 S. Guinea Savannah 5-9 5-12 3-7 Groundnut Groundnut NA Groundnut 6.6 Sudan Savannah 7-13 1.5-3 North Basin 4-6 N. Guinea Savannah 9-17 8-15 11-13 Central Basin 5-8 S. Guinea Savannah 10-21 9-10 North Sine-Saloum 7-9.5 South Sine-Saloum 8-11 Millet Millet NA Millet 7.03 Sudan Savannah 3-11 2.5-6 2.5 North Basin 14-20 N. Guinea Savannah 7-13 3-9 Central Basin 15-17 S. Guinea Savannah 13-21 North Sine-Saloum 17-20 South Sine-Saloum 17 Rice (Upland) 7-12 6 Rice (HYV) Rice Derived Savannah 4-11 Northern Plain 12-39 Casamance 5.60 Forest 3-13 Rice (Swamp) S. Guinea Savannah 4-7 3-8 5-6 Wheat (Irrigated) Coffee Sudan Savannah 3-11 Arabica 5-6 Robusta 2-3 Yam 30 14 Cassava 46 20-32 Cowpeas 9-16 2-13 15-18 1-3 Notes: ' See Appendix 11 for an explanation of the actual nutrients. LP - Local Practice. IP - Improved Practice. Sources: See Appendix 11. often average responses (derived from the difference in Despite the incomplete state of information on food and yield between the fertilized and control plot) and IFDC export crop research in Africa, including trial results that data that are marginal responses. The approach used in have not been made publicly available, the crop response determining physical response by IFDC has been to take data on the MADIA countries (see Tables 17 and 18) allow raw agronomic results from research stations and develop some insights into the agronomic circumstances of country response curves (simple quadratic production function) and/or region as well as the level of available agricultural using regression technique. The differences in the FAO and technology. Generally, these data convey that for maize, IFDC crop responses can also be attributed to the differ- responses to fertilizer use in the high potential areas of ences in their objectives for conducting fertilizer trials. Most East Africa are similar to those in certain areas of the FAO trials are conducted and reported in terms of yield highlands of Cameroon and the rainforest zone of Nigeria. increase over the unfertilized "control" plot for specified With respect to arabica coffee, however, varieties in Kenya fertilizer treatments. Trials by IFDC, on the other hand, are are reported to be roughly twice as responsive to fertilizer primarily for decisions concerning intensity of fertilizer use as in Cameroon. It is difficult to gauge the extent to which or alternative allocations of fertilizer use among competing the favorable soil and climatic conditions in Kenya or the crops. Because the FAO and IFDC figures are not compara- excellent research system for coffee explains these high ble, we have presented response data for the countries responses. from all sources, presenting a range of values rather than a Regional variation in responses within each country single value. means that, in Kenya, maize response coefficients in the 38 high potential areas are twice those in the low potential per ha. Fewer trials have been possible in Arusha or semiarid areas (see Tables 17 and 18).84 Sumbawanga but indications are of responses to both In Malawi, there is less reported regional variation in nitrogen and phosphate in Sumbawanga. In Songea response coefficients than in either Kenya or Tanzania, responses to high levels of N have been obtained. In perhaps reflecting Malawi's single rainfall season compared the Uyole area responses have been significant up to to the bimodal distribution in Tanzania and Kenya. None- 80 kg N per ha. with a greater response when N and theless, as stated earlier, maize as a cash crop is mostly P were both applied. No response was obtained to concentrated in the central part of the country, where phosphate alone or to potassium (Spurling 1982). responses are higher (i.e., Lilongwe and Kasungu districts). With respect to wheat, response coefficients are believed to Since socioeconomic conditions (the existence of a larger be low although no concrete data are available. Even in number of more commercial smallholders) are also more 1982, the NAFCO farm used no fertilizer on its 20,000 suitable than in the Southern region, it is difficult to say hectare mechanized wheat complex at Hanang in Southern precisely the role of hybrid responses and economic ability Arusha (World Bank 1983a). Trial results for cotton in Geita to undertake risks in adoption. Data also suggest that and Sengerema districts-the predominantly cotton-grow- response to fertilizer on hybrid maize in Malawi is higher ing areas-showed significant responses to nitrogen but not than in many of the high potential districts of Kenya- enough to make fertilizer use economic, even at the notable exceptions being Embu, Muranga, and Kiambu, subsidized prices that prevailed before 1984 (World Bank where the reported maize response coefficients are compa- 1979). Yet other experimental station data at Ukiruguru rable to those in Malawi. The fact that so little hybrid maize show a sharp decline in cotton yields on plots that have is grown in Malawi-less than 8 percent of the total output been continuously cultivated for well over a decade, compared to 60 percent in Kenya-must thus be explained suggesting acid soils and a need for a soil management by several factors discussed earlier, rather than the respon- program to maintain soil fertility beyond the simple siveness of hybrid maize.85 application of fertilizer. The decline in soil fertility and average yields in Malawi Data on tea responses to fertilizers are also very poor. referred to earlier also seems to apply to the response The available information indicates that fertilizer recom- coefficients on hybrid maize which declined from 23 to 13 mendations on tea were not based on any formal research between 1957-62 and 1982/83-1984/85. Similarly, Kasunga, trials in Tanzania, but on general recommendations made Salima, and Mzuzu Agricultural Development Districts by the Tea Research Foundation for East Africa (Project (ADDs) showed declines in the response coefficients of Completion Report 19791. For a country like Tanzania where hybrid maize from 24 to 18, 25 to 17, and 32 to 18, the soils in most parts of the country have pronounced respectively, over the same period (Twyford 1988). This topsoil and/or subsoil acidity and are of low fertility (except decline has again been attributed by certain agronomists to for the volcanic areas of Kilimanjaro and Meru and the a change from hybrid variety SR52 (Zimbabwe origin) to alluvial valleys of Ruaha-Kilombero-Rufiji), there is an MH 12 (Malawi origin) or solely to decreasing organic matter urgent need for systematic research on soils and fertilizer and phosphates in the soil.86 trials as well as improved resource management to lay a Reflecting the short history of agronomic research in sound basis for the formulation of fertilizer policy. Tanzania, as compared to Kenya, and the subsequent In Senegal, crop response data in the farmers' fields are erosion of agricultural research, data sources on fertilizer primarily available for the Groundnut Basin from IFDC for responsiveness are the weakest for Tanzania and are not 1976 and 1977 and from research station trials conducted by available by agro-ecological zones. The weakness of some ISRA. Some results of fertilizer demonstration trials con- of the trial data can be seen in the FAO experimental plots ducted on rice by FAO in the 1960s outside the Groundnut for maize, rice, and wheat. There, the responses to fertiliz- Basin are also available. The two main issues on crop ers-even in the medium to high potential regions-under response are: (1) the obvious problems of yield variability improved management are less than under existing prac- and farmer unwillingness to take risks when response tices, although in terms of total output, yields under coefficients are positively correlated with the declining, improved management are higher (Mhella 1985).87 A recent variable rainfall levels; and (2) the differences between ISRA study by CYMMIT in the less fertile regions of the Southern and IFDC on the question of crop responses to certain Highlands, however, suggests that hybrid maize responses nutrients and the differing fertilizer recommendations that to fertilizers were 1:9. On soils of higher fertility in the result. Southern Highlands the responses are thought to be much An example is the response of millet and groundnuts to higher. For example, a significant response to fertilizers in phosphates. In the north (Louga) and central (Diourbel and the high altitude regions, including the Southern Highlands, Thies) areas of the Groundnut Basin with 350-600 milli- is also noted by the national maize research program from meters of annual rainfall, the yield is one-third to one-half its fertilizer trials: less than that in the southern part of the Groundnut Basin In the western part of the country, nitrogen has given (e.g., Kaolack/Fatick), where rainfall averages 600-800 milli- a consistent response but the only regular response meters per annum. Crop response data for the Ziguinchor/ to phosphate has been in the Tabora area. In the Kolda and Tambacounda regions, where rainfall levels are Kilimanjaro area at an altitude above 1,000 m.a.s.l. a greater than 800 millimeters, are not available to determine significant response to nitrogen was obtained in the the extent to which responses increase in the assured and intensively cropped NAFCO farms. In Iringa and high rainfall areas, relative to the areas in the southern part Mpwapwa areas both nitrogen and phosphate had to of the Groundnut Basin. There are fertile lands concen- be applied to give the maximum response with trated in small river valleys which are believed to be of very economic levels at 30 kg P2O, and 20 kg N per ha. In high agricultural potential and richer in organic matter than the high altitude high rainfall areas both N and P have most parts of the Groundnut Basin.88 The information on given economic responses, the former up to 120 kg N their productivity is crucial for planning future agricultural 39 development policies. whereas the FAO estimates refer to average responses, The issue of crop responses and fertilizer recommenda- those from IFDC relate to responses at the margin, esti- tions in Senegal is more complex than in the other mated through yield functions. Thus the translation of countries, with IFDC and ISRA making different fertilizer results into location-specific fertilizer recommendations is recommendations for millet and groundnuts (see Appendix still at an early stage of development in Cameroon. As in 11, Tables I and 2).89 The reduced doses recommended by Senegal, these data problems underscore the need for IFDC on these crops were criticized by ISRA on the grounds extensive long-term trials using a comparable methodology. that IFDC results were based on short-term trials (two Without such information even the most elementary judg- years) and on the principle of profit maximization but were ments about fertilizer policy are at risk of being wholly indifferent to the long-term impact of fertilizer use on soil ineffectual or perhaps even damaging. In this context the fertility. There has since been some reconciliation, with issue of changing donor priorities must be reiterated. ISRA moving in the direction of IFDC.90 The issue of USAID's food crop research project with the participation of groundnut responses to potash, however, remains unre- ITTA is doing an excellent job of helping the Cameroonian solved, with scientists at ISRA strongly urging potassium government in technology development. It is hoped that application and IFDC recommending total elimination of both USAID and ITTA will maintain the long-term horizon potassium in all regions except southern Sine-Saloum. ISRA needed to support Cameroon's efforts, but this is by no also strongly favors the application of sulphur for both means certain. ITTA does not perceive its mandate as one groundnuts and millet. Such differences in recommenda- of developing Cameroon's research capacity rather than tions must lead to the conclusion that a long-term research carrying out its own research. It is also likely that the recent effort is necessary to determine appropriate types and focus on privatization and policy reforms within USAID will amounts of fertilizers. Short-term trials, such as those divert its attention to new priorities. conducted by IFDC and supported by donors, frequently Nigeria has perhaps the most complex and controversial raise more questions than they answer, especially as to the set of issues concerning crop responses to fertilizers. This weight to be given to soil maintenance objectives and is not only because the data from different sources are short-run economic considerations. Unfortunately, despite a inconsistent, but also because there has until recently been $105 million agricultural research project funded by donors a general disagreement among experts on how to interpret in Senegal, owing to the extreme shortage of recurrent the data that are available. Further, the number of agrocli- resources and the lack of identification of research priori- matic zones and crops serves to complicate the formulation ties by Senegalese scientists, it has not been possible to of recommendations and policy. (A summary of the coeffi- undertake fertilizer response trials on farmers' fields where cients from various sources for crops and regions is declining soil fertility is a serious problem (Jammeh and provided in Table 18.) The crop responses reported by FAO Lele 1988; Khan and Palmier 1989). (based on trials conducted under actual field conditions) The IFDC/ISRA disputes on responses are part of a and Falusi (who has compiled the most systematic informa- broader malaise in donor assistance of hiring external tion on a range of coefficients from different sources, agencies on a short-term basis to carry out analysis and including documents' on the World Bank's projects and resolve policy disputes on national issues in which an trials carried out by IAR) are similar (Falusi 1987), and show! understanding of complex and long-term interactions that fertilizers are more responsive in the Guinea Savannah between soils, climate, and farmer practices is needed. region (with rainfall between 1,000 and 1,500 millimeters) African governments must share responsibility for not than in the Sudan Savannah (with rainfall ranging between supporting their own indigenous research efforts and 500 and 1,000 millimeters).92 93 In contrast the estimates scientists on a continuous basis and for not providing made by the ADP project staff and accepted by the Bank at financial and other incentives to their nationals for conduct- different times are presented separately, in order to ing research and participating in the complex process of demonstrate the cycles of optimism and pessimism that policy formulation and refinement. It is this situation that have characterized perceptions about technology availabil- has led the Senegalese scientists to conclude that they ity in Nigeria, and in which views on the fertilizer responses essentially serve the function of skilled labor furnishing of individual crops have played a central part. The response international agencies rather than pursue their own pro- coefficients applied by the Bank during the 1976-80 period, gram of work (Khan and Palmier 1989). Unless donor when most of the enclave and statewide ADPs were attitudes change drastically in favor of long-term indigenous appraised and/or implemented, characterize the optimism capacity building and unless African governments take that prevailed at the time about the availability of improved primary responsibility for protecting and nurturing their technologies, and the potential ability of extension to own science and technology capacity by giving it the convert farmers to sole cropping (Lele, Oyeiide, et al. necessary esteem, the lack of knowledge on fertilizer 1989).94 On the other hand, the coefficients presented for responses and more generally on technology issues will 1985/86 characterize the pessimism introduced by the mid- continue. term reviews of the Bauchi, Kano, and Sokoto statewide In Cameroon food crop research is relatively new. ADPs.95 These reviews concluded that the response of Response trials have been carried out on major cash and fertilizer was low in crop mixtures dominating Nigerian food crops by FAO, IFDC, and IAR over the last decade but, farming practices. The response coefficients formulated by as elsewhere, trial results are not comparable. A partial the mid-term review were 2.5 for millet, 2.5-3 for sorghums, explanation for this is the tendency of donors to support and 1-3 for groundnuts and cowpeas. The FAO/Falusi short-term trials that produce a limited amount of fertilizer responses, in comparison, depending on the area, ranged response data (IFDC 1985). Most often the results were not between 3 and 11 for millet, 3 and 8 for sorghum, 7 and 13 replicated or standardized and show vast differences for groundnuts in the Sudan Savannah region, and 9 and 16 among sources.91 .- - not possible to explain the precise for cowpeas.96 These low responses of the mid-term reviews cause of the differences because the seed varieties used were contested by the Nigerians (Idachaba 1987).97 (for example, in the FAO trials) are not reported. Also, The inability of the Bank and the government of Nigeria 40 to reach agreement on this issue hindered the formulation itself.) This includes the cost of transporting fertilizer to the of policies that are needed to address pressing problems farm gate and the additional labor time needed to apply faced by farmers. For instance, as stated earlier, compound fertilizer and for incremental weeding due to fertilizer use. fertilizers (15:15:15) continue to be predominant. However, Therefore, the critical or threshold level of the benefit-cost recent studies demonstrate a phosphorus deficiency in ratio neded to make fertilizer use attractive is higher than most parts of Africa, including Nigeria, and the need for otherwise. more phosphatic fertilizers. IFDC observed, "Phosphorus A fourth factor affecting the reliability of benefit-cost deficiency in tropical African soils is a major factor limiting ratios is the risk of nutrient price and/or output price food production" (1985b, p. 151.98 Recently the Bank has changes over time. The magnitude of the risk can be voiced serious concerns about declining soil fertility and demonstrated by calculating the coefficient of variation for the need to supply appropriate types of fertilizers. The international fertilizer prices and using this as a proxy for mid-term reviews argued that the present fertilizer (15: 15:15) the standard deviation of domestic fertilizer prices.'02 This, was inappropriate for the soils in northern Nigeria and that in turn, can be used to estimate the impact of variations in the government should use a composition with more international fertilizer prices on benefit-cost ratios. Table 19 nitrogen and less potash recommended by its research gives the coefficient of variation for the price of major institutions.99 In Nigeria, as elsewhere, systematic long-term fertilizer types between 1975 and 1985, indicating an agricultural research and analytical capacity for addressing average deviation of about 23 percent of price. these problems could help resolve some of the contentious Using the coefficient of variation of international fertilizer issues surrounding technological packages and the efficient prices as a proxy for the standard deviation of domestic use of fertilizer. The high priority to agricultural research- fertilizer prices, it is possible to compute a range of rather than ad hoc trials in ADPs-is long overdue given that benefit-cost ratios. It is then possible to assign a probability large amounts of resources (over $1.7 billion between 1971 (associated with an observation being I or 2 standard and 1988) have been committed to Nigerian agriculture by deviations from the mean) to the occurrence of a particular the World Bank alone.'00 In total, the Nigerian government benefit-cost ratio in a particular year owing to fertilizer has spent It billion naira since the oil boom. A research price fluctuations (see Table 201. (These benefit-cost ratios project is currently under preparation but the most recent use 1986 fertilizer and crop prices as mean values without agricultural sector report (1989) has not emphasized the adjusting for subsidies or currency overvaluation.) This fundamental importance of agricultural research. exercise demonstrates that there is a very significant effect Greater efficiency of fertilizer use can also be achieved on the profitability of fertilizer use of a relatively large by promoting congruence between the types of fertilizer coefficient of variation for fertilizer prices (25 percent). produced and consumed in Nigeria (Lele, Oyejide, et al. A similar exercise performed for output prices (see 1989). The fertilizer plant at Onne (Port Harcourt), which Tables 21 and 22) by using international primary commodity went into operation in 1987, has the capacity to produce price fluctuations for tradeable goods and domestic pro- 220,000 tons each of ammonia and NPK, and 495,000 tons of ducer prices for crops traded regionally shows substantial urea. Notably, in the context of the IFDC recommendation fluctuations. concerning the greater use of phosphatic fertilizers, the Despite these weaknesses, benefit-cost ratios are useful plant's capacity can be converted to produce over 600,000 for assessing lil the impact of subsidy removal and currency tons of DAP per year (IFDC 1988). revaluation on the economic returns to fertilizer use, and (ii) the general profitability of fertilizer use. They are computed Benefit-Cost Ratios for Fertilizer Use using nominal input and the output price data that reflect We now turn to the benefits of fertilizer use. There are taxes and/or subsidies and are shown in Tables 23 and 24 several limitations to this analysis. First, the focus is on the for all six countries.'03 Where applicable, a second benefit- immediate economic benefits. The long-run benefits are cost ratio is computed (labeled "Without Explicit Subsidy" not easily quantifiable. Even the short-term analysis has in Tables 23 and 24), which demonstrates the effect of many limitations and this must be stressed at the outset. removing the explicit subsidy. In those countries where For example, all price data are national and only provide currency overvaluation is serious (Tanzania, Nigeria, Camer- national values. The benefit-cost ratios presented here do oon, and Senegal), a third set of benefit-cost ratios (labeled not explicitly consider transport costs, of either fertilizer or "Without Explicit or Implicit Subsidy") takes into consider- agricultural output to the farm gate. This can have a ation the removal of both explicit and implicit subsidies.'04 significant impact on the actual value of the benefit-cost Because of the relatively high cost of nutrients in Malawi, ratios where depot to farm gate costs are substantial.'0I even though the response coefficients for hybrid maize are Second, although a benefit-cost ratio greater than I higher than those for the low and medium potential areas indicates that fertilizer use is profitable, it is difficult to ascertain the threshold value that makes fertilizer use attractive enough to farmers to compensate for the inherent Table 19 risk in its use, for example, the failure to get enough rain to Coefficient of variation in international fertilizer prices, make fertilizer use profitable. Typically a ratio of 2 is used 1975-85 as the critical value under favorable conditions; however, as Type of Fertilizer Coefficient is discussed in this paper, the number of factors affecting farmers' perceptions of risk is large and often difficult to Urea 0.21 quantify (World Bank 1986b). DAP 0.20 A third factor that influences the interpretation of the Rock Phosphate 0.26 benefit-cost ratios is the fact that the substantial additional TSP 0.24 labor cost of using fertilizer is not included. (Carr (1989) Potash 0.25 estimates the cost of incremental labor for Tanzania in 1989 Source: International Fertilizer Prices for 1975-85 from British Sulphur to be as high as 50 percent of the full cost of the fertilizer Corporation 1987. 