V7 G70 Education and Social Policy Department The World Bank September 1993 ESP Discussion Paper Series No. 10 A Review of the World Bank's Efforts to Assist African Governments in Reducing Poverty Tony Addison The ESP Discusion Paper evies serves s an internal system of record keeping, reference, and reealfor de products of dw Education and Social PoUcy Deparanent's wot* program The viewr erpressed here are der of the esahors anod oud no be anfibaed to *he World Bank or Is Bowad of Eecudve Directors or the coraaIes they reprea. Education and Social Policy Department The World Bank September 1993 ESP Discussion Paper Series No. 10 A Review of the World Bank's Efforts to Assist African Governments in Reducing Poverty Tony Addison he ESP Discmion Paper Serfe av as inenal system of rcord keeping, refeswes, ad raevalfor se product f e Educades ad Social Po&tey Depaem's wo*k prgsrn. Tr views Cprearedlhere are lioe of the eat&or mad shoWld nwb e rths&ed o dhe Wsd Bwak or its Bard of Exeatyer Diretors or lie comvaoes liey presa. A REVIEW OF THE WORLD BANK'S EFFORTS TO ASSIST AFRICAN GOVERNMENTS IN REDUCING POVERTY Tony Addison University of Warwick October 13, 1993 A Review of the World Bank's Efforts to Assist African Governments in Reducing Poverty Abstract Recent activity by the World Bank's Africa region is reviewed, with a focus on information collection and analysis, strategies for poverty reduction, and Bank supported policies and programs. New trends are highlighted and ways of enhancing the effectiveness of Bank operations in reducing poverty are discussed. It is emphasized that widespread political uncertainty and civil conflict, economic decline as a result of inappropriate policies and weak institutional capacities provide a difficult context for achieving poverty reduction in Africa. The information base on poverty is being improved by Bank supported surveys of households and communities, together with beneficiary and participatory poverty assessments. Identifying the constraints which the poor face, the regional characteristics of poverty, and the causes of female impoverishment can guide the design of market liberalization, the restructuring of public expenditures, and the choice of investment projects. Bank studies such as those of food security, human resources development, and women and -development help to define the Bank's country strategies for poverty reduction. These strategies provide precise recommendations for redirecting social expenditures toward basic services, but measures to improve the economic opportunities of the poor (especially their employment) need to be more clearly defined to make them fully effective as guides to operations. In particular, further analysis of the impact of reform on poverty can strengthen the poverty focus of the relevant policies and programs. In addition, public expenditure reviews should assess how best to redirect investments in infrastructure towards meeting the needs of the poor. Issues of poverty are increasingly discussed in dialogue with governments. Conditionalities to improve the benefits of reform to the poor are now more common in structural adjustment loans, and further analysis can help to guide the choice of conditionalities. Market reforms are improving living standards, but slow progress in reforming agricultural marketing is acting as a brake on rural poverty reduction, and earlier liberalization should therefore be encouraged. Agricultural projects can reduce poverty, but their effectiveness requires further assessment. The redirection of social expenditures towards the poor is being encouraged, but chaotic budgetary processes and weak institutional capacities hamper progress. Social funds incorporate lessons learned from social action programs, and as a result they are quicker to implement and are better able to reach the rural poor. The paper concludes that increasing the poverty focus of structural adjustment and sector lending is one of the most effective ways for the Bank to help Africa's poor. For such a focus, further information collection and analysis is necessary. Finally, the Bank can encourage governments to make poverty and social data more widely available, thereby stimulating a more informed debate within countries. This would help to establish the social consensus necessary for firm action on poverty reduction. A REVIEW OF THE WORLD BANK'S EFFORTS TO ASSIST AFRICAN GOVERNMENTS IN REDUCING POVERTY Table of Contents A cronym s ........................................................ ii Preface ......................................................... iii Summary ........................................................ v I. Introduction.................................................. II. The Economic and Social Context for the World Bank's Operations in Africa ......... 1 III. The Collection of Information on Poverty and the Analysis of Poverty ............. 4 IV. World Bank Supported Policies and Programs ........................... 19 V. Conclusions................................................. 28 List of Boxes Box 1. Zimbabwe: Inequalities in Household Access to Productive Assets within a Rural Economy ................................................. 4 Box 2. Chad: Collecting Priority Data on Urban Poverty ........................ 7 Box 3. Ethiopia: Rapid PovertyAssessment ................................ 8 Box 4. Participatory PovertyAssessment ................................. 9 Box 5. Nigeria: Assessing the Extent of Food Poverty ......................... 10 Box 6. C6te d'Ivoire: Using Household Data to Clarify the Benefits of Social Expenditures to the Poor ...................................... 13 Box 7. Tanzania: Analyzing Adjustment's Rural-Poverty Impact .................. 14 Box 8. Togo: Assessing the Impact of Agricultural Policy Reform ................. 16 Box 9. Malawi: Improving the Labor Market for the Poor ...................... 17 Box 10. Zimbabwe: The Informal Sector and the Role of Women .................. 18 * Box 11. Ghana: Cocoa Price Reform and the Poor ........................... 23 Box 12. Burkina Faso and Rwanda: Translating Food Security Strategies into Action .............................................. 25 Acronyms AgSAC Agricultural Sector Adjustment Credit CEM Country economic memorandum EdSAC Education Sector Adjustment Credit ESW Economic and sector work HRD Human resources development IDA International Development Association LSMS Living Standards Measurement Study OD 4.15 Operational Directive 4.15: Poverty Reduction PER Public expenditure review PFP Policy framework paper PHN Population, health, and nutrition PIP Public investment program SAL Structural Adjustment Loan SAP Social Action Program SDA Social Dimensions of Adjustment SSA Sub-Saharan Africa WDR90 World Development Report 1990 WID Women in development Preface This paper was prepared for the Poverty Analysis and Policy Division of the Population and Human Resources Department during the summer of 1992, as an input to the progress report Implementing the World Bank's Strategy to Reduce Poverty - Progress and Oallenges. The paper was updated and revised during 1993 for the Education and Social Policy Department. It is based on a series of discussions with staff in the Bank's Africa region, together with a survey of relevant reports and other documentation. The author is indebted to Helena Ribe for her insights and encouragement. The paper could not have been written without the willingness of Bank staff to give generously of their time to discuss the issues. Especial thanks are due to the poverty coordinators of the Africa Region: Christine Allison, Sarah Keener, Lynne Sherburne-Benz, Neeta Sirur, Tom Stephens, Roger Sullivan, and Mark Woodward. Comments on early outlines by Emmanuel Jimenez, Lyn Squire, and Michael Walton helped to focus the paper. Discussions with Colin Bruce, Jim Coates, Soniya Carvalho, Robert Crown, Adriana De Leva, Lionel Demery, Graeme Donovan, Jean-Luc Dubois, Ed Elmendorf, Peter Fallon, Paul Glewwe, Chris Grootaert, Charles Humphreys, Christine Jones, Steen Jorgensen, Kapil Kapoor, Kathie Krumm, Alexandre Marc, Tim Marchant, Oey Meesook, Mike Mills, Michel Noel, Eric Rice, Larry Salmen, Antoine Simonpietri, Gurushri Swamy, Rogier Van Den Brink, Jack van Holst Pellekaan, and Harry Walters were particularly useful. Julia Dayton provided much able assistance in gathering materials. Fiona Mackintosh provided considerable help with the editing, and Alexander Arenas and Eileen Dempsey helped to process the document. The views expressed in this report are those of the author and should not be attributed to the World Bank. Summary Introduction. The purpose of this paper is to review recent activity by the World Bank's Africa region on poverty issues. Its aim is to highlight new trends and to identify ways of strengthening the Bank's assistance to governments in the reduction of poverty. The Economic and Social Context for the World Bank's Operations in Africa. The Bank's efforts to assist African governments in reducing poverty are constrained across the region by widespread political uncertainty and civil conflict, economic decline as a result of policy distortions, political patronage, and weak institutional capacities. Even after policy distortions have been removed, long-term development solutions will still be required, particularly for agriculture. Moreover, because of rapid population growth, both employment opportunities and social services are under intense pressure, making further progress more difficult to achieve. The Collection of Information on Poverty and the Analysis of Poverty. Analysis using information collected from households and communities has an important role to play in helping to incorporate the objective of poverty reduction into the policy dialogue between governments and the Bank. On the basis of this dialogue, and the information and analysis supporting it, the benefits of policies and programs to the poor can be enhanced in three ways. First, while measures that. benefit agriculture and the informal sector will help many poor people, some of the poor face constraints that limit their market opportunities and therefore their gains from such measures. Information on poverty can help to identify the constraints which the poor face, thereby assisting the design of measures to increase their gains from policies and programs of benefit to agriculture and the informal sector. Second, the poor are often disproportionately concentrated in regions having the least economic and social infrastructure. Understanding the characteristics of regional poverty is important to eliminating policy biases against poor regions, and to redirecting public expenditures towards the poor. Third, understanding the determinants of female impoverishment is important to the design of policies and programs of benefit to women. The Bank has contributed to a range of capacity-building exercises aimed at helping national statistical offices build an information base on households and communities. Despite the weak capacities of most statistical offices, household surveys supported by the Bank have yielded results, albeit more slowly than planned. The Bank has developed a wide range of survey instruments, including the Priority Survey which is designed to be implemented quickly and is suitable for statistical offices with limited capacities. The Bank is also supporting the use of rapid appraisal methodologies such as beneficiary and participatory assessments which collect the views of poor people regarding the causes of their poverty and the effectiveness of measures designed to help them. Studies conducted by the Bank of such issues as the benefits of public expenditures, food security, human resources development, and women in development constitute a useful information base on poverty. They present well-defined priorities for policy action within sectors. However, these studies have not taken a cross-sectoral view, which is why poverty assessments will be' crucial for identifying cross-sectoral priorities. Three major gaps in the analysis need to be filled to strengthen the favorable impact of policies and programs on the poor. First, the impact of reform on poverty requires more analysis. There has been a tendency for the available household data to be underused in operational analyses, and more use can be made of other socio-economic indicators when household data is scarce. Second, a higher priority should be given to analyzing employment issues in order to strengthen policies aimed at achieving labor- intensive growth. Third, the analysis of public expenditures, while comprehensive on the access of the vi poor to social services, has not sufficiently analyzed the impact on poverty of public investments, particularly investments in infrastructure. World Bank Supported Policies and Programs. The Bank's strategies for reducing poverty in each country are based on the themes set out in the World Development Report of 1990. Country strategies for poverty reduction emphasize the redirection of public expenditures in health and education towards services of most benefit to the poor. Their weakest element is in defining measures to improve the economic opportunites of the poor, in particular by adding specific pro-poor components to policy reform. Bank operations in those countries for which poverty assessments have been prepared have the strongest and most well-defined measures for incorporating poverty reduction into policies and programs. Poverty assessments are now being prepared for more countries, and these will help to strengthen the link between the analysis of poverty and the formulation of strategy, and will thereby enhance the poverty focus of Bank operations. Efforts to improve the sensitivity of governments to the need to reduce poverty are increasing. Conditionalities relating to poverty are now more common in structural adjustment loans. Since loans to support structural adjustment are the most high profile component of the Bank's lending program, this conditionality acts as a signal to governments that the Bank has increased its commitment to poverty reduction. However, overcrowding the SAL with too many conditionalities, thereby slowing the implementation of the program itself, should be avoided. With more analysis of poverty, the most effective reforms to help poverty reduction can be identified, and suitable conditionalities put in place. Where market reforms have been implemented, living standards are beginning to improve, but difficulties in achieving further improvements in agricultural marketing (particularly for traditional export crops) are acting as a brake on rural poverty reduction in some countries. For individual countries and crops, the speed at which the reform of marketing proceeds partly depends on the scope for private marketing and the influence of vocal rent-seekers opposed to reform. Governments should be encouraged to implement the rapid liberalization of crop marketing in order to accelerate the recovery of rural incomes. The Bank's agricultural projects (particularly those which improve rural infrastructure) have an important role to play in reducing poverty, but more analysis of their effectiveness in achieving this objective will help to improve their focus on poverty. The Bank is using leverage to encourage governments to restructure their public expenditures so that social expenditures can be reallocated to the poor. The Bank is working with governments to improve the cost effectiveness of services, but this is inevitably a complex process. Chaotic budgetary processes, weak institutional capacities, and the opposition of vocal groups are constraining the speed of progress. Lessons learned from the experience of implementing social action programs have been incorporated into the design of social funds. Social funds are more able to reach the rural poor than the social action programs, and are quicker to implement. Conclusions. Increasing the poverty focus of structural adjustment and sector lending is one of the most effective means for the Bank to help reduce poverty. Increasing the poverty focus of Bank lending requires