Report No. 54491-BF Burkina Faso Social Safety Nets January 31, 2011 Human Development Department Social Protection Unit Africa Region Document of the World Bank GOVERNMENT FISCAL YEAR January 1 to December 31 CURRENCY EQUIVALENTS Currency Unit : CFA Franc (CFAF) (as of December 31, 2009) US$1 : CFAF 480.00 WEIGHTS AND MEASURES Metric System ABBREVIATIONS AND ACRONYMS ACT Artemisinin-based Combination Therapy AFD French Development Agency (Agence Française de Développement) AIDS Acquired Immune Deficiency Syndrome AMBC Assurance Maladie à Base Communautaire APE Association de Parents d'Elèves (Parent­Teacher Association) ART Antiretroviral Therapy CAS Country Assistance Strategy CAST Compte d'Affectation Spéciale du Trésor CCT Conditional Cash Transfer CFAF CFA Franc CILSS Permanent Interstate Committee for Drought Control in the Sahel (Comité Permanent Inter-Etats de Lutte contre la Sécheresse dans le Sahel) CNLS-IST Conseil National de Lutte contre le Sida et les Infections Sexuellement Transmissibles CNRST Centre National de Recherche Scientifique et Technologique CNSA Food Security National Committee (Conseil National de Sécurité Alimentaire) COGES Comité de Gestion des CSPS CONASUR Conseil National de Secours d'Urgence et de Réhabilitation CRB Croix Rouge Burkinabè CRCHUM Centre de Recherche du Centre Hospitalier de l'Université de Montréal CRS Catholic Relief Services CRSN Centre de Recherche en Santé de Nouna CSB Corn Soya Blend CSPS Center de Santé et de Promotion Sociale CST Sector-based and Thematic Commission (Commission Sectorielle et Thématique) CWR Community Wealth Ranking DAF Directorate of Administrative and Financial Affairs (Direction Administrative et Financière) DECRG Development Economic Research Group, The World Bank DFID United Kingdom's Department for International Development DGB Directorate General for the Budget (Direction Générale du Budget) DGCOOP Directorate General for International Cooperation (Direction Générale de la Coopération) DGEP Director of Economy and Planning DGPER Direction Générale de la Promotion de l'Économie Rurale DHS Demography and Health Survey DMEG Dépôt de Médicaments Essentiels Génériques DNEB National Directorate for Basic Education (Direction Nationale de l'Education de Base) EC European Commission ECHO European Commission Humanitarian Office ECOWAS Economic Community Of West African states EICVM Enquête Intégrale sur les Conditions de Vie des Ménages FFT Food-For-Training FFW Food-For-Work FNS National Solidarity Fund (Fonds National de Solidarité) GDP Gross Domestic Product HDI Human Development Index HIPC Heavily Indebted Poor Countries HIV Human Immunodeficiency Virus ILO International Labor Office IMF International Monetary Fund IRD Research Institute for Development (Institut de Recherche pour le Développement) IRSS Institut de Recherche en Sciences de la Santé du CNRST M&E Monitoring and Evaluation MAHRH Ministère de l'Agriculture, de l'Hydraulique et des Ressources Halieutiques MAMS Maquette for MDG Simulation MASSN Ministry of Social Action and National Solidarity (Ministère de l'Action Sociale et de la Solidarité Nationale) MDG Millennium Development Goal MEBA Ministry of Basic Education and Literacy MEF Ministry of Economy and Finance MESSRS Ministry of Secondary and Higher Education and Scientific Research MID Ministère des Infrastructures et du Désenclavement MJE Ministry of Youth and Employment (Ministère de la Jeunesse et de l'Emploi) MoH Ministry of Health MPF Ministère de la Promotion de la Femme MSF Médecins Sans Frontières MTEF Medium-Term Expenditure Framework MTSS Ministry of Labor and Social Security (Ministère du Travail et de la Sécurité Sociale) NGO Non-Governmental Organization OECD Organization for Economic Cooperation and Development OVC Orphans and other Vulnerable Children P4P Purchase for Progress PADS Program d'Appui au Développement Sanitaire PAP-PRSP Priority Action Program for the implementation of the PRSP PER Public Expenditure Review PNDS Health Development National Plan (Plan National de Développement Sanitaire) PNOCSUR Plan National d'Organisation et de Coordination des Secours d'Urgence et de Réhabilitations PrEst Program Pistes Rurales ­ Désenclavement à l'Est PRSC Poverty Reduction Support Credit PRSP Poverty Reduction Strategy Paper QUIBB Questionnaire des Indicateurs de Base du Bien-être SAP Early Warning System (Système d'Annonce Précoce) SCADD Strategy of Accelerated Growth and Sustainable Development (Stratégie de Croissance Accélérée et de Développement Durable) SNS National Food Security Stock (Stock National de Sécurité) SNSA Food Security National Strategy (Stratégie Nationale de Sécurité Alimentaire) SONAGESS Société Nationale de Gestion du Stock de Sécurité SONU Soins obstétricaux et néonataux d'urgence SP Social Protection SSA Sub-Saharan Africa SSN Social Safety Net TB Tuberculosis TB-DOTS Tuberculosis Direct Observed Therapy Short-term TdH Terre des Hommes TFP Technical and Financial Partners THR Take-Home Ration UN United Nations UNDP United Nations Development Program UNICEF United Nations Children's Fund USAID United States Agency for International Development VAT Value Added Tax WAEMU West African Economic and Monetary Union WB World Bank WFP World Food Program WHO World Health Organization Vice President : Obiageli K. Ezekwesili Country Director : Madani M. Tall Country Manager : Galina Y. Sotirova Sector Director : Ritva S. Reinikka Sector Manager : Lynne D. Sherburne-Benz Task Team Leader : Setareh Razmara ACKNOWLEDGEMENTS This report was prepared by the following team members: Azedine Ouerghi (Lead Economist, AFTSP), Cécile Cherrier (Consultant), Pascale Kervyn de Lettenhove (Senior Economist, ECAPR), Carlo del Ninno (Senior Economist, AFTSP), and Setareh Razmara (Lead Social Protection Specialist, Task Team Leader, AFTSP). On gender issues valuable inputs were provided by Dominique van de Walle (Led Economist, PRMGE) and Harounan Kazianga (Consultant). The report also benefited from the technical support of Sarah Hague (Social and Economic Policy Specialist, UNICEF Burkina Faso) and from a background paper prepared by Professor Kimseyinga Savadogo (National Consultant). Peer reviewers were John Elder (Lead Social Protection Specialist, HDNSP) and Damien de Walque (Senior Economist, DECRG). Darcy Gallucio provided valuable editorial support and processed the document. Administrative support was provided by Bintou Sogodogo (Team Assistant, AFMBF) and Josiane Luchmun (Program Assistant, AFTSP). Significant suggestions and comments were provided by Galina Sotirova (Country Manager, AFMBF), Kalanidhi Subbarao (Lead Social Protection Specialist, SASHD), Gilles Alfandari (Senior Economist, AFTP4), Siaka Coulibay (Senior Economist, AFTP4), and Mama Keita (Poverty Economist, AFTP4). To prepare this report, the team worked closely with the government counterparts and other Development Partners (particularly UNICEF, WFP, and IMF). The team expressed its gratitude to the Government of Burkina Faso for its support and cooperation in providing information and data reviewed in this report. In particular, valuable support from the Ministry of Economy and Finance, Ministry of Basic Education and Literacy, Ministry of Social Action and National Solidarity, and Ministry of Health are gratefully acknowledged. The team would like to thank specially Mr. Yamsékré Tiendrebeogo, General Director of Economy and Planning (DGEP) who lead the preparation of this work in Burkina, and Mrs. Eugénie Malgoubri (SCADD). Preliminary results of the report were presented in Ouagadougou (December 2009) to government officials as well as development partners and NGOs, and valuable comments were received from participants. There was consensus on the recommendations and strategic messages of the report. At the request of the DGEP, the team presented the findings and suggestions of this report during the review of the PRSP on April 16, 2010, chaired by the Prime Minister and during the National Social Protection Technical Forum (April 27-29, 2010). The findings of this report, including the action plan for the development of an effective social safety net system, provides valuable inputs to the SCADD prepared recently by the government (September 2010). TABLE OF CONTENTS EXECUTIVE SUMMARY ....................................................................................................... i CHAPTER I: INTRODUCTION .......................................................................................... 1 A. Rationale and Objectives ............................................................................................. 1 B. Definition of Social Safety Net Used in this Study ..................................................... 2 C. Methodology of the Study............................................................................................ 6 D. Structure of the Report ................................................................................................. 6 CHAPTER II: POVERTY, VULNERABILITY, AND RISKS IN BURKINA FASO ..... 7 A. National Context and Macroeconomic Outlook .......................................................... 8 B. Poverty and Human Development Issues .................................................................... 9 C. Poverty Trends, Vulnerability and Poverty Profile .................................................... 10 D. Magnitude of Poverty and Cost of Making a Meaningful Difference ....................... 21 CHAPTER III: GOVERNMENT SOCIAL SAFETY NETS POLICIES AND INSTITUTIONS ..................................................................................................................... 24 A. The Strategic Role Given to Social Safety Nets within the Larger Development Policy ......................................................................................................................... 24 B. Institutional Arrangements of the Social Safety Net System ..................................... 32 C. Financing of the Social Safety Net System ................................................................ 35 CHAPTER IV: REVIEW OF INDIVIDUAL EXISTING SOCIAL SAFETY NET PROGRAMS ......................................................................................................................... 56 A. Cash and Near-Cash Transfers ................................................................................... 56 B. Food Transfers ........................................................................................................... 61 C. Universal Food and Fuel Subsidies ............................................................................ 70 D. Labor intensive Public Works and Cash-/Food-for-Work ......................................... 72 E. Fee Waivers for Health .............................................................................................. 75 CHAPTER V: RECOMMENDATIONS FOR A MORE EFFICIENT SOCIAL SAFETY NET SYSTEM ....................................................................................................... 82 A. Summary of Findings ................................................................................................. 82 B. Policy Recommendations ........................................................................................... 89 REFERENCES ..................................................................................................................... 102 List of Annexes Annex 1: Glossary of Terms ................................................................................................. 107 Annex 2: Good Practice Design Features for Direct Support drawn from International Experience ............................................................................................................. 109 Annex 3: Some Country Examples of Good Practice in Safety Net Programs in Africa ..... 113 List of Tables Table 1: Poverty Indices Based on Place of Residence .......................................................... 10 Table 2: Monetary Poverty Trend Estimates Based on National Accounts ............................. 11 Table 3: Poverty Incidence and Food Vulnerability by Administrative Regions ................... 13 Table 4: Poverty by Area of Residence and Administrative Region, 2003 ............................ 18 Table 5: Poverty by Selected Characteristics of Household Heads, 2003 .............................. 21 Table 6: Estimated Aggregate Amounts of Bringing All the Poor and the Poorest and Most Vulnerable to the Poverty Line through Cash Transfers (2003, 2005, and 2007) ... 22 Table 7: The "Social Safety Net" Component of the Crisis Action Plan 2009-2010 (CFAF million) .................................................................................................................... 31 Table 8: The "Social Protection of Vulnerable Groups" Sub-Program of the Crisis Action Plan 2009 ................................................................................................................. 31 Table 9: List of Main Social Safety Net Programs Currently in Place in Burkina Faso, 2008-2009 ................................................................................................................ 37 Table 10: Total Spending on Social Safety Net Programs, 2005-2009 .................................. 41 Table 11: Social Expenditure (including HIPC and Excluding External Financing) ............. 43 Table 12: Donors and Government's Priorities by Program for 2008 ..................................... 45 Table 13: Beneficiaries and Budgetary Allocation by Program in Burkina Faso, 2009 .......... 47 Table 14: Options for Scaling Up SSN Programs .................................................................... 49 Table 15: Options for Increasing Safety Net Budgets: Advantages and Disadvantages ........ 52 Table 16: Beneficiary Groups in CNLS-IST Research-Action ............................................... 58 Table 17: Main Characteristics of Recent Food Voucher Programs ....................................... 58 Table 18: Intervention Stock, Situation as of 31 December 2008 .......................................... 62 Table 19: National Food Security Stock, Situation as of 31 December 2008 ......................... 62 Table 20: Indicative Numbers of Potential Beneficiaries of National Food Stocks ............... 63 Table 21: CONASUR's Assistance in Cereals to Vulnerable Populations ............................. 64 Table 22: CONASUR's Assistance in Cereals to Disaster-Affected Persons ........................ 64 Table 23: Indicative Number of Beneficiaries of Targeted Food Distributions, 2006-2010 .. 65 Table 24: Main Characteristics of School Feeding Programs in Primary Schools ................. 68 Table 25: Price implications of exemptions on targeted staple products, 2008 ...................... 71 Table 26: Price Implications of Exemptions on Targeted Fuel Products, 2008 ...................... 71 Table 27: WFP Food for Asset Programs (Beneficiaries and Budgets), 2006-2010 .............. 74 List of Figures Figure 1: Position of Social Safety Nets in Larger Development Policy .................................. 5 Figure 2: Regional Typology of Burkina Faso........................................................................ 14 Figure 3: Household Typology in Burkina Faso (National) ................................................... 16 Figure 4: Household Typology in Burkina Faso (Rural/Urban) ............................................. 17 Figure 5: Past and Current Widows Have Lower Body Mass Indices Than Women of Other Marital Status ................................................................................................. 19 Figure 6: Rural Children of Currently Widowed and Previously Widowed Women Are Less Likely To Be in School ................................................................................... 20 Figure 7: Health Insurance National Scheme Project's Institutional Arrangements .............. 29 Figure 8: Evolution of Spending on Social Safety Nets Over Time, 2005-2009 .................... 42 Figure 9: Share of Basic Education, Health and SSN in Total Domestically And Externally Financed Expenditures ........................................................................... 44 Figure 10: Indicative Annual Number of Beneficiaries of the Existing Social Safety Net Programs, 2009 ........................................................................................................ 47 Figure 11: Share of Subsidy Benefits ...................................................................................... 72 Figure 12: Average Number of Monthly Consultations of Children Under 5 by CPSP in the Four HELP/TdH Intervention Districts and Four Control Districts .................. 79 Figure 13: Rates of Return of Human Capital Investment Initially Setting Investment To Be Equal across All Ages ........................................................................................ 91 List of Boxes Box 1: Outline of Social Protection National Policy Drafted under MTSS Leadership (Not Adopted) ........................................................................................................................... 27 Box 2: The Social Protection Floor Initiative........................................................................................ 35 Box 3: Information Issues Related to Social Safety Nets...................................................................... 36 Box 4: Budget Management Issues in Burkina Faso............................................................................. 45 Box 5: Lessons of the Analysis of Fiscal Space in Burkina Faso, Using the MAMS Model ............... 51 Box 6: WFP's Urban Voucher Program................................................................................................ 60 Box 7: Transfers In Cash and In Kind: Alternatives or Complements? ................................................ 66 Box 8: Planning for Going to Scale with Nutrition Interventions ......................................................... 67 Box 9: Public Works Programs ­ Elements Required for Reaching the Poor ...................................... 73 Box 10: Targeting Effectiveness of Public Works Programs: International Experience ...................... 75 Box 11: A Community-Based Targeting Approach to Exempt the Worst-Off From User Fees in Burkina Faso ............................................................................................................................. 77 Box 12: The Debate over User Fees ...................................................................................................... 79 Box 13: Principles of Child-Sensitive Social Protection ....................................................................... 90 Box 14: Does and Dont's of Female-Sensitive Social Protection ......................................................... 97 Box 15: To Target or Not To Target? ­ The Costs and Political Economy of Being Selective.......... 100 IBRD Map No. 33379 .......................................................................................................................... 122 EXECUTIVE SUMMARY Burkina Faso is a poor landlocked country with a narrow natural resource base and a rapidly expanding population of 15.8 million. Despite structural reforms and sound macro- economic policies, the economy remains highly vulnerable to adverse shocks (climatic and external, including food and fuel price increases and terms of trade for cotton). Suffering from extreme poverty and difficult living conditions, the country is ranked one of the lowest in the world (177 out of 182 countries in 2009) according to the United Nations Development Program's Human Development Index (UNDP HDI). In 2008, average per capita income was US$480 and over 40 percent of the population lived under the poverty line. Despite the government's efforts to improve overall living standards, gaps in access to social services, a high demographic growth rate of 3.1 percent, extreme poverty, and high vulnerability to a wide variety of shocks continue to plague its populace. The ninth annual review of the Poverty Reduction Strategy Paper in 2008 revealed that economic growth had not translated into the expected poverty reduction, mainly due to: the recent food crises; strong demographic growth; and the inefficiency of wealth redistribution mechanisms. More recently, the world's financial and economic crisis and the severe flood of 2009 have impeded economic growth. High vulnerability of the country to a variety of shocks (environmental, social, and economic), particularly affecting the poorest Burkinabes, has increased the demand for social safety net (SSN) programs. Earlier World Bank work (2004) has shown that SSN programs are administered by several ministries, have limited coverage, are heavily dependent on external financing, and require further assessment to better understand the role of the existing safety net programs and particularly to review the functioning of selected programs. In this context, in response to a government request, this report, with the technical support of UNICEF, provides a detailed, updated inventory of the existing social safety net programs and suggests policy measures that could improve their coverage, efficiency, relevance, and financial sustainability. These findings are expected to help the government extend safety nets to reach the poorest (chronic poor) and the most vulnerable populations and to provide valuable inputs to its 2010 Strategy of Accelerated Growth and Sustainable Development (Stratégie de Croissance Accélérée et de Développement Durable or SCADD). A. Main Findings of the Report This report shows that the scope and coverage of the existing social safety net system is too limited and that most interventions are fairly small in scale and designed as temporary programs. On average, excluding fuel subsidies, spending on social safety net programs was about 0.6 percent of GDP from 2005 to 2009 ­ from 0.3 percent in 2005 to 0.9 percent in 2009, while about 20 percent of the population is food insecure and lives permanently in chronic poverty. Universal fuel subsidies are very expensive (0.7 percent of GDP in 2007) and have a very limited impact on the poorest decile (84 percent of the benefits go to the non- poor). Among the remaining programs, food transfers are the main form of social safety net programs in Burkina Faso, accounting for 69 percent of total SSN spending and over 80 percent of all estimated SSN beneficiaries in 2009 (excluding fuel subsidies). However, most of the financing for social safety net programs comes from external and ad hoc resources. i Definition of Social Safety Net The term social safety net refers in this report to publicly funded non-contributory transfer programs targeted in some manner to poor or vulnerable individuals or households, and aimed at directly increasing consumption and/or access to basic social services. Social safety nets can play four roles in development policy: (i) they redistribute income to the poor and most vulnerable, with an immediate impact on poverty and inequality; (ii) they can enable households to make better investments in their future ­ both in the human capital of their children and in the livelihoods of the earners; (iii) they help households manage risk ­ both protecting households and promoting their independence; and (iv) they allow governments to make choices that support efficiency and growth ­ by freeing other sectors from any redistributive role and letting them concentrate on the efficient provision of services. Thus, social safety nets can contribute to the overall development of the country. Like other social protection instruments, social safety nets are demand-side interventions aimed at acting in synergies with supply-side interventions (e.g., provision of quality education and health services, agricultural production, and provision of microfinance services) for human capital development, rural development, etc. and ultimately poverty reduction and social cohesion. Source: Grosh et al. 2008. Poverty and Vulnerability Although poverty incidence has declined as a result of sustained economic growth over the last decade, and important progress has been made on the access to basic services by households, significant challenges remain and the rapid population growth is likely to prevent the country from reaching its Millennium Development Goals (MDG). Based on the available survey data and national accounts projections, the proportion of the population living below the poverty line fell from 46.4 percent in 2003 to 43.2 percent in 2009. In terms of assets, although poverty declined (from 46.3 percent in 2003 to 40.4 percent in 2005 and to 38.5 percent in 2007), the share of chronic poor increased as a result of shocks: their share dropped from 25.4 percent in 2003 to 18.8 percent in 2007 after peaking at 31.1 percent in 2005 following the weather-related shocks. In terms of social indicators, access to health services has become more accessible due to a regular increase in public health expenditures as a share of the overall budget (from 7.4 percent in 2004 to 9.9 percent in 2006). However, (i) the quality of education remains low; (ii) gender gaps remain high; and (iii) the rapid population increase (3.1 percent per annum) has major negative consequences on growth, living standards, and poverty reduction. Based on the 2003 priority survey, a typical poor household in Burkina Faso has seven or more members, is headed by a male who is polygamous, illiterate, or has a lower primary education level, and is a farmer living in the rural areas of the Northeast Region. Over 92 percent of the population lives in rural areas where the poverty incidence is more than two times higher than in urban areas (52.3 percent versus 19.9 percent). In terms of regional poverty, the Center Region is among the poorest with over half of the poor. Male- headed households, whose total share in the incidence of poverty represents 95.6 percent, tend to be poorer than households headed by females both in terms of incidence and gap. However, gender is also a key poverty correlate, and as in many parts of Sahelian West Africa, there are reasons to believe that some groups of women may be particularly poor and vulnerable (widows, remarried widows, divorced women, and their children). Moreover, ii households whose head is illiterate or educated on a low primary level also have a notably higher poverty incidence (51.0 and 41.1 percent respectively). Government Strategy, Institutional Setup and Expenditures for Social Safety Nets Although Burkina Faso has not yet adopted a consolidated national social protection policy, social safety net programs are playing an important role in Burkina's human capital development and crisis response strategies. Social safety net programs appear in many sector-based strategies: health, education, food security, and employment. Yet, their potential to reduce poverty and vulnerability through income redistribution, promotion, and transformation is less often recognized in medium-term sectoral strategies. Moreover, as a result of the lack of a comprehensive strategy and the lack of an appropriate institutional setup, inter-ministerial coordination mechanisms for social protection and social safety nets are weak. The Government is taking several actions to address the need for more comprehensive social protection, and a more effective social safety net system in particular. In February 2010, a Joint Ministerial Committee on Social Protection was established by decree and its first meeting took place in July 2010. In addition, in April 2010, the Ministry of Economy and Finance (MEF) organized a national technical forum on social protection with the explicit goal to outline a national and consolidated vision of social protection, including the social safety net component, and to directly inform the drafting of the SCADD. Excluding the cost of general subsidies for food and fuel, total spending on social safety net programs over 2005-2009 averaged about 0.6 percent of GDP. Currently, the social safety net programs in Burkina can be classified in five categories: (i) cash and near-cash transfers; (ii) food transfers (subsidized food sales, targeted food distributions, nutrition programs, and school feeding); (iii) universal subsidies (food and fuel); (iv) public works; and (v) fee waivers. This level of spending on SSNs corresponded to roughly one-third of the amount needed to bring all the poor to the poverty line through cash transfers. Since 2006, the financing for SSN programs has become increasingly donor-dependent. Excluding universal subsidies, the external financing share in total SSN financing increased from about 57 percent in 2005 to 69 percent in 2008 and 76 percent in 2009. In the meantime, government spending increased more slowly, from 0.1 percent of GDP in 2005 to 0.2 percent of GDP in 2009 (excluding fuel and food subsidies). There are important priority differences between the government and donors: donors focus on nutrition, while the government concentrates financing mainly on universal fuel subsidies. The coverage of existing social safety net programs remains limited compared to the needs. Theoretically, the estimate of the total number of beneficiaries of social safety net programs exceeded 3.9 million individuals in 2009 (or about 25 percent of the total population). In reality, the coverage for most programs is temporary (reactions to shocks); and the level of benefits in some programs is very small and insufficient to help the poor smooth their consumption. This result is primarily due to the limited financial allocation and lack of implementation capacities ­ in particular, difficulties in identifying and reaching the poorest and measuring cost-effectiveness. Moreover, due to lack of investment on monitoring and evaluations (except for few recent donor-financed initiatives, such as cash transfers, food vouchers, and school feeding), there is no information on the actual impact of social safety net programs that would facilitate informed policy choices. iii Existing Safety Nets Programs The review of existing social safety net programs confirms that despite substantial needs, few programs assist the chronic poor and provide regular and predictable transfers. Existing interventions are mainly implemented through projects, often during periods of shocks, and focused on one geographic area. No systematic approach exists to assist poor and vulnerable households. The main challenges for such a system to be defined are: the definition of priority target groups; the choice of adequate instruments; and the establishment of solid monitoring and evaluation systems to inform policy decisions. Given the recent innovative initiatives, such as food vouchers, cash transfers, and health fee waivers, lessons can be learned to improve the efficiency of the existing social safety net programs and potentially scale them up. Cash and Near-Cash Transfers Valuable lessons can be learned from recent cash transfer programs introduced in Burkina Faso: Since 2008 three pilot cash transfer programs have relied exclusively on external funding: a pilot cash transfer program under the CNLS-IST (Conseil National de Lutte contre le Sida et les Infections Sexuellement Transmissibles) and two food voucher programs under the Catholic Relief Services (CRS) and the World Food Programme (WFP). The results of the impact evaluation of the CNLS-IST project on 3,900 households are expected late 2010. A mid-term evaluation of the WFP program, benefiting over 30,000 households, has shown an efficient implementing process and positive results in terms of food consumption both in quantity and quality, despite challenges of targeting in urban areas. Food Transfers Food transfers are the main form of social safety net programs in Burkina Faso, accounting for 69 percent of total SSN spending over the period 2005-2009, and over 80 percent of all estimated SSN beneficiaries in 2009 (excluding fuel subsidies). Four types of food transfers are currently in place: (i) targeted subsidized food sales, (ii) targeted free food distributions, (iii) nutrition programs, and (iv) school feeding programs. Targeted subsidized food sales: As a result of weak monitoring and evaluation of the subsidized food sales program, information on the actual number, profile, and poverty level of the beneficiaries is lacking. However, it is feared that the poorest may not have the financial resources to access the subsidized cereals. Targeted free food distributions: Free food distributions are provided through three mechanisms: (i) distributions of free food by the Conseil National de Secours d'Urgence et de Réhabilitation (CONASUR); (ii) Food-for-Education/Training programs by the World Food Programme (WFP); and (iii) general relief programs by the Catholic Relief Services (CRS). Iin 2008, the CONASUR provided food assistance to about 44,000 persons on an ad-hoc basis (particularly individuals affected by flooding or small-scale disasters). The WFP food transfers, conditional on attending literacy or training courses, has gradually refocused its efforts to the Sahel Provinces, covering about 36,000 beneficiaries in 2008/2009. Finally, CRS annually assists about 14,000 vulnerable persons (people living with HIV, orphans, the elderly, and the disabled). However, because food insecurity is primarily an access problem, cash-based transfers might be more appropriate. Nutrition programs: In response to important needs, the government and its partners, iv UNICEF and the World Bank in particular, have made substantial efforts since 2003 to increase nutrition interventions. UNICEF, WFP, ECHO, and their partner NGOs are particularly active in the treatment of moderately and severely malnourished children under 5 and pregnant and lactating mothers. Nevertheless, the coverage of existing programs remains inadequate, even though severe acute malnutrition is one of the main challenges in reaching the nutrition MDGs. School feeding: School feeding represents one of the main social safety net programs currently in place in Burkina Faso. In 2009, it accounted for 24 percent of total spending on SSN and covered over 27 percent of the estimated total number of SSN beneficiaries (excluding general subsidies). Over 50 percent of primary schools in the country have school feeding activities, and over 50 percent of spending is externally funded by USAID and multilateral funding through the WFP and CRS. The Ministry of Basic Education and Literacy (MEBA) promotes endogenous school feeding programs, providing only a three- month supply of food, while donor-funded programs provide support for the whole school year in the most food insecure areas. A recent impact evaluation (2008) reported that both forms of assisted school feeding, in-school meals and take-home rations for girls, increased girls' enrolment by about 6 percent. Neither intervention showed a significant positive impact on learning outcomes, but a positive impact of take-home rations on the nutritional status of younger siblings was observed. These mixed results invite to reconsider the choice of in- school meals to increase school enrolment rates. In-school meals programs are costly and greater impacts (on poverty and nutrition) may be achieved with targeted (possibly conditional) rations with similar results on school enrolment ­ and the possibility to target girls in priority and reduce gender disparities. The government is planning to scale up its school feeding program (in-school meals) for the 2010-2011 School Year in order to cover all primary schools. It is unclear, nevertheless, whether the planned expansion of school feeding activities is a temporary measure to mitigate the effects of the recent crisis or a long-term strategy. While scaling up school feeding might be a legitimate practical temporary response to the food crisis, the authorities need to further investigate the most appropriate social safety net instrument in the long run, including targeting mechanisms to reach the poor, to increase school enrolment rate and contribute to poverty reduction. Universal Food and Fuel Subsidies To mitigate the negative effects of high food and fuel prices, the government has provided universal subsidies that are expensive and inefficient in reaching the poor. The universal food subsidies that were introduced in 2008 in response to the crisis induced by high prices worldwide proved very expensive and less than efficient in reaching the poor. Therefore, the program has been terminated. In addition, a number of fuel products have been subsidized by the government for years, despite a high fiscal cost and very limited impact on the poorest households (except lamp oil). Overall, the direct fiscal cost of exoneration on fuel import taxes amounted to an estimated 0.7 percent of GDP in 2007 and in 2008 and had a very limited impact on the poor: over 84 percent of the benefits went to the non-poor. Labor Intensive Public Works and Cash/Food-For-Work Much is to be learned from the positive experience of the ongoing public works programs. The PrEst (Program Pistes Rurales ­ Désenclavement à l'Est), implemented by the NGO Helvetas since 2002 under the supervision of the Ministry of Infrastructure (MID), uses a labor intensive approach for infrastructure creation in rural areas. Other experiences of public works include the WFP's Food-for-Assets program, which focuses mainly on building quality v assets. Building on PrEst and WFP's Food-for-Assets experiences, other types of public works programs could be introduced as effective social safety net instruments. In fact, a recent review of the experience with public works programs in several countries shows that well-designed and implemented public works programs can help mitigate income shocks and be used as an effective anti-poverty instrument. That being said, effectiveness highly depends on the ability of the program to provide additional sources of income to the most vulnerable population when most needed. And, further attention would need to be put on targeting methods, length and timing of work, specific design features that can increase the participation of women, community participation, and the choice of remuneration. Fee Waivers for Health Although several national initiatives intended to provide free health care to the poor and vulnerable, they are not operational for lack of implementation mechanisms. The SONU ( emergency obstetric and neonatal care) subsidy policy of 2006 entitled indigent (poor) women to free health services, yet it failed to specify the parameters of qualification and, thus, prevented these women from receiving services. There are three main issues with this program: (i) the definition of who is indigent (or poor)? (ii) who decides upon this status? and (iii) who absorbs the cost of the fee waivers? Although international experience suggests that abolishing user fees for the poorest is equitable, further work is needed to define the implementation mechanisms. Particularly in the context of health system financing in Burkina Faso, fee waivers for health need to be considered in a broader health policy strategy in order to identify the model to be promoted and determine the appropriate implementation mechanisms. A couple of externally-funded pilot projects providing free health services to pregnant and lactating women and children under 5 showed very good results on the use of health services. These initiatives, implemented by the NGOs HELP (in two districts in the Sahel since September 2007) and Terre des Hommes (in two districts in the North since October 2008) in collaboration with the COGES, echo the international experience which suggests that abolishing user fees is not only fair and equitable but also potentially feasible both technically and financially. The debate surrounding fee abolition or expansion of fee waivers posits a reconsideration of the health sector strategy and financing and its social protection measures as a whole. Further research is needed to inform the following key questions: (i) if the government wants to abolish user fees, how should it be done? (ii) how can partners support the government? and (iii) how can the abolition of user fees be an entry point to service quality improvement? B. Main Policy Recommendations The main recommendation of this report is to develop an efficient social safety net system that adequately responds to the needs of the poor in Burkina Faso. Building on the government's commitment to building its social protection base, the priority actions toward the development of a more efficient and cost-effective social safety net system need to focus on: (i) Strengthening the strategic, institutional, and financial framework for designing, implementing, managing, monitoring, and evaluating safety net programs; and vi (ii) Developing a plan for improving the effectiveness of the safety net system by reforming existing programs and, based on recent experiences, designing new ones. In particular, the reform of existing social safety net programs would involve reducing very small or ineffective programs while strengthening a few viable programs with better targeting and outcomes. Strengthen the Strategic, Institutional, and Financial Framework (a) Adopt a National Social Protection Strategy including Social Safety Nets Develop a comprehensive social protection strategy. The Social Protection National Policy drafted in 2007 represents a good starting point. It should be updated considering the current challenges faced by the country. Given the huge needs in health and education, and in the context of limited financial resources, social protection instruments should directly contribute to human capital development. Synergies and economies of scale should be promoted between the different social protection instruments and other social policies. Clarify the objectives of the social safety net system, within a broader social protection strategy. The objectives of the social safety net component are to: (i) directly support the consumption of the chronically poor and vulnerable populations; (ii) ensure access to basic social services to poor and vulnerable populations, in order to promote human investment; and (iii) provide temporary support to poor and vulnerable populations affected by shocks. Therefore, the priority principles of the social safety net system should be to: (i) ensure that chronically extreme poor and vulnerable populations receive regular and predictable support along with complementary programs to escape poverty traps and break the intergenerational transmission of poverty; (ii) provide temporary income to vulnerable groups in case of shocks; (iii) pay particular attention to the needs of vulnerable children (e.g., nutrition, education and conditional cash transfers) and the needs of poor and vulnerable women (i.e., minimize potential negative impacts, optimize positive impacts on women and gender equity). Propose priority actions and clarify links among social policies. The different existing sectoral strategies tend to refer to the whole spectrum of people in need, without clearly setting priorities. The strategy should outline priorities and clarify links among social policies. Social safety net programs are meant to act in conjunction with other poverty reduction programs, and are typically used to complement supply-side interventions and fill in where other policies cannot deliver sufficient immediate results. Therefore, in Burkina Faso, social safety net programs could complement other social programs like ensuring education and health spending become pro-poor, providing food security, complementing health insurance and population policies, etc. Coordinated support would be needed among various sectoral ministries. vii (b) Reinforce the Institutional Framework for Social Protection and Social Safety Nets Support the permanent inter-ministerial committee for social protection, recently created. As of February 2010, the government set up a permanent Committee for Social Protection. It will be responsible for revisiting the social protection strategy (including social insurance and social safety net programs), supervising/coordinating the various initiatives, and ensuring not only cross-sectoral dialogue among the ministries but also tangible outputs resulting from the dialogue. This committee needs to play an active role in designing and supervising the implementation of the social protection strategy incorporating social safety nets. Clarify the role and responsibilities of the different national institutions engaged in social safety nets. Once the Strategic Framework for Social Protection is defined and priority actions for social safety nets are clarified, the respective functions and roles of the main structures responsible for social safety nets need to be defined (MOH, MOE, MASSN, FNS, and CONASUR). This includes: (i) defining the roles of centralized and decentralized authorities; (ii) defining the appropriate implementation arrangements and ensuring a separation of duties (inclusive of the NGOs and private sector involved in the delivery of SSNs); and (iii) providing capacity-building support. The definition of any new responsibility will have to be reflected in budget allocations (e.g., train staff, build human and material resources, and strengthen institutional setup) and be supported by better coordination mechanisms among the various sectoral ministries. Provide capacity-building support. The concept of social safety nets as a necessary social investment (regular and predictable) is largely new to Burkina Faso. Awareness efforts and training are required both at the national and local levels. Other initiatives like on-the-job training and study tours may be useful to increase the general understanding of and interest in social safety nets and social protection. (c) Strengthen the Financial Framework Secure the financial resources needed to ensure that, in the short run, expenditures for social safety net programs are kept at least at their current levels (around 1 percent of GDP in 2009) and are later expanded to accommodate larger coverage. To this end, fiscal arbitrage will be needed to retain or scale up the most cost-effective social safety net programs. Burkina Faso needs to carefully consider the role safety nets should play in the development strategy of the country, the desired number of beneficiaries for each type of program and their expected cost given the poverty and vulnerability profile, performance of existing programs, international experience, and national institutional capacity. To bring this financing onto a more sustainable basis, the following steps could be considered: Determine the overall envelope of the government's and partners' budget needed for the desired level of social safety net coverage and make adequate provision each year in the budget. Seek budget support in the context of a Poverty Reduction Support Credit from IDA and similar operations from other external partners. This will mean a step-up in the management of the safety net system and an improvement in such aspects as fiduciary arrangements, procurement, and audits as well as results monitoring and evaluation. Savings can be achieved through better targeting, streamlining costs, and