56129 Distortions to Agricultural Incentives in Sub-Saharan and North Africa Kym Anderson and William A. Masters University of Adelaide kym.anderson@adelaide.edu.au Purdue University wmasters@purdue.edu Agricultural Distortions Working Paper 64, September 2008 This is a product of a research project on Distortions to Agricultural Incentives, under the leadership of Kym Anderson of the World Bank's Development Research Group. The authors are grateful for the distortions estimates provided by authors of the focus country case studies, for assistance with spreadsheets by Johanna Croser, Esteban Jara, Marianne Kurzweil, Signe Nelgen, Francesca de Nicola, Damiano Sandri and Ernesto Valenzuela, and for funding from World Bank Trust Funds provided by the governments of Ireland, the Netherlands (BNPP) and the United Kingdom (DfID). This paper draws on the chapters in Distortions to Agricultural Incentives in Africa, edited by K. Anderson and W. Masters, Washington DC: World Bank (forthcoming 2009). A revised version without the Appendix will appear as Ch. 8 in Distortions to Agricultural Incentives: A Global Perspective, 1955 to 2007, edited by K. Anderson, London: Palgrave Macmillan and Washington DC: World Bank (forthcoming 2009). This is part of a Working Paper series (see www.worldbank.org/agdistortions) that is designed to promptly disseminate the findings of work in progress for comment before they are finalized. The views expressed are the authors' alone and not necessarily those of the World Bank and its Executive Directors, nor the countries they represent, nor of the institutions providing funds for this research project. Distortions to Agricultural Incentives in Sub-Saharan and North Africa Kym Anderson and William A. Masters1 In the 1960s and 1970s, many African governments had macroeconomic, sectoral and trade policies that increasingly favored urban employees at the expense of farm households, and favored the production of importable goods at the expense of exportables (Krueger, Schiff and Valdes 1988, 1991). Similar biases were also prevalent elsewhere, but rarely to the same extent as in Africa. The magnitude of pro-urban (anti-agricultural) and also pro-self- sufficiency (anti-trade) intervention matters greatly for economic development, because agriculture is the main employer for the poor and is often a key export sector. Changes in these biases could help explain Africa's development experience, including the continent's slow pace of poverty alleviation and economic growth especially in the 1970s and 1980s, and its subsequent recovery since then. Much progress has been made in recent years to reduce the anti-agricultural and anti- trade biases of policy in Africa, and these changes have been associated with faster economic growth and poverty alleviation. Many price distortions remain, however, and with 60 percent of Sub-Saharan Africa's workforce still employed in agriculture and more than 80 percent of the region's poorest households depending directly or indirectly on farming for their livelihoods (World Bank 2007, Chen and Ravallion 2007), agricultural and trade policies are still key influences on the pace and direction of change in Africa. This chapter summarizes a set of case studies measuring distortions within and across countries over time. We make no attempt to summarize the voluminous literature on policy and economic growth in Africa, the most recent major continental study being Ndulu et al. (2008). This chapter also makes no attempt to summarize the literature dealing with public investment or economic growth strategies more broadly, which was addressed recently by Spence et al. (2008). Our goal is more narrowly defined, simply to compare quantitative indicators of past and recent agricultural price policies. 1 This chapter draws on the introductory and country chapters in Anderson and Masters (2008), with data updated using Anderson and Valenzuela (2008). 2 Including Africa in this global study is crucial for several reasons. First, the continent is home to many of the world's poorest people. In 2006 Sub-Saharan Africa accounted for less than 2 percent of global gross domestic product (GDP) and exports and just 4 percent of agricultural GDP, but it also accounted for 12 percent of the world's farmers, 16 percent of agricultural land, and 28 percent of those living on less than US$1 a day (World Bank 2008). Second, it is the region where output and income growth has been slowest over the past half- century, especially on a per capita basis. And third, it is where sectoral and macro (including exchange rate) policies have been among the most heavily interventionist, dampening the contribution of market incentives to growth. There is thus much to be learned from examining the policy history of the region, and there is great potential for poverty alleviation if market-friendly, growth-enhancing policies were to be adopted and the recent large increase in development assistance funds were to be used wisely to complement and strengthen market forces. The African part of this study is based on a sample of 21 countries. It includes Egypt, the largest and poorest country in north Africa, plus five countries of eastern Africa (Ethiopia, Kenya, Sudan, Tanzania and Uganda), five countries in southern Africa (Madagascar, Mozambique, South Africa, Zambia and Zimbabwe), five large economies in west Africa (Cameroon, Cote d'Ivoire, Ghana, Nigeria and Senegal), and five smaller economies of west and central Africa for which cotton is a crucial export (Benin, Burkina Faso, Chad, Mali and Togo, for which we estimate price distortions for just cotton and four nontraded food staples). In 2000­04 these economies (leaving aside Egypt) together accounted for around 90 percent of the agricultural value added, farm households, total population and total GDP of Sub- Saharan Africa. Estimates of distortions are provided for as many years and products as data permit, amounting to an average of 43 years and 9 crop or livestock products per country. The covered products account for more than two-thirds of the value of most countries' agricultural production. Our 21 focus economies in Africa accounted for only 1.3 percent of worldwide GDP but 11 percent of the world's farmers in 2000-04. These and related shares are detailed in table 1, which reveals the considerable diversity within the region in terms of stages of economic development, resource endowments, trade specialization, poverty incidence and income inequality. The countries are also very diverse in political and social development terms, and thus and thus offer important opportunities for comparative study. Our averages all include South Africa, whose per-capita national income is more than four times larger 3 than the other focus countries, but whose income inequality is among the highest in the world. The extent of poverty decline in Sub-Saharan Africa (SSA) has been disappointing relative to other developing country regions. Over the 1981-2004 period, the number of SSA people living on less than $1/day (in 1993 PPP terms) grew from 168 million to 298 million. As a percent of the population, the number of people in such extreme poverty rose to 47 percent in 1990, then stabilized and eventually declined to 41 percent by 2004, marginally below the 42 percent level of 1981 (Appendix table 1). More than two-thirds of that decline in poverty incidence over the past