further analysis of poverty issues in country economic memoranda and public expenditure reviews. Improving the analysis of the impact of market reforms on the poor and the benefits of public expenditures to the poor requires the collection of information on poor people and their communities. More comprehensive assessments of the poverty effects of market reforms on poverty will improve dialogue with governments and the donor community. vii The Bank should consider further how best to achieve poverty reduction in rural regions which contain high concentrations of the poor. It is too often assumed that migration to more prosperous regions can be easily undertaken by the poor of regions with limited agricultural potential. Further analysis of instruments (such as improved education) to enhance their mobility is required. The Bank must also be wary of allowing government 'sensitivity' to block appropriate dialogue and operations to reverse ethnic discrimination. The Bank has an important role to play in encouraging a more active debate on poverty issues within Africa. One means to this end is to press governments to make poverty and social data more widely available. This would help to stimulate a more informed debate within countries, thereby helping to establish the social consensus necessary for firm action on poverty reduction. I. Introduction This paper reviews recent action taken by the Bank's Africa region on poverty issues.: It was prepared in the midst of much self-analysis within the Bank itself, a process that began with the World Development Report of 1990 on poverty, and accelerated with the issue of Operational Directive 4.15: Poverty Reduction (O.D.4.15). While any review must necessarily have a broad scope - since nearly all the Bank's work may have some impact on the poor - the aim here is not to itemize every activity, but rather to highlight new trends and to identify ways of strengthening the Bank's focus on poverty. The paper begins by summarizing the economic and social context within which the Bank's operations in Africa take place (Section II). This context provides an agenda for the Bank, but it also imposes constraints on what can be achieved. The paper then turns to the activities of the Bank. The Bank's understanding of poverty depends upon the information available, including the information the Bank itself collects, and the resulting analyses. These, in turn, help to define the Bank's strategy for assisting governments to reduce poverty. These strategies then guide operations at the program and policy levels. Needless to say, these processes do not always connect smoothly day-by-day, but together they do provide a framework within which to discuss the range of the Bank's activities. Accordingly, the paper discusses information collection and analysis (Section III), and then policies and programs (Section IV). Section V offers some concluding observations on what will be necessary if poverty is to be reduced further in Africa in the 1990s. II. The Economic and Social Context for the World Bank's Operations in Africa 1. Civil Conflict and Political Uncertainty are Widespread. The last two decades have seen wars and political instability throughout the region. Ethiopia and Uganda are now rebuilding after years of strife, but others - such as Somalia - are mired in civil war. The security situation in Mozambique, Sudan, and much of the Sahel remains precarious. Transitions to multi-party politics have now begun. These have been successful in Tanzania and Zambia, but can be fragile, as the situations in Kenya and Nigeria show. Established elites are being challenged in Cameroon, C6te d'Ivoire, and Madagascar. In some countries - most notably Malawi and Zaire - uncertainty prevails as the era of personal rule draws to a close. In these situations, priority has too often been given to buying off political challenges, while long-term development strategy has been neglected. The policy instability associated with personal rule has encouraged rent-seeking to the detriment of productive investment and employment. Many of the rural poor have withdrawn from the market economy and have retreated into subsistence activities. New groups of the poor - such as refugees - have also been created by civil strife. Many African countries present a difficult political environment for any donor agency that is intent on reducing poverty. The Africa region of the Bank is hereafter referred to as 'the Bank'. 2 2. Policy Distortions Leave a Legacy of Economic Decline. A policy history of overvalued exchange rates, inefficient state marketing, and low producer prices have depressed farm earnings across the region, thereby exacerbating the plight of the rural poor? Inefficient but protected manufacturing industries have imposed high costs on poor consumers, while failing to generate employment growth. Mismanagement of public expenditures and revenues has drained resources from basic education, health care and infrastructure. Africa's decline contrasts strongly with Asia's success where appropriate policies, together with human capital investments, have accelerated growth, and thereby reduced poverty. The creation of a policy framework that is conducive to agriculture and export manufacturing will eventually reverse the decline in per capita incomes. But achieving a political consensus on reform remains politically difficult; inefficient but influential manufacturers lobby against devaluation and import liberalization, while overstaffed government agencies block efficiency improvements. Creating an efficient economic system, and thereby new economic opportunities for the poor, is inevitably a slow process. 3. Many Governments Lack Commitment to Reducing Poverty. Taking action to reduce poverty requires governments to redistribute public expenditures towards basic services, to pay farmers more for their crops, and to introduce competition into consumer markets. But, many of the policies that benefit the poor work against the interests of higher-income groups, who are politically influential ("vocal") in their demands. These are often the very people who constitute the political power base of the region's governments. Thus governments often direct public expenditures to the benefit of higher-income groups, both as government employees and as users of services, rather than to services that are important to poor people such as preventive health care and primary education. Paying farmers a low share of the world price of their commodities taxes the rural poor but creates public revenues with which to buy political support among the vocal. Import controls, by limiting competition, create profits for favored manufacturers and economic rents for those who trade in scarce goods but tax the rest of society, especially those on low-incomes who have the least access to scarce goods. As a result, many governments are reluctant to define a poverty agenda (a concrete set of actions to reduce poverty) in order to avoid sacrificing political patronage. Instead, they use rhetoric such as "everyone is poor" to avoid making a political commitment to poverty reduction that would be unpopular with those who keep them in power. 4. The Institutional Capacities of Government are Poorly Developed. The capacities of governments for designing and implementing programs to benefit the poor are often very weak at both central and local levels. Budgetary processes are chaotic in many African ministries. Budgetary planning is haphazard, decision making is cumbersome and over-centralized, and expenditures are poorly controlled. Education and health ministries are often the weakest, due to underinvestment in 2 The combined effect of these measures is estimated to amount to a 50 percent tax on farm incomes, by far the highest rate of any developing region. See M. Schiff, 1991, "Sectoral, Macroeconomic and Industrial Trade Policies: Measuring their Effect on Agricultural Incentives and Distortions in Africa and Asia." World Economy, Vol. 14 No.4, December. 3 their institutional capacities. Consequently, redirecting public expenditures toward services that benefit the poor is frequently a slow and cumbersome process. Non-governmental organizations (NGOs) provide important help to donors but this is not an adequate substitute for an effective government capacity to implement. 5. Agricultural Systems are Fragile. Since agriculture provides the livelihood of most of Africa's poor and about 75 percent of all employment, its revitalization is crucial to reducing poverty. The Bank's Long-term Perspective Study (LTPS) sets a target of 4 percent per annum agricultural growth.' This would, given 3 percent per annum population growth, restore growth in rural incomes and help to achieve food security. Expansion of acreage - rather than higher productivity - accounts for most of agriculture's modest output growth (2 percent per annum) since 1960.' The 4 percent target is achievable, but only if reforms in agricultural marketing and pricing policy are adopted to provide a conducive environment for investments that will increase productivity. In addition to reforms, the difficult tasks of encouraging the adoption of hybrid varieties, improving farm technologies, and sustaining farmers in hostile agro-climatic conditions will have to be undertaken to achieve the 4 percent target. Policy reform, while crucial to rural poverty reduction, is only a first step in the complex task of managing Africa's difficult agricultural conditions. 6. Social Services and Employment Opportunities Remain Under Pressure from Rapid Population Growth. * Population growth of 3 percent per annum is putting health and education services under intense pressure.5 With only 2 percent per annum agricultural growth, rural underemployment is increasing. With only 1 percent growth in modern sector employment, underemployment is growing among the young, who are thus being absorbed into overcrowded informal sectors. As a result of population pressures, it is difficult to make progress when social demands for employment and services are growing so fast. In summary, Africa - of all the developing regions - presents the most challenges to the Bank in its efforts to help governments to reduce poverty. While the Bank's assistance to regenerate agriculture, create new manufacturing sectors, and improve public administrations all help to reduce poverty, the issue of governance remains the most important constraint on the Bank's efforts in Africa. Having set the context for the Bank's operations in the region, we now turn to the activities of the Bank itself. 3 Sub-Saharan Africa From Crisis to Sustainable Growth, Washington DC., World Bank, 1989. 'This is the lowest agricultural growth of any developing region. In Central Africa and much of East Africa there is scope for more agricultural growth by expanding agriculture into underused fertile land. But arable land is under intense pressure in the Sahel, the highlands of East Africa, and Southern Africa. 5 The spread of AIDS is imposing additional demands on already meager health care resources, especially in Eastern and Southern Africa. As a result, the social fabric of communities is under pressure, new social problems such as orphaned children are arising, and the economic burdens of women are growing. 4 III. The Collection of Information on Poverty and the Analysis of Poverty 1. The Collection of More Information on the Poor Will Help to Strengthen the Poverty Focus of Policies and Investments. The poor in Africa typically depend on agriculture or the informal sector for their livelihoods. Policies and programs that benefit these sectors, therefore, potentially benefit many poor people. For this reason, the argument is sometimes made that collecting more information about poor households and communities, and using this information to undertake more analysis, will not greatly increase the poverty focus of policies and programs that are already of benefit to agriculture and the informal sector. However, an important argument in favor of collecting more information arises out of the complexity of the livelihoods of both rural and urban communities. Although many small farmers are poor, their incomes nevertheless vary considerably - even within the same agro-climatic zones - reflecting differences in access to productive assets, education, and market opportunities. This is true both in countries with commercialized agricultural systems (see Box 1 on Zimbabwe, for instance), and in areas (such as the Sahel and the Horn of Africa) where average incomes are among Africa's lowest. The incomes of informal entrepreneurs also vary widely from those able to use their skills and market access to produce high-value goods and services (many of whom consequently earn more than formal sector employees) to those without education, capital, or market opportunities. Box 1 - Zimbabwe: Inequalities in Household Access to Productive Assets within a Rural Economy. Rural Zimbabwe illustrates the variation in incomes that can exist within Africa's rural economies. While per capita income is clearly higher in high-potential regions than in low-potential regions, incomes vary considerably within ecological zones, districts within zones, and villages. In one area studied by the 1991 Agricultural Sector Memorandum, per capita incomes in the highest income quartile are 500 percent greater than in the lowest quartile. Reaching the poor, therefore, requires more than simply assisting the low-potential regions. Access to productive assets accounts for much of the income inequality; the landholdings per capita of low-income households are half those of high-income households, and they have significantly less equipment and fewer draft animals. As a consequence, the grain sales of high-income households are ten times those of low-income households. Those households that have insufficient land for subsistence rely on wage income from employment on commercial farms. They are the poorest people in their communities, and malnutrition among their children is high. Moreover, high-income households have four times more nonfarm income per capita, thereby giving them more resources for making farm investments and making it easier for them to bear the risks to their incomes of adopting high-yielding seed varieties. This study shows the importance of programs to enhance the productive assets of the poor. Measures to diversify their nonfarm incomes - such as more and better education - will also help to provide investment funds for their farms. Source: Zimbabwe: Agriculture Sector Memorandum, Report No.9429-ZIM, May 1991. 5 It cannot, therefore, be assumed that policies and programs that have direct or indirect benefits for agriculture and the informal sector will always result in large reductions in poverty. Much depends on the severity of the constraints which face the poor. The collection of information on poverty can identify those constraints on the livelihoods of the poor which it is feasible to reduce, thereby assisting the design of measures to increase their gains from policies and programs of benefit to agriculture and the informal sector. Such measures can include improving their access to productive assets (through land-tenure reforms and credit schemes), the availability of inputs which are essential to their enterprises (through improvements to distribution systems), and their participation in the markets for their goods and labor (through improved marketing facilities and the elimination of labor market distortions). The collection of information on poverty by region is especially important for policy design. Large differences in living standards across regions are to be found in most African countries. Some of this regional variation is accounted for by differences in agro-climatic conditions which determine the crops that farmers can grow, and therefore their incomes. But living standards are also strongly influenced by the availability of economic infrastructure (particularly transport infrastructure which determines the access of a region to national and international markets) and the provision of education and health facilities which influence the returns to household labor. The poor are often disproportionately concentrated in regions having the least economic and social infrastructure. The collection of regional poverty information enables analysis to be undertaken of the ways in which limited access to infrastructure contributes to household poverty. Such analyses can help to direct the regional allocation of public investments to the benefit of the poor. These investments could include improved transport and marketing facilities to increase the access of the poor to markets for their outputs and inputs. Improvements in a region's education facilities can help the poor to improve their incomes from nonfarm employment, and can enhance their mobility to more prosperous regions. Government policies with regard to agricultural marketing and the distribution of consumer goods and inputs can also have strong effects on living standards across regions, often to the disadvantage of remoter regions, which contain high concentrations of poor people.