public expenditure reallocation by reducing the cost of small or inefficient programs while viii strengthening a few viable programs with better targeting and outcomes. In this context, spending needs to become more efficient and pro-poor in general, scaling down poorly targeted subsidies and focusing on high-priority sectors like health and education. Gains in discretionary expenditures, furthermore, can produce efficiency gains. (d) Improve Program Monitoring and Evaluation Systematic monitoring of the overall set of safety net programs and evaluation of individual programs are needed to judge how well resources are being used. This is a precondition for the piloting and/or scaling up of any social safety net program. In particular, six systemic actions could be considered: Establish a rigorous classification of social protection expenditures and a comprehensive list of public social safety net programs; Set up minimum reporting requirements for social safety net programs to allow proper monitoring and assess effectiveness, with costs broken down between service delivery and overhead, sources of financing, etc.; Systematically transmit program monitoring reports to the sectoral ministries responsible for social protection and social safety nets and maintenance of a database on programs; Set up proper evaluation of programs and especially pilots during the introduction of new interventions and/or the expansion of existing interventions to new categories of beneficiaries; Involve civil society in monitoring and evaluation; Strengthen the capacities of the sectoral ministries for monitoring and evaluation; and Provide more training for program managers in monitoring and evaluation techniques coupled with a mechanism for the exchange of experiences across programs. Planning the Improvement of the Effectiveness of the Social Safety Net System Once the policy, institutional, financial frameworks, and monitoring and evaluation system have been defined, there is need to: (i) identify appropriate social safety net instruments based on needs; (ii) improve the efficiency and targeting effectiveness of existing programs; and (iii) expand successful ongoing pilot interventions. Identify the appropriate set of social safety net instruments: In this context, first, based on the data from the ongoing Enquête Intégrale sur les Conditions de Vie des Ménages (EICVM) (available in 2010-2011), the poverty analysis needs to be updated to identify priority target groups and poverty maps used to identify the respective geographical locations. Second, define the type, role, scale, and frequency of social safety net instruments for each priority target group. Based on the updated poverty analysis, and further feasibility analysis, it is proposed to consider the following set of instruments on a permanent basis to tackle chronic poverty: (i) nutrition supplement programs for pregnant and lactating women and children under 5; (ii) targeted school feeding programs for children aged 6-14 to increase school enrolment and attendance rates for poor children; (iii) regular cash transfers to households ix living in chronic poverty to increase the real income or poor households; and (iv) seasonal labor intensive public works to provide a source of income to poor workers and to construct public infrastructure or provide community services. The scope and the scale of each of the instruments proposed should be decided based on the results of the analysis, fiscal availability, and implementation capacity. Once an appropriate permanent safety net system is set up, selected mechanisms could be considered to be scaled up to respond to crisis, complemented by other temporary instruments. Improve the efficiency and targeting of social safety net programs. Effective targeting, at the moment, faces many constraints, including ­ among others ­ administrative capacities and quality of governance, negative perceptions about public programs, and the family obligations of women. In this context, international experience indicates the following best practice approach: Monitor processes and costs of the programs. Information on the resources used to manage programs is crucial to improve the quality of their delivery and to reduce costs over time. However, a larger initial investment might be needed in administrative tools and capacity. Ensure that better information is collected among potential and current beneficiaries to facilitate targeting and assess results. Currently, detailed data on the beneficiaries is in deficit. Develop effective targeting tools to redirect the flow of resources toward the poor. The government needs to develop and apply common targeting criteria and instruments at two levels: (i) geographical level, to allocate social public expenditures in general and social safety net programs in particular where the largest number of chronic poor are located; and (ii) household level, with the development of common proxy means test indicators that can be applied objectively across a range of programs. Establish appeals and grievances mechanisms. More transparency in program standards is needed and high standards of governance need to be set and maintained. Retain or expand effective and efficient programs. Further assessments and analysis are needed to make detailed proposals on the expansion or reduction of the scope of individual social protection programs. That being said, initial recommendations can be made as a result of the review presented in this report: (i) Expand efficient cash transfers taking into account the upcoming impact evaluations (cash transfers and food voucher); (ii) Further review of subsidized food sales and targeted food distributions to assess their cost-effectiveness and evaluate their impact on beneficiaries and consider their reduction in favor of program with better targeting outcomes; (iii) Strengthen and expand nutrition programs, particularly given the existing poor nutritional outcomes; (iv) Explore geographical targeting mechanisms for school feeding to ensure that they benefit poor children and particularly learn lessons from the ongoing experiences of WFP and CRS which are already geographically targeted; (v) Avoid use of general subsidies, except as an instrument of last resort in time of crisis, and particularly use commodities consumed primarily by the poor; x (vi) Introduce public works targeted to the rural poor such as using the wage rate and other possible targeting criteria to temporarily generate income during seasonal shortages of jobs and in times of shocks; (vii) Carefully review the waiver of fees and the abolishment of user fees in health care for the rural poor in the context of a broader health care system and health financing reforms, in order to establish compensation mechanisms for the effective implementation of the program. xi Burkina Faso: POLICY ACTION PLAN FOR AN EFFECTIVE SOCIAL SAFETY NET SYSTEM Policy Recommendations Actions and Time Frame Actors Monitoring Indicators 2010-2011 2012-2014 Policy Objective 1: Strengthen Strategic Framework to Design, Coordinate, Manage, and Finance the National Social Protection System, Including Social Safety Nets A national consultation A permanent SP inter-ministerial committee is Structures in charge of SP and SSNs DGEP leading the Adoption of national framework involving set up to design and monitor SP strategy, are operational inter-ministerial SSN strategy (2011) various social sector including SSNs. Dissemination of national SP strategy committee with Adoption of national agencies provides policy A sub-committee responsible for social including SSNs representatives from SP strategy (2012) guidance insurance system is set up to follow up reforms sectoral ministries, Annual reports on in health insurance and pension system decentralized results of SP collectivities, civil A sub-committee responsible for SSNs is set up including SSNs society, and the to define the type, role, and instrument to technical and financial address the needs of the poor and vulnerable partners (TFPs) Institutional framework for Clarify the role and responsibilities of the Ensure coordination between the state DGEP/ inter- Institutional capacities SP and SSN is strengthened different national institutions engaged in SSNs: and TFPs ministerial committee are improved define roles and appropriate implementation for SP/SSNs arrangements and Provide capacity-building support A sustainable financial Establish rigorous tracking of SP expenditures Multiyear program budgeting of MEF/DGEP and Reporting system on framework is set up for and of public SSNs. SSNs sectoral ministries, spending on SSNs financing SP programs Determine budget envelop needed for a TFPs (including budget and including SSNs comprehensive SSNs external funding) Identify sources of sustainable financing (budget, development partners, local collectivities, NGOs, and private sector) xii Policy Recommendations Actions and Time Frame Actors Monitoring Indicators 2010-2011 2012-2014 A robust monitoring and Invest in program monitoring and evaluation to Transmit annual program evaluation MEF/Sectoral Annual monitoring evaluation system for SSNs assess cost-effectiveness of SSNs reports to the sectoral ministries ministries and report for each SSN is in place to facilitate Set up minimum reporting requirements for responsible for SP and SSNs agencies, FTPs program informed policy decisions SSNs Strengthen the sectoral ministries Impact evaluation of Begin implementing systematic monitoring of capacities for monitoring and most important SSNs evaluation (training, exchange of programs Involve civil society in monitoring and evaluation experience across programs). Policy Objective 2: Support Consumption of the Poor and Vulnerable and Increase Their Access to Basic Social Services through Efficient Social Safety Net System Effectiveness of current Define priority groups benefiting from SSNs Develop effective targeting tools to Sub-committee Assessment reports SSNs is strengthened and based on the results of the 2010 household redirect the flow of resources toward responsible for SSNs Criteria for targeting effective programs are budget survey the poor (prepare TOR for Monitoring indicators scaled up. Define priority instruments to address needs of Explore geographical targeting program assessment) Evaluation report on priority groups mechanisms for school feeding to Technical ministries the results of pilot Develop a targeting system for populations with ensure that they benefit poor children implementing the programs chronic poverty (food insecurity) Review feasibility of fee waivers for programs Prepare feasibility analysis for expanding cash health and abolishment of user fees TFPs transfers based on recent experience (cash for the poor in the context of health transfers and food voucher) financing reforms Prepare feasibility analysis for introducing Test pilot programs (cash transfers public works targeted to the poor and based on and public works) and monitor and ongoing experience assess them Review cost-effectiveness of subsidized food sales and targeted food distributions Review mechanisms for strengthening and expanding nutrition programs xiii CHAPTER I: INTRODUCTION 1. This Chapter (i) presents the background, rationale, and objectives of this study; (ii) clarifies the definition of social safety net as understood in this review; and (iii) discusses the methodology and the structure of this report. A. Rationale and Objectives 2. Burkina Faso suffers from extreme poverty and difficult living conditions. Burkina Faso is a poor landlocked country with a narrow natural resource base and a rapidly expanding population of 15.8 million (mid-2009).1 Despite structural reforms and sound macro-economic policies, the economy remains highly vulnerable to adverse shocks (climatic and external, including terms of trade for cotton, and food and fuel price increases). The economy is largely agricultural: over 80 percent of the active population derives its income from agriculture, and the sector accounts for over 40 percent of the Gross Domestic Product (GDP). The country is ranked one of the lowest in the world (177 out of 182 countries in 2009) according to the United Nations Development Program Human Development Index (UNDP HDI). In 2008, the per capita income was US$480 (Atlas method) and about 40 percent of the population lived under the poverty line. Similar to other countries, poverty is particularly prevalent in rural areas, where over 90 percent of the poor are located. 3. Despite government efforts to improve overall living standards, gaps in access to social services, a high demographic growth rate of 3.1 percent, extreme poverty, and high vulnerability to a wide variety of crises continue to plague Burkina Faso's population. The rapid demographic growth has jeopardized the provision of human capital investments and has important consequences for the economy and food insecurity. According to the Country Assistance Strategy (CAS), Burkina Faso is unlikely to reach several Millennium Development Goals (MDGs) by 2015, such as halving the proportion of people with income less than US$1.25 a day, ensuring completion of the primary education cycle for all children, and eliminating the gender disparity in primary and secondary education. In education, the adult literacy rate was only 28.7 percent in 2007, and the quality of education remained low. Health outcomes for children and women have not shown any improvement: only 41 percent of births were attended by trained medical personnel, the maternal mortality ratio was estimated at a very high 700 per 100,000 child births, and infant mortality remained at around or slightly above 100 per 1,000 child births between 2000 and 2008. However, some progress has been made in other areas, specifically those related to HIV prevalence and environmental sustainability. Given the deteriorating economic conditions, the availability of public resources for basic services and progress toward the MDGs may be challenging. Like most other Sub-Saharan African countries, Burkina Faso has to contend with severe seasonal fluctuations in rainfall, and associated price (terms of trade) and output shocks impinging harshly on the poor's consumption (welfare). The poor in Burkina are also more exposed to health shocks. Both covariate shocks induced by fluctuating weather and the macro and external environment, and idiosyncratic risks associated in particular with health, render most of the poor and near-poor vulnerable to severe consumption shortfalls. 1 Source: Population Reference Bureau, 2009 Population World Data Sheet. 1 4. In this context, social safety net (SSN) programs are of particular importance. The extent of high vulnerability among the population has increased the demand for social safety net measures. Earlier World Bank work provided a broad review of existing publicly funded risk management and safety net programs in Burkina Faso [World Bank 2004]. This analytical work has focused mainly in programs on health, education, food security, the labor market, and social assistance programs targeted to women and children. The broad findings are the following: (i) programs are administered by several ministries (e.g., health, education, social affairs, agriculture, women and children, etc.); (ii) programs are either universal or address specific vulnerable groups (women, children, the disabled, repatriates, etc.), and are heavily dependent on external financing; (iii) most programs have limited coverage, mainly because of inadequate financing, lack of institutional structures to implement programs, and lack of coordination between various ministries; and (iv) further assessment of the existing programs is needed to better understand the role of the existing safety nets programs and particularly to review the functioning of selected programs. 5. The government requested technical support from the World Bank to improve the efficiency of safety nets programs for the most vulnerable populations. In this context, the present report aims at: (i) providing a detailed updated inventory of the existing social safety net system ­ analyzing their costs, financing mechanisms, coverage, and effectiveness whenever data is available; (ii) identifying the shortcomings of the current social safety net system; and (iii) making suggestions, based on international experience, for improving the coverage, efficiency, relevance, and financial sustainability of the social safety net system and of selected programs, including improvements in their monitoring and evaluation. 6. This work is part of the technical assistance provided by both the World Bank and United Nations Children's Fund (UNICEF) to the Government of Burkina Faso. The findings of this study will directly support the revision of the Poverty Reduction Strategy Paper (PRSP) that will be replaced in 2010 by the Strategy for Accelerated Growth and Sustainable Development (Stratégie de Croissance Accélérée et de Développement Durable or SCADD),under the Ministry of Finance supervision, as well as efforts initiated by the Ministry of Social Action and National Solidarity (Ministère de l'Action Sociale et de la Solidarité Nationale or MASSN) to develop a social protection strategy, and by the Ministry of Youth and Employment (Ministère de la Jeunesse et de l'Emploi or MJE) to investigate the use of labor intensive public works programs (programs HIMO) in Burkina Faso. B. Definition of Social Safety Net Used in this Study 7. There is no overall consensus on a universal definition of social safety nets, on what they should address and on how to tailor safety net programs to local circumstances.2 Some players may use different terminologies ­ social protection, social security, social assistance, social safety nets, and social transfers ­ interchangeably. 8. In the present report, the term social safety nets refers to non-contributory transfer programs targeted in some manner to the poor or vulnerable [Grosh et al. 2008]. Defined in this way, one might think that the term "social safety nets" is analogous 2 Appendix 1 clarifies key social policy concepts used in this report to ensure a common understanding of key terminology and ideas. 2 to the Burkinabe term "social action." However, in practice, this Burkinabe concept appears to cover a much broader array of activities, including the provision of social services (e.g., support to disabled associations, access to justice, etc.) and income- generating activities. 9. Social safety nets as defined in this report, aim at increasing and stabilizing consumption as well as supporting the use of basic social services ­ either directly or indirectly by lowering the cost of food and other basic commodities and essential services, and not at increasing resources per se. Income-generating activities and other livelihood programs thus fall outside the scope of this study. Such programs are important poverty reduction instruments but may not ensure a direct increase in consumption, and are not classified as social safety net programs. 10. Social safety nets are targeted in some manner to the poor and vulnerable, that is, individuals living in poverty and unable to meet their own basic needs, or in danger of falling into poverty, whether because of an external shock or socio-economic circumstances, such as age, illness, disability, or discrimination. Social safety nets may serve one or a combination of the following groups [Grosh et al. 2008]: Chronic poor, defined as people who lack the assets to earn sufficient income, even in good years; Transient poor, defined as people who earn sufficient income in good years but fall into poverty, at least temporarily, as a result of idiosyncratic or covariate shocks ranging from an illness in the household or the loss of a job to drought or macroeconomic crisis; Vulnerable groups, commonly including ­ but not limited to ­ the disabled, the elderly, orphans, the displaced, refugees, and asylum seekers; and Losers in reforms. 11. Policies and programs intended to improve access to basic services for the entire population (free primary education) thus fall outside the scope of the present report. And so do transfer programs targeted at communities and associations (to build social assets in vulnerable communities) since they are not targeted at poor and vulnerable individuals or households directly. General subsidy programs (price subsidies) may be considered as social safety nets if they are introduced with the intention of increasing the consumption of vulnerable households (households affected by global high food prices, for example) by lowering the price of the basic commodities that they consume. 12. Instruments used to increase consumption include direct transfers, subsidies, and fee waivers. Common types of social safety net programs may be classified as follows [Grosh et al. 2008]: Programs that provide unconditional transfers in cash or in kind: a. Cash transfers (e.g., child benefit, family allowances, and social pensions) and near-cash transfers (e.g., food stamps and commodity vouchers); b. In kind food transfers (e.g., school feeding and take-home rations) and other in kind transfers (e.g., school supplies); and 3 c. General subsidies meant to benefit households, often for food, energy, housing, or utilities. Programs that provide an income: a. Public works in which the poor/vulnerable work for food or cash. Programs that protect and enhance human capital and access to basic services: a. Conditional transfers (i.e., transfers in cash or in kind to poor/vulnerable households subject to compliance with specific conditions in relation to education and/or health); and b. Fee waivers for health and education ensuring access to essential public services (e.g., fee waivers for health care services, scholarships, etc.), and thus often referred to as demand-side interventions in health and education. 13. Social safety net systems are usually woven of several programs, ideally complementing each other as well as complementing other public or social policies. They can be long-term predictable transfers or short-term emergency transfers. A good social safety net system is more than a collection of well-designed and well-implemented programs. The social protection "systemic effect" can trigger more than the sum of the individual social programs. 14. Social safety nets form a subset of broader social protection policies and programs along with social insurance and social legislation, which ensures minimum civic standards to safeguard the interests of individuals (e.g., labor laws, and health and safety standards). Social protection is a basic human right that directly tackles poverty and food insecurity and contributes to economic growth and human development. 15. Social safety nets are part of a broader poverty reduction strategy. Social safety nets interact with and work alongside of social insurance; health, education, and financial services; the provision of utilities and roads; and other policies aimed at reducing poverty and managing risk (Figure 1). Poverty reduction requires ensuring people's access to consumption and food security, health, education, rights, voice, security, dignity, and decent work. It involves a political process and requires dedicated efforts to empower the poor by strengthening their voice and fostering democratic accountability. In recent years, the concepts of social protection and social safety nets have increasingly become a central component of poverty reduction and food security strategies in developing countries. 4 Figure 1: Position of Social Safety Nets in Larger Development Policy Source: Grosh et al. 2008. 16. At the core of widespread debate surrounding social safety nets is the question of predictability and sustainability. An increasing number of development actors argue that social transfers should be predictable meaning, paid or distributed regularly or in a predictable manner (e.g, whenever climatic conditions impede solid agricultural production), not as an ad-hoc reaction to a crisis. As a pre-emptive initiative, this type of social transfer allows recipients to prepare for and protect themselves in an effective way against unforeseeable catastrophes. If in the past safety nets were perceived as simple relief transfers that helped poor people to alleviate the worst effects of shocks, it is now increasingly recognized that a social safety net is to be distinguished from individual social projects by the integration of many activities into a predictable, institutionalized social protection system capable of responding to the vulnerabilities of risk, and supported by a rights-based approach. Simple relief transfers proved to have limited long-term benefits and involved the danger of creating dependency. Safety nets, if correctly implemented, have the potential not only to protect, but also to significantly promote the livelihoods of poor people. 17. Finally, because the main realm of public action is via the public sector, the present report concentrates on publicly financed social safety nets or, those funded by national or local government or by official international aid. In most developing countries, there are three basic forms of social transfers provisioning: (i) formal mechanisms that are provided by governments and are prescribed by law; (ii) semi-formal support provided by UN agencies or NGOs; and (iii) informal mechanisms supplied by households and communities. The present report does not cover informal social safety nets. 5 C. Methodology of the Study 18. The present study was completed using existing poverty analysis, administrative data, and household survey data, in close collaboration with various ministries as well as with donors engaged in social protection. The main sources of data used in this report are: Core Welfare Surveys. The needs analysis involved an econometric analysis of the data from the 2003, 2005, and 2007 CWIQ surveys, along with a review of the 2003 Risk and Vulnerability Assessment. See Dynamique de la pauvreté et de la précarité au Burkina Faso (2003-2007) undertaken by a joint team from the Ministry of Economy and Finance of Burkina Faso and the World Bank (Nouvé et al. 2009 unpublished). A new detailed household survey for 2009-2010, with in- depth information on households consumption, is currently in the field (over 12 months with four visits to households). The complete results of this survey, expected in 2011, will provide more accurate estimates of poverty. Inventories of various existing safety net programs (funded by the government, international organizations/bilateral donors, and NGOs). Administrative data as well as any existing program evaluations have been used to assess the outreach and effectiveness of existing programs and their relevance to the main risks faced by vulnerable households. The detailed inventory of the safety nets programs, which was prepared by Professor Kimseyinga Savadogo (National Consultant), was the key source of information for this study. D. Structure of the Report 19. The rest of this report is organized as follows: Chapter II presents the nature and magnitude of the problem, describing what is known of poverty and vulnerability in Burkina Faso. Chapter III provides an overview of governmental social safety net policies, institutional setups, and financial arrangements. Chapter IV reviews the performance of individual existing social safety net programs, considering in particular their adequacy, equity, cost-effectiveness, and sustainability. Finally, Chapter V provides key policy recommendations to improve the effectiveness of the existing social safety nets in order to protect the poor and vulnerable and respond to future crises. 6 CHAPTER II: POVERTY, VULNERABILITY, AND RISKS IN BURKINA FASO As a result of sustained economic growth over the last decade, the monetary poverty incidence is estimated to have declined in Burkina Faso from 46.4 percent in 2003 to 43.2 percent in 2009. In terms of assets as determined by the QUIBB surveys, poverty declined from 46.3 percent in 2003 to 38.5 percent in 2007. However, there are reasons to believe that recent deceleration in economic growth and heightened fuel and food prices may have resulted in real income drop and reversed the positive trends in poverty reduction. Although different surveys give rise to different poverty estimates, it can be confirmed that poverty has decreased over the last decade. The upcoming household consumption surveys for 2008 and 2009-2010 should provide more accurate information on poverty trends. Based on available data, a household's vulnerability to risks in Burkina Faso depends on such factors as the existing health and nutritional status of individuals, gender, physical assets such as housing, infrastructure, and household location, as well as on educational level, access to information, and cultural and behavioral practices. The poor are more vulnerable than other population groups because they are typically more exposed to risk, face many risks simultaneously, and have fewer assets and resources to be able to cope with downside risks. Based on QUIBB surveys, it is estimated that chronic poverty dropped from 25.4 percent in 2003 to 18.8 percent in 2007 after peaking at 31.1 percent in 2005 following the 2004 weather- related shock. Moreover, during period of shocks, it is estimated that an additional 20 percent of the population are transient poor. According to the 2003 priority survey, a typical poor Burkinabe household has seven or more members, is headed by a male who is polygamous, illiterate, or educated on a lower primary level, and is a farmer living in the rural areas of the Northeast Region. There are also indications that widows and women who have been widowed in the past may be significantly worse off, as well as more vulnerable to downside risk than many other women in rural Mali. Thus safety net policies should take this into account. Given the high incidence of poverty, the share of the population that should benefit from social safety nets is substantial (over 40 percent of the total population). Based on available poverty data, the financial cost of public transfers needed to close the poverty gap could range between 2.3 and 2.6 percent of GDP. However, given budget constraints, there is need for a sensible safety net strategy that aims to reduce the most extreme forms of destitution and food insecurity and helps the poorest and most vulnerable to smooth consumption through shocks. Moreover, pro- poor public expenditures through well-targeted social safety nets could provide the resources for the poor to make the necessary investments in human capital development. 20. A good understanding of the population groups that need safety net programs on a permanent basis is crucial to guide policy-makers in defining an appropriate mix of safety net policies. This Chapter thus takes stock of the poverty and vulnerability situation in the country using existing reports and available data. Section A reviews the national macroeconomic context; sections B and C focus on the extent of poverty, access to basic services, poverty evolution over time, and main sources of risks and vulnerability. Finally, section D provides a rapid assessment of the financial dimensions of poverty alleviation. 7 A. National Context and Macroeconomic Outlook 21. Burkina Faso is a small, land-locked, predominantly rural West African country. The rapidly expanding population of 15.8 million inhabitants (2009) relies upon a very narrow natural resource base, with cotton as the only major existing export commodity.3 The agriculture sector employs over 80 percent of the population and contributes to 40 percent of national GDP. The cotton sector is, and will remain, a vital element of Burkina's economy in the short to medium term; it provides 700,000 jobs, employs 17 percent of the population and constitutes the only source of cash revenue for many farmers. However, the agrarian economy is particularly vulnerable to climatic and external shocks. Burkina's fragile Sahelian environment is drought-prone and susceptible to flooding. In addition to climatic conditions, external factors, such as international prices of cotton, and the fluctuating fuel and food prices also impact the country. As a land- locked country, bordered by six neighbors, Burkina Faso is highly dependent on good relations with and the stability of its neighbors for transit, trade with external markets, and political stability. 22. Since the early 1990s, the country has benefited from relative political stability and a progressive platform of economic and political reforms. Over the last two decades, the Government of Burkina Faso has embarked upon a dramatic economic and political reform agenda, involving reforms in trade, price liberalization, and tentative moves toward decentralized governance. Following the devaluation of the currency in 1994 and on the back of this reform agenda, stable real GDP growth of 5.6 percent on average was achieved between 1994 and 2004. Between 1998 and 2006 Burkina Faso enjoyed a high average economic growth rate of about 6 percent per annum. Between the 1980s and 2008, its per capita income increased from US$270 to US$480. Most of this achievement is an outcome of the efficiency gains earned through the implementation of structural reforms, which included the 1994 CFAF devaluation. The significant flow of external aids, estimated at between 8 and 9 percent of GDP since 1996, contributed strongly to this success as well. In fact, Burkina Faso became one of the first countries to be eligible for the Heavily-Indebted Poor Countries Initiative. 23. However, since 2007, variable rainfall, fluctuating commodity prices, increases in fuel and food prices, and the recent global financial crisis have resulted in a decline in growth rates. Low cotton prices in 2004-2006 (a drop of over 40 percent in the world's reference price during this period), combined with the depreciation of the dollar against the euro, further exacerbated by a drop of 44 percent in cotton production in 2007-2008 and rising food and fuel prices in the last two years, all had an effect on Burkina's growth. The global financial crisis is negatively affecting Burkina's growth and fiscal performance. Although the country has been resilient to the impact of the financial crisis due to limited cross-border linkages in its banking system, the economic outlook has worsened as the lagging effects on the real activity are feeding through the economy. A dramatic reduction in cotton export prices and lower demand are the main transmission mechanisms, although lower oil and high gold prices are partly offsetting them. The lower global growth and the decline in domestic cotton prices have significantly affected the profitability of ginning companies, placed pressure on government finances, and created risk causing a drastic cut in production by small cotton farmers potentially leading to a significant increase in 3 In 2008 total population was 14.7 million. With a population growth of 3.1 percent per annum, it is estimated that the total population reached 15.8 million in 2009. 8 poverty. With the tightening of credit conditions, foreign direct investment is also expected to decline, especially in mining. Moreover, official aid may be reduced as donors tackle financial crises at home. 24. As a result, GDP growth for 2009 was revised from 6.2 percent prior to the crisis to an estimated 3.1 percent. Growth projections for 2010 were similarly revised from 6.0 percent to 4.2 percent. Growth is expected to come mainly from the agriculture sector, as productivity measures introduced in 2008 in response to the food crisis are becoming effective, and in mining (albeit at a lower level), as gold prices have mostly resisted the downturn in commodity markets while more mines are moving to the exploitation stage. Growth is now not expected to recover before 2012. Based on the IMF staff report (January 2010), in the 2009 supplementary budget, the overall deficit (commitment basis, including grants) is estimated at 6.7 percent of GDP. With the accelerated repayment of domestic payment arrears, the overall deficit on a cash basis (including grants) would widen to 7.3 percent of GDP. Despite the impact of the global economic slowdown on economic activity, revenue is estimated at 13.2 percent of GDP thanks to administrative measures taken by the authorities to step up controls and enhance efficiency in revenue-collecting agencies. Total expenditure is estimated at 27 percent of GDP in 2009, up from 21.7 percent in 2008. This significant increase is due to the cost of measures against adverse shocks, and to the return to the trend in externally financed investment. The latter is forecast to reach 7 percent of GDP in 2009, increasing by more than 3 percentage points, as some of the projects delayed in 2008 would be completed in 2009. The current account balance is not expected to be significantly affected for the moment because the drop in cotton exports is offset by high gold prices and low oil prices. Also, inflation declined to 2 percent in 2009 stemming from the recent declines in food and fuel prices. B. Poverty and Human Development Issues 25. Past good economic growth and increased investments in social services have resulted in a decline in the incidence of poverty. Based on the available survey data and national accounts projections, the proportion of the population living below the poverty line is estimated to have decreased from 46.4 percent in 2003 to 43.2 percent in 2009. However, there are reasons to believe that recent deceleration in economic growth and heightened fuel and food prices may have resulted in a real income drop and reversed the decreasing trend of poverty reduction. The 2009 Country Assistance Strategy (CAS) estimated the drop in average real income by 1.1 percent in 2007 and 3 percent in 2008, and estimated that the share of the population living below the poverty line in 2008 reverted back to above 45 percent.4 The CAS also argues that while the immediate impact of the cotton crisis on producers is being softened by the "smoothing fund," poverty upward trends could be exacerbated if the impact of the global financial crisis, particularly on the cotton sector, were prolonged due to the close linkages between cotton production, rural revenue generation, and food production in Burkina Faso. 4 Reported in the 2009 CAS and citing the 2009 CEM. However, no source was cited and it is very difficult to ascertain whether this increase is actually due to the crisis or simply because of the use of different methods to calculate poverty incidence. In fact the issue of data comparability has already been flagged in the 2003 Risk and Vulnerability Assessment for Burkina Faso (see section on poverty trends). 9 26. Similarly, Burkina has made satisfactory progress on households' access to basic services. Indicators show improvements in drinking water and health. Overall access to drinking water has improved to 79 percent in 2007 from 54 percent in 2004 with more modest gains in rural areas. Health services have become more accessible on average due to a regular increase in public health expenditures as a share of overall budget (from 7.4 percent in 2004 to 9.9 percent in 2006). Vaccination rates remained high, resulting in marked improvements in fighting contagious diseases with a reduction in the mortality rate for meningitis from 21 percent in 2005 to 8.5 percent in 2006. The prevalence of HIV/AIDS remains low at 2.03 percent in 2006. 27. Significant challenges, nonetheless, remain--particularly given the rank of Burkina Faso on the UNDP HDI as one of the lowest countries in the world (177 out of 182 countries in 2009). The HDI reveals that: (i) the literacy rate is still low (28.7 percent in 2007) and the quality of education is preoccupant; (ii) the gender gaps persist in school enrolment, literacy, and primary completion rates (only 40 percent overall) as well as to access to lower secondary, even though they are increasing; and (iii) the Gross Intake Rate at first grade is 82 percent. In addition to children who are unable to enter into primary school, more than a third of those attending cannot continue into secondary. 28. Furthermore, the rapid population growth is likely to prevent Burkina Faso from reaching its MDGs. Despite impressive declines in mortality levels, fertility has remained very high at 6.2 children per woman on average. As a result of its delayed fertility transition, Burkina Faso has experienced a phenomenal population increase. This is exemplified by the densification of the population, in particular on arable land. Since independence, the urban population has increased eleven fold, while the rural population has tripled in size. Today, Burkina Faso must accommodate 600,000 live births every year, as compared to 200,000 in 1950. Finally, the rapid population increase also has far reaching consequences for the economy in the areas of labor productivity, savings, growth, living standards, and poverty reduction. C. Poverty Trends, Vulnerability and Poverty Profile 29. Tracking poverty trends in Burkina Faso is a challenge in spite of the relatively large number of surveys. Since 1994, four priority surveys (1994, 1998, 2003, and 2009-2010, although results are available only for the first three surveys) and several QUIBB (2003, 2005, and 2007) were undertaken by the INSD. Qualitative surveys also were conducted to assess the risks faced by households and their vulnerability. The different surveys, however, used different methodologies and were not directly comparable with each other. The findings based on the 1994, 1998, and 2003 priority surveys undertaken by the INSD revealed an increase in key poverty indicators, particularly in urban areas as summarized in Table 1. Table 1: Poverty Indices Based on Place of Residence Incidence Depth Severity Share 1994 1998 2003 1994 1998 2003 1994 1998 2003 1994 1998 2003 Urban 10.4 16.5 19.9 2.5 4.0 5.5 0.9 1.5 2.2 3.8 6.1 7.8 Rural 51.0 51.0 52.3 16.1 15.7 17.9 7.0 6.8 8.2 96.2 93.9 92.2 National 44.5 45.3 46.4 13.9 13.7 15.5 6.0 5.9 7.0 100.0 100.0 100.0 Source: The National Institute of Statistics and Demography (INSD), 2003. 10 30. However, during the period from 1994 to 2003, the growth rate in output outpaced population growth by 3 percentage points per annum. Subsequently, per capita GDP was 14 percent higher in 1998 than in 1994, and 13 percent higher in 2003 than in 1998. Inequality was stable from 1994 to 1998 and fell from 1998 to 2003. Similarly, the trends in non-monetary welfare indicators showed positive improvement (education, health, and nutrition). The combination of robust growth, unchanged or declining inequality, and several improved non-monetary welfare indicators should have made a significant impact on poverty. 31. Despite various inconsistencies in surveys data, it is safe to conclude that poverty has declined over the last 15 years in Burkina Faso, and that despite recent shocks, poverty incidence seems to have remained around 40 percent of the population. The 2004 Risk and Vulnerability Assessment for Burkina Faso study concluded that taking into account data inconsistencies, such as: (i) changes in the recall period, which has triggered substantial and spurious changes in reported consumption, and hence poverty; and (ii) the timing of the survey, which has substantial impact on poverty trends especially in the arid West-African context, given the large differences in consumption in the pre-harvest (lean-season) and the post-harvest seasons, poverty might actually have decreased between 1994 and 1998. The above-mentioned 2004 study undertook adjustments to correct for the inconsistencies identified. The adjusted results suggest that poverty had fallen in Burkina Faso from 44.5 percent in 1994 to somewhere between 33 and 36 percent in 1998, consistent with the macroeconomic trends and dynamics of key non-monetary welfare indicators. Similarly, other adjustments carried within the context of the macroeconomic simulation model for poverty analysis (PAMS) suggested that poverty may have declined from 46.4 percent in 2003 to 39.2 percent in 2007 (cited in Nouvé et al. 2009). According to the poverty estimated projected based on the national accounts, it is also estimated that the national poverty incidence declined in the period 2003-2009 from 46.4 percent to 43.2 percent (See Table 2). Nevertheless it is important to ensure that further poverty analyses take into account the comparability among surveys data. A comprehensive household survey was launched in 2009-2010 and will be completed by the end of 2010. This survey, which covers a period of 12 months and is based on four visits to households, and which is comparable with 2003 survey, should provide reliable information on monetary poverty and trends. The results of this survey are expected to be available in 2011. Table 2: Monetary Poverty Trend Estimates Based on National Accounts 2003 2005 2007 2009 National 46.4 45.4 43.6 43.2 Rural 52.3 51.4 50.4 48.8 Urban 19.9 20.9 19.2 20.6 Source: The 2003 data is based on Household Budget survey, and estimates for 2005-2007-2009 are based on economic projections from DGEP. 32. Finally, a new study has attempted to shed some light on poverty trends in Burkina Faso.5 It takes advantage of a series of household surveys called Questionnaire des Indicateurs de Base du Bien-être (QUIBB), designed to monitor the implementation of the Poverty Reduction Strategy Paper (PRSP) and the fact that the INSD priority surveys of 2003, 2005, and 2007 are also based on the QUIBB model. The main difference resides 5 Dynamique de la pauvreté et de la précarité au Burkina Faso (2003-2007) undertaken by a joint team from the Ministry of Economy and Finance of Burkina Faso and the World Bank (2009 unpublished). 11 in the fact that the priority surveys contain information on income and expenditure whereas the QUIBB relies on assets and wealth on the one hand and on households' perceptions of their own well-being on the other hand. The 2003 priority survey is also a QUIBB, which provided the study with the possibility of using the 2003 monetary poverty as a baseline. It appears at this stage that it would be conceptually more appropriate to use the QUIBB series to track the poverty trends (and probably also the vulnerability of households) and to use the 2003 priority survey (awaiting the 2008 survey results) to identify the poverty profile at a particular point in time. Poverty trends6 33. In terms of assets, as determined by the QUIBB surveys, poverty declined from 46. 