4 Table 20 Table 21 Estimated range of benefit-cost ratios due to changes in Coefficient of variation in commodity prices international fertilizer prices Commodity Country Coefficient Range of benefit-cost ratios Country Commodity ± 1 S.D. ± 2 S.D. Maize Maiawi (1971-87) 0.53 Kenya (1971 -87) 0.52 Malawi: Hybrid Maize 1.6-2.7 1.4-4.7 Tanzania (1971-87) 1.09 Local Maize 1.0-1.6 0.8-2.4 Cameroon (1971-82) 0.45 Kenya: Hybrid Maize' 4.7-7.8 3.9-11.7 Senegal (1971-85) 0.39 Hybrid Maize2 3.5-5.9 2.9-8.8 Nigeria (1976-87) 0.46 Hybrid Maize3 2.6-4.3 2.1-6.4 Rice Tanzania (1971-87) 1.14 Tea (green leaf) 8.1-13.6 6.8-20.4 Cameroon (1975-85) 0.30 Coffee 39.3-65.6 32.8-98.5 Senegal (1971-86) 0.38 Tanzania: Maize 1.2-2.0 1.0-3.0 Nigeria (1976-87) 0.76 Rice 1.8-3.1 1.5-4.6 Millet/Sorghum Senegal (Mil/Sor) (1971-86) 0.34 Wheat (H) 1.4-2.3 1.2-3.5 Nigeria (Mil) (1976-87) 0.47 Wheat (L) 0.6-1.0 0.5-1.4 Nigeria (Sorg) (1976-87) 0.53 Cameroon: Maize 11.2-18.7 9.3-28.0 Groundnuts Senegal (1971-87) 0.44 Rice (H) 18.7-31.2 15.6-46.8 Nigeria (1971-86) 0.61 Rice (L) 7.5-12.5 6.2-18.7 Tea International (1 975-85) 0.23 Coffee 7.4-12.3 6.2-18.5 Coffee International (1975-85) 0.29 Senegal: Sorghum 1.6-2.6 1.3-3.9 Groundnuts (H) 4.0-6.7 3.3-10 Wheat Tanzania (1971-87) 0.94 Groundnuts (L) 1.6-2.7 1.3-4.0 Source: International crop prices for tea and coffee from Commodity Millet 4.7-7.9 3.9-11.7 Trade and Price Trends 1986. Commodity prices for nontradeable Rice 1.9-3.2 1.6-4.7 goods are producer prices from World Bank Database. Nigeria: Maize (H) 5.7-9.5 4.8-14.3 Maize (L) 2.9-3.8 1.9-5.7 for mainly because of the failure to convert very small Sorghum (H) 2.7-4.5 2.3-6.8 Sorghum (L) 0.9-1.5 0.8-2.3 farmers to using hybrid maize. Groundnuts (H) 20.7-34.5 17.2-51.7 In Kenya, all of the regions demonstrate profitability in Groundnuts (L) 10.3-17.2 8.6-25.9 using unsubsidized fertilizers. In the arid and semiarid Millet (H) 5.2-8.7 4.3-13 areas, however, the profitability is marginal, which as noted Millet (L) 1.6-2.6 1.3-3.9 earlier, raises questions about how to encourage intensifi- Notes: (H) Range of benefit-cost ratios for high crop response. cation of food production in these areas. A subsidy on the (L) ( Range of benefit-cost ratios for low crop response. transport of fertilizer to these areas may be justified. Benefit-cost ratios for crop response in the high potential districts. In Tanzania, fertilizer and crop prices have moved up 2 Benefit-cost ratios for crop response in the medium potential districts. drastically since 1984. Benefit-cost ratios for three different Benefit-cost ratios for crop response in the low potential districts. years are computed to show the extent to which profitabil- Source: Crop prices from World Bank Database. Crop responses: See . year are r ts how thes exten to which profitabl Tables 17 and 18. International Fertilizer Prices from British Sulphur Ity In fertilzer use Is sensitive to changig fertilzer and Corporation 1987. Local fertilizer prices are from the following crop prices. The first set of benefit-cost ratios is for 1984, sources: Malawi: World Bank Internal Memo dated August 23,1988; when an explicit subsidy on fertilizers prevailed in Tanzania. Kenya: Tisminieszky and Kimuyi 1986; Tanzania: FAO and World Since then the direct official subsidy has been abolished Bank 1987; Cameroon and Senegal: Personal communicationwith and fertilizer grants, which cover virtually all imported government officials; Nigeria: IFDC 1988. wt n etlzrgat,wihcvrvrulyalipre fertilizers, were counted as commercial imports.'05 The second set of benefit-cost ratios are therefore computed for of Kenya, the benefit-cost ratios (using 1987 official maize the price of fertilizer in 1986/87, the same year that prices and subsidized fertilizer prices) are only compar- producer prices of crops increased threefold over their 1984 able to Kenya's low potential areas (see Table 23). The levels. After the abolition of the subsidy in 1984, the full dramatic differences in the profitability of hybrid and local cost of fertilizer was passed to producers. Subsequently, maize are also seen from Table 23. The benefit-cost ratios however, prices have not been adjusted upward for deval- for local maize (using 1987-88 official crop prices and uation. A combination of currency devaluation, increases in subsidized fertilizer prices) are less than 2, which helps to the c.i.f. price of fertilizers, and increases in internal costs explain the low application rate on local maize. For a high that have not been reflected in fertilizer prices has meant analysis fertilizer, for example, urea used with DAP, which that a significant subsidy is again in place in Tanzania. The has been promoted for maize since 1987/88, fertilizer use third set of benefit-cost ratios are computed for 1988/89 would be marginally profitable, both with and without a fertilizer prices that farmers actually pay-the subsidized subsidy-ranging between 2.1 and 2.4 with subsidy and 1.6 price-and for the estimated full cost of grant fertilizer to and 1.9 without a subsidy (see Table 25). The substantial the TFC depots.106 increases in the official maize price for the 1988/89 crop In 1984, for the subsidized price of fertilizer and producer resulted in a benefit-cost ratio of above 2 for local maize, price of crops, the benefit-cost ratios were more than 2 for even without a subsidy. For hybrid maize, use of high hybrid maize, rice, and wheat in most regions. Fertilizer use analysis fertilizer gives a benefit-cost ratio above 3.5, even on local maize, which forms 95 percent of the area under without a subsidy, for the years 1987/88 and 1988/89, maize in Tanzania, however, was unprofitable. Recognizing stressing the need to understand urgently the factors that the market prices of crops are significantly higher than explaining slow adoption. In Malawi, a subsidy is thus called official prices, fertilizer use on local maize may have been 42 Table 22 In Nigeria, the profitability of unsubsidized fertilizer Estimated range of benefit-cost ratios due to (15:15:15) use on maize at 1987 prices depends on the area changes in commodity prices of the country and the set of response coefficients that Range of benefit-cost ratios applies. This is true for the other two major food crops, Country Commodity I 1 S.D. ± 2 S.D. sorghum and millet, as well. With the current rates of subsidy in Nigeria, however, the computed benefit-cost Malawi: Hybrid Maize 1.1-3.7 0.0-4.9 ratios for maize are the highest among the MADIA countries Local Maize 0.5-2.2 0.0-3.0 (taking the highest response coefficients reported by the Kenya: Hybrid Maize' 2.8-9.0 0.0-12.0 FAO)-about four times the ratios at subsidized prices for Hybrid Maize2 2.1-6.7 0.0-9.0 hybrid maize in Malawi and even greater than the benefit- Hybrid Maize3 1.6-4.9 0.0-6.6 cost ratios for the highest potential areas in Kenya. This is Tea (leaf) 4.6-15.9 0.0-21.5 because in Nigeria the average market price of maize has Coffee 12.3-59.8 0.0-83.6 been significantly higher, and fertilizer prices much lower Tanzania: Maize 0.0-6.5 0.0-9.9 because of the subsidy, than in Kenya or Malawi. In 1987, Rice 0.0-11.7 0.0-17.9 the benefit-cost ratios for Nigeria, at the lower end of the Wheat (H) 0.0-6.6 0.0-4.1 reported range of response coefficients of FAO, are near 3 Wheat (L) 0.0-2.8 0.0-9.8 for maize and sorghum and 2 for millet (see Table 24). (If the Cameroon: Maize 3.9-10.2 0.0-13.3 benefit-cost ratios are computed only in terms of the Rice (H) 8.2-15.2 4.7-18.7 nitrogren component of 15:15:15, then the benefit-cost ratio Rice (L) 3.3-6.1 4.7-18.7 would decline to slightly above I for maize, i.e., 1.1-1.3, to Coffee (Robusta) 6.6-11.9 3.9-14.6 0.5-0.6 for sorghum, and 0.7 for millet.) Of the total Senegal: Sorghum/Millet 3.8-7.8 1.9-9.8 cultivated area in Nigeria, these three crops constitute Groundnuts (H) 2.8-7.2 0.5-9.4 nearly 70 percent. Without the fertilizer subsidy the benefit- Groundnuts (L) 1.1-2.9 0.0-3.7 cost ratios for all these crops can be less than 1, depending Rice 1.5-3.2 0.5-4.1 on the estimate of response coefficient that is used. This Nigeria: Maize (H) 4.2-11.3 0.5-14.8 suggests that a subsidy of the order of 50 percent of the Maize (L) 1.7-4.5 0.0-5.9 fertilizer price may be adequate to provide an incentive for Sorghum (H) 2.1-6.9 0.0-9.3 - all crops. Underlying the complexities in the range of Sorghum (L) 0.7-2.3 0.0-3.1 benefit-cost ratios resulting from the widely different Groundnuts (H) 16.1-66.6 0.0-91.9 responses for each crop reported by different sources, Groundnuts (L) 8.1-33.3 0.0-45.9 there is the issue of whether fertilizer use is profitable at Notes: (H) = Range of benefit-cost ratios for high crop response. all under the conditions of mixed cropping predominant in (L) = Range of benefit-cost ratios for low crop response. Nigeria. For instance, there is no consensus over the 'Benefit-cost ratios for crop response in the high potential districts. question of whether existing hybrid varieties respond well 2 Benefit-cost ratios for crop response in the medium potential districts. to fertilizers under mixed cropping or if new varieties have 3 Benefit-cost ratios for crop response in the low potential districts. Source: International crop prices for tea and coffee from Commodity to be developed, emphasizing the fundamental importance Trade and Price Trends 1986. Commodity prices for nontradeable of research. goods are producer prices from World Bank Database. Fertilizer Comparing Kenya and Cameroon for coffee, the impact of prices: Malawi: World Bank Internal Memo dated August 23,1988; Kenya's higher response coefficients and Cameroon's low Kenya: Tisminieszky and Kumuyi 1986; Tanzania: FAO and World producer prices is evident from the computed benefit-cost B3ank 1987; Cameroon and Senegal: Personal communication wfhh government officials; Nigeria: IFDC 1988. Crop responses: See ratio. Fertilizer use on robusta coffee in Cameroon at actual Tables 17 and 1 8. producer prices is barely profitable without a subsidy. The return to fertilizer use on arabica coffee in Kenya, on the profitable in certain areas. At 1987 fertilizer prices, without other hand, is twelve times higher than in Cameroon at the an explicit or implicit subsidy, fertilizer was marginally unsubsidized fertilizer price, but at the subsidized fertilizer profitable on hybrid maize, but for local maize it was only price in Cameroon, the benefit-cost ratios for Kenya are 0.8. Rice continued to be profitable with fertilizer use at only three times higher. 1987 prices without an explicit or implicit subsidy on In Senegal, the consequences of subsidy removal differ fertilizer. In 1988, for the fertilizer prices that farmers between regions because the crop responses are different. actually pay, including an implicit subsidy, benefit-cost IFDC data reported for major crops in different parts of the ratios were above 2 only for hybrid maize, rice, and wheat Groundnut Basin indicate that millet is the only crop that is in some areas. For local maize the benefit-cost ratio was profitable at unsubsidized prices in all parts of the only 1.3. The ratios when computed for the estimated full Groundnut Basin. For sorghum, using FAO crop response cost of fertilizer, referred to as "without subsidy" in the data, with no explicit or implicit subsidy, the benefit-cost table, were less than 2 for all crops including hybrid maize. ratio is below 2 for all of Senegal. IFDC data for sorghum, For local maize it was as low as 0.5. (Recall also that these available for North and South Sine-Saloum, also show data do not take into consideration the incremental labor similar results.l08 Removal of the explicit subsidy means costs associated with fertilizer use.) Therefore, especially fertilizer use on groundnuts may be totally uneconomic in when the great majority of farmers use local maize and the the northern Groundnut Basin but marginally profitable in existing local maize responses to fertilizers are low, the the South, where it could be given a priority which is not elimination of the subsidy would seriously affect fertilizer now well articulated in Senegalese policy.'09 For rice, the use. A recent case study on the economics of fertilizer use removal of subsidies will result in. a benefit-cost ratio below in the Iringa and Mbeya regions-the predominantly maize 2 in Casamance, but a more appropriate exchange rate in growing areas-have come out with similar results, suggest- Senegal that would increase prices of imported rice may ing the need for reviewing the roles of subsidies on change this picture.110 Because of lack of crop response transportation so as to encourage private trade.'07 data for groundnut, millet, and sorghum for the Ziguinchor/ 43 Table 23 Benefit-cost ratios for fertilizer use in the East African MADIA countries Kenya' 1987 Estimated by: Maize Green tea Arabica coffee Goverment 37 East of Rift - 12-14 West of Rift - 6-8 High potential land 3.4-5.92 - Medium potential land 2.3-4.8 - Low potential land 2.0-3.2 - FAO Western & Nyanza 3.5 - Rift Valley 2.7-4,9 - Central & Eastern 3.6-5.7 - Malawi 1987-88 Estimated by: Local maize Hybrid maize With subsidy Without explicit With subsidy Without explicit subsidy subsidy Government (ASA) 2.0 (2.6)3 1.6 (2.0) 2.6 (4.6) 2.9 (3.7) FAO - 1.9-3.6 (2.4-4.5) 1.5-2.9 (1.9-3.6) World Bank 1.3 (1.7) 1.1 (1.3) 2.9 (3.7) 2.3 (2.9) Tanzania 1984 Esimated by: Local maize Hybrid maize Rice Wheat with subsidy with subsidy with subsidy with subsidy FAO - 3.4-4.0 6-7 1.9-4.8 World Bank 1.5 3.9 Tanzania 1987 Estimated by: Local maize Hybrid maize Rice Wheat Without Without Without Without Without Without Without Without explicit explicit explicit explicit explicit explicit explicit explicit subsidy or implicit subsidy or implicit subsidy or implicit subsidy or implicit subsidy subsidy subsidy subsidy FAO - - 2.7-3.1 1.5-1.8 4-4.6 2-3 1-3 0.7-1.8 World Bank 1.2 0.8 3.2 2.1 - - - Tanzania4 1988 Estimated by: Local maize Hybrid maize Rice Wheat With Without With Without With Without With Without subsidy subsidy subsidy subsidy subsidy subsidy subsidy subsidy FAO - - 2.9-3.4 1.0-1.2 4.7-5.3 1.6-1.8 1.2-3.0 0.4-1 World Bank 1.3 0.5 3.5 1.4 - Notes: I Since Kenya has considerable variation in the quality of its land, the benefit-cost ratios are computed for high (H), medium (M), and low (L) potential land. With respect to the geographical division provided by FAO, the Western, Nyanza, and Rift provinces are medium potential and Central and Eastern provinces are high potential land. 2 When benefit-cost ratio for a crop is presented as a range, it represents profitability within different regions of a country. As Kenya does not subsidize fertilizer prices, benefit-cost ratios for Kenya are calculated for prevailing fertilizer prices and producer price of crops. I Figures in parentheses are ratios computed using the producer price of maize and fertilizer prices for 1988/89. 4 The subsidy on fertilizer for 1988 in Tanzania is an explicit and implicit subsidy resulting from devaluation and increasing intemal costs that have not been reflected in the prices that farmers pay for fertilizer. Sources and Notes: See Appendix 12. 44 Table 24 Benefit-cost ratios for fertilizer use in the West African MADIA countries Cameroon 1987 Estimated by: Maize Arabica Coffee Robusta Coffee Rice With Without Without With Without Without With Without Without With Without Without subsidy explicit explicit subsidy explicit explicit subsidy explicit explicit subsidy explicit explicit subsidy or implicit subsidy or implicit subsidy or implicit subsidy or Implicit subsidy subsidy subsidy subsidy FAO 5.1 1.5 1.3 - - - - - - - - - IFDC 7-11 2-3 1.9-2.9 12-15.5 3.8-4.6 3.3-3.9 4.5-6.6 1.3-2.0 1.1-1.7 5-15 1.4-4.5 1.2-3.9 (28-34)1 (4-5) (3.5-4) (8-13) (2-3) (2-3) Nigeria 1985 Estimated by: Maize Rice Sorghum Groundnuts Millet With Without Without With Without Without With Without Without With Without Without With Without Without subsidy explicit explicit subsidy explicit explicit subsidy explicit explicit subsidy explicit explicit subsidy explicit explicit subsidy or implicit subsidy or implicit subsidy or implicit subsidy or implicit subsidy or implicit subsidy subsidy subsidy subsidy subsidy FAO 6-26 3-12 1-6 8-35 4-16 2-8.5 6-18 3-8 1.5-4.0 13-38 6-17 3-9 6-39 2.5-18 1.5-9.5 A.0. Falusi 9-20 4-9 2-5 13-16 6-7 3-4 4-11 2-5 1.0-2.5 20-23 9-11 5-5.5 5-17 2-8 1-4 World Bank 7-9 3-4 1.8-2.0 16 7 4 4-5 1.7-2.0 0.9-1.1 3-5 1-2 0.6-1.3 5 2 1.0 Nigeria 19872 FAO 3-12 0.5-2.0 - 11-48 2-9 - 3-8 0.5-1.4 - 14-41 2.5-7 - 2-17 0.4-3.0 - A.O. Falusi 4-9 0.7-1.7 - 19-22 3-4 - 1.5-4.5 0.2-0.8 - 21-23 4-5 - 2-7 0.3-1.3 - World Bank 3-4 0.5-0.7 - 22 4.0 - 1.5-2 0.3-0.4 - 3-6 0.5-1 - 1.9 0.4 - Senegal 1987 Estimated by: Rice Sorghum Groundnuts Millet With Without Without With Without Without With Without Without With Without Without subsidy explicit explicit subsidy explicit explicit subsidy explicit explicit subsidy explicit explicit subsidy or implicit subsidy or implicit subsidy or implicit subsidy or implicit subsidy subsidy subsidy subsidy FAO 2.6 1.9 1.8 1.5 1.0 1.0 - - - 2.7 2.0 1.8 IFDC North Basin - - - - - - 1.1-1.6 0.8-1.2 0.7-1.1 3.0-4.3 2.2-3.1 2.0-2.9 Central Basin - - - - - - 1.4-2.2 1.0-1.6 0.9-1.5 3.2-3.7 2.3-2.6 2.2-2.5 North Sine-Saloum - - - 1.7 1.2 1.1 1.9-2.6 1.4-1.9 1.3-1.8 3.7-4.3 2.6-3.1 2.5-2.9 South Sine-Saloum - - - 2.2 1.6 1.5 2.2-3.1 1.6-2.2 1.5-2.1 3.7 2.6 2.5 Figures in parentheses are ratios computed for producer prices of coffee with taxes. 2 Based on the second-tier exchange market, assuming 4 Naira = 1 US$ Source and Notes: See Appendix 12. Kolda (Casamance) and Tambacounda (Eastern Senegal) Table 24 shows that the lower harvest prices do have a regions, the profitability of fertilizer use cannot be deter- significant impact on the profitability of fertilizer, for mined, but the fertile lands in the river valleys are bound example, in the case of maize a difference of 10 to 20 to have higher benefit-cost ratios than in Sine-Saloum. percent, suggesting a justification for the government Thus, while subsidy removal may not have an adverse effect becoming a buyer and seller of last resort for output, if on profitability in certain regions, in the vital Groundnut rapid technological change and intensification are the Basin it still raises some questions. objectives." II The benefit-cost ratios for all the East African MADIA On balance, the link between the benefit-cost ratios for countries have been computed in terms of the official individual crops and the change in fertilizer use is difficult producer price of crops. For others, average yearly prices to establish, because in most of the MADIA countries the have been used. Harvest prices would affect profitability nonprice constraints also affect fertilizer use. For instance, because they are often lower than average. Nigeria is the the availability of fertilizer at the right time in the right only MADIA country for which harvest prices of crops for the place is often a most important constraint, whether due to most recent year were available. Estimates of benefit-cost foreign exchange shortages or lack of credit to importers ratios using these prices and the range in benefit-cost and wholesalers. The location of retail outlets is also crucial ratios for intrayear crop price variations are presented in to the attractiveness of fertilizer use, given the lack of Table 26. Comparing these benefit-cost ratios with those in reliability and high cost of transport in most African 45 Table 25 countries. Most of the fertilizer retail outlets, especially Benefit-cost ratios with high analysis fertilizer in Malawi private outlets, are clustered around large towns because of Country Maize marketing and profitability constraints. As farmers in rainfed areas usually delay their fertilizer purchase until the first Malawi 1987/88 rain sets in, when climatic factors are favorable, the Local maize proximity of distribution outlets and, therefore, transport With subsidy costs are an important consideration in assessing the ASA Government 2.4 (3.1) attractiveness of fertilizers. The data suggest that on World Bank 2.1 (2.7) Without explicit subsidy average farmers travel 20 to 25 kilometers in some coun- ASA 1.9 (2.4) tries to buy fertilizers, suggesting the need to extend the World Bank 1.6 (2.1) retail networks in these countries (see Appendix 13).1 12 With respect to benefit-cost ratios, it should be clear that With subsidy there are comparatively few crops and areas where fertilizer ASA 4.3 (5.6) use is not profitable. There are several examples of crops FAO 5.5(7.1) that have benefit-cost ratios greater than 1, but less than World Bank 4.5 (5.8) the levels needed to overcome the risk and costs of Without explicit subsidy fertilizer use, in other words, between 2 and 3, especially in ASA 3.4(4.4) areas where sorghum and millets account for a sizable FAO 4.3(3.5) share of the area under cultivation. Consequently, even World Bank 3.5(4.5) though excess demand for fertilizer may exist in several Notes: Benefit-cost ratio for Malawi is computed in terms of DAP and areas due to supply constraints, the long-term sustained urea. Figures in parentheses are ratios computed using producer growth of fertilizer may require subsidies until research can price of maize and fertilizer prices for the year 1988/89 for Malawi. increase response coefficients and the risk of costs asso- Sources: Fertilizer price and crop price for Malawi from World Bank ciated with fertilizer use can be reduced. Internal Memo dated August 23,1988. Crop responses: see Tables 17 and 18. Table 26 Benefit-cost ratios for harvest prices in Nigeria and estimated range of benefit-cost ratios due to intrayear crop price variations Maize Rice Sorghum Groundnuts high low high low high low high low Benefit-cost ratios using harvest prices With Subsidy 8.2 2.4 33.0 7.6 6.4 1.3 47.8 3.4 Without Subsidy 1.5 0.4 6.0 1.4 1.2 0.2 8.7 1.0 Benefit-cost ratios with seasonal low crop prices 1 S.D. 3.8-10.3 1.5-4.1 3.6-26.5 1.5-26.5 1.9-6.3 0.5-2.1 13.5-55.5 6.7-27.8 2 S.D. 0.5-13.5 0.2-5.4 0.0-38.0 0.0-15.2 0.0-8.5 0.0-2.8 0.0-76.6 0.0-38.3 Benefit-cost ratios with seasonal high crop prices * 1 S.D. 8.7-23.6 3.5-9.4 7.3-53.6 2.9-21.4 4.9-15.8 1.6-5.3 27.5-113.4 13.7-56.7 * 2 S.D. 1.3-31.0 0.0-12.4 0.0-76.7 0.0-30.7 0.0-21.3 0.0-7.1 0.0-156.4 0.0-78.2 Notes: Harvest prices are for Kaduna and Bida ADPs in Nigeria for 1987 and fertilizer prices refer to the year 1987. The highest crop price usually refers to June-July, and the lowest price to January-February. The highest and lowest crop prices are averages for the year 1984-86. Source: Fertilizer and crop prices: Lele and Bindlish 1988. Crop responses: See Table 18. 