- Information on households and communities facilitates the analysis of the poverty impacts of these policies, and can thereby contribute to their reform. The third argument in favor of collecting information on poverty is the need to incorporate the needs of women into policy and program design. Women are disproportionately found among both the rural and urban poor. The gender division of labor within agriculture often confines women to subsistence food production or limits their income from cash crop production. Their possibilities for taking advantage of new opportunities are often limited by discrimination in the markets for credit and inputs, and in the allocation of land within communities. Within the informal sector, women are often confined to the least profitable occupations, which reflects their limited access to education and vocational Iraining. The disaggregation of information by gender is therefore essential if the needs of women are to be incorporated into the redesign of policies and programs. When a policy or program has potentially large benefits to the poor, investing in information collection at regular intervals to monitor its impact (and to enable subsequent improvements to be made to it) can be justified. The regular collection of basic anthropometric indicators is often sufficient to indicate trends in the welfare of communities in those areas for which it is particularly important to understand the impact on the poor of policies and programs. In summary, there are strong arguments in favor of the view that the collection of information on poverty is not merely of academic interest, but is an exercise that is likely to improve our understanding 6 of the constraints that poor people face, thereby helping the design of policies and investments intended to reduce poverty. Thus it merits the investment of country and donor resources. 2. Progress is Being Made in Providing Household and Community Information. Helping African governments to build an information base on households and communities presents a formidable challenge. The limited capacities of most statistical offices - due to years of underfunding - make it difficult to set up and sustain data collection efforts. The Bank has contributed to a range of capacity-building exercises aimed at making national statistical systems more efficient.' Despite the adverse circumstances, household surveys supported by the Bank have yielded results, albeit more slowly than planned. Living Standards Measurement Surveys (LSMS) have been conducted in COte d'Ivoire, Ghana and Mauritania." Such multi-topic surveys provide a better information base than was possible with traditional single-topic surveys, which do not allow the analysis of poverty as a multi-dimensional phenomenon., Under the Social Dimensions of Adjustment (SDA) program, two types of household survey have been developed and implemented. The Integrated Survey is designed to collect, for a relatively small national sample, comprehensive data on incomes, expenditures, assets, and indicators of health and welfare." The integrated survey enables a detailed picture of poverty to be constructed. The second type of survey is the Priority Survey.'o Priority surveys are designed to be implemented quickly, and are suitable for statistical offices with limited capacities. To achieve these objectives, priority surveys collect less information than integrated surveys by focusing on key socio-economic indicators. These indicators include the education and employment of household members, key items of household expenditure such as food, sources of income, and anthropometric information. Data from priority surveys is now available for some 11 countries (see Box 2 for the example of a priority survey in Chad). 3. Poverty Analysis Requires Flexibility in the Choice of Information Collection Methods. The development of the integrated and priority surveys was a first step in providing a wider choice of instruments for data collection. Making available a selection of survey instruments enables Bank 6For ways to help national statistical systems focus on the needs of users see M. Gallagher and T. Marchant A Strategy for Enhancing Statistical Capacities and Information Management in Africa, World Bank, 1993. 7 LSMS household surveys provide information on a broad range of topics, including household income and expenditures, education, household composition, sources of income, employment, health and nutrition, and fertility. 8 Refinements to the design of the LSMS are considered in C. Grootaert The Evolution of Weyare and Poverty Under Structural Change and Economic Recession in C6ted'Ivoire, 1985-88, Washington D.C., World Bank, Policy Research Working Paper WPS 1079, 1993. 9 See G. Delaine, et al. The Social Dimensions of Adjustment Integrated Survey, World Bank, SDA Working Paper No.14, 1993, for an overview of the design of integrated surveys. 10 See C. Grootaert and T. Marchant The Social Dimensions ofAdjustment Prority Survey, World Bank, SDA Working Paper No. 12, 1993. 7 support for data collection to be tailored to Box 2 - Oiad. Collecting Priority Data on Urban the capacities of statistical offices. In Poverty. providing such support it is important to focus data collection on the main issues of A successful household survey has been importance to policy and program design. conducted in Chad despite the fact that it has ine of The link between data collection and policy the weakest statistical capacities in the world. Civil and program design can be strengthened in conflict and drought have combined to swell the three ways. population of Chad's capital city (N'Djamana), * By using policy priorities to narrow putting intense pressure on urban employment and Bywn us ang o prriesfservices. In 1991 a priority survey for N'Djaniens down the range of variables for which estimated that 20 percent of all households are poor, data are collected in multi-topic surveys. that 26 percent of households with informal This will free resources to increase the livelihoods are poor, and that 7.5 percent of frequency of surveys, thereby providing households with public sector incomes are poor. more insight into poverty trends. Some 8 percent of all children (under 5 yearm) are . Bmalnourished; malnutrition is highest among children * By increasing the use of fast mucro- from informal sector households (10.5 percent of all surveys that focus on specific policy children from such households), and more boys are issues. The rapid overview of poverty malorished (10.2 percent) than girls (7.5 percent). that was carried out in Ethiopia is one Overall, a very serious nutrition problem has example (see Box 3). Another is the emerged in urban areas. rapid employment survey in Malawi, which has led to a revision of the The primary enrollment rate is 53 percent for minimum wage policy (see Box 9 boys and 45 percent for girls, although illiteracy is below). higher for men (60 percent) than women (39 percent). Among informal sector households * By attaching program components and illiteracy is very high (61 percent), indicating that the loan conditionalities to sector operations primary enrollment of poor children is low. requiring the implementation of Increasing the enrollment of poor children may household and community surveys to improve informal livelihoods, if the curriculum is monitor the operation's impact on appropriate and the education-productivity bak is poverty. strong. In addition to quantitative information on poor people (their incomes, expenditures, household characteristics, etc.), the collection of more qualitative information on poor households and communities improves our understanding of the determinants of their livelihoods. Such qualitative information includes the views of poor people themselves regarding their situation and how to improve it, their expectations regarding policy changes, and their assessment of the effectiveness of development programs to them. Over the last two years, the use of techniques of rapid appraisal has grown within the Bank. This trend began with the use of Beneficiary Assessments to provide a more thorough understanding of the livelihoods of communities, and to help to design and evaluate the impact on poor people of projects and sector operations." Rapid appraisal techniques are now expanding on this tradi- tional focus to include assessments of the extent and causes of poverty, as well as of the link between existing policies or policy reforms and poverty. This will help to incorporate the poverty reduction nSee L. Salmen Reducing Poveny: An Institutional Perspective, World Bank, Poverty & Social Policy Series Paper No.1., 1992 and L. Salmen Beneficiary Assessment: An Approach Described, Poverty & Social Policy Division, Working Paper No.1, 1992. 8 objective into the design of policies and programs." The technique of participatory poverty assessment is now being used to support Bank operations in both Ghana and Zambia (see Box 4). Box 3 - Ethiopia: Rapid Poverty Assessment. At the request of the transitional govemment in Ethiopia, the Bank has undertaken a rapid assessment of poverty to improve the information base on urban poverty. A small survey of households in two poor neighborhoods in Addis Ababa was conducted. This yielded results within one month. Using the survey's results, together with additional information on the urban poor, the Bank mission estimated that some 4 million people living in urban areas are below the poverty line. This is about 58 percent of the urban population. Estimating poverty in rural areas is even more difficult in Ethiopia given the problems created by the aftermath of war. To provide an estimate of rural poverty the Bank mission updated information collected during the preparation of the Bank's 1991 Food Security Report, and information gathered by govemment agencies and NGOs. This process resulted in an estimate that over 12 million people in rural areas are poor. This is about 30 percent of all farm households. In addition, some 7.4 million people in rural areas are close to being poor as a result of drought and war. These results are tentative, but they do provide an information base until further data collection is undertaken. On the basis of its findings, the mission was able to recommended measures to improve the employment of the poor, increase their access to social services, and improve the targeting of food assistance. Sourcet Ehiopia Toward Poverty Alleviation and a Social Action Program, Agriculture and Environment Operations Division, Eastern Africa Department, 1993. 4. The Analysis of Poverty Issues is Undertaken in a Wide Range of Bank Reports. Reports which contain analysis of poverty issues include, public expenditure reviews, food-security studies, Women in Development (WID) country assessments, studies of human resources development, SDA country assessment reports, and agricultural sector reports. Country Economic Memoranda (CEMs) also sometimes discuss poverty (CEMs are assessed in sections 6 and 7 below). Each kind of report is discussed below. Public expenditure reviews (PERs). These reports have a major role to play in redirecting public provisions to the poor. They are a key input into the dialogue with governments and into the design of major Bank operations such as SALs. To exploit this instrument to the full, they need to have a clear poverty focus. 12 Recent topics include: the demand for rural education in rural Mali; village access to services in Malawi; monitoring the impact of reform and the Zambia Social Fund; the determinants of fertility, and access to family planning for the Afican Agenda on Population; access of the poor to health, education, and agriculture services in the Central African Republic; informal township enterprises in South Africa. 9 Box 4-Participatory Poverty Assessment. The objective of a participatory poverty assessment is to ensure that the experiences and concerni; of poor people are incorporated into the design of policies and programs aimed at helping them. Participatory assessments have a wide scope. They can help policy-makers understand how poor people perceive their problem and its causes, how they seek to reduce their vulnerablity through informal safety nets, their perceptions regarding the impact upon them of reforms in economic incentives, and their views as to the effectiveness of measures intended to help them. The methodology of participatory assessment draws on earlier work within the Bank on beneficiary assessment, together with the now well-established techniques of rapid appraisal. The cornerstone of the methodology is the conversational interview in which local researchers, having been briefed on the issues of policy importance, engage individuals from selected communities in discussions regarding their circumstances. The researchers do not administer a pro-determined questionnaire to which individuals respond. Instead, the interviews are semi-structured, in that the researcher has an agenda of issues to discuss, but not a complete set of questions. The information gathered by this means is both qualitative and quantitative in nature. Such interviews can also be conducted with groups of people ("focus groups") rather than with individuals, in order to gather the views of communities on key issues. When participant observation is used, researchers undertake a period of residence in the community, enabling discussion of the issues, together with observation of community life, to occur. Finally, interviews can be conducted with institutions such as NGOs engaged in helping the poor. PERs often draw on studies of human resources development for their discussion of social expenditures and the poor. This gives PERs a useful focus on the education and health of the poor. But, in using the results of those studies, there is a tendency for PERs to downplay some of the important results of the education and health-sector studies. For example, PERs tend to assume that increasing spending on basic services will result in proportionate benefits to the poor. However, they put little effort into assessing the incidence of actual spending and on establishing ways to achieve and track progress. They should take account of the analyses by sector reports of the demand side constraints - especially the opportunity costs of foregone income - that can limit the usage by poor people of the available services.13 A major gap in the poverty focus of PERs is their neglect of the importance of improving the availability and effectiveness of public investments in the economic infrastructure which serves poor communities. As a priority, PERs should assess the impact on poverty of public infrastructure investments. Understanding the impact on households and communities of investments in agricultural infrastructure and supporting services (such as transport) is especially important given the potential of these investments to improve rural livelihoods. As a first step, correlating the ranking of regional living standards (measured by per capita incomes when available or more indirect indicators otherwise) with the regional expenditure allocations can be undertaken. A second step is to deploy small surveys and rapid appraisal techniques during the preparation of the PER, focusing these measures of information collection on the poorest regions. These can contain questions to determine the use by poor households and communities of key items of infrastructure and problems in their access. 