3 percent in 2003 to 38.5 percent in 2007 (Table 3). This is consistent with the pattern of economic growth, non-monetary welfare indicators, and inequality trends. A similar trend is observed in both urban (from 19.9 percent to 13.0 percent) and rural areas (from 52.2 percent to 43.9 percent) for the same period of time. Similar declining trends have been observed across most administrative regions although with varying starting levels, speed in decrease, and in some cases reversal at the mid-term (2005). In terms of assets, Northeast and the Southwest Regions have the highest poverty incidence (above 60 percent). Not only did these regions show very little improvement across the survey dates, their situation also worsened in 2005. The observed worsening may be explained by the effects of 2004 low rainfall and locust attacks that affected agriculture especially in the agro-ecological fragile areas, which in turn contributed to higher prices of food items observed in 2005. This phenomenon is quite noticeable in the data depicting food vulnerability (Table 3). The information gathered from the households regarding their ability to satisfy their food needs clearly show a spike in 2005 before reverting back to lower levels in 2007. The effects vary depending on the geographic location and to some extent the initial level of household assets. Urban households saw their food vulnerability increase by 5.6 percentage points between 2003 and 2005 whereas rural households suffered a much higher increase, amounting to 18.5 percentage points for the same period. The Northeast, Southwest, and Cascades Regions were the most affected by the 2004-2005 shock and did not seem to fully recover in 2007 when compared to other regions. Although affected like other groups, households with higher asset levels in 2003 were able to relatively weather the 2004-2005 shock. This is in particular true for urban centers and some of the areas in Center Region. 34. The 2009 study (Nouvé et al.) combined assets and food vulnerability indicators over time in an effort to determine a regional typology of households. This approach identified three distinct regions reflecting the household level of expenditure, their asset, their vulnerability to food shocks, and their resilience to shocks in the 2003- 2007 period. The regional typology includes: (i) Northeast Region characterized by moderate consumption, limited assets, high vulnerability, and stagnating situation; (ii) Center Region characterized by low consumption, moderate assets, moderate vulnerability, and an improving situation; and (iii) Capital and South Regions characterized by high consumption, high assets, limited vulnerability, and improving situations (Table 3). The underlying message of this approach is that factoring in risks and household resilience to 6 This section draws largely on the analytical work undertaken by Nouvé et al. (2009) cited above. It uses a consistent approach across the three surveys (2003, 2005, and 2007) to determine trends over time but also across geographic zones and socioeconomic groups. 12 shocks are key to sound policy formulation and the choice of interventions to address poverty and vulnerability. Table 3: Poverty Incidence and Food Vulnerability by Administrative Regions Monetary Poverty Based on Food Vulnerability Poverty Assets or Wealth (Inability to Satisfy Food Needs) 2003 2003 2005 2007 2003 2005 2007 NATIONAL 46.4 46.3 40.4 38.5 49.1 65.4 38.2 Areas of residence Rural 52.3 52.2 45.5 43.9 50.7 69.2 39.4 Urban 19.9 19.9 16.3 13.0 41.5 47.4 32.1 Northeast Regions Sahel 37.2 62.7 83.7 67.1 50.2 86.8 52.5 Center North 34.0 64.0 55.4 50.5 45.0 90.6 42.4 East 40.9 76.6 63.1 73.2 45.3 72.0 34.7 Center Region North 68.6 43.8 37.7 23.6 70.9 72.1 23.7 Boucle du Mouhoun 60.5 51.3 42.5 35.8 46.4 52.5 29.0 Center West 41.3 44.7 40.2 39.4 61.3 64.9 45.8 Center Plateau 58.6 37.5 27.1 36.4 71.3 64.9 38.3 Center South 66.1 45.1 29.6 26.4 54.9 65.7 43.5 Center East 55.1 40.2 40.7 38.4 43.6 73.7 48.9 Southwest 56.6 69.3 63.5 60.2 58.6 75.5 66.9 Capital/South Regions Center 22.3 10.5 7.1 7.8 41.7 50.6 31.9 Hauts Bassins 34.8 28.7 21.8 17.0 27.0 44.0 22.7 Cascades 39.1 54.2 23.4 47.4 34.9 39.6 47.7 Source: Nouvé et al., unpublished 2009; based on the QUIBB of 2003, 2005, and 2007 surveys of INSD. 13 Figure 2: Regional Typology of Burkina Faso Source: Nouvé et al., unpublished 2009; based on the QUIBB of 2003, 2005, and 2007 surveys of INSD. Risks, Shocks, and Vulnerability 35. Households in Burkina Faso face several categories of risks, such as: natural risks and food security, economic and social risks, regional stability, health, and other individual risks. These risks are a major factor in exacerbating the vulnerability of the poor to shortfalls in consumption and seasonal increase in the incidence of poverty. Natural Risks (Food Security) 36. The geographic location of Burkina is a source of vulnerability in itself. Located in Sub-Saharan Africa, Burkina is generally dry and its soil is infertile. In the north the climate is a semiarid steppe (Sahelian) that is characterized by three to five months of­often erratic­rainfall. To the south it becomes increasingly of the tropical wet- dry type (Sudanic), characterized by greater variability of temperature and rainfall and greater total rainfall. As in most other Sub-Saharan countries, the correlation between agricultural output and rainfall levels in Burkina is strong. High risks are associated with agricultural activities within the country because of the intra- and inter-year fluctuations. These fluctuations have immediate impacts on agricultural outputs and hence a farmer's income and welfare levels. Furthermore, Burkina Faso has historically relied predominantly on extensive agriculture to meet the food needs of its population. 14 Economic Risks (Terms of Trade Shock) 37. Agriculture constitutes more than one-third of Burkina Faso's GDP. Until the recent mining boom, cotton and livestock were Burkina's two major export items. The cotton sector provides the main livelihood for around 3 million people and accounts for around 60 percent of the export earnings. Recently the global market witnessed a downturn in cotton prices. This price shock was further compounded by a currency mismatch between export prices and production cost. Cotton is traded in dollars, but the farmers in Burkina Faso are paid in CFAF. Hence, as Burkina's currency is pegged to the euro, the recent euro appreciation compared to the dollar has had direct implications for cotton producers through lower profit margin. Therefore, the country's limited export capacity makes it vulnerable to volatility in world prices for agricultural products and leads to unpredictable swings in household incomes in rural areas. It also induces volatility in the country's export earnings and government finances. Such economic risks can lead to reduced income, either as a result of unemployment or of lesser income-generating activities in farming, small trade, and small-scale manufacturing. Geographical risks (regional stability dependency) 38. Burkina is a landlocked country that depends on its neighbors for the delivery of goods to and from the country and in some cases, for labor migration. This geographical dependence makes the country's economy vulnerable to any political instability in the region. For example, the conflict in Côte d'Ivoire affected the economy of Burkina Faso in many ways: (i) trade disruption due to the difficulty of transporting goods from and to the port of Abidjan; (ii) government revenue shortfalls due to lower tax revenues (customs); (iii) higher expenditures to cope with returning Burkinabés; and (iv) reduced worker remittances. Health Risks (Cost and Forgone Income) 39. Poverty and health are interrelated. Not only does poverty lead to poor health status, but poor health also contributes to monetary poverty. The country's poor health status is characterized by high rates of child and infant mortality, maternal mortality, and HIV/AIDS prevalence. Malnutrition is also widespread. Malaria is the main source of infant morbidity and mortality in Burkina, with diarrhea-related diseases and acute respiratory infections being the next most common causes. Recurrent epidemics of measles and cerebrospinal meningitis are also major concerns in Burkina Faso. Health risks imply loss of labor time either periodically, or permanently, leading to lower incomes and reduced abilities to cope with other shocks. Social Risks 40. Households in Burkina Faso face several lifecycle events, such as funerals and weddings as well as other risks like disability and injuries. This often requires significant extra expenditures and/or forgone income, thus reducing the resources available for essential needs. Households that are excluded or lack access to common resources and informal networks face serious difficulties. 15 41. Using the 2003, 2005, and 2007 QUIBB surveys Nouvé et al. (unpublished 2009) showed that across time, vulnerability varies across groups depending on household endowment and ability to satisfy their food needs and by location. The severity of the shocks and their frequency as well as their compounding factors can impact groups differently. The study was able to identify four different groups of households depending on their level of asset poverty and ability to satisfy their food needs: (i) households not poor and not vulnerable to lower food consumption; (ii) households not poor but vulnerable to lower food consumption; (iii) households poor but not vulnerable to lower food consumption; and (iv) households poor and vulnerable to lower food consumption. This typology could be used to identify the extent of chronic poverty. The category designated as poor and vulnerable to lower food consumption during the period of 2003-2007 could be considered as a proxy of chronic poverty. In this context, the chronic poor are those individuals, not only vulnerable to low food consumption, but also have low capacity of increasing their income, thus are expected to remain poor in the near future. Their share dropped from 25.4 percent in 2003 to 18.8 percent in 2007 after peaking at 31.1 percent in 2005 following the 2004 weather-related shock. Overall the percentage of poor went from 46.4 percent in 2003 to 40.4 percent in 2005 and 38.5 percent in 2007. Thus, it was the share of chronic poor to poor and vulnerable to poverty that changes in when shocks occur (Figure 3). Figure 3: Household Typology in Burkina Faso (National) Evolution des différentes formes de pauvreté au Burkina Faso: Ensemble du pays, 2003-07 100% 25.2 Proportion des différentes formes 30.0 42.2 de pauvreté (%) 23.7 34.3 19.4 21.0 9.3 19.7 31.1 25.4 18.8 0% 2003 2005 2007 Année Ménages à l'abri de la pauvreté et de la précarité Ménages plus aisés mais en précarité Ménages pauvres et non en précarité Ménages pauvres et en précarité Source: Nouvé et al., unpublished 2009; based on QUIBB 2003, 2005, and 2007 INSD surveys. 42. The picture is very different when comparing households in rural and urban areas. Chronic poverty in urban areas was almost cut by half (from 12.5 percent in 2003 to 6.7 percent in 2007), although it stagnated in 2005. However, chronic poverty did not decrease in rural areas at the same rate as in urban areas. It only decreased from 28.2 percent to 21.4 percent, with a peak of 35 percent in 2005 following the 2004 shock (Figure 4). 16 Figure 4: Household Typology in Burkina Faso (Rural/Urban) Evolution des différentes formes de pauvreté au Burkina Faso: Evolution des différentes formes de pauvreté au Burkina Faso: Milieu rural, 2003-07 Milieu urbain, 2003-07 100% 100% 20.4 25.3 Proportion des différentes formes Proportion des différentes formes 38.1 51.1 48.6 61.6 de pauvreté (%) de pauvreté (%) 22.5 34.2 18.1 24.0 10.4 22.5 29.0 35.1 25.4 35.0 7.4 28.2 4.1 21.4 6.3 12.5 12.3 6.7 0% 0% 2003 2005 2007 2003 2005 2007 Année Année Ménages à l'abri de la pauvreté et de la précarité Ménages à l'abri de la pauvreté et de la précarité Ménages plus aisés mais en précarité Ménages plus aisés mais en précarité Ménages pauvres et non en précarité Ménages pauvres et non en précarité Ménages pauvres et en précarité Ménages pauvres et en précarité Source: Nouvé et al., unpublished 2009; based on QUIBB 2003, 2005, and 2007 INSD surveys. Poverty Profile 43. As mentioned above, the 2003 priority survey has been used to determine the poverty profile of Burkina Faso. The 2003 survey provides the most recent available information on household consumption. The ongoing 2009-2010 household survey also includes information on household consumption and expenditures, but the data are not yet available. According to the 2003 priority survey, a typical poor household has seven or more members, is headed by a male who is polygamous, illiterate, or educated on a lower primary level, and is a farmer living in the rural areas of the Northeast Region. 44. Poverty remains a predominantly rural phenomenon. The proportion of poor people living in rural areas amounted to 92.2 percent, although their total share of the population was close to 10 percentage points lower (81.8 percent). Furthermore, poverty incidence in rural areas is more than two times higher than in urban areas (52.3 versus 19.9 percent). At 17.9 percent, the poverty gap in rural areas is more than three times the level observed in urban areas (5.5 percent). 45. The regional dimensions of poverty show that the Center Region is among the poorest, with a poverty rate well above the national average. This region houses over half of the poor in Burkina Faso (share of population) and also has the highest poverty gap. The Northeast and the South Regions are second, with poverty rates slightly below the national average, whereas the Capital area has the lowest poverty incidence at about half the national average (Table 4). 17 Table 4: Poverty by Area of Residence and Administrative Region, 2003 Poverty Incidence Share of Poverty Poverty Gap % % % NATIONAL 46.4 100 15.5 Area 100 Rural 52.3 92,2 17.9 Urban 19.9 7,8 5.5 Regions 100 Northeast Region Sahel 37.2 7.8 12.6 Center North 34.0 11.4 8.2 East 40.9 14.1 12.3 Center Region North 68.6 8.1 24.7 Boucle du Mouhoun 60.5 13.5 21.3 Center West 41.3 8.4 14.1 Plateau Central 58.6 4.9 20.3 Center South 66.1 4.2 26.0 Center East 55.1 7.2 19.7 Southwest 56.6 7.3 17.5 Capital /South Region Center 22.3 2.3 7.1 Hauts Bassins 34.8 6.7 10.6 Cascades 39.1 4.3 14.6 Source: Nouvé et al. unpublished 2009; based on the Priority/QUIBB 2003 INSD survey. 46. The socio-economic and educational level of the household head are key determinants of poverty. Male-headed households tend to be poorer than households headed by females both in terms of incidence (46.9 versus 36.5 percent) and gap (15.8 versus 12.7 percent). With a headcount of 55.1 percent, polygamous households are poorer than other groups classified by the marital status of the head. Households whose head is illiterate or with only a lower primary education level have notably high poverty incidence of 51.0 and 41.1 percent respectively. Wage salaried head of households whether in the public, formal, or informal private sector have considerably lower poverty incidence than their counterparts working in agriculture and specifically in the cultivation of cotton. Those who are engaged in independent activities (entrepreneurship) are also better off than those in agriculture but slightly worse off than wage salaried earners (Table 4). 47. Gender is also a key poverty correlate. Studies around the world have shown that individuals within the same household do not always have the same standard of living: income and resources are not necessarily pooled and members do not share in them equally. Differentiation within the household is typically most pronounced by age and gender. As in many parts of Sahelian West Africa, there are reasons to believe that some groups of women may be particularly poor and vulnerable. In much of Burkina Faso they are dependent on men. Single women have limited rights. They are married young to a husband that is on average nine years older. Marriage accords them protection and some rights like the use of land. Yet, such rights are contingent on marital status and are lost through divorce and widowhood. Given the large age gap between spouses, many young women find themselves widows with few rights once a husband dies. Almost all remarry, often into their husband's lineage which allows them to retain access to some land and 18 provides continued support to her and her children. This tradition of levirate marriage, whereby a widow marries a relative of her deceased husband, is slowly being undermined without adequate opportunities for women to support themselves otherwise. 48. Unfortunately household consumption surveys do not collect individual-level data detailing the intra-household allocation of resources, consumption, work, and time allocation. Thus, the poverty status or vulnerability of individual household members cannot be directly calculated from household surveys. One often-used approach is to compare female- and male-headed households to infer effects of an individual's gender on welfare. However, as in many countries, female-headed households in Burkina Faso are found to be extremely heterogeneous. Although as mentioned earlier, the overall incidence of poverty is lower for female-headed households, an analysis using the 2003 priority survey confirms that after controlling for various characteristics that affect living standards, there are few differences between female- and male-headed households. However, Demographic and Health Surveys (DHS) do contain information on individuals including some individual level welfare indicators. They have drawbacks for such an analysis given that they survey many more women than men and only collect detailed information on individuals aged 15 to 49. Figure 5: Past and Current Widows Have Lower Body Mass Indices Than Women of Other Marital Status Source: van de Walle, 2010. 19 Figure 6: Rural Children of Currently Widowed and Previously Widowed Women Are Less Likely To Be in School Source: van de Walle, 2010. 49. An analysis of the 2003 DHS for Burkina Faso reveals that widowed and currently married but previously widowed or divorced women may be particularly vulnerable groups in Burkina Faso. Unfortunately, the DHS cannot shed light on the large group of widows and other single women who are older than 49. However, we examine a few individual welfare indicators for women aged 15 to 49 by their marital status. Controlling for age, we find that in rural Burkina, widows and women married more than once are an especially vulnerable group compared to women who have been married once, or are currently divorced. Figure 5 shows that they have significantly lower body mass indices than these other women, although no such differences are apparent in urban areas. The data also indicate that the children of currently and previously widowed women are less likely to be in school than the children of women of other marital status (Figure 6). 20 Table 5: Poverty by Selected Characteristics of Household Heads, 2003 Poverty Incidence Share of Poverty Poverty Gap % % % NATIONAL 46.4 100 15.5 Gender Male 46.9 92.2 15.8 Female 36.5 7.8 12.7 Marital status Single 20.8 1.0 12.7 Monogamous 40.4 49.4 19.6 Polygamous 55.1 44.9 9.5 Divorced 28.7 0.5 13.1 Widow 39.9 3.9 5.6 Partners 19.4 0.3 - Education level None 51.0 93.7 17.2 Primary- lower 41.1 2.8 13.0 Primary - upper 24.8 2.7 7.6 Secondary ­ lower 19.4 0.8 6.3 Secondary ­ upper 3.1 0.1 0.9 University 0.3 - - Employment status Wage salaried ­ public 4.7 0.2 1.2 Wage salaried - private formal 6.5 0.1 1.9 Wage salaried - private informal 18.3 1.0 4.5 Cotton cultivation 45.5 19.3 14.8 Farmer 55.6 71.7 19.0 Independent (entrepreneurs) 21.5 2.4 5.9 Voluntary worker 39.2 0.7 14.1 Inactive 39.0 0.8 13.3 Unemployed 33.0 3.7 12.2 Household size 1 ­ 2 members 10.6 3.2 2.3 3 ­ 4 members 24.5 14.1 6.6 5 ­ 6 members 38.0 21.4 11.0 7 or more members 55.5 61.3 19.6 Source: Nouvé et al. unpublished 2009; based on the Priority/QUIBB 2003 INSD survey. D. Magnitude of Poverty and Cost of Making a Meaningful Difference 50. In assessing the financial cost of reducing poverty using transfers, a first question to ask is how much it would cost, at a very aggregate level and on an annual basis, to close the gap between the income of the poor and the poverty line. For illustrative purposes, the cost of bringing the income of all the poor to the poverty line is estimated in this section, using the poverty profile data for 2003 discussed earlier in this chapter, and based on the World Bank's Poverty Assessment Report (2005). The numbers below are built on a series of very simple assumptions and may contain a wide margin of error. Specifically, the estimates assume an average CPI-based progression of the poverty line since 2003, and a 3.1 percent increase of the population over that period. Poverty 21 depth is based on the 2003 national poverty depth (15.5) as shown in Table 5 above and is assumed constant. The estimates also quantify only the total amount of benefits needed to close the poverty gap. These calculations thus ignore issues such as administrative costs and possible efficiency losses. They also assume away the costs of targeting problems, giving only a rough baseline for the costs of fully effective cash transfers. 51. A minimum of 2.5 and 2.7 percent of GDP would have been needed in 2007 and 2008 respectively to pull all the poor to the poverty line through cash transfers. These estimates are based on the poverty lines and poverty gaps described in Table 6. It is also estimated that a minimum amount of 1.4 and 1.6 percent of GDP in 2007 and 2008 would have been needed to pull the chronic poor (the poorest and most vulnerable) to the poverty line through cash transfers (Table 6). These estimates for the level of chronic poverty in 2008 were derived assuming a constant ratio of 49 percent in 2007 and 2008 of chronic poor with respect to poverty, and an increase in poverty incidence to 45 percent in 2008. Thus, the number of the poorest and most vulnerable would increase to 22.1 percent of the population. These estimates are likely to be too conservative as the poverty depth of the poorest and most vulnerable group has been aligned on poverty depth in the rural areas where poverty is more acute (Table 6). Table 6: Estimated Aggregate Amounts of Bringing All the Poor and the Poorest and Most Vulnerable to the Poverty Line through Cash Transfers (2003, 2005, and 2007) 2003 2005 2007 2008 GDP (CFAF billion, current prices) (1) 2513.8 2862.8 3239.4 3647.9 Population (millions) (2) 12.6 13.4 14.3 14.7 Per capita GDP (CFAF) 199,206 213,428 227,200 248,155 Dollar/CFAF (annual average) 519.3 449.9 481.5 Per capita GDP (USD) 505.0 515.4 CPI index (average) 2.04 6.41 -0.25 10.67 Poverty line, basic needs method (3) 82,672 88,298 94,308 97,464 Poverty incidence, wealth/asset-based (4) 46.4% 40.4% 38.5% 45% Number of Poor (millions) 5.9 5.4 5.5 6.6 Poverty depth assumed constant (percent) (5) 0.155 0.155 0.155 0.155 Poverty gap (CFAF billions) 75 74 80 100 Poverty gap as % GDP 3.0% 2.6% 2.5% 2.7% Proportion of poor and vulnerable (% population) (6) 25.4% 31.1% 18.8% 22.1% Number of poor and vulnerable (millions) 3.21 4.17 2.68 3.24 Poverty depth (7) 0.18 0.18 0.18 0.18 Poverty gap 47.7 66.3 45.5 56.9 (1) IMF ; (2) CEM, WB ; (3) 2003: PA WB, p.16 (graph); indexation from 2003 based on average CPI 2003- 2009; (4) 2003-2007: QUIBB; 2008: based on CAS but methodology uncertainties; (5) INSD, 2003, assumed constant; (6) 2003-2007: Nouve 2009 (unpublished), based on QUIBB; 2008: assumed same ratio as in 2007 between very poor and poor;(7) Bank staff estimates based on rural poverty depth 2003 (INSD). Source: Authors. 52. These estimates, while tentative, should be considered as a minimum. Viewed in relation to existing budgetary numbers, these costs would amount to significant financial commitments by the government. For example, the minimum cost of fully closing the poverty gap in 2007, estimated at 2.5 percent of GDP, was roughly equal to less than 10 percent of the total budget in 2007. This amount also is equivalent to less than half the 22 government wage bill in 2007 (6 percent of GDP) or to a little more than total spending on health (2.3 percent of GDP) for the same year. 23 CHAPTER III: GOVERNMENT SOCIAL SAFETY NETS POLICIES AND INSTITUTIONS Burkina Faso has not adopted a consolidated national Social Protection policy, which would provide a framework for social safety nets. Nevertheless, social safety net programs play an important role in Burkina's human capital development and crisis response strategies. Social safety nets are identified as necessary social and economic investments in the country's long-term vision of society. Yet, their potential to reduce poverty and vulnerability through income redistribution, promotion, and transformation is often not recognized in medium-term strategies. Overall, the existing social safety net system has a limited impact on the poorest and most vulnerable. The main reasons are limited coverage due to low financial allocation and poor targeting efficiency for lack of implementation capacities ­ in particular, difficulties in reaching the very poorest and measuring cost-effectiveness. Until recently, there has been a lack of leadership in the domain, as well as weak cross-sectoral coordination mechanisms. Of late, however, the context of repeated crises has brought social protection and its social safety net component to the fore and the Government has taken several actions to address the need of developing an effective social protection policy such as the recent creation of the Joint Ministerial Committee on Social Protection and the perspective of the integration of a multisectorial social protection component into the upcoming SCADD, thus opening up opportunities to see social protection, and social safety nets, play a more pronounced and efficient role in poverty reduction strategies. Excluding the cost of general subsidies for food and fuel, total spending on social safety net programs over 2005-2009 averaged about 0.4 percent of GDP. As a result of greater engagement by the government and donors, the spending on social safety net programs has continuously increased from 0.3 percent of GDP in 2005 to about 0.9 percent in 2009. This increase is largely explained by expenditures made in food-based programs. Since 2006, the financing for SSN programs has become increasingly donor-dependent; yet donors focus predominantly on nutrition programs while government financing mainly supports universal food and fuel subsidies 53. Following the poverty, vulnerability and risks diagnostic presented in Chapter II, this Chapter explores whether the existing social safety net system constitutes an appropriate response to poverty and vulnerability. The first section reviews the objectives of Burkina Faso's development strategy, looking at their interaction and complementarities with other public policy responses to poverty and vulnerability. Institutional arrangements for social safety nets are reviewed in the second section. The following section considers the financing of the system, estimating total spending on social safety nets and discussing funding sources. A. The Strategic Role Given to Social Safety Nets within the Larger Development Policy Social Safety Nets within Burkina's Medium-Term Poverty Reduction Strategy and Long-Term Development Vision 54. The latest Poverty Reduction Strategy Paper (PRSP) includes social protection as a clear aspiration in its second pillar, "Promoting access by the poor to basic social services and social protection" [MEDEV 2004]. In 2000, Burkina Faso was the first 24 country in the sub-region to adopt and implement a PRSP, with a strong focus on the education and health sectors.7 The second and current PRSP developed in 2003, along with 13 regional PRSPs, broadened the social sectors covered and specifically included social protection in its second pillar.8 The latter aims at ensuring access by the poor to basic social services and social protection by strengthening human capital (i.e., education, health, nutrition, HIV/AIDS, drinking water, sanitation, etc.), promoting access of the poor to education, health services, nutrition programs, the fight against HIV/AIDS, drinking water, housing, and social protection. Moreover, national solidarity appears at the core of the long-term vision of society that the country developed in 2005.9 Beyond social and political gains, solidarity, and the extension of social safety nets, is expected to provide the poorest citizens with the means to contribute to production and consumption, and social safety nets are identified as necessary social and economic investments. 55. However, social safety nets are not explicitly considered as a key element of the poverty reduction strategy, which remains focused on supply-side interventions. To promote social protection, four strategic areas for action were identified in the PRSP: (i) strengthening action to fight all forms of social exclusion; (ii) encouraging actions to promote solidarity; (iii) increasing the incomes of the poorest; and (iv) promoting health micro-insurance for the poorest and indigent population groups [MEDEV 2004]. In this context, measures largely focused on the provision of social welfare services for vulnerable groups, the promotion of rights and adoption of new legislation, the provision of income- generation activities, and contributory social insurance schemes (e.g., health mutuals). Thus, in this strategy, the use of direct non-contributory transfers to individuals or households is limited to a few vulnerable groups (e.g., vulnerable school children, HIV- affected children, and the disabled) on a pretty ad-hoc basis. 56. Poor performance of the previous PRSPs and recent crises stressed the need for more comprehensive social protection, and social safety net programs in particular. The PRSP currently under revision is to be replaced in 2010 by the Strategy for Accelerated Growth and Sustainable Development (Stratégie de Croissance Accélérée et de Développement Durable or SCADD). As of February 2010, five pillars were identified to guide the development of this envisioned five-year strategy: accelerate economic growth; reduce poverty; manage demography; preserve the environment; and reduce inequalities between men and women [GoBF 2010]. The extension of social safety nets ­ along with improved access to basic services, universal school enrolment, access to employment, etc ­ is a priority action explicitly identified under the second pillar of the upcoming SCADD. In April 2010, the Ministry of Economy and Finance (MEF) organized a national technical forum on social protection with the stated goal to outline a national and consolidated vision of social protection, including the social safety net component, and to directly inform the drafting of SCADD. 7 The Letter of Intent on Sustainable Human Development Policy (LIPDHD) adopted in 1995 served as the basis for the Poverty Reduction Strategy Paper. 8 Other PRSP pillars are: 1: Accelerating broad-based growth; 3: Increasing employment and income- generating activities; and 4: Promoting good governance. 9 Strategies identified in the long-term vision of society "Burkina 2025" to make national solidarity effective are: equitable redistribution of wealth; fight against social discriminations; education and employment; balanced and harmonious territorial development; and citizen mobilization and moral rearmament [MEDEV 2005]. 25 Social Safety Nets and Burkina's Approach to Social Protection 57. Burkina Faso has no consolidated social protection policy that has been adopted by the government. In 2006, the Ministry of Labor and Social Security (Ministère du Travail et de la Sécurité Sociale or MTSS) led an inter-ministerial process to develop a social protection national policy (Box 1). This document presents an interesting multisectoral approach, in which social protection is seen as an investment rather than a form of relief. Recognizing that needs for social protection are huge but resources limited, the document proposes a number of initiatives aimed at ensuring the most vulnerable groups are really the ones benefiting, and seeking maximum program cost-effectiveness in the development of the system [MTSS 2007]. Social safety net programs appear at the core of the proposed policy. Although this document was finalized in January 2007, it was never adopted and the government requested each ministry to develop its own ministerial strategy ­ the MASSN thus developed its Social Action National Policy. 58. Social safety nets do not appear as a key strategic element in the Social Action National Policy. This ministerial policy (Politique Nationale d'Action Sociale or PNAS) developed in 2007 by the MASSN in line with the PRSP's second pillar, aims at ensuring the protection and promotion of specific social groups, and at creating favorable conditions to a consolidation of solidarity within the populations. It builds around five components: improvement of household living conditions; promotion of the culture of national solidarity,10 protection and promotion of specific groups; contribution to the fight against HIV/AIDS; and institutional capacity strengthening [MASSN 2007]. The proposed activities, which are restricted to the MASSN's mandate, focus on education and awareness-raising, provision of social welfare services (e.g., for the disabled, vulnerable children, etc.), legal protection, and support to income-generating activities. Therefore, little consideration is given to social safety nets. The policy does consider the provision of financial support to Orphans and other Vulnerable Children (OVC), disaster-affected persons, or as part of solidarity actions but does not elaborate on any form of institutionalized regular and predictable social transfers to poor or vulnerable individuals or households.11 59. The Government is taking several initiatives recently toward the development of a national social protection strategy, and the prospect of the integration of a social protection component into the upcoming SCADD presents an opportunity to again promote the required multisectoral approach to social protection and social safety nets. As discussed in the next section, social protection is gaining momentum, and both national and international stakeholders are trying to build national consensus among various ministries and agencies. They are also mobilizing efforts to promote an integrated social protection strategy, which will also include an efficient social safety net system. 10 Solidarity is defined as a "humanist action" in the recently adopted Solidarity National Charter, and is at the core of the MASSN's action. This approach to solidarity may differ somehow from the notions of income redistribution and necessary social investment supported by social safety nets as defined here. 11 In 2009, the MASSN proposed a new program entitled, Social Safety Nets [MASSN, 2009a], but this draft proposal received serious criticism from both national and international stakeholders. It was largely judged as ill-designed to constitute an effective safety net, and overambitious considering the MASSN's current financial, technical, and human resources. The program was thus not considered in the 2010 budget. 26 Box 1: Outline of Social Protection National Policy Drafted under MTSS Leadership (Not Adopted) Overall objective: Mitigate the impact of shocks that directly threaten the life of the most vulnerable groups of the population. Specific objectives: Ensure for the most destitute or vulnerable groups access to basic social services (mainly health and education) and temporary incomes in case of shocks threatening the survival of the household; and Create an environment favorable to the development of social protection through market mechanisms for those who can contribute. Guiding principles: National solidarity; equity and equality of treatment; diversification of instruments and actors; overall responsibility of the state; and democratic management. Pillars: 1. Ensuring access for indigent populations and particularly vulnerable groups to basic social services; 2. Ensuring an adequate social cover for the actors of the informal sector; and 3. Reforming the institutional social security and strengthening its legitimacy. Proposed activities (for the 2006-2008 transitional period): Continuation of existing programs and evaluation of their cost-effectiveness: Health: malaria and HIV/AIDS programs, introduction of subsidies for obstetrical care; Education: school feeding, water supply, community-based schools, provision of school materials and waivers for parental contributions in destitute areas, and MASSN-supported education programs (for street children, children with learning difficulties, pre-primary education); Income-generation and job creation: programs' efficiency improvement; Food security: national food security stock; and Risk management and assistance to natural disaster-affected persons: program against locust invasions. Feasibility studies and pilot projects to test: Subsidized health services for the most destitute groups of the population (under the supervision of MoH); Nutrition minimum package for children under 5 (under the supervision of MoH); Conditional cash transfers to improve school attendance (under the supervision of MEBA); and Labor intensive public works to respond to situations of high vulnerability (under the supervision of MJE). Program proposed by the MASSN to assist OVCs. Cross-cutting activities to improve existing programs: Definition and identification of the indigent (under the supervision of MASSN); Review of the mutual benefit insurance sector (under the supervision of MTSS); Preparation of a reform of the formal social security system (under the supervision of MTSS); and Evaluation of existing programs (under the supervision of MEDEV). Estimated annual budget: CFAF 37.9bn (1.1 percent of GDP). Institutional arrangements: Social Protection Steering and Monitoring Committee (lead: Prime Minister's representative; members: representatives of relevant ministries and organizations); Permanent Secretariat (hosted by MEDEV); execution by respective agencies, through a contractual approach whenever possible. Source: MTSS, 2007. Social Safety Nets within Burkina's Sector-Based Social Policies 60. Despite the absence of a national social protection strategy and lack of consideration of social safety nets in the social action national policy, social safety net programs appear in many sector-based strategies: health, education, food security, and employment. 27 Health Policy 61. Demand-side interventions, such as fee waivers, are included in the strategy to ensure an efficient access to health services, and act in synergy with efforts to improve the quality and availability of health services. The Health Development National Plan (Plan National de Développement Sanitaire or PNDS, 2001-2010) includes specific activities to: (i) promote the preventive and curative care of vulnerable groups defined as children, women, the youth, the aged, and HIV-infected persons (under specific Objectives 3 and 4); and (ii) ensure access to medical care for the poor (under specific Objective 6) [MoH 2000]. 62. A number of services and medicines are proposed at subsidized prices for all. The direct payment for health services by users has existed in Burkina Faso since the 1980's. It was reinforced with the implementation of the Bamako Initiative (adopted in 1987), which became effective in Burkina Faso in 1993. In the following years, a number of measures were adopted to abolish or subsidize user fees, particularly for women and young children, as a strategy to reduce maternal and neonatal mortality rates and promote access to health services for the poor. In particular, the government adopted in 2006 a subsidy policy based on the principle that it would meet 80 percent of the cost of deliveries and emergency obstetric and neonatal care ­ up to 100 percent for destitute (indigent) women. A total budget of CFAF 30bn was allocated for this over the period 2006-2015, including CFAF 5bn for additional subsidies to the indigent. General health subsidies were not considered as social safety nets because they are not specifically targeted at the poorest and most vulnerable, only the additional subsidies to the indigent were included as part of the social safety net programs. 63. Mechanisms are considered to provide additional subsidies to the indigent. The indigent are entitled to free heath care as per the measures of the Kiti n°An-VIII- 0202/FP/SAN-AS dated 8 February 1991 relative to the general pricing basis. However, national mechanisms are yet to be clarified to make this right effective. A few isolated initiatives exist, but serious challenges remain. This is particularly true in regard to identifying the indigent (criteria and mechanisms) and financing of these additional subsidies. 64. The envisioned health insurance national scheme also considers contribution waivers for the indigent (Figure 7). However policy-makers are faced with the same key challenges of: how to identify the indigent, finance the system, and then implement it. 28 Figure 7: Health Insurance National Scheme Project's Institutional Arrangements Légende : CNAM : Caisse Nationale d'Assurance Maladie CRAM : Caisse Régionale d'Assurance Maladie AMO : Assurance Maladie Obligatoire AMV : Assurance Maladie Volontaire SI : Système d'Information Source: Permanent Secretariat of the Health Insurance Steering Committee. Education Policy 65. Demand-side measures like education fee waivers are at the core of the Education for All policy, and children aged 6-16 are entitled to free education, as stipulated in the Education Reform Act adopted in 1996. Concrete measures to ensure free education include fee abolition in public schools and free distribution of essential school materials in both public and private schools. These measures were initiated in all provinces by 2007-2008, with a total annual budget of CFAF 6.5bn (funded by both the state and the CAST). In addition, the state has been covering the contribution made by APE12 for all girls enrolled in public school in the CP1 grade, with an annual average of 192,700 girls benefiting from the CFAF 1,000 fee waiver, for a total annual cost of about CFAF 190m (largely supported by the CAST). This policy has resulted in a significant increase in the school enrolment rate: (i) overall enrolment rose from 47 percent in 2002- 2003 to 71 percent in 2008-2009; (ii) the enrolment rate of girls progressed significantly, from 47.7 percent in 2002-2003 to 72.4 percent in 2009-2010; and (iii) the girl­boy parity index increased from 0.72 in 2002-2003 to 0.86 in 2009-2010. Overall, the measures to provide free primary education for all, has played a relevant role in terms of improving access to education. 12 Contribution made by the Parent­Teacher Association (Association de Parents d'Elèves or APE). 66. A number of social safety net programs complement these universal measures and supply-side interventions to further increase school enrolment, attendance, and completion rates while reducing disparities [MEBA 2000]. These include: school feeding in both primary and secondary schools, scholarship programs, and a pilot (conditional) cash transfer program. The programs will be further discussed in Chapter IV. Food Security Policy 67. Some social safety net programs appear explicitly in the Food Security National Strategy. This strategy (Stratégie Nationale de Sécurité Alimentaire or SNSA), developed by the Ministry of Agriculture in 2001 and revised in 2003, includes the provision of nutrition programs and access to health care for malnourished pregnant and nursing mothers and children under 5 (under its Objective 3, "Improve in a sustainable manner economical and nutritional conditions of poor and vulnerable populations"). The emergency food distributions through decentralized mechanisms and the national food security stock constitute the core of the strategy to protect populations affected by natural disasters (under its Objective 4, "Strengthen the prevention and management of cyclical crises in consistency with structural food security building"). Employment Policy 68. The promotion of a (systematic and wide) use of labor intensive approaches is set as a clear ambition of the National Employment Policy adopted in 2008, which aims to strengthen targeted measures for employment promotion. The Public Investment program offers a broad range of infrastructures that could be realized and/or maintained using labor intensive approaches. However, while the policy aims at creating job opportunities through this approach, it is not clear if the design of the programs would facilitate participation of the poorest and most vulnerable groups in temporary work activities. Social Security Nets within Burkina's Crisis Responses 69. Direct transfers to individuals and households are at the core of Burkina's crisis response strategies. The emergency food security and nutrition plan that the government adopted in 2008 in response to the high world food and fuel prices includes, overall the same types of programs, such as short-term nutrition interventions. The country's contingency plan includes, for instance, food distributions to disaster-affected households, nutrition programs for malnourished children and women, school feeding programs for displaced pupils, and child and mother health care services, among others. 70. However, the action plan 2009-2010 developed to mitigate the financial and economic crisis is actually the first strategic document to incorporate a specific focus on social safety nets. The fourth pillar of this crisis action plan ("Strengthening of social protection programs") includes three components: (i) implementation of social safety nets and innovative mechanisms to reduce household charges; (ii) supervision of a pricing structure; and (iii) implementation of a special program for employment and labor intensive public works [MEF 2009b]. More specifically, the proposed activities under the "social safety net" component of the crisis action plan include both existing activities (SNS and school feeding) and new activities (free prenatal consultations) (Table 7). Nevertheless, as in the PRSP and Social Action National Policy, the term "social protection" is used in a very narrow way, and the Social Protection of Vulnerable Groups 30 program mostly refers to social welfare services ­ rather than covering the larger spectrum of social insurance, social safety nets, policy and regulations, and social welfare services (Table 8). It should be noted that this package of measures was voted with the supplementary 2009 budget intended to be disbursed before the end of February 2010. Whether the resources were made available in 2010 remains unclear. Table 7: The "Social Safety Net" Component of the Crisis Action Plan 2009-2010 (CFAF million) Activity 2009 2010 2010 required required secured supplementary supplementary funding as of budget budget December 2009 Program 1: Food and nutrition security 8,446.8 31,466.2 175.2 Food distributions and construction of food 8,446.8 30,050.9 175.2 storage rooms Program to fight mother and child malnutrition - 1,415.2 - Program 2: Education 12,582.0 12,582.0 4,348.2 Subsidies for girls newly enrolled in primary 260.0 260.0 - school MEBA's school feeding program 12,322.0 12,322.0 4,348.2 Program 3: Health 2,000.0 26,602.3 19,477.9 Subsidy program for obstetric and neonatal care 1,702.0 1,162.4 2,500.0 Program to fight TB - 3,521.8 50.0 Program to fight malaria - 8,636.7 55.0 Extended immunization program - 6,901.5 1,847.3 Free prenatal preventive consultations - 400.0 469.1 Program to reduce morbidity associated to - 690,2 200.0 lymphatic filariasis ART for HIV-infected persons - 3,517.0 13,378.9 Others (supply-side interventions) 298.0 5,289.7 977.6 Program 4: Social protection of vulnerable groups 865.0 2,702.6 1,038.1 Total 23 893.9 73 353.3 25 039.5 Source: MEF, 2009b. Table 8: The "Social Protection of Vulnerable Groups" Sub-Program of the Crisis Action Plan 2009 Activities of Program 4 "Social Protection of Vulnerable Groups" Share of Program Budget Allocated (%) Food and kits support to preprimary schools 31 Materials for the disabled 26 Income-generating activities in specialized education centers 18 School fees, school kits, and bikes to 1,500 OVCs 18 Food support to orphanages 4 Food support to cours de solidarité 1 Administrative cost 2 Total 100 Note: The 2010 Social Protection for Vulnerable Groups program include the same activities, as well as food support to OVCs and the indigent sick in hospitals. Source: MEF, 2009b; Authors. 