46 Summary, Conclusions, and Implications The removal of fertilizer subsidies, and privatization of grant aid fertilizer has increased sharply. Since donor importation and distribution networks have been promi- support for fertilizer imports tends to be ad hoc and year- nent features of recent policy reforms in MADIA countries to-year, this makes it difficult for recipient governments to to reduce budget deficits and the role of the public sector. plan for growth in fertilizer use on a long-term basis. This paper reviews the reform policies implemented during Recipients also have to make do with fertilizers that donors the 1980s in the MADIA countries and their impact on the are willing to provide, which are not necessarily those best development of fertilizer use. In particular, it explores the suited to the particular soil, climatic, and cropping condi- supply and demand contraints that hinder the process of tions that prevail in the parts of each country. rapid growth and diffusion of fertilizer use. Supply constraints Further, the location-specific knowledge about fertilizer are analyzed in terms of: 1 ) macroeconomic factors, i.e., responsiveness that is needed to formulate sound policy is foreign exchange and budgetary constraints and (2) institu- lacking, which makes efficient intensification difficult. Given tional factors, especially those political factors that affect that fertilizer responsiveness is lower and more variable the stability and predictability of institutional arrangements under rainfed conditions with few new high-yielding tech- for the distribution of fertilizers and the regional priority in nologies, the issue of whether fertilizer is being allocated fertilizer supply. With regard to institutional factors, those of where the marginal responses to its use are the greatest especial importance include: changes in import licensing becomes a particularly important issue from the viewpoint systems, lack of working capital for importers, wholesalers, of maximizing growth in production. In reality this issue is transporters, and retailers, officially fixed distributive complicated by the fact that population densities are not margins, poor transport facilities, and remote areas or areas necessarily the greatest in areas where responses to of low physical response and high population densities. In application are high. Indeed, for a variety of reasons addition, the weaknesses of the cooperative sector, includ- densities have been greater in drier areas and in some ing the reasons for the failure of governments to promote countries the proportion of population in the marginal decentralized commercial institutions that represent the areas with low and variable responses is increasing. Given interests of small farmers and the private sector are con- poorly developed transport networks and the growing sidered. market dependence of these populations, fertilizer policy Demand for fertilizer is analyzed in terms of the level and requires considerations of growth and equity. Finally, it variability of fertilizer prices and output, physical responses must be recognized that fertilizers alone cannot solve the to fertilizer application in different locations as they relate range of complex agronomic problems affecting agriculture. to land potential, population densities, and transportation The use of fertilizers must be combined with improved networks, the availability of working capital to small farmers, crop rotation and the application of organic matter to and the ability of small farmers to undertake risks under maintain soil quality. rainfed agriculture. Supply constraints The paper shows that the development projects funded Supply constraints during the 1970s did facilitate the process of diffusion of Supply constraints are by far the most significant in fertilizer use among a large number of small farmers, expanding fertilizer use on a sustained basis, and are also fertilizeruse amojecs had largeiumiios small oftearmer. the area where policy can make a difference. Foremost Although the projects had limitations, many of these are aogte r hrae ffrinecag n ek now being addressed through policy reform measures among them are shortages of foreign exchange and weak- These measures include liberalization of import restric- nesses in the domestic procurement and distribution tions, increased access of private importers and distributors network. in both Nigeria and oCameroon, oil revenues to working capital, the more extensive use of high analysis allowed for adequate supplies of foreign exchange that fertilizer, packaging of fertilizer in small bags, and generally contributed to their record of the most rapid growth of familiarizing farmers with different types of fertilizers. The fertilizer use among the MADIA countries. In Tanzania and paper argues, however, that neither these reforms nor the Senegal, which have received more foreign assistance than parlier argues, assiner, t have adequately taken into the other MADIA countries, there has been little growth in earlier project assistance haver-tely tatinto fertilizer use. Tanzania's performance is especailly striking account the much broader and longer-term implications of since it has been the largest recipient of grant aid fertilizer. fertilizer's role in agricultural intensification. The disappointing growth in Tanzania and Senegal was also The need for intensification of agriculture due in part to the collapse of internal distribution networks, Increasing population pressure on the land, low productivity in turn the result of unpredictable government policies and of agriculture, increasing reliance on food imports, rapid unstable institutions that donors condoned and even movement of population to the areas of marginal physical facilitated. Malawi ranked third in growth of fertilizer use potential, and rapid degradation of soils due to the decline despite being the poorest of the MADIA countries, and in bush fallow all contribute to the need for agricultural having the largest current account deficits as a share of intensification. Despite the rapid increase in foreign aid to GNP An important feature of Malawi's performance in Africa, average per hectare use of fertilizer on arable land recent years has been the Smallholder Fertilizer Revolving is the lowest in the world. Even more striking, while the Fund, supported by IFAD and IDA. The SFRF was designed developing world's share in fertilizer consumption has to ensure the availability of foreign exchange for fertilizer doubled (from 19 to 38 percent between 1970/71 and 1986/ import and in turn the reliability of supplies. 87), Africa's share in the developing world total has In Kenya, despite its superiority in achieving broad- declined from 11.9 percent to 7.0 percent, and the share of based development in smaliholder agriculture, the growth 47 in fertilizer use ranked only fourth during this period. ture since the oil boom and another $1.7 billion committed Although Kenya did not subsidize fertilizer, it did regulate by the World Bank between 1971 and 1988, the establish- fertiiizer prices. This contributed to problems of availability ment of firm information on technology packages for in rural areas. Further, restrictions on import licensing complex mixed cropping practices of small farmers has not caused problems with the availability and timeliness of yet received the urgent priority it deserves. The inevitably fertilizer deliveries. subjective assessment of technology in Nigeria has gone De-mand constraints through cycles of optimism and pessimism leading to Country differences in demand-related factors and their substantial uncertainty about the profitability of fertilizer impact on fertilizer use are striking. The relative price of use, and hence the demand for fertilizer use. The Nigerian fertilizer in terms ofmaize, the food crop mostcommonl case illustrates the much broader phenomenon docu- fertilizer in terms of maize, the food crop most commonly mented in the case of Senegal, and the semiarid parts of grown among MADIA countries, has been highest in Malawi Kenya and Tanzania. and lowest in Nigeria. Malawi's landlocked nature and the In several of the MADIA countries there exist compelling hostilities in neighboring Mozambique make Malawi the country with the highest external transportation costs-$137 aents ior oetizer sub. Thes wincude .. A^_. r . 1 A _ . ~~~benefit-cost ratios between l and 2, combined with the per ton (1987) of nutrient, compared to $15-17 per ton for poverty of a significant number of households, growing Nigeria 11986) or $9 per ton for Senegal (19841. In addition, household dependence on the market for food that limits the frequent devaluations of the Malawian kwacha have resulted in the highest nutrient price/crop pi ratios... output price increases as a means to ensure profitability of resuAtediAth countries. InutNigentria, /crop triaeuation fertilizer, lack of access to seasonal credit, the high cost of aurnong the highest Nigeria, among thenMADIAe fertilizer, and the serious land shortage. Malawi is such a currency, the highest rate of subsidy among the MADIA case. Problems with leakages of subsidized fertilizer to the countries, and food prices that tend to be well above world estate sector have made the issue of continuing with a market prices have combined to yield nutrient price/crop generalized subsidy to the smallholder sector difficult. Of price ratios that are the lowest in the sample and one-tenth the reported fertilizer use in 1988, 40 percent was in the those inMaawi. . .estate sector, which cultivates less than 20 percent of the The sharp differences in physical response of crops per total land but where returns to fertilizer use are far higher; unit of nutrients among countries and regions within each 60 percent of use was reported in the smallholder sector, country affect profitability. For example, responses of but leakages of between 5-25 percent suggest the share of hybrid maize per kilogram of nutrient are four times (20-30 t estates of be even highe s a shar k..grms highe in Maaw tha in nothr Niei (6 the estate sector use to be even higher. Malawi is a clear kilograms). Within Malawi there are dramatic differences in case where demand constraints have been as binding as kilograms). Within Malawi there are dramatic differences in supply constraints in the smallholder sector. Targeting responsiveness. Fertilizer response on local maize is only fertilizer subsidies to the poorest households through half that on hybrid maize, yet only 5 percent of farmers grow specific programs (e.g., packaging in small bags) is under- hybrids; the remainder grow local maize due to its pre- ferred storage, milling, and cooking qualities. In Nigeria, way and fertilizer for work programs are under considera- two-thirds of the area is under sorghum and millets on ton.eBu -problem of fertilizer leakage to the more which response coefficients, while positive, are so low that fertilizer use does not seem to be profitable without a Because so little knowledge exists on how targeted subsidy fertilizer use dloes not seem to be profitable without a prgaswkinrcte,hslckokoldewlled subsidy on the order of 50 percent. programs work in practice, this la.k of knowledge will need W..ithn tea cout of et lerc ue s to be recognized explicitly, monitored carefully, and mod- within each country fertilizer use should be given priority ife inlgto.xeine Ifiled in light of experience. in areas and on crops where the marginal value of its use is highest. Only in some countries has fertilizer application Roles for the public and private sectors reflected such high marginal productivity in use. Political The paper points out the different roles of the private and considerations, poor transport networks, taxation of export public sectors in regard to fertilizer. The private sector can crops, excess demand for food, tying of fertilizer access to and must play an important and growing role in distribu- certain projects and areas, and restrictions on distributive tion. It operates most effectively in the areas of established mrargins are all factors that have distorted application to demand, easy access to transportation, and assured prof- less productive uses. While reform measures are alleviating its.' 3 This leaves the public sector to establish new some of these constraints others, such as inadequately demand, especially among low income producers with little articulated location-specific technical packages, poor trans- or no access to credit. This paper illustrates how and why port networks, and growing market dependence of house- the public sector has fulfilled this important developmental holds in marginal areas for food may well result in con- task in countries with stable institutional arrangements, and tinued suboptimal application of fertilizer on sociopolitical the extent to which climatic, political, and financial prob- and other grounds. lems have hindered the operations of the public sector. It illustrates the dilemma faced in planning for the acceler- Fertilizer subsidies ated use of fertilizers in countries with unstable institutions. Subsidies in Nigeria now amount to nearly one billion naira It also demonstrates the important contributions made by cr 71 percent of the budget devoted to agriculture in 1987. many public sector institutions in promoting the growth of Clearly the more permanent investments in agricultural use stressing that the effectiveness of public and private research, small-scale irrigation, transportation, and credit institutions must be assessed in the context of the partic- are needed to replace fertilizer subsidies. Despite nearly 11 ular circumstances in which they operate. billion naira spent by the Nigerian government on agricul- a, Policy recommendations fertilizer subsidies, nor can a single policy be adopted * The paper documents an urgent need for donors to across countries, areas, or income groups. Merits of fertilizer undertake long-term, untied import support for fertilizers as subsidies must be considered in the broader context of a way to ensure supply, and to improve the research, development objectives, the risk of returns, the adoption of planning, implementation, monitoring, and policymaking new technology, and the macroeconomic feasibility of a capacity of African governments in order to promote sound subsidy. In particular, the relative importance of fertilizer intensification of fertilizer use on a sustained basis. subsidies must be compared to other investments that * It recommends the financing of food and fertilizer stocks increase fertilizer use. The paper argues that in some of the at the national and regional levels as a way of encouraging MADIA countries there are compelling arguments in favor of governments to remove intra- and intercountry restrictions a fertilizer subsidy based on the need to increase the on trade, to broaden markets, and to increase the profita- quantity of fertilizer demanded. The rationale for subsidies bility of fertilizer use over time. reflects the need for household food security, as well as * It stresses the urgent need to improve the knowledge market imperfections, e.g., failure of credit and insurance base on a location-specific basis, especially the relative role markets. Specific examples of these circumstances include of fertilizers vis-a-vis other more complex resource manage- benefit-cost ratios greater than I (but less than the critical ment needs. value of 2 needed to make fertilizer use attractive); growing * It is clear that privatization offers great potential for household dependence on the market for food, which limits improving procurement and distribution of fertilizer. The output price increases as a means to ensure the profitabil- need for complementary public sector involvement-in ity of fertilizer; lack of access to credit; and the increasing terms of supporting the private sector and in meeting the scarcity of arable land. Given the imperfect knowledge needs of farmers beyond the reach of the private sector- about fertilizer responsiveness, the impact of price changes must also be recognized. and other intentions (on inputs use and outputs) must be * Finally, the paper concludes that simple judgments such monitored regularly, based on field-level investigations in as "good" or "bad" cannot be made with respect to order to fine-tune policy. 49 Appendices APPENDIX 1: Gross food imports (cereals) in the MADIA countries, 1971-86 Year Kenya Malawi Tanzania Cameroon Senegal Nigeria -------------------------'000 Metric Tons--------------------- 1971 62.8 28.5 365.1 112.9 57.4 467.3 1972 71.8 26.1 287.8 96.7 183.6 375.6 1973 81.1 26.9 457.6 91.2 42.8 451.2 1974 15.3 17.0 340.9 81.1 430.7 389.3 1975 86.1 41.3 220.1 68.8 461.0 447.5 1976 11.6 43.2 429.9 74.5 97.4 831.6 1977 34.4 27.0 415.3 116.2 134.8 1320.3 1978 99.7 13.7 453.9 123.6 123.4 2007.7 1979 22.4 14.8 502.8 163.6 58.9 1652.8 1980 366.7 36.3 451.9 140.3 398.6 1827.6 1981 149.1 74.2 470.9 107.2 266.1 2215.7 1982 274.9 26.5 494.2 117.1 324.5 2156.7 1983 160.9 19.3 544.4 202.0 235.6 1494.3 1984 556.4 52.7 661.5 122.1 266.3 1353.9 1985 364.7 22.3 496.0 139.3 259.4 1956.8 1986 189.1 5.9 544.1 149.4 243.5 1596.4 Source: World Bank Database. APPENDIX 2: Table 1: Ratios of producer prices to international prices for major smallholder crops in East African countries, 1970-85 (using nominal exchange rates) Kenya Malawi Tanzania Year Coffee Tea Tobacco Croundnuts Tobacco Cotton Coffee 1970 0.91 - 0.30 0.73 0.78 0.73 - 1971 0.90 0.79 0.33 0.81 0.84 0.61 - 1972 0.98 0.77 0.29 0.81 0.84 0.57 0.57 1973 0.96 0.77 0.27 0.58 0.84 0.35 0.43 1974 0.97 0.67 0.20 0.45 0.68 0.33 0.46 1975 1.01 0.75 0.20 0.60 0.70 0.52 0.36 1976 0.96 0.74 0.23 0.63 0.65 0.42 0.30 1977 0.93 'b.89 0.22 0.59 0.63 0.46 0.35 1978 1.02 0.85 0.30 0.56 0.70 0.56 0.39 1979 0.99 0.75 0.24 1.03 0.51 0.51 0.29 1980 1.04 0.83 0.23 1.03 0.47 0.53 0.41 1981 0.89 0.89 0.20 0.80 0.50 0.62 0.53 1982 0.82 0.86 0.18 1.28 0.50 0.73 0.52 1983 0.90 0.68 0.31 1.45 0.70 0.67 0.47 1984 0.83 0.98 0.28 0.92 0.55 0.65 0.47 1985 - - 0.29 1.11 0.72 1.03 0.50 Sources: International Prices: World Commodity Trade & Price Trends 1985b. Producer Prices: World Bank Database. 50 APPENDIX 2: Table 2: Ratios of producer prices to international prices for major smallholder crops in West African countries, 1970-85 (using nominal exchange rate) Cameroon Nigeria Senegal Coffee Year Arabica Robusta Cocoa Cotton Cocoa Palm Kernel Groundnut Cotton 1970 0.63 0.37 0.45 0.18 0.44 0.52 0.42 0.16 1971 0.63 0.46 0.58 0.15 0.56 0.57 0.41 0.13 1972 0.59 0.45 0.55 0.16 0.80 0.82 0.57 0.13 1973 0.57 0.41 0.37 0.11 0.93 0.95 0.43 0.09 1974 0.57 0.37 0.29 0.12 0.80 0.64 0.26 0.15 1975 0.62 0.44 0.46 0.17 0.64 1.40 0.63 0.16 1976 0.31 0.19 0.28 0.11 0.66 1.50 0.59 0.11 1977 0.22 0.15 0.19 0.15 0.55 0.85 0.44 0.13 1978 0.40 0.31 0.30 0.21 0.52 0.67 0.39 0.13 1979 0.44 0.34 0.38 0.20 0.60 0.78 0.35 0.14 1980 0.49 0.42 0.53 0.19 0.65 0.86 0.54 0.14 1981 0.44 0.42 0.53 0.17 0.98 1.03 0.37 0.16 1982 0.34 0.32 0.55 0.20 1.08 1.31 0.66 0.14 1983 0.33 0.32 0.44 0.17 0.85 1.10 0.57 0.12 1984 0.30 0.28 0.36 - 0.74 1.00 0.40 0.12 1985 - - 0.41 - - - 0.38 - Sources: International Prices: World Commodity Trade and Price Trends 1985b., Producer Prices: World Bank Database. APPENDIX 2: Table 3: Ratios of producer prices to international prices for major smallholder crops in East African countries, 1970-85 (using purchasing power parity exchange rates) Kenya Malawi Tanzania Year Coffee Tea Tobacco Groundnut Tobacco Cotton Coffee 1970 0.85 - 0.30 -.72 0.74 0.69 - 1971 0.88 0.78 0.33 0.80 0.82 0.59 - 1972 0.98 0.77 0.29 0.81 0.84 0.57 0.57 1973 1.02 0.82 0.29 0.62 0.85 0.35 0.44 1974 0.99 0.68 0.21 0.68 0.64 0.31 0.44 1975 1.03 0.76 0.21 0.64 0.60 0.45 0.31 1976 1.00 0.77 0.24 0.66 0.59 0.38 0.27 1977 0.93 0.89 0.24 0.65 0.56 0.41 0.31 1978 0.96 0.80 0.33 0.62 0.60 0.48 0.33 1979 0.98 0.74 0.28 1.19 0.47 0.47 0.27 1980 1.05 0.84 0.27 1.21 0.38 0.43 0.33 1981 0.94 0.94 0.23 0.93 0.31 0.39 0.33 1982 0.83 0.88 0.21 1.50 0.26 0.38 0.27 1983 0.96 0.72 0.35 1.63 0.33 0.32 0.22 1984 0.83 0.98 0.30 0.97 0.25 0.30 0.21 1985 - - 0.32 1.23 0.27 0.39 0.19 Sources: International Price of Crops: World Commodity Trade and Price Trends 1985b. Producer Price of Crops: World Bank Database. Real Effective Exchange Rate Index: Seka and Fishstein 1988. 51 APPENDIX 2: Table 4: Ratios of producer prices to international prices for major smallholder crops in West African countries, 1970-85 (using purchasing power parity exchange rates) Cameuoon Nigeria SenegaL Coffee Year Arabica Robusta Cocoa Cotton Cocoa Palm Kernel Groundnut Cotton 1970 0.64 0.38 0.46 0.18 0.47 0.56 0.42 0.16 1971 0.66 0.48 0.60 0.16 0.55 0.56 0.42 0.13 1972 0.59 0.45 0.55 0.16 0.80 0.82 0.57 0.13 1973 0.54 0.39 0.35 0.11 0.99 1.01 0.41 0.09 1974 0.54 0.35 0.28 0.11 0.81 0.65 0.25 0.16 1975 0.56 0.40 0.42 0.15 0.54 1.17 0.50 0.13 1976 0.28 0.17 0.26 0.10 0.46 1.04 0.51 0.09 1977 0.19 0.13 0.17 0.13 0.39 0.60 0.38 0.11 1978 0.34 0.27 0.26 0.18 0.34 0.66 0.35 0.12 1979 0.38 0.29 0.32 0.17 0.39 0.51 0.32 0.13 1980 0.44 0.36 0.46 0.16 0.41 0.55 0.52 0.13 1981 0.42 0.39 0.50 0.16 0.55 0.58 0.38 0.17 1982 0.33 0.31 0.53 0.19 0.61 0.74 0.66 0.14 1983 0.30 0.30 0.41 0.16 0.40 0.52 0.57 0.12 1984 0.30 0.25 0.33 - 0.25 0.34 0.38 0.12 1985B - - 0.38 - - - 0.34 - Sources: International Prices: World Commodity Trade and Price Trends 1985b. Producer Price of Crops: World Bank Database. Real Effective Exchange Rate Index: Seka and Fishstein 1988. APPENDIX 3: Volume of and growth rates for fertilizer consumption in the MADIA countries Year Malawi Kenya Tanzania Senegal Cameroon Nigeria --------Metric Tons of Nutrient--------------------- 1970 21,900 1971 14,800 1972 16,374 29,400 19,558 1973 17,634 31,200 15,200 1974 13,861 52,700 31,000 44,000 16,180 29,000 1975 14,196 44,494 30,000 43,900 16,654 54,300 1976 16,309 53,896 31,000 105,000 12,200 79,000 1977 24,100 51,472 28,370 116,000 13,049 74,000 1978 27,000 51,002 26,783 74,600 34,000 71,400 1979 27,500 38,000 30,203 110,000 35,600 108,300 1980 31,900 61,600 32,327 57,900 32,608 173,900 1981 28,600 82,000 28,768 102,000 35,682 213,200 1982 32,400 64,855 24,201 50,000 40,200 201,800 1983 31,800 91,308 27,183 21,000 39,900 166,000 1984 38,200 75,130 35,674 39,000 38,300 221,300 1985 37,600 97,100 41,253 41,000 49,000 205,160 1986 39,900 101,689 45,473 23,000 57,000 229,740 1987 46,300 102,715 46,330 25,000 NA 262,960 Average Annual Growth Rate: 7.7% 6.7% 2.9% 0.8% 11.7% 18.0% Coefficient of Variation 0.4 0.3 0.2 0.6 0.4 0.6 Note: The fertilizer consumption data in this table are based on local government data, as opposed to FAG data, because for most countries the former is thought to be more accurate. (FAO acknowledges that its data are based on trade and production estimates.) Sources: Malawi: Nathan 1987; Ministry of Agriculture, Malawi 1987. Kenya: IFDC, 1986b; Ministry of Agriculture, Kenya 1987. Consumption data for 1987 from USAID office, Nairobi, Kenya. Tanzania: Mhella 1985; Rioseco 1989. Cameroon: IFDC 1986a. Senegal: Jammeh 1987a. Senegal consumption figures for 1987/88 are estimates from the Senegalese government. Nigeria: IFDC 1988. 52 APPENDIX 4: Growth rate of export and food crop production in the MADIA countries, 1970-1985 East Africa West Africa Kenya Malawi Tanzania Cameroon Nigeria Senegal Food Crops Maize 3.9% 1.5% 2.1% 4.1% 6.1% 5.6% Rice 16.5% 10.5% 1.2% Millet/Sorghum 1.3% 0.2% 0.2% Export Crops Smallholder Coffee 6.0% 2.3X 1.0% Tea 13.5% 13.7% Tobacco 3.0% -4.8% Cotton 4.9% 1.1% 1.6% 8.3% Groundnuts -7.4% Sugar 16.9% Cocoa 2.0% Estates Coffee 1.0% -4.1% Tea 5.5% 1.0% Tobacco 12.9% -7.5% Cotton Sugar 14.71 0.8% Rubber 2.7% Palm Oil 4.2% General Cotton -6.1% 6.8% Groundnut s -3.2% Cocoa -4.9% -0.4% Source: Compiled from MADIA Country Highlights, CECSS, World Bank 1987. APPENDIX 5: Table 1: Trend in the ratio of producer prices for export crops to food crops in the East African MADIA countries Kenya Malawi Tanzania Coffee/ Tea/ Tobacco/ Coffee/ Groundnut/ Cotton/ Cotton/ Tobacco/ Cashewnuts/ Coffee/ Year Maize Maize Maize Maize Maize Maize Maize Maize Maize Maize 1975 15.3 11.6 6.05 11.19 3.70 3.77 1.73 14.29 1.87 7.00 1976 32.9 13.8 5.40 8.75 3.11 2.25 2.50 9.66 1.29 10.00 1977 44.7 24.2 6.24 8.70 3.39 3.52 2.50 10.90 1.33 18.75 1978 31.7 17.8 7.80 11.28 3.70 3.94 2.71 10.67 1.31 12.81 1979 36.8 17.6 7.88 12.54 5.81 4.19 2.82 10.51 1.92 10.67 1980 27.6 16.7 6.31 0.40 4.60 3.25 3.00 8.95 1.73 11.42 1981 22.6 17.7 6.53 7.58 4.65 3.24 3.20 9.64 2.75 12.36 1982 25.8 18.0 4.03 4.50 2.87 2.45 2.47 7.41 3.09 9.93 1983 22.7 14.2 7.56 9.35 4.64 3.39 2.69 9.96 2.65 8.67 1984 22.0 29.6 6.61 8.33 4.89 3.31 2.73 7.61 2.95 10.40 1985 21.2 18.0 8.11 NA 5.57 3.56 2.10 6.30 2.42 6.75 Source: Crop prices from World Bank Database. 53 APPENDIX 5: Table 2: Trend in the ratio of producer prices for export crops to food crops in the West African MADIA countries Senegal Cameroon Nigeria Robusta Arabica Groundnut/ Cotton! Coffee/ Coffee/ Cocoa/ Cotton/ Cocoa/ Groundnut/ Cotton Year Maize Maize Maize Maize Maize Maize Maize Maize Maize 1975 1.19 1.34 3.86 5.43 3.43 1.23 NA NA NA 1976 1.19 1.34 4.53 7.34 4.06 1.72 4.4 1.67 2.07 1977 1.12 1.32 4.33 6.11 3.33 1.44 4.48 1.09 1.43 1978 1.12 1.32 5.0 6.5 4.40 1.30 3.55 0.97 1.14 1979 1.23 1.49 4.67 6.0 4.33 1.17 5.45 1.32 1.5 1980 1.35 1.62 5.17 6.5 4.83 1.33 5.0 1.35 1.54 1981 1.49 1.45 5.08 5.38 4.77 1.38 3.9 1.27. 