13 With the increasing availability of household data, it is becoming possible for PERs to analyze the effectiveness of public expenditures in raising the incomes and improving the 'capabilities' of the poor. 10 Analysis of the incidence of consumer and producer subsidies, taxes, and user charges is generally limited in PERs. This is likely to improve as more income/expenditure data become available. Until then, rapid surveys of consumer budgets across areas with different living standards (to establish differences in expenditure shares, thereby providing approximate indications of the impact of subsidies/taxes) would help to fill this analytical gap. Beneficiary assessments have also proven to be useful in analyzing the impact of user fees and are cost effective in collecting this type of information. Food security reports. As a result of the Bank's Box 5 - Nigeria: Assessing the Extent of Food Poverty. Food Security Initiative, food security studies exist for Household income/expenditure data are very limited in Nigeria. some 20 countries. Each The 1991 Food Security Study used evidence from existing small study estimates the surveys, and information on food intake, calorie availability, and food percentage of the population prices to calculate the income required to satisfy the food requirements that has insufficient income of an average sized urban or rural household. These food poverty lines to meet its nutritional show that 22 percent of urban households and 17 percent of rural 14 households spend less than this amount and are therefore food insecure. needs. Since the needsiit Sinathie These are rough estimates given the sparseness of the data. They do availability of data varies nevertheless indicate the existence of significant food poverty. The substantially across seriousness of this situation is confirmed by Nigeria's nutrition status countries, the studies deploy indicators, which are as bad as those of most other SSA counies- for a wide variety of methods to example height-for-age shows an incidence of stunting of between 22 calculate food insecurity (see and 54 percent among children. This study has made an important Box 5 on Nigeria, which has contribution to the Bank's dialogue with the government, in a country one of the most thorough whose governments have not traditionally acknowledged having a assessments). These studies poverty problem. This dialogue is continuing with the poverty have resulted in several assessment that is currently being prepared. projects aimed at improving Source: Nigeria Strategy for Food and Nutrion Secrity, Report food security (examples of No. 9040-UNI, June 1991. such projects in Burkina Faso and Rwanda are given in Box 12 in section IV). WID Assessments. Analysis in WID country assessments focuses on the livelihoods of women (particularly on discrimination in asset and labor markets) and on gender bias in access to public services. Their poverty focus is generally implicit - the incidence of poverty is high among African women - but can be made more explicit, as data to carry out poverty assessments that are disaggregated by gender are collected from new surveys. An example of the policy recommendations which emerge from a WID study is given in Box 10 in section 9 below. Human Resources Development. The number of studies of education, health, population and nutrition has grown rapidly in parallel with the major expansion of lending for human resources development (HRD) over the last three years. They are given an implicit poverty focus by their analysis of the 14 The 1988 Report of the World Bank Task Force on Food Security defines food security as "access by all people at all times to sufficient food for a healthy and productive life." Households below the poverty line are usually food insecure in either normal or abnormal years. Bad harvests may push some households that are normally above the poverty line below it; after large food price increases, their incomes become insufficient to purchase enough food. 11 composition of public expenditures and the constraints on the demand side which can limit the use of services. This is made explicit when household and community level data are analyzed. The link between HRD studies and policy is discussed further in section 6 of part IV. SDA Country Assessments. SDA country assessments focus on the impact of adjustment on poverty, covering both the livelihoods of the poor and the incidence of public expenditures. Among the most thorough are those for Central African Republic, Madagascar, Senegal and Togo, in which existing information is collated to construct a basic poverty profile, thereby providing the foundation for a more comprehensive poverty assessment. Agriculture Sector Studies. Agricultural sector memoranda analyze the sector in which most of the poor make their livelihoods. Their focus on issues of agricultural pricing policy, infrastructure, support services, and food-security therefore has significant bearing on the problems of rural poverty. Since the poor are frequently concentrated in particular regions (often semi-arid) and grow a limited range of crops, discussions of regional and crop issues have an implicit poverty focus. Some of these reports make the poverty focus more explicit by disaggregating the broad socio-economic category of small farmers. The memoranda for Uganda and Zimbabwe, for example, contain in-depth discussions of poverty issues based on disaggregated information." Taken together, these different types of study constitute a large information base on poverty and related issues. While individual reports concentrate on defining priorities within their sectors of concern, priorities at the country level across the topics and sectors tend to be left undefined." These studies tend to be topic or sector specific and thus other issues are incorporated only when they are relevant to the main issue. For example, food security studies will cover household income issues only when discussing the ability of the poor to purchase food. HRD studies discuss household income only as a determinant of the demand for education and health services. WID studies discuss household income only as it arises in the gender division of labor. Similarly, it is possible to consider public expenditures solely from the perspective of targeting food subsidies, targeting social services, or targeting expenditures towards women. 5. Poverty Assessments will Help to Identify Cross-Sectoral Priorities. It is clear that there is a need for an overview of poverty at the country level, which sets priorities for action across sectors and identifies complementarities and trade-offs between actions. This is important both as a focus for Bank-government dialogue and as the basis of a country strategy for poverty reduction to guide operations. This is the task of the poverty assessments, the preparation of which is now underway across all of the Bank's regions. The process of preparing a poverty assessment is as important as the report itself, since it provides a framework within which the Bank and governments can work towards a consensus on the actions that need to be taken. is Uganda: Agricultural Sector Memorandum, Report No. 10715-UG, June 1992. 'o Until the introduction of poverty assessments, the SDA country assessments came closest to providing a single overview, as did those HRD studies that adopted a very broad scope (the studies for Kenya and Malawi are examples). However, these studies have been necessarily limited in their focus, since they emerged out of specific initiatives, rather than constituting broader country reviews. 12 Of all the Bank's regions, Africa presents some of the most difficult problems in preparing poverty assessments. First, the region contains a large number of countries, each of which requires a separate Bank-government dialogue during the preparation of the poverty assessment, as well as the information collection and analysis. Taken as a whole, the exercise is very intensive in Bank staff time. Second, in many countries, the government is not fully committed to reducing poverty. Careful dialogue is required to convince governments of the importance of establishing agendas for action in reducing poverty. Third, only limited information is available. To improve the information base, surveys and participatory poverty assessments are now underway in countries such as Ghana, Madagascar, and Zambia. Previous single- topic studies - those on food security and human resources development especially - provide good sources of information and analysis, and some of the SDA country assessments (that of the Central African Republic, for example) are being updated to produce poverty assessments. Some 8 poverty assessments have now been completed for SSA." Some 16 Poverty Assessments are scheduled to be completed during FY 1994. By FY 1997, a total of 35 poverty assessments are scheduled for completion."' The value of poverty assessments in guiding operations is demonstrated by the two earliest poverty assessments, those for Malawi and Mozambique." The poverty assessment for Malawi was one of the Bank's first, reflecting the early priority given to poverty and human resources development in the Bank's Southern Africa region. It helped set the standard for later poverty assessments in other countries. Subsequent lending to Malawi has included an explicit poverty focus (especially in operations in agriculture and human resources development) building on the priorities identified in the poverty assessment. The Mozambique poverty assessment is less comprehensive - it was prepared as documentation for a consultative group meeting - but provided a useful framework for dialogue and set the agenda for later in depth studies such as the food security report which have guided subsequent country operations. 6. Use of Household-Data Analysis in Operational Analyses is Now Increasing. Given the desirability of collecting household data to guide policy, an important question is how far the earliest Bank-supported household surveys in Africa - the Living Standards Measurement surveys (LSMS) for C6te d'Ivoire and Ghana - have been used in key documents such as CEMs, PERs, and reports on HRD, which underpin the Bank-government dialogue.' Evaluating how useful the LSMS for COte d'Ivoire has been for operations is problematic given the slowdown in country operations (and therefore economic and sector work which could use the data) since " Poverty Assessments have been completed for Ethiopia, The Gambia, Ghana, Malawi, Mali, Mozambique, Sierra Leone, and Uganda. I8 AF2 (Eastern Africa) has prepared short poverty status reports for some countries until their poverty assessments become available, a useful initiative that should be followed in other areas of the Bank. ' Malawi: Growth Through Poverty Reduction, Report No. 8140-MAI, March 1990, Mozambique Poverty Reduction Framework Paper, October 1990. a The Mauritania survey is more recent, and the results of the data have not been available long enough to judge how useful it has been. 13 the mid-1980s. There is no recent Box 6 - CMte d'Ivoire: Using Household Data to Carify the CEM, for instance. The 1990 PER Benefits of Social Expenditures to the Poor. makes no reference to the studies -.. using the LSMS data which were then In C6te d'Ivoire, the share of health and education in available, but relies mostly on the government expenditures is one of the highest in the world (50 functional and geographic percent of recurrent expenditures), but C6te d'Ivoire's social distributions of services to discuss indicators are significantly worse than those of countrie with issues of access (the standard method similar per capita incomes. This discrepancy is explained by for countries lacking household the low priority given to primary education and basic health- data).21 The survey for COte care in expenditure allocations. d'Ivoire appears, therefore, to have had limited value for the PER, Analysis of LSMS data for C6te d'1voire (collected in perhaps because social expenditure annual surveys over 1985-87) shows that existing social allocations in Mte dIvoire are so expenditure allocations have few benefits for the poor. Whereas 68 percent of households in Abidjan consult a health care distorted towards services that benefit worker, only 36 percent of those in the Savannah region (C6te the non-poor, that this information is d'lvoire's poorest region) do so. Preventative health care is sufficient to indicate the necessary underfunded in all regions; only 32.5 percent of households in reforms. The 1991 project document the Abidjan region and 9.9 percent of households in the on human resources development Savannah region have access. The data reveals similar makes more use of the data (see Box inequalities in education; the primary enrollment rate ia the 6) to reinforce conclusions reached by Savannah is only 30 percent. analysis of the expenditure allocations and to emphasize the fact that use of Subsequent surveys indicate that the situation of the poor public education and health by the deteriorated over the period. Only 40.7 percent of poor ill- poor is declining. people consulted a health worker in 1985, but this declined to 33.8 percent in 1987, while the proportion of poor 6-11 year Operational use of the studies of olds enrolled in primary education fell from 38.7 percent in the LSMS for Ghana lagged 1985 to 33.3 percent in 1987. These data have been used to somewhat behind the availability of reinforce the 1991 HRD Project which contains conditionAities the studies themselves. The 1991 on shifting resources to primary education and basic healih. Ghana CEM does not fully use the han pvEy pofe, ptiulrly the Source: C6te d'Ivoire Human Resources Development regioA poverty drofisagprtiou y te Program. Report No. P-5653-IVC, November 1991. regional poverty disaggregations or in the discussion of adjustment's poverty impact.' The 1990 PER makes no use of the data.m Use of the data is now increasing, starting with the recently completed 'Accelerated 21 C6te d'Ivoire: Public Expenditure Review, Report No. 8748-IVC, July 1990. The studies using ISMS data are distributed through the LSMS working paper series. 2 Ghana: Progress on Adjustment, Africa Region, Report No.9475-GH, April 1991. E.O. Boateng, K. Ewusi, R. Kanbur, and A. McKay A Poverty Profile for Ghana, 1987-88, SDA Working Paper No.5, 1990. This uses the first year of the Ghana Living Standards Survey (GLSS) for 1987-88. The LSMS profile (which also uses 1987-88 data), is published as P. Glewwe and K.A. Twum-Baah The Distribution of Welfare in Ghana, 1987-88, LSMS Working Paper No.75, 1991. 2 Ghana: Public Expenditure Review, Report No.8616-GH, June 1990. 