31 71. Overall, the review of existing policies and strategies reveals that policy- makers in Burkina Faso do recognize a significant role for SSNs in development policy, and that further work is still needed to define a comprehensive and concrete action plan. Among the four roles SSNs can play in development policy, their role in human capital development is probably the most recognized one in Burkina Faso ­ as illustrated by the country's health, nutrition, and education policies. The protective role SSNs can play to mitigate the negative impact of shocks also appears largely acknowledged ­ as reflected in Burkina's crisis response plans. Nevertheless, further work is still needed to define and adopt a concrete action plan for social safety nets. One that presents time-bound targets with corresponding human and financial resources allocated to each task and the designation of institutions responsible for implementation. 72. Moreover, this strategic analysis also suggests that SSNs have not been used to their full potential. Up to this point, the scope of the existing social safety nets has been limited and the level of transfers has been small. The approach to improve economic conditions of poor and vulnerable households has been centered on the promotion of a culture of national solidarity and the support to income-generating activities. The potential role of SSNs to improve livelihoods (through better risk management and promotion of independence) is yet to be incorporated. Finally, the lack of a national strategy for SSNs illustrates the fact that the government has yet to see the role that a coherent and efficient SSN system can play to support efficiency and growth ­ by freeing other sectors from any redistributive role and letting them concentrate on the efficient provision of services. B. Institutional Arrangements of the Social Safety Net System At the National Level 73. The absence of a national strategy for social safety nets goes along with a lack of institutional coordination in the domain. A draft social protection national policy, developed by an interministerial committee under the coordination of the Ministry of Labour and Social Security (MTSS), had identified priority areas in regard to safety nets. Unfortunately, this draft was not nationally adopted and each Ministry involved opted for a sectoral policy focused on its own mandate. Thus, the MTSS, in its draft labor national policy focuses on labor regulation and social insurance. As part of this social insurance, it is envisioned to set up a health insurance system designed to become universal which will include safety net programs ­ inter alia health subsidies for vulnerable people, cash transfers for the elderly and households with children. As for the MASSN, whose mandate13 suggests a strategic focus on social safety nets, it has actually designated no clear priority to SSNs in its social action national policy and does not manage any significant social safety net program. For instance, in 2008, the FNS assisted only 4,562 persons (less than 0.07 percent of the monetary poor) and about 10 organizations through one-off in kind or cash transfers for a total cost of CFAF 38.8m. MoH, MEBA, and MAHRH appear to be the most active in the implementation of SSNs, despite the fact that their mandate suggests that they concentrate on the provision of basic service delivery. 13 The MASSN has six domains of intervention: child and adolescent protection and promotion; family protection and social promotion; promotion of national solidarity; protection and social promotion of specific groups (the disabled, aged, and socially excluded); promotion and socioeconomic and psychosocial support to HIV/AIDS-infected and -affected persons; and institutional competence and capacity building. 32 74. Overall, inter-ministerial coordination mechanisms on social protection and social safety nets are weak. The PRSP Sector-based and Thematic Commission (Commission Sectorielle et Thématique or CST) for social protection has not been very active over the past few years, and has not been aligning ministries that manage the social safety net programs.14 In fact, until recently there has been no institutional setup for sectoral ministries to coordinate their social safety net strategies and exchange their experiences.15 75. The repeated crises of have lent momentum to the uptake of social protection and particularly to social safety nets. Late in 2009, MEF coordinated the development of an action plan to mitigate the financial and economic crisis, which incorporated social safety net programs explicitly within a priority social protection pillar. In January 2010, a training workshop on social protection was offered to 30 technicians from 11 ministries, with UNICEF support. In February 2010, a Joint Ministerial Committee on Social Protection was established by decree. In April 2010, MEF organized, with UNICEF support, a National Technical Workshop on Social Protection in order to: (i) disseminate the basic principles of social protection; (ii) share information on the existing social protection programs in Burkina Faso; and (iii) outline a consolidated and multisectorial vision of the role of social protection in the country. The workshop gathered about 90 persons and marked a significant step toward the development of a national social protection strategy, incorporating social safety nets, and its integration into the upcoming SCADD. 76. The creation of the Joint Ministerial Committee on Social Protection and the perspective of the integration of a national social protection strategy into the upcoming SCADD has created opportunities for social protection, and social safety nets, to play a more prominent and efficient role in poverty reduction strategies. The integration of a social protection component into SCADD is an important opportunity to adopt the necessary multisectoral approach; one that largely reverses the sectoral approaches that have prevailed in the past. Currently, implementing agencies face common issues such as difficulties to identify the poor, introduction of innovative program types, and monitoring and evaluation of the programs. The establishment of the Committee will support cross-ministerial dialogue on SSN design and implementation issues, enable the capitalization of national learning, and promote economies of scale. At the Decentralized Level 77. As outlined in the recent Public Expenditure Review, program and expenditure management remains largely centralized and ineffective. Despite a rapid increase since 2005, the resources of local authorities accounted for only 0.8 percent of GDP in 2007 and 4.0 percent of all state revenue [World Bank 2009a]. Centralized expenditure management is one explanation of the "Burkina Paradox" that is, an apparent 14 It is only five ministries: the Ministry of Social Action (Ministère de l'Action Sociale et de la Solidarité Nationale or MASSN), the Ministry for Women Promotion (Ministère de la Promotion de la Femme or MPF), the Ministry of Labor and Social Security (Ministère du Travail et de la Sécurité Sociale or MTSS), the Ministry of Employment and Youth (Ministère de l'Emploi et de la Jeunesse or MEJ), and the Ministry of Sports and Leisure (Ministère des Sports et Loisirs or MSL). 15 The Inter-ministerial Committee for the Monitoring of Economic Conditions (Comité Interministériel de Suivi de la Conjoncture or CISC) created in July 2009 to coordinate the crisis action plan takes place at ministerial level under the leadership of the prime minister. 33 disconnection between good economic performance and management and poor results in terms of poverty reduction. Creating the conditions of effective decentralization is key to improving public expenditure efficiency in terms of poverty reduction. Under the National Plan for the Organization and Coordination of Emergency Relief and Rehabilitation (Plan National d'Organisation et de Coordination des Secours d'Urgence et de Réhabilitation or PNOCSUR), Councils for Emergency Relief and Rehabilitation16 are meant to be functional at the national, regional, provincial, departmental, and village levels to coordinate crisis responses and prevent the exclusion of the vulnerable. At the Donor Level 78. Burkina Faso is a pilot country for the Global Social Protection Floor Initiative. Major donors engaged in social protection in Burkina Faso ­ namely, ILO, UNICEF, WFP, World Bank, IMF, EC, and UNDP ­ have been coordinating efforts closely. In 2009, in a context of a global crisis threatening to roll back decades of investment in favor of human development and in pursuit of the MDGs, the United Nations established the Social Protection Floor Initiative aimed at building a basic set of social protection guarantees for all citizens (Box 2). Burkina Faso was selected as one of the pilot countries along with Cambodia, Ethiopia, and Honduras.17 16 For example, Conseil National de Secours d'Urgence et de Réhabilitation or CONASUR at the national level. CONASUR is an inter-ministerial structure composed of about 15 ministerial departments, NGOs, and the national Red Cross Society, reporting to the MASSN. 17 This is supported by an ILO and European Commission project, managed from Geneva, which aims to provide a platform for awareness-raising and the exchange of good practice in social protection and employment. 34 Box 2: The Social Protection Floor Initiative This global financial and economic crisis will have dramatic social, health, hunger, and education effects unless decisive action is taken. The global crisis threatens to roll back decades of investment in favor of human development and in pursuit of internationally agreed development goals, including the MDGs. In crisis conditions, social security benefits, public health and nutrition programs, and social services act as social, health, and economic stabilizers thereby curtailing the potential social and economic depth of the recession, through avoiding poverty, ensuring continuity in services, and stabilizing aggregate demand. The international community should not just repair the problems identified by the crisis in global financial, monetary, and economic systems, but should protect people during the crisis, and thereafter. The crisis is also an opportunity to build a basic set of social protection guarantees for all citizens. A social protection floor could consist of two main elements that help to realize respective human rights: (i) Essential Services: ensuring the availability, continuity, and access to public services (such as water and sanitation, health, education, and family-focused social work support); (ii) Social Transfers: a basic set of essential social transfers, in cash and in kind, paid to the poor and vulnerable to enhance food security and nutrition, and provide a minimum income security and access to essential services, including education and health care. Calculations by various UN agencies including ILO, UNAIDS, UNICEF, and WHO show that a basic floor of social transfers is globally affordable, even if the funding is not yet available everywhere. It would also have a major impact on poverty, access, and use of key services including those for AIDS, tuberculosis, and malaria, and on child labor and child trafficking. When properly implemented, already-existing cash transfer and basic health systems in many developing countries have positive impacts on poverty, child labor, health and nutrition, education, social status of recipients, and economic activity. The social protection floor provides a conceptual catalyst for a UN-led global coalition to safeguard the attainment of the Millennium Development Goals (MDGs). Source: ILO and WHO, 2009. C. Financing of the Social Safety Net System 79. The main social safety net programs currently in place in Burkina Faso are summarized in Table 9. Design, implementation, and financing issues of each program type will be discussed individually in Chapter IV. As shown in Table 9, over thirty social transfer programs are currently funded by various donors. While each of these has merits, most do not reach more than a few tens of thousands of beneficiaries, nor do they deliver sustained benefits to the poor. In this section their financing will be reviewed. 80. Quantifying spending on safety nets is difficult. Currently the SSN programs are managed and implemented by various ministries, and there is no centralized monitoring system in place. Therefore tracking spending on safety nets programs is difficult and the lack of detailed data raises challenges in estimating the respective level of spending (Box 3). 35 Box 3: Information Issues Related to Social Safety Nets Overall data on social safety nets (SSN) is incomplete. Any attempt to cost out existing SSN programs confronts serious problems of data availability and reliability. Specifically: Data on extra-budgetary expenditure by donors and NGOs may at times not be in a comparable format. In particular, data on some non-governmental spending cannot be annualized, thus making it hard to conduct any time series analysis. The multitude of donors in the sector is also striking and complicates data collection and analysis. As a result, government spending, which is presented on an annual budget basis, may be over-represented relative to other contributors. Moreover, even for donor-financed programs, it is likely that spending appears smoother than it is in reality. For budgeted expenditure, as with social protection expenditure, it is apparent that the budget classification system is not well adapted to any decision-making in social safety nets. For example, no estimate is available for free health care for the poor or for the amounts spent on the indigent. These expenditures are not budgeted, and unless a specific reporting is organized, financial information on the amounts represented by the subsidy and cash transfers are not available. Whenever possible, amounts are presented in a net basis that is, the actual amount of the benefits paid to target beneficiaries. In some cases, however, program costs cannot be broken down between benefits and other costs (management costs, transport costs, etc.). In this case, data are mostly provided on a gross basis, which may lead to overestimating actual benefits. Source: Authors. 36 Indicative number Indicative Funding Implementation Geographical Program title, Year(s) Target group(s) of beneficiaries annual source(s) agency(ies) area(s) spending Table 9: List of Main Social Safety Net Programs Currently in Place in Burkina Faso, 2008-2009 CASH & NEAR-CASH TRANSFERS Cash transfers to OVCs (PASS component SP-CNLS-IST for OVCs Province of 6,500 children CFAF 270m World Bank 2) Oct 2008-Sept 2010 / Research-action on interventions, and World Nahouri (2009) (2009) community-based care of OVCs in the Bank for research Nahouri Urban voucher program (EMOP 10773.0) ­ WFP with MASSN, CRS, Urban extreme poor Towns of 338,915 persons CFAF 8,501m Multilateral Jan 2009 to June 2010 BRC Ouagadougou & (2009) (2009) Bobo-Dioulasso Food vouchers to HIV-infected women and CRS with OCADES Pregnant women and Regions of Center, 2,220 persons (2009) CFAF 176m Global Fund children) ­ Dec 2008-Aug 2009 children under 2 infected Center South, (2009) by HIV Center East, Center West, High Basins, Cascades Food voucher to urban poor ­ Sept 2008- CRS with OCADES Urban poor households Towns of 1,900 households CFAF 112m Gates Feb 2009 Ouagadougou & (2008-2009) (2008-2009) Foundation Bobo-Dioulasso FOOD TRANSFERS Targeted subsidized sales of food Intervention Stock SONAGES with As per Food-insecure 1,800,000 persons CFAF 1.9bn State CONASUR COPROSUR/CODESUR- areas (2008) (2008-2009) defined criteria Targeted food distributions General Relief CRS with partners PLHIV, orphans, 23 provinces 14,850 persons CFAF 242m USAID malnourished children, (2008) (2008-2009) aged, widows, etc. Food assistance to vulnerable populations MASSN with CONASUR Vulnerable populations and National 44,300 persons CFAF 31m State and disaster-affected households disaster-affected (2008) (2002-2008) households 37 Food for Education (CP10399.0 World Food Program with Persons attending literacy 11 provinces 20,300 women CFAF 224m Multilateral, component 1: ii. Literacy) ­ 2006-2010 partners centers (Sahel, East, 16,348 men (2009) (2009) State Center and Center North) Nutrition Nutritional supplement to malnourished World Food Program with Moderately malnourished Sahel, North, East, 312,000 persons FCFA 4,900m Multilateral, children and pregnant and nursing partners children and pregnant and Southwest, (2009) State mothers (PRRO 10541.0 2007-2009 and nursing mothers Center-North and PRRO 20054 2010-2011) Center-South Nutritional support to vulnerable groups World Food Program with Moderately malnourished Center, North, 20,500 persons FCFA 2,175m Multilateral, and PLHIV (CP10399.0 component 2) ­ partners persons Center-North, (2009) (2009) State 2006-2010 Center-South, East, Southwest, Cascades, High Basins, Mouhoun) Treatment and prevention of acute UNICEF with partners Severely malnourished 5 regions most 17,000 children FCFA 3,790m Multilateral malnutrition for child development and children under 3 and affected by (2009) (2009) survival ­ 2006-2010 pregnant and nursing malnutrition mothers Treatment of severely malnourished Médecins Sans Severely malnourished 23 CSPS in the 18,000 children FCFA 1,476m MSF, ECHO children ­ 2007-2010 Frontières-France with children aged 6-59 months Passoré and (2009) CSPS Loroum Provinces Reduction of child mortality and Burkinabe and Belgian Pregnant and lactating 180 villages in n/a FCFA 472m ECHO malnutrition ­ 2007-2010 Red Cross women and children North, Sahel and (2009) Southwest Regions Reduction of acute malnutrition and Terres des Hommes and Malnourished women and Tougan, n/a n/a ECHO mortality among children under 5 and Helen Keller International children Séguénéga, Gayéri pregnant and lactating women ­ 2007- and Fada Districts 2010 Prevention and treatment of acute Action Contre la Faim Severely malnourished Diapaga health n/a FCFA 527m ECHO malnutrition ­ 2008-2010 children aged 6-59 months District, Tapoa (2009) Province Nutritional emergency for children ­ Save the Children UK Severely malnourished Kaya District, n/a FCFA 412m ECHO 2008-2010 children aged 6-59 months Center-North (2009) region 38 School feeding (and take-home rations) Endogenous school feeding in primary MEBA Pupils in public primary 22 provinces 711,000 pupils in CFAF 4.5bn State, CAST, schools ­ since 1988 schools 3,565 schools (2008- (2008-2009) Communities 2009) Endogenous school feeding in secondary MESSRS Students in public National 20,700 students in 310 CFAF 750m State, schools secondary schools schools (2008-2009) (2008-2009) Communities Assisted school feeding in primary CRS with partners Pupils in public, private, 19 provinces 245,900 pupils CFAF 2.3bn USAID, schools ­ since 1962 and community primary in 1,328 schools (2008- (2009) Communities schools 2009) Assisted school feeding in bisongo CRS with partners Children attending bisongo n/a 4,029 children in 41 CFAF 48m USAID, bisongo (2009) (2009) Communities Assisted school feeding in BRIGHT CRS with Tin Tua and Girls attending BRIGHT 10 provinces 3,000 girls in 13 CFAF 13m USAID, schools FAWE schools schools (2009) Communities Assisted school feeding (CP10399.0 World Food Program with Pupils in primary schools 4 Sahelian 86,000 pupils in 640 CFAF 864m Multilateral, component 1) ­ since 2002 partners Provinces schools (2009-2010) (2009) State, Communities FOOD & FUEL SUBSIDIES Universal, indirect price support for food National Security Stock SONAGES n/a n/a n/a n/a State, Donors Tax and duty exemptions MEF/DGD n/a National n/a CFAF 4.3bn State (2008) Universal, indirect price support for fuel Tax exemptions on fuel products Customs with SONABEL, n/a National n/a CFAF 21.7bn State SONABHY and private (2009) retailers PUBLIC WORKS PrEst (Program de désenclavement de Helvetas with the MIS Youth in priority 9 communes in 865 persons (2005- CFAF 478m Swiss l'Est) ­ 2002-2011 Gnagna, Gourma, 2007) (2002-2011) Cooperation and Tapoa Provinces Food for Assets (CP10399.0 component World Food Program with Population of targeted 18 provinces 30,800 women CFAF 294m Multilateral, 3) ­ 2006-2010 partners villages 31,400 men (2009) State (2009) FEE WAIVERS FOR HEALTH Additional subsidy of emergency MoH Indigent pregnant women National n/a CFAF 500m State, World obstetric and neonatal care for indigent (allocated, not Bank (CASRP) women (SONU) ­ since 2006 spent) 39 Fee waivers for the indigent ­ since 2007 MoH with CBOs Indigent PLHIV National n/a CFAF 37m State (incl. IGA) Treatment of ART patients National Solidarity Fund Indigent ART patients National 200 persons CFAF 40 m State Free treatment for pregnant women and HELP Children under 5 and Seba and Dori 100,000 persons CFAF 500m ECHO children under 5 ­ since Sept 2008 pregnant and nursing health districts in women the Sahel Free treatment for pregnant women and Terre des Hommes Children under 5 Tougan, n/a CFAF 640m ECHO children under 5 Séguénéga Free treatment for PLHIV ­ 2001-2009 Médecins Sans Frontière- People living with HIV Ouagadougou n/a n/a n/a Luxembourg District Free malaria treatment ­ 2008-2009 Médecins Sans People suffering from Titao District 40,000 cases CFAF 49m MSF, ECHO Frontières-France with malaria (2008-2009) CSPS Note: This list does not mean to be exhaustive, though the major social safety net programs currently in place in Burkina Faso have been included. There may be many other small NGOs' or local authorities' initiatives throughout the country that provide social transfers to poor/vulnerable individuals or households. Most of these initiatives were considered to be too small and isolated. Yet, some small projects were included because they appear to be innovative and noteworthy initiatives. Source: Authors. . 40 Social Safety Net Spending Levels 81. Excluding the cost of general fuel subsidies and temporary food subsidies (2008), total spending on SSN programs over 2005-2009 averaged 0.6 percent of GDP. This compares to average spending on SSN in Mali in the period 2006-2009.18 A recent global study concluded that safety net programs in developing countries typically represent about 1-2 percent or less of GDP [Grosh et al. 2008]. Burkina thus appears to be on the low side of the spectrum in terms of spending on social safety nets. At around 0.6 percent of GDP, spending on SSNs thus corresponds to less than one-fourth the illustrative minimum amount needed to bring all the poor to the poverty line through cash transfers in 2007 ­ equivalent to less than 50 percent the minimum amount required to bring the poorest and the most vulnerable to the poverty line. Similarly as shown in Table 10, spending per poor person (excluding the universal food and fuel subsidy) was at about 3 percent of the poverty line in 2007. Still excluding subsidies, the amount spent on SSN per capita amounted to about 3 percent of per capita GDP. Table 10: Total Spending on Social Safety Net Programs, 2005-2009 Type of program 2005 2006 2007 2008 2009 2010 (*) 1. Cash and near-cash transfers - - - 284 9,021 4,359 2. Food transfers 7,378 11,195 13,760 22,130 25,047 34,714 2.1 Targeted subsidized food sales - 248 239 1,644 2,144 2,144 2.2 Targeted food distributions 273 1,400 558 407 497 563 2.3 Nutrition - 1,940 5,873 12,001 13,771 12,692 2.4 School feeding 7,106 7,608 7,090 8,079 8,635 19,315 3. Universal subsidies 16,726 21,999 23,757 26,340 21,730 24,500 3.1 Universal food subsidies - - - 4,263 - - 3.2 Universal fuel subsidies 16,726 21,999 23,757 22,077 21,730 24,500 4. Public works 945 865 651 1,075 772 913 5. Fee waivers - - 311 1,162 1,249 1,118 5.1 Fee waivers for health - 311 1,162 1,249 1,118 TOTAL (million CFAF) 25,050 34,059 38,478 50,991 57,818 65,603 TOTAL (% of GDP) 0.88% 1.13% 1.19% 1.40% 1.52% 1.62% TOTAL excluding universal fuel and food subsidies (million CFAF) 8,324 12 060 14,721 24,651 36,088 41,103 TOTAL excluding universal fuel and food subsidies (% of GDP) 0.3% 0.4% 0.5% 0.7% 0.9% 1.0% Per capita spending in nominal terms (CFAF), incl. universal subsidies 487 1,944 2,366 3,188 3,591 Per capita spending in real terms (2005=100) (CFAF) 487 1,886 2,226 2,909 3,178 Per poor spending in nominal terms (CFAF) 1,205 6,145 7,084 Per poor spending in real terms (CFAF): - including general food/fuel subsidies 1,205 5,781 6,464 - excluding general food/fuel subsidies 1,205 2,054 2,708 (*) Figures for the year 2010 are indicative only. Source: Authors. 18 Mali Social Safety Nets, World Bank, 2010. 41 82. Spending on SSN has increased steadily (Table 10 and Figure 8). Spending on SSN increased from 0.3 percent of GDP in 2005 to about 0.9 percent of GDP in 2009 excluding subsidies. With the tax and duty exemptions on fuel imports and the temporary 2008 exemptions on food imports, spending on social safety nets reached 1.4 percent and 1.5 percent of GDP in 2008 and 2009 respectively. Tax exemptions on various food imports implemented in response to escalating food prices cost reached about 0.1 percent of GDP in 2008. Thus the cost of the general food subsidy absorbed less than 8 percent of total SSN spending in 2008. However, for 2008, the cost of these general food subsidies accounted for about 13 percent of all government-financed social safety nets. 83. Overall increases in SSN levels in real terms show the rapid response of both the government and the donors to the 2008 food and fuel crisis. Measured in 2005 CFAF, total spending on SSN increased by 10 percent in 2007, 29 percent in 2008, and 10 percent in 2009. However, when discounting the universal subsidies, the growth is far more spectacular: 18 percent in 2007, 63 percent in 2008, and 42 percent in 2009. 84. The increase in SSNs spending is largely explained by significant efforts made in food transfer programs and specifically, subsidized food sales, nutrition programs, as well as targeted food distribution. A cash transfer and a public works program, both with significant disbursements in 2008, compounded the growth of SSNs for that year. In contrast, the subsidy to consumers on petroleum products (VAT exemption) increased only marginally in 2008. These subsidies, nevertheless, remain the most important component of government-financed SSNs, or 65 percent (see below for a detailed discussion). Figure 8: Evolution of Spending on Social Safety Nets Over Time, 2005-2009 70,000 60,000 3. Universal subsidies 50,000 5.1 Fee waivers for health 4. Public works 40,000 2.4 School feeding 30,000 2.3 Nutrition 20,000 2.2 Targeted food distributions 10,000 2.1 Targeted subsidized food sales - 1. Cash and near-cash transfers 2005 2006 2007 2008 2009 2010 NB: Figures for the year 2010 are indicative only. Source: Authors. 85. Spending on SSN is dwarfed by spending on health and education combined as well as spending on other social sectors (all activities defined as contributing to poverty reduction19) (Table 11). Given the importance of all social expenditure in the budget 19 Other than health and basic education, these include social programs by Ministry of Women's Welfare, Ministry of Youth, Labor, Employment, Infrastructure (rural roads), Ministry of Agriculture, Ministry of Animal Resources, Ministry of Environment, Ministry of Information, Ministry of Justice, Ministry of Mining and Energy, Ministry of Economy and Finance, and some common expenditures. 42 (including health and education), the amounts spent on SSNs appear insignificant. Spending on health and education combined in 2008 was about 14 times the amounts spent on SSNs, excluding food subsidies. However, domestically financed expenditures in health, education, and SSNs, while showing significant progression in nominal terms and as percentage of GDP between 2005 and 2009 as shown in Table 11, have not progressed equally when compared to total domestically financed expenditures (including HIPC financing). Expenditures on basic education have remained on average a little above 14 percent of total domestically financed expenditure. Health expenditures have hovered below 10 percent, and domestic spending on SSN below 1 percent of total expenditure. This suggests that the growth in these sectors' expenditures is commensurate to the growth of overall domestic resources. 86. Specifically, the above implies that the importance of the social sectors (education, health, and SSN) has remained largely unchanged in the government priorities agenda. External support to health and education has shown variations (Figure 9). High dependency on external financing characterizes all three sectors, but SSN financing seems to be increasingly donor-dependent. Table 11: Social Expenditure (including HIPC20 and Excluding External Financing) 2005 2006 2007 2008 2009 (*) 2010 (**) Basic ducation (CFA billions) 59.6 74.7 74.9 81.9 98.5 109.2 as percentage of GDP 2.1% 2.5% 2.3% 2.2% 2.6% 2.7% as pecentage of domestically financed expenditure and HIPC 14.5% 15.9% 13.7% 13.4% 14.0% 15.3% Health (CFA billions) 37.1 47.9 56.9 57.7 69.8 83.8 as percentage of GDP 1.3% 1.6% 1.8% 1.6% 1.8% 2.1% as pecentage of domestically financed expenditure and HIPC 9.0% 10.2% 10.4% 9.5% 9.9% 11.8% Other social expenditure (CFA billions) 65.1 44.6 50.8 58.9 91.2 81.5 o/w Rural roads 3.7 3.9 2.7 3.8 3.4 4.1 o/w Ministry Action Soc, Solidarite Nat 5.1 5.1 5.4 5.4 n/a n/a as percentage of GDP 2.3% 1.5% 1.6% 1.6% 2.4% 2.0% as pecentage of domestically financed expenditure and HIPC 15.9% 9.5% 9.3% 9.7% 13.0% 11.4% domestically financed SSN (excl. universal subsidies) (CFA billions) 2.8 3.7 4.1 5.9 7.2 4.5 As percentage PIB 0.1% 0.1% 0.1% 0.2% 0.2% 0.1% as pecentage of domestically financed expenditure and HIPC 0.7% 0.8% 0.8% 1.0% 1.0% 0.6% (*) projected, (**) budgeted Source: Public Expenditure Review, World Bank, 2009; IMF; Authors. 20 HIPC or PPTE (Pays Pauvres Très Endettés). The amounts spent under this title correspond to government expenditures equivalent to the debt relief under HIPC initiative. 43 Figure 9: Share of Basic Education, Health and SSN in Total Domestically And Externally Financed Expenditures 30.0% 25.0% 20.0% SSN (domestic and external, excl. universal subsidies) 15.0% Health 10.0% Basic eduction 5.0% 0.0% 2005 2006 2007 2008 Source: Authors. Safety Net Funding Patterns 87. Financing for SSN programs is heavily and increasingly donor-dependent. Excluding universal subsidies, the external financing share in total SSN financing increased from about 57 percent in 2005 to 69 percent in 2008, and 76 percent in 2009. Including fuel and food subsidies, external spending in total SSNs increased from less than 20 percent of the total in 2006 to more than 48 percent in 2009. Excluding fuel and food subsidies, government spending increased slowly in 2005 and 2009 as a share of GDP, from 0.1 to 0.2 percent of GDP. 88. Donors are showing a clear commitment toward increased financing of SSN. The increase in donor financing has been large (as shown in the above numbers) from about 0.2 percent of GDP in 2005 to 0.7 percent in 2009. In real terms (2005 CFAF), donor financing has thus increased by almost five times. 89. Excluding food and fuel subsidies, donors and the government give priority to food-based programs. Based on available data, targeted food distribution, and sales, nutrition and school feeding programs (all food-based programs) represented 90 percent of total SSN spending in 2005 and still hovered around 70 percent in 2009. Within these programs, donors are massively invested in school feeding and nutrition while the government spends on school feeding but hardly on nutrition. If data on food sales and distribution were to be completed, this picture would probably be somewhat altered. It is interesting to note that cash and near-cash transfers appeared in 2008 in Burkina while spending on public works, exclusively financed by donors, has decreased as a share of the total from about 10 percent in 2005 to about 2 percent in 2009. 90. Spending patterns reveal different priorities between the government and donors. For available years (2005-2009), the government's priorities have been overwhelmingly toward universal subsidies. Expenditures for fuel subsidies increased in 2008 following the hike in fuel prices. In 2005, the share of the subsidy in total 44 government-financed SSNs amounted to more than 80 percent. In 2008, this share still stood at 77 percent of the total (Table 12). Table 12: Donors and Government's Priorities by Program for 2008 Government (%) Donors (%) 1. Cash and near-cash transfers n/a 2 2. Food transfers 19 92 2.1 Targeted subsidized food sales 5 n/a 2.2 Targeted food distributions 0 2 2.3 Nutrition 1 67 2.4 School feeding 13 23 3. Universal subsidies 77 0 3.1 Universal food subsidies 13 0 3.2 Universal fuel subsidies 65 0 4. Public works n/a 6 5. Fee waivers (health) 3 40 Total 100 100 Source: Authors. Box 4: Budget Management Issues in Burkina Faso Irrespective of the levels of SSNs funding, issues with disbursement of funds, tracking, monitoring, evaluation, and accountability have been raised in other reports. The disbursement of government funds is hampered by a slow, protracted process marred by multiple ex-ante controls, partly responsible for uneven, yet improving, execution rates. As shown in the table below for the Ministry of Health, the Ministry of Basic Education, and the Ministry of Social Action and National Solidarity. Disbursement issues also seem to prevail in the case of the supplementary 2009 budget comprising CFAF 25.7 billion in social measures, including CFAF 8.4 billion which, as of April 2010, has not yet been disbursed. Particularly noteworthy are slow disbursement rates on externally financed investments, averaging a little over 60 percent between 2004 and 2008. This is of particular concern in sectors such as SSNs where external financing is largely dominant. Budget Execution Rates 2004-2008 2004 2005 2006 2007 2008 (%) (%) (%) (%) (%) Total budget (incl. external financing and HIPC) 95 93 96 98 97 Total externally financed investment budget 66 60 70 74 43 Ministère de l'Enseignement Base et l'Alphabétisation (incl. external financing and HIPC) 83 78 120 112 93 Ministère de la Santé (incl. external financing and HIPC) 99 100 104 102 96 Ministère des Affaires Sociales et de la Solidarité Nationale (domestic financing only) n/a 98 100 98 95 The budget nomenclature does not allow for tracking SSN-related expenditure, and no mechanisms are currently in place to track social protection and social safety net expenditures. The budget nomenclature is based on the UMOA directive of 1998. While a new functional classification was introduced in 2007 (based on the International Financial Statistical Manual), it does not allow for singling out SSN programs. These are the shared responsibility of different ministries, including small ministries like MASSN which do not have the capacity or the resources to monitor much less evaluate social safety net programs. Additionally, as mentioned in other reports, fiduciary concerns remain important in Burkina, owing to the weakness of control mechanisms. As stated in the Public Expenditure Review, efficiency and effectiveness of internal controls by the Inspection Générale des Finances (IGF) and the Authorité 45 Supérieure du Contrôle d'Etat are hampered by lack of means and, in the case of the IGF, insufficient independence. The Supreme Audit Institution (Cour des Comptes), created in 2000, has not yet been able to operate to its full potential owing to limited resources and dependency on the Ministry of Finance for staffing. The 2007 Public Expenditure and Financial Accountability Assessment (PEFA) rating of internal controls of public resources in Burkina shows that challenges remain to create a low risk fiduciary environment in Burkina. Internal Control over Public Expenditure Control over payment of salaries (rating C+): While employment files and payment data are well synchronized and changes well documented, delays in adjustments to employment or pay lead to significant retroactive adjustments and verifications of effective payments are partial. Control of non-salary expenditure: (Rating C+) Commitment control mechanisms are in place but are not comprehensive. While rules are usually well understood and applied, simplification and emergency procedures arise without justification and raise concerns of irregularity. Transparency of public procurement (rating C+): While effective complaints mechanisms are in place, little information is available to evaluate procurement processes and justifications for less competitive methods are scant. Efficiency of Internal controls (rating: D+): Control coverage is partial and does not meet international norms. In addition, reports are few and strictly internal. Recommendations from internal controls are usually ignored. Note: Public Expenditure and Financial Accountability Assessment (2007). (Ratings are from A to D, A being the best performance, and D the worst.) Source: GoBF: PER, World Bank, 2009; and Public Expenditure and Financial Accountability Assessment (GoBF 2007). 91. The coverage of existing social safety net programs is limited compared to the needs. The theoretical estimate of the total number of beneficiaries of social safety net programs, assuming no overlap between programs and excluding general subsidies, exceeded 3.9 million individuals in 2009, that is, about 25 percent of the total population of Burkina Faso.21 This figure is however misleading. Most of the beneficiaries are from targeted subsidized food sales (Figure 10). And the coverage of these subsidized sales is only temporary and the level of benefits are very small (Table 13). Thus, there are many issues that have to be taken into account to estimate the effective coverage of the existing safety nets programs. 21 Based on a total population of 15.8 million (mid-2009). 46 Figure 10: Indicative Annual Number of Beneficiaries of the Existing Social Safety Net Programs, 2009 5.1 Fee waivers for 4. Public works health 1. Cash and near- cash transfers 2.4 School feeding 2.1 Targeted subsidized food sales 2.3 Nutrition 2.2 Targeted food distributions Table 13: Beneficiaries and Budgetary Allocation by Program in Burkina Faso, 2009 CFAF per % of Number of Annual Budget Program Beneficiary Poverty Beneficiaries Million FCFA per Year Line 1. Cash and near-cash transfers 359,062 9,021 25,128 27% 2.1 Targeted subsidized food sales 1,800,000 2,144 1,191 1% 2.2 Targeted food distributions 95,878 497 5,184 6% 2.3 Nutrition 377,362 13,771 36,493 39% 2.4 School feeding 1,070,649 8,635 8,065 9% 4. Public works 63,065 772 12,241 13% 5.1 Fee waivers for health 140,200 1,249 8,909 9% Total (potential coverage) 3,906,216 36,089 9,239 10% % of population covered 25% Total (potential coverage) - 2,106,216 33,945 16,116 17% Excluding subsidized sales % of population covered 13% % of the population below the 43% poverty line Note: The level of transfer does not include administrative cost that might be large for nutrition programs. The poverty line considered in 2009 was FCFA 93,949 persons per year. Source: Authors. Options for increasing SSN budgets 92. As other reviews have pointed out, supporting the current income of the poor is a sensible policy choice but the existing social safety net system appears too limited to constitute an appropriate response to poverty and vulnerability. Viewed against the needs, the level of SSNs coverage of the poor needs to pick up significantly to make a difference in the lives of the poor. In addition, large donor dependency on SSN programs suggests that the state needs to play a greater role in the organization, consolidation, and 47 perpetuation of the different components of the social safety net system with the view of creating a sustainable SSN system. 93. Within the constraint of its budget profile, the government needs to allocate limited resources among programs that support the income of the poor, promote education for girls, or expand the still-limited stock of public physical capital, such as roads. In a larger debate, these considerations have to be borne in mind, particularly in the context of the current economic crisis where spending on human capital may be threatened, due to pressures on financial resources. 94. The costs of leaving the level and SSNs unchanged (at about 1 percent of GDP in 2009) may be high in the long term even if this option may seem attractive in the short term.22 In that case, the focus of the fiscal policy would be to preserve expenditures in the social sectors at their current levels as a percentage of GDP. Arbitrage should focus on promoting or scaling up the most cost-effective programs (Table 14). 22 Indeed, as a World Bank report recently noted [Background paper prepared by World Bank Group staff for the G-20 Leaders' Meeting, Pittsburgh, USA, September 24-25, 2009]: "Even mild downturns can have costly and long-lasting effects on human welfare, as families with few alternative employment opportunities and little or no access to credit are forced to reduce food intake, even for very young children, or pull children out of school. Evidences from past crises shows that children who experience short-term nutritional deprivation can suffer long-term harm. Such possible adverse outcomes highlight the importance of protecting core spending, including on health and education, in the face of sharply declining revenues." 48 Table 14: Options for Scaling-Up SSN Programs Type of SSN program Target group All bottom Urban bottom Rural bottom 10% 20% 20% 1. Direct transfers Beneficiaries ­ Households 200,000 28,770 357,135 Beneficiaries ­ Population 1,400,000 201,390 2,499,943 Transfer/household/month (CFAF) 10,967 10,967 10,967 Transfer/person/year (CFAF) 18,800 18,800 18,800 Total cost of transfers (billion CFAF per year) 26.30 3.80 47.00 2. Public works Beneficiaries ­ Workers 200,000 85,705 Beneficiaries ­ People (7 persons per household) 1,400,000 599,932 Wage rate / day (CFAF) 1,000 1,000 Days (5 months at 20 days/month) 100 50 Transfer/worker/year (CFAF) 100,000 50,000 Transfer/person/year (CFAF) 14,286 7,143 Total cost of transfers (billion CFAF per year) 20.00 4.30 Taking into account 20% inputs 24.00 5.10 People per project 100 100 Number of projects 2,000 857 3. Conditional cash transfers: scholarships Scholarship/year on a 10-month basis (CFAF) 15,000 Number of children 5-9 not in school 690,780 20% poorest children 5-9 not in school 138,156 Total cost of transfers (billion CFAF per year) 2.07 As percentage of GDP (2009) 0.05% As percentage of government expenditure (2009) 0.2% 4. Nutrition targeting poor children 0-5 years Program average cost per capita/year (CFA) 7,000 Number beneficiaries (poor children under 5) 1,339,650 Total cost of transfers (billion CFAF per year) 9.40 As percentage of GDP (2009) 0.2% As percentage of government expenditure (2009) 0.9% Notes: (i) Annual cash transfer/person of CFAF 18,880 (20% of 2009 poverty line); (ii) Wage rate based on 2/3 of minimum wage (about CFAF 20,000 a month) and 20 days of work a month. Source: World Bank, UNICEF, Authors. 95. The first alternative, the "incremental option" would consist of gradually expanding some effective and complementary SSN programs (nutrition programs and public works) as well as introducing permanent cash transfer programs. An indicative increase in SSN spending for this option could be, for example, doubling the amount (as percent of GDP) spent on SSN programs in the medium term. This would bring total spending to around 2 percent of GDP. The benefits of that option is that costs of SSN financing would remain controlled and their manageability would be better secured while fiduciary risks associated with large-scale programs would be contained. The disadvantages are that this option would not alleviate poverty entirely, even if a well- designed SSN program to the tune of 2 percent of GDP could significantly improve extreme poverty. Small-scale programs should also be protected from the stop-go 49 approach (to mitigate the stop-go approach risk, a choice could be made to build and expand existing programs). Table 14 shows some hypothetical examples of programs that could be financed under this option in order to provide an idea of the number of people that could be covered by each type of programs: Costs could be reduced by selecting the more cost-effective programs or modifying the amount of resources across programs and the number of beneficiaries to be covered. It is important to remember that the amounts estimated are the net benefits, which means they exclude all of the costs associated with program management. A margin should thus be added to them to estimate their operational costs. 96. The second alternative would be a more aggressive approach to poverty alleviation through national-scale programs. The obvious benefits of this approach would be to address poverty in a significant manner. The challenges are numerous and significant, including financing and fiduciary issues. Overall, this option may not yet be realistic particularly as it raises the need for a better-targeted safety net program that reaches at least a portion of the poorest and would involve significant resource mobilization while running the risk of jeopardizing other important sectors such as health and education. 97. Financing options for increasing SSNs raises the issue of fiscal space. The various options for enlarging fiscal space have been reviewed by the Bank (Public Expenditure Review or PER) in the context of trade-offs for supporting increased resources in order to reach the MDGs (with specific focus on health, education, and agriculture). The recent PER relied on the MAMS23 model to review alternatives (Box 5), as this discussion is also pertinent to the issue of fiscal space for additional SSN financing. 98. Four basic options are usually considered for creating fiscal space: (i) reallocation of expenditure under existing fiscal constraints, (ii) increased donor financing on concessional terms domestic revenue; (iii) borrowing; and (iv) increasing domestic revenue (either through taxation or improved revenue collection). The advantages and disadvantages of each is summarized in the table below and briefly discussed in the context of Burkina. 23 Maquette for Millennium Development Goals Simulation. 