decade or so has been in rural areas, while most of the rest is explained by the rural poor moving to urban centers where their incomes may rise above the dollar-a-day threshold but many remain very poor. The African experience contrasts strongly with that of Asia, where even in South Asia the proportion of the population living on less than $1 a day has fallen from one-half to less than one-third (Chen and Ravallion 2007). Policy choices have played an important role in observed rates of economic growth, structural change and poverty alleviation. Many African countries had increasingly severe anti-agricultural and anti-trade biases in the 1960s and 1970s, contributing to farmers' poverty especially in the 1970s. Subsequent reforms varied widely in terms of starting date, speed and extent of policy change. The switch to policies that are less biased against farmers and trade began in some countries by the late 1970s but in many others only in the 1980s or even later ­ and the transition is still on-going, often with periods of stalling and even reversals, the most notable recent example being Zimbabwe. Agricultural price distortions are not the only target of policy reform of course, but they are a key aspect of economic policy in most African countries. This chapter begins with a brief summary of economic growth and structural changes in the region since the 1950s and of agricultural and other economic policy developments as they affected the farm sector at the time of and in various stages after independence from colonial powers. The chapter then summarizes estimates of the nominal rate of assistance (NRA) and the relative rate of assistance (RRA) to farmers delivered by national farm and nonfarm policies over the past several decades, as well as the impact of these policies on the consumer prices of farm products, using the project's methodology (Anderson et al. 2008). The final sections point to what we have learned and draw out implications of the findings, including for poverty and inequality and for possible future directions of policies affecting agricultural incentives in Africa. 4 Growth and Structural Change2 Between 1980 and 2004, per capita GDP for our 21 focus countries of Africa grew at just 0.7 percent per year (Appendix table 2). This was half the global average of 1.4 percent and a small fraction of Asia's 5.5 percent, so per capita incomes in Africa have fallen well below the income levels of other countries, especially those in Asia. The difference is due mainly to non-farm growth, since agricultural GDP growth per capita was about the same in Africa as in other regions (0.6 percent in Africa compared with 0.5 percent for the world as a whole). The aggregate, long-term experience hides large variation over time and across countries. Most notably, Africa experienced a sharp decline in agricultural output per capita from the early 1970s through the late 1980s, followed soon thereafter by a decline in total national income per capita, both of which were then stabilized and reversed in the 1990s. Most recently, during the 2000-06 period, per capita GDP growth averaged 4.7 percent in Sub-Saharan Africa compared with 3.0 percent for the world as a whole (World Bank 2007, page 341). Trends in GDP are closely linked to changes in Africa's export volumes. These grew at relatively slow rates compared with the global average of 6.1 percent (last column of Appendix table 2), causing the region's share of global exports to halve. However, as economies have gradually opened up, the share of exports in GDP has reversed its decline and begun rising in several African countries. African economies are slowly recovering from their decline during the 1970s and 1980s, but only a few countries have achieved substantial restructuring away from agriculture and towards other activities. In fact, about one-quarter of our focus countries have seen their share of agriculture in GDP actually increase over the entire 1965-69 to 2000-04 period (Appendix table 4). Agriculture's share of GDP is above 25 percent in nearly three-quarters of our focus countries, and is above 40 percent in Cameroon, Chad and Ethiopia. The share of overall employment accounted for by farming activities has fallen in all focus countries but generally remains above 50 percent (Appendix table 5), which is much higher than farmers' 2 The economic indicators quoted in this section are from the first ten tables in the Appendix, based predominately on the compilation of data from the World Bank's World Development Indicators and the UN's FAOSTAT databases by Sandri, Valenzuela and Anderson (2007). 5 share of GDP. These data underscore the relatively low incomes of farm households, and hence the continued importance of agricultural prices for social welfare. Agriculture is particularly important as a source of exports, accounting for over 70 percent of merchandize exports in Benin, Burkina Faso, Ethiopia, Tanzania and Uganda during the 2000-04 period (Appendix table 6). Agriculture's share of merchandise exports has actually risen in three of our focus African countries (Benin, Zimbabwe and Zambia), and has declined elsewhere partly because of rises in other primary exports such as petroleum in Sudan, and partly because of growth in exports of manufactured goods as for example in Kenya, Madagascar and Senegal. Such nonfarm exports have grown even faster in other regions, however, so the index of revealed agricultural comparative advantage (defined as the share of agriculture and processed food in national exports as a ratio of the share of such products in worldwide merchandise exports) has risen in most of our focus countries (Appendix table 7). The exceptions are Nigeria and Sudan, which have newly exploited mineral or energy deposits. While most African countries have an increasing level of revealed comparative advantage in agricultural exports, there is also rising domestic demand for farm output. During colonial times, production was heavily export oriented. At the start of the independence period in 1961-64, the total value of agricultural output was about 120 percent of consumption, and that ratio has since declined to around 105 percent. The share of farm production that is exported has fallen from nearly 20 percent to just 8 percent, and the share of imports in domestic consumption of farm products has doubled, from 2 to 4 percent (Appendix table 8). The Evolution of Agricultural Trade Policies The trends in growth and development described above are closely linked to economic policies pursued by African governments. Before independence, most of Africa had been ruled since the 19th century by foreign powers whose explicit objective was to control trade, for political reasons and to extract revenue. Interventions were typically managed through licensed monopolies and marketing boards, as well as restrictions on Africans' labor mobility, property ownership and market participation. The few countries not ruled by a foreign power were controlled by a local aristocracy or immigrant minority whose economic 6 policies were similarly repressive, as in Ethiopia, South Africa and Zimbabwe among our focus countries. Majority rule came to Africa much later than any other major region of the world, and it arrived in the 1960s at a time when central planning was widely seen as a promising strategy for economic development. The newly independent, elected governments