1.42 1982 1.49 1.66 5.00 5.29 4.71 1.50 4.8 1.67 1.89 1983 1.49 1.66 NA NA NA NA 4.67 1.5 1.87 1984 1.17 1.30 NA NA NA NA 2.05 0.89 0.96 1985 1.29 1.43 NA NA NA NA 2.62 1.23 1.39 Source: Crop prices from World Bank Database. Appendix 6: Fertilizer use by crop in the MADIA countries Kenva Malawi Tanzania Cameroon Ni eria Senegal (1982 83) (1984/85) (1985) (1985) (1979/80) (1985) Relative Share of Fertilizer Use Coffee 25.6 - - 47.0 - - Maize 19.5 83.5 47.0 - 11.7 - Tea 17.6 - - - - - Sugar 16.7 - - - - - Cotton - - - 20.0 - 21.0 Tobacco - - - - - - Millet - - - - 15.7 - Sorghum - - - - 23.7 - Rice - - - - 11.1 30.0 Groundnuts - - - - - 32.0 Food Crops - - - 15.0 - - Millet/Sorghum - - - - - 17.0 Note: Data for Mala%wi refer to smallholders only. Sources: Kenya: World Bank 1986b; Tanzania: Mhella 1985; Nigeria: IFDC 1981; Malawi: Nathan 1987; Senegal: Jammeh and Ranade 1987; Cameroon: IFDC 1986a. 54 Appendix 7: Malawi's Smallholder Fertilizer Revolving Fund The main objectives of the Smallholder Fertilizer Project were to "increase the productivity of the smallholder sub-sector through improvements in policy decisions, the restoration of sufficient and timely financing for fertilizer procurement and technical support to the sub-sector" (World Bank 1983, p.23). These goals were to be accomplished through the implementation of four components: (i) improving ADMARC's procedures for procuring and distributing fertilizer; (ii) establishing a fertilizer revolving fund in the Reserve Bank of Malawi; (iii) strengthening the institutional capacity of MOA and ADMARC; and (iv) setting in motion a process to bring about a reform of all relevant policies through agreements on subsidies and regular consultations on resource allocations, crop and input pricing, and measures to strengthen agricultural research (World Bank 1983b, p.23). The appraisal report notes that adequate supplies of foreign exchange are vital to improving the procurement and distribution system."... to provide foreign exchange: ... a revolving fund would be established by Government and ADMARC in the Reserve Bank of Malawi with funds from Government, ADMARC, IFAD and IDA. While ADMARC would be responsible for the physical implementation of the procurement program, a joint management committee (consisting of representatives from the Ministry of Finance (MOF), the Economic Planning Division office of the President, Department of Statutory Bodies (DSB), MOA and ADMARC) would be responsible for control of the revolving fund. The Fund would be used exclusively to finance the import of adequate fertilizer for smallholders over the four year Project period. Finance would be subscribed to cover the foreign exchange cost of importing the total fertilizer requirement (less the carry- forward stocks) in the first year and the incremental amounts of fertilizer thereafter over the four-year period of the Project (US$24.47 million). All funds generated internally by way of fertilizer sales to smallholders would be paid into the Fund (World Bank 1983b, p.24). The critical issue of foreign exchange availability and the government's commitment to guarantee adequate supplies of foreign exchange is explicitly addressed: Government would make available, through the Reserve Bank, adequate foreign exchange to enable the Fund to purchase foreign exchange each year with local currency for the importation of fertilizer. During negotiations, the Government confirmed that the necessary foreign exchange would be made available (World Bank 1983o, p. 25). In addition, the project sought to strengthen institutional capacity by providing for technical assistance and staff training. The Project would provide eleven man years of technical assistance for contract staff and four man-months of consultancy services. The technical assistance needs for ADMARC were identified through a management study of ADMARC financed under the first Technical Assistance Loan (Ln 2027-MAI). The study recommended, inter alia: (a) improving MOA and ADMARC's pricing policies through enhanced planning and analysis; (b) re-establishing ADHARC's management and financial control; (c) reducing marketing costs by cutting back on the number of markets and (d) improving monitoring, evaluation and training. The report was accepted by Government in December 1982 and provided the basis for the institutional improvements (World Bank 1983b, p.25). It would also include provision for two years of external specialist training for local staff, as needed. The emphasis on training will, however, be on in-service, on-the-job training and technical assistance staff assigned to ADMARC will be expected to assist with the teaching of courses at ADMARC's training facility in Kanjedza (World Bank 1983b, p.26). 55 Although the Smallholder Fertilizer Project has not resulted in dramatic increases in nutrient uptake within the smallholder sector, it is necessary to evaluate its performance against the likely counterfactual case. Malawi has, since 1984, experienced serious foreign exchange shortages and disruption of its least cost rail links to ocean ports. Despite these circumstances, the SFRF has managed to procure an increasing amount of fertilizer each year. Further, it has been responsible for marketing improvements (small packets) and providing a basis (data and reports) for analyzing fertilizer policy in Malawi. The following comments by Gert Stern, however, indicate many of the weaknesses in its design and implementation which offer important lessons for donors and governmerts interested in establishing such a fund. Overall SFRF was a success in meeting an emergency situation in Malawi, caused by ADMARC's liquidity crisis, but did it really provide much needed foreign exchange adequately or free Malawi from the need to rely on variable donor fertilizer grants? To some extent, yes and to some extent, no. The project set up a special fund to make it more difficult for ADHARC to use fertilizer funds for non-fertilizer purposes. Despite this, ADMARC very nearly succeeded in such fund diversion. The situation was saved by "energetic" supervision action, a Government rescue operation and unexpected US AID help (subsidy removal program). Nevertheless, but for the project, the chances are that there would not have been any fertilizer distribution that year. With regard to FE, the IFAD/IDA injection helped only to some extent. As at March 31, 1987, the joint contribution represented 18 percent of SFRF's financial resources consisting of 15 percent US AID contribution; 25 percent EEC contribution (Buffer stock); 11 percent ADMARC contribution; 17 percent COM contribution; and 14 percent other donor contributions. The last IDA/IFAD disbursement for fertilizer purchases - a mere US $300,000 equivalent was made in 1987 and the last sizeable disbursement was US $1.35 million equivalent in 1986, in support of total purchases of US $16 million and $20 million, respectively. The project could, however, claim some credit for FE injection through the covenant, honored absolutely by COM, to match the Kwacha revolving fund by FE as needed and by SFRF becoming a focus for donor assistance, the major EEC buffer stock input in particular. However, at the end of the day, SFRF has chosen to gamble on donor grant assistance for securing adequate fertilizer supplies rather than, for instance, using the available IDA ITPAC Credit to offset any FE shortage or resort to short-term borrowing to supplement the revolving fund. The gamble has paid off this year, but I am not sure how much long-term coamitment there is or whether COM will continue to play annual "roulette" to procure fertilizer supplies. What was SFRF good at? Procurement using FINCOM (ADMARC subsidiary) and international transport (using MITCO) both expatriate managed - and keeping the accounts up-to-date and straight. What was not improved by the project? Distribution, in particular, having the right fertilizer at the right place at the right time - as yet all done by ADMARC to date. The main problem is transport, an aging, inadequate fleet being offered inadequate rates, particularly for less attractive routes. With regard to institution building, the project benefitted from the long-term service of two competent consultants, but both the newly independent Trust Corporation and the MOA Fertilizer unit are "delicate plants" and the latter withered very quickly, after the consultant left. GOM had no urge whatever to retain him. The Trust Fund Corporation is a major institution handling the US $50 million or so fertilizer purchase annually, using rather sophisticated financial management recently introduced by EEC financed consultants. At this juncture, EEC has become the main SFFRF benefactor by providing 70,000 tons buffer stock fertilizer and about MK 12 million worth of storage. One must hope that EEC will be willing to sustain management capabilities until a competent local team is in place. Perhaps the lesson is not to "spawn" a new institution unless one is willing to stay with it for say 10-15 years. With regard to whether this is a model to be copied by other countries . . . it served its purpose at the time in Malawi, but it has entrenched a new parastatal as smallholder fertilizer importer. With management talent as yet scarce in many of our client countries, do we really favor parastatal proliferation? (Personal communication with the authors). 56 In view of these comments, the concept of long-term import support can be pursued without following SFRF in all its details. APPENDIX 8: Producer prices for primary nutrient types used in the MADIA countries, 1971/72-1987/88 (converted to US$ per ton using official exchange rates) Malawi Kenya Tanzania Cameroon Senegal Nigeria Year 20:20:0 CAN A/S DAP CAN A/S TSP 20:10:10 A/S 6:20:10 15:15:15 1971/72 NA NA NA 179 315 NA NA NA NA 120 NA 1972/73 215 317 327 245 364 NA NA NA NA 132 NA 1973/74 211 310 320 344 458 NA NA 269 NA 199 NA 1974/75 535 777 759 337 449 NA NA 239 395 185 NA 1975/76 492 690 606 639 942 NA NA 525 800 259 NA 1976/77 466 653 574 561 770 NA NA 471 598 291 128 1977/78 471 660 580 447 805 676 431 366 581 283 124 1978/79 504 707 621 520 995 725 464 355 548 308 125 1979/80 520 730 641 537 966 681 436 376 582 326 133 1980/81 523 995 762 841 1206 682 437 379 630 329 146 1981/82 475 902 957 797 1194 983 504 294 490 256 261 1982/83 597 947 947 669 1100 965 495 266 506 211 238 1983/84 617 917 973 660 743 804 412 262 500 364 245 1984/85 619 844 910 626 946 586 300 229 436 445 550 1985/86 596 850 970 NA NA 1278 803 223 424 371 471 1986/87 564 748 921 441 681 967 597 NA NA 521 331 1987/88 611 792 992 NA NA 640 395 NA NA NA 155 Sources: Fertilizer Prices for Malawi: Nathan Report 1987; Kenya: World Bank 1986b; Tisminieszky and Kimuyi 1986. Tanzania: Mhella 1985; World Bank and FAO 1987; Nigeria: Lele and Bindlish 1988; Senegal: Kelly 1987; Cameroon: IFDC 1986a. Note: Exchange rates used are the Par rate as given by IMF International Financial Statistics Yearbook 1988. APPENDIX 9: Table 1: Producer prices for maize in MADIA countries, 1971-87 Year Cameroon Senegal Nigeria Malawi Kenya Tanzania International US$ per metric ton using purchasing power parity exchange rates 1971 66 66 NA 36 44 35 58.4 1972 72 71 NA 37 53 34 56.0 1973 86 82 NA 39 56 37 98.0 1974 95 98 NA 51 58 43 132.0 1975 148 128 NA 62 88 46 119.6 1976 123 126 172 58 86 85 112.4 1977 163 130 262 61 107 83 95.3 1978 192 149 316 66 115 89 100.7 1979 241 159 240 99 97 90 115.5 1980 246 167 288 101 120 92 125.3 1981 223 179 297 142 111 68 130.8 1982 205 143 226 122 97 74 109.3 1983 NA 122 195 114 109 66 136.0 1984 NA 131 323 90 111 58 135.9 1985 NA 139 260 79 105 78 112.2 1986 NA 172 257 79 120 55 87.6 1987 NA 212 190 103 150 96 NA Sources: Producer Price from World Bank Database. Purchasing power parity exchange rates: Seka and Fishstein 1988. International Maize prices from "Commodity Trade and Price Trends" 1987. 57 APPENDIX 9: Table 2: Producer prices for seed cotton in MADIA countries, 1970-86 (converted to US$ per kilogram using purchasing power parity exchange rates) Year Cameroon Kenya Senegal Nigeria Tanzania Malawi 1970 0.11 0.13 0.10 0.16 0.15 0.15 1971 0.12 0.14 0.10 0.14 0.15 0.15 1972 0.12 0.16 0.12 0.19 0.15 0.17 1973 0.16 0.19 0.13 0.22 0.16 0.20 1974 0.17 0.22 0.12 0.25 0.15 0.23 1975 0.18 0.27 0.12 0.43 0.16 0.24 1976 0.21 0.28 0.16 0.35 0.22 0.24 1977 0.24 0.36 0.16 0.37 0.23 0.28 1978 0.25 0.39 0.19 0.36 0.28 0.32 1979 0.28 0.43 0.21 0.36 0.29 0.33 1980 0.33 0.44 0.26 0.45 0.33 0.33 1981 0.31 0.39 0.24 0.42 0.27 0.35 1982 0.31 0.32 0.21 0.42 0.21 0.42 1983 0.28 0.29 0.21 0.42 0.22 0.40 1984 0.30 0.30 0.21 0.31 0.19 0.34 1985 0.32 0.29 0.16 0.36 0.21 0.31 1986 0.38 0.38 0.25 - - 0.26 Sources: Nominal producer prices: World Bank Database. Purchasing power parity exchange rates: Seka and Fishstein 1988. APPENDIX 9: Table 3: Producer prices for coffee in Kenya and Cameroon, 1971-87 (converted to US$ per kilogram using purchasing power parity exchange rates) Year Cameroon Kenya 1971 0.63 0.90 1972 0.69 0.84 1973 0.86 0.91 1974 0.75 1.04 1975 1.00 1.12 1976 1.17 1.32 1977 1.18 2.66 1978 1.38 1.96 1979 1.41 1.80 1980 1.39 2.12 1981 1.27 2.01 1983 1.01 1.71 1984 0.94 3.50 1985 1.00 2.00 1986 1.18 4.12 1987 1.44 2.87 Sources: Producer prices from World Bank Database. Purchasing power parity exchange rates: Seka and Fishstein 1988. APPENDIX 10: Analysis Of Nutrient Price/Crop Price Ratios The net effect of the different nutrient and output prices can be seen through a comparison of the cost of nutrients relative to the producer price--the nutrient price/crop price ratio (see Tables 15 and 16). As the exchange between fertilizer and output takes place internally, we have calculated nutrient price/crop price ratios in local currencies. Although there is only one crop (maize) for which there are complete cross-country data, it is still clear that the relative cost of nutrients differs significantly both across and within countries. For example, in 1986 the nutrient price/crop price ratio for maize in Malawi was higher than other East 58 African countries and higher still than in the West African countries, for example, nine times that in Nigeria in 1986/87. In general, fertilizer was more expensive relative to maize prices in East Africa th!I) in West Afriia and reflects the generally lower maize prices in the former. - The ratios for arabica coffee in Cameroon and Kenya are similar, indicating that the effects of import taxes and low quality in Cameroon are offset by the subsidy on fertilizer applied to coffee. For cotton, the ratios in East African countries tend to be higher than2 n West Africa, although this trend is not as strong for cotton as for maize. - Differences among countries need to be treated with caution as they reflect different commodities and quality. Nonetheless, some surprising variations exist. In Kenya, a kilogram of fertilizer is considerably more expensive in terms of maize as compared with coffee. In Nigeria, the difference in the price of fertilizer (in international price terms) relative to maize compared to a higher value crop such as cotton is relatively small. The nutrient price/official crop price ratios are based on the subsidized price of fertilizer and the official or market producer price of crops. The fertilizer price is based on the cost of the nutrient content. The nutrient price/crop price ratios are computed as the ratio of the price of one kilogram of nutrient to the price of one kilogram of output. Wherever possible the price of the fertilizer that is typically applied to each crop is used to calculate the ratios. When those data are unavailable, we substitute the price of the type of fertilizer that is predominantly used in the country. In Malawi, prior to 1982/83, the ratios are computed in terms of A/S and, thereafter, in terms of CAN prices. Data for producer prices of maize, groundnuts (chalimba), and Tobacco (grade C) are from the World Bank Database. Data for fertilizer prices are from Nathan Associates 1987. In Kenya, the ratios for cotton are in terms of ASN prices; coffee, tea, and sorghum in terms of CAN prices; maize and wheat in terms of DAP. Data for producer prices of crops are from the World Bank Database. Data for fertilizer prices are from World Bank 1986b. In Cameroon, the computed ratio is based on the subsidized nutrient price of A/S for all crops. Data for producer prices are from the World Bank database and Elliot Berg Associates 1983. Data for fertilizer prices are from Berg 1983 and IFDC 1986a. For Senegal, the ratio is computed based on the subsidized nutrient price of 6:20:10. It should be noted that, since 1984, the majority of fertilizer sold in the country has been at unsubsidized prices. Data for producer prices are from the World Bank Database. Data on fertilizer prices for 1980-84 are from Kelly 1988 and prices for the years 1984-86 are from personal communication with the Government of Senegal. Subsidized fertilizer prices for 1987 are estimates based on USAID's limited subsidy of CFA 16,000 per tonne on cash sales. In Nigeria, the ratios are based on the subsidized nutrient price of 15:15:15. Cash crop prices for 1987 are postdevaluation prices. Data for producer prices are from the World Bank Database. Data for fertilizer prices are from Lele and Bindlish 1988. In Tanzania, the ratios are based on the subsidized nutrient price of A/S for foodcrops, cotton, and tobacco. The ratio for tea is based on the price of 25:5:5. Data for producer prices and fertilizer prices are from Mhella 1985 and Rioseco 1989. 1/ If parallel market prices were used instead of official producer prices in Tanzania, the relative cost of nutrients would be lower. 2/ It is important to note that if nutrient price/crop price ratios were computed using the purchasing power parity exchange rate, the resulting ratios would be different. It is assumed that overvaluation or devaluation in currency has an impact only on fertilizer prices. As it is difficult to anticipate the extent to which crop prices will be adjusted as exchange rates change, we have assumed that crop prices are invariant with respect to such fluctuations. Depending on whether the country's currency has generally been overvalued, undervalued, or relatively stable, the ratios at real effective exchange rates may be inflated or deflated to that extent from the ratios computed at local prices. 59 APPENDIX 10: Table 1: Nutrient price/crop price ratios for selected crops in MADIA countries (using purchasing power parity exchange rates) Country Maize Rice Tobacco Coffee Cotton Tea Groundnuts Millet Cocoa Arabica Robusta Malawi 1980/81 7.6 0.9 1981/82 7.1 1.3 1982/83 7.8 1.4 1983/84 7.9 0.9 1984/85 9.5 1.0 1985/86 11.6 1.0 1986/87 10.1 0.8 1987/88 7.6 0.8 Kenya 1980/81 6.3 0.4 2.6 3.6 1981/82 7.0 0.5 3.3 4.0 1982/83 4.6 0.4 3.3 3.3 1983/84 4.8 0.3 2.9 1.3 1984/85 5.4 0.4 3.0 2.6 1985/86 NA NA NA NA 1986/87 2.8 0.2 2.6 2.0 1987/88 3.7 NA NA NA Tanzania 1980/81 6.2 3.6 0.7 2.1 3.5 1981/82 7.9 5.1 1.2 3.7 5.5 1982/83 9.5 5.5 1.3 4.5 7.7 1983/84 7.8 4.3 1.0 3.7 6.0 1984/85 4.5 3.0 0.7 3.0 4.5 1985/86 9.9 6.5 1.4 6.2 9.0 1986/87 6.3 4.2 0.8 3.1 NA 1987/88 5.1 2.9 NA NA NA Cameroon 1980 2.6 0.4 0.5 1.9 0.5 1981 2.2 0.4 0.4 1.6 0.5 1982 2.5 0.5 0.5 1.6 0.5 1983 NA 0.5 0.5 1.8 0.6 1984 NA 0.5 0.5 1.6 0.5 1985 NA 0.4 0.5 1.4 0.5 Senegal 1980 1.9 1.7 1.3 1.5 1981 1.4 1.3 1.1 1.1 1.8 1982 1.4 1.3 1.0 1.1 1.3 1983 2.9 2.6 1.7 1.1 1.4 1984 2.8 2.8 2.2 2.8 23.7 1985 2.3 2.4 2.1 2.8 1986 2.7 2.3 1.9 2.1 2.7 1987 3.2 2.7 NA 251 2 25 3.2 Nigeria 1980 0.5 0.3 0.30. 05 01 1981 0.9 0.5 0.6 0.4 0.5 0.1 1982 1.1 0.6 0.6 0.7 0.7 0.2 1983 1.3 0.7 0.7 0.6 0.7 0.2 1984 1.7 1.2 1.8 10.98 1.1 0.3 1985 1.8 1.0 1.3 1.9 1.5 0.8 1986 1.7 0.6 0.8 1.7 1.4 0.7 1987 0.8 0.2 0.3 0.9 1.8 0.4 NA 0.8 0.1 60 APPENDIX 11: Notes and Sources for Crop Response Data The crop response data are from the following sources for each country: Nigeria: FAO and Falusi's crop responses are mean responses to total nutrients. Falusi's response figures are based on the Institute of Agricultural Research (IAR), World Bank Projects and others, and are either from field trials or actual farms depending on whether they are derived from IAR or from World Bank projects. FAO crop responses are from actual field conditions. World Bank crop response data for the period 1976-80 are based on the Staff Appraisal Reports (SARs) for the Funtua, Cusan, Gombe, Lafia, and Ayangha projects, and the recent estimates (1984/85) from the mid-term review of Kanu, Bauchi, and Sokota Agriculture Development Projects (ADPs). These responses are for nitrogen in the presence of P and K. Mean Annual Rainfall in the different regions of Nigeria: Sudan Savanna--500-1000 millimeters; North and South Guinea Savannah--1000-1500 millimeters; Rain Forest--1500-4000 millimeters. FAO 1974; Falusi 1987; World Bank Files. Cameroon: IFDC data are drawn from IAR and the SEMRY project. The responses are from one year trials except for rice responses (SEMRY) which are from three year trials. IFDC's sorghum responses are marginal responses to total nutrients, rice and maize responses are to nutrient N, and coffee responses are average response to nutrient N. FAO figures are average responses to total nutrients. FAO 1974; IFDC 1986. Senegal: FAO crop responses are average responses to total nutrients. IFDC figures for groundnut and millet are marginal responses to nutrient P. IFDC trials in Senegal were conducted in farmers' fields during the years 1976 and 1977 under ideal crop management. The results were recorded for three different rainfalls in four regions of the Groundnut Basin. Mean rainfall in the groundnut regions of Senegal: North Basin--up to 350 millimeters; Central Basin--350-600 millimeters; North Sine-Saloum--600-800 millimeters; South Sine-Saloum--Above 800 millimeters. FAO 1974; IFDC 1980. Malawi: Crop responses for Malawi from ASA and World Bank are average response to nitrogen and phosphorus. FAO response data for hybrid maize are average responses to nitrogen. Ministry of Agriculture, Malawi 1987. Kenya: -/ Crop responses for maize and sorghum in Kenya are average responses to nitrogen and phosphorus. Coffee and tea response data are from the field trials conducted by the Coffee and Tea Research Foundations, respectively. 1/ The land classification given here provides only general tendencies and should not be considered as definitive. There are also great variations within each region. For instance, G. Stern observes ". . . Siaya, Busia, Elgeyo Marakwet and the Uasin Gishu [are classified] as high potential districts, while Trans Nzoia and Nakuru [are] classified as medium potential districts. I cannot imagine why Siaya and Busia should be thus classified . . . [Trans Nzoia] is every bit as good as the Uasin Gishu . . .. Both districts contain some of the most fertile land in the world (Mt. Elgon and adjoining plains in Trans Nzoia - and the "Burnt Forest Area" in Uasin Cishu) and both contain less fertile [land], but still highly productive areas capable of producing 8-10 tons maize/hectare (Personal communication with the authors). 61 They are average responses to nitroRen. IFDC 1986b; World Bank 1986b; Ministry of Agriculture, Kenya 1983. Classification of High, Medium, and Low Potential in Kenya The districts of Kenya have been classified as high, medium, and low based on their agroclimatic characteristics and the crop response to fertilizers. The high potential regions are predominantly in the humid West and Central Highlands where there is abundant rainfall and the low potential regions in the semiarid uplands and arid low lands. Rainfall level: Humid West--900-2500 millimeters; Central Highlands--600-2500 millimeters; Coastal--500-1400 millimeters; SemiArid--400-1000 millimeters; Arid Low lands--150-500 millimeters. Classification of Districts: High Potential Districts: Kisii, Siaya, Busia, Kakamega, Nandi, Kericho, Uasin Gishu, Elgeyo Marakwet, Nyandarau, Kiambu, Nyeri, Muranga, Kirinyaga, and Meru. Medium Potential Districts: South Nyanza, Kisumu, Bungoma, Trans Nzoia, Nakuru, Narok, West Pokot, Embu, Taita, Kiwale, and Kilifi. Low Potential Districts: Laikipia, Kajiada, Baringo, and Machakos. Tanzania: FAO crop responses are average responses to nitrogen and phosphorus. The response data from the World Bank is mean response to nitrogen. Hhella 1985; Rioseco 1989. APPENDIX 11: Table 1: Fertilizer recommendations of IFDC and ISRA (1980 and present) for millet in Senegal kilograms/hectares Region N P K IFDC: North Basin 30-49 27-32 0 Central Basin 67-85 32-37 0 N. Sine-Saloum 51-70 40-45 12-23 S. Sine-Saloum 34-53 34-39 40 ISRA: Prior to 1980 North Basin 21 10 10 Central Basin 21 10 10 South Basin 60 31 31 ISRA: Post 1980* N P K S Millet recommendations were stated in terms of production objectives Production of 1 ton of millet 60 31 0 10 Production of 1 ton of millet 60 31 30 10 (Recommendations were not specific to any region for ISRA's post 1980 recommendations.) 