50 Box 5: Lessons of the Analysis of Fiscal Space in Burkina Faso, Using the MAMS Model With regard to which fiscal space options are best for reaching the MDGs, none of the scenarios dominates the others, which makes it necessary to consider trade-offs. For example, poverty is likely to be most reduced by 2015 through an aid-financed increase in infrastructure spending. However, focusing the aid resources on human development spending would yield a better outcome for the education and health MDG indicators, thus a trade-off between MDGs. There are also trade-offs regarding the sources of fiscal space. For example, mobilizing additional aid will allow an increase in domestic absorption and avoids the political resistance that may emerge as domestic sources of fiscal space are generated--taxpayer lobbying against higher taxes or changes in the tax structure--but it is potentially less reliable (and permanent) than domestic sources of fiscal space, and it does make the government more dependent on donors. Ultimately, these trade-offs need to be resolved at the country level, taking into account country-specific preferences. The model-based analysis of fiscal space, though, can help to identify and quantify likely trade-offs. Increasing education and health services in real terms takes time. Burkina Faso has in recent years already made substantial progress in these areas, and the baseline projects very large gains in education and health MDG indicators in coming years. But expanding these services substantially over baseline levels can prove challenging because doing so requires capacity, which takes time to build up. Consequently, expanding expenditure in these areas requires careful preparation to align the pace of expenditure increases with the ability of education and training programs to deliver suitably educated workers. There may also be a need to monitor wage pressures to avoid large increases in the wage bill that could crowd out other expenditures. Infrastructure spending promotes not only growth but also other MDG objectives. In MAMS infrastructure has a direct positive impact on education and health MDG indicators because it facilitates the delivery of these services. There is also an indirect effect through higher growth--higher per capita income increases demand for these services. In the simulations, infrastructure spending yields very substantial improvements in the education and health MDG indicators through these two channels. However, they take some years to effectively improve education and health MDG indicators. Source: Gottschalk et al., 2009. 51 Table 15: Options for Increasing Safety Net Budgets: Advantages and Disadvantages Financing methods Advantages Challenges Reallocation of Finances programs within No additional funds relative to the budget expenditure budget constraints Depending on the amounts to be freed Increases overall productivity up, this options could require significant of government outlays commitment from the government to Efficiency gains by cutting implement trade-offs into unproductive expenditure Can be difficult to implement if large- Feasible in the short term on a scale reallocations are necessary ­ that small-scale basis particularly option would require a detailed analysis if "low hanging fruits" can be of public expenditure programs and identified for cuts medium-term commitment by government Donor financing Concessional financing or Budget support required for more flexible grants use of funds Increases the overall envelope Cyclicality of funding, and medium-term downward trend Donor coordination issues Needs significant increase in fresh budget support Needs a significant and durable donor commitment toward harmonization and continuity, for gradual or larger-scale programs Borrowing Finances temporary expansion Burkina at high risk of debt distress of programs during crisis Agreement IMF on non-concessional Less reliance on donors borrowing Overhang impact on growth More expensive than concessional financing Increased domestic May be more sustainable than If done through increased taxation, revenue other options economic costs may have direct and Need to be trough improved indirect effects on the economy, which tax management could harm economic growth and Provide secured financing in ultimately the poor; needs tax incidence the short-term analysis May have redistributive effect Tax increase may have limited return given Burkina's narrow fiscal base and low revenue to GDP ratio, be politically unpopular, and may hurt the poor Source: Authors. Option 1: Expenditure Reallocation 99. The first option, expenditure reallocation, under existing fiscal constraints requires an in-depth analysis of expenditure allocations and effectiveness, which is beyond the scope of this report. However, given the huge needs in sectors like education and health, and other sectors important to poverty alleviation and growth, the premise is that allocations to these sectors should at minimum be preserved while efficiency and equity gains should be sought within them (see Public Expenditure Review, World Bank, 2009). 52 100. Nonetheless, within priority sectors such as education, public expenditures should become more efficient and pro-poor. Specifically, the Public Expenditure Review (World Bank 2009) showed that, while public expenditures in education have increased continuously (from 10 percent of total expenditure in 2000 to 16.6 percent in 2008) and the gross enrolment rate soared accordingly from 40 percent in 2000 to 65 percent in 2007, gender inequalities remain substantial in access to education. In addition, access to education is limited for children in rural areas and poor families. These inequalities reinforce each other so that among the poorest families in rural areas, the gross enrolment rate is only 40 percent. The analysis of the poverty profile suggests that spending in education needs to become pro-poor. Restructuring spending toward improving the supply of education is not enough. Demand for education also needs to be supported. In this context, to boost demand for education, SSNs raising the income of parents, (e.g., through transfers in cash or in kind) and other programs supporting school enrolment could be important, particularly for girls. 101. Similarly, in the health sector, despite a continuous increase in health expenditures (from 7.7 percent of the budget to 9.2 percent in 2008), and the improvement in some health indicators (e.g., assisted deliveries and vaccination coverage), inequalities prevail between regions in the supply of health services. Access to health services has even declined in 7 out of 13 regions. Public health expenditures could be directed toward explicitly pro-poor programs, such as nutrition and fee waiver programs covered under SSNs. For example, based on the costs of nutrition programs discussed in the previous section,24 an increase in nutrition programs to benefit poor children under 5 (estimated at about 1.3 million) would be of the order of CFAF 9.2bn (assuming an annual cost of CFA 7,000 per child), or about 0.9 percent of total government expenditure in 2009. 102. To finance these needs as well as a progressive expansion of SSN programs, the government could turn toward less productive expenditure, including other "social" sectors, and identify programs that have low or little effectiveness in alleviating poverty. Such is the case of the subsidy to fuel imports and social housing programs, which seem to be poorly targeted to the neediest. In functional terms, this could imply scrutinizing such categories as unallocated expenditure and General Services and Public Administration, which covered 27 percent of the budget in 2007-2009, equivalent to the health and education expenditures combined. Based on the PER, because of high levels of Operations and Maintenance expenditures under these categories, there is risk of losing control over costs such as vehicles, gasoline, travel, etc. Based on the 2009 budget numbers, a 10 percent efficiency gain on these expenditures could yield about 0.6 percent of GDP, roughly equivalent to the estimated annual cost of a cash transfer of CFAF 18,800 to the poorest decile of the population (excluding administration and targeting costs). 103. More generally, very rough estimates of efficiency gains in public expenditure, such as, for example, a 10 percent efficiency gain in discretionary expenditure (goods and services, current transfers, and domestically financed investments) could yield up to 1.5 percent of GDP, which would go a long way toward establishing SSN programs as described above and would place the government in a leadership position in promoting and financing a coherent SSN system. 24 This is based on assuming an average cost of CFAF 12,000 per year per beneficiary for nutrition programs of children less than 5 years old and mothers. 53 Option 2: Increase External Support on Concessional Terms 104. Increased donor funding may be an option in the short term as Burkina is likely to remain heavily donor-dependent in the medium term and donors of late have taken greater interest in SSN financing. Evidence of donor interest in SSNs has also been highlighted in the composition and evolution of donor-supported programs in Burkina since 2005, as discussed in the previous section. 105. Issues with external financing include the fact that foreign aid is still dominated by projects as opposed to budget support. Based on IMF estimates and projections, the ratio of projects to programs was about 1 in 2008. However, this ratio is projected to increase to 2.5 in the coming years.25 This trend points to aid fragmentation and makes it more difficult to envisage sector reforms in any sector largely dominated by foreign aid. Another concern with external financing has been the low levels of disbursement on investments as discussed above (see Box 4). On average between 2004 and 2008, the disbursement of external funds corresponded to about 50 percent of projected amounts. 106. While donor financing may increase in the future as the recession eases in industrialized countries, its present volatility is projected to decrease in the medium term. Donor financing is highly susceptible to business cycles. As a result, unpredictability of external financing may increase in the coming years. This volatility may further aggravate a stop-and-go approach issuing from unpredictable funding. Over the medium term, projections point to a decrease in foreign aid as a percentage of GDP: based on IMF projections,26 external financing would decrease from 11.4 percent of GDP to 10.8 percent in 2013. Furthermore, and more importantly, a further increase in the financing imbalance between donors and the government in SSN funding may well be an unsustainable proposition in the medium term, not only financially but also from the point of view of coherence and manageability of the overall SSN system. Option 3: Borrowing 107. Borrowing on non-concessional terms is a very risky proposition in Burkina at this point. The 2008 joint IMF/World Bank Debt Sustainability Analysis (DSA) classified Burkina as being at high risk of debt distress. Implementing a prudent fiscal policy, limiting external borrowing to concessional loans, improving public debt management, and setting a sustainable deficit target would be required to stabilize the debt ratios in the long term [IMF 2009a]. Option 4: Increased Domestic Revenue 108. The government voted measures in 2009 to reform taxation that are expected to raise revenue, albeit on a declining basis as a percentage of GDP. Such measures cover the introduction of the corporate tax, the rationalization of the investment code, and the strengthening of the VAT system. Based on IMF projections, these measures are expected to raise domestic revenue by 0.9 percent of GDP in 2011, 0.6 percent of GDP in 25 IMF: Burkina Faso: Staff Report for 2009 Article IV Consultation, Fifth Review Under the Poverty Reduction and Growth Facility, and Requests for Augmentation of Access and Modification of Performance Criteria, January 2010. 54 2012, and 0.4 percent of GDP in 2013. Allocating a portion of this increase to SSNs could be envisaged, with the understanding that additional measures would be needed to secure sustainability of the SSN programs. Additionally, many countries, developed and developing, exempt the poorest households from paying taxes. Tax exemption for the poorest could actually be a useful measure to support the consumption of poor households. But of course, this again raises the issue of targeting capacity. 55 CHAPTER IV: REVIEW OF INDIVIDUAL EXISTING SOCIAL SAFETY NET PROGRAMS Most existing social safety net programs in Burkina Faso provide temporary assistance and are mainly introduced during periods of shock, which make them less appropriate to tackle chronic poverty. The existing interventions often are based on a project approach, focused on one geographic area, and there is no systematic approach to assist poor and vulnerable households. The main challenges are: difficulties in defining the target groups (poor and vulnerable); selection of adequate instruments; and establishment of solid monitoring and evaluation systems to inform policy decisions. Given the recent innovative initiatives, such as food vouchers, cash transfers, and health fee waivers, important lessons can be learned to improve and potentially scale up the existing social safety net programs. Food transfers are the main form of social safety net programs in Burkina Faso, accounting for 69 percent of total SSN spending and over 80 percent of all estimated SSN beneficiaries in 2009 (excluding fuel subsidies). Among these programs, school feeding, supported almost equally by both the government and development partners, represents one of the main social safety net programs, accounting for 24 percent of total spending on SSN and covering about 38 percent of the estimated total number of SSN beneficiaries. Nutrition programs are implemented with donor support, but they are still insufficient to address the needs of acute childhood malnutrition: More proactive national policies are needed. Universal fuel subsidies, on the other hand, are very expensive (0.7 percent of GDP in 2007) and have a very limited impact on the poorest decile (84 percent of the benefits go to the non- poor). The public works programs, supported by external financing, provide interesting experiences that could help the development of effective social safety net instruments. Finally, implementation of national health subsidy/fee waiver program is impeded by the lack of clear targeting criteria and of explicit implementation mechanisms. 109. This chapter describes how each individual program within the existing system could be made more efficient and potentially scaled up. The programs discussed are: (i) cash and near-cash transfers; (ii) food transfers (subsidized food sales, targeted food distributions, nutrition programs, and school feeding); (iii) universal subsidies (food and fuel); (iv) public works; and (v) fee waivers. The design and performance of each program type is discussed, considering international good practices and a set of key performance criteria, such as appropriateness, adequacy (coverage, benefit level, and duration), equity, cost-effectiveness (efficiency and effectiveness, adequate funds for administrative costs), efficiency, and sustainability (fiscal, political, and administrative) [Grosh et al. 2008]. A. Cash and Near-Cash Transfers 110. International evidence shows that social cash transfers can have a positive impact on education, health, nutrition, food security, and overall poverty reduction [Grosh et al. 2008]. Cash transfers have the potential to enable the poor to better manage social risks and to generate a range of positive impacts. First, by providing cash income, they directly reduce both income and expenditure poverty over the short term. Second, given typical consumption patterns, poor households allocate a significant proportion of the expenditure to food, improving nutritional outcomes. Depending on intra-household allocation decisions, much of this spending is likely to benefit children. In most cases, social transfers will also support children's human capital accumulation, particularly in 56 terms of increased school attendance and educational outcomes as well as increased access to primary health care. In addition, an evolving evidence base demonstrates that social transfers contribute to pro-poor and inclusive economic growth. The range of outcomes contributes to long-term poverty reduction, particularly by breaking the inter-generational transmission of deprivation. Annex 2 presents good practices in the design of social cash transfer programs, and Annex 3 presents a couple of good examples of cash transfer programs implemented in Africa. 111. Learning from international experience, a few cash transfer programs were recently introduced in Burkina Faso: CNLS-IST's pilot cash transfer program, and CRS and WFP's food voucher programs. These projects rely exclusively on external funding and are implemented with the support from international partners. Conditional and Unconditional Cash Transfers 112. The CNLS-IST research action is expected to bring valuable learning on the role of conditionalities and gender on the impact of cash transfers. In 2008, the CNLS-IST (Conseil National de Lutte contre le Sida et les Infections Sexuellement Transmissibles) and the World Bank launched a research-action in the Nahouri Province aimed at testing the role of conditionalities and gender on the impact of cash transfers targeted at orphans27 and other vulnerable children28 (OVC) on health, education, and well- being. In total the program covers 2,600 households (approximately 18,000 people). Villages of the intervention area were randomly split into five groups, and 650 households were randomly selected among eligible households in each of these groups to participate in the research-action (Table 16). This project is one of the few initiatives in Africa aimed at testing the comparative advantage of conditionalities ­ similar research is taking place in Zambia and Kenya. About 40 percent of the total budget was devoted to impact evaluation. Initial results are expected in the course of 2010. The final results are expected in 2011 and will be extremely useful to inform the design of appropriate mechanisms to assist OVC, as well as poor households in general. 27 The CNLS-IST defines an orphan as a child under 15 who lost at least one of his/her biological parents. 28 The CNLS-IST defines a vulnerable child as a child living in a household whose revenue per capita is below the poverty line, or in a household in which at least one member is HIV positive. 57 Table 16: Beneficiary Groups in CNLS-IST Research-Action Group 1 Group 2 Group 3 Group 4 Group 5 Size 15 villages 15 villages 15 villages 15 villages 15 villages 650 households 650 households 650 households 650 households 650 households Entitlement to Yes Yes Yes Yes No (control) cash transfers Conditionality No No Yes Yes n/a Benefit level for children 1,000 FCFA/quarter 1,000 FCFA/quarter conditional n/a 0-6 years to visits to the health center for children 2,000 FCFA/quarter 2,000 FCFA/quarter conditional n/a 7-10 years to school attendance (90%) (grades CP1 to CE2) for children 4,000 FCFA/quarter 4,000 FCFA/quarter conditional n/a 11-15 years to school attendance (90%) (grades CM1 to 4th) Frequency quarterly quarterly quarterly quarterly n/a Recipient mother father mother father n/a Source: SP/CNLS-IST, 2008. Food Vouchers 113. The urban food voucher is an innovative program with the purpose of assisting the extreme poor in urban settings. Building on their experience with vouchers in seed fairs, and then food and building material fairs to assist flood-affected households, Catholic Relief Services launched in 2008 a food voucher program to respond to increased food insecurity in urban areas induced by high food prices. The World Food Program followed in 2009 with a larger program currently providing food vouchers to over 30,000 households in the towns of Ouagadougou and Bobo-Dioulasso (over 180,000 persons) (Table 17). Table 17: Main Characteristics of Recent Food Voucher Programs Program Food Vouchers to Urban Food Vouchers to HIV- Urban Voucher Program Poor Infected Persons Implementing agency CRS with OCADES CRS with OCADES WFP with MASSN, CRB network network and health district and CRS Funding source Gates Foundation Global Fund Multilateral Intervention area Urban: Ouagadougou and Health centers in six Urban: Ouagadougou and Bobo-Dioulasso regions Bobo-Dioulasso Type of beneficiaries 56% of female-headed HIV-infected pregnant Households living in households + aged + women and children under extreme poverty precarious professions 5 (mason, mechanic, watchman) Number of beneficiaries 3,800 households 2,200 persons 31,500 households Benefit level 9,000 per household n/a 1,500 per person with a (FCFA/month) ceiling of 9,000 per household Duration 6 months 8 months 18 months Sept 2008-Feb 2009 Dec 2008-Aug 2009 Jan 2009-June 2010 Source: CRS; WFP; Authors. 58 114. The mid-term evaluations of the WFP's Urban Voucher Program reported a rather efficient implementing process and positive results, yet emphasizing the serious challenge of targeting in urban settings. The census of 142,000 households conducted prior to the intervention provided the basis for beneficiary selection at a cost of FCFA 1,700 per beneficiaries. The targeting evaluation reported disappointing results and recommended exploring alternative, and possibly simpler, methods for beneficiary selection (Box 6). A joint mid-term evaluation (qualitative only) conducted in June 2009 reported that the large majority of beneficiaries were satisfied with the range of commodities offered as well as with the delivery mechanism. Shopkeepers also indicated being pleased with the program. And the WFP reported the voucher program to be 40 percent cheaper than classic food distribution. These positive results supported the extension of the program for an additional six months, until June 2010. And the WFP has been gradually increasing the number of beneficiaries by 30 percent since January 2010. The program was terminated in June 2010 due to lack of funding. 115. The mid-term evaluation team recommended maintaining the program on a permanent basis for the extreme poor who cannot work, suggesting this group also receives health vouchers. Furthermore, they called for the strengthening of the National Social Action Policy and overall social protection system.29 A tool kit was developed to capitalize the experience and support a gradual handover of the program to the national institutions, and as of April 2010, a total of 142 social workers had already been trained on social safety nets and food voucher implementation. 29 In addition, the team recommended a number of complementary activities, including: malnutrition screening and potential expansion of the WFP and UNICEF nutrition programs to Ouagadougou and Bobo- Dioulasso; school feeding activities (possibly with vouchers to expressly target indigent pupils); income- generating activities (for women especially); vocational training and microcredit programs for youth; support to small industries in Bobo; support to agricultural production; and fee waivers for health. They also suggested vouchers could be made conditional to school attendance in a second phase of the program, although recognizing that additional schools would need to be built (there are currently 100 to 200 pupils per class in public schools). They also called for an assessment of the situation in other towns of the country. 59 Box 6: WFP's Urban Voucher Program In June 2008, the mission conducted jointly by the Government of Burkina Faso, the United Nations System, and Save the Children to assess the impact of the world high food prices in the towns of Ouagadougou and Bobo-Dioulasso reported an alarming situation. Responding to an official government request, the WFP launched an emergency program providing food vouchers to 30,000 very poor households to make up for the loss of purchasing power and nutritional supplements to both poor and very poor households in targeted urban neighborhoods to prevent malnutrition. Objectives ­ The program has three specific objectives: (i) protect livelihoods by providing beneficiary households a compensation of the loss of purchasing power through a distribution of unconditional food vouchers; (ii) reduce chronic hunger and under nutrition of children 6-24 months, and pregnant and lactating women through the distribution of fortified food; and (iii) strengthen the capacity of the government to fight hunger through the establishment of social safety nets. Targeting ­ The targeting exercise took place in August 2008, using the vulnerability score developed by the Burkinabe Red Cross (CRB). Since surveying the whole population of Ouagadougou and Bobo- Dioulasso would have required time and resources, a pre-selection of the poorest neighborhoods was organized based on economic indicators. Within these neighborhoods, a total of 89,835 households in Ouagadougou and 48,290 households in Bobo-Dioulasso were surveyed and ranked. The list of beneficiaries was then cross-checked by local committees and partners. But when the operation actually started in February 2009, about 10 percent of selected households could not be found (2,300 out of 20,000 households in Ouagadougou) and were replaced by the next-eligible households on the list. The targeting evaluation conducted by the IRD in April-May 2009 reported disappointing results in view of the efforts engaged, although concluding that all beneficiary households were needy. Considering food expenditures as the indicator for food vulnerability, the evaluation reported that a third of the beneficiary households were very poor (targeted group), a third were poor, and a third were less poor. The IRD explains this inclusion error by the fact that the vulnerability score (based on structural criteria) is not appropriate to estimate food poverty (largely cyclical). Budget constraints also led to an important exclusion error. Considering that about 60 percent of the population of Ouagadougou had not been visited, the IRD estimated that more than half of the poor households had de facto been excluded. The evaluation also reported serious issues with data collection and data entry. The questionnaire was complex and ill- adapted for a large-scale emergency intervention, and no adequate software for data management was used. Another highlighted issue is the fact that urban households move and their status evolves. The mid-term evaluation noted that many households that had been classified as poor in 2008, had fallen in the category very poor in 2009 becoming eligible for assistance. This suggests the introduction of mechanisms for re- targeting along the way. The IRD recommended exploring alternative methods, such as geographical targeting at a lower scale, combined with community-based targeting and a light questionnaire for pre- selected households. Transfer form, level, and delivery ­ A market analysis conducted in September 2008 indicated that markets were functioning and integrated, and the risk of inflation was low, suggesting a cash-based intervention was feasible. The WFP choose to use vouchers rather than cash to ensure that transfers are spent on food, for better security of beneficiaries, and for fear of mismanagement of the cash (as expressed in focus group discussions with women). The voucher option also allows restricting the list of accessible commodities to locally produced goods. Rice might be preferred by some beneficiaries, but it is imported and much more expensive than maize. All the cash injected (CFAF 2,600m in 2009) is spent in the country. The voucher value was set at CFAF 1,500 per person per month (up to a maximum of 6 persons per household ­ in polygamous households, each wife and her children would be considered as one household, which the WFP estimated to correspond to correspond to 22 percent of the poverty line, or 15-18 days of cereal needs). Accessible commodities are maize, vegetable oil, sugar, salt, and soap. The targeting evaluation suggested this level is appropriate. Beneficiaries (women) collect their vouchers monthly in one of the 20 distribution sites, and redeem them in one of the participating shops (one for 200 beneficiary households on average). Post-distribution monitoring revealed that 99 percent of beneficiaries redeemed all of their vouchers within the two days following the voucher distribution. Participating shops were selected in collaboration with the Ministry of Trade and are paid by the microfinance institution (Microfi). Nutritional supplements are distributed by MASSN agents through the health centers of the intervention area: Plumpy-Doz (or enriched flour Nutrifaso-Gret) to all children aged 6-24 months with an up-to-date health record, and CSB to all pregnant and lactating women during the lean season. 60 Impact ­ The mid-term evaluation (rapid qualitative review) indicated that savings on food expenditures generated by the program were too limited to allow beneficiary households to invest in income-generating activities, but had improved their food consumption both in quantity and quality, and to a smaller extent their access to health and education. Participating shops reported a positive impact on their turnover, as well as the creation of temporary jobs. No inflation was induced and most non-participating shops did not suffer a loss in business. Source: WFP, 2008a, 2008b, 2009a, 2009b; IRD, 2009; Authors. 116. Further analysis is expected in the first semester of 2010 to inform the design of effective social safety net mechanisms for urban populations. In late 2009, the WFP expanded its voucher program to assist another 6,562 households affected by the floods. On top of the vouchers, flood-affected households (8,906 households) received monthly cash transfers of CFAF 1,500 per person (with a ceiling of CFAF 9,000 per household) over four months. By providing this additional cash assistance, the WFP intended to prevent households from reselling their vouchers or food to access materials for housing repairs and essential non-food items. Both vouchers and cash were delivered to women. The post-distribution monitoring (PDM) that took place in March 2010 will provide information on how the cash was used and may shed some new light on the previously anticipated cash management issues. The final evaluation of the Urban Voucher program (both qualitative and quantitative, with a baseline and control group) launched in April 2010 is also expected to bring valuable additional evidence and learning for the design of efficient social safety nets in urban settings. B. Food Transfers 117. Food transfers are the main form of social safety net programs in Burkina Faso, accounting for 69 percent of total SSN spending and over 80 percent of all estimated SSN beneficiaries in 2009 (excluding fuel subsidies). The 2003 survey on household living conditions reported that 49 percent of the rural population was unable to meet a satisfactory food ration. This resulted in a high prevalence of chronic malnutrition, affecting nearly 38.7 percent of the children, along with a high child mortality rate (two children out of 10 die before the age of five). This alarming situation motivated many stakeholders to invest more in food transfer programs. Four types of food transfers are currently in place: (i) targeted subsidized food sales, (ii) targeted free food distributions, (iii) nutrition programs, and (iv) school feeding programs. Targeted Subsidized Food Sales 118. The government provides subsidized food sales using the Intervention Stock. The government Intervention Stock30 of 10,000 MT is aimed at assisting populations living in remote areas or affected by a disaster through targeted subsidized food sales (e.g., CFAF 10,000 instead of CFAF 35,000 per 100 kg-bag of sorghum or mil). The use of the Intervention Stock is decided at the central level ­ defining the areas, subsidized price, and period of operation. One of the objectives of these subsidized sales is to contribute to cereal price stability in deficit provinces. During the agricultural year 2007-2008, the 30 As of 30 June 2009, the level of the Intervention Stock was 10,404.14 MT, in line with its usual level of 10,000 MT, and made of 91 percent of maize, mil, sorghum, 5 percent of niebe, and 4 percent of local rice. 61 government estimated such effects to have been reached in several provinces, but only over a short period of time [RVCC 2008]. Another objective is to assist the households that are vulnerable to food insecurity. 119. There is no information on the actual number and profiles of beneficiaries of the subsidized food sales.31 Neither CONASUR nor SONAGESS, the agencies responsible for management and distribution of food stock, have information on the beneficiaries of their programs. In 2008, SONAGESS reported that the entire stock was used, along with an additional 20,000 MT borrowed on the National Security Stock (Stock National de Sécurité or SNS). The sale of these 30,000 MT of cereals could potentially have benefited up to 136,800 households if one considers the average household quarterly needs in cereals32, or up to 300,000 households if sales were limited to 100 kg per household. However, these indicative figures appear poorly consistent with related figures provided by the Food Security National Committee (CNSA)33 and the CONASUR34. . Table 18: Intervention Stock, Situation as of 31 December 2008 Center Mil (MT) Sorghum (MT) Maize (MT) Total (MT) Ouagarinter 122.40 507.60 119.92 749.92 CNSAO - 307.30 - 307.30 Kaya 406.50 - 152.10 558.60 Ouahigouya - 360.70 - 360.70 Titao Djibo 450.00 - - 450.00 Dori 976.20 - - 976.20 Total (MT) 1,955.10 1,175.60 272.02 3,402.72 Total (%) 57% 35% 8% 100% Source: SONAGESS. Table 19: National Food Security Stock, Situation as of 31 December 2008 Center Mil (MT) Sorghum (MT) Maize (MT) Total (MT) Ouagarinter - - 1,350.54 1,350.54 CNSAO - 1,950.00 738.09 2,688.09 Kaya 500.00 500.00 - 1,000.00 Ouahigouya 500.00 1,000.00 - 1,500.00 Titao - 500.00 - 500.00 Djibo 525.00 - - 525.00 Gorom-Gorom 1,050.00 - - 1,050.00 Aribinda 1,557.10 - - 1,557.10 Dori 691.80 1,243.00 - 1,934.80 Total (MT) 4,823.90 5,193.00 2,088.63 12,105.53 Total (%) 40% 43% 17% 100% Source: SONAGESS. 31 The COPROSUR is receiving only CFAF 100,000 per month for M&E activities. 32 Considering a household of six persons and a daily ration of 400 g of cereals per person. 33 The CNSA estimated that over the period 2002-2008, the number of supplementary daily food rations (400g of cereals) distributed through both subsidized sales and free food distributions exceeded 100 million, suggesting a total of 40,000 MT (only) of cereals was distributed over the period [SE-CNSA 2009]. 34 The CONASUR reported for 2005 a total of 13,342.5 MT distributed through subsidized sales to a total of 552,611 households [CONASUR 2008a]. This would represent only 24 Kg per household per year i.e. covering only 10 days of the average household needs. No figures on subsidized sales' tonnage and beneficiaries are provided for the years 2006, 2007 and 2008 [CONASUR 2008a and 2008b]. 62 120. These subsidized sales may not reach the poorest households since the very poorest groups may simply not have the financial resources to access the proposed subsidized cereals. Eligibility criteria are defined by COPROSUR (committee for emergency assistance at the provincial level). Interested eligible heads of households need to register at their village hall and collect their cereals at SONAGESS warehouse (at the provincial or departmental level). It is unclear whether sales are done on a first-come first- serve basis, and to what extent the criteria for vulnerability are applied. In the absence of strong monitoring and evaluation, and with the lack of arrangements for the effective identification of vulnerable households, it is not possible to assess the effectiveness of the Intervention Stock in support of the food consumption of the poor. 121. The CNSA envisions to further support the decentralization process with the establishment of cereal banks. Post-operation reports tend to indicate that quantities received are insufficient to cover needs. And CNSA is calling for an increase of the level of national stocks35 (Table 20) and a further decentralization of the process to create commune/community-based stocks. The Food Security Emergency Plan developed for the period 2008-2012 includes the establishment of one cereal bank in each commune in the medium and long term. The proposal includes CFAF 1,510m for bank construction and CFAF 7,550m for initial stock.36 Table 20: Indicative Numbers of Potential Beneficiaries of National Food Stocks Number of Number of Number of individual individual household Tonnage (MT) monthly rations quarterly rations quarterly rations (*) (*) (**) Intervention Stock (SI) 10,000 822,000 274,000 45,600 National Security Stock (SNS) 35,000 2,877,000 959,000 159,600 Financial Stock (SF) 25,000 2,055,000 685,000 114,000 Total 70,000 5,754,000 1,918,000 319,200 (*) Considering a daily ration of 400 g of cereals. (**) Considering an average household size of six members. Note: The central system has a delivery capacity of 365 MT (26 trucks). Source: Authors. Targeted Free Food Distributions 122. Because the country has been producing surpluses for the past 10 years, the National Security Stock has never been used to distribute free food on a large scale. The use of the National Security Stock (SNS) is to be agreed by both the government and international donors. According to criteria set in the PNOCSUR (Plan National d'Organisation et de Coordination des Secours d'Urgences et de Réhabilitation), once there is a national deficit of at least 5 percent, free food is distributed as a response to the emergency. On average, Burkina's cereal production increased by 7 percent a year over 35 The 70,000 MT level defined 10 years ago was never revised despite huge demographic growth. 36 The food security emergency plan developed for the period 2008-2012 includes in the short term: targeted food distributions, school feeding, nutrition, mother and child health care; and in the medium and long term: cereal banks; program of public purchase to increase market demand for nutritive food; expansion of school feeding programs based on local production; investments in the nutrition of mothers; integration of nutrition in child health promotion programs (under "improve the use of food commodities"); increased employment opportunities, esp. for women; construction of 10,516 km of rural trails (MID, CFAF210,720m 2009-2012). 63 the period 1997-2007, with an average surplus of 447,570 MT and an average annual cover rate of 115 percent ­ down to 92 percent in 1997-1998. Yet, some areas remain structurally in deficit; 14 deficit areas were identified in 2009, most of them in the North (poor pluviometry) and Kadiogo Province (capital province where trade prevails over agriculture). These areas are supplied by traders, humanitarian organizations,37 and the national system. The needs were estimated at 22,354 MT (for three-month consumption) in 2008-2009. Donors and partner NGOs covered 5,819 MT, and the state covered 13,435 MT [SE-CNSA 2009]. 123. The national system has been providing food assistance to a few tens of thousands of people every year, on an ad-hoc basis. Assistance to vulnerable populations is made based on requests received from charities; in 2008, CONASUR responded positively to 342 requests out of 414. Its assistance to vulnerable populations consists of food and non-food items (Table 21). The quantity of cereals provided to vulnerable populations in 2008 was equivalent to 31,800 individual monthly rations. In addition, CONASUR provides assistance to disaster-affected persons; the quantity of cereals it provided to disaster-affected persons in 2008 was equivalent to 12,550 individual monthly rations (Table 22). Most responses were to flooding or small-scale disasters (farmer-breeder conflict, fire). Table 21: CONASUR's Assistance in Cereals to Vulnerable Populations Item 2007 2008 Cereals (MT) 204.37 381.93 Maize (MT) 101.10 235.56 Sorghum (MT) 3.60 4.75 Mil (MT) 43.09 - Rice (MT) 56.58 141.62 Note: This table only presents the quantity of cereals distributed as an indication. Other food items were also distributed (oil, dates, biscuits, and milk in 2007; sardines, corned beef, and sugar in 2008) along with non- food items. Source: SP/CONASUR 2008a and 2008b. Table 22: CONASUR's Assistance in Cereals to Disaster-Affected Persons Item 2007 2008 Cereals (MT) 1,121.28 150.63 Maize (MT) 824.78 127.53 Sorghum (MT) 91.50 2.50 Mil (MT) 100.00 - Rice (MT) 105.00 20.60 Note: This table only presents the quantity of cereals distributed as an indication. Other food items were also distributed (corned beef, sardines, and oil in 2008; BP5 biscuits, dates, canned food, and sugar in 2007) along with non-food items. Source: SP/CONASUR, 2008a and 2008b. 124. In the future, CONASUR would like to provide disaster-affected populations with full rations. The food security objective of the new national contingency plan, adopted in February 2009 and managed by CONASUR, is to maintain a response capacity 37 In terms of food aid, over the period 2006-2009, the country registered 18,324 MT of sorghum and maize, 5,003.1 MT of niebe from WFP, and 14,532.484 MT of rice from Japan. Most of it was distributed through free distributions or FFW programs. 64 to adequately feed 300,000 persons over a month. The plan foresees the proposition of 4,545 MT, consisting of 3,600 MT of cereals, 540 MT of pulses, 225 MT of oil, 135 MT of sugar, and 45 MT of salt (i.e., a full ration of 2,100 Kcal/person/day) in the 13 regions to reduce delivery time [GoBF 2009a]. The mobilization of the stock is decided at the regional level, when the number of affected persons does not exceed 5,000; above this threshold, the central system would be activated. As of September 2009, this plan had yet to be funded. 125. Overall, targeted food distributions (by CONASUR, WFP, and CRS) may currently reach less than 100,000 persons every year, which is much less than the Urban Voucher program alone (Table 23).38 Through its Food-for-Education/Training program, the WFP provides food transfers conditional on attendance at literacy centers or training courses (e.g., on asset maintenance, improved agricultural techniques, etc). The WFP program has been gradually refocused on the Sahel Provinces, which were being scaled down from 21 provinces in 2005 to 9 provinces in 2010. Through its General Relief program, CRS assists over 13,000 vulnerable persons (e.g., people living with HIV, orphans, the elderly, and the disabled) every year. Table 23: Indicative Number of Beneficiaries of Targeted Food Distributions, 2008/2009 Program 2008/2009 WFP Food for Education (2009 data) 36,648 Female beneficiaries 20,300 Male beneficiaries 16,348 CRS General Relief (2008 data) 14,842 (*) CONASUR Vulnerable Populations (2008 data) 31,800 (**) CONASUR Disaster-Affected Persons (2008 data) 12,550 Total 95,840 (*) Equivalent in individual monthly rations of quantity of cereals provided. (**) Equivalent in individual monthly rations of quantity of cereals provided. In 2008, CONASUR registered 27,383 disaster-affected persons. Source: WFP; CRS; CONASUR. 126. Food insecurity is primarily an access problem in Burkina Faso, and the choice of cash-based transfers might be preferred. The agriculture production is dominated by traditional cereals, which are largely for self-consumption. In the North and the Sahel, where production is low, households need sources of income to buy cereals on the market. Gross production per capita increased by 20 percent between 1996 (0.236 MT) and 2009 (0.283 MT). Accessibility is being improved through market economy development and efforts to expand to remote areas. Cash-based interventions coupled with market interventions may have a positive impact on local development (Box 7). Food transfers may undermine local markets and further impede the development of private traders. Food is sometimes provided due to a generally higher probability of control over food being retained by women, which is expected to result in a greater impact on family welfare. Whether these benefits are real or worth the extra costs, remains questionable. It is clear that food transfers may still be needed in drought years if the market fails to react efficiently. Finally, programs should consider switching to cash vouchers whenever possible, especially in response to chronic food insecurity/poverty. 38 This does not include food transfers provided under nutrition and food-for-work interventions. 65 Box 7: Transfers In Cash and In Kind: Alternatives or Complements? When are food transfer programs appropriate? What are the criteria to keep in mind when deciding how much to distribute in the form of rations and how much as cash. Program designers should keep the following four key considerations in mind when deciding if food transfer programs are appropriate or necessary: The functioning of food markets, including access, transport, and storage, and how this is reflected in the prices of staples. If markets are well integrated across regions, cash transfers have an advantage because of the private sector's superior ability to move food and other goods more efficiently than the public sector. Furthermore, some argue that providing cash can have a positive impact on smaller trade and other economic activities (Devereux 2000). However, if markets are thin, poorly integrated across regions, or monopolistic, the provision of cash may increase prices, which reduces the value of the transfer and may cause additional hardship to those poor households that do not receive any transfers (Devereux, Mvula, and Solomon 2006). A close monitoring of prices, not of production, is needed to assess the situation. The level of transaction costs for the program and for beneficiaries. Most of the argument about transaction costs refers to the high cost of distributing food provided by donors compared with the relatively lower cost of distributing cash. Food distribution takes time to organize, requires storage and transport, and is subject to losses and pilferage; and the public sector tends not to be efficient at keeping costs down. However, in some places where marketing and transport channels are not developed, only the public sector can provide adequate supplies in local markets. Beneficiary transaction costs also need to be taken into account. These costs include the time and expense of going to local markets, which might increase if places are far or unsafe. The impact of the form and size of the transfer in determining the level of food consumption. Poor households are more likely to consume food and to eat good food if they receive a small transfer. Some claim that men might use cash transfers to purchase such commodities as cigarettes or alcohol, and the literature indicates that small food transfers result in higher food consumption than cash transfers (del Ninno and Dorosh 2003; Fraker 1990). Moreover, Hoddinott and Islam (2007) and Jacoby (2002) show that households are more likely to stick (the so-called flypaper effect) to consumption patterns and intra- household distributions that have a positive impact on the nutrition of children if they have access to small transfers of good food. The preferences of the beneficiaries. Beneficiary preferences may vary depending on circumstances. Even though beneficiaries may prefer cash simply because it is more flexible, they still want to maximize the level of the transfer and their control over it. This is why women in certain circumstances might prefer food to cash (see Ahmed, Quisumbing, and Hoddinott 2007 on Bangladesh and Sharma 2006 on Sri Lanka). Source: Grosh et al., 2008. Nutrition programs 127. The government and its partners, UNICEF and the World Bank in particular, made substantial efforts since 2003 to increase nutrition interventions. UNICEF, WFP, ECHO, and their partner NGOs are particularly active in the treatment of moderately and severely malnourished children under 5 and pregnant and lactating mothers. The WFP also provides nutritional support to people living with HIV. While the national health information system has yet to track nutrition data, UNICEF estimated the number of severely malnourished children treated in 2009 at roughly 25,000 ­ out of which 90 percent recovered, 2 percent defaulted, and 2 percent died (information missing for 6 percent). Médecins Sans Frontières in the Titao and Yako Districts reported over 20,000 severely malnourished children treated in 2008. 