typically kept the marketing boards and other instruments for intervention that had been developed by previous administrations, and simply expanded their mandate to cover more people and larger regions of the country. Their stated goals were to be more inclusive and serve a larger fraction of the African population than the exclusive licensing and limited mandates of colonial institutions. The new governments also adopted new criteria for public employment, with staffing priorities that reflected electoral politics instead of colonial interests. Both changes led to large increases in the public payroll and fiscal expenditure. These steps were often underwritten by foreign donors, including the former colonial powers plus other industrialized countries and oil exporters. Project aid and budget support grew rapidly, especially in the 1970s when loans were available at zero or negative real interest rates. These capital inflows covered growing fiscal deficits, current-account imbalances and increasingly overvalued foreign exchange rates. Inflation was usually kept low, as governments chose to ration credit and foreign exchange rather than expand the money supply, although a few countries such as Ghana and Zimbabwe have experienced hyperinflation. African governments' use of externally funded, state-controlled development strategies seemed promising at the time. Many countries around the world were adopting similar approaches. Western aid to support economic interventions also helped counter the growing influence of the Soviet Union, which had supported African liberation movements against colonialism. In retrospect, we can say that the communist powers helped Africans pursue political freedoms they denied to their own people, while the aid donors and lenders helped Africans maintain economic controls they would never have tolerated at home. The net result was a substantial rise in the degree of African governments' economic intervention during the 1960s and 1970s, from the severe but targeted controls of colonial administrations to the more generalized attempts at state-led development of independent elected governments. The growth of African government intervention during the 1960s and 1970s had two major consequences. First, it fueled political instability, offering the incentives and the means for incumbents and their rivals to seize power and exploit government institutions. Some elected leaders were overthrown by force, while others became increasingly despotic and 7 only a few allowed peaceful transitions. Political opportunism among both elected and self- appointed leaders compounded the second consequence of economic intervention, which was to weaken market institutions, distort economic incentives and slow the pace of poverty alleviation. We do not know how fast African economies might have grown under different economic policies in the 1960s and 1970s, but the nature and extent of historical interventions was clearly associated with some degree of reduced growth and worsening poverty.3 In the 1980s, African governments faced mounting pressures for public-sector reform. The need to reform was triggered by a sudden rise in world real interest rates, combined with global recession that worsened Africa's terms of trade. Domestic political concerns intensified, and governments found it increasingly difficult to finance the growing fiscal deficits associated with intervention. Lenders of last resort were the World Bank, IMF, USAID and others who made their aid conditional on devaluation, deregulation, privatization and retrenchment. The three big Washington-based institutions used similar criteria for their clients around the developing world, following the "Washington consensus" reform agenda described by Williamson (1990). Trade policy reforms in the 1980s and 1990s were heavily influenced by structural adjustment programs sponsored by the World Bank and the IMF. Loan conditions were widely debated and often blamed for the economic stresses which accompanied them, but the actual implementation of reforms was typically slow and often subject to reversal or offsetting policy changes. Senegal, for example, took out the first World Bank Structural Adjustment Loan (SAL) in 1980 and received its last such loan in 1992 before switching to other instruments, such as a "Private Sector Adjustment Credit" received in 2004. The last African loan to have the term "structural adjustment" in its title was made to Mali in 2005.4 By then, the focus of World Bank-IMF conditionality had shifted to national Poverty Reduction Strategy Papers (PRSPs), a mechanism designed to involve a broader range of stakeholders and a wider variety of government activities than had been involved in the SALs. The process was initiated in 1999, and as of mid-2008 a total of 33 African countries 3 Recent studies attempting to measure the magnitude of various constraints on growth have addressed the direct effects of exogenous factors such as unfavorable demographic conditions and transport opportunities (Bloom and Sachs 1998), unfavorable temperature conditions and economic scale (Masters and McMillan 2001) and declining rainfall during the mid-1960s through the late 1980s (Barrios, Bertinelli and Strobl 2009), as well as choice variables such as institutions (Rodrik, Subramanian and Trebbi 2004), policies (Glaeser et al. 2004), and inequality (Easterly 2007). A synthesis approach allows for simultaneous determination of government choices and economic outcomes, in models that link exogenous conditions to an equilibrium level of tax rates and public investment which in turn drives growth (e.g. McMillan and Masters 2003). 4 A detailed listing of World Bank projects is available at http://go.worldbank.org/0FRO32VEI0. 8 had some sort of PRSP on record with the World Bank and the IMF.5 Of our 21 focus economies, the only countries without one are Egypt, Sudan and Zimbabwe. Africa is a large and diverse continent, divided into over 50 sovereign nations with widely varying circumstances. Some of the smaller countries have had very distinctive policy and growth experiences. For example, Botswana, Lesotho, and Swaziland had no choice but to maintain free trade in a customs union with their much richer and more powerful neighbor South Africa. This enforced openness probably facilitated convergence towards South Africa's income level, helping them achieve Africa's fastest rates of poverty alleviation through the 1970s and 1980s. Other small countries such as Cape Verde and Mauritius enjoyed high levels of migration, remittances or capital flows and experienced rapid economic growth. Africa's larger countries, including all of our 21 focus economies, have had relatively interventionist governments and slow poverty alleviation in this period, followed by reform and a degree of recovery. Studies of African economies customarily emphasize the diversity among them, which is extremely important. There are also striking patterns across countries, as found in previous studies such as Ndulu et al. (2008). The new data presented below reveals both diversity and clear trends in policy choices. Measuring Rates of Assistance and Taxation The magnitude of government interventions affecting farmers and food consumers is quantified here using the common methodology (Anderson et al. 2008) that has been adopted by the authors of this volume and the four preceding regional volumes. After a brief description of that methodology, a summary of results follows.6 Methodology The nominal rate of assistance (NRA) is defined as the percentage by which government policies have raised gross returns to farmers above what they would be without the government's intervention. Similarly, the consumer tax equivalent (CTE) is the percentage by which policies have raised prices paid by consumers of agricultural outputs. Negative values 5 A detailed listing of countries' PRSP documents is available at http://www.imf.org/external/np/prsp/prsp.asp. 6 Annual estimates and additional details may be found in the appendix. 