14 Growth Study,' which is a key document in the Bank-government dialogue, and forthcoming studies for the next poverty assessment in Ghana, which will draw extensively on the LSMS survey.24 7. More Comprehensive Analyses of the Effects on Poverty of Market Reform Can Improve the Poverty Focus of Policy Reforms. Analysis of the macroeconomic and sector impact of policy reform appears regularly in both operational output and in-depth studies. This has not been matched by a comparable amount of analysis of the impact of reform on poverty. This relative neglect results from the assumption - widespread across the Bank - that policy reform is generally pro-poor and that, therefore, no additional analysis is needed. Until recently, most CEMs have made very general statements about reform's social impact, tending to conflate groups who are most vocal in their demands - public employees, especially - with the poor. Box 7 - Tanzania: Analyzing Adjustment's Rural-Poverty Impact. Given insufficient household incomelexpenditure data, the 1991 CEM for Tanzania uses indirect indicators to analyze adjustment's impact on poverty. The income from crop sales of small farmers (the group containing the most poor) did not increase over the period of the reform (1986-89), nor did their per capita consumption of non-food goods or their real cash balances (both of which indicate that their cash incomes did not increase). However, rural food consumption rose by 50 percent over 1983-88, as the liberalization of food marketing encouraged more production and real food prices fell (which benefited food-deficit households). The relaxation of consumer goods rationing also improved welfare. The CEM concludes that reform arrested the rapid decline in rural incomes that occurred in the years prior to 1986. However, except in the first year after liberalization, rural incomes have not increased significantly. Real income from export cropping has risen only modestly, since traditional export crops remain under monopsonistic marketing boards, which have absorbed much of the gain from devaluation. Liberalizing export marketing is therefore a priority not only to generate foreign exchange, but also to reduce poverty. Source: Tanzania Economic Report: Towards Sustainable Development in the 1990s: Volume 1 (Main Report), Volume H (Bacground Papers), Report No. 9352-TA, June 1991. The situation is now improving. Recent CEMs take a comprehensive look at the impact of reform on poverty, including the 1991 Tanzania CEM (see Box 7), and the 1992 Uganda CEM (which is also a poverty assessment).' This is a welcome development. Although a sound case can be made that -policy reforms reduce poverty in Africa - especially in the early stages when efficiency gains recover lost output - modifications to the design of reforms should still be explored to make them even more effective in reducing poverty. 2 Ghana 2000 and Beyond. Lessons of International Experience for Accelerated and Equitable Growth, West Africa Department, 1993. 2s Ugandw Growing Out of Poverty, World Bank Country Study, 1993. 15 Analyses that look closely at poverty and reform can be useful in guiding operations in six areas: * by increasing the productive assets of the poor (through projects and redirections of public investments) so that they can take advantage of changes in incentive policies (such as increased producer prices for farmers). * by prioritizing actions to remove policy-induced distortions in land, credit and input markets (such as providing new market institutions and facilities and introducing legislation). * by adjusting agricultural reforms, for example, through tighter and earlier loan conditionalities to ensure that output markets of particular importance to poor farmers are liberalized faster. * by modifying trade reforms to prioritize the liberalization of imports of essential inputs used by small farmers and informal entrepreneurs, and by prioritizing improvements in the efficiency of domestic manufacturers who also produce such inputs. * by prioritizing improvements in wage and employment legislation to reduce labor market segmentation, thereby reducing the crowding out of women and young people from the formal labor market. * by targeting input subsidies to poor farmers as 'second-best' measures when first-best solutions (liberalization and institutional reform) are delayed or are not fully effective, and by using price support mechanisms when monopsonies persist in product markets and minimum-wage mechanisms when monopsonies persist in labor markets. Each of these issues requires analysis at the country level to determine the pay-off from focusing more on the poor. The purpose of more analysis should not therefore be to take action in every one of these areas, but rather to guide policymakers in selecting those areas with the most potential for reducing poverty. Given that policy reform is the centerpiece of the Bank's activity in Africa, attaching to operations more conditionalities in favor of the poor (after analysis to estimate their probable effectiveness in reducing poverty) has considerable potential to increase the poverty focus of operations. Such conditionalities are further discussed in section 3 of part IV below. 8. More Thorough Assessments of the Impact of Market Reforms on the Poor can be Undertaken with Existing Information. Given the scarcity of information, it is important to make the most of what is available, until more information on households and communities is collected. To strengthen the analysis of the impact of reform on poverty, three existing sources of information can be exploited more fully. * Survey data collected by governments, development agencies, and academics. Such studies are often small scale, but careful sifting of the evidence they present can build a picture of trends.' * Nutrition and health indicators collected by other development agencies (for example, UNICEF) and NGOs, using these as 'output' indicators of changes in household incomes. * For example, the 1991 Tanzania CEM uses data on household living standards collected by the Centre for the Study of African Economies at Oxford University, and the Bank's work in Mozambique has drawn on studies of economic reform and food security prepared by the Food and Nutrition Policy Program of Cornell University. 16 * Indirect economic indicators such as agricultural value added in the crops grown by the poor (and disaggregated by regions and districts), terms of trade indices based on those outputs and inputs of most importance to poor farmers (disaggregated by region if regional market price information is available), real wages for unskilled urban and farm labor, and movements in key consumer prices for such commodities as food and clothing.' Box 8 - Togo: Assessing the Impact of Agricultural Policy Reform. Poor farmers grow cotton alongside food crops, and the country's two poorest regions produce 32 percent of total cotton output. Agricultural reform, in raising cotton and food prices, has benefited poor producers. Collating anthropometric data collected by NGOs shows that rural malnutrition has declined in the 1980s, even in the poorest regions. The evidence, therefore, suggests that agricultural reform has had a favorable impact on the rural poor, although the welfare of the very poorest - who rely on subsistence food production and have no marketable surpluses - has changed little either way. improving the productive assets of the poor, and their productivity, could do much to spread the benefits of price reforms. Source: Povey Profile of Togo: Country ALsessment and Policy Issues, Report No. 7825-TO, April 1989. The SDA Country Assessments for Madagascar, Senegal, and Togo (see Box 8), a number of food security studies (that of Nigeria for example), and the 1993 Uganda CEM, all use data from sources like these, particularly information on real prices. When income/expenditure data are available for only one or two years, which is usually the case, other indicators are needed to provide some analysis of probable trends in living standards (since the survey data will provide only poverty cross-sections). 9. More Analysis of the Relationship between Employment and Poverty is Needed to Strengthen the Benefits to the Poor of Policies and Programs. Analysis of the relationship between issues of employment and poverty is to be found in several different types of report produced within the Bank. These include studies of human resources development, food-security studies, and women in development (WID) studies.' Each takes a specific focus on the employment issue. Reports on human resources development usually concentrate on the returns to education in the labor market. Their discussion of poverty is generally implicit in their treatment of the market for unskilled labor. The 1991 report on human resources development in Kenya, which is a comprehensive analysis of the issues, is one example of this approach.' The 1990 report for Malawi, which discusses rural and 27 See the World Bank policy paper Assistance Strategies to Reduce Poverty, 1991, on the interpretation of these variables. 2 The Bank has conducted evaluations of public employment policy in a number of countries. Since most public employees are non-poor, these studies are not considered here. * Kenya Human Resources: Improving Quality and Access, Report No.9023-KE, June 1991. 17 Box 9 - Malawi: Improving the Labor Market for the Poor. Malawi's growth has been based on estate agriculture, which has benefited from favorable prices and large investments. The estates are the largest employers in the rural labor market, and such employment is one of the few available wage sources for the landless (including migrants from the poorest areas) and food-deficit farmers. Moreover, estate workers do not have an organized forum for bargaining in contrast to the strong associations of employers. As a consequence, employers have traditionally forced wages below competitive levels, thereby exacerbating poverty and malnutrition. In such situations, setting a minimum wage at a level consistent with a competitive market. and therefore above the wage offered by monopsonistic employers, improves welfare. The governient's pohcy has been to set rural minimum wages at levels equivalent to those generated by competitive labor markets. A Bank-assisted study concluded that the 1989 adjustment to the rural minimum wage was, in approximating a market solution, not detrimental to rural employment. Follow-up assessments by the Bank have concluded that minimum-wage levels have been revised too infrequently to make them effective in altering the behavior of employers. The 1991 Entrepreneurship Development and Drought Recovery Program commits the government to review labor market imperfections and minimum-wage legislation to ensure that the real value of the minimum wage is maintained. Source: Malawi Entrepreneurship Development andDrought Recovery Program, Report No. P-5461- MAI, June 1992. urban employment in detail, is a model of how to analyze poverty-employment links, and has had a major impact on operations in the country (see Box 9).3 Food security studies regard employment as a means of generating income with which households can buy food; those for Mozambique and Nigeria are good examples of in-depth analysis of income in determining "entitlements" to food. Rural employment sometimes features in agricultural sector memoranda; the Uganda study is the most extensive, being based on a detailed study of the rural labor market.1 WID studies pay specific and consistent attention to the gender dimensions of employment (see Box 10 on Zimbabwe, for example), but the focus on poverty is generally implicit when data are unavailable to decompose employment patterns and payments by gender. Official employment statistics concentrate on formal occupations which are dominated by men. Since the employment of women is concentrated in household enterprises (both farm and nonfarm), information collection to understand the determinants of the demand and supply for female labor is especially important. These studies amount to a significant body of work. However, although the Bank's strategies for poverty reduction frequently urge the need for labor-intensive growth to reduce poverty, many are un- specific about how to accomplish this. 3 Malawi: Human Resource Development Study, Report No.7854-MAI, April 1990. 31 Uganda: Agricultural Sector Memorandum, Report No. 10715-UG, June 1992. 18 Box 10 - Zimbabwe: The Informal Sector and the Role of Women. Informal employment in Zimbabwe constitutes less than 10 percent of total employment, a much smaller share than the SSA average of 21 percent. A small survey conducted by the Bank - focusing on women who comprise two-thirds of those in the informal sector - confirmed that entrepreneurs are discouraged by minimum-wage and labor regulations together with over-zealous business licensing. In addition, inadequate access to credit (due to weak collateral) and materials/equipment (due to the import licensing system that favors large firms) have limited the expansion of the informal sector. Removing such barriers to informal cntrepreneurship will make it easier for the sector to absorb a rapidly growing labor force into remunerative employment. Accordingly, the Bank recommended that minimum wage and labor regulations for enterprises with ten employees or less should be relaxed, that there should be more flexibility in business licensing, that lending by the Small Enterprise Development Corporation (SEDCO) should be extended to informal entrepreneurs, that suitable sites for informal enterprises should be provided and equipped with child care facilities, that training should be expanded, and that appropriate technology should be imported freely without license. Since workers in the unskilled and semi-skilled informal occupations tend to be poor women, efforts to expand their facilities and skills will help to reduce the extent of female poverty. Sources: The Informal Secor in Zimbabwe: The Role of Women, Report No.9006-ZIM, August 1991. Zimbabwe. Strategy for Women in Development, Report No.9204-ZIM, March 1991. If employment-poverty links are not fully analyzed, this will diminish the quality of the Bank's advice to governments in the dialogue on poverty. Although farm self-employment will dominate total employment well into the foreseeable future, more analysis of poverty and nonfarm employment is important for the following reasons: * In order to reduce poverty rapidly, Africa must follow Asia into export manufacturing. How to maximize job creation for the poor, given the impediments of labor market segmentation and low quality education systems, is a major challenge for Africa in the 1990s. The Bank must ensure that it offers appropriate advice and programs by giving its industrial sector analyses a clear poverty focus. * In order to follow the path taken by Asian countries, Africa's nonfarm rural enterprises must expand to diversify the source of people's livelihoods. Understanding how agricultural reform improves not just farm incomes but also nonfarm opportunities for the poor can guide rural project investments, particularly in infrastructure. * With rapid urban population growth (and allowing for some reversal in traditional rural-urban migration patterns as rural incomes improve), most poor job-seekers in urban areas will start basic microenterprises or engage in casual labor. Understanding how the relaxation of microenterprise con- straints helps those microentrepreneurs who are poor (as well as microentrepreneurs in general) can guide project investments and legislative reforms. * The spread of the benefits of reform from high potential to resource-poor regions by means of migration can be constrained by markets, cultures, legislation, and social and physical infrastructure (especially urban housing and sanitation). Using appropriate policies and investments to eliminate or circumvent these problems requires an understanding of the determinants of migration and how 19 (or whether) the poor gain as migrants or from remittances by members of their households who do migrate. * Countries with especially limited natural resources (the Sahel, for instance) will face the most difficulties in achieving agricultural growth rates of 4 percent per annum on a sustained basis (the target of the Bank's Long Term Prospective Study). Given an average of 3 percent population growth in the Sahel region, it is especially important that the Bank should be able to assist these countries to maximize the growth of off-farm employment opportunities. * In areas where land is under intense pressure due to rapid population growth (for example, Southern Africa and the East African highlands) poverty will increase rapidly unless opportunities for nonfarm employment expand. Each of these points is especially important in the Bank's aim to increase the opportunities that are available for women in poverty. The livelihoods of Asian women are now being transformed as rural off-farm employment expands - the result of the productivity revolution in Asia's agriculture -- and as women are absorbed into export manufacturing. Asia shows that increasing women's access to productive assets and education