62 APPENDIX 11: Table 2: Fertilizer recommendations of IFDC and ISRA (1980 and Present) for groundnuts in Senegal kilograms/hectares Region N P K IFDC: North Basin 0 10-21 0 Central Basin 10 19-31 0 N. Sine-Saloum 0 30-41 20-36 S. Sine-Saloum 0 40-51 32-39 ISRA: Prior to 1980 North Basin 15 15 12 Central Basin 9 30 15 South Basin 12 27 40 ISRA: Post 1980 N P K S North Basin 0 15 12 12 Central Basin 0 30 12 12 South Basin 0 27 40 15 Source: IFDC 1980; Kelly 1988. APPENDIX 12: Notes and Sources for Tables 23 and 24 Benefit-cost ratios were computed primarily for food crops using subsidized and unsubsidized prices for fertilizer. The fertilizer price for countries .with considerable overvaluation in the exchange rate has been corrected using purchasing power parity exchange rates. The effects of removing the implicit subsidy were estimated. Benefit-cost ratio with a subsidy The benefit-cost ratios with a fertilizer subsidy are computed using the subsidized price of fertilizers and the producer price of crops for the specific year. Crop response coefficients from several sources were used. (For more details see Tables 17 and 18, and Appendix 11). Benefit-cost ratio without an explicit subsidy The benefit-cost ratios without an explicit subsidy were computed using the unsubsidized price of fertilizers, without adjusting for an overvaluation or undervaluation in the exchange rate, and the producer price of crops for the specific year. The unsubsidized fertilizer price was estimated for each country as follows: Malawi (1987/1988) The subsidy element of 22 percent was removed from the subsidized fertilizer price. This figure was the proposed rate of subsidy for 1987 based on the Fertilizer Subsidy Removal Program that emerged from the negotiations of the Third SAL (Nathan Associates 1987). Tanzania (1987) Subsidies in Tanzania were abolished in 1984. Fertilizer prices for 1987 are thus unsubsidized prices. For 1988/89, a subsidy to the extent of 60-66 percent has been estimated (Rioseco 1989). Cameroon (1987) Unsubsidized prices are estimates from IFDC (IFDC 1986a). Senegal (1987) Fertilizers sold are predominantly unsubsidized. Price data obtained from Government of Senegal. Nigeria (1985) Unsubsidized prices are estimated by IFDC (IFDC 1985a). (1987) The subsidy of 82 percent was removed from the price of fertilizer. Benefit-cost ratio without an implicit subsidy Cameroon, Senegal, and Tanzania showed significant overvaluation in their exchange rates for 1987. Benefit-cost ratios without an implicit subsidy were computed for Cameroon, Senegal, and Tanzania for 1987 and for Nigeria for 1985. In 1987, the currencies in the francophone countries were generally overvalued by 20 percent. In Senegal, the overvaluation is estimated to be over 10 percent. The equilibrium rate for Tanzania in 1987 was Tsh 100 per US$ (Personal communication with the staff of AF5AG and AF6AG, World Bank). 63 Nigeria (1985) The unsubsidized fertilizer price for 15:15:15 adjusted for overvaluation in the exchange rate was calculated using the border price (in naira) adjusted for overvaluation plus 20 percent of intermediary costs (defined as the difference between the c.i.f. price and the farm gate price), also in naira adjusted for overvaluation, plus 80 percent of intermediary costs in naira not adjusted for overvaluation (this represents internal costs). Based on the distribution cost components of fertilizers in Nigeria for 1985, it is assumed that 20 percent of the intermediary cost of storage and transportation is paid in foreign exchange (IFDC 1985a). Tanzania (1987) The unsubsidized price of A/S corrected for overvaluation in the exchange rate is equivalent to the border price of A/S (in Tsh) adjusted for overvaluation, plus 30 percent of intermediary costs in Tsh not adjusted for overvaluation. A large part of the intermediary costs in Tanzania is transport cost, and it is assumed that this part of the cost is paid in foreign exchange (FAO/World Bank 1987). Senegal (1987) Due to the lack of reliable data on the breakdown of fertilizer distribution costs in Senegal, we assumed that the intermediary costs were 35 percent of the unsubsidized farm gate price. The border price was estimated using this assumption as well. In Senegal, transportation costs are relatively economical and are estimated to be 20 percent of the intermediary costs. The fertilizer price was corrected for overvaluation by adjusting the border price and the transportation (Jammeh 1987b; the transport cost components in Jammeh's study relate to paddy). Sources: Producer prices for crops are from the World Bank database. For Nigeria, the prices are from World Bank 1988, and crop prices used were the prices in 1987, compiled by the Bank mission from the Kaduna Markets. Coffee prices for Kenya are the net prices paid by the Coffee Board of Kenya to the Kenya Planters Cooperative Union (KPCU), but data on the amount paid in 1987 was not available. In 1987, there was a decline of 28 percent in Kenyan coffee prices compared to international prices; it is assumed that the price paid to farmers was affected at the same rate. Therefore, the 1987 prices were derived by reducing the 1986 prices by 28 percent. This data came from personal communication with the World Bank, Eastern Africa Department (AF2AG). Coffee prices for Cameroon came from personal communication with the World Bank, Occidental and Central Africa Department, Agriculture Operations Division. Subsidized and unsubsidized fertilizer prices come from: Malawi: Nathan Associates 1987. Tanzania: Mhella 1985; FAO/World Bank 1987; Rioseco 1989. Kenya: Tisminieszky and Kimuyi 1986; Ministry of Agriculture 1987. Cameroon: IFDC 1986a; personal communication with the Government of Cameroon. Senegal: Personal communication with the Government of Senegal. Nigeria: IFDC 1985a; Lele and Bindlish 1988. 64 APPENDIX 13: Distance, number of outlets, and areas served by retail outlets in the HADIA countries Average Distance Cultivated from Farm to Retail Estimated Area Served Country Outlet (kilometers) Number of Outlets per Outlet (hectares) Nigeria 22 1,315 9,500 (10 - 15) 1, Kenya 10 - 13 2/ Malawi 5 - 7 3 1,349 2,698 4/ Senegal 20 - 25 - (7.5-10) 5/ Tanzania 23 - 1/ Figures in parentheses are for areas covered by the Farm Service Centers (FSC) in the Nigerian ADPs. The role ADPs have played in setting up a reliable retail distribution network for fertilizers thereby making fertilizer more accessible to farmers has been significant. The average distance of FSCs and their estimated number in each of the ADPs is shown below: Average Distance of FSC from Estimated Number of ADP Farms (kilometers) Outlets Ayangba 11 34 Bauchi 11 184 Bida 10 53 Ekiti-Akoko 12 12 ilorian 9 45 Kaduna 10 213 Kano 9 163 Lafia NA NA Oyou 21 9 Sokoto 14 143 Futana NA NA Cusua NA NA Gombe NA NA 2/ Figures are maximum distance from rural households to the market center in most parts of Kenya except the coastal regions. 3/ These data are somewhat misleading as some farmers travel up to 30 kilometers. The percentage of farmers affected by the recent nationalization of ADMARC is not known, but the GOM has commissioned a study on options for liberalizing fertilizer sales and utilizing outlets supplementary to ADMARC to better serve farmers. 4/ Figures for Senegal are for the Groundnut Basin and for fertilizers bought at CEPA--Agricultural Products Distribution Center. Figures in parenthesis represent the distance when purchased from SECCOS--Local Fertilizer and Seed Distribution Point. Nearly 80 percent of fertilizers are bought from CEPA. 5/ The number of retail outlets in Tanzania was reported to be only 13 in 1980. Since 1984, the 23 cooperatives are playing a major role in distribution of fertilizers to small farmers, and they are to be assisted by the primary cooperatives who will serve as subretailers. Since there is no information on the primary societies that are active in fertilizer distribution, their number has not been included in the above table. Sources: Nigeria: IFDC 1985a. Tanzania: FAO and World Bank 1987. Malawi: ADMARC 1986; Twyford 1988. Senegal: Kelly 1988. Area under cultivation: Malawi: National Sampale Survey of Agriculture and Department of Lands and Valuation. Nigeria: Lele, Oyejide, et al. 1989. 65 9. For further discussion of the efficiency gains resulting from Notes privatization, see van de Walle 1989, and Vernon-Wortzel and Wortzel 1989. 1. A recent FAO report stressed the problem posed by land IC. For export crops, the correction of producer prices through pressure in Africa. For example, it has been estimated that 28 of removal of distortions has already caused several countries to 51 African countries will have a population in excess of their overshoot output price adjustments. In the face of declining real agricultural capacity if the present levels of input use continue (G. world prices for many primary commodities (for example, ground- M. Higgins 1982, p. 101. For more details, see Lele and Stone 1989. nuts in Senegal, cotton in Cameroon), some countries have had to 2. FAO has constructed estimates on the amount of per capita reduce their producer prices IR. E. Grilli and Maw Cheng Yang rainfed land available and the minimum per capita land required 1988). under low and intermediate input technologies. The pressure on 11. The extent to which Kenya's monopoly marketing arrange- rainfed land, especially in Kenya, Senegal, and Nigeria, if low ment for maize has been a constraint to growth of maize inputs are used is clear from the details below: production and fertilizer consumption is likely to be a controver- Rainfed Land Requirement Available Rainfed Land sial issue. Between 25 and 32 percent of Kenya's maize has been by Government purchased by the marketing board during the last fifteen years, Estimates and its share has remained constant except for a decline in the Low Intermediate 1985 2000 years 1978/79 to 1980/81. Some observers have argued that, in years of a good crop, the maize board has lacked the necessary hectares/person cash to purchase all maize offered to it, and the producers have Kenya 2.8 0.6 0.7 0.4 not had an alternative marketing channel owing to restrictions on Senegal 2.7 0.5 25 1.0 interdistrict movements of grain. Similarly, the argument goes, Nigeria 0.9 0.2 0.7 0.4 'producers have not benefited from higher prices in years of Malawi 0.6 0.2 0.8 0.5 shortages by being able to carry maize across regional bounda- Cameroon 0.4 0.2 3.5 2.0 ries. While these observations are correct, year-to-year fluctua- Source: Personal communication with Mr. G. M. Higgins, FAO. tions in producer prices and in the profitability of fertilizer use are likely to be substantially greater in the absense of a price 3. The food self-sufficiency ratio is defined as the ratio of support and, thus, in turn are likely to offset demand for fertilizer domestic production to total availability, in other words domestic use. production + net imports. 12. The fact that Kenya's policies require only small changes 4. Raising agricultural productivity through increased use of should not be cause for complacency. Growing land pressure in fertilizer also offers the prospect of relief in the longer term from that country combined with the absence of easy options for some environmental problems, for example, declining soil fertility further increasing smallholder production means that increasing and deforestation, created by the overuse of land that accompan- fertilizer use in Kenya is especially urgent (Lele and Meyers 1986). ies population pressure. Increased agricultural productivity 13. It should be noted that actual fertilizer use is especially reduces the amount of land needed for food production and, difficult to estimate. The data used here are based largely on furthermore, can be accompanied by a transfer of labor to government estimates which, in turn, typically rely on import data nonagricultural sectors. adjusted for stock holding. Therefore, the use of consumption 5. ODA to Sub-Saharan Africa has tripled in the last fifteen data cited here more correctly reflect what is available for use by years. From about $4 billion in 1973, donor assistance increased farmers. to nearly $11 billion (at constant prices) in 1986, (OECD, Geograph- 14. The growth rates in Figure 2 and Appendix 3 are for ical Distribution of Financial Flows, Various Issues). aggregate fertilizer use, rather than for per hectare consumption, 6. Africa's share of the developing world's fertilizer consump- because reliable annual data on arable land are difficult to obtain. tion was 11.9 percent in 1970/71 and 7.0 percent in 1986/87, while In order to present a comprehensive picture of fertilizer use, data Sub-Saharan Africa's share of developing world consumption was are presented for as many years since 1970 as possible. However, 3.0 percent in 1970/71 and 2.2 percent in 1986/87. Over this problems with the availability of data in certain countries mean period, the developing world's share of world consumption that the growth rates were computed for different periods: Kenya, increased from 19.6 percent in 1970-71 to 37.8 percent in 1986/87. 1974-87; Cameroon, 1974-86; Tanzania, 1974-87; Senegal, 1970-87; 7. The Bank's justification for eliminating subsidies addressed Nigeria, 1972-87; and Malawi, 1972-87. Even when the growth rates each of the five points in favor of subsidies that were presented are calculated fdr the years in common (1974-1985/861, the above. It was argued that, first, the benefits of fertilizer use are rankings do not change and the values do not change significantly. well known even in the least dynamic agricultural systems and a The growth rates of fertilizer use were estimated using a log-linear subsidy does little to teach the appropriate use of fertilizer; regression equation. second, the alleged risk associated with the use of fertilizer is 15. The interactions between rainfall and human and animal overstated and, with regard to credit constraints, it is better to health have resulted in high population densities in drier areas of correct the credit market imperfection directly rather than Africa where not only are health hazards limited, but so are attempt to offset it with an input subsidy; third, fertilizer production possibilities. In contrast, in millions of acres of well- subsidies frequently help the richer peasants more than poorer watered land, hazards to human and animal health keep popula- ones, since the former are better able to afford the subsidized tion densities low. In drier, semiarid areas, greater concentration fertilizers and are more likely to have access to supplies of of excessive population has already begun to threaten fragile fertilizers; fourth, the argument that fertilizer subsidies promote ecologies. In a separate paper, Lele and Stone (1989) investigate soil conservation is only valid ". ..for a temporary subsidy where the relationship of population densities to the process of population growth has accelerated and farmers may not learn intensifying agriculture and achieving food security. about fertilizers fast enough to prevent severe damage to soil 16. In Tanzania, for instance, the requirements per ton of quality" (World Bank, World Development Report 1986, p. 96); fifth, transporting bulky maize was much higher compared to export fertilizer subsidies intended to offset an implicit tax on small- crops such as coffee, cotton, or tobacco (Schluter and Sackett holder output may not offset the distortion but may introduce 1982). For this reason a policy of concentrating food production in greater distortions in the longer term as that tax changes. See also high potential areas in Kenya is likely to be successful, given its Repetto 1988, pp. 19-25. well-developed and well-maintained transport network; while in 8. For a detailed account of agricultural marketing reform in the Tanzania, with its widely dispersed population and poor transport MADIA countries, see Lele and Christiansen 1989. infrastructure, such a policy is probably not feasible. 66 17. See Lele 1988; Lele and Gaviria 1989. in the central or northern Basin such as Thies, Diourbel, and 18. In 1986/87, of the total maize (white) sold by NCPB, nearly Louga. As in Kenya, fertilizer policy has not adequately been 72 percent was in the Coastal province and about 9 percent in the focused on priorities in fertilizer promotion. Policies should relate Rift Valley. to the physical responsiveness to achieve rapid growth and to the special dispensation which may be necessary in the areas of low 19. The data on fertilizer use by smallholders and estates are and variable responses for populations that have relatively few available on a yearly basis for Malawi: income generation and employment alternatives. Fertilizer use by smaliholders and estates in Malawi, 1972/73 to 1987/ 25. Agricultural production units may be classified by size into 88 industrial plantation, large family farms, small market-oriented Year Fertilizers in Nutrients (Tons) Smaliholder Percentage farms, and small farms (marginal or below subsistence). Lele and Smaliholders Estates of Total Agarwal 119891 have pointed out that definitions of large and small 1973 5,602 10,772 34 farmers vary considerably among countries. Whereas in Kenya 1974 7,793 9,841 44 farm size is defined in terms of the area cultivated, in Malawi it 1975 3,561 10,300 25 is defined by the differential rights to grow and sell crops in 1976 5,130 9,066 36 specified markets. Thus, estates in Malawi are those that can grow 1977 7,030 9,279 43 export crops and sell in auctions, while smallholders are those 1978 10,500 13,600 44 who live on customary land and sell their produce to ADMARC at 1979 1.,300 15,700 42 fixed prices, which are one-half to two-thirds of the price paid to 1980 12,100 15,400 44 estates in auction. 1981 15.800 16,100 50 1982 13,900 14,700 49 26. Per hectare fertilizer use (in terms of product) among 1983 17,000 15,400 52 smallholders in Kenya is less than among the smaliholders in 1984 17,800 14,000 56 Malawi (assuming there is no leakage to the estate sector). This 1985 21,200 17,000 55 evidence needs to be interpreted cautiously as the years are 1986 19,000 17,700 53 different, and comparisons are for product fertilizers used and not 1987 22,200 17,700 56 in nutrients; details were not available separately for smaliholders 1988 28,200 18,100 61-- and estates in Kenya. Source: Nathan Report 1987; Agriculturist's Report No.1, GOM 1986; Economic Report, GOM 1988; USAID Office in Malawi. Product Area Cultivated Fertilizer Used Kilograms 20. The smaliholder sector has been able to obtain nutrient at (hectares) (tons) per Hectare lower prices than the estate sector, because of the competitive Kenya (1985/86) bidding process used by the SFRF; the estate sector has been Large Farms 1,015,900 136,635 134 forced to rely on a single local firm (Opticheml that has a South Small Farms 2,693,600 101,760 38 African parent company. Malawi 119811 Large Farms 319,364 44,679 140 21. Small Farms 1,332,000 64,448 48 Fertilizer use in Tanzania by region for selected years Source: Data on area under cultivation and fertilizer used are from the Region 1986/87 1985 1980 1975 ~~~~~~following sources: Region 1986/87 1985 1980 1975 Malawi: Smailholder area from National Sample Survey of Agriculture, Tonnes 1980/81 and estate area from Department of Lands and Valuation, Iringa 35,792 25,233 2,224 17,625 1981. Fertilizer used from World Bank Internal Memorandum, dated Mbeya 25,055 27,232 16,085 7,875 3/1/89. Ruvuma 20,743 22,655 17,675 6,891 Kenya: Smallholder and intermediate area from CBS Int. Rural Survey, Tabora 15,546 13,217 12,334 17,767 1976-79 and estate area from Large Farm Survey 1978. Fertilizer used Kilimanjaro 7,163 7,982 5,901 4,645 from Ministry of Agriculture, Kenya 1987. Rukwa 9,976 4,393 6,540 2,258 Morogoro 1,556 2,902 4,653 2,831 27. For a more detailed discussion of the performance of estate Tansgha 13,268 2,317 3929 27303 and smallholder agriculture in the three East African MADIA Kigoma 763 2,033 1,661 3,427 countries, see Lele and Meyers 1987. Shinyanga 1,951 1,112 2,606 2,207 28. With respect to export crops, the impact of an overvalued Mara 727 1,158 2,840 727 exchange rate depends on adjustments to producer prices. If Mwanza 2,086 1,642 2,923 5,859 producer prices are increased, the effects of the overvaluation on Source: FAO/World Bank 1987, Annex 3. production incentives can be neutralized. In the case of food 22. Thirty-five percent of the population lives in the Northern crops, incentives are chiefly influenced by the prices that prevail region. In thie Midl Belt, where rainfall is higher and more in informal (or, where controls exist, parallel) markets and the reliable and fertilizer response greater, population densities have ability of governments to fix statutory prices and exert monopoly been much lower i53 persons per square kilometer compared to control over purchases. Witness the government's control of maize 88 in the Northern Belt, and 236 in the Southern Belt). This has prices and marketing in the three East African countries com- caused problems in intensification owing to labor and manure pared to the absence of government influence (with the exception shortages and has led to a drive for larger farms. The absence of of rce) i the three West African countries. higher populations in the Middle Belt has historical reasons. 29. The proportion in Figure 3 for Malawi is for smallholder 23. Fertilizer use is now diversified among different crops. For agriculture. For more details, see Nathan 1987. instance, in 1980, groundnuts and millet/sorghum accounted for 30. Considering only smallholder fertilizer use 143 percent of 81 percent of fertilizer use (42 and 39 percent, respectively), total use in 1985/86), approximately 29 percent was on coffee, 15 whereas rice and cotton accounted for only 19 percent. In 1982, the percent on tea, and 20 percent on sugar. Only a fifth of total shares for groundnuts and millet/sorghum began to decline and smaliholder fertilizer use was on maize, although the area kept were only 50 percent by 1985. under maize by smallholders is eight times that under tea and 24. As in the case of Kenya, fertilizer responsiveness varies coffee. substantially among regions, with better watered areas in the 31. The 1984 Agriculture Census data (based on where fertilizer southern Groundnut Basin, such as Sine-Saloum, showing greater is used) indicates that the percentages of fertilizer consumed in response to fertilizers for groundnut and millet than the regions Littoral, Northern, and Western regions in 1984 were 19.2, 25.1, and ,67 37.4, respectively, compared to the IFDC's estimates (based on 1977, the MRP has been based on the c.i.f. price of fertilizer at where fertilizer is purchased) of 27.2, 20.0, and 28.5 percent for the Mombasa plus 30 percent plus KSH 100 per ton for the f.o.r. same period. IFDC presumes that from 6,000 to 12,000 metric tons Mombasa price. There is, however, typically more discretion in the of fertilizer purchased in the Littoral province was actually used price setting exercise than this would indicate- To this price is in the West. Field reports of coffee fertilizers being used on maize added transport costs between Mombasa and district centers. or applied to coffee for the benefit of maize (generally inter- The selection of importers has been a source of variability in cropped with other crops in Cameroon) are widespread in the volume of imports, because many license holders fail to Cameroon. exercise their quota rights by not importing fertilizer. Since the 32. The principal crops in the remaining provinces are cocoa, in approval of import licenses is carefully tailored to estimated the Center/South, and robusta in the Southwest. One explanation demand in order to conserve foreign exchange, the failure of even for this concentration is that most fertilizer is distributed in a small importer to exercise quota rights can lead to a fertilizer connection with development programs for individual crops, shortage. Among the reasons for importers not exercising their especially export crops such as cotton and coffee. quota rights are lack of assured customers, lack of financing, Data on fertilizer sales provided in the 1984 Agriculture Census uncertainty about the price level to be set by the government, give a slightly different picture than the use data, since sales are and the late allocation of import quotas IB. Tisminieszky and P relatively concentrated, with approximately 80 percent in 1985 Kimuyi 1986, pp. 3-5; Agriconsult 1988, p.3). occurring in three provinces, West, Northern, and Littoral, (see 38. For instance, Agriconsult 1988; B. Tisminieszky and R table) that contain about 56 percent of the population. Approxi- Kimuyu 1986; and World Bank 1986b. mately 75 percent of subsidized fertilizer was sold in West and 39. Increasing the number of importers has raised questions Littoral provinces, which are primarily coffee growing areas, while about the loss of scale economies in import procurement about half (54 percent) of unsubsidized fertilizer was sold in the resulting in increasing unit costs, It is still too early to ascertain Northern provinces. the consequences of liberalization in this area. 40. The KTDA and coffee cooperatives alone accounted for Fertilizer sales by province, 1981-85 approximately 22 percent of the fertilizer consumed in 1984/85 Province 1981 1982 1983 i9984 1985 (World Bank 1986b). Percent 41. The level of commercial fertilizer imports in 1985/86 was Central/South 0.83 0.65 0.85 1.35 0.95 190.282 metric tons; in 1986/87 it was 142,849 metric tons; and in West 27.21 33.43 23.27 28.47 19.89 1987/88 it totalled 83,908 metric tons, which was only 35 percent East 9.92 8.43 7.02 7.11 6.17 of the total imports. Northern 20.70 19.12 19.54 19.98 25.88 42. For instance, districts like Machakos and Kitui, which have Littoral 27.07 24.24 33.22 27.24 34.75 over 71 percent and 50 percent, respectively, of the areas Southwest 7.78 8.95 7.49 7.03 7.87 cultivated under maize, have yields less than half the average for Northwest 6.49 5.18 9.40 8.82 4.50 the country. Source: IFDC 1986a. 43. For instance (even in the high and medium potential The discrepancy between the sales data and the use data in regions) the gap between actual input use and recommended the 1984 Agriculture Census suggests that much fertiiizer is not levels varies between 3 percent and 5 percent of recommended used in the province where it is sold. Further, it suggests that rates in Nyanza province to the highest level of 43 percent and 60 much of the fertilizer bought in the Littoral and Southwest is percent in Trans Nzoia. A recent World Bank estimate puts the actually resold for use in the West and Northwest. Despite the gap between the present and recommended levels of fertilizer inconsistencies in data on the geographical distribution of use on maize around 100,000 tons. Closing this gap is estimated fertilizer, it still seems likely that most fertilizer is applied to to bring an increase in maize production by 400,000-600,000 tons coffee and cotton since these are the principal crops of these (World Bank 1986b). provinces. 