9 of the NRA and CTE imply net taxation of farmers or net subsidies to consumers. The NRA and CTE will be identical if the sole source of government intervention is a trade measure and the two are measured at the same point in the value chain, but in general there will also be some domestic producer or consumer taxes or subsidies to differentiate them. The NRAs are based on estimates of assistance to individual industries at the farmgate. The targeted degree of coverage of the products for which agricultural NRA estimates are generated is 70 percent of the gross value of farm production at undistorted prices. The authors of the country case studies also provided guesstimates of the NRAs for noncovered farm products. For countries with non-product-specific agricultural subsidies or taxes, such net subsidies are then added to product-specific assistance to obtain NRAs for total agriculture and also for tradable agriculture for use in generating a relative rate of assistance (RRA) as defined below. Farmers are affected not only by the prices of their own outputs, but also--albeit indirectly because of the changes to factor market prices and the exchange rate--by the incentives nonagricultural producers face. In other words, not just absolute but relative prices and, hence, relative rates of government assistance affect producer incentives. If one assumes that there are no distortions in the markets for nontradables and that the value shares of agricultural and nonagricultural nontradable products remain constant, then the economy- wide effect of distortions to agricultural incentives may be captured by the extent to which the tradable parts of agricultural production are assisted or taxed relative to producers of other tradables (Vousden 1990, pp. 46-47, following Lerner 1936). By generating estimates of the average NRA for nonagricultural tradables, it is then possible to calculate an RRA, which is defined in percentage terms as: RRA = 100[(1+NRAagt/100)/(1+NRAnonagt/100) ­ 1], where NRAagt and NRAnonagt are the weighted average percentage NRAs for the tradable parts of the agricultural and nonagricultural sectors, respectively. Since the NRA cannot be less than -100 percent if producers are to earn anything, neither can the RRA. And, if both these sectors are equally assisted, the RRA is zero. This measure is useful in that, if it is below (or above) zero, it provides an internationally comparable indication of the extent to which a country's policy regime has an anti- (or pro-) agricultural bias. In calculating the NRA for producers of agricultural and nonagricultural tradables, the methodology seeks to include distortions generated by dual or multiple exchange rates. These have been important in many African countries, particularly during the 1970s and 1980s, making their estimated (typically) positive NRAs for importables and (typically) negative NRAs for exportables larger than they otherwise would have been. 10 Dollar values of farmer assistance and consumer taxation are obtained from multiplying the NRA estimates by the gross value of production at undistorted prices, to obtain an estimate in US dollars of the direct gross subsidy equivalent of assistance to farmers (GSE). This is then added up across products for a country and across countries for any or all products to get regional aggregate transfer estimates for the studied economies. These GSE values are calculated in constant dollars, and are also expressed on a per-farm- worker basis. To obtain comparable dollar value estimates of the consumer transfer, the CTE estimate at the point at which a product is first traded is multiplied by consumption (obtained from the FAO's supply and utilization database) valued at undistorted prices to obtain an estimate in constant US dollars of the tax equivalent to consumers of primary farm products (TEC). This too is added up across products for a country, and across countries for any or all products, to get regional aggregate transfer estimates for the covered farm products of our focus countries. Estimates of NRAs in agriculture Agricultural price, trade and exchange rate policies have reduced the earnings of African farmers quite substantially.7 The average rate of taxation on all agricultural production, as measured by our weighted average NRA, was less than 8 percent at the time many African countries achieved independence in the early 1960s, and then almost doubled to a peak around 15 percent in the 1970s as interventions became more severe (table 2). Reforms have since reduced the average extent of taxation to below its level of the early 1960s, including a brief period in the late 1980s when a combination of policy reforms and low international commodity prices brought the weighted average NRA to near zero. Such averages hide considerable diversity within the region, including particularly South Africa whose trend of rising net protection of farmers during the 1970s and early 1980s, followed by declining support, was opposite to trends in the Africa-wide average . A visual impression of the variation across countries and the extent of reforms between 1975-79 and 2000-04 is provided in figure 1, showing clearly the major reduction in taxation rates facing farmers in such countries as Ghana,Uganda, Tanzania, Cameroon, Senegal and Madagascar. That figure also shows the transition from taxation to support of 7 Recall that our sample covers around 90 percent of Sub-Saharan Africa's economy. For North Africa, the sample includes only Egypt, which accounts for almost half the population of North Africa but only 37 percent of its GDP. 11 farmers in Mozambique and Kenya, as well as the transition from slight support to slight taxation in Nigeria, and the continuing heavy degree of taxation still in Cote d'Ivoire, Zambia and Zimbabwe. One important type of variation in distortions is the within-country dispersion of product NRAs, as measured in table 3 by their standard deviation around the weighted mean NRA for covered agricultural products in each period. This dispersion was highest in 1985-89 when many reforms were only partly completed, but even after the recent reforms it is no lower than it was at the beginning of the period. The dispersion of NRAs within African countries is an important target for reform, whatever the level of average NRA. Variation among products has a somewhat similar pattern across countries. Figure 2 shows the pattern of dispersion in the region-wide average NRA among the key farm commodities in the late 1970s and a quarter-century later. As in other regions of the world, assistance is among the highest for sugar and milk, and is most negative for tropical cash crops such as coffee, cotton, cocoa and tobacco. The dispersion over a wider range of products and the full time period