are crucial to achieving the equitable distribution of the benefits of growth. Similar opportunities will arise in Africa to reduce female poverty if structural transformation of the region's economies can be achieved. IV. World Bank Supported Policies and Programs 1. Country Poverty Reduction Strategies should Strengthen their Focus on Ways for Improving the Economic Opportunities of the Poor and for Enhancing the Benefits that Reform Bring; to the Poor. At present, the poverty reduction strategy in most country strategy documents consists of a repeat of the WDR90 themes of economic opportunities for the poor, increased access to social services, and effective transfers/safety nets. While these themes constitute the Bank's poverty strategy across the developing world, their restatement in country strategies could be more effective if it included a set of country-specific priorities to guide policy dialogue and operations in each area of poverty reduction. As poverty assessments become available it will become clearer which priorities are being chosen in each country. Using the strategy outlined in WDR90 as a framework, it is possible to evaluate the strengths and weaknesses of the current strategies for poverty reduction of the departments comprising the Africa region of the Bank. It is apparent that they tend to prioritize the delivery of social services to the poor (because they aim to reallocate resources to primary education and basic health), and they vary in the degree of emphasis that they put on public works programs (these are given the most priority in the Sahel's strategy). Where they are weakest is in defining ways to improve the economic opportunities of the poor, beyond the overall reforms and investments supported on growth and efficiency grounds. Several of the strategies regard structural adjustment as being equivalent to the poverty reduction strategy, as the objective of adjustment is to restore growth and growth reduces poverty. Although the evidence shows that adjustment improves the lives of most of the poor compared to the absence of 20 adjustment, this cannot be considered to be a sufficient strategy unless the feasibility of ways to enhance this beneficial effect have been investigated. It is clear from most of the country strategy documents that the Bank has not yet done enough analysis to be able to conclude that no further steps can be taken to help the poor beyond the adjustment process itself. Typically, the strategy documents spell out the benefits of reform to, for example, the agricultural sector, but they leave implicit the impact of reform on rural poverty and they fail to explore ways in which agricultural growth could be enhanced to increase its beneficial impact on poverty. The strategy documents also present labor intensive growth as a key component in accordance with WDR90, but there is minimal discussion of whether employment growth can be increased by altering the reform strategy. 2. Efforts to Improve the Sensitivity of Governments to the Need to Reduce Poverty are Increasing. Dialogue in the areas of human resources development and food security is increasingly being used to bring poverty to the negotiating table. Bank staff also increasingly discuss poverty issues with officials from finance and planning ministries, thereby broadening out this dialogue from its traditional focus on macroeconomics. There are several examples of this trend. In Malawi, the 1990 HRD study and CEM have helped to focus the government's attention on poverty as a 'legitimate' policy issue, resulting in several measures to improve farm livelihoods and to increase wage employment (see Box 9 in section Ell above).3 Dialogue with the Kenyan government on poverty issues has built on the 1991 HRD study, which shows that economic growth and public services have bypassed the most deprived regions.3 In Nigeria, the 1991 Food Security Report (see Box 5 in section M above), and the 1992 Population Report constituted important first steps in initiating a poverty dialogue.' The Ghana 'Accelerated Growth Study' refocused the country dialogue on poverty by highlighting the private and social returns to improving the human capital of the poor. In Zimbabwe, dialogue on HRD has resulted in the geographic targeting of social expenditures, as a way of concentrating public resources on the needs of the poor. No primary school fees are charged in rural areas, and fees are lower in poor low-density urban areas than in wealthier high-density areas. Introducing poverty into the Bank's dialogue with finance and planning ministries - and therefore into the design of structural adjustment - is a new initiative with potentially high returns.' These core ministries are at the center of policy formulation, and their decisions regarding incentive policies and public expenditures are critical to the welfare of the poor. 3 Malawi Human Resources Development Study, Report No.7854-MAI, April 1990. 3 Kenya Human Resources: Improving Qualty and Access, Report No.9023-KE, June 1991. 3 Nigeria. Implementing the National Policy on Population, Sector Report No.9306-UNI, May 1992. 3 To coordinate government and donor poverty strategies, the Bank is encouraging the establishment of poverty units in the finance and planning ministries of Ghana, Mali, and Sierra Leone. These units will help to incorporate the objective of poverty reduction into the formulation of macro-policy. 21 Zimbabwe is an example of a country where the Bank's dialogue on poverty has included the core ministries in addition to the social sector ministries. The dialogue began during the preparation of Zimbabwe's 'Economic and Social Adjustment Programme' (ESAP) in 1990. The Bank assisted the Government to focus the objectives of ESAP on poverty reduction, and to set out proposals for action to reduce any adverse impact of ESAP on the poor. The first structural adjustment credit supported the first phase of ESAP, an Emergency Drought Recovery and Mitigation Project helped maintain food- security during the 1992 drought, and a second structural adjustment credit is now in place.' The dialogue on poverty has helped to identify measures for improving incentive policies - particularly improved crop marketing and reducing controls on informal activity - which will be of especial benefit to the poor. These measures are incorporated as conditionalities in the second structural adjustment credit (see point 3 below). To strengthen this new trend in the Bank's relationship with governments, the poverty dialogue should begin before the stabilization process is started. Initial negotiations are critical in convincing governments that the Bank's commitment to pro-poor policy reform is firm. Since the implementation of macro-reform involves a great deal of the time of both national and Bank staff, the poverty dialogue with core ministries should focus on those instruments that have the highest potential for reducing poverty such as the Public Investment Program. Finally, dialogue is more effective when CEMs and PERs, which are both key consensus building documents, focus on poverty. This in turn requires the kind of analysis discussed in Section III above. 3. Conditionalities Relating to Poverty are now more Frequent in Structural Adjustment Loans. Conditions relating to poverty issues are now more common both as pre-conditions to SACs, SALs, and Economic Recovery Loans and as triggers for the release of second tranches. Since loans to support structural adjustment are the most high profile component of the Bank's lending program, this conditionality acts as a signal to governments that the Bank has increased its commitment to poverty reduction. It also encourages governments to make pro-poor choices. Care should, however, be exercised in evaluating the presence or absence of pro-poor conditionalities in SALs. If the dialogue has been especially successful, the necessary measures may already have been taken prior to the SAL or may be imminent. Sometimes pro-poor policy changes are implemented as a condition for the appraisal of the Bank loan. While adding a relevant conditionality to the loan reminds the government of its commitment to poverty reduction, overcrowding the SAL with too many conditionalities, thereby slowing the implementation of the program itself, should be avoided. With more analysis of poverty, the most effective reforms to help poverty reduction can be identified, and suitable conditionalities put in place. - Preserving and raising social expenditures of benefit to the poor is one of the most suitable conditionalities for SALs. The objective is high-profile, has a potentially strong impact on the poor, and is relatively easy to monitor. Recent SALs for Burkina Faso and Togo have shown how this can be done. The release of the second SAL I tranche in Burkina Faso depends on the government raising the share of expenditures on primary education and health as a proportion of total expenditures. This reinforces the commitment made in the 1990 Policy Framework Paper to improve expenditure management, including making cuts in military expenditures, to end the crowding out of primary education and - Zimbabwe Presidents Report for a Second SAC, Report No.P-6085-ZIM, June 1993. 22 preventive health care.' Likewise, the Togo SAL IV makes adequate budget allocations to basic education and health a condition for the second and third tranche releases. Social sector conditionality is also being introduced into other program lending. The release of the second tranche of the Malawi Entrepreneurship Development and Drought Recovery Program (EDDRP) is dependent on the government making allocations to education of at least 15 percent of total recurrent and development expenditures for 1992/93 and annual increases in the share of primary health care in the total health budget. The aim is to reverse the very low shares that these basic services have received in past budget allocations. Higher surtaxes on utilities are incorporated to finance this increased expenditure. This conditionality reinforces the general stance in favor of the poor of the Bank's sector loans (such as EdSAC II) to Malawi. Program lending can also be used to encourage governments to take actions that increase economic opportunities for the poor. SALs have traditionally committed governments to implement incentive reforms, particularly in agriculture, which are potentially pro-poor. However, this aspect has been underplayed in past SAL President's reports. More detailed discussion of the consequences for the poor of changes in incentive policies is now beginning to appear in documents relating to program lending, and poverty reduction is now more often given as an objective of specific policy changes. Two examples illustrate this. In the case of Zimbabwe, the second SAC emphasizes the benefits to the poor of improving crop marketing (especially of maize, which is the staple food) and reducing burdensome controls on informal activity. Agricultural marketing reform is a condition for tranche release in the second SAC, and the Small and Microenterprise Development Project, which is complementary to the SAC, will assist informal entrepreneurs. In the case of Malawi, the EDDRP commits the government to review labor market imperfections (including tenancy contracts for small farmers) as a condition of release of the second tranche and to making regular upward adjustments of the minimum wage to preserve its real value. This follows extensive Bank analysis of the labor market (see Box 9 in section III above). 4. Market Reforms are Improving Living Standards, but Difficulties in Achieving Improvements in Agricultural Marketing Systems are Delaying Further Poverty Reduction in Rural Areas. Adjustment programs, while macroeconomic and sectoral in nature, are a potentially powerful means for reducing poverty, especially among the rural poor.m The degree to which the poor benefit as producers depends on the extent to which changes in 'border prices' - through such measures as devaluation and import liberalization - are translated into price improvements at the local level. Whether prices improve for producers at the local level depends on the extent to which market distortions are removed. There is some evidence that prices at the farm gate for tradable crops are not increasing in proportion to increases in border prices due to continued inefficiencies in crop marketing. Accordingly, the Bank may need to encourage governments to eliminate these inefficiencies at a faster pace than is presently the case. 7 Burkina Faso SAL I. President's Report No.P-5562-BUR, June 1991. * The principles are well-known. In competitive markets, devaluation raises the domestic prices of agricultural tradables - both exportables and import-substituting foods - while the liberalization of manufactured imports tends to reduce the relative prices of consumer and intermediate goods purchased by farmers. The most favorable outcome for poor farmers is a large rise in output prices and minimal increases in consumer and input prices. 23 In Tanzania for example, the real prices of major export crops have increased little, since the crop marketing boards - being statutory monopsonies - have absorbed much of the increased shilling value of export earnings induced by devaluation, and thus little of the gain has been passed on as higher farm prices (see Box 7 in section M above). The elimination of high-cost marketing is crucial to raising the farmer's share of export earnings within the context of weaker international prices (which have fallen for Tanzania's principal export crops since the late 1980s). The persistence of this inefficiency - after nearly seven years of adjustment - continues to act as a brake on the growth of rural income. Significant proposals for reforming the marketing of major export crops were made in the 1990 AgSAC, but the conditionality proved to be ineffective." The same issue has arisen in the case of Ghana's cocoa marketing system. During the 1980s, improvements were made to the cocoa marketing system, including savings in the costs of the cocoa board, which benefited farmers (see Box 11). However, the process of liberalizing cocoa marketing slowed down, and did not pick up momentum until 1992. In contrast, the liberalization of Uganda's export crop marketing system has been a much faster process and, now that remaining export taxes have been eliminated, more of the world price is passed on to the farmer. Box 11 - Ghana: Cocoa Price Reform and the Poor. Considerable progress has been made in Ghana in revitalizing the production of cocoa, which is the country's principal export crop. Devaluation of the cedi, lower inflation, economies in the Cocoi Board, and a better international price all contributed to a doubling of the real producer price of cocoa between 1983 (the start of the adjustment process) and 1988. Analysis of the Ghana Living Standards Survey shows that a quarter of every cedi earned by cocoa goes to poor farmers. Price reform is, therefore, beneficial to their livelihoods. Starting in 1988, however, the international cocoa price began to decrease, thereby removing by 1990 half of the previous gain in the real producer price. Farmers have been partly protected by a reduction in the amount retained from export earnings to finance the marketing system. The amount retained as a contribution to general government revenues has also been reduced as tax reform has diversified the tax base. Nevertheless, more cost-cutting is needed in cocoa marketing if farm incomes and incentives are to be protected further. While the farmer's share of the world price is now some 50 percent - more than double the share 10 years ago - competitors in Asia and Latin America pay their farmers over 85 percent of the world price, and further reductions in marketing margins are needed. Sources: Ghana- Progress on Adjustment, Report No. 9475-GH, April 1991. E.O. Boateng et ad A Poverty Profile for Ghana, SDA working paper No.5, 1990. Ghana, Tanzania, and Uganda are examples of the variation in speed with which marketing reform is proceeding across Africa. The speed at which a country proceeds