44. The TFC is a parastatal and is under the supervision of the 33. For our purposes, we define a subsidy in terms of whether Ministry of Industries. Although the TFC plant was designed to the pricing policy requires government expenditures. In other produce 125,000 tons of product annually (AS, TSP, and MPK words, this is a fiscal or financial definition as opposed to an compounds), historically it has operated well below that limit economic definition. (between 13,000 and 69,000 tons) for a variety of reasons. With the 34. For example, lapan, Germany, Finland, France, the United exception of phosphate, all materials are imported. Since 1983, States, Norway, Sweden, and Switzerland have indicated that they phosphate rock has been supplied to TFC from a mine near will untie at least percentoftheirassistaArusha. However, this rock is of poor quality and is unable to will untie at least 50 percent of their assistance,.ettecunr' ead 35. Using development projects to acquire fertilizer helps to meet the country2s demand. explain the relatively low reported use, but perhaps higher actual 45. In 1975 only 12 percent of imports were grants, while in use, of fertilizers on food crops in Cameroor, observed by the 1985/86 and 1986/87 the total quantity of imports was supplied World Bank's supervision missions to coffee areas. Because the from grants (see table below). Since 1980, an average of 94 percent fertilizer provided for use on coffee (ammonium sulfate, urea, or of Imported fertilizer has been covered by grant aid agreements 20:10:10) is not appropriate for maize in all regions, especially as ,FAO/World Bank 1987, p. 6). a basal dressing, there is obvious inefficiency in fertilizer use with negative consequences for soil fertility. Obviously, availability of a Sources of fertilizer supply wider range of fertilizers (both on cash and credit terms) Local Total combined with an improvement in farmer knowledge of the Year Imports Grants Produced Supplies Offtake correct application of fertilizers to specific crops would improve 1975 56,612 6,860 59,571 116,183 93,541 efficiency. 1980 68,641 63,141 50,852 119.493 107,091 36. IFDC recognizes DAP to be the fertiiizer of choice for iood 1981 54,449 44,599 69,031 123,480 96,569 crops in Cameroon. However, since sulphur deficiency is becom- 1982 85,783 79,783 13,662 99,445 82,409 ing a constraint to crop production in Cameroon, 1FDC has 1983 57,433 51,433 31,237 88,670 89,874 proposed a complex fertilizer 10:30:10:5S, taking into account the 1984/85 87.030 82,230 48,230 135,260 109,675 varied soils and crop growth conditions in Cameroon. 1985/86 1128276 112,276 33,960 146.236 118.436 varied soils and crop growth ~~~~~~~~~~1986/87 98,950 98,950 46.790 145,740 130,124 37. Officially, the government controls the orice through the Source FAO/Worid Bank 1987, p. 6. Price Controller who sets the Maximum Retail Prices NM.RP). Since 68 46. For more details on cooperatives and their relations with 52. Malawi's transport problems stem from its landlocked governments in MADIA countries, see Lele and Christiansen 1989. position and the civil war in neighboring Mozambique, which has 47. Lengthy procedures of loan application, poor logistic caused the closure of the traditional Nacala and Beira routes. As planning, and input supply distribution are some of the inherent a result, the government must rely on domestic production to problems limiting the performance of cooperatives. The preferen- insure food security. tial retail traders, like the TFA, Agricultural and Industrial Supplies 53. The SFRF was established in 1983 after the marketing Company (AISCOJ, and others, whose market share was 13.5 parastatal ADMARC was unable to provide timely supplies of percent in 1986/87 mainly distribute to medium- and large-scale fertilizer because of liquidity problems. Currently, the SFRF, in farmers, with both their outlets more concentrated in major conjunction with the Ministry of Agriculture, estimates small- townships. A major constraint to private entrepreneurs in operat- holders' fertilizer needs and is solely responsible for procurement ing retail outlets is their inability to obtain necessary trade of supplies on the basis of international competitive bidding. The licenses from local authorities. SFRF has also been instrumental in improving the packaging of 48. As with most other farm implements, fertilizer prices and fertilizers (smaller bags to meet the needs of very small farmers), margins are set by the government based on the recommenda- increasing the number of outlets to improve accessibility, and tions of TFC. Historically, these prices have represented a (together with the EEC) developing storage facilities for a fertilizer significant subsidy. In 1973, in order to boost food production, free buffer stock-a need in Malawi, given the transportation prob- fertilizer was offered to maize growers. This policy lasted for two lems. As a result of the SFRF, fertilizer imports have been timely seasons 11973/74 and 1974/75) and was replaced in 1975/76 with and average procurement costs have been lower than those of the a 75 percent subsidy (World Bank 1982, p. 4). In recent years, the private company, Optichem, which imports fertilizer from its subsidy covered internal transport costs in addition to a 40 to 50 parent company in South Africa. A crucial element in this success percent reduction in ex-factory/landed costs depending on the has been access to foreign exchange, and the capacity being built type of fertilizer. Beginning in 1984, the fertilizer subsidy was to in the Ministry of Agriculture to carry out timely and accurate be removed, a step that resulted in an increase in average estimates of fertilizer demand. The creation of SFRF as a separate fertilizer prices of 150 percent. At present, fertilizer subsidies are entity insures the liquidity needed to procure supplies for the officially abolished, but prices remain controlled and are uni- subsequent year. As discussed earlier, one drawback to the formly fixed by fertilizer type (FAO/World Bank 1987, p. 14). design of the SFRF is that its funds are denominated in local 49. The correlation between farm size and the ability to bear currency lkwacha) with the result that devaluation as mandated in risk is due to the ability of farmers with more than 1.5 hectares to the SAC agreements eroded the purchasing power of the SFRF bear the risks of technological innovation better than farmers with For more details on the SFRF, see Appendix 7. less land. As a result, these farmers are regarded as more 54. Between 1979/80 and 1986/87, the price of the two most creditworthy than smaller-scale farmers and, therefore, have widely used fertilizers increased by approximately 150 percent, greater access to credit, extension, and purchased inputs in while the producer price of maize increased by 85 percent. An comparison to farmers with less land. For example, the high evaluation of the subsidy removal program (Nathan Associates correlation between credit use and size of land holding is due in 1987) acknowledged that the elimination of the subsidy accounted part to the method by which credit is supplied to farmers, in for only a small portion of the increase in the price of fertilizer other words, through credit groups sponsored by the National and that the more important factors were the devaluation of the Rural Development Program (NRDP). At present, the entire local currency and the increased external transport costs. As an membership of a credit group is held responsible for the failure incentive to encourage the subsidy removal, USAID offered $15 of any individual member to repay a loan. Although this approach million in tranches to supplement the SFRF. to credit effectively reduces the lender's risk, it also discourages 55. The process of acquiring fertilizer through the government the group from admitting farmers who are viewed as less channels is cormplicated. Subsidized fertilizer is supplied by the creditworthy than the average member, thereby discriminating Ministry of Agriculture and a number of specialized agencies that against small-scale farmers. Further, many credit clubs require collectively are responsible for estimating fertilizer needs and that members pay a fixed membership fee, a device that acts to allocating import quotas. The Ministry of Commerce and Industry restrict membership to more affluent farmers. Through this historically has been responsible for granting fertilizer import program, NRDP provided credit to about 16 percent of farmers in licenses. In addition, two other ministries have been regularly 1985/86 and recorded repayment rates as high as 95 percent. The involved in procurement, and on occasion, approval from the data on the quantity of fertilizer applied by participants in NRDP president's office is required. The process of estimating fertilizer are from sample surveys in only three districts, and they show that needs begins in November, with delivery to farmers planned for farmers receiving credit account for over 75 percent of fertilizer the following june. This schedule, however, is frequently missed, consumption, while 65 percent of farmers who do not use fertilizer with the result that all or part of the fertilizer imports arrive late. cite a lack of credit as the primary reason. Once the fertilizer arrives, it is usually necessary for FONADER 50. The cost of fertilizer use can be seen by comparing the cost (Fonds National pour le Developpement Rural) to allocate quotas to the of a fertilizer package to average income. In 1986/87, the cost of various categories of users, since imports are commonly short of fertilizer recommended by the extension service for one hectare estimated needs, and arrange for transport to regional depots. In of hybrid maize (150 kg) was K 100 ($50) while the per capita princple, the responsibility for arranging transport rests with each income of most small farmers is less than $100. institution. In practice, few organizations have the financial 51. As discussed earlier, the reported growth of smaliholder resources to prefinance fertilizer purchases and arrange for fertilizer consumption has been more rapid (about 11.7 percent delivery to regional distribution centers. between 1972/73 and 1987/88) than that of the estate sector 56. Although Cameroon has a large arable land area with a wide (about 4.5 percent), although the problem of leakage of subsi- range of climatic conditions, suggesting a great potential for dized smaliholder fertilizer to the estate sector suggests that this agricultural development, the soils of Cameroon are of relatively disparity is overstated. Estates constitute about 40 percent of the low fertility, with the exception of the Western Highlands. There total fertilizer use and purchase the majority of their supplies are two factors that account for this: fi) the low level of phosphates from a private firm lOptichem). However, the higher prices in most soils (nearly 80 percent of the arable soils are deficient charged by Optichem and its small number of outlets (2) create in phosphates), and (ii) a widespread sulphur deficiency in the strong incentives to buy from outlets intended for the exclusive soils. use of smaliholders. The leakage has been variously estimated to Crop intensification and increased use of ammonium sulfate, be between 17 and 25 percent of smaliholder sales. which is, along with 20: 10: 10, the most commonly used fertilizer in Cameroon. also contribute to the problem of soil acidification. 69 The solution to the problem of acidification lies in an increased the subsidy will be reduced to 10 percent on a maximum of 60,000 use of calcium carbonate in conjunction with the nitrogenous tons. For all subsequent years, the subsidy is to be eliminated. fertilizers. Each kilogram of nitrogen (if urea or ammonium nitrate Donor concern over the budgetary cost of the subsidy is is the nitrogenous fertilizer used) requires 1.8 kilograms of somewhat misplaced since the cost is small. The forecast budget calcium carbonate to neutralize the soil acidity. However, if the for 1986/87 was CFA 800 billion. The subsidy costs approximately fertilizer used is ammonium sulfate, then 5.2 kilograms of calcium CFA 10.5 billion, less than 2 percent of the total budget. carbonate is needed. 59. The FSSRP (Fertilizer Sub-Sector Reform Program) includes Many countries have started to replace ammonium sulfate and a credit line for the importation of fertilizers and another credit ammonium nitrate with urea because of its delivered cost line for distribution of fertilizers. The commercial banks will now advantage. In Cameroon, however, all subsidized fertilizers are grant loans to importers and distributors by drawing on funds put priced the same, irrespective of their nutrient content. In terms of at their disposal by the Bank of Credit and Commerce- the acidification problem, urea is preferable to either ammonium Cameroon (BCCC) which manages the Credit Fund and Subsidy sulfate or ammonium nitrate, but ammonium sulfate and ammo- Fund. nium nitrate continue to be the fertilizers predominantly imported and supplied by the government. In part this is 60. One report estimates that 20 percent of the coffee and explained by concerns about the sulphur deficiency (ammonium cocoa grown in the country is not harvested because a lack of sulfate contains sulphur); it is also based in part on IAR's soil credit prevents farmers from hiring the necessary labor. This is studies. This, however, ignores the more serious problem of similar to the situation among tea growers in Kenya where it is acidification of the soils by ammonium sulfate. estimated that as much as 50 percent of tea is not picked because It is worth noting the actions taken by SODECOTON in this of a lack of cost for hiring labor (Schluter 1985). matter. SODECOTON has been responsible for importing the 61. Historically SODECOTON has subsidized the fertilizer sold appropriate complex fertilizers (IFDC 1985), which have the to farmers, but in a way that has been indirect and difficult to necessary amount of phosphorous, nitrogen, potassium, and the estimate. The question of the size of its subsidy paid by micronutrients sulphur and boron. SODECOTON has adequately parastatals is complicated. The subsidy on fertilizer was financed demonstrated to its farmers the need for the micronutrients and out of the general subsidy paid to SODECOTON by the govern- the additional phosphorous required for the Cameroonian soil. ment. Due to lack of detail on these arrangements and the SODECOTON is also said to ensure that fertilizer recommenda- assertion by SODECOTON that the subsidies are to be elimi- tions are carefully followed by all cotton growers. nated, they are not addressed here. On the other hand, the use on coffee of ammonium sulfate, 62. SEMRY was an autonomous development society whose which also finds its way to food crops (primarily maize), solves the main shareholders were the government, ONCPB, and BCD. It was sulphur deficiency problem, but ignores acidification. Moreover, charged with the promotion of modern rice cultivation in the the need for a high-phosphate starter fertilizer (such as DAP) for irrigation perimeters of the Longone River in the North. SEMRY maize must be recognized. Replacing ammonium sulfate with DAP was recently dissolved by the government at the request of the alone cannot give an answer. DAP would take care of the farmers in the area, who agreed to take over its development phosphorus needs of maize and the soil acidification problem but functions. not sulphur deficiency (DAP unlike ammonium sulfate has no 63. For a more detailed account of the cost of operations for sulphur in it), which is being increasingly recognized as a SODECOTON compared to similar organizations in other franco- constraint to crop production in Cameroon. phone African countries, see Lele, van de Walle, and Gbetibouo It is evident that Cameroon presents a more complicated 1989. situation than appears at first. The fertilizer products currently. 64. The history of fertilizer subsidy removal began with SAL 1, used, the rates at which they are used, and the way they are which called for the liberalization of fertilizer prices. Because of applied are not the most effective in terms of crop yields and problems with policy decisions by the Government of Senegal maintaining soil fertility. In the semiarid areas that are poor in soil (GOS), the planned liberalization was not implemented. Following fertility and have risky climatic conditions, there is reason for the cancellation of the first SAL, there was a period of consolida- greater concern. Farmers who are not adequately advised natu- tion and reassessment on the part of donors and the GOS. As rally "exploit" the land, and the natural fertility of the soil and the part of this retrenchment and the donors' search for a means to long-term adverse effects of agricultural exploitation on soil promote policy reform, USAID moved to concentrate its efforts on productivity are of major concern. This helps to make the point the fertilizer subsector. Partly as a result of this decision by that research, extension, effective distribution, and institutional USAID, but also because of the importance of fertilizer, it was the agencies, along with appropriate prices, are equally important focus of much dialogue between donors and GOS. Fertilizer was factors in making fertilizers a successful input. The subject of emphasized by USAID for several reasons: (i) the assumption that fertilizer should not be perceived in terms of a subsidy or price empirical evidence (IFDC 1986) indicated that fertilizer is econom- factors alone. ically feasible, especially for food crops; (ii) the subsector was 57. Of the $9 million investment in 1987 by USAID, $5 million seen as an area with potential for saving foreign exchange; (iii) will be used as credit to facilitate and partially finance private fertilizer distribution was perceived as inefficient, thereby offering bulk importation and initial warehouse storage, $2.5 million will the potential for substantial benefits from a reform program; [iv) be used as credit to encourage and partially finance private more efficient use of fertilizer would be reflected in increased sector distribution and retail sales operations, and the remaining rural incomes in a relatively short period of time; and Iv) the amount will be used to conduct fertilizer related studies. The elimination of subsidies would help reduce the government's obligation of the $11 million program in 1989 will be contingent budget deficit. In addition to these specific reasons for focusing upon continued subsidy reductions and farm gate price adjust- on fertilizer, USAID was concerned about its concentration of ments for the 1990 1991, and 1992 crop years, with total elimina- assistance in project lending. USAID argued that one of the tion of subsidies in 1992, and continued implementation and lessons to emerge from the experience of the failed SAL was that refinement of reforms in procurement. Further implementation of the success of project lending is heavily dependent on the reforms to liberalize and privatize the distribution system will broader policy environment. USAID therefore sought a strategy take place as agreed upon in Phase I of the two-phase program. that would allow it to influence the broader policy environment. 