is summarized in table 4. A third type of variation is cross-country diversity of national average NRAs. This is evident from the bottom of table 2: NRA averages for the agricultural sector became more similar between the latter 1950s and the early 1970s, then less similar through to the latter 1980s, and then more similar again so that by 2000-04 this type of dispersion was back to what it had been in the early 1960s. The fourth important type of variation is differential treatment of import-competing and exportable products, in a way that often favors self-sufficiency. The extent of anti-trade bias is shown in figure 3, as the gap between the average NRAs for import-competing and exportable products. This gap grew from the 1950s through to the 1980s. It has since narrowed again, due mainly to changes in taxation of exportables, but the gap is still sizeable. This is summarized in the Trade Bias Index (TBI) reported for Africa as a whole in the middle row of table 5. Decomposing the NRA into components reveals a subtle but important influence on the aggregate average. Since the late 1970s, the share of tradable farm products that are exportables has fallen from two-thirds to just over one-half (from 67 to 54 percent). Many governments tax trade in both directions, with negative NRAs for exportables and positive NRAs for importables, so the changing composition of African agriculture from exportable to importable helps drive the aggregate NRA towards zero. This compositional effect adds to the changes within the exportables and import-competing subsectors illustrated in figure 3. 12 In the African context, product-specific input price distortions contributed very little to the sectoral NRA estimates, and in many cases the case-study authors reported no values at all. Interventions in domestic markets also contributed relatively little. Most of the region's measured NRA is due to border measures (see Appendix table 12), which are largely trade taxes, quantitative trade restrictions and the operations of parastatal trading companies. In absolute terms, the total value of taxes on farming has been substantial. Africa's anti-agricultural bias in NRA terms peaked in the late 1970s, but the sector has grown and so in constant (2000) US dollars the total value of annual transfers from farmers has risen from around $2 billion in the early 1960s (taking account of the fact that NRAs were available for only four-fifths as much agricultural production then as from 1980) to $10 billion in the 1970s, and back to around $6 billion in the 1980s (ignoring the mid-1980s period when international prices were at record lows), 1990s and 2000-04 (see bottom row of table 6(a)). The distribution across countries is shown in figure 4(a), where it is clear that the major transfers in recent years have been from farmers in Ethiopia and Sudan in the east, Zimbabwe in the south, and Cote d'Ivoire and Nigeria in the west. What is also clear from that figure is how much decline there has been since the latter 1970s in such transfers, particularly in Egypt and Tanzania but also in many smaller African economies. For Africa as a whole, the latest estimate is equivalent to a gross tax of $40 per year for each person engaged in agriculture, down from more than three times that amount in the 1970s (bottom row of table 6(b)), but still larger than government investment or foreign aid targeted to agriculture (Masters 2008, Figure 9). As shown in table 7 and figure 4(b), the burden of taxation was imposed mainly through the three major export cash crops (cocoa, coffee and cotton) plus groundnuts, beef, rice, and sugarin the 1970s. Three decades later those cash crops are still the main source of transfer from agriculture, while sugar and milk have become positively assisted. In summary, the level and dispersion of agricultural NRAs confirm that there has been substantial reform towards less distortion of incentives. However, they also suggest that there are still many opportunities for policy changes that would be both pro-poor and pro-growth, raising income for low-income farmers and improving resource allocation within and between countries. Assistance to non-farm sectors and relative rates of assistance The anti-farm policy biases of the past were due not just to agricultural policies, but also to policies affecting mobile resources engaged in other sectors. For example, to the extent that 13 protection to manufacturing also has declined over time, the relative burden on agriculture has diminished even more than the agricultural NRA suggests. This study aims to capture inter-sectoral effects through using the NRA also on non- agricultural products to generate the relative rate of assistance (RRA) between farm and nonfarm activities. The case studies were focused mainly on agricultural policy, and their NRAs for the nonfarm sector typically were measured simply using data on applied trade taxes rather than price comparisons. As a result, unlike for farm NRAs, the estimated nonfarm NRAs usually do not include the effects of quantitative trade restrictions which were important in earlier decades but have been relaxed in recent times. The nonfarm NRAs also do not capture distortions in the services sectors, some of which now produce tradables or use resources that are mobile between sectors. We can therefore be confident that the estimated NRAs for non-farm activities are smaller and decline less rapidly than in fact was the case, and that our RRA estimates understate the past level of anti-farm bias. Even though the estimates of the NRA for non-farm tradables should be considered lower-bound estimates, they turn out to be quite large. Their unweighted average among the African focus countries rose from around 12 percent in the 1960s to 27 percent during 1975- 84 before declining to around 15 percent during the most recent decade or so. As a result, the unweighted RRA is lower and dips even more (to -42 percent) in the middle of the studied period than does the NRA for agriculture, before returning at the end of the period to around the -20 percent it was in the early 1960s (figure 5(a)). Consumer tax equivalents of agricultural policies If there were no farm input distortions and no domestic output price distortions so that the NRA was entirely the result of border measures such as an import or export tax or restriction, and there were no domestic consumption taxes or subsidies in place, then the CTE would equal the NRA for each covered product. But such domestic distortions are present in several African countries. Also, the value of consumption weights used in getting the CTEs are quite different from the value of production weights used for getting weighted average NRAs (both measured at undistorted prices). Hence the average CTEs are quite diferent from the average NRAs for numerous countries, particularly those exporting cash crops in order to import staple foods. This can be seen by comparing the country and product CTEs in table 7 with the corresponding NRAs in table 2. Nonetheless, the weighted average CTE for the region has moved much like the NRA: starting at around -10 percent at the time of independence, falling 14 to -17 percent (that is, a 17 percent consumer subsidy equivalent) by the early 1970s, and then gradually lessening and eventually reaching close to zero (with a blip in the latter 1980s when Egypt overshot in its reform efforts to