with reform depends on how much scope it has for private crop marketing, the extent to which the public crop authorities listen to vocal rent seekers, and the extent to which the Bank has leverage over marketing. In Uganda, the government was 3 Tanzania: President's Report on the Agricultural Adjustment Program, No.P-5200-TA, March 1990. Marketing reforms for food grains and minor export crops have moved at a faster pace. 24 eventually persuaded to disregard rent seeking interests and to liberalize, while, in Ghana, the strength of these interests blocked further reforms of cocoa marketing until recently. The literature on poverty and adjustment frequently refers to the need to explore choices in the timing and sequencing of reform to maximize the gains to the poor without incurring undue delays in achieving reform itself. Nevertheless, few recommendations have yet been made that do not require unrealistic amounts of external finance. However, the early timing of liberalization of crop marketing would accelerate the recovery of rural incomes, and further consideration should be given to ways of achieving this. 5. The Bank's Agricultural Projects Aim to Reduce Poverty, but their Overall Impact on Rural Poverty has not been fully Assessed. The region has numerous projects in the areas of agricultural services, food security, natural resource management, investment and rehabilitation. While in principle these aim to reduce poverty, their effectiveness in improving rural livelihoods is in many cases unknown. This applies particularly to rural infrastructure projects, which do not have a direct poverty focus but which nevertheless can provide considerable benefits to the poor. Most project documents approach poverty indirectly by using categories such as 'small farmers' and 'women farmers'. Given the scarcity of information, this is sometimes unavoidable, but (as was indicated in Section III) care must be exercised in using these categories since they cover not only the poor, but also significant numbers of non-poor. Further collection of information on how these projects benefit the poor should therefore be undertaken as a component of operations. Given the evidence that the rural poor are often women, the Bank's agricultural (and related) projects with a specific focus on women could have a potentially significant impact on poverty (the extent of that impact is, however, largely unknown). Recent projects that focus on women include the Gambian women-in-development project, which enables women to diversify their farm livelihoods; the Congo national extension project, which focuses particularly on women farmers; and the Mozambique agricultural rehabilitation and development project, which also provides extension services that are targeted to women farmers. While Africa's poor farmers grow some export crops, most concentrate on food crops for both subsistence and domestic sale. Therefore, they can expect to benefit from the Bank's food security projects. The Burkina Faso and Rwanda food security projects fall into this category (see Box 12). Other projects that focus on food include the Benin agricultural services project, which improves the access of the least privileged farmers; the Cameroon food security project, which improves extension services for food crops; the Malawi fisheries development project, which emphasizes income-generating activities; and the Uganda livestock services project which is aimed to assisting small dairy farmers. Regions with fragile ecosystems tend to contain large concentrations of poor people. Some recent Bank projects that concentrate on providing assistance in these areas include the Ghanaian second transport rehabilitation project, which includes a labor-intensive construction component for the poor Northern region; the Mali natural resource management project, which includes village-level investments through labor-intensive works to prevent environmental degradation; and the Ugandan northern reconstruction project, which invests in infrastructure in the poorest region of Uganda. 25 Box 12 - Burkina Faso and Rwanda. Translating Food Security Strategies into Action. Burkina Faso The 1990 food security study found that over half of Burkina's population remain chronically food insecure; some 30-50 percent of pre-schoolers suffer from chronic malnutrition, and 12 percent of children are bom under-weight. Policy reforms under SALI and the AgSAC have increased food production per capita, but raising incomes is essential to securing household food access. The most drought-prone provinces will be targeted through food-for-work programs to reduce their reliance an ran-fed agriculture. For women in marginal areas, microprojects will be established with NGO assistance. Finally, the government's capacity to respond more rapidly to acute food shortages will be improved. Rwanda As a result of inappropriate policies that have undermined agriculture and intense population pressure on land resources, the number of hungry people is increasing. One third of rural households have access to insufficient land for subsistence, and landlessness is increasing. A 1991 Poverty Profile classifies as poor some 41 percent of rural households and 11 percent of urban households. The incidence of poverty is highest in the South-Central Zone (the location of recent famines) and among female-headed households (who have the highest dependency ratios). In response to this situation, the 1992 Food-Security Project is undertaking two measures. First, it is providing direct food transfers to the chronically food insecure and to those in acute transitory food insecurity. Target groups include the very poorest people from among the landless, poor female-headed households, families afflicted by AIDs, and orphans. Second, microenterprise development and public works in the poorest communities (focusing on infrastructure construction and land improvement measures) will be undertaken. The project also includes a survey component to monitor poverty. Sources: Burkina Faso: Towards Achieving Food Security, Report No.8761-BUR, June 1990. Brina Faso: Food Security and Nurition Projea, Report No. 10615-BUR, June 1992. A Poveny Profle for Rwanda, 1983-85, AFTSP, October 1991, Rwanda: Food Secrity and Social Action Project, Report No. 10373-RW, May 1992. 6. The Bank is Using Leverage to Shift Social Expenditures towards the Poor Rebuilding effective education and health systems undermined by economic decline and civil conflict is an immense task. But progress is being made, to which the Bank's public expenditure reviews, together with social-sector lending, are contributing. The shares of social sectors in total public expenditures are now rising (particularly the shares of primary health care and education). Three examples - Ghana, Uganda, and Zambia - illustrate this modest success. In Ghana, successive PERs have highlighted inadequate expenditures on primary schooling. Accordingly, two EdSACs (the latest for 1990-93) have helped to raise the share of education in the recurrent budget from 27 percent to 37 percent over 1984-90. Primary education now accounts for some 60 percent of the education budget compared to 44 percent in the mid-1980s. The process of cost- recovery is also now sensitive to the implications for the poor. For example, food subsidies for primary school students have been continued in order to encourage poor children to attend, while subsidies to secondary and tertiary students (who tend to be from higher-income families) have been withdrawn. 26 Primary enrollment has risen, even in poor rural regions, although the quality of the schooling that is provided is still too low.' In Uganda, the 1991 PER, studies of HRD, and conditionality attached to the first structural adjustment credit, have helped to persuade the government to raise the budget allocations for primary education and health. Total central government expenditure on preventive health care rose by 54 percent in real terms between FY 1989/90 and 1991/92, and further increases are budgeted. This goes some way to reversing the over-concentration of public resources on services providing curative care which mainly benefit higher income groups in urban areas. Likewise, the government is now intent on raising the budget share of primary education, especially in rural areas.". With the Bank's encouragement, the Zambian government has begun to allocate more resources to primary education (starting with the 1992 budget). Together with similar efforts in health care, this will reverse the erosion of basic services that occurred during Zambia's economic decline.' The Bank is reinforcing the government's new public expenditure priorities by focusing its rehabilitation projects on basic services. The 1992 Education Rehabilitation Project targets low-income, peri-urban areas, and includes a beneficiary assessment to monitor the project's impact. Pro-poor conditionality, on the tranche releases, is now appearing in social sector loans. Making pro-poor public expenditure reform a condition for the release of SAL tranches is increasing within the Bank (see point 3 above), thereby reinforcing the Bank's message on developing the human resources of the poor. The effectiveness of the Bank's leverage does, however, vary from country to country within Africa, being potentially strongest in the poorest countries that are dependent on IDA lending. Much also depends on whether all of a country's aid donors adopt the same position. Despite significant progress on social spending, real social expenditures are low in Africa, and will remain so until revenue mobilization - neglected during years of macro-instability - is improved. The revenue position is worst in economies that are rebuilding after civil strife such as Uganda." This leaves little room for simultaneously raising the budget share of basic services and maintaining some public provision of benefit to non-poor (but vocal) users, unless the cost-effectiveness of social provisions improves. The Bank is working with governments to improve cost-effectiveness, but this is inevitably a complex process. In the meantime, while budgetary and delivery systems remain disorganized, governments may 40 Ghana.- Public Erpenditure Review, 1989-91, Report No.7673-GH, October 1991. "' Public Choices for Private Initiatives: Proritizing Public Expenditures for Sustainable and Equitable Grow1h in Uganda, Report No.9203-UG, February 1991. Uganda: Managing Public Expenditure, Report No. 10512-UG, May 1992. Uganda. Social Sector Strategy, Report No. 10765-UG, June 1992. ' By 1990, primary education's share of the education budget was only 29 percent; expenditure for each university student was 120 times more than for each primary student, Sources: Zambia Education Rehabilitation Project, Staff Appraisal Report No.10843-ZA. June 1992, and Zambia: Public Expenditure Review - Volume I: Executive Summary, June 1992. 43 See, for example, Guinea: President's Report for an Education Sector Adjustment Credit, Report No.P-5288- GUI, May.1990. '4In Uganda domestic revenues accounted for only 7 percent of GDP in 1990 compared to the SSA average of 20 percent. 27 be committed to new policies that favor the poor, but find it difficult to translate these into actual public services. This is true even in the three countries - Ghana, Uganda, and Zambia - cited above for having been successful in reallocating public expenditures. It is also true in COte d'Ivoire where the 1991 HRD Program redirects social expenditures towards basic services, but ineffective budgetary management - and the resulting complexity of the public expenditure accounts - hinders the implementation (and monitoring) of these reallocations.' 7. Lessons Learned from Social Action Programs have been Incorporated into the Design of Social Funds. The Bank has helped to finance social action programs (SAPs) in Africa since the launch of Ghana's PAMSCAD in 1987.46 The original objective of SAPs was to ameliorate some of the adverse impact of adjustment on the poor, but this eventually broadened out to encompass poverty reduction in general. SAPs have been an important experiment in how to direct micro-assistance to the poor in the context of macro-policy change. Reaching the poor through SAPs has proved to be a challenging task given the weak institutional capacities of governments and the large number of people in need of assistance.4 Some of the early SAPs were difficult to implement, including those in Cameroon, Ghana, and Madagascar. Government agencies were mainly used to implement the project components, and their weak institutional capacities often caused the projects to fall behind schedule. These capacities were also overstretched as most SAPs included a large number of components covering many sectors. It has been especially difficult to implement SAPs in rural areas as the implementation capacities of government agencies are concentrated in towns. As a consequence, urban projects have received more funding than rural projects. For example, most of the labor-intensive works programs were undertaken in urban areas. How much the poor benefit from SAPs is uncertain. The area in which they clearly benefit is in those project components of SAPs that help to rehabilitate the infrastructure of education, health, and sanitation services in poor communities. However, the extent to which the labor-intensive public works projects provide employment for the poor is unknown. Projects such as Senegal's AGETIP -paid the market wage for unskilled workers as a way of targeting the poor (through "self-selection"), but how many of the poor took up such employment is unknown. To achieve more rapid implementation, the Bank has now begun to support a different kind of institution for reaching the poor - social funds.' Social funds respond to funding requests by local governments, NGOs, and other agencies that have the capacity to implement projects at the level of the community. The Zambia social fund, for example, began disbursing money soon after the Bank 4 Cate d'lvoire: Public Ependiture Review, Report No.8748-IVC, July 1990. 4 PAMSCAD - Program of Action to Mitigate the Social Costs of Adjustment (PAMSCAD). 47 For a recent and comprehensive review, see A. Marc, M. Schacter, and C. Graham Social Action Programs and Social Funds: A Review of Design and Implementation in Sub-Saharan 4ica, Africa Technical Department, Human Resources and Poverty Division, Technical Note No. 9, June 1993. a The Bank's involvement with social funds began with Bolivia's Emergency Social Fund (ESF), which was launched in 1986. 