128. Yet, inadequate coverage of screening and management of severe acute malnutrition is identified as one of the main challenges in reaching the nutrition MDG (Box 8). Despite significant efforts, the nutrition situation remains alarming. Nearly two out of five children under 5 suffer from acute malnutrition, particularly children aged 24-59 months; over one out of three children suffer from stunting ­ with 66 enormous geographical disparities; and in rural areas, one out of seven girls aged 15-19 suffers malnutrition. The rise in food prices since 2006 was a major risk factor for increased malnutrition, particularly in urban areas. One challenge is to show how the leadership at MoH in favor of nutrition can establish efficient links with other sectors relevant to nutrition, such as social protection, water and environment, and community development. Box 8: Planning for Going to Scale with Nutrition Interventions The government and its partners, UNICEF and the World Bank in particular, have made substantial efforts since 2003 to increase nutrition interventions. From 1993 to 2003, underweight increased by 0.5 pp/year, compromising achievement of the MDG, reaching 35 percent of underweight in 2003. Redressing the status of malnutrition gained prominence in the government's agenda, and thereafter encouraging outcomes resulted. Figure B6.1: Turning Point in Underweight Prevalence Source: WHO 2009, reference WHO 2005. Meeting the MDG target for nutrition is possible. There is an urgent need to accelerate efforts to bring underweight prevalence further down. Today, the country has: an appropriate organizations framework; a well-defined vision; and, a results- based National Nutrition Strategic Plan, including a detailed costed action plan. The main sources of funding for nutrition are pooled funding (43 percent, of which 8 percent comes from World Bank nutrition fund) and UNICEF (31 percent). Burkina Faso presents a favorable environment for success, with: commitment to International Health Partnership (IHP+) and preparation of COMPACT; well-established mechanisms for pooled funding for health; strong support for new approach of contracting with NGOs to accelerate community health and nutrition programs; and, development of a performance-based financing system. The main challenges in reaching nutrition MDG are: inappropriate infant and young child feeding practices (exclusive breastfeeding for six months and complementary feeding practices); inadequate coverage of screening and management of severe acute malnutrition; inadequate utilization of zinc in treatment of diarrhea in young children; low utilization of long-lasting insecticide treated nets by women and children; low sanitation rate and inadequate hygiene practices; inadequate qualified human resources for nutrition and behavioral change communication; and, inadequate funding for prevention of malnutrition. To scale up nutrition interventions and reach the MDGs, Burkina Faso still needs to: reinforce nutrition capacity at all levels; strengthen involvement of NGOs and CBOs in community nutrition activities; and, ensure adequate and constant investments from national budget and international development partners. Source: MoH, 2009. 67 School Feeding 129. School feeding represents one of the main social safety net programs currently in place in Burkina Faso. In 2009, this program accounted for 24 percent of total spending on SSN, and about 38 percent of the estimated total number of SSN beneficiaries (excluding food and fuel subsidies). MEBA alone spent an average of CFAF 4.2bn annually between 2005 and 2009 for "endogenous" (initiated by the communities and supported by the state) school feeding activities in primary schools. In addition, CRS and the WFP run "assisted" school feeding programs in the most vulnerable areas. At present, schools receive only a three-month supply of food from MEBA, while the CRS and WFP programs provide support for the whole school year. MEBA program is complemented by community contributions (providing support for endogenous school feeding models). In 2005-2006, the DAMSE reported that a total of 4,636 schools ­ slightly less than half the total number of primary schools ­ benefitted from school feeding programs supported by CRS (46 percent), MEBA (45 percent), and WFP (9 percent). The school feeding strategy planned a gradual handover of the CRS school feeding programs to MEBA and maintenance of the assisted school feeding programs in the Sahel, since the latter is a food- insecure area (Table 24). MESSRS also supports school feeding in secondary schools.39 In 2007-2008, 20,700 students benefited from subsidised meals in 310 schools. Even though demand is rising, not even 10 percent of the students were covered by the 2009 budget of CFAF 760m in 2009 covered. The program also suffers from low recovery rates (45.73 percent in 2007), and various implementation issues [MESSRS 2009]. Table 24: Main Characteristics of School Feeding Programs in Primary Schools Assisted Endogenous Program Type School Feeding School Feeding Implementation CRS CRS WFP and NGOs MEBA Funding USAID (FFP) USAID (MCA) Multilateral State (CAST) Hot meal at school Lunch Lunch Lunch and Lunch breakfast Take-home ration 8 kg of fortified 5 kg of rice per girl 10 kg of cereals per none maize flour per girl per month to all girl per month to per month to all girls attending girls attending the girls attending schools two last grades of primary school primary school Origin of In kind In kind Local and Community commodities contributions from contributions from international contributions and the USA the USA purchases local purchases Dynamic Phased withdrawal/ Stable Stable Expansion/taking handover to MEBA (BRIGHT schools) (Sahel) over from CRS Source: CRS; WFP; MEBA; Authors. 130. A large scale up of school feeding activities in primary schools is planned for the 2010-2011 school year. Under the Crisis Action Plan, MEBA was allocated supplementary annual budget of CFAF 12.3bn for school feeding activities for the years 39 The DAF/MESSRS procures for beneficiary schools ­ both public and designated private schools can apply ­ kits and food commodities ­ rice, pasta, sardines, tomatoes (four imported products), niebe, and yam ­ for one-third of the students. Schools are responsible for transporting the goods and serving meals at a subsidized price: CFAF 100 in Ouagadougou and Bobo-Dioulasso, CFAF 50 in the Sahel, and CFAF 75 elsewhere. Schools have to pay CFAF 35 per meal to the Treasury for administrative costs. 68 2010 and 2011 (on top of the CFAF 5,270bn already allocated in 2010). The Crisis Action Plan project aims at providing a meal every school day to 100 percent of the children attending a public school versus the current coverage of 40 percent. On top of the planned procurement of 28,548 MT of food commodities (cereals and oil), the following activities are envisioned: (i) constructing/renting regional storage; (ii) purchasing pallets for storage; (iii) purchasing stationary and computers; (iv) transporting food commodities from regional stores to schools; (v) monitoring and evaluation of commodity management; and (vi) training controllers and storekeepers in stock management and storage maintenance. It is still unclear whether the planned expansion of school feeding activities to all primary schools for the whole year is a temporary measure to mitigate the effects of the recent crisis or a long-term strategy. Sustaining such a strategy will require substantial resources. While scaling up school feeding might be a legitimate response to the food crisis, authorities need to further investigate the most appropriate social safety net instrument over the long run. Moreover, given the financial constrains, to ensure the sustainability of the program, targeting mechanisms and the role of community contributions need to be investigated. 131. In 2008, an evaluation of the impact of two school feeding schemes on educational and health outcomes of children from low income households reported mixed findings. The two forms of school feeding programs under consideration were school meals where pupils are provided with lunch each school day, and take-home rations (THR) which provide girls with 10 kg of cereal flour each month, conditional on 90 percent attendance rate (both provided under the WFP program). After the program ran for one academic year, the study found that both school feeding programs increased girls' enrolment by 5 to 6 percent ­ with no evidence that THR targeting girls crowds out boys registration. A negative effect of school meals on raw scores on mathematics was observed for 13-15 year old boys, suggesting that school meals may have a negative impact on schools which are less well organized. No other significant impact on raw scores on mathematics was observed, but the time-adjusted scores on mathematics improved slightly for girls in school meals villages. An unexpected lower average attendance was observed ­ increased enrolment may be accompanied by lower attendance. Authors argue that this reflects the absence of an active labor market and the fact that households are labor constrained and/or child labor is complementary to adult labor. The interventions appear to have caused attendance to decrease in households with a low in child labor supply while attendance improved for households with a relatively large child labor supply, consistent with the labor constraints. This, in turn, explains the mixed impacts on learning outcomes that were observed. Finally, for younger siblings of beneficiaries, aged between 12 and 60 months who were not in school, take-home rations increased weight-for-age by .38 standard deviations and weight-for-height by .33 standard deviations. In contrast, school meals did not have any significant impact on the nutrition of these younger children [Kazianga et al. 2008]. 132. These mixed results invite potential reconsideration of the choice of the most cost-effective model to increase school attendance rates among the poorest and most vulnerable populations. In recent years, improvements have occurred in the primary school enrolment, attendance, and achievement indicators. However, it is unclear whether this is due to school feeding or other Education For All initiatives, and whether greater results could have been achieved in a more cost-effective manner with alternative instruments. The following options may be considered to increase school attendance rates: in-school meals only; take-home rations only; conditional cash transfers; and non- conditional cash transfers. To inform such a reflection, more information and evidence on 69 the cost40 and impact (on school attendance, local development, etc.) of the different types of programs is needed, along with demand-side and supply-side intervention strategies becoming consistent. 133. Despite the global evidence on the positive impact of school feeding, further research is required to assess the longer-term relative merits of in-school feeding versus take-home rations or other social safety net instruments (conditional cash transfers). School feeding programs may increase school attendance, cognition, and educational achievement, particularly if supported by complementary actions such as de- worming and micronutrient fortification or supplementation [Bundy et al. 2009]. In discussing the effectiveness of school feeding modalities, Bundy et al. [2009] recognize a particular need for better data on the cost-effectiveness of the available school feeding approaches and modalities. 134. School feeding programs may not reach the poorest and most vulnerable. In the poorest areas, where school enrolment is low, school feeding may not reach the poorest people. First, the poor are less likely to be in school than the non-poor. Second, it is extremely difficult to target benefits to the poor within a school, except with take-home rations which are not that different from conditional cash transfers. Third, school feeding activities are expensive and need to be geographically targeted, but then the program does not provide benefits to the majority of the poor who live in areas not covered. 135. Food assistance programs could be better linked to local production. Increasingly, local purchasing has been promoted and better linkages between food assistance programs and support programs to small farmers have been established. Globally, local procurement is being actively evaluated as a means to achieve sustainable programs and, at the same time, to use the purchasing power of the program as a force multiplier and a stimulus for the local agricultural economy [Bundy et al. 2009]. Yet, in USAID-funded programs, food is provided in kind from the United States. This approach does not support local, national, or regional markets. In contrast, the WFP introduced the Purchase for Progress (P4P) program, which aims to procure a significant part of the food from associations of small farmers.41 The promotion of local produce is also part of "Burkina 2025" and the emergency plan for the achievement of food and nutrition security in Burkina Faso (2008-2012). C. Universal Food and Fuel Subsidies 136. The universal food subsidies introduced by the government in 2008 in response to the crisis induced by high prices worldwide were very expensive and inefficient in reaching the poor. To mitigate the negative effects of high world food prices globally, the government suspended the VAT or import tariff on a number of basic food commodities in March 2008 (Table 25). The loss generated by these measures was estimated to over CFAF 4bn (0.1 percent of GDP). In terms of the impact on food prices, the following months showed a price increase for rice and no price reduction for the other 40 Indicatively, between 2006 and 2010, the average cost of WFP school feeding program was CFAF 21,000 per pupil per year. 41 The WFP's Purchase for Progress (P4P) initiative aims at ensuring that 10-15 percent of commodities are purchased from local producers. This initiative was launched in January 2009 with the support of the Bill and Melinda Gates Foundation, and aims at strengthening production channels. 70 subsidized commodities. In terms of the impact on the well-being of households, the two poorest quintiles were hit the hardest by high food prices yet benefited very little from these tax exemptions: Only 10 percent of the benefits went to the poorest quintile. Within six months (October 2008) the measures were rescinded, the poorly performing program terminated. Table 25: Price implications of exemptions on targeted staple products, 2008 Product Type of exemption Rate of tax suspended (%) Rice Import tariff 10 Salt Import tariff 5 Milk and milk products Import tariff 9 Pasta VAT 18 Soap VAT 18 Vegetable oil VAT 18 Source: IMF, 2008. 137. The government has been subsidizing a number of fuel products for years, despite a high fiscal cost and very limited impact on the poorest households.42 Since July 2007, the government used a pass-through mechanism for oil products, more on an ad- hoc basis than in an automatic manner. The fiscal cost generated was estimated to 0.8 percent of GDP. To stem the use of charcoal and promote environmental conservation, LPG (gaz butane) is subsidized through subventions to distributors at a cost of 0.3 percent of GDP. And, in an attempt to maintain low electricity prices, the fuel used in electricity- producing generators is also subsidized, at a cost of 0.4 percent of GDP. Overall, 2007 fuel subsidies represented a cost of 0.73 percent of GDP.43 This is an extremely high cost for a very limited impact on the poor. In fact, it was estimated that over 84 percent of the benefits went to the non-poor, 98 percent when considering the butane subsidy benefits alone. Kerosene (pétrole lampant) subsidies proved better suited to the poor, with 38 percent of benefits reaching the two poorest quintiles. Table 26: Price Implications of Exemptions on Targeted Fuel Products, 2008 Gasoline LPG Kerosene Diesel 1 Diesel 2 (premium) (butane) (pétrole) (gasoil) (DDO) Explicit subsidy (to distributors) Implicit subsidy (SONABHY) Suspension customs tax Suspension excise petroleum tax Suspension VAT Total subsidy, % of current price(*) 10% 240% 81% 22% 75% (*) As of September 2008 Source: IMF, 2008. 42 An explicit subsidy for LPG was introduced before 2007 as a long-term measure. Other measures to offset rising fuel prices were introduced in July 2007, January 2008, and July 2008. 43 According to IMF, taking into account losses in consumption taxes due to these measures, the total cost of fuel subsidies are estimated at 2.9 percent of GDP. 71 Figure 11: Share of Subsidy Benefits Source: IMF, 2008. D. Labor intensive Public Works and Cash-/Food-for-Work 138. The Ministry of Youth and Employment, through the Employment National Policy, aims at promoting the use of labor intensive public works approach. Recent simulations suggest that investing in infrastructure may have a substantial positive impact on economic growth and progress toward the MDGs [World Bank 2009b]. The Ministry of Youth and Employment (MJE) has launched a critical review of past experiences in public works projects ­ many unsuccessful ­ to draw lessons learned, and envisions the design of two new pilot projects: one in urban settings, one in rural settings.44 139. Even though positive, the experience of the ongoing PrEst, which uses a labor intensive approach for infrastructure creation in rural areas although not designed as a public works program to specifically reach the poor, did not produce replicable knowledge.45 The program reported positive impacts on poverty reduction and local development like increased use of health services, increased school enrolment, slight reduction of child mortality, reduction of youth migration, and poverty reduction among direct beneficiaries [Helvetas 2008; Balkouma Boursin 2008]. It has been implemented by the NGO Helvetas since 2002 under the supervision of the Ministry of Infrastructure, and is expected to continue until 2011. Between 2003 and 2008, over 200 km of rural trails in the Gourma, Gnagna, and Tapoa Provinces were built, benefiting 42 villages with an estimated population of 492,000 inhabitants.46 Between 2005 and 2007, the project employed an annual average of 865 persons (mainly youth aged 20-35, of which 16 percent were women) per year for about six months. The average remuneration was CFAF 44 In 2009, the MJE allocated a budget of CFAF 10m for such an activity, and established an inter- ministerial committee to supervise the review. 45 Program Pistes Rurales ­ Désenclavement à l'Est. 46 The project contributes to the National Strategy for Rural Transport of 2003, which set the objective of rehabilitating 10,700 km of rural trails for a total cost of CFAF 58bn within 10 years. 72 130,950 per annum (fluctuating from CFAF 19,275 to 26,437 per month), slightly below the guaranteed minimum wage (FCFA 33,139 per month in 2008) but well above the poverty line (CFAF 82,672 per annum in the Eastern Region) [Helvetas 2008]. PrEst was designed to create locally managed rural trails, focusing on local development as a strategy for poverty reduction; and it also offered complementary activities such as training on gender issues, environment, and HIV-AIDS. It was, however, not designed as a safety net program per se. Yet, the program managed to build the capacity among local communities, authorities, and private enterprises for the labor intensive approach. The fundamentals of PrEst offer national stakeholders a blueprint for public works programs that could be expanded to reach the poor (Box 9). Box 9: Public Works Programs ­ Elements Required for Reaching the Poor Self-targeting by setting the wage rate at an appropriate level. In a context where poverty targeting appears particularly challenging, and where financial and administrative capacities remain limited, relying on self-targeting is attractive. However, this will only be possible if the market wage is above the minimum wage. Indeed, the publicly funded program wage cannot be lower than the minimum wage and would hence be higher than the local wage for unskilled labor, thus likely to attract the non-poor to the public works program. So if the minimum wage is equal or above the market wage and restrictive employment laws prevent setting the wage below the minimum level, the possibility of using self-targeting is hindered and other targeting mechanisms need to be introduced. The use of pure self-selection might also be insufficient in reaching vulnerable groups in poor areas or when demand for participation is very large and some form of employment rationing is needed. The fact that youth aged 15-24 represent a third of the unemployed ­ with young women even hit harder than young men ­ may also suggest adopting targeting methods to reach these categories specifically. In addition, setting the program wage too low also presents the risk of excluding poor households that have higher opportunity costs of labor ­ if the program wage is below the reservation wage or the risk of missing program objectives (e.g., nutrition objective if the program wage is far below the cost of the minimum basket). It is crucial to ensure the program wage is set in relation to the project goals. Provision of quality public goods is crucial. Based on international experience, public works should only be promoted as a social safety net instrument if the public goods generated have a positive impact on the community and are built at a cost similar to that charged using hired contracting procedures. It should not be introduced as strategies to provide social transfers to "deserving" poor. Public works projects may include traditional infrastructures or public environmental improvement projects (e.g., sanitation projects to roll back malaria, natural disaster risk reduction projects), but also social activities (e.g., South Africa's home-based care workers and early childhood development workers) or economic activities (e.g., small businesses and cooperatives). The public goods produced if relevant, well executed, and maintained, could have an important role in alleviating constraints to higher returns for poor people, regardless of their participation in the program. Since 2004, the WFP has promoted synergies between food-for-work programs and school feeding and nutrition programs (e.g., building classrooms, storage rooms, latrines, etc.). Community projects benefiting women are also given priority by the World Food Program. To address chronic poverty, public works programs should run throughout the year with varying degrees of intensity. A program run during agricultural slack seasons only, when the opportunity cost of labor is low, would provide "consumption-smoothing" for poor households but no assurance of finding a job whenever it is needed. A program operating throughout the year with varying degrees of intensity will provide both "insurance" and "consumption-smoothing" for poor households. In countries with widespread levels of unemployment and underemployment, standard short-term public works programs proved unable to lift the chronic poor out of poverty. Brazil, Argentina, India, and Bangladesh represent some good practice examples where the program served the functions of assurance, consumption-smoothing and poverty reduction. To ensure additional coverage, the number of days worked can be rationed and a rotation system applied. For instance, India provides a legal guarantee of 100 days of employment a year to any rural household willing to do public work for a statutory minimum wage, and Ethiopia assists over 7 million chronically food-insecure people ­ about 10 percent of the population ­ through its Productive Safety Net program's employment schemes and food or cash transfers. This being said, high labor intensive public works projects can also be effectively used in the aftermath of natural disasters for the rehabilitation and reconstruction of damaged or destroyed infrastructures. Source: Grosh et al., 2008; del Ninno et al., 2009. 73 140. Lessons can also be learned from the WFP's Food-For-Asset program, although again, the focus of this program is primarily put on the building of quality assets and to a lesser extent, on the labor intensity of the project. Food aid is used as an incentive to carry out community works such as small anti-erosion dykes and embankments, or work whose results are only apparent after a year or so, such as swamp drainage and dykes. The program contributes to mitigating ongoing soil degradation in arable or potentially arable lands and to support initiatives aimed at settling or farming highly productive farmlands ­ market gardening on reclaimed land or with small-scale irrigation. The resources and reach of the program have increased over time (Table 27). Table 27: WFP Food for Asset Programs (Beneficiaries and Budgets), 2006-2010 2006 2007 2008 2009 2010 Female beneficiaries 14,880 13,870 21,606 30,808 35,300 Male beneficiaries 9,920 14,880 33,794 31,408 35,300 Total beneficiaries 24,800 28,750 55,400 62,216 70,600 Activity costs (USD) 730,964 327,121 1,127,271 1,022,408 1,436,820 Administrative costs (USD) 51,167 22,898 78,909 71,568 100,580 Total costs (USD) 782,131 350,019 1,206,180 1,093,976 1,537,400 Share of administrative costs 6.5% 6.5% 6.5% 6.5% 6.5% Average cost per beneficiary (USD) 31.54 12.17 21.77 17.58 21.78 Source: WFP. 141. Building on the PrEst and the Food-For-Asset programs experiences and capacities, other types of public works programs could be introduced as effective social safety net instruments.47 A recent review of the experience with public works programs in several countries shows that well-designed and implemented public works programs can help mitigate income shocks and be used as an effective anti-poverty instrument (Box 10). However, the effectiveness of public works as a safety net instrument highly depends on the ability of the program to provide additional sources of income to the most vulnerable population when most needed. Moreover, further attention would need to be put on the targeting methods (Box 10), length and timing of work, specific design features that can increase the participation of women, and community participation. In particular, the choice of the remuneration method can affect the targeting and outcomes of the public works program. Task-based payment provides flexibility and may attract more women to worksites. In addition, community involvement in the selection of public works projects is crucial to ensure the most-needed assets and to create ownership. 47 In Ouagadougou, about 500 women are employed in the Brigades Vertes and are paid less than the minimum wage. 74 Box 10: Targeting Effectiveness of Public Works Programs: International Experience While there is a need to keep in mind the ultimate objectives of static and dynamic poverty reduction, targeting effectiveness clearly remains an important factor in assessing cost-effectiveness of public works programs. In general, targeting performance has been good relative to other targeted public interventions: · In Chile, almost 100 percent of participating households were poor. · In Argentina's Trabajar program, around 80 percent of households were from the poorest quintile, and 60 percent from the poorest decile. In addition to the low program wage rate, this outcome was achieved by including the local level poverty rate in the scoring formula for determining program allocations. · Various researchers have found positive targeting outcomes on the India's Maharashtra Employment Guarantee Scheme (MEGS), at least prior to the 1988 increase in minimum (and hence program) wages. A dedicated survey of MEGS participants from 1978-79 found 90 percent of workers living below the poverty line, even including EGS earnings in the calculation, at a time when the poverty headcount similarly calculated was 49 percent. A later survey in 1985-86 found mean incomes of participating MEGS households to be around 20 percent below the poverty line. The positive targeting outcomes are confirmed by other analyses for 1979-85. There is also evidence that the scheme's targeting performance deteriorated when the wage rate rose substantially in 1988. · In Bangladesh, targeting of the Food-for-Work program was also found to be good, with program participants coming from poorer households than the rural population as a whole, and the total number of person days of work created even more pro-poor. While around 25 percent of all rural households had incomes below 1,500 taka in 1981-82, around 60 percent of FFWP participants had incomes below this level. Equally, around 70 percent of program employment went to the 25 percent of rural households with incomes below this level. · Under the Bolivia Social Investment Fund, 77 percent of participants came from the poorest 40 percent of the population. Assessments of other SIFs in Honduras, Nicaragua, and Peru have also found pro-poor incidence of SIF benefits. · In Philippines, most public works participants were from marginally poor and non-poor households rather than the very poor. This outcome was largely driven by program compensation that provided for official minimum wage in cash plus in kind benefits which took the total compensation package above the market wage. · In Indonesia, assessment of safety net programs introduced after the Asian crisis found that public works schemes that relied on self-targeting were much more likely to reach households that had suffered large shocks than programs that relied on administrative targeting methods, including subsidized rice, scholarships, and health subsidies. · Analysis from all-India NSS data on public works schemes for 1993-1994 indicates a pro-poor targeting performance but also significant inclusion of errors. Targeting performance was also better in comparison to both PDS (then untargeted) and the micro finance program (IRDP). Sources: K. Subbarao, 2003; C. del Ninno, K. Subbarao, and A. Milazzo, 2007. E. Fee Waivers for Health 142. Several initiatives intend to provide free health care to the poor and vulnerable. The Health Development National Plan (PNDS 2001-2010) includes specific activities to: (i) promote preventive and curative care for vulnerable groups (children, women, youth, the elderly, and PLWHIV); and (ii) ensure access to health services to the poorest. As discussed in Chapter III, a number of measures were adopted to abolish or subsidize user fees for specific medical care, in particular those used by women and young children. These universal health subsidies were excluded from the scope of the present study. Instead, the review is focused at health subsidies specifically targeted at the indigent. 75 143. Measures to provide free health care specifically to the indigent are not enforced due to a lack of mechanisms to identify the indigent. The Kiti n°An-VIII- 0202/FP/SAN- as dated 8 February 1991 states that the indigent are entitled to free health services. However, in practice, 20 years later, national mechanisms have yet to be clarified to make the measure effective. The SONU subsidy policy of 2009 entitled indigent women to free health services, yet it failed to specify the parameters of qualification and, thus, prevented these very women from receiving services [AI 2009].48 Three key questions remained for the establishment of fee waivers for health: (i) who is indigent (or poor)?; (ii) who decides upon this status?; and (iii) who will pay for their fee waivers?. In 2007, a Joint Ministerial Committee on Indigent Care was established under the supervision of MoH to look at these issues, but it has never been very active; its reactivation is in MoH agenda for 2010. In late 2009, a directive from MoH called on health centers to actually use the CFAF 200,000 allocated for indigent care; yet with the defining criteria still not address, the directive cannot be operationalized. Nevertheless, a few interesting initiatives for the identification of the indigent/poor are being conducted, especially with: the University of Montreal (Centre de Recherche du Centre Hospitalier de l'Université de Montréal or CRCHUM) in collaboration with the IRSS/CNRST (Institut de Recherche en Sciences de la Santé du CNRST) (Box 11), and the Nouna Research Center on Health (Centre de Recherche en Santé de Nouna or CRSN) in rural areas; and the WFP and the IRD (Institut de Recherche pour le Développement) in urban settings (Box 6). The University of Montreal research has brought some valuable elements on critical points: possible identification methods of the indigent and possible financing sources for such a system. These research findings also showed that fee waiver systems for the indigent may actually support the strengthening of the Bamako Initiative pillars: cost recovery, community participation, and equity. 48 When the SONU subsidy policy was defined in June 2006, the 80 percent subsidy option was chosen ­ versus 60 or 100 percent ­ not for financial reasons but for fear that gratuity would encourage the population to use the services irresponsibly [AI 2009]. In February 2010, the President of Burkina Faso committed to lift all financial barriers to emergency obstetrical care and access to family planning in the strategy to fight maternal mortality [AI 2009]. It is worth noting here that subsidizing emergency obstetrical care at 100 percent for all would not actually require any additional financial resources. A miscalculation of the cost of the intervention (estimated at CFAF 4,500 when it costs only CFAF 2,000 in reality) when designing the SONU subsidy policy meant that when MoH reimburses CFAF 3,900 (80 percent) and users pay CFAF 900, health centers actually make money ­ not even considering that the health staff receives 20 percent of the cost of each intervention [Ridde and Bicaba 2009b]. The CFAF 30bn allocated for the SONU subsidies over the period 2006-2015 would be enough to provide full subsidy to all. 76 Box 11: A Community-Based Targeting Approach to Exempt the Worst-Off From User Fees in Burkina Faso To contend with the risk of exclusion created by user fees, those implementing the Bamako Initiative (BI) were asked to organize exemption schemes for the indigent. But those exemption schemes were never put in place in Africa due to difficulties identifying the indigent. The University of Montreal implemented an action research in collaboration with the IRSS/CNRST to test the hypothesis that a community-based process for selecting beneficiaries of user-fee exemptions is feasible. Three mechanisms were reviewed: · Intervention A ­ In 2005, a one-off donation of medicines (purchased on PPTE funds) were donated to health centers for distribution to the indigents. A total of 2,700 persons were said to have benefited in the intervention area. The list of only 297 indigents could be found in the 11 CSPS, and only 48 were found in 2007 to test their poverty status. The population had not been informed of the measure. No criteria were set. The action was not sustainable. · Intervention B ­ The study was carried out in 10 primary health centers (CSPS) in Burkina Faso. Village selection committees (VSC) made lists of those worst-off, which were then validated by village chiefs, mayors, and health committees (COGES). The 124 VSCs selected 566 persons. The 10 COGESs retained 269 persons (48 percent), i.e., 2.81 per 1,000 inhabitants. Except for one CSPS, the annual profits from the user fee schemes could support on average six times more indigents than the mean number selected by the VSCs. · Intervention C ­ A list of 20 criteria was developed with the Social Workers. Health staff ran the criteria only for 0.007 percent of the patients (72 patients in 11 CSPS over 18 months). Only 33 patients were then given the status of indigent. Table B11.1: The three interventions of the University of Montreal research-action Intervention A Intervention B Intervention C Districts Ouargaye Ouargaye, Dori, Sebba Ouargaye Fully subsidized Medicines at the CSPS All services at the CSPS All services at the CSPS services and district hospital and district hospital Beneficiary selection Health staff with no Community with no Health staff with 20 imposed criteria imposed criteria, but criteria consensus definition Beneficiary information None, at the point of Distribution of indigent None, at the point of service cards with Health Staff service and MoH Financing Ministry grant Endogenous Endogenous Source: Ridde et al., 2009. In the rural African context, villagers are capable of selecting those who should be exempted from user fees according to their own perspective. Thanks to the BI, health centers have a certain financial capacity to take care of the indigent. In a community-based targeting approach using endogenous resources generated from BI profits, local perceptions of the health centers' financial viability, coupled with the hierarchical social context, led to a very restrictive selection of candidates for exemption. Sources: Ridde et al., 2009; Authors. 77 Box 12: Using Community Wealth Ranking to Identify the Poorest Quintile Eligible to Subsidized Contribution to the Community-Based Health Insurance In line with the PRSP's objective to improve access to health, the Nouna research center on health (CRSN), in collaboration with the University of Heidelberg, launched in 2004 a Community-Based Health Insurance (Assurance Maladie à Base Communautaire or AMBC) project in the district of Nouna. In 2006, only 5.1 percent of the households had subscribed to the AMBC. Lack of financial resources appeared as a major explanatory factor of this low enrolment rate. In 2007, a donor accepted to subsidize the contribution to the AMBC for the poorest quintile of the district over two years (CFAF 750 instead of CFAF 1,500 for an adult, and CFAF 250 instead of CFAF 500 for a child). A community wealth ranking (CWR) approach was designed to identify the poorest members of the population based on local perceptions and relative levels of wealth within a given community. Process ­ The CWR exercise took place in the 41 villages where the AMBC had been launched, and the 7 sectors of the town of Nouna (1 rural and 6 semi-urban). Focus group discussions were organized in each village or sub-sector. They aimed at: i) understanding the local perceptions and criteria of poverty; ii) defining the number and label of local wealth groups; and iii) choosing three key representatives. These representatives were then responsible for assigning a wealth group to each household (defined as those who share and live on the same grain stock), and then ranking them within each group. This process was first to be done individually by each of the three representatives, and then discussed among them until they reach consensus. Results ­ It took a month and a week to run the exercise in the 41 villages, and two months to run it in Nouna. Experience showed that more time and diversified tools should have been used to inform households about the initiative in advance. In regards to the definition of poverty, households mainly described it as: a privation of capacities; a lack of basic needs; indecent living conditions; and a lack of social capital (social network that can be called upon in time of hardship). On average, communities came up with three wealth groups: very poor, middle and rich. Yet, the concept of wealth group proved difficult to apprehend by many participants. Also, people could not come up with quantified poverty criteria, and "the rich in a given village may be considered as poor if the reference changes i.e. in another village". Using the CWR approach proved the easiest in villages with 50 households or less, but might not be relevant given the low number of households to rank (observations tended to match). A number of households between 50 and 150 is more appropriate to test the approach, and proved manageable. Villages with over 150 households were multiethnic, and interethnic links appeared too weak to allow one person from one ethnic group to rank a household from another ethnic group. The process also proved too time-consuming for the representatives and members of the focus group. In semi-urban areas, community mobilization proved challenging (no respect of set meetings, lack of punctuality, poor attendance), and many households could not be ranked because they were unknown from people participating in the focus group. In both rural and semi-urban areas, some of the resource persons chosen by the community actually did not know much about the households' socioeconomic status. Reaching a consensus among the three key persons also proved challenging. Next steps ­ Overall, the process was evaluated as being rapid (2 hours in a 100-household village), democratic (with some caution required), and cost-efficient. It is an alternative method which is not based on monetary income and expenditures (hard to estimate in rural areas), but based on poverty criteria defined by those who live in poverty. To validate the classification, the CRSN calculated the kappa coefficient for each village. The next challenge will be to correlate these results with household survey data. Source: Savadogo and Souares, 2009. 144. The question of user fee abolition has been receiving increasing attention. A report published in June 2009 entitled "The Global Campaign for the Health Millennium Development Goals 2009 ­ Leading by Example ­ Protecting the Most Vulnerable during the Economic Crisis" was endorsed by a number of countries, UN agencies, and institutions. It stated: "Evidence is now mounting for the efficacy of a package of free quality services at the point of care to overcome the inequity that fee-for service inevitably breeds. This is one effective, evidence-based and equitable way to expand access to services to a greater proportion of the population." [Cited in AI 2009]. In Burkina Faso, a couple of pilot projects provide free health services to pregnant and lactating women and 78 children under 5. Implemented by the NGOs HELP (in two districts in the Sahel since September 2007) and Terre des Hommes (in two districts in the North since October 2008) in collaboration with the COGES, these projects reported very positive results (Figure 12). In the intervention districts, the use of health services by children under 5 more than doubled and assisted deliveries by qualified staff increased by 50 percent, with comparisons made 10 months before and 10 months after the project. After 10 months of intervention, the cost of each consultation was estimated at CFAF 1,200. These initiatives echo the international experience which suggests that abolishing user fees is not only fair and equitable but also potentially feasible both technically and financially (Box 13). Figure 12: Average Number of Monthly Consultations of Children Under 5 by CPSP in the Four HELP/TdH Intervention Districts and Four Control Districts Source: Ridde and Queuille, 2009. Box 13: The Debate over User Fees Globally, user fees have increasingly come into question. Research in Mali, Burkina Faso, and elsewhere showed that user fees reduced the access of vulnerable populations to health services, leading to a reduction in service use, particularly among women and the poorest groups [James et al. 2006 cited in Ridde and Haddad 2009]. In its 2008 annual report, the World Health Organization (WHO) urged countries to "resist the temptation to rely on user fees" [2009:26]. User fees were successfully abolished in Uganda, Ghana, South Africa, Madagascar, and Niger. In Uganda, abolishing fees doubled the number of people going to clinics, and more than doubled immunization rates for children. James et al. [2005] estimated that more than 230,000 children's lives could be saved each year if fees were abolished in 20 African countries. An increasing number of donors support governments willing to abolish user fees for basic health services [DFID 2006; Ridde and Haddad 2009]. Now, while it is clear user fees must be abolished, how to accomplish this is not really known. A recent study on the effect of removing direct payment for health care on utilization and health outcomes in Ghanaian children [Ansah et al. 2009 cited in Ridde and Haddad 2009] showed that pre-payment schemes are not pro-poor, because the worst-off are rarely enroled. And Ridde and Haddad [2009] stress that local health insurance systems have shown limits in Africa, where the penetration rate, after more than 15 years of promotion by their organizations, remains very low (5 percent). They also point out the considerable gap between "the enthusiasm generated by pre-payment schemes and the scientific evidence to support their use." Still, as long as there is no evidence that health insurance schemes are ineffective, they call for the protection of families against catastrophic health care costs and the removal of financial barriers to health care as a health system priority. Recent and ongoing studies of the abolition of user fees may shed light on the debate. Studying the effects 79 of operating the Bamako Initiative in one district of Burkina Faso, Ridde [2003] found that the study district was in a position to bear the financial cost of taking care of the poor and that the community was able to identify such people. He thus recommended the state introduce incentives so that the communities agree to enforce a more equitable health system. A three-year follow-up study launched in 2008 will test the feasibility and efficiency of different models aimed at ensuring health care to the poorest (community-based targeting with and without a set percentage of beneficiaries, endogenous/exogenous financing, etc.). This study is to lead to the drafting of a practical guide for the implementation of fee exemptions in other parts of the country. And in June 2009, a three-year research program (funded by the CRDI and the AFD) was launched to document policies on the abolition of health user fees in Mali, Niger, and Burkina Faso following an interdisciplinary approach (anthropology, political sciences, epidemiology, and public health). Source: Authors. 