reduce the suppression of domestic food prices just when the international price of food fell to record low levels). The variance in national CTEs within countries also rose before the reforms and fell after the latter 1980s (see table 7 including the bottom row). In dollar terms the subsidies to consumers of farm products in Africa are largest in Sudan and Ethiopia while the tax on consumers historically has been largest in Nigeria and South Africa. Egypt prior to its reforms in the 1980s was also a huge subsidizer of food consumers. The transfer on average from producers to consumers in the region amounted in 2000-04 to around $1.7 billion per year, which is only one-third (when expressed in 2000 US dollars) the annual average transfer in the 1970s (Appendix Table 16(a)). Among the covered products, the diversity in measures across the continent means that there are no obvious stand-out products (Appendix Table 16(b)), unlike in other regions where the biggest transfers are from consumers to producers of milk, rice and sugar. The link between anti-farm and anti-trade policies A visual picture of the overall finding ­ that distortions have been reduced substantially since the 1970s ­ is provided in figure 6. That figure shows values of agriculture's trade bias index (TBI) on the horizontal axis and relative rate of assistance (RRA) on the vertical axis. An economy with no anti-agricultural bias (RRA = 0) and no anti-trade bias within the farm sector (TBI = 0) would be located at the intersection of the two axes in the upper right-hand corner. In 1975-79, South Africa was the only economy anywhere near that point, and most other Sub-Saharan African economies were far to the southwest of it. In 2000-04, by contrast, Kenya and Nigeria were also close to that neutrality point, and all the other countries shown were far closer than they were in the 1970s. This is not to say there are few distortions left within the agricultural sector though, because RRA and TBI values in the ranges -20 to -40 and -0.2 to -0.4, respectively, are not small ­ and because within most countries' agricultural sector there is still a wide dispersion of product NRAs. Note also from Figure 10 that the 2000-04 values fit roughly along a 45-degree line, as the tax burden on agriculture in these countries consists primarily of taxes on trade. 15 International spillovers and multilateral agreements Our distortion estimates take each country's border prices as given, but in reality each country's policies do have some small effect on other country's prices. An import restriction that raises domestic prices will lower prices elsewhere, and an export tax that lowers domestic prices will raise them elsewhere. In addition, attempts by one country to stabilize its domestic prices over time will reduce the stability of international prices. As a result, each country's openness to trade contributes to an international public good, offering other countries more favorable and often more stable border prices. This is a classic collective action problem, calling for a multilateral agreement to lock in freer trade policies. Collective action to stabilize world prices is precisely what was sought during the GATT's Uruguay Round Agreement on Agriculture, via tariff bindings and disciplines on administered domestic prices. Tariff bindings can reduce the extent of spillovers by restricting the range over which tariffs can increase in response to low prices. But WTO bindings are now so far above applied import tariffs that this discipline on food-importing members in years of low international prices is very weak. The most recent stage of the Doha round of WTO-sponsored multilateral trade negotiations broke down in mid-2008 because many developing countries were calling for policy space in the form of a Special Safeguard Mechanism which would have allowed even more scope for limiting imports ­ something richer members including the United States were not willing to sanction in a new agreement. Moreover, there is no corresponding GATT/WTO discipline on food export restrictions, which ­ as 2008 has starkly revealed ­ can be the problem in years of high international prices. Africa's share of world trade is so small that its policies contribute relatively little to the collective-action problem described above, except to the extent that African governments have sided with such countries as Indonesia and India in demanding special safeguards and thereby delayed or prevented the emergence of a new WTO agreement. As the victim rather than perpetrator of international agricultural-policy spillovers, however, Africa could benefit greatly from a more effective system of multilateral trade rules. International agreements may also help African governments undertake reforms that would not otherwise be possible, allowing them to make commitments and assemble coalitions that cannot otherwise be sustained. The details of WTO and other international agreements are outside the scope of this book, but generally our results regarding national policies suggest that multilateral agreements can help each government deliver more favorable market conditions for 16 agricultural development at the very least by limiting the rise of import restrictions in other countries. In addition, following the imposition by numerous food-exporting developing countries in 2008 of export restrictions that harmed food importers, perhaps WTO members may eventually agree to limit export restrictions as well. Summary: What have we learned? Each of the case studies presented in this volume provides detailed insights into Africa's wide variety of country experiences. Aggregating their results to characterize all of Africa necessarily obscures as much as it reveals. Making generalizations is sometimes useful, however, if only to allow comparison with other regions, and to detect common trends that cannot be seen in individual cases. Averaging over the 21 African countries considered in this study, our principal findings are the following. African governments have removed much of their earlier anti-farm and anti-trade policy biases. Government policy biases against agriculture had worsened in the late 1960s and 1970s, primarily through increased taxation of exportable products. Reforms of the 1980s and 1990s reversed that trend, and average rates of agricultural taxation are now back to or below the levels of the early 1960s. Most of this gain has come from reduced taxation of farm exports. Substantial distortions remain, and still impose a large tax burden on Africa's poor. In constant (2000) US dollar terms, the transfers paid by farmers in our 21 focus countries peaked in the late 1970s, at over $10 billion per year or $134 per farm worker. In 2000-04 the burden of taxation averaged $6 billion per year, or $41 per person working in agriculture. Even this lower amount is appreciably larger than public investment or foreign aid into the sector. The continuing taxation in Africa contrasts with both Asia and Latin America, where the average agricultural NRAs and RRAs reached zero by the early 21st century, although like Africa those regions still have a wide dispersion of NRAs across products and countries. African farmers have become less taxed in part because of the changing trade orientation of African agriculture. Reduced taxation of farmers has occurred in part because of a decline in the share of output that is exportable, and a corresponding rise in the share that is import-competing. The rate of protection from imports for these products has fluctuated but remains positive. This helps only the few farmers who are net sellers of the protected products, however, and does so in a way that is less efficient and less equitable than many other possible interventions. 