28 completed the project identification process. Since social funds are "demand driven," they can respond to community needs more flexibly than SAPs could and, because communities are involved in designing and implementing the social fund projects, the projects are probably more sustainable than those of SAPs. Since social funds involve communities and NGOs more than SAPs did, and rely less on government agencies to implement, they are more likely to be able to reach poor communities in rural areas. Beneficiary assessments are now being used to assess the effectiveness of social funds in reaching the poor. This is an important step, given the uncertainties regarding the cost-effectiveness of the earlier SAPs in helping the poor. Well-designed SAPs and social funds should continue to be an important part of the Bank's assistance to governments in reducing poverty. Their potential to help local communities is significant. However, in helping people at the local level, it is important for the Bank not to lose sight of the need to continue increasing the poverty focus of macro and sectoral policies and public expenditure programs. Well- designed policies and programs provide an economic environment that supports initiatives at the local level. Among donor agencies, the Bank has the most influence over the use of these instruments by governments. Social funds should, therefore, be viewed as complements to the many forms of assistance that the Bank can offer. V. Conclusions 1. Increasing the Poverty Focus of Structural Adjustment and Sector Lending is an Effective Means for the Bank to Help Reduce Poverty. Having set out a new agenda for poverty and development in WDR90, the Bank must press forward to identify for each of its lending instruments, the most effective way of using this instrument to reduce poverty. By this means actions can be taken to make concrete the WDR90 strategy of reducing poverty through encouraging labor-intensive growth, shifting social expenditures towards the poor, and creating safety nets. Loans to support structural adjustment and the reform of sectoral policy (particularly for agriculture) represent a powerful instrument available to the Bank for improving the welfare of large numbers of people. These loans encourage and support market reforms and redirections of public expenditures which can create new economic opportunities for the poor, increase their access to economic infrastructure, and improve their human capital. To realize this potential, the strategy guiding such lending must have a clear set of priorities regarding the best means to reduce poverty in the country concerned. Negotiations and loan conditionalities can then focus on delivering policy changes which will yield the desired benefits for the poor. In the case of structural adjustment and sector loans, negotiations and subsequent loan conditionalities can prioritize the removal of market distortions which have especially harmful effects on the poor. Such distortions include measures of import protection which raise the prices of important consumer goods and inputs to the poor, burdensome regulations which inhibit the growth of micro and small scale enterprises, agricultural marketing boards which impose low producer prices on farmers, distribution systems which fail to deliver essential inputs, and industrial and labor regulations which limit employment. Negotiations and conditionalities for structural adjustment loans can also prioritize the redirection of public investments towards infrastructure of most use to the poor, the redirection of social expenditures to services of most 29 benefit to the poor, and the redesign of consumer subsidies when they are ineffective in targeting the poor. 2. Increasing the Poverty Focus of Bank Lending Requires More Analysis of Poverty Issues in Country Economic Memoranda and Public Expenditure Reviews. Improving strategies for poverty reduction, and therefore operations, requires a better understanding of how the poor participate in markets and what impact policies and programs have on their livelihoods. Given the central importance of CEMs and PERs to defining Bank country assistance strategies, it is critical that these studies should have a clear poverty focus. The appearance of poverty assessments in the form of CEMs is therefore a welcome development in reversing the neglect of poverty issues which has traditionally characterised these reports. The CEMs for Malawi and Uganda, which focused on poverty, have proven to be effective means for guiding and reinforcing the poverty focus of lending operations in these countries. A focus on poverty is still to be implanted in Bank reviews of public expenditures. Such reviews should provide a comprehensive analysis of poverty issues not only with regard to social expenditures, but also with regard to all categories of public expenditure. In particular, public investments in economic infrastructure should be assessed to determine their benefits for the poor. Through these reviews, the Bank has a unique influence among donors on the composition and volume of public expenditures. 3. Improving the Analysis of the Impact of Market Reforms on the Poor and the Benefits of Public Expenditures to the Poor Requires the Collection of More Information on Poor People and their Communities. To support analysis of the effectiveness of market reforms and redirections of public expenditures in reducing poverty, the Bank should increase its support to the gathering of both quantitative and qualitative information on households and communities. The Bank now deploys a wide variety of survey instruments to match the capacities of statistical offices, and has begun innovative developments in the techniques of rapid appraisal. Quantitative information includes such variables as the incomes and expenditures of the poor, their educational attainments, the sources of their employment, and anthropometric information to determine their nutritional status. Disaggregation by region and gender is necessary to direct policies and programs to the benefit of disadvantaged regions and to improve the livelihoods of women. Such information can be provided by both sample surveys and rapid appraisal methods such as beneficiary and participatory assessment. The latter techniques can also provide qualitative information on how the poor themselves perceive their problems, and their assessment of the effectiveness of measures intended to help them. With this information analyses can identify those market reforms and public expenditures of most benefit to the poor. Such information will also assist in designing programs to reduce the constraints which imprison households and communities in poverty, and which will enable poor people to benefit fully from the new opportunities now being created by economic reform. 30 4. More Comprehensive Assessments of the Poverty Effects of Market Reforms on Poverty will Improve Dialogue with Governments and the Donor Community. Persuading governments and the international community that market reform is strongly pro-poor remains an urgent task. Distrust of the Bank's efforts in Africa continues to be both widespread and resilient among many governments and some aid donors, resulting in public opposition to reform and scepticism among donors and NGOs about IDA's effectiveness. To convince this audience that modest welfare improvements under reform represent major achievements - given a starting point of economic decline and high population growth - requires detailed analysis of the impact of reform. Regular repetition of the statement that "adjustment promotes growth, and growth reduces poverty" will prove to be an inadequate basis for dialogue with governments in the 1990s, and an inadequate means for achieving the support and assistance of the international donor community. 5. Priorities among Poverty Groups Need to be Clearer to Guide Bank Operations. The new mandate given by the Operational Directive 4.15 is to reduce poverty in a cost-effective manner. Ultra poor people-such as the disabled and the elderly without family support-are dispersed across communities and it may be cost-ineffective to try to reach them directly. Therefore supporting NGO endeavours is the best route for the Bank to take, and the social funds provide a mechanism for the Bank and other donors to support NGO help to the poor. More assistance to entire communities of the "very poor"-such as farmers in resource-poor regions-is valid and it can have adequate economic returns. However, it needs to be carefully considered to ensure that the instruments used will be effective. At present the populations of resource-poor regions mainly benefit from Bank supported policies and programs when such regions produce an export crop, and the marketing and pricing of that crop is reformed. The Bank has become somewhat pessimistic about its abilities to provide other forms of assistance to such regions, due to weaknesses in its integrated rural development projects of the 1970s. Nevertheless, the search for cost-effective means to achieve poverty reduction in these regions should continue. In particular, arrangements for marketing food crops in resource-poor regions are often weak, and more investment in transport, marketing, and storage infrastructure may be viable as a means for improving both intra-regional and inter-regional food trade, which are of benefit to food security. Moreover, it is now too often assumed that migration to more prosperous regions can easily be undertaken by the populations of resouce-poor regions. To help improve the mobility of the poor of these regions (thereby providing new opportunities for the migrants and remittance income for household members who remain behind), further analysis of instruments (such as improved education in resource-poor regions) is required. One cause of extreme poverty is ethnic discrimination. The Bank should be wary of allowing government 'sensitivity' to block appropriate dialogue and operations to reverse this discrimination, if these have considerable potential for benefiting the poor. If governments find this unacceptable, then the Bank should consider reducing the country operation, and should shift resources to countries that are willing to reverse ethnic discrimination in the interests of the poor. There is a parallel here with the way in which the Bank used to treat gender issues as a 'cultural' constraint, and then moved to a more pro- active policy in the 1980s. 31 6. The Bank can Encourage Local Debate on Poverty by Pressing Governments to Make Poverty and Social Data More Widely Available. Issues of poverty and governance are interlinked in Africa. The poor have paid a heavy price for the resources that have been squandered by elites, who have little interest in making known the Irue costs - especially to the poor - of policy distortions. For these reasons, economic and social data are often not made available or are left unprocessed, and critical scholarship is discouraged. When information is not disseminated, public debate - particularly about policy reform - is easily manipulated by vested interests. Throughout Africa, there is an urgent need for a free and active debate about poverty and the misallocation of public resources that underlies it. The contrast with India - where vibrant policy debates take place - is striking. In India economic and social information is more widely available than in Africa, and is used by scholars, the press, and political parties to attack or defend government actions. As a consequence - and despite all the imperfections of Indian policy making - democratic pressures force politicians to acknowledge the needs of the poor. Until poverty issues are more freely debated within African societies, the Bank's effectiveness will be constrained. The Bank can help to widen the debate by insisting on the dissemination of social and poverty data when the Bank has financed its collection. Governments often use the confidentiality of data as a means for restricting local and independent scholarship on poverty issues. This works against the interests of the poor. Ultimately, poverty reduction must be built on a social consensus about priorities, and priorities can only be established on the basis of sound information and analysis. The Banik has an important role to play in encouraging debate to help establish that consensus in Africa. Bibliography 1. E.O. Boateng, et al., (1990) A Poverty Profile for Ghana, 1987-88, SDA Working Paper No.5. 2. G. Delaine, et al. (1992) The Social Dimensions ofAdjustment Integrated Survey, SDA Working Paper No.14. 3. P. Glewwe and K.A. Twum-Baah (1991) The Distribution of Welfare in Ghana, 1987-88, LSMS Working Paper No.75. 4. C. Grootaert (1993) The Evolution of Welfare and Poverty Under Structural CMange and Economic Recession in COte d7voire, 1985-88, Washington D.C., World Bank, Policy Research Working paper. WPS 1079. 5. C. Grootaert and T. Marchant (1991), The Social Dimensions ofAdjustment Priority Survey, SDA Working Paper No. 12 6. A. Marc, M. Schacter, and C. Graham Social Action Programs and Social Funds: A Review of Design and Implementation in Sub-Saharan Africa, AFTSP (June 1992, Preliminary Draft). 7. S. Reutlinger, "Ethiopia: Coping with Poverty under Policy Reform," Mission Background paper (June 1992). 8. L. Salmen (1992), Reducing Poverty:An Institutional Perspective, World Bank, Poverty & Social Policy Series Paper No. 1. 9. M. Schiff, 1991, "Sectoral, Macroeconomic and Industrial Trade Policies: Measuring their Effect on Agricultural Incentives and Distortions in Africa and Asia." World Economy, Vol. 14 No.4. December). 10. World Bank, A Poverty Profile for Rwanda, 1983-85, AFTSP (October 1991). 11. World Bank policy paper (1991) Assistance Strategies to Reduce Poverty on the interpretation of these variables. 12. World Bank, 1992, Beneficiary Assessment: An Approach Described, Poverty & Social Policy Division, Working Paper No. 1. 13. World Bank, Burkina Faso: Food Security and Nutrition Project, Report No.10615--BUR (June 1992). 14. World Bank, Burkina Faso SAL 1: President's Report No.P-5562-BUR (June 1991). 15. World Bank, Burkina Faso: Towards Achieving Food Security, Report No.8761-BUR (June 1990). 16. World Bank, Cbte d'Ivoire: Public Expenditure Review, Report No.8748-IVC (July 1990). 17. World Bank, Ethiopia: Towards Poverty Alleviation and a Social Action Program: Preliminary Mission Conclusion, AF2 (May 1992) 18. World Bank, Ghana: Progress on Adjustment, Africa Region, April 16 1991 (Report No.9475- GH - Grey Cover). See pp.33-35. 19. World Bank, Ghana: Public Expenditure Review, Report No.8616-GH (29 June 1990), Green Cover. 20. World Bank, Ghana: Public Expenditure Review, 1989-91, Report No.7673-GH (October): grey cover. 21. World Bank, Ghana 2000 and Beyond: Lessons of International Experience for Accelerated and Equitable Growth, West Africa Department, (Yellow cover, August 1992). 22. World Bank, Guinea EdSAC, President's Report No.P-5288-GUI (May 1990). 23. World Bank, The Informal Sector in Zimbabwe: The Role of Women, AF6 and PHR Report No.9006-ZIM (August 1991), Grey Cover. 24. World Bank, Kenya Human Resources: improving Quality and Access, Report No.9023-KE (June 1991), Grey Cover. 25. World Bank, Kenya: Presidents Reportfor the SecondAgSAC, Report No.P-5415-KE (December, 1990), Grey Cover. 26. World Bank, Malawi: Human Resource Development Study, Report No.7854-MAI (April 1990), Grey Cover. 27. World Bank, Nigeria: Implementing the National Policy on Population, Sector Report No.9306- UNI (May 15, 1992), Green Cover, Two Volumes. 28. World Bank, Nigeria Strategy for Food and Nutrition Security, Report No. 9040-UNI (June 1991). 29. World Bank, Poverty Profile of Togo, Country Assessment and Policy Issues, Report No.7825-TO (April 1989). 30. World Bank, President's Report on the Agricultural Adjustment Program, No.P-5200-TA (March, 1990) Marketing reforms for food grains and minor export crops have moved at a faster pace. 31. World Bank, President's Report No.P-5461-MAI (June 1992) on Malawi Entrepreneurship Development and Drought Recovery Program. 32. World Bank, President's Report No.P-5653-IVC (November 1991) on C6te d'Ivoire Human Resources Development Program. 33. World Bank, Public Choices for Private Initiatives. Prioritizing Public Expenditures for Sustainable and Equitable Growth in Uganda, Report No.9203-UG (February 1991), (Three Volumes). 34. World Bank, Report of the World Bank Task Force on Food Security (1988). 35. World Bank, Rwanda: Food Security and Social Action Project, Report No.10373-RW (May 1992). 36. World Bank, Tanzania Economic Report: Towards Sustainable Development in the 1990s: Volume I (Main Report), Volume II (Background Papers) Report No.9352-TA (11 June 11, 1991), Grey Cover. 37. World Bank, Uganda: Agricultural Sector Memorandum, Report No.10715-UG (June 1992, Yellow). 38. World Bank, Uganda: Managing Public Expenditure, Report No. 10512-UG (May 1992), Green. Uganda: Social Sector Strategy, Report No. 10765-UG (June 1992), Yellow Cover, Two Volumes. 39. World Bank, Zambia Education Rehabilitation Project, Staff Appraisal Report No.10843-ZA (June 1992) 40. World Bank, Zambia: Public Expenditure Review - Volume I: Executive Sunmary, White Cover Draft (June 1992). 41. World Bank, Zambia Social Recovery Project, StaffAppraisal Report No.9471-ZA (May 1991). 42. World Bank, Zimbabwe: Agriculture Sector Memorandum, Report No.9429-ZIM (May 1991) Grey Cover, Three Volumes. 43. World Bank, Zimbabwe Presidents Report for a Second SAC, Report No.P-6085-ZIM (June, 1993), Grey Cover. 44. World Bank, Zimbabwe: Strategy for Women in Development, AF6 and PHR Report No.9204- ZIM (March 1991), Green Cover.