145. Further research is needed to inform the following three key questions: (i) if the government wants to abolish user fees, how should it be done? (ii) how can partners support the government? and (iii) how can the abolition of user fees be an entry point to service quality improvement? There are possible synergies between abolition of user fees and output-based financing. The SONU subsidy mechanisms, if improved, could be a first step toward performance-based financing (PBF). Already the PADS49 is working on building its capacity to adopt results-based financing. Yet, abolishing user fees is not simple, and the technical dimension of any reform should not be underestimated. The public health perspective should not be lost (e.g., the need for a referral system, transport means, and opportunity costs) to ensure an actual impact on MDG 5. One needs to keep in mind that user fee abolition does not remove all of the barriers on the demand side, and other actions must be taken. The real effect of such a reform could be on the supply side (if efficient output-based payment mechanisms are promoted). The incentives on the supply side deserve serious attention [Meessen 2009], and the experiences in Rwanda, Uganda, Burundi, and Niger have much to offer. 146. The debate surrounding fee abolition or expansion of fee waivers posits a reconsideration of the health sector strategy and financing and its social protection measures as a whole. The initiatives described above are implemented in parallel of other social protection measures aimed at ensuring access to health services to all. Firstly, there are the universal subsidies of specific health services, such as the SONU subsidies, coordinated by the Ministry of Health. Mechanisms for communication and reliable financial compensations are yet to be developed for a better efficiency of these measures. Secondly, the project to establish a national health insurance system aims at ensuring access to health services to a large part of the population by 2015 ­ by subsidizing contributions for the poorest ­ and an inter-ministerial committee is responsible for exploring ways to establish such a system. A first step, particularly important, will be, in the context of the national health strategy, the definition of the minimum package that will be covered by the insurance system. How to finance the system (particularly if it is expected to eventually cover the whole population), and how to identify persons entitled to an exemption on their contribution are difficult questions yet to be studied. Thus, when 49 The PADS (Program d'Appui au Développement Sanitaire) is a pooled funding mechanism for support to the national health development plan. The PADS is defined as "a joint initiative of the Ministry of Health and certain partners in the development of the health sector, for a better coordination and a rational management of the resources mobilized for the implementation of the PNDS." In September 2009, the donors supporting this pooled funding mechanism were the Netherlands (acting as lead donor in coordinating this funding mechanism), Sweden, France, the World Bank, UNFPA, UNICEF, the Global Alliance for Vaccines and Immunisation (GAVI), and the Global Fund to Fight AIDS, Tuberculosis and Malaria. This "basket funding" is managed by the Ministry of Health. 80 revisiting the health system financing and its social protection measures, the following critical points need to be considered: (i) Which model should be promoted, and what role should be given to each of the different approaches currently being tested or envisioned: health subsidies (MoH), health insurance (MTSS), fee waivers for the indigent (MASSN), fee waivers for vulnerable groups (NGOs), cash transfers (CNLS), etc.? (ii) How to finance the health sector: user fees­currently, Burkinabe households are the main funding source, contributing for 40 percent, national expenditure reallocation, new funding sources (tax on mobile phones), international funding, etc? (iii) What implementing mechanisms: identification of the indigent, financial compensation of health centers, etc.? 81 CHAPTER V: RECOMMENDATIONS FOR A MORE EFFICIENT SOCIAL SAFETY NET SYSTEM The analysis presented in previous chapters leads to the conclusion that the existing social safety net system is inadequate and particularly insufficient to respond to poverty and vulnerability. Coverage of existing programs is limited, and interventions are fairly small in scale and are designed mainly as temporary programs. Yet Burkina is committed to building its social protection system. There is the need for translating existing policy documents, based on a few existing initiatives, into a broad-based social safety net system as part of the social protection strategy. The analysis of public expenditures suggests that, in the short run, expenditures should be kept at least at their current levels (around 1 percent of GDP in 2009) and the gradual expansion of selected effective and complementary social safety net programs--nutrition programs, public works, and cash transfers--would constitute an appropriate policy choice. To ensure financial sustainability of social safety net spending, the primary sources of financing are: expenditure reallocation, improved efficiency gains in and decisively more pro-poor spending on education and health, and reallocation in discretionary expenditures. The report thus suggests priority actions toward the development of a more efficient and cost- effective social safety net system: (i) strengthening the strategic, institutional, and financial framework for designing, implementing, managing, monitoring, and evaluating safety net programs to create the culture of informed policy choices among decision-makers; and (ii) developing a plan for improving the effectiveness of the safety net system by reforming existing programs and, based on recent experiences, designing new ones. A. Summary of Findings Poverty and Vulnerability 147. Despite the existence of alternative poverty estimates, which make tracking poverty trends in Burkina Faso difficult, the poverty incidence has declined as a result of sustained economic growth over the last decade. Based on the available survey data and national accounts projections, the proportion of the population living below the poverty line fell from 46.4 percent in 2003 to 43.2 percent in 2009. In terms of assets, poverty also declined from 46.3 percent in 2003 to 40.4 percent in 2005 and to 38.5 percent in 2007. However, during the same period the share of chronic poor increased as a result of shocks: their share dropped from 25.4 percent in 2003 to 18.8 percent in 2007 after peaking at 31.1 percent in 2005 following the weather-related shocks. There are, in fact, reasons to believe that recent deceleration in economic growth and heightened fuel and food prices may have resulted in a drop in real income and reversed the decreasing trend of poverty reduction. The ongoing Household Consumption Survey (2009-2010) will provide more accurate and robust information on the trend of poverty in Burkina. 148. Burkina has also made important progress on the access to basic services by households, but significant challenges remain and the rapid population growth is likely to prevent the country from reaching its MDGs. Available indicators show improvement in access to drinking water (from 54 percent in 2004 to 79 percent in 2007) and health services have become more accessible on average due to a regular increase in public health expenditures as a share of the overall budget (from 7.4 percent in 2004 to 9.9 percent in 2006). However, the UNDP HDI reveals: (i) the literacy rate is still low (28.7 82 percent in 2007) and the quality of education is low; (ii) gender gaps remain in school enrolment, literacy, and primary completion rates and in access to lower secondary; (iii) fertility has remained very high at 6.2 children per woman on average; and (iv) the rapid population increase (3.1 percent per annum) has major negative consequences on growth, living standards, and poverty reduction. 149. Based on the 2003 priority survey, a typical poor household in Burkina Faso has seven or more members, is headed by a male who is polygamous, illiterate, or has a lower primary education level, and is a farmer living in the rural areas of the Northeast Region. Similar to other countries, poverty remains a predominantly rural phenomenon, where over 92 percent of the population lives: the poverty incidence in rural areas is more than two times higher than in urban areas (52.3 percent versus 19.9 percent). In terms of regional poverty, the Center Region is among the poorest with a poverty rate well above the national average. This region houses over half of the poor in Burkina Faso and also has the highest poverty gap. The Northeast and the South Regions are second, with poverty rates slightly below the national average; whereas the Capital area has the lowest poverty incidence (about half the national average). The socio-economic and educational level of the household head are key determinants of poverty. Male-headed households, whose total share in the incidence of poverty represents 95.6 percent, tend to be poorer than households headed by females both in terms of incidence and gap. Moreover, households whose head is illiterate or educated on a low primary level also have a notably higher poverty incidence (51.0 and 41.1 percent respectively). 150. Based on available data, household vulnerability to risk depends on various factors like the existing health and nutritional status of individuals, gender status, physical assets, infrastructure, location, educational levels, and cultural and behavioral practices. The poor are more vulnerable than other population groups because they are typically more exposed to risk, face many risks simultaneously, and have fewer assets and resources to provide a means of coping. 151. Gender is also a key poverty correlate. As in many parts of Sahelian West Africa, there are reasons to believe that some groups of women may be particularly poor and vulnerable. Unfortunately household consumption surveys do not collect individual- level data detailing the intra-household allocation of resources, consumption, work, and time allocation. However, as in many countries, female-headed households in Burkina Faso are found to be extremely heterogeneous. Although as mentioned earlier, the overall incidence of poverty is lower for female-headed households, an analysis using the 2003 priority survey confirms that after controlling for various characteristics that affect living standards, there are few differences between female- and male-headed households. The 2003 Demographic and Health Surveys (DHS) reveals that widowed and currently married but previously widowed or divorced women may be particularly vulnerable groups in Burkina Faso. Controlling for age, widows and women married more than once are an especially vulnerable group compared to women who are once married, or currently divorced. The data also indicate that the children of currently and previously widowed women are less likely to be in school than the children of women of other marital status. 152. Because over 40 percent of the population in Burkina is poor, the financial cost of closing the poverty gap will be substantial (ranging between 2.5 and 2.7 percent of GDP) and unrealistic in calling for a sensible safety net strategy that targets the poorest and most vulnerable population. Viewed in relation to existing budgetary numbers, these estimated costs would amount to significant financial commitments by the government. For example, the minimum cost of fully closing the poverty gap in 2007 (estimated at 2.5 percent of GDP) is roughly equal to less than 10 percent of the total 83 budget in 2007. This amount is also equivalent to less than half of the government wage bill in 2007 (6 percent of GDP) or to a little less than total spending on health (2.3 percent of GDP) for the same year. Government Strategy, Institutional Setups, and Expenditures for Social Safety Nets 153. Burkina Faso has not yet adopted a consolidated national social protection policy, incorporating a social safety net framework. In 2000, Burkina Faso was the first country in the Sub-Saharan Region to adopt and implement a Poverty Reduction Strategy Paper (PRSP), with a strong focus on supply-side interventions (education and health). Nevertheless, social protection including social safety nets was not explicitly considered as a key element of this medium-term strategy. In 2006, the Ministry of Labor and Social Security (Ministère du Travail et de la Sécurité Sociale or MTSS) led an inter-ministerial process to develop a social protection national policy. Although this document was finalized in January 2007, it was never adopted and the government requested each ministry to develop its own ministerial strategy, instead of focusing on a multisectoral approach. 154. Despite the lack of a national strategy, social safety net programs are playing an important role in Burkina's human capital development and crisis response strategies. Social safety nets are identified as necessary social and economic investments in the country's long-term vision of society. Social safety net programs appear in many sector-based strategies: health, education, food security, and employment. Yet, their potential to reduce poverty and vulnerability through income redistribution, promotion, and transformation is less often recognized in medium-term sectoral strategies. Moreover, as a result of the lack of a comprehensive strategy and the lack of an appropriate institutional setups, inter-ministerial coordination mechanisms on social protection and social safety nets are weak. 155. The Government is taking several actions to address the need for more comprehensive social protection, and a more effective social safety net system in particular, that can better respond to crises. The PRSP that is currently under revision will be replaced in late 2010 by the Strategy for Accelerated Growth and Sustainable Development (Stratégie de Croissance Accélérée et de Développement Durable or SCADD). The extension of social safety nets ­ along with improved access to basic services - is a priority action explicitly identified under the second pillar of the upcoming SCADD. In February 2010, a Joint Ministerial Committee on Social Protection was established by decree and in July the first meeting of the committee took place. In addition, in April 2010, the Ministry of Economy and Finance (MEF) organized a national technical forum on social protection with the explicit goal to outline a national and consolidated vision of social protection, including the social safety net component, and to directly inform the drafting of SCADD. Presently, national and international stakeholders are acting concertedly to build national consensus among ministries and agencies while mobilizing efforts to promote an integrated social protection strategy incorporating an efficient safety net system. 156. Excluding the cost of general subsidies for food and fuel, total spending on social safety net programs over 2005-2009 averaged about 0.6 percent of GDP. Currently, the social safety net programs in Burkina can be classified in five categories: (i) cash and near-cash transfers; (ii) food transfers (subsidized food sales, targeted food distributions, nutrition programs, and school feeding); (iii) universal subsidies (food and 84 fuel); (iv) public works; and (v) fee waivers. A recent global study concluded that safety net programs in developing countries typically represent about 1-2 percent or less of GDP [Grosh et al. 2008]. While Burkina appears on the lower range of the spectrum, spending on SSNs almost tripled in constant 2005 FCFA between 2005 and 2010: soaring from 0.3 percent of GDP in 2005 to 0.9 percent of GDP in 2009 (excluding general subsidies on food and fuel). This spending on SSNs corresponded to more than one-third of the illustrative minimum amount needed to bring all the poor to the poverty line through cash transfers in 2007. Spending per poor person (excluding the universal food and fuel subsidy) was at about 8 percent of the poverty line in 2007. 157. Significant efforts made in food transfer programs ­ specifically subsidized food sales, nutrition programs, and targeted food distribution ­ are attributed to the continuous increase of government and donor spending on social safety nets. Since 2006, the financing for SSN programs has become increasingly donor-dependent. Excluding universal subsidies, the external financing share in total SSN financing increased from about 57 percent in 2005 to 69 percent in 2008 and 76 percent in 2009. While excluding fuel and food subsidies, government spending increased more slowly, from 0.1 percent of GDP in 2005 to 0.2 percent of GDP in 2009. There are important priority differences between the government and donors: Donors focus on nutrition, while the government concentrates financing mainly on universal fuel subsidies. 158. The coverage of existing social safety net programs remains limited compared to the needs. Theoretically, the estimate of the total number of beneficiaries of social safety net programs exceeded 3.9 million individuals in 2009 (or about 25 percent of the total population). In reality, the coverage for most programs is temporary (reactions to shocks); and the level of benefits in some programs is very small and insufficient to help the poor smooth their consumption. This result is primarily due to the limited financial allocation and lack of implementation capacities ­ in particular, difficulties in identifying and reaching the poorest. 159. In terms of measuring the actual impact of social safety net programs, investments have been scarce. Except for a few recent donor-financed initiatives (cash transfers, food vouchers, and school feeding) for which in-depth evaluations were prepared or are under preparation, there is no sufficient data and evidence, particularly on programs financed by the government, to inform policy-makers on the characteristics of beneficiaries, costs of programs, mechanisms for targeting, and actual outcomes in terms of helping the poor and vulnerable. A monitoring system to capture programmatic cost- efficiencies and to promote evidence-based data on the impact of social protection policies would facilitate informed policy choices. Existing Safety Nets Programs 160. The review of existing social safety net programs confirms that despite substantial needs, few programs assist the chronic poor and provide regular and predictable transfers. Existing interventions are often implemented through projects and during periods of shocks, focused on one geographic area; no systematic approach exists to assist poor and vulnerable households. The main challenges are: the lack of a clear definition of target groups; choice of adequate instruments; and establishment of solid monitoring and evaluation systems to inform policy decisions. Given the recent innovative initiatives, such as food vouchers, cash transfers, and health fee waivers, lessons can be learned to improve the efficiency of the existing social safety net programs and potentially scale them up. 85 Cash and Near-Cash Transfers 161. Valuable lessons can be learned from recent cash transfer programs introduced in Burkina Faso: Since 2008 three cash transfer pilot programs have set up using external funding: a pilot cash transfer program under CNLS-IST (Conseil National de Lutte contre le Sida et les Infections Sexuellement Transmissibles) and food voucher programs under Catholic Relief Services (CRS) and World Food Program (WFP). The CNLS-IST research action, introduced in 2008, is testing the role of gender and conditionality in the impact of cash transfers targeted at orphans and other vulnerable children (OVC) on health, education, and well-being in the Nahouri Province. The results of the impact evaluation of this program on 3,900 households are expected in late 2010. Urban food vouchers were introduced in 2008 (CRS) and 2009 (WFP) in Ouagadougou and Bobo-Dioulasso as an emergency measure to respond to increased food insecurity due to high food prices. The mid-term evaluation of the WFP program, benefiting over 30,000 households, has shown an efficient implementing process and positive results, despite challenges of targeting in urban areas. The evaluation has also shown that the transfers, and the resulting savings on food expenditures generated by the program, improved beneficiary households food consumption both in quantity and quality, and to a smaller extent their access to health and education. Food Transfers 162. Food transfers are the main form of social safety net programs in Burkina Faso, accounting for 87 percent of total SSN spending over the period 2005-2009, and over 80 percent of all estimated SSN beneficiaries in 2009 (excluding fuel subsidies). Four types of food transfers are currently in place: (i) targeted subsidized food sales, (ii) targeted free food distributions, (iii) nutrition programs, and (iv) school feeding programs. Targeted subsidized food sales: In principle, the objectives of the subsidized food sales program are to: (i) contribute to cereal price stability in deficit provinces; and (ii) assist households vulnerable to food insecurity. However, the poorest may not have the financial resources to access the proposed subsidized cereals. Moreover, the actual number, profile, and poverty level of the beneficiaries as well as the impact of the program on the food security situation are not know due to the lack of proper monitoring and evaluation. Targeted free food distributions: Free food distribution is provided through three mechanisms: (i) Conseil National de Secours d'Urgence et de Réhabilitation (CONASUR) distribution of free food; (ii) World Food Program (WFP) Food-for- Education/Training program; and (iii) Catholic Relief Service (CRS) general relief program. Because the country has been producing surpluses for the past 10 years, the National Security Stock (SNS) has never been used for its intended objective. Nevertheless, since some areas remain structurally in deficit (mainly in the North and Kadiogo Provinces), about 22,354 MT of cereals (for three-month consumption) was distributed in these regions in 2008-2009: donors and partner NGOs covered about 40 percent of the needs and the state covered the remaining 60 percent. Overall, in 2008, the national system has been providing food assistance to about 44,000 people on an annual and ad-hoc basis (particularly individuals affected by flooding or small-scale disasters). The WFP food transfers conditional on attending literacy or training 86 courses, has gradually refocused its efforts to the Sahel Provinces, and has scaled down its coverage from 21 provinces in 2005 to 11 provinces in 2010, covering about 36,000 beneficiaries. Finally, CRS annually assists about 14,000 vulnerable persons (people living with HIV, orphans, the elderly, and the disabled). However, because food insecurity is primarily an access problem, cash or voucher programs might be preferred. Nutrition programs: Nearly two out of five children under 5 in Burkina Faso suffer from acute malnutrition, in particular children aged 24-59 months, and over one out of three children suffer from stunting, with enormous geographical disparities. In rural areas, one out of seven girls aged 15-19 suffers malnutrition. Given the needs, the government and its partners, UNICEF and the World Bank in particular, have made substantial efforts since 2003 to increase nutrition interventions. UNICEF, WFP, ECHO, and their partner NGOs are particularly active in the treatment of moderately and severely malnourished children under 5 and pregnant and lactating mothers. The WFP also provides nutritional support to people living with HIV. The coverage of existing programs, despite significant efforts, is inadequate, even though severe acute malnutrition is one of the main challenges in reaching the nutrition MDGs. UNICEF estimated the number of severely malnourished children treated in 2009 at only 25,000, far below the needs (over 1million children malnourished). School feeding: School feeding represents one of the main social safety net programs currently in place in Burkina Faso; in 2009, it accounted on average for 24 percent of total spending on SSN (excluding general subsidies) and covered over 27 percent of the estimated total number of SSN beneficiaries. Over 50 percent of primary schools in the country have school feeding activities, and over 50 percent of spending is externally funded by USAID, WFP, CRS, and NGOs. The Ministry of Basic Education and Literacy (MEBA) manages endogenous school feeding programs and provides only a three-month supply of food, while donor programs are focused on assisted school feeding programs that provide support for the whole school year. Findings of a recent impact evaluation (2008) reported that both forms of assisted school feeding, in-school meals and take-home rations for girls, increased girls' enrolment by about 6 percent. Neither intervention showed a significant positive impact on learning outcomes, but a positive impact of take-home rations on the nutritional status of younger siblings was observed. These mixed results invite to reconsider the choice of in-school meals to increase school enrolment rates. In-school meals programs are costly and greater impacts (on poverty and nutrition) may be achieved with targeted (possibly conditional) rations with similar results on school enrolment ­ and the possibility to target girls in priority and reduce gender disparities. The government is planning to scale up its school feeding program (in-school meals) for the 2010-2011 School Year in order to cover all primary schools. It is unclear, nevertheless, whether the planned expansion of school feeding activities is a temporary measure to mitigate the effects of the recent crisis or a long-term strategy. While scaling up school feeding might be a legitimate practical temporary response to the food crisis, the authorities need to further investigate the most appropriate social safety net instrument in the long run, including targeting mechanisms to reach the poor, to increase school enrolment rate and contribute to poverty reduction. Universal Food and Fuel Subsidies 87 163. To mitigate the negative effects of high food and fuel prices, the government has provided universal subsidies that are expensive and inefficient in reaching the poor. The universal food subsidies were introduced in 2008 in response to the high world prices. They proved to be very expensive and less than efficient in reaching the poor. The two poorest quintiles were hit the hardest by high food prices yet benefited very little from the tax exemptions; only 10 percent of the benefits went to the poorest quintile. Within six months the measure was rescinded, and the program terminated. A number of fuel products, on the other hand, have been subsidized by the government for years, despite a high fiscal cost and very limited impact on the poorest households (except lamp oil). It is estimated that over 84 percent of the benefits go to the non-poor. Overall, the direct fiscal cost of exoneration on fuel import taxes amounted to an estimated 0.7 percent of GDP in 2007 and 2008. This is a high cost for a very limited impact on the poor and vulnerable: Labor Intensive Public Works and Cash-/Food-for-Work 164. Much is to be learned from the positive experience of the ongoing public works programs. The PrEst (Program Pistes Rurales ­ Désenclavement à l'Est), implemented by the NGO Helvetas since 2002 under the supervision of the Ministry of Infrastructure (MID), uses a labor intensive approach for infrastructure creation in rural areas. Although the project is not designed as a safety net to specifically reach the poorest able-bodied persons, the results are showing positive impacts on poverty reduction and local development (e.g., increased use of health services, increased school enrolment, and reduction of youth migration). Between 2005 and 2007, the project, which uses remuneration below the guaranteed minimum wage, employed on average 865 persons (mainly youth aged 20-35, of which 16 percent were women) a year for about six months. Other experiences of public works include the WFP's Food-For-Assets program, which focuses mainly on quality assets building and to a lesser extent on the labor-intense activities. Building on PrEst and WFP's Food-for-Assets experiences, other types of public works programs could be introduced as effective social safety net instruments. Fee Waivers for Health 165. Although there are several initiatives intended to provide free health care to the poor and vulnerable, they are not operational for lack of implementation mechanisms. For example, the SONU (emergency obstetric and neonatal care) subsidy policy of 2006 entitled indigent (poor) women to free health services, yet it failed to specify the parameters of qualification and, thus, prevented these women from receiving services. There are three main issues with this program: (i) the definition of who is indigent (or poor)? (ii) who decides upon this status? and (iii) who absorbs the cost of the fee waivers? A few interesting initiatives, nevertheless, were conducted in rural areas by the University of Montreal (in collaboration with the IRSS/CNRST) and by the Nouna Health Research Center , and in urban areas, by the WFP and the IRD. They used a community-based process to select the beneficiaries of user fee exemptions. Although international experience suggests that abolishing user fees for the poorest is equitable and feasible both technically and financially, further work is needed to define the implementation mechanisms. Particularly in the context of health system financing in Burkina Faso, the following critical points need to be considered: (i) which model should be promoted: health subsidies (Ministry of Health or MOH), health insurance (Ministry of Labor and Social Security or MTSS), fee waivers for the indigent (poor) (Ministry of Social Action and National Solidarity or MASSN), fee waivers for vulnerable groups (NGOs), cash transfers (Conseil National de Lutte contre le Sida or CNLS), etc.? (ii) how should the health sector be financed: user fees, national expenditure reallocation, new 88 funding sources (e.g., tax on mobile phones), international funding, etc.? and (iii) what implementing mechanisms should be used to identify the poor and provide financial compensation to health centers? B. Policy Recommendations 1. Strengthening the strategic, institutional, and financial framework for designing, implementing, managing, monitoring, and evaluating safety net programs (a) Adopt a National Social Protection Strategy including Social Safety Nets 166. Develop a comprehensive social protection strategy that gives priority to an efficient social safety net component (non-contributory transfers). Where pervasive poverty touches all corners of a country, it is crucial that a comprehensive social protection strategy be developed, one that focuses on both contributory and non-contributory schemes. The Social Protection National Policy drafted in 2007 represents a good starting point, and should be updated considering the current challenges faced by the country. Given the huge needs in health and education, social protection instruments should directly contribute to human capital development. This invites to give particular attention to the needs of children, who represent nearly half of the population50 and constitute a priority target group for human capital investment (Box 14). The acute demographic challenge51 also needs to be taken into consideration and, for instance, suggests investing more in girls' education. 50 46.6 percent of the population is 15 years old or younger [INSD 2006]. 51 The 2006 population and housing general census revealed a demographic growth rate of 3.1 percent versus 2.4 percent over the previous period, which adds to the challenge of meeting needs in basic services. 89 Box 14: Principles of Child-Sensitive Social Protection The following principles should be considered in the design, implementation, and evaluation of child- sensitive social protection programs: Avoid adverse impacts on children, and reduce or mitigate social and economic risks that directly affect children's lives. Intervene as early as possible where children are at risk, in order to prevent irreversible impairment or harm. Consider the age- and gender-specific risks and vulnerabilities of children throughout the life-cycle. Mitigate the effects of shocks, exclusion, and poverty on families, recognizing that families raising children need support to ensure equal opportunity. Make special provision to reach children who are particularly vulnerable and excluded, including children without parental care and those who are marginalized within their families or communities due to their gender, disability, ethnicity, HIV and AIDS, or other factors. Consider the mechanisms and intra-household dynamics that may affect how children are reached, with particular attention paid to the balance of power between men and women within the household and the broader community. Include the voices and opinions of children, their caregivers, and youth in the understanding and design of social protection systems and programs. Source: DFID UK et al., 2009. 167. Promote synergies and economies of scale. In a context of limited financial resources, synergies and economies of scale need to be promoted between the different social protection instruments and other social policies like education, health, and employment, through joint targeting approaches and safety net programs directly supporting the demand for education and health. This will need to be clearly articulated in a cross-ministerial social protection action plan. 168. Clarify the objectives of the social safety net system. Within a broader social protection strategy, the objectives of the social safety net component are to: (i) directly support the consumption of the chronically poor and vulnerable populations; (ii) ensure access to basic social services to poor and vulnerable populations, in order to promote human investment; and (iii) provide temporary support to poor and vulnerable populations affected by shocks. Therefore, in Burkina Faso, the priority principles of the social safety net system should be to: (i) ensure that the chronically extreme poor and vulnerable populations receive regular and predictable support along with complementary programs to escape poverty traps and break the intergenerational transmission of poverty; (ii) provide temporary income to vulnerable groups in case of shocks; (iii) pay particular attention to the needs of vulnerable children (e.g., nutrition, education, and conditional cash transfers) and the needs of poor and vulnerable women (i.e., minimize potential negative impacts, optimize positive impacts on women and gender equity); and (iv) facilitate access of the poor and the vulnerable to basic social services. 169. Develop a shared vision. A shared national vision of a good social safety net system is essential to work toward an efficient social safety net system that is, a system woven of several programs, complementing each other as well as complementing other public social policies, and meeting the following criteria [Grosh et al. 2008]: Appropriate: customized to best fit with the circumstances; 90 Adequate: covering the various groups in need of assistance; Equitable: treating beneficiaries in a fair and equitable way; Cost-effective: running efficiently with the minimum resources required to achieve the desired impact, but with sufficient resources to carry out all program functions well; Incentive compatible: not causing disincentives (e.g., labor participation); Sustainable: pursued in a balanced manner with other aspects of government expenditure, and both financially and politically sustainable; and Dynamic: evolving over time. 170. Agree on priority actions. The different strategic documents tend to refer to the whole spectrum of people in need, without clearly setting priorities. Given financial constraints, it is necessary to agree on shared priorities if any significant impact is to be seen. As mentioned above, it is recommended to prioritize investment in human capital. This, along with the concern to invest in the most cost-effective activities, suggests prioritizing actions that support early childhood development (Figure 13).52 Figure 13: Rates of Return of Human Capital Investment Initially Setting Investment To Be Equal across All Ages Source: Carneiro and Heickman, 2003. 171. Clarify links with other policies. Social safety net programs are meant to act in conjunction with other poverty reduction programs, notably, labor market programs, pensions, health insurance, policies to ensure macroeconomic stability, rural development, and human capital formation. Social safety nets are typically used to complement supply- side interventions and fill in where other policies cannot deliver sufficient results in the short run. In Burkina Faso, they could prove particularly useful to ensure education and health spending become pro-poor. Complementarities and synergies with other social policies ­ food security, education, health, employment, health insurance, population, etc. 52 "Ceteris paribus the rate of return to a dollar of investment made while a person is young is higher than the rate of return to the same dollar made at a later age. Early investments are harvested over a longer horizon than those made later in the life cycle. In addition, because early investments raise the productivity (lower the costs) of later investments, human capital is synergistic." [Carneiro and Heickman, 2003:7]. 91 ­ need to be made clear to support coordinated efforts. 172. Agree on a common language. Today the lack of common language for issues related to social protection is a major problem, and renders cross-ministerial dialogue and dialogue with international partners difficult. It seems crucially important for the Burkinabe government to come up with a social protection glossary common to all ministries. (b) Reinforce the Institutional Framework for Social Protection and Social Safety Nets 173. Ensure that the permanent inter-ministerial committee for social protection is operational. Social protection consists of cross-ministerial issues and its supervision should not be limited to one ministry. Given the mulit-sectoral aspect of the social protection, the government set up in February 2010 by decree a permanent Social Protection Committee that will be responsible for revisiting the social protection strategy, supervising/coordinating the various initiatives ­ studies, pilot projects, etc. ­ that have been launched or are expected to be launched, and ensuring the required dynamic cross- sectoral dialogue among various sectoral ministries. It is also expected that two sub- committees will be put in place, each responsible for (i) social insurance programs (health insurance and pension system) and (ii) social safety net programs. This is an important step that requires further strengthening of inter-, intra- and extra-ministerial coordination mechanisms. The reasons of the limited impact of past social protection initiatives in Burkina Faso underline the priority to ensure sectoral consistency between the various departments involved, as well as the need to promote policy impact evaluations and evidence-informed policy decisions. 174. Clarify the role and responsibilities of the different national institutions engaged in social safety nets. Once the Strategic Framework for Social Protection is defined and priority actions for social safety nets are clarified, the respective functions of the main structures created to provide social assistance ­ e.g., MASSN, FSN, and CONASUR at the central level ­ will need to be revisited to avoid gaps and duplication, and come up with a more cost-efficient setup. 175. Define the appropriate implementation arrangements for new social safety net programs. Defining who, at an institutional level, will manage the design, implementation, and ongoing operation of a social transfer program is a crucial first step upon adoption of a social safety net program. The institution that gathers the following characteristics will be the best one to manage the program [Samson et al. 2006]: (i) a durable political commitment to social protection; (ii) the political influence to secure resources and defend the program's priority; and (iii) the institutional capacity to deliver an efficient social safety net system. That being said, it is often impossible to find all three qualities in one single institution. The choice of the managing institution often goes to: the relevant social development ministry (the most committed one); the ministry responsible for finance (the most powerful one); or a separate agency that reports to a committee of related ministries (bringing together commitment, influence, and capacity). A reassignment of responsibilities over time is also possible as observed in South Africa (from provincial governments to a national social security agency), Bangladesh (from the Ministry of Social Welfare to the Ministry of Women and Children Affairs), Namibia (from the Ministry of Labor to the Ministry of Health and Social Services). In India, a share of responsibility is in place where the Ministry of Labor supervises pensions and the 92 National Family Benefit Scheme administers the grants. Each of these models presents advantages and disadvantages and institutional arrangements need to be informed by a review of relevant institutions, the primary objective of the program (e.g., poverty reduction versus education), and any longer-term vision for social protection in Burkina Faso. 176. Ensure a separation of duties. Institutionalized program does not mean that all duties will be performed by one single national institution or by national institutions only. The key to a successful design and implementation is to delegate the responsibility of each duty to the formal or non-formal institution for which it is the core activity, and to establish strong control mechanisms. For instance, cash transfer delivery is the core business of banks, and traditionally it is civil society's role to ensure people's rights are respected. Such an implementation strategy (along with an appropriate design) will contribute to minimize fiduciary and corruption risks, and maximize efficiency. 177. Provide capacity-building support. The concept of social safety net as necessary social investment (regular and predictable) is largely new in Burkina Faso. Awareness efforts and training is required both at national and local levels.53 Other initiatives may be useful to increase understanding, interest, and capacities in social safety nets and social protection in general ­ for example, on-the-job training and study tours. 178. Engage more with local authorities for an effective implementation of national policies. Decentralized authorities will need to be involved in the whole policy formulation process. Their roles and responsibilities in providing social assistance will need to be clarified and strengthened. The definition of any new responsibility will have to be reflected in budget allocations (e.g., to train staff, build human and material resources, and strengthen institutional setup) and be supported by better coordination mechanisms with the various sectoral ministries. 179. Further explore the role of NGOs and the private sector in the delivery of social safety nets. The capacities of governmental and local authorities remain somewhat limited on the ground, particularly in remote areas that are most in need of assistance. NGOs and the private sector may be useful partners to intervene in these areas, and a contract-based solution needs to be further explored. The contract-based solution currently being developed under the PADS may bring useful lessons as well as capacity for an application beyond the health sector. (c) Strengthen Financial Framework 180. With a per capita income of US$480 equivalent, the surplus available to redistribute under any sustained safety net program is relatively small. The large proportion of the poor in Burkina means that: (a) any program large enough to have a substantial impact would be extremely costly; and, (b) affordable options will likely only be able to reach some fairly limited portion of the population in need, and/or to have a limited effect on household incomes. Under these conditions the challenge for Burkina Faso is to carefully consider the options for public policy, and to be highly selective in choosing interventions that are cost-effective in delivering benefits to the poorest. Central 53 In January 2010, 30 technicians from 11 ministries were introduced to the fundamentals of social protection and social safety nets in a training workshop supported by UNICEF. 93 questions to be asked are: (i) what are the realistic objectives of a publicly funded safety net ­ given the nature and characteristics of the poor; and revealed political and social preferences? (ii) which groups (or sub-groups) among the poor should benefit? and (iii) what choice of safety net programs is most cost-effective in terms of achieving the desired objectives? Therefore, to assess the feasibility of expanding social safety nets, the government needs to consider the role safety nets should play in the development strategy of the country, how much it makes sense to spend on them, and what the best choice of programs might be ­ looking at poverty and vulnerability profile, performance of existing programs, international experience, and national institutional capacity. 181. Through better targeting, efforts to streamline costs, and public expenditure reallocation, it seems reasonable to expect a gradual increase of coverage and therefore spending for social safety nets in the next two to three years. Since revenues will not be able to create substantial fiscal space, it seems more realistic to expect the increase in SSN spending from a combination of (i) reallocation between expenditures from less efficient programs, and (ii) more efficient public expenditure management. The previous analysis of the poverty profile and the public expenditure review suggest that gradually expanding some effective and complementary SSN programs, such as nutrition programs, public works, and recent experience in cash transfers will be an appropriate policy choice. Increasing SSN coverage and spending raises the issue of fiscal space and requires an in-depth analysis of expenditure allocations and effectiveness. In this context, within priority sectors, like education and health, spending needs to become more efficient and pro-poor. Moreover efficiency gains can be obtained through gains in discretionary expenditures. Based on international experience, spending on SSN in Burkina could be around 1.5 percent of GDP. A more ambitious expansion in social safety net spending is not advisable for several reasons. The first reason is the need to avoid a situation in which too great an increase in spending in this sector leads to a relative decline in spending in other sectors that are just as vital to the effort to reduce the poverty level and vulnerability of the population, such as health and education. The second reason is that it is important to see to the efficiency of expenditures already made before increasing them too rapidly. The third reason relates to implementation capacity in the field, which is limited. The efficiency of expenditures allocated to social safety nets needs to be strengthened through better targeting of beneficiaries, and a relative reduction in management costs. In addition, design and piloting of potential flagship programs, like cash transfers and public works, may change the relative priority of programs and call for a reconsideration of the expenditure mix and overall amounts. 