17 Trade restrictions continue to be Africa's most important instruments of agricultural intervention. Domestic taxes and subsidies on farm inputs and outputs, and non-product- specific assistance, are a small share of total distortions to farmer incentives in Africa. As a result, policy incidence on consumers tends to mirror the incidence on producers, with fiscal expenditures playing a much smaller role than in more-affluent regions. Differences in NRAs and RRAs across commodities and countries are still substantial. Dispersion rates, as measured by the standard deviation in NRAs and RRAs across commodies and countries, rose and then fell over time. Looking forward, whatever the overall level of taxation or assistance, moving towards more uniform rates within the farm sector and between countries within the region could still yield substantial increases in efficiency of resource use. Where to from here? Every reader of this volume will draw their own conclusions as to what these findings imply about the future of agricultural policy in Africa. We expect that the policy choices of African governments will continue to vary, but we hope that the overall trend towards reform will continue. Despite difficult conditions, many African governments will continue to reduce taxation of agricultural exports, improve market institutions, and invest in rural public goods. In response, we expect that producers will continue to respond in ways that generate faster economic growth and sustained poverty alleviation. That has been the pattern in other regions, and African countries have shown their willingness and ability to begin these changes. Our hopes are tempered by experience, however, including particularly the experience of agricultural policy transition in other regions. A fundamental concern in agricultural policy over time as economies join the middle-income group is `overshooting'. In response to rural poverty and inequality, many countries start protecting agriculture soon after they stop taxing it. This imposes large costs on consumers, and slows national economic growth. Countries that lock in relatively efficient and equitable policies as soon as they are attained can therefore enjoy a high payoff relative to those that allow farm support policies to become increasingly costly over time. In particular, policies that raise the prices of staple foods 18 impose serious costs on the urban poor and on rural net buyers of these products, as has been demonstrated by recent increases in their prices for other reasons (Ivanic and Martin 2008). Rural-urban poverty gaps can be addressed in far more efficient ways than by subsidizing production or raising food prices. For example, rural poverty can and has been alleviated in parts of Africa and Asia by the mobility of some members of farm households who work full- or part-time off the farm and repatriate part of their higher earnings back to those remaining on the farm (Otsuka and Yamano 2006, World Bank 2007). Concerted government interventions through targeted social policy measures can also be an efficient and effective way to reduce gaps between rural and urban incomes and raise national incomes overall (Winters, McCulloch and McKay 2004). Efficient ways of assisting the left-behind groups of poor (nonfarm as well as farm) households include public investment measures that have high social payoffs such as basic education and health, rural infrastructure and agricultural research and extension. References Anderson, K. (ed.) (2009a), Distortions to Agricultural Incentives: A Global Perspective, 1955 to 2007, London: Palgrave Macmillan and Washington DC: World Bank (forthcoming). Anderson, K., M. Kurzweil, W. Martin, D. Sandri and E. Valenzuela (2008), "Methodology for Measuring Distortions to Agricultural Incentives", Agricultural Distortions Working Paper 02, World Bank, Washington DC, revised January. Posted at www.worldbank.org/agdistortions and reproduced as Appendix A to this volume. Anderson, K. and W. Masters (eds.) (2009), Distortions to Agricultural Incentives in Africa, Washington DC: World Bank. Anderson, K. and E. Valenzuela (2008), Global Estimates of Distortions to Agricultural Incentives, 1955 to 2007, data spreadsheet available at www.worldbank.org/agdistortions. Barrios, S., L. Bertinelli and E. Strobl (2009), "Trends in Rainfall and Economic Growth in Africa: A Neglected Cause of the African Growth Tragedy", Review of Economics and Statistics, forthcoming. Bloom, D.E. and J.D. Sachs (1998), "Geography, Demography, and Economic Growth in Africa", Brookings Papers on Economic Activity 29(2): 207-96. 19 Chen, S, and M. Ravallion (2007), "Absolute Poverty Measures for the Developing World, 1981-2004", Policy Research Working Paper 4211, World Bank, Washington DC, April. Easterly, W. (2007), "Inequality Does Cause Underdevelopment: Insights From a New Instrument", Journal of Development Economics 84(2): 755-76. Glaeser, E.L., R. La Porta, F. Lopez-de-Silanes and A. Shleifer (2004), "Do Institutions Cause Growth?" Journal of Economic Growth 9(3): 271-303. Ivanic, M. and W. Martin (2008), "Implications of Higher Global Food Prices for Poverty in Low-Income Countries" Policy Research Working Paper 4594, World Bank, Washington DC, April. Krueger, A.O., M. Schiff and A. Valdés (1988), "Measuring the Impact of Sector-specific and Economy-wide Policies on Agricultural Incentives in LDCs", World Bank Economic Review 2(3): 255-272. Krueger, A.O., M. Schiff and A. Valdés (1991), The Political Economy of Agricultural Pricing Policy, Volume 3: Africa and the Mediterranean, Baltimore: Johns Hopkins University Press for the World Bank. Lerner, A. (1936), "The Symmetry between Import and Export Taxes", Economica 3(11): 306-13, August. Masters, W.A. (2008), "Beyond the Food Crisis: Trade, Aid and Innovation in African Agriculture", African Technology Development Forum 5(1) (forthcoming). Masters, W.A. and M.S. McMillan (2001), "Climate and Scale in Economic Growth", Journal of Economic Growth 6(3): 167-86. McMillan, M.S. and W.A. Masters (2003), "An African Growth Trap: Production Technology and the Time-Consistency of Agricultural Taxation, R&D and Investment", Review of Development Economics 7(2): 179-91. Ndulu, B., S.A. O'Connell, R.H. Bates, P. Collier, C.C. Soludo, J.-P. Azam, A.K.Fosu, J.W. Gunning and D. Njinkeu (2008), The Political Economy of Economic Growth in Africa, 1960­2000, Volumes 1 and 2, Cambridge and New York: Cambridge University Press. Otsuka, K. and T. Yamano (2006), "Introduction to the Special Issue on the Role of Nonfarm Income in Poverty Reduction: Evidence from Asia and East Africa", Agricultural Economics 35 (supplement): 373-97, November. 