58. The fertilizer subsidy, which is currently approximately 65 Although USAID and the World Bank appear to have worked percent, is to be removed in four stages. During 1988, the subsidy closely on the reform of fertilizer policy, there is little mention of will be reduced to 60 CFA per kilogram (approximately 45 percent) the issue in connection with SAL 11 beyond some broad guidelines on a maximum of 60,000 tons. During 1989, the subsidy will be for fertilizer policy that the GOS was expected to follow (World reduced to 30 percent on a maximum of 50,000 tons. During 1990, Bank 1986, p. 13). These guidelines reflect many of the points 70 made in the government's New Agricultural Policy (NPA), which 70. In 1984, in an effort to address the problem of declining was launched in April 1984. As part of the NPA, it was agreed that credit availability and lower repayment rates, the government USAID would make a subsidy payment of CFA 24 per kilogram on created the National Agricultural Credit Fund (CNCAS). It was all fertilizer sold by ICS to private traders or producers groups capitalized by a combination of government, bank, and other (World Bank 1986d, p.1 office memo). Interestingly, the Bank private sources and was set up as an autonomous and decentral- seems to have accepted the argument that fertilizer subsidies ized body. The scope of operation of the CNCAS, though limited, were necessary in Senegal "... since at current market prices, is expected to ease the agriculture credit situation in Senegal. It imported fertilizers (primarily ureal or the special blends pro- has started operating only in selected regions on a pilot basis and duced in Senegal by ICS ... are not attractive to farmers" (World the expansion is expected to be slow. Under conditions prevailing Bank 1986, p.13). in Senegal, the private sector was not ready to undertake the risks 65. USAID describes its objectives with respect to fertilizer involved in selling fertilizers in a situation of no credit and policy as including: frequent droughts. SONACOS and UNCA, operating at selected (11 institutional reform, in other words, privatizing fertilizer points during the marketing period, were entrusted with distrib- distribution as part of a larger policy of supporting "disen- uting the new and less expensive binary fertilizers, which were gagement" by the State; not well-received by the farmers. SONACOS reported sales of 189 (2) reduction of public sector spending and debt through tons and UNCA 10 tons in 1986. elimination of fertilizer subsidies; 71. Kelly describes USAID's policy as follows: (3) increasing fertilizer consumption and presumably agricul- Despite the strong pressure against fertilizer subsidies, the tural production, particularly food production (Personal United States Agency for International Development IUSAID) communication with the authors). agreed to a limited subsidy for 1985-88. The objective is 66, The use of fertilizer on rice and cotton continued to twofold-to encourage fertilizer consumption and to discour- increase between 1982 and 1985 relative to that on groundnuts. age costly government credit and distribution programs. The This was because the projects working with rice and cotton were subsidy is limited to fertilizer that is distributed by the able to secure supplies of fertilizer and provide farmers with private sector and sold on a cash-and-carry basis. Fertilizer credit at a time when levels of fertilizer use in the Groundnut sold by government agencies will not benefit from the Basin had been sharply reduced as a result of the policy reform subsidy. This means that SODEVA's program to introduce measures (subsidy removal and privatization of distribution farmers to hybrid maize cultivation will not benefit even networks) and a lack of credit for small farmers (Jammeh 1987b; though maize is a crop which the World Bank credits with a lammeh and Lele 1988). sufficiently high fertilizer response to warrant temporary 67. The growth rates of rainfall (and t values) for each province subsidies. See Shalit and Binswanger 1984. Given the low are as follows: Cap Vert, -4.09 percent (-2.71; Casamance, -1.49 priority farmers place on fertilizer compared to food, seed, percent (-2.6); Diourbel, -3.16 percent 1-4.2); Louga, -2.72 percent and equipment investment, it is likely that much of the (-3.7); Eastern Senegal, -1.17 percent (-1.8); Sine-Saloum, -1.54 financing for this subsidy will go unspent (Kelly 1988, p. 22). percent 1-2.0); Thies, -3.07 percent (-3.2); Fleuve, -3.12 percent 72. For more details, see Kelly 1988. (-4.0). The Ziguinchor/Kolda and Tambacounda (Eastern Senegal) 73. The argument in favor of subsidy removal rests on the level regions, where the rainfall is longer, have experienced less of response coefficients in different parts of the country and on variability in rainfall. The coefficient of variation for each region for . accepting a benefit-cost ratio of less than 2 for many parts of the the period 1960-83 is: Cap Vert, 52 percent; Ziguinchor/Kolda Groundnut Basin. We are skeptical that such a low benefit-cost (Casamance), 20.5 percent; Diourbel, 31 percent; Louga, 32.4 ratio will be adequate. However, in order to fully assess such an percent; Tambacounda (Eastern Senegal), 23.0 percent; Kaolack/ argument it is necessary to rely on a detailed discussion of Fatick (Sine-Saloum), 30.4 percent; Thies, 36.1 percent; St. Louis, response coefficients. Therefore, we postpone the discussion until 31.5 percent; total Senegal, 23 percent. the section on benefit-cost ratios. 68. Another consequence of the declining rainfall pattern has 74. This section draws heavily on Lele, Oyeiide, et al. 1989. been a shortened growing season in the northern areas of the country, a factor that has contributed to migration from the North 75. Nearly two-thirds of the fertilizer use in Nigeria is in the to areas in the South that have higher rainfall levels. Northern states, where the first ADPs were located. Three factors account for the large impact of the ADPs on the growth of fertilizer 69. Throughout the life of Programme Agricole, from the beginning use: Ii) the network of Farm Service Centers and feeder roads of the 1960s until 1981/82, agricultural credit was provided by the established within the ADPs is convenient (in the case of the early government's development bank, La Banque Nationale pour le ADPs, farmers had to travel no more than 15 kilometers to Developpement du Senegal (BNDS), acting alone or in concert with a purchase inputs), (ii) the input supply companies of the ADPs are consortium of local banks. This was due in part to the absence of relatively efficient in comparison with the distribution procedures well-organized rural financial markets and to the unwillingness of of the ministries, and (iii) the ADPs have been given preferential private banks to make loans to smallholders. Commercial lending treatment by the government in the allocation of fertilizer to smallholder agriculture was regarded as high risk because of supplies In 1978, the four northern ADPs in existence accounted the farmers' inability to put up the necessary collateral. Land for 30 percent of fertilizer use but only 3 percent of total land could not be used to secure loans because farmers enjoyed only area. As the area encompassed by the ADPs has increased, the usufruct rights on what, since 1964, had become National Land. At discrepancy between the share of land area and total fertilizer the same time low farm incomes and lack of valuable assets made use has decreased. As recently as 1984, the ADPs accounted for 70 it less likely that the private bankers would extend credit to small percent of total use and only 50 percent of total area. Neverthe- farmers. less, some Bank evaluation reports argue that given the existing The BNDS provided two types of loans to the cooperative knowledge of fertilizers among farmers, fertilizer subsidies alone organization: short-term credit (1-2 years) and medium to long- would have increased its use, and that much of the ADP apparatus term credit 12-5 years and up to 10 years). Short-term credit, which was unnecessary or redundant. It is uncertain if fertilizer would constituted an average of 70-75 percent of total agricultural credit, have been as available and accessible without the network of was extended under the common credit guarantee scheme of the Farm Service Centers in the ADPs. See Lele, Oyejide, et al. 1989. cooperative system. State subsidies to cooperatives were paid 76. Idachaba argues that the public sector distribution through through the Rural Mutual Development Fund (FMDR) to decrease Farm Service Centers of ADPs also played a very important role the difference between the factory price of fertilizers and the in improving the quality and reliability of the retail distribution of price to farmers. These subsidies accrued to the state-owned fertilizers. The uniform price of fertilizers avoided the internal fertilizer plant, Societe lndustrielle d'Engrais du Senegal (SIES) (lammeh cross state transportation of fertilizers that had taken place when 1987b). individual states and ADPs fixed their own fertilizer prices (fIdachaba 1987). 71 77. The rate of explicit subsidy in Nigeria was reduced from 83 helped to keep fertilizer prices relatively low in Nigeria, Senegal, percent to 50 percent in 1984; it was further reduced to 34 percent and Cameroon. See Jaeger 1988; World Bank 1987b; IFDC 1986a; in 1985 and to 28 percent in 1986. With the introduction of the lammeh 1987b. second-tier foreign exchange market in October 1986, however, 81. When using purchasing power adjusted exchange rates for the subsidy is once again about 82 percent since fertilizer is being making cross-country comparisons of food costs, it is necessary to imported and priced at the first-tier exchange rate. assume that the crop is a traded good. To the extent that maize 78. Before 1982 more than 90 percent of Malawi's foreign trade cannot be treated as a traded good, these comparisons of prices and all overseas trade moved by rail through the Mozambique are less valid. ports of Beira and Nacala (Alternative Institutional Arrangements 82. Although it is tempting to correlate the nutrient price/crop for the Smallholder Fertilizer Revolving Fund 1987). The war in price ratio with trends in fertilizer use, this ignores the role of Mozambique has, however, caused the deterioration of the crop response coefficients and, therefore, requires restrictive transportation routes from these ports. Malawi has been forced to assumptions about the balance of these coefficients over time resort to the longer routes through Durban and Harare or Lusaka and the comparability of agricultural production functions across and, occasionally, through Dar es Salaam. The closing of the countries. traditional routes (Beira/Nacala) has increased the external 83. All MADIA countries are severely handicapped in formulat- transportation for Malawi considerably. The cost of transporting ing sound policies because of a weak database. Even basic fertilizer in bags between Blantyre and Beira (or Nacala), which are information (such as area, production, and yield by crops and the traditional routes, was $35 per metric ton. The cost of regions/districts, market prices of inputs and outputs by regions/ alternatives is substantially higher. For example, the external districts, and crop response by regions/districts) that is essential transport cost from Durban via Lusaka to the three regional for agricultural policy decisions is amiss. It is unfortunate that not headquarters-Mzuzu, Lilongwe, and Blantyre-is $180 per metric much attention or resources have been given either by the ton, $129 per metric ton, and $145 per metric ton, respectively, national governments or the donors in setting this right. The and the costs from Dar es Salaam (via Mbeya) are, respectively, success of the green revolution in India could partly be attributed $67 per metric ton, $82 per metric ton, and $83 per metric ton to the detailed knowledge of regions at the micro level on the (World Bank 1987a). basis of which production and project planning were undertaken. 79. The comparatively high transport costs in Tanzania, unlike Through departments exclusively set up at the central and state Malawi, are associated with the longer distances between level-Central Statistical Organization, Directorate of Economics producing and consuming or export areas and an inefficient and and Statistics-India periodically releases a detailed breakdown poorly managed state transport network. For example, Staab of data on all aspects of the economy. For fertilizers, the effort of argues that new road construction was directed by neither the Fertilizer Association of India, one of whose primary objec- settlement problems nor the location of agricultural activity: tives is to compile and disseminate international, national, state, Across regions one finds the curious result that, in relation to and district level fertilizer and soil statistics, is noteworthy. The rural population, road densities are relatively low in some of Fertilizer Statistics, published annually, contributes to the knowl- the most populated regions such as Mwanza (1.9 kilometers), edge of sound agricultural practices and is a useful tool for Shinyanga (1.9 kilometers), Kagera (2.1 kilometers), and policymakers, government officials, fertilizer manufacturers, and Dodoma (3.1 kilometers) and relatively high in some of the consumers alike. least populated regions such as Coast (6.7 kilometers), Tabora 84. Kenya has comprehensive and regional data on crop (6.9 kilometers) and Rukwa (5.8 kilometers). Following a responses. Despite this knowledge, the fertilizer recommendation similarly unexpected inverse relationship, some of the made by the government is still a blanket one that does not take highest densities in relation to available agricultural land are into consideration regional variation. The high potential districts found in those regions with relatively little agricultural land in Kenya are Kisii, Siaya, Busia, Kakamega, Nandi, Kericho, Uasin such as Kilimanjaro (22.0 kilometers), Tanga (15.7 kilometers), Gishu, Elgeyo Marakwet, Nyandarau, Kiambu, Nyeri, Muranga, and Coast (12.4 kilometers), while regions with relatively large Kirinyaga, and Meru. For details on classifications of high, agricultural areas have relatively low road densities such as medium, and low potential districts in Kenya, see Appendix 11. Arusha 15.0 kilometers), Shinyanga (7.9 kilometers), and 85. See also Lele 1988, Carr 1988. Mbeya (8.8 kilometers). It would appear, therefore, that 8 S neither the settlement patterns of the population nor 86. Personal communication with Mr. Andrew Spurling. available agricultural land, which are indices of the present 87. It is also possible for crop response to be higher under and potential spatial demand for transport services, have existing practices with the right crop husbandry, than under been used in determining new road construction priorities improved practices if the technical packages in the latter are not (Staab 1982, p. 10). right. 80. The transport cost data for Nigeria is for the postdevalua- 88. The Ziguinchor/Kolda and Tambacounda regions are tion period in 1986. In the predevaluation period, it could have believed to have appreciable untapped agricultural potential. But been as much as four times higher. Cameroon's transportation there is no consensus among agronomists on this issue. A SONED cost (weighted transport costs to all destinations) refers to the study estimated in 1978 that there were nearly 1,265,000 hectares year 1984/85. However, to serve the Extreme North and North of average to good soils in this region, but it included the 813,000 provinces in Cameroon, a combination of rail and road transpor- hectares of land within the Niokolo-Koba National Park, not used tation would cost between $70 and $110 per ton. Senegal's for agriculture or animal husbandry. A more recent study by the transportation cost is for the year 1984 in the Groundnut Basin. French described Eastern Senegal as "appearing in effect as an Malawi has had the highest transportation costs among the immense table of hardpan, notched by a fossilized hydrographic MADIA countries. For Senegal, because of the proximity of the system. The latter is a network of valley bottom lands, where the Groundnut Basin to the markets and a relatively developed bulk of arable land is located" (Abt Associates 1984). Most of the transportation infrastructure, transport costs have been the land in Upper Casamance is of similar type. According to ISRA, lowest. In Tanzania, there has been considerable deterioration of prospects for bringing additional land under cultivation seems the transport infrastructure, and maintenance operations have bleak, as they believe "..... clearing wide areas of the Upper been insufficient in some areas to keep the roadways passable. Casamance would cause considerable loss of the thin layer of The fertilizer subsidy in Tanzania, therefore, effectively covered topsoil that lies over the prevalent hardpan" (Senegal Agricultural the high transportation costs. Nigeria, helped by the oil boom in Policy Analysis, USAID 1984). The scope for extensification is the 1970s, has a well-developed transportation system. It can be therefore limited because the fertile land lying in the river valleys said that the geographical size, location, extent, and quality of is either isolated or already heavily cultivated. transport systems, and the proximity of markets and ports have 72 89. There is a consensus among agronomists that ISRA and 98. IFDC's trials on maize with DAP, 15:15:15, SSP, and partially IFDC represent two different schools of thought, IFDC recommen- acidulated phosphate rock (PA50) in Nigeria showed that dations emphasizing quick returns and profit maximization and response was highly significant for the partially acidulated ISRA recommendations emphasizing environmental concerns and phosphate rock and SSP (both are straight phosphatic fertilizers), impact of fertilizer use on soil fertility in the longer run. whereas no significant response was observed for DAP and 90. In 1980, the IFDC formulated recommendations for ground- 15:15:15. IFDC states, however, that the absence of sulphur in DAP nuts and millet for the Groundnut Basin. For groundnuts, IFDC and 15:15:15 could have resulted in their poor performance (IFDC recommended reduced levels of nitrogen and potassium applica- 1985bl. tions, and for millet, reduced levels of potassium. The ISRA 99. Nigeria's Onne fertilizer plant has been producing and scientists criticized the IFDC recommendations on two major exporting significant amounts of nitrogenous fertilizers since 1987. grounds: Il) the IFDC trials encompassed too short a period (two This policy of promoting more nitrogen aims to bring about a yearsl to be able to generate meaningful results, and (21 the congruence between the types of fertilizer produced in Nigeria recommendations based on economic analysis did not consider and those consumed. the soil as "capital" which would be irredeemably depleted if 100. For a more detailed explanation of why technological IFDC recommendations were adopted. Since 1980, on the basis of issues have been so controversial and have remained unresolved, a variety of unspecified research programs. ISRA recommended see Lele, Oyeiide, et al. 1989. elimination of the "starter" dose of nitrogen on groundnuts. Another important recommendation was the elimination of 101. The impact of transport costs on benefit-cost ratios is potassium for millet on land of lower productivity. The decision to difficult to compare across countries as the data are often recommend the total elimination of nitrogen for groundnuts is unavailable. Some examples from individual countries help to stronger than the position taken by IFDC, which recommended demonstrate the importance of transport costs. In Malawi, if pan- some use of nitrogen in the Thies/Diourbel zone. For more territorial pricing of fertilizer were eliminated, it is estimated that details, see IFDC 1980; Kelly 1988. the benefit-cost ratio would decline by approximately 10 percent for every 200 kilometers that the fertilizer is transported. 91. For instance, maize response to nitrogen in Yaounde is Therefore, in the remote parts of the country the benefit-cost seven times that in Ntui, although both areas are located in the ratio would be 20 to 30 percent lower than ratios given in Table same ecological zone 80-100 kilometers apart. There are two 23. Even this estimate tends to be conservative, because it uses possible explanations for this difference, both of which lead to the haulage rates that are weighted toward the cost of transport over same conclusions for policymakers. The easiest explanation is that paved roads and because it assumes that pan-territorial pricing data are unreliable. Alternatively, it may be that the range of for outputs is maintained. If the latter assumption is relaxed, the variables affecting yield is so great that, even in the same benefit-cost ratio for fertilizer use would decline by another 10 ecological zone, substantial differences may exist. For example, percent for each 200 kilometers. Even with a system of pan- Cameroon primarily distributes A/S, urea, and complex fertilizer territorial pricing, the uniform price only applies at depots. 20:10:10 for food crops; the practice of issuing standard fertilizer Farmers must bear the cost of transport to and from the farm gate, recommendations (60:50:0 for ferralitic soils, 60:100:0 for ferralitic which typically means hiring a bicycle, ox cart, or head loading. In soils with ash, and 60:0:0 on brown and black soils-all dosages Cameroon, because of the rail networks and the lower cost of with an equivalent amount of nitrogen) in all parts of the.country imported fuel and vehicles as compared to Malawi, the benefit- raises serious questions about the economics of fertilizer use and cost ratio decreases by approximately 3 percent for every 100 the impact on soil fertility. kilometers that fertilizer is transported outside the Douala region. 92. Nigeria has three distinct agroclimatic zones: Sudan Savan- 102. We are grateful to M. Agarwal for suggesting this dimension nah-rainfall level: 500-1,000 millimeters (Kano and Sokoto, of the analysis. Bauchi, Borno, Kaduna); North and Southern Guinea Savannah- of Benefis. rainfall level: 1,000-1,500 millimeters (Niger and Plateau, Benue, 103. Benefit-cost ratios have been computed for the year 1987. Gangoia, and Kwara), Rain Forest zone-rainfall level: 1,500-4,000 In the case of Nigeria, the ratios are for both the predevaluation millimeters limo, Lagos, Ogun, Anambra Bendel, Cross Rivers, and postdevaluation years, i.e., 1985 and 1987. As crop response Ondo, Oyom and Rivers), data are primarily available for food crops, with the exception of Ondo . Oyo. and Rivers. . coffee, benefit-cost ratios are computed mainly for food crops. 93. Despite the relatively higher crop responses in the Guinea The output prices are the official producer prices for the East Savannah zone, fertilizer consumption has been heavily concen- African MADIA countries and market prices for the West African trated in the northern states of Kano, Sokoto, Bauchi, Borno, and MADIA countries. With regard to fertilizer cost, the ideal data Kaduna, which are in Sudan Savannah region, where the first would be the unit cost of nutrients in terms of the type actually Bank-supported ADPs were funded. applied, but in the absence of such detail. we have used the unit 94. The SARs for most projects do not explicitly state their cost of nutrients that are predominantly used on a crop or within assumptions about the response of various crops to fertilizers, but a country. When crop responses are in relation to a particular the average response coefficients that are assumed in them can nutrient, the cost of that nutrient alone has been considered. be derived by combining the information on projected yields and More recently, some of the MADIA countries have made major recommended fertilizer application rates. efforts to introduce high analysis fertilizers as a way to reduce the 95. The reasons indicated for crop responses being less than cost per tonne of nutrient. Kenya has been especially successful what the Bank earlier expected were (I) overestimated responses in promoting DAP In the last two years, Malawi has encouraged to fertilizer by the Bank because of the presumption of sole farmers to shift from conventional to high analysis fertilizers. For cropping, whereas in reality farmers had preferred to grow crops this country, benefit-cost ratios using HAF have also been in mixtures; and (2) inefficient compounds, resulting in poor yield computed to understand the extent to which it can alter response. See Lele and Bindlish 1988 for a more detailed profitability in fertilizer use. As a caution it must be remembered discussion of the Bank's assessment on the effects of fertilizer use that unless the existing import policies on fertilizers are revised on crop production in each of the ADPs, and the position taken and the effectiveness of research and extension agencies by the World Bank regarding crop responses and fertilizer use improved, it is likely to take several years for these fertilizers to before and after the mid-term review of the ADPs. be widely accepted. 96. Unlike the FAO/Falusi crop response coefficients, which are 104. Only Tanzania, Senegal, and Cameroon had significant primarily for crops grown solely, the World Bank's recent estimates overvaluation in their currency' in 1987. In Nigeria, the naira was are for crop mixtures and actual field conditions of farmers. overvalued in 1985, but had a relatively less distorted exchange 97. Correspondence between F. S. Idachaba, Head of the rate in 1987. For details on the estimation of implicit and explicit Agricultural Coordinating Unit, Nigeria, and Alan Denness, AF4AG, subsidies, see Appendix 12. on the ADP review mission, dated 24/6/85. 73 105. The grant fertilizers were considered as commercial 112. In Malawi and Kenya there is a relatively widespread imports since TFC reimbursed the C&F value in local currency to distribution of outlets. In Tanzania, the problem of inadequate the Treasury (World Bank 1987). and inefficient fertilizer retail outlets is not new. In the early 1980s 106. Recent studies have estimated the level of explicit or nearly 70 percent of the fertilizer distributed went to five regions, implicit subsidy in Tanzania to range between 60-66 percent. See three of which had no retail outlets. But regional consumption was Rioseco 1989; Carr 1989. closely linked with agricultural potential and fertilizer-responsive 107. See for instance, Carr 1989. areas, suggesting that despite the poor distribution system, 107. See for instance, Carr 1989. ~fertilizer had been getting through to consumers, but inefficiently. 108. Benefit-cost ratios for sorghum when computed even for With the reemergence of cooperatives since 1984, there are 23 maximum market prices are less than 2 for these regions. cooperative unions operating at the regional level, accounting for 109. A recent study also highlights that fertilizers would be 73 percent of the distribution in 1986/87. The primary societies uneconomical in most regions if unsubsidized. The benefit-cost (numbering over 2,000) are expected to perform the role of ratio for groundnut (computed for an entire crop rotation and subretailers. Without information on the primary societies that are accounting for economic returns from groundnut hay) based on AF really active in the distribution business, it is premature to data was greater than 2 in Nioro for subsidized and unsubsidized conclude that the distribution network in Tanzania is preparing to prices. But for Boulel it was in the range of 1.36-1.73 (exclusive of meet the fertilizer needs in all parts of the country. The groundnut hay) for 1987 unsubsidized prices. Reviewing the Agriculture Products Distribution Centers ICEPA), where the results with ISRA agronomists, the author of the study observes majority of fertilizer sales in Senegal take place, are located as far ".. .we found general agreement that the estimated yields for all as 20-25 kilometers from farms. Fertilizers often have to be treatments (of fertilizers) in Nioro were higher than generally transported nearly 120 kilometers (because the shorter route obtained by farmers. There was a general acceptance of the becomes inaccessible by truck after the first rains) in Senegal. In Boulel results, which many consider to be a reflection of recent the Nigerian ADPs, the average distance to a Farm Service Center declines in agricultural productivity due to lower rainfall and is less than 15 kilometers, but on a national level retailers are declining soil fertility." The author also concludes that " . . .current both inadequate and concentrated in certain regions, making analysis of AF data suggests that value/cost ratios are now less farmers travel nearly 23 kilometers to buy fertilizers. than 2 in zones that were previously thought to exhibit profitable In India, to deal with the problem of concentration of retail responses." For more details, see Kelly 1988. outlets in specific regions, the government has emphasized the 110. Senegal imports a significant amount of rice. In a bid to need for increased fertilizer distribution points in the remote encourage rice production within the country, the government rural areas by way of a transport subsidy. According to a recent recently increased the producer price of rice. Even at the import policy of the Indian government, all fertilizers are to be delivered parity price of rice, the benefit-cost ratios are unfavorable without by the manufacturers to the retail channels, with the freight paid an explicit or implicit subsidy. to the block headquarters. The Block Delivery Scheme, as it is Ill. The price changes from year to year and within a year are commonly known, provides a subsidy or secondary transport considerable for certain crops. For instance, the ratio of harvest to rebate for transporting the fertilizers to the block headquarters postharvest price in 1987 was as follows: maize, 0.68; millet, 0.80; from the primary supply point of the manufacturers, the rakeload rice. 0.88; and groundnuts, 0.57. Similarly, the postharvest price for point, or the warehouse. the year 1986 to the year 1985 and for the year 1987 to the prices 113. Desai, drawing on Asian experience, notes that private for the year 1986 were as follows: sector involvement in input supply is concentrated where turnover 1986/85 1987/86 is high. Using evidence from a few Sub-Saharan countries, he confirms that the distribution system in the private sector has Maize 0.62 0.92 effectively catered only to large commercial farms. Desai indicates Millet 0.46 0.99 from the Indian experience that the number of private sector oghum 0.96 0.60 outlets grows in years of tight availability and shrinks in years of Groundnuts i0.3 1.25 easy availability. See Desai 1987. Bibliography Carr, S. 1988. "Modification and Extension of the National Rural Development Program." Paper presented at Symposium on Abt Associates. 1984. "Senegal Agricultural Policy Analysis." Report Agricultural Policies for Growth and Development, Mango- prepared for USAID. chi, Malawi, October. ADMARC. 1986. "Annual Accounts and Report," Malawi, 1986. . 1989. "The Role of Fertilizer in Maize Production in Agriconsult. 1988. "Fertilizer Marketing Development Programs Southern Tanzania." AFTAG. World Bank. Impact Study." Christiansen, R.E. and L.A. Stackhouse. 1989. "The Privatization of Ahmad, R. 1988. "Fertilizer Use in Asia: Lessons from Selected Agricultural Marketing in Malawi." World Development 17 (May). Country Experiences." Paper presented for IFPRI/IFDC and VR. Southworth. 1988. "Agricultural Pricing and Market- Workshop on Fertilizer Policy in Tropical Africa, Lome, Togo, ing Policy in Malawi: Implications for a Development April 5-7, 1988. Strategy." MADIA Working Paper. Barker, R. and Y Hayami. 1976. "Price Support Versus Input Cleaver, K. and M. Westlake. 1987. "Pricing, Marketing and Subsidy for Food Self-Sufficiency in Developing Countries." Agricultural Development in Kenya." MADIA Working Paper. American Journal of Agricultural Economics 58 (November)., Crawford, E. and V. Kelly. 1987. "A Field Study of Fertilizer Berg Associates, Elliot. 1983. "Agricultural Input Supply in Camer- Distribution and Use in Senegal, 1984: Summary Report." oon." Report prepared for the World Bank, Washington, D.C. MSU International Development Paper, no. 8. East Lansing: Berg, E. 1989. "The Liberalization of Rice Marketing in Madagas- Michigan State University. car." World Development 17 (May). Delgado, C. and C. Ranade. 1987. "Technological Change and Bloch, P. 1985. "Wage Policy, Wage Structure and Employment in Agricultural Labor Use." In Accelerating Food Production in Sub- the Public Sector of Senegal." World Bank CPD Discussion Saharan Africa. Ed. Mellor, Delgado, and Blackie. Baltimore: Paper no. 1985-41. Johns Hopkins University Press. Bumb, B. 1988. "Fertilizer Supply in Sub-Saharan Africa: An Desai, G. M. 1987. "Commentaries on Marketing System." In Analysis." Paper presented at IFDC-IFPRI Workshop on Accelerating Food Production in Sub-Saharan Africa. Ed. Mellor, Fertilizer Policy in Tropical Africa, Lome, Togo, April 5-7, Delgado, and Blackie. Baltimore: Johns Hopkins University 74 1988. Press. _ . 1988. "Development of Fertilizer Markets in South Asia- _ . 1980. "Response of Groundnut and Millet to NPK in the A Comparative Overview." Paper presented at the 1988 IFA Groundnut Basin of Senegal." Muscle Shoals, Alabama: loint Middle East/South Asia Fertilizer Conference. Lahore, IFDC. Pakistan. IFPRI/IFDC. 1988. "Proceedings of the Workshop on Fertilizer . 1988. "Price Environment, Government Policies and Growth Policy in Tropical Africa." Tshikala Tshibaka. Of Fertilizer Use in india." Paper prepared for the 1988 laeger, W 1988. "Agricultural Incentives and Policy in Sub-Saharan Annual Seminar of the Fertilizer Association of India, New Africa: Some Recent Trends and Methodological Issues." Delhi. ~~~~~~~~~~~~~~~Washington, D.C.: World Bank. . 1982. "Sustaining Rapid Growth in India's Fertilizer: A Washintn D.C.: Wor ank. Persectie Baed o Comositon o Use" Wahingon, Jammeh, S. 1987a. "Distribution and Consumption of Agricultural Perspective Based on Composition of Use." Washington, Inputs." MADIA Working Paper. D.C.: IFPRL. Eicher, C. K. 1982. "Facing Up to. Africa's Food Crisis." Foreign Affairs1986. "The Evolution of Marketing and Pricing Policy in Eicher, C. K. 1982. ";Facing Up to Afrca's Food Crsis. Foreign Afairs Senegal. MADIA Working Paper. Falusi, A.O. 1987. "Economic and Technical Efficiency of Fertilizer . 1987b i "State Intervention in Agricultural Pricing and Use in Food Production in Nigeria." Paper presented at the Marketing in Senegal." Ph.D. dissertation, tohns Hopkins National Fertilizer Seminar, Port Harcourt. _____ and C.G. Ranade. 1987. "Agricultural Development and FAO. 1974. FAO Fertilizer Programme, The First Decade. Zurich Centre Foreign Assistance: A Review of the World Bank's Experi- d'Etude de l'Azote; London, International Superphosphate ence in Senegal 1969-86." MADIA Working Paper. and Compound Manufacturers' Association Ltd; and Berne _ and U. Lele. 1988. "Building Agricultural Research Capacity International Potash Institute.inSegl"MDAWrngPp. - . 1986. Fertilizer Year Book 1986. Rome: FAO. Kelly, V A. 1988. "Farmers' Demand for Fertilizer in the Context of /FADINAP. 1987. "Fertilizer Marketing Costs and Margins in Senegal's New Agricultural Policy: A Study of Factors Developing Countries 1985/86." Rome: FAO. Influencing Farmers' Fertilizer Purchasing Decisions." MSU FAO and World Bank. 1987. "Tanzania-Input Supply and Credit International Development Paper, no. 19. East Lansing: Pre-Identification Mission." Working Papers I and 2. Michigan State University. Geib, Alan and Associates. 1988. Oil Windfalls: Blessing or Curse. A Khan, 0. and Harry Palmier. 1989. "An Interim Report on National World Bank Research Publication. New York: Oxford Univer- Agricultural Research Systems in Five African Countries." sity Press. Kydd. 1. G. 1989. "Building Agricultural Research Capacity: The Githongo and Associates. 1983. "Grain Marketing Study." Interim Case of Malawi." MADIA Working Paper. Report, Volume 1, Nairobi. Kydd, 1. G. and R. E. Christiansen. 1982. "Structural Change in Grilli, R. E. and Maw Chung Yang. 1988. "Primary Commodity Malawi Since Independence: Consequences of a Develop- Prices, Manufactured Goods Prices, and the Terms of Trade ment Strategy Based on Large-Scale Agriculture." World of Developing Countries: What the Long Run Shows." World Development 10: 355-376. Bank Economic Review (January). Lele, U. 1988. "Agricultural Growth, Domestic Policy, and External Government of Cameroon. 1984. Agriculture Census 1984. Ministry of Assistance to Africa: Lessons of a Quarter Century." In Trade, Agriculture. Aid, and Policy Reform: Proceedings of the Eighth Agriculture Sector Government of Malawi. Economic Report 1988. Office of the Presi- Symposium. Ed. Colleen Roberts. Washington, D.C.: World dent and Cabinet, Zomba, Malawi. Bank. Government of Senegal. 1985. "Programme de Sauvegarde du _ . 1987. "Aid Effectiveness in Africa: Some Implications for the Coton." Rural Development Agency for Cotton (SODEFITEXI. Bank's Reorganization." MADIA Working Paper. Haggblade, S., PB. Hazell, and I. Brown. 1987. "Farm/Nonfarm _ . 1979. The Design of Rural Development: Lessons from Africa. Linkages in Rural Sub-Saharan Africa: Empirical Evidence Baltimore: Johns Hopkins University Press. and Policy Implications." AGRAP Economic Discussion . 1989. "Sources of Growth in East African Agriculture." WoHld Paper, no. 2. Washington, D.C.: World Bank. Bank Economic Review 3 (January): 119-144. , and PB. Hazell. 1988. "Prospects for Equitable Growth in and L.R. Meyers. 1986. "Agricultural Development and Rural Sub-Saharan Africa." AGRAP Economic Discussion Foreign Assistance: A Review of the World Bank's Experi- Paper, no. 3. Washington D.C.: World Bank. ence in Kenya, 1963-1986." MADIA Working Paper. Hedley, Douglas and Steven R. Tabor. 1989. "Fertilizer in Indone- _ , M. Gbetibouo, and N. van de Walle. Forthcoming. "Agricul- sian Agriculture: The Subsidy Issue." Agricultural Economics tural Development in Cameroon: A Review of the World (March). Bank's Experience 1967-1987." MADIA Working Paper. Higgins, G.M. 1982. "Potential Population Supporting Capacities of - N. van de Walle, and M. Gbetibouo. 1989. "Cotton in Africa: Lands in the Development World: Land Resources for An Analysis of Differences in Performance." MADIA Working Populations of the Future: A Technical Report of the Paper. Project." FAO and International Institute for Applied Sys- _ and A Goldsmith. 1989. "The Development of National tems Analysis. Agricultural Research Capacity: India's Experience with the Idachaba, F.S. 1987. "State-Federal Relations in Nigerian Agricul- Rockefeller Foundation and Its Significance for Africa." ture." MADIA Working Paper. Economic Development and Cultural Change 37 (lanuary): 305-343. IFDC. 1988. "An Assessment of Fertilizer Demand and the _ and L.R. Meyers. 1987. "Growth and Structural Change in Efficiency of the Fertilizer Distribution System in Nigeria." East Africa: Domestic Policies, Agricultural Performance, and Muscle Shoals, Alabama: IFDC. World Bank Assistance, 1963-1986, Parts I and If." DRD i__._ 1986a. "Cameroon Fertilizer Sector Study." Muscle Shoals, Discussion Paper, nos. 273 and 274. Washington, D.C.: World Alabama: IFDC. Bank. _ . 1985a. "A Commercial Fertilizer Marketing System for _ and V. Bindlish. 1988. "How Important are the Relative Nigeria." Muscle Shoals, Alabama: IFDC. Effects of Economy-Wide and Sector-Specific Policies in . 1986b. "Fertilizer Marketing and Economics of Use in Explaining the Past Performance of Nigerian Agriculture?" Kenya." Muscle Shoals, Alabama: IFDC. MADIA Working Paper. . 1985b. "Fertilizer Research Program for Africa." Muscle and R. E. Christiansen. 1988. "Markets, Marketing Boards, Shoals, Alabama: IFDC. and Cooperatives: Issues in Adjustment Policy." Paper . 1981. "Nigeria Fertilizer Sector: Present Situation and presented at the conference on "Agricultural Growth and Future Prospect." Muscle Shoals, Alabama: IFDC. Market Town Development," Lilongwe, Malawi. 75 -___* AT. Oyeiide, V. Bindlish, and B. Bumb. 1989. "Nigeria's Siamwallah, A. 1981. "Security of Rice Supplies in the ASEAN Economic Development, Agriculture's Role, and World Bank Region." In Food Security for Developing Countries. Ed. Alberto Assistance, 1961-1988: Lessons for the Future." MADIA Valdes. Westview Press: Boulder, CO. Working Paper. Spurling, Andrew. 1982. "Problems of Fertilizer Distribution." _____ and S. Stone. 1989. "Population Pressure, the Environment, Background Paper no. X, Tanzania Agricultural Sector and Agricultural Intensification: Variations on the Boserup Report, World Bank. Hypothesis." MADIA Working Paper. Srinivasan, TN. 1986. "Fertilizer Pricing in Developing Countries: A _____ and luan Gaviria. 1989. "Rural Road Transport and Agricul- Market Based Approach" In Fertilizer Producer Pricing in tural Development in Malawi." joint MADIA/SSATP Study. Developing Countries. Ed. E. L. Segura, Y T. Shatty, and M. _____ and J. W Mellor. 1981. "Technological Change, Distributive Nishimizu. Washington, D.C.: World Bank. Bias and Labor Transfer in a Two Sector Economy." Oxford Staab, M. 1982. "Transport and Agriculture." Background Paper, Economic Papers 33 (November): 426-441. Tanzania Agricultural Sector Report, World Bank. Mellor, 1. 1985. "Determinants of Rural Poverty: The Dynamics of Staatz J., l. Dione, and N. Nango Dembele. 1989. "Cereals Market Production, Technology and Price." In Agricultural Change and Liberalization in Mali." World Development 17 (May). Rural Poverty. Ed. Mellor and Desai. Baltimore: lohns Stanley, M. K. 1987. "An Evaluation of Tea Productivity on Hopkins University Press. Smaliholdings in Kenya." M. Phil dissertation, Centre for Mhella. 1985. "Effects of Agricultural Policy Alternatives with Development Studies, University of Glasgow. Emphasis on Fertilizer Pricing in Tanzania." Dar es Salaam: Timmer, P 1974. "The Demand for Fertilizer in Developing Maireting Development Bureau. Countries." Food Research Institute, Stanford University. Ministry of Agriculture, Kenya, in cooperation with the German Tisminieszky, B. and P Kimuyi. 1986. "Fertilizer Infrastructure Agricultural Team (GAT) of the German Agency for Technical Improvement Support." Nairobi, Kenya: Chemical Engineer- Cooperation (GTZI. 1983. "Farm Management Hand Book." ing Consultants. Ministry of Agriculture, Kenya. 1987. "National Policy for Fertilizer ing Cosutants. Piigand Marketing." Tourte, R. et a]. 1971. -Thiemes Legers-Thi6mes Lourcds-Systl~mes Pricing and Mareting."Intensifs, Voies Differentes Ouvertes au D6veloppement Ministry of Agriculture, Malawi. 1987. "MOA Report no. 2." Agricole du Senegal." Agronomie Tropicale, Vol. XXVI. Nathan Associates, R.R. 1987. "The Impact of the Fertilizer Subsidy Twyford, IT 1988. "Development of Smallholder Fertilizer Use in Removal Program on Smallholder Agriculture in Malawi." Malawi." Paper presented at FAO/FIAC meeting, Rome. Report presented to Ministry of Agriculture, MalawU and USAID/Cameroon, Africa Economic Policy Reform Program. 1987. USAID. "Fertilizer Subsector Reforms Program." Washington, D.C.: Nigerian Central Bank. Various years. "Annual Report." USAID. O'Brien, Donol Croise. 1975. Saints and Politicians: Essays in the USAID/Senegal. 1983. "Program Assistance Approval Document Organization of a Senegalese Peasant Society. Cambridge: Cam- Supplement-Sahel Development Fund." Washington, D.C.: bridge University Press. USAID. OECD. Various Issues. "Geographical Distribution of Financial Van de Walle, N. 1989. "Privatization in Developing Countries: A Flows." Review of the Issues." World Development 17 (May). Pieri, C. 1984. "Conduite De La Fertilisation Des Cultures Vivrieres Vernon-Wortzel H. and L. Wortzel. 1989. "Privatization: Not the En Zones Semi-Arides" (Fertilization of Food Crops in Only Consumer." World Development 17 (May). Semiarid Zones). Trans. S. Jammeh. IRAT/CIRAD unpub- Waterbury, John. 1989. "Agricultural Policymaking and Stagnation in lished working paper. Senegal." MADIA Working Paper. Pieri, C. 1989. Fertilite des terres de savanes. Paris: Ministere de la Senegal. MAD8 WoringtPaper. Coop6atio et d 136elopprnen et CRAD-RAT.World Bank. 1988. "Agricultural Growth in Nigeria: Sources and Cooperation et du Developpement et CIRAD-IRAT. Recommended Actions." : Repetto, R. 1988. "Economic Policy Reform for Natural Resource _ _. 1979. Geita Cotton Project, Supervision Mission Report. Conservation." Environment Department Working Paper no. 4. Washington, D.C.: World Bank. . 1986a. Internal Memorandum. 117 July 1986). Republic of Kenya. 1985. Statistical Abstract. Central Bureau of . 1985. "Kano Agricultural Development Project." Statistics, Kenya. , Eastern and Southern Africa Projects Department. 1986b. Rioseco, G. 1989. "Input Supply and Distribution-Overview of "Kenya: Agricultural Inputs Review." Major Issues." AFTAG, World Bank. . 1988a. "Malawi-Public Expenditure Review." Volume 1, Samki, J.K. et al. 1982. "Fertilizer Recommendations Related to Executive Summary. Ecological Zones in Tanzania." Milingano: National Soil . 1983b. "Malawi, Smaliholder Fertilizer Project." Service Agricultural Research Institute. . 1987a. "Study of Alternative Institutional Arrangements for Schluter, M. 1985. "Size and Deployment of Public Sector the Smallholder Fertilizer Revolving Fund." Manpower in Agriculture, 1974-75 to 1981-82." , Eastern Africa Projects Department, Southern Agriculture Schluter, M. and M. A Sackett. 1982. "Estimates of 1981/82 Import Division. 1983a. "Tanzania: Agricultural Sector Report." Requirements for the Production, Processing and Marketing _. 1986c. "Tanzania-Multi-Sector Rehabilitation Credit." of Major Crops in Mainland Tanzania." Background Paper _ /OED. 1985. "Tenth Annual Review of Project Performance no. XIV, Tanzania Agricultural Sector Report. Audit Results 1984." Schmacher, E. J. 1975. Politics, Bureaucracy and Rural Development in _ . 1985. World Commodity Trade and Price Trends. Baltimore: Johns Senegal. Berkeley: University of California Press. Hopkins University Press. Seka and Fishstein. 1988. "Real Effective Exchange Rate Index for _ . 1987c. World Commodity Trade and Price Trends. Baltimore: the MADIA Countries." Unpublished MADIA Worksheet. Johns Hopkins University Press. Seth, A. K. 1988. "Nigeria: Fertilizer Sector Loan--Revised." EPS. _ . 1986a. World Development Report, 1986. New York: Oxford Shalit, H. and H. Binswanger. 1985. "Is There a Theoretical Case for University Press. Fertilizer Subsidies?" Agriculture and Rural Development . 1987b. World Development Report, 1987. New York: Oxford Department. Washington, D.C.: World Bank. University Press. Shepherd, 1988. "Approaches to Privatization of Fertilizer Market- . 1988b. World Development Report, 1988. New York: Oxford ing in Developing Countries." Paper presented at FAO/FIAC University Press. 76 meeting, Rome. THE MADIA STUDY Although many generalizations have been made about the agricultural crisis in Africa, relatively few detailed country and cross-country studies of African agriculture based on systematic data analysis have been conducted. Similarly, although foreign aid has constituted a large part of total government expenditures in Africa for close to fifteen years, there has been little analysis of the role of external assistance in African countries that goes beyond political criticism of official assistance or the alleged self- serving objectives of donors. The impetus for the study "Managing Agricultural Development in Africa" (MADIAI was to begin the process of filling this gap and to explain the nature and sources of the agricultural crisis, particularly the extent to which it originated in resource endow- ments, historical and contemporary events, external and internal policies, and the economic and political environment. The MADIA study involved detailed analysis of six African countries- Kenya, Malawi, Tanzania, Cameroon, Nigeria, and Senegal. In addition to the World Bank, seven donors, USAID, UKODA, DANIDA, SIDA, the French and German governments, and the EEC participated in the study. The analysis of country policies and performance during the last 20-25 years was carried out with the benefit of substantial input from the governments and nationals of each of the countries represented. The study had three main areas of focus: (I) the relationship between domestic macroeconomic and agricultural policy and agricultural performance, 12) donors' role in the development of agriculture, and (3) the politics of agricultural policy. The MADIA study was the result of encouragement and support from many people. Anne Krueger, former Vice President for Economic Research Staff in the World Bank, encouraged the establishment of these studies on aid and development in 1984. Gregory Ingram, former Director of the Development Research Department, provided unstinting support for the study. During the reorganization of the World Bank in 1986, the strong support from Benjamin King, then acting Vice President for Economic Research Staff, proved invaluable. Barber Conable, President of the World Bank, and Mr. Edward V. K. Jaycox, Vice President for the Africa Region, have played a key role by ensuring support for the study's completion, as did Stanley Fischer, the Vice President for Development Economics. Yves Rovani, Director General of the Operations Evaluation Department, was particularly helpful as the MADIA study drew heavily on the works of OED. A special debt of gratitude is owed to the World Bank's Research Committee, which provided the initial funding for the study, and to the MADIA Steering Committee. In particular the strong support of the chair of the Steering Committee, Stephen O'Brien, has been of critical importance. Finally, without the active and continued encouragement of many African policymakers and donor officials, including numerous colleagues in the World Bank, this study would not have provided new perspectives. This support has taken the form of numerous reactions to written and oral presentations, and refinement of the analysis to identify the areas of consensus and continuing controversy. r- (D 2 The World Bank Headquarters 1818 H Street, N.W. Washington, D.C. 20433, U.S.A. Telephone: (202) 477-1234 D Facsimile: (202) 477-6391 rD Telex: WUI 64145 WORLDBANK RCA 248423 WORLDBK Cable Address: INTBAFRAD WASHINGTONDC > European Office 66, avenue d'1ena 75116 Paris, France Telephone: (1) 40.69.30.00 Facsimile: (1) 47.20.1 9.66 Telex: 842-620628 Tokyo Office CD Kokusai Building O 1-1, Marunouchi 3-chome Chiyoda-ku, Tokyo 100, Japan Telephone: (3) 214-5001 Facsimile: (3) 214-3657 > Telex: 781-26838 CD 0 0-8213-1321-5