20 Rodrik, D., A. Subramanian and F. Trebbi (2004), "Institutions Rule: The Primacy of Institutions Over Geography and Integration in Economic Development", Journal of Economic Growth 9(2): 131-65. Sandri, D., E. Valenzuela and K. Anderson (2007), "Economic and Trade Indicators for Africa" Agricultural Distortions Working Paper 21, World Bank, Washington DC, December. Posted at www.worldbank.org/agdistortions Spence, M. et al. (2008), The Growth Report: Strategies For Sustained Growth and Inclusive Development (Report of the Commission on Growth and Development, Chaired by Michael Spence), Washington DC: World Bank. Valenzuela, E., J. Croser. M. Kurzweil, S. Neglen and K. Anderson (2008), "Annual Estimates of African Distortions to Agricutural Incentives", Appendix B to this volume. (A fuller version appears at www.worldbank.org/agdistortions as Agricultural Distortions Working Paper 55, World Bank, Washington DC, December 2007.) Vousden, N. (1990), The Economics of Trade Protection, Cambridge: Cambridge University Press. Williamson, J. (1990), "What Washington Means by Policy Reform", Ch. 2 in J. Williamson (ed.), Latin American Adjustment: How Much Has Happened? Washington DC: Institute for International Economics, April. Winters, L.A., N. McCulloch and A., McKay (2004), "Trade Liberalization and Poverty: The Empirical Evidence", Journal of Economic Literature 62(1): 72-115, March. World Bank (1986), World Development Report 1986: Trade and Pricing Policies in World Agriculture, New York: Oxford University Press. World Bank (2007), World Development Report 2008: Agriculture for Development, Washington DC: World Bank. World Bank (2008), World Development Indicators, Washington DC: World Bank. WTO, ITC and UNCTAD (2007), World Tariff Profiles 2006, Geneva: World Trade Organization. 21 Figure 1: Nominal rates of assistance to agriculture, individual African focus countries and unweighted regional average, 1975-79 and 2000-04a (percent) 20 0 -20 -40 -60 Mozambique Kenya Madagascar Uganda Cameroon South Africa Ghana Nigeria 20 0 -20 -40 -60 Egypt Senegal Ethiopia Sudan Tanzania Cote d'Ivoire Zambia Zimbabw e 1975-79 2000-04 Source: Anderson and Valenzuela (2008) based on estimates reported in the Appendix and in Anderson and Masters (2008). a Ethiopia data for the first period refer to 1981-84 as 1975-79 data are unavailable. 22 Figure 2: Nominal rates of assistance, key covered product, African focus countries, 1975-79 and 2000-04 (percent, weighteda average across 21 countries) Sugar Sorghum Milk Poultry Banana Plantain Wheat Millet Cassava Yam Sunflow er Maize Rice Coffee Palmoil Vanilla Tea Sheepmeat Bean Beef Cocoa Sesame Groundnut Cotton Soybean Tobacco -100 -50 0 50 1975-1979 2000-04 Source: Anderson and Valenzuela (2008) based on estimates reported in the Appendix and in Anderson and Masters (2008). a. Weights based on gross value of agricultural production at undistorted prices, with each NRA (by country, by product) weighted by the country's value of production of that commodity in a given year. 23 Figure 3: Nominal rates of assistance to exportable, import-competing and alla agricultural products, African region, 1955 to 2004 (percent, weighted averages across 16 countries) 80 60 40 20 perc ent 0 -20 -40 -60 1955-59 1960-64 1965-69 1970-74 1975-79 1980-84 1985-89 1990-94 1995-99 2000-04 year exportables import-competing total Source: Anderson and Valenzuela (2008) based on estimates reported in the Appendix and in Anderson and Masters (2008). a. The total NRA can be above or below the exportable and importable averages because assistance to nontradables and non-product specific assistance is also included. 24 Figure 4: Gross subsidy equivalents of assistance to farmers, African focus countries,a 1975- 79 and 2000-04 (constant 2000 US$ billions) (a) Total per country 1.5 1.0 0.5 US$, billions 0.0 -0.5 -1.0 -1.5 -2.0 -2.5 Kenya Mozambique Uganda South Africa Madagascar Cameroon Ghana 1.5 1.0 US$, 0.5 billions 0.0 -0.5 -1.0 -1.5 -2.0 -2.5 Senegal Zambia Tanzania Egypt Zimbabwe Cote d'Ivoire Nigeria Ethiopia Sudan 1975-79 2000-04 25 Figure 4 (continued): Gross subsidy equivalents of assistance to farmers, African focus countries,a 1975-79 and 2000-04 (constant 2000 US$ billions) (b) Total per product Sugar Milk Maize Sorghum Poultry Millet Wheat Coffee Palmoil Tea Rice Bean Yam Cassava Tobacco Sheepmeat Groundnut Cotton Cocoa Beef -3500 -2500 -1500 -500 500 1500 1975-79 2000-04 Source: Anderson and Valenzuela (2008) based on estimates reported in the Appendix and in Anderson and Masters (2008). 26 Figure 5: Nominal rates of assistance to agricultural and non-agricultural tradable products and relative rate of assistance,a Africa region, 1955 to 2004 (percent, weighted averages across 16 countries)a 30 20 10 0 percent -10 -20 -30 -40 1955­59 1960­64 1965­69 1970­74 1975­79 1980­84 1985­89 1990­94 1995­99 2000­04 year NRA agriculture NRA nonagriculture RRA Source: Anderson and Valenzuela (2008) based on estimates reported in the Appendix and in Anderson and Masters (2008). a. The RRA is defined as 100*[(100+NRAagt)/(100+NRAnonagt)-1], where NRAagt and NRAnonagt are the percentage NRAs for the tradables parts of the agricultural and non- agricultural sectors, respectively. The 5 small cotton-exporting countries of West and Central Africa are not included here. 27 Figure 6: Relationship between RRA and the trade bias index for agriculture, African focus countries, 1975­79 and 2000­04 a. 1975­79 20 Nigeria 10 Relative Rate of Assistance, % South Africa 0 -10 -20 - Sudan -30 -40 - Senegal Egypt -50 Ghana Cote d'Ivoire Madagascar -60 -70 Tanzania Zimbabwe Zambia -80 Mozambique Uganda -1 -0.8 -0.6 -0.4 -0.2 0 0.2 Trade Bias Index b. 2000­04 20 10 Relative Rate of Assistance, % Mozambique Kenya 0 South Africa -10 Ghana Nigeria Madagascar -20 Sudan Senegal Uganda -30 Tanzania Egypt Cote d'Ivoire -40 Zambia -50 -60 -70 -80 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 Trade Bias Index Source: Anderson and Valenzuela (2008) based on estimates reported in the Appendix and in Anderson and Masters (2008). 28 Table 1: Key economic and trade indicators, African focus countries, 2000-04 Share (%) of world: National rel. to TSIb Povc Gini world (=100) Indexd Pop'n Total Agri GDP Ag RCAa 2004 GDP c per land ag & GDP capita per food capita Benin 0.12 0.01 0.09 7 55 1034 na 31 39 Burkina Faso 0.19 0.01 0.09 5 111 953 na 29 40 Cameroon 0.25 0.03 0.38 13 74 445 na 15 45 Chad 0.14 0.01 0.07 5 695 na na na na Cote d'Ivoire 0.28 0.04 0.21 12 139 722 na 18 48 Egypt 1.13 0.26 1.11 23 6 175 na 2 34 Ethiopia 1.08 0.02 0.23 2 58 958 na 12 30 Ghana 0.33 0.02 0.2 6 88 748 na 17 41 Kenya 0.52 0.04 0.29 8 103 636 na 12 43 Madagascar 0.28 0.01 0.1 5 202 670 0.94 63 47 Mali 0.2 0.01 0.1 5 353 624 na 39 40 Mozambique 0.3 0.01 0.08 4 324 359 -0.03 30 47 Nigeria 1.98 0.15 1.09 8 73 3 na 71 44 Senegal 0.17 0.02 0.09 10 94 444 na 13 41 South Africa 0.73 0.42 0.39 59 275 134 0.52 9 58 Sudan 0.55 0.05 0.5 8 490 209 na na na Tanzania 0.58 0.03 0.33 5 166 800 0.73 56 35 Togo 0.09 0 0.05 5 80 407 na na na Uganda 0.42 0.02 0.15 4 60 938 0.8 83 46 Zambia 0.18 0.01 0.07 7 398 194 0.35 60 51 Zimbabwe 0.21 0.04 0.14 18 200 602 0.83 62 50 African focus countries 9.73 1.21 5.74 13 145 na na na na All Sub-Saharan Africa 9.37 0.98 4.93 10 164 na 0.55 41 na All North Africa 2.34 0.70 2.81 30 84 na -0.78 na na All Africa 11.7 1.67 7.74 14 148 na 0.20 32 na 1 Source: Sandri, Valenzuela and Anderson (2008), compiled mainly from World Bank's World Development Indicators. a. Revealed Comparative Advantage = share of agriculture and processed food in national exports as a ratio of that sector's share of global exports b. Primary Agriculture Trade Specialization = (X-M)/(X+M), 2000-02 (world av =0). c. Percentage of population living on