182. Moreover, fiscal arbitrage is needed to focus on promoting or scaling up the most cost-effective social safety net programs. To bring the financing of social safety net programs onto a more sustainable basis, the following steps could be considered: Establishing a rigorous classification of social protection expenditures and a comprehensive list of public safety net programs coverage. This is essential to deducing a clear estimate of what level of expenditure is justified and what financing is needed to achieve the objectives of the system. Determining the overall envelope of the government's budget needed for the desired level of safety nets coverage and making adequate full provision each year in the budget. In this context, there is need for assessing how much it is reasonable and affordable to spend on social safety nets. 94 Establishing that the government can seek to obtain funding for safety nets on a non-project basis, for example by seeking budget support in the context of a Poverty Reduction Support Credit from IDA and similar operations from other external partners. This will mean a step-up in the management of the safety net system and an improvement in such aspects as fiduciary arrangements, procurement, and audits as well as results monitoring and evaluation. 183. Savings can be achieved through better targeting, streamlining costs, and public expenditure reallocation by reducing very small or inefficient programs, while strengthening a few viable programs with better targeting and outcomes. In this context, spending needs to become more efficient and pro-poor in general, scaling down poorly targeted subsidies, and focused on high-priority sectors like health and education. Gains in discretionary expenditures, furthermore, can produce efficiency gains. (d) Improve Program Monitoring and Evaluation 184. Promote robust and independent program process and impact evaluations and evidence-informed policy decisions. The specific objectives of an M&E system are to: inform the implementation of the program and any necessary adjustments in a timely manner; demonstrate program impact to policy-makers, development partners, and general public; and feed into the global lessons of experience. The evaluation function is particularly critical to inform an evidence-based policy development. As raised throughout this report, very little solid evidence on the actual characteristics of the beneficiaries, the costs, and the impact of existing social safety net programs has been collected. This lack of evidence impedes a greater mobilization of political and financial support to these programs. Robust monitoring and evaluation will be particularly crucial in envisioned pilot projects (e.g., on cash transfers, public works, and fee waivers). 185. Therefore a systematic monitoring of the overall set of safety net programs is needed, to judge how well resources are being used. This is a precondition for the piloting and/or scaling up of any social safety net program. In particular five systemic actions can be considered: Establish a rigorous classification of social protection expenditures and a comprehensive list of public safety net programs. Set up minimum reporting requirements for safety net programs to allow evaluation of effectiveness, costs broken down between service delivery and overheads, the sources of financing, etc. In this context, it is important to clarify monitoring and evaluation mechanisms of social safety net programs under a common framework. The monitoring and evaluation system will need to monitor several indicators: gender, health, education, poverty, nutrition, economic growth, social cohesion, etc. Working under a shared framework will ease program cost-efficiency as well as comparisons among programs. Broaden evaluation to incorporate rigorous impact measurements, emphasizing the piloting during the introduction of new interventions and/or the expansion of extension of existing interventions to new categories of beneficiaries. Involve the civil society in monitoring and evaluation. Civil society is currently left out of the process of monitoring and evaluating safety nets provided by the state and its partners. Efforts should be made to support civil society engagement in the 95 monitoring of the programs, for example, supporting engagement with the National Assembly or community budget-tracking. Transmit systematically program evaluation reports to the sectoral ministries responsible for social protection and social safety nets and maintenance of a database on programs. This will facilitate better-informed decision-making for policy-makers. Strengthen the sectoral ministries capacities for monitoring and evaluation and provide training for program managers in monitoring and evaluation techniques coupled with a mechanism for exchange of experience across programs. 2. Improve Effectiveness of the Social Safety Net System (a) Define the Appropriate Set of Social Safety Net Instruments 186. Update poverty analysis to clarify the priority target groups. Available data on poverty is outdated and might not reflect the actual conditions prevailing after the 2008 crisis induced by high food and fuel prices. Data of the ongoing EICVM54 are expected to be made available in the course of 2010. The poverty analysis will need to be updated based on this data and poverty maps could be constructed based on the 2006 census. 187. Identify appropriate instruments for each priority target group. Type, role, scale, and frequency of social safety net instruments need to be defined for each of the set priorities. Based on the available poverty analysis, it is proposed to consider the following set of instruments as a basis for discussion and further feasibility studies on the appropriate mix of programs to be implemented on a permanent basis to tackle chronic poverty: (i) Nutrition supplement programs for pregnant and lactating women and children under 5, including widows and their children, to ensure nutritional needs of these particularly vulnerable groups are covered; (ii) Targeted school feeding programs for children aged 6-14 to increase school enrolment and attendance rates for poor children, including the children of current and ex-widows; (iii) Regular cash transfers to women in households living in chronic (extreme) poverty to increase the real income and food spending of poor households; and (iv) Seasonal labor intensive public works to provide a source of income to poor men and women and to construct public infrastructure or provide community services. 188. Even when they are not explicitly targeted to women, safety net programs should ensure that they do not reinforce gender disparities and biases in society. Box 14 outlines some clear does and don'ts. 54 Enquête Intégrale sur les Conditions de Vie des Ménages. 96 Box 15: Does and Don'ts of Female-Sensitive Social Protection Exact policy details will vary a lot depending on circumstances and objectives. However, there are some clear principles that should guide thinking about social protection policy from a gender perspective. The key overall policy objective should be to help poor and vulnerable women, whether they are explicitly targeted by a policy or not. In order to meet this policy objective, policies should: Not assume a unitary model of the household and be mindful that the beneficiary of a transfer matters: transfers should go to women. Not be limited to heads of households, men, or the unemployed; Look for feasible mechanisms for targeting the poor as individuals without undue costs in reaching them. Not be biased against women ­ do not exacerbate inequities. Try to compensate for any pre-existing biases against women. Consider the form of transfer: Share of wages or transfers paid in kind is often preferred by women since cash is more easily expropriated by male household members; conditionality may often be desirable for this same reason. Take account of transaction costs: women may face higher time constraints, lower mobility. Not forget about the many social constraints faced by women. Provide childcare. Do not forget that responses to policies may differ by gender: for example, foregone incomes and incentive effects may differ. Remember that programs can have unintended consequences: transfers impact labor supply but differently by gender; they may cause re-allocations of work within the household to children. Do not assume that equality in the law is enough. Affirmative action may be needed for both efficiency (potential externalities, for example, through benefits to children from targeting women: gender of transfer recipient matters to household welfare) and equity arguments: (when women are poorer or more vulnerable). Source: van de Walle 2010. 189. Improve mechanisms to scale-up and -down programs in case of shocks. Once an appropriate permanent safety net system is set up, selected mechanisms could be considered to be scaled up to respond to crisis, complemented by other temporary instruments. The rules for scaling up programs will need to be incorporated into the national food security system (i.e., early warning system,55 contingency plans, experience in emergency responses, etc.). (b) Improve the Efficiency of Selected SSN Programs 190. Improve cost-effectiveness of programs. An in-depth critical review of each program (largely lacking in Burkina Faso) will better inform necessary adjustments. Based on the current program review presented in this report, initial recommendations are: 55 The early warning system (SAP) is fed by several sectoral information systems centralized at the DGPER. A survey on agricultural security is run every year, with 1,000 persons surveyed, to determine the vulnerability level and caseload. 97 (i) Cash transfers ­ Upcoming robust impact evaluations of the ongoing pilot programs will inform the potential expansion of cash-based programs. Further appropriateness and feasibility studies will be required to explore the use of cash transfers at a larger scale, in both urban and rural areas, to address access to food for extremely poor families. The piloting of regular and predictable cash transfers to female members of poor households through a multiannual approach (allowing for successive program adjustment and learning), reaching a large enough population (to collect solid evidences) in close collaboration with national authorities will be extremely useful to reflect on design and implementation arrangements in particular. Integration of incentives for mothers of infants to enroll in supply-side nutrition interventions into such a program can also be explored. (ii) Subsidized food sales and targeted food distributions ­ Further review is needed to assess their cost-effectiveness and evaluate the impact of these programs on beneficiaries. However, a reduction of their scope in favor of program with better targeting outcomes should be considered. (iii) Nutrition programs ­ Given the exiting needs, the coverage of the programs needs to be expanded and the strategic and institutional framework of the programs needs to be strengthened. (iv) School feeding ­ Based on existing evaluations, further work is needed to evaluate the relative cost-effectiveness of assisted school feeding programs compared to other forms of social safety nets ­ considering costs, social cohesion impact, human development, etc. ­ and to explore an improved integrated school feeding model to ensure that poor children are better targeted and there is contribution to local development through better ownership and local procurement. (v) General subsidies ­ Such measures should be adopted only as an instrument of last resort in time of crisis and particularly target commodities that are primarily consumed by the poor. (vi) Public works ­ The appropriateness and feasibility of introducing programs targeted to the rural poor, like using a low wage rate and other possible targeting criteria, need to be assessed. This type of program has proved particularly appropriate in situations of high vulnerability to support regional development and community-based structures (Annex 3), and a public works/cash-for-work program could become a productive safety net by incorporating good workfare design principles (low wage below market rate, selection of projects, etc.) and could have a direct effect on income generation during seasonal shortages of jobs and in times of shocks. (vii) Fee waivers for health ­ Abolishment of user-fees and fee-waivers for the rural poor (and in particular for children under 5, pregnant and nursing mothers, vulnerable widows, remarried widows and their children) should be carefully reviewed in the context of a broader health strategy and financing reforms, in order to establish compensation mechanisms for the effective implementation of the program. Critical issues to be addressed are: identification of beneficiaries, financing schemes, and implementation mechanisms. 191. Reinforce links between social safety nets and other social services (i.e., education and health). The following investments are important for an improved coverage, efficiency, and impact of social transfers: 98 (i) Social welfare services (for which the role of the civil society is important): Community-based social workers to assist households access their entitlements and create opportunities to connect households with other available services (e.g., income-generation activities); communication and public education on eligibility criteria and entitlements; parenting support programs; and quality community-based health and education services to enable beneficiaries to effectively invest in human capital; and (ii) Capacity-strengthening measures: For national data management system (e.g., civil registry) and for decentralized government social services to ensure supervision and coordination of the different NGOs engaged in social protection activities. 192. Expand efficient programs. The current coverage of social safety net programs remains minimal compared to needs. Once the cost-effectiveness of the different types of safety net programs will have been defined, and the impact evaluation of ongoing pilot programs (e.g., cash transfers and food voucher) and any newly introduced pilot programs (e.g., public works) are available, it will be advisable to expand efficient programs. (c) Improve and Harmonize Approaches to Targeting 193. Agree on general targeting principles. Policy-makers might want to take into consideration the following principles: Consider mixing multiple targeting methods ­ There is evidence showing that the use of multiple targeting methods ­ geographical, community-based, categorical, self-targeting, proxy-means test ­ makes the identification of the neediest more accurate and comprehensive, improving the targeting performance [Coady et al. 2004a]. Consider simplistic but practical and transparent targeting mechanisms ­ The development of a fair, transparent, scalable, and efficient poverty targeting system in Burkina Faso is crucial and particularly challenging given political and administrative difficulties. Poverty targeting aims to economize on program resources by directing transfer benefits only to the poor. The savings in social transfers must be balanced against the costs of the targeting processes ­ which include not only the direct costs to the government from administering the targeting mechanisms, but also the private costs to program participants they incur in complying with the targeting requirements, as well as a range of social, political, and other costs (Box 14). Establish appeals and grievances mechanisms ­ A program without a way to address targeting issues runs the risk of wrecking its reputation. Establishment of appeals and grievances mechanisms can ensure that programs are accessible, simple, transparent, fair, and prompt. Therefore, more transparency in program standards is needed and high standards of governance need to be set and maintained. 99 Box 16: To Target or Not To Target? ­ The Costs and Political Economy of Being Selective As the diagram below illustrates, there is a trade-off between universal coverage and narrow targeting. The question is to determine where along that curve is the `right' place to be. The level of data currently available does not allow to rigorously evaluate those trade-offs for Burkina Faso, but there are a number of considerations for policy-makers to bear in mind: · Universal programs may enjoy wider political and popular support; · Universal programs are expensive (even in the richest countries, there is intense debate regarding their affordability), and provide benefits to many people who do not need them; · Targeting may can cause perverse incentive effects; · Information is expensive; · Community targeting may reduce both the information and administrative costs of targeting, but may put (opportunity, social, etc) costs on communities; and · It may not be administratively possible to target. Source: Smith, 2001; Authors. 194. Clarify eligibility criteria for each type of program (and exit criteria). Currently many constraints inhibit effective targeting, including information about poverty, administrative capacities, political choices, and quality of governance. Despite the fact that the term "indigent" is often mentioned in policy documents in Burkina Faso, there is no common understanding of what describes an indigent person. 195. Research innovative poverty-based targeting methods. To ensure that effective targeting tools are developed to redirect the flow of resources toward the poor, further research is need. In this context, the following approach is necessary: Ensure better information is collected to facilitate targeting and assess results ­ Currently there is a lack of detailed data on the beneficiaries and costs of the programs. Moreover, poverty information is often not sufficiently disaggregated for fine targeting. Carry out targeting efficiency evaluations ­ For most programs, it is not possible to conclude whether adopted targeting methods are efficient or not. Only a few initiatives actually made significant effort to research on targeting and measure inclusion and exclusion errors (the Urban Voucher program, the University of Montreal, and the Nouna Health Research Center research works). Research innovative poverty-based targeting methods ­ It seems particularly difficult to target in an efficient manner the poorest given huge similarities in observable characteristics, making difficult the use or proxy-means test for instance. On the other hand, community-based mechanisms may be difficult to implement on a large scale [Coady et al. 2004; Save the Children UK et al. 2005]. These challenges call for further research on the efficiency of various targeting mechanisms (community-based, household surveys, categorical, self-targeting, etc.) in the specific socio-economic context of Burkina regions. Research social dynamics potentially impacting targeting efficiency ­ More research is needed to better understand intra- and inter-household redistribution mechanisms, which may be common and strongly rooted in the culture in some communities, and how they may impact targeting efficiency and overall program impact. Also, since polygamous households are common in Burkina Faso, a common approach needs to be defined on how to approach these households (as one or several households). Finally, more research on the most appropriate targeting unit (household or individual) is needed. Ongoing research conducted by the University of Montreal reported that community-based targeting (to identify indigents eligible to fee waivers) led to the identification of poor individuals living in non-poor households. These are elements that would be missed in a proxy- means test (household-based) approach. 196. Work toward a common registry. The identification of chronically poor individual/households should (eventually) inform different social safety net programs like cash transfers and fee waivers as well as the envisioned National Health Insurance scheme. It could also inform other initiatives such as tax exemptions. This will greatly contribute to ensure better synergies and economies of scale within the social protection system. 101 REFERENCES Amnesty International (AI). 2009. `Donner la vie, risquer la mort. La mortalité maternelle au Burkina Faso'. December. Balkouma Boursin, F. 2008. `L'impact des activités du Program Pistes Rurales (PrEst) sur la migration des jeunes dans la région de l'Est du Burkina Faso'. Rapport d'Etude. Version synthétique. Helvetas Burkina Faso. December. Ouagadougou, Burkina Faso. Bundy, D., C. Burbano, M. Grosh, A. Gelli, M. Jukes, and L. Drake. 2009. `Rethinking School Feeding: Social Safety Nets, Child Development, and the Education Sector'. June. World Food Program & World Bank: Washington, DC. Carneiro, P. and J. Heckman. 2003. `Human Capital Policy'. IZA Discussion Paper No. 821. July. Coady, D., M. Grosh, and J. Hoddinott. 2004a. `Targeting of Transfers in Developing Countries. Review of Lessons and Experience'. October. World Bank: Washington, DC. ______. 2004b. `Targeting Outcomes Redux'. The World Bank Research Observer, Vol. 19, No.1, pp.61-85. World Bank: Washington, DC. del Ninno, C., K. Subbarao and A. Milazzo. 2009. `How to Make Public Works Work: A Review of the Experiences'. Social Protection Discussion Paper No. 0905. World Bank: Washington, DC. Dembele, S. 2010. `Diagnostic de la commission sectorielle et thématique chargée de la protection sociale'. Draft 01. January. Ouagadougou, Burkina Faso. Department for International Development, United Kingdom (DFID UK), HelpAge International, Hope & Homes for Children, Institute of Development Studies, International Labor Organization, Overseas Development Institute, Save the Children UK, United Nations Children's Fund (UNICEF), United Nations Development Program (UNDP) and the World Bank. 2009. `Joint Statement on Advancing Child- Sensitive Social Protection'. August. Fonds National de Solidarité (FNS) [Burkina Faso]. 2009. `Rapport d'activités 2008'. Ministère de l'Action Sociale et de la Solidarité Nationale. January. Ouagadougou, Burkina Faso. Grosh, M., C. del Ninno, E. Tesliuc, and A. Ouerghi. 2008. `For Protection and Promotion: The Design and Implementation of Effective Safety Nets'. World Bank: Washington, DC. Gottschalk J., V. Manh Le, H. Lofgren and K. Nouve. 2009. `Analyzing Fiscal Space Using MAMS: An Application to Burkina Faso'. International Monetary Fund Working Paper 09/227. Government of Burkina Faso (GoBF). 1996. `Loi n°013/96/ADP portant orientation de l'éducation au Burkina Faso'. Ouagadougou, Burkina Faso. ______. 2008a. `Politique nationale de l'emploi'. March. Ouagadougou, Burkina Faso. ______. 2008b. `Plan d'Action Opérationnel (PAO) pour la mise en oeuvre de la politique nationale de l'emploi'. March. Ouagadougou, Burkina Faso. 102 ______. 2008c. `Plan d'Urgence pour la réalisation de la sécurité alimentaire et nutritionnelle au Burkina Faso'. June. Ouagadougou, Burkina Faso. ______. 2009a. `Plan National Multiriques de préparation et de réponse aux catastrophes. préparation, interventions de premiers secours, réhabilitation et reconstruction. période couverte : janvier à décembre 2009'. Ouagadougou, Burkina Faso. ______. 2009b. `Rapport national sur la sécurité alimentaire'. Document prepared for the World Summit on Food Security. October. Ouagadougou, Burkina Faso. ______. 2010. `Stratégie de croissance accélérée et de développement durable (SCADD) 2011-2015. Canevas de rédaction'. February. Ouagadougou, Burkina Faso. Helvetas. 2008. `Rapport de fin de phase du Program Pistes Rurales - Désenclavement à l'Est (PrEst) : Octobre 2004-Mars 2008'. March. Ouagadougou, Burkina Faso. Institut de Recherche pour le Développement (IRD). 2009. `Évaluation du processus d'identification des ménages bénéficiaires dans le cadre de l'opération d'urgence d'assistance aux ménages pauvres et très pauvres de la ville de Ouagadougou (Burkina Faso ­ EMOP 10773.0)'. Preliminary Report. 2nd version. November. Ouagadougou, Burkina Faso. International Labor Office [ILO] and World Health Organisation [WHO]. 2009. `The Social Protection Floor. A Joint Crisis Initiative of the UN Chief Executives Board for Co-ordination on the Social Protection Floor'. October. Geneva, Switzerland. International Monetary Fund (IMF). 2010. `Politiques de protection des pauvres contre l'impact de la flambée des prix des produits alimentaires et énergétiques'. Ouagadougou, Burkina. James, C., et al. 2005. `Impact on child mortality of removing user fees: simulation model'. British Medical Journal, 2005. 331(7519): p. 747-9. Kazianga, H., de Walque, D., and Alderman, H. 2008. `Educational and Health Impact of Two School Feeding Schemes: Evidence from a Randomized Trial in Rural Burkina Faso'. December. Kildal, N. and S. Kuhnle. 2002. `The Principle of Universalism: Tracing a Key Idea in the Scandinavian Welfare Model'. Paper presented at the First Conference of the European Social policy Research Network, Social Values, Social Policies. Tilburg University, Tilburg, The Netherlands, 29-31 August 2002. Meessen, B, et al. 2009. `Removing User Fees in the Health Sector in Low Income Countries. A Multi-Country Review'. September. Working Paper prepared for UNICEF. Ministère de l'Action Sociale et de la Solidarité Nationale (MASSN) [Burkina Faso]. 2007. `Politique nationale d'action sociale'. April. Ouagadougou, Burkina Faso. ______. 2008. `Charte nationale de solidarité'. Ouagadougou, Burkina Faso. ______. 2009a. `Program d'introduction d'un filet social de sécurité pour la protection et la promotion des enfants des ménages démunis au Burkina Faso'. First Draft. June. Ouagadougou, Burkina Faso. ______. 2009b. `Projet de protection et de promotion sociale des personnes défavorisées au Burkina Faso 2009'. October. Ouagadougou, Burkina Faso. 103 ______. 2009c. `Projet de protection et de promotion sociale des personnes défavorisées au Burkina Faso 2010'. November. Ouagadougou, Burkina Faso. Ministry of Health (MoH) [Burkina Faso]. 2000. `Plan National de Développement Sanitaire 2001-2010'. Ouagadougou, Burkina Faso. ______. 2009. `Planning for Going to Scale with Nutrition Interventions. An Experience Worth Sharing'. Presentation made at the World Bank in Washington on November 11, 2009. Ministère de l'Economie et des Finances (MEF) [Burkina Faso]. 2009a. `Document de STRATEGIE DE CROISSANCE ACCELEREE ET DE DEVELOPPEMENT DURABLE : 2010-2015. Note Conceptuelle'. Février. Ouagadougou, Burkina Faso. ______. 2009b. `Plan d'actions 2009-2010 du gouvernement pour faire face aux effets de la crise financière et économique sur l'économie Burkinabé'. December. Ouagadougou, Burkina Faso. Ministère de l'Economie et du Développement (MEDEV) [Burkina Faso]. 2004. `Poverty Reduction Strategy Paper'. July. Ouagadougou, Burkina Faso. ______. 2005. `Etude Nationale Prospective « Burkina 2025 ». Rapport sur la Vision Prospective et les Stratégies'. April. Ouagadougou, Burkina Faso. Ministère de l'Enseignement de Base et de l'Alphabétisation (MEBA) [Burkina Faso]. 2000. `Plan Décennal de Développement de l'Education de Base'. Ouagadougou, Burkina Faso. Ministère des Enseignements Secondaire, Supérieur et de la Recherche Scientifique (MESSRS) [Burkina Faso]. 2009. `Les cantines scolaires du secondaire'. April. Ouagadougou, Burkina Faso. Ministère du Travail et de la Sécurité Sociale (MTSS) [Burkina Faso]. 2007. `Politique nationale de protection sociale'. January. Ouagadougou, Burkina Faso. Nouvé, K., J. Kafando, A. Savadogo, D. Sebre, C. Tsimpo, and Q. Wodon. 2009 (unpublished). `Dynamique de la pauvreté et de la précarité au Burkina Faso (2003- 2007)'. Draft World Bank Working Paper `Croissance, pauvreté et secteurs sociaux au Burkina Faso: les dynamiques à moyen terme'. Chapter 1. Organisation for Economic Co-operation and Development (OECD). 2006. `Paris Declaration Monitoring Survey'. Réseau de Veille sur la Commercialisation des Céréales (RVCC) [Burkina Faso]. 2008. `Evaluation de l'opération de vente de céréales à prix social dans les provinces déficitaires : cas de la province de l'Oudalan'. Ouagadougou, Burkina Faso. Ridde, V. 2003. `Fees-for-Services, Cost Recovery, and Equity in a District of Burkina Faso Operating the Bamako Initiative'. Bulletin of the World Health Organization 2003, 81 (7): p. 532-8. Ridde, V. and S. Haddad. 2009. `Abolishing User Fees in Africa'. PLoS Med 6(1): e1000008. doi:10.1371/journal.pmed.1000008. January. Ridde, V. and F. Morestin. 2009. `Une recension des écrits scientifiques sur l'abolition des frais aux usagers dans les services de santé en Afrique'. January. 104 Ridde, V. and A. Bicaba. 2009a. `Revue des politiques d'exemption/subvention du paiement au Burkina Faso : La stratégie de subvention des soins obstétricaux et néonataux d'urgence'. University of Montreal. January. ______. 2009b. `The National Subsidy for Deliveries and Emergency Obstetric Care in Burkina Faso'. Presentation made on 4 November 2009 at the UNICEF Social Policy Network Meeting in Dakar. Ridde, V. et al. 2009. `A Community-Based Targeting Approach to Exempt the Worst-Off from User Fees in Burkina Faso'. Journal of Epidemiology and Community Health. January 2010. Volume 64. Issue 1:10­15. Samson, M., van Niekerk, I., Mac Quene, K. 2006. `Designing and Implementing Social Transfer Programs'. EPRI: Cape Town, South Africa. Save the Children UK (SCUK), HelpAge International (HAI), and Institute of Development Studies (IDS). 2005. `Making Cash Count. Lessons from Cash Transfer Schemes in East and Southern Africa for Supporting the Most Vulnerable Children and Households'. London, UK. Save the Children UK and FEG Consulting. 2008. `The Household Economy Approach. A Guide for Program Planners and Policy-Makers'. March. London, UK. Secrétariat Exécutif du Conseil National de Sécurité Alimentaire (SE-CNSA) [Burkina Faso]. 2009. `Informations sur la sécurité alimentaire au Burkina Faso : Etat des lieux ­ perspectives. Session interparlementaire sur la sécurité alimentaire mondiale'. Document prepared for the World Summit on Food Security. September. Ministère de l'Agriculture, de l'Hydraulique et des Ressources Halieutiques. Ouagadougou, Burkina Faso. Secrétariat Permanent du Conseil National de Lutte contre le Sida et les Infections Sexuellement Transmissibles (SP/CNLS-IST) [Burkina Faso]. 2008. `Recherche action sur la prise en charge communautaire des orphelins et enfants vulnérables du VIH/SIDA dans les Provinces du Nahouri et du Sanmatenga'. Septembre. Ouagadougou, Burkina Faso. Secrétariat Permanent du Conseil National de Secours d'Urgence et de Réhabilitation (SP/CONASUR) [Burkina Faso]. 2008a. `Première session ordinaire du CONASUR. Rapport d'activités 2005-2007 du SP/CONASUR'. Décembre. Ouagadougou, Burkina Faso. ______. 2008b. `Deuxième session ordinaire du CONASUR. Rapport d'activités 2008 du SP/CONASUR'. February. Ouagadougou, Burkina Faso. Smith, W. and K. Subbarao. 2003. `What Role for Safety Net Transfers in Low Income Countries'. Social Protection Discussion Paper No. 0301. January. World Bank: Washington, D.C. van de Walle, Dominique. 2010. "Widows and Social Protection in Mali and Burkina Faso." Mimeo, PRMGE, World Bank, Washington, DC. World Bank. 2004. `Burkina Faso Risk and Vulnerability Assessment'. Report No. 28144-BUR. January. Washington, D.C. ______. 2005. `Burkina Faso: Reducing Poverty through Sustained Equitable Growth. Poverty Assessment'. Report No. 29743-BUR. June. 105 ______. 2009a. `Au-delà du Paradoxe Burkinabé : Feuille de route pour une dépense publique plus efficace, équitable et de meilleure qualité'. Burkina Faso Public Expenditure Review. Report No. 50354-BF. September. ______. 2009b. `Burkina Faso: Promoting Growth, Competitiveness and Diversification'. Country Economic Memorandum. Volume III: Enhancing Growth Factors. Report No. 51815-BK. November. World Food Program (WFP). 2008a. `Transferts monétaires en zones urbaines : villes de Ouagadougou et Bobo-Dioulasso. Analyse de marché opérationnelle'. September. Ouagadougou, Burkina Faso. ______. 2008b. `Emergency Response to High Food Prices in Burkina Faso Cain Cities'. EMOP 10773.0. ______. 2009a. `Résultats du post distribution monitoring (PDM) de l'EMOP 10773.0'. April. Ouagadougou, Burkina Faso. ______. 2009b. `Evaluation à mi-parcours des interventions de l'opération d'urgence d'assistance aux ménages très pauvres et pauvres des villes de Ouagadougou et Bobo-Dioulasso'. Mission conjointe Gouvernement/CONASUR ­ Program Alimentaire Mondial ­ Catholic Relief Servoces ­ Croix Rouge Burkinabè. Preliminary Report. July World Health Organization (WHO). 2008. `Primary Health Care. Now More Than Ever'. The World Health Report 2008. Geneva, Switzerland. 106 Annex 1: Glossary of Terms For many specific terms used in this report, there is no overall consensus on a universal definition. In order to avoid misunderstanding, definitions used in this report are presented below. Poverty and vulnerability Chronic poverty Poverty that endures year after year, usually as a result of long-term structural factors faced by the household, such as low assets or location in a poor area remote from thriving markets and services. Transient poverty Poverty among households who are poor in some years but not all. They may be poor in some years due to idiosyncratic or covariate temporary shocks ranging from an illness in the household or the loss of a job to drought or macroeconomic crisis. Vulnerability The likelihood or probability that a household will pass below the defined acceptable threshold of a given indicator and fall into poverty. Social protection Social protection The set of public interventions aimed at supporting the poorer and more vulnerable members of society, as well as helping individuals, families, and communities manage risks. Social protection includes safety nets (social assistance), social insurance, labor market policies, social funds, and social services. Social assistance Synonymous to "social safety net". Social safety net Noncontributory transfer programs targeted in some manner to the poor and those vulnerable to poverty and shocks. Social insurance Contributory programs designed to help households insure themselves against sudden reductions in income. Types of social insurance include publicly provided or mandated insurance against unemployment, old age (pensions), disability, the death of the main provider, and sickness. Program evaluation Effectiveness The extent to which the program objectives were achieved, or are expected to be achieved, taking into account their relative importance. 107 Efficiency An economic term which signifies that the intervention is using the least costly resources possible to achieve the desired results. Efficiency measures qualitative and quantitative outputs in relation to results. Impact Long-term effects, positive or negative. Sustainability Continuation of benefits after the end of the intervention. Probability to obtain benefits in the long term. 108 Annex 2: Good Practice Design Features for Direct Support drawn from International Experience Based on good practices in the design of social assistance (direct support) programs, four design features are fundamental to the discussion: Selection of Households for Direct Support It is common knowledge that the more generous the definition of eligibility, the larger will be the pool of applicants for social assistance and the cost of the program. The cost of the program also depends on the level and duration of benefits (discussed in the section below). For example, for the old age social assistance pension program, Nepal defined the cut-off age limit for eligibility as 75. This undoubtedly restricted the pool of beneficiaries and kept the program within the limits set by the available budget, but the program could not reach some of the critical vulnerable groups even slightly below the age 75. On the other hand, some countries set the limit for eligibility low at 60 (or even lower), leading to the opposite effect of too many beneficiaries and a very large budget. To overcome difficulties of this kind, many countries now resort to limiting the cash transfer social assistance to, say, the poorest 10 percent of the population. One such example is the Kalomo District Pilot Social Cash Transfer program in Zambia, which limited the outreach to the bottom 10 percent of the population. Determining the eligibility and selection of eligible beneficiaries has varied a great deal from one country to another, depending upon (i) administrative feasibility, and (ii) the available information. Where both these sets of conditions are weak, countries (such as Rwanda) have resorted to community targeting approaches. However, good practice dictates validation of such selections via a transparent communication system. For example, in Mexico's conditional cash transfer program, Oportunidades, beneficiary lists are presented at community meetings which has given communities a chance to pick both exclusion and inclusion errors. In all cases, a complaints mechanism is critical for ensuring community satisfaction with the targeting approach. Where the information constraint is less severe, countries have adopted a proxy means test (PMT). This is a targeting method by which a score for each applicant is generated based on household characteristics that are fairly easy to observe, usually non-income characteristics such as the location and quality of housing unit, ownership of durable assets, number of children, level of education, etc. A threshold score level is set below which a household becomes eligible for the benefit. When a community targeting approach such as the one described in paragraph 3 above is adopted, it is still possible to move gradually to a PMT method, short-list households for the benefit, and then use community meetings to ensure transparency and avoid exclusion and inclusion errors. Many low income countries are resorting to this combination of a PMT and community validation as a means to select beneficiaries for social safety net programs generally. A recent example is Bangladesh, which is now sponsoring a social safety net program using a combination of PMT and community validation. Determining Benefit Levels Determining the size of social assistance direct support is a tricky issue in all countries. It is hard to provide clear-cut policy advice based on international experience, but some guiding principles can be offered. Typically, in last resort programs such as the one 109 proposed for Rwanda, which aims to reduce extreme poverty, the benefit levels are set as a fraction of the income gap of target beneficiaries. How high or low that fraction should be depends on the available budget and the number of people in extreme poverty. Using a proxy means test, Armenia and Georgia have used this principle. In this regard, some number crunching might be helpful with the available household level information. For example, information on the number of extremely poor households can be combined with information on their income (poverty) gap and, from both these sets of information, one can derive the financial requirements for a given level of benefit. One can then see the feasibility (affordability) of alternative benefit levels and decide on the level that can be defended within the available budget envelope. Benefit levels need not be fixed at a flat level for all types of households. Instead, the levels can be varied. Variable benefit formulas are often the norm in many countries, mainly because such formulas allow for variation in household circumstances (such as number of children, presence of a disabled child or person, long-term sickness of a household head, etc.). A very good (successful) example of such a variable benefit formula is that of Brazil's Bolsa Familia program. It provides two types of benefits: a base benefit to all families in extreme poverty, and a variable benefit that depends on family composition and income. Whatever method is adopted to determine the benefit level, it is useful to assess the level of benefit as a percentage of the consumption expenditure of extremely poor households. How generous the program is can be assessed from this proportion. The higher the benefit level as a percentage of the household's consumption expenditure, the more generous the program. Maintaining a generous benefit level is likely to impact on labor supply through adverse disincentive effects; that is, households, even when provided with opportunities to work in the labor market may opt to stay in the "generous" program. While this concern for the impact of benefit levels on work disincentives is theoretically valid, it does not apply to programs that target extremely poor households with no adult labor to participate in the labor market, the disabled, or the elderly. Delivery Mechanisms and Payment Modalities Four principles generally guide the delivery mechanism: (i) ensuring reliability and regularity of payments; (ii) maintaining accountability (governance issues) and prevention of fraud; (iii) reducing transaction costs to the beneficiaries; and (iv) minimizing the administrative cost of delivery. While a number of delivery agencies or routes are available ­ bank branches, mobile banks, post offices, decentralized government agencies, NGOs ­ the selection of the delivery mode eventually must satisfy the above four principles, and be available and suitable for a given country situation. Not surprisingly, countries have varied a great deal in this regard. If contractors or a specific agency are/is selected, performance-based incentive contracts can be developed as was done in the Brazil's Bolsa Familia program. Kenya has tried a pilot program for a cash transfer to orphans and vulnerable children through a bidding process to select a lowest-cost service agency. The advantage of these contract-based service agencies is that the contracts can be revised based on performance. In countries with somewhat developed IT infrastructure, debit cards and smart cards are being used to transfer cash assistance. In the state of Gujarat in India, a pilot program is being tried to use smart cards to transfer in kind social assistance. It is hard to recommend one specific option or delivery mechanism: the main 110 challenge is to adapt any one reliable mechanism to country circumstances and avoid any unintended adverse effects. Monitoring and Evaluation of Cash Transfer Programs Program monitoring is extremely important for any safety net program and especially for cash transfer programs. Systematic monitoring helps one to assess how well the program is being implemented at all levels, and helps mid-course correction in the event of poor implementation. Evaluation complements the monitoring system, inasmuch as it allows an assessment of the distributive effects of cash transfer programs. Despite the critical importance of monitoring and evaluation, unfortunately most safety net programs lack a credible monitoring and evaluation system in place. Monitoring is a continuous activity, and is typically done at all levels ­ village, district, and national. Its main role is to assess whether or not the program is being implemented in accordance with its design with outcomes as expected. Its annual cost must be factored into the program costs and it must become an integral part of the programmatic framework. A good monitoring system must collect information on the program's key outcomes. A good practice procedure is that monitoring should be done by an independent agency, outside of the agency or institution implementing the program. A well-documented international experience of good monitoring is from Zambia. The Kalmo District Pilot Social Cash Transfer program (which operated with technical assistance from Germany) implemented third-party monitoring that focused on the quality of program management, the effectiveness of targeting, regularity of transfer payments, and even beneficiaries' use of the transfers. It is not enough to know the program's outcomes; it is important also to know the impact of the program on household welfare, which is the ultimate goal of a cash transfer program. Several techniques are available to do an impact evaluation. Two approaches can be distinguished. One is a quantitative approach that collects information on a random sample of households belonging to both the treatment group and the control group, both at the launch of the program and after a given period of time (say, one year). Econometric techniques are then used to assess the impact of the program.56 A complementary approach is qualitative evaluation, which is based on focus group interviews, key informant interviews, and direct observation. Though qualitative evaluations are not representative, they do offer rich information on the program's functioning, merits, and shortcomings. A variant of descriptive evaluation is "process evaluation," which is probably the most common evaluation technique followed in many countries. Its approach is to assess and document how each of the processes underlying a cash transfer program is being implemented. It helps address the question, What is happening through the program? Process evaluation strongly complements, but does not substitute for, an internal monitoring system and other evaluations mentioned above. For example, the Zambia example is worth repeating: It included a process evaluation by external evaluators in 56 For a thorough understanding of the techniques and applications of the impact evaluation, visit the World Bank website on Impact Evaluation: http://web.worldbank.org/WBSITE/EXTERNAL/TOPICS/EXTPOVERTY/EXTISPMA/0,,menuPK:384336 ~pagePK:149018~piPK:149093~theSitePK:384329,00.html 111 addition to other evaluations which helped improve the monitoring capacity as it revealed specific flaws in specific processes underlying the program's implementation. Source: Grosh et al. 2008. `The Design and Implementation of Effective Safety Nets: For Protection and Promotion'. World Bank. 112 Annex 3: Some Country Examples of Good Practice in Safety Net Programs in Africa Recent financial crises and price hikes have increased policy-makers' interest in finding ways to address persistent, and often deepening, vulnerabilities. The success of cash transfer programs in many parts of the world has led many leaders to ask whether cash transfer programs could be successful in addressing the major challenges present in Sub- Saharan Africa (SSA). This section examines how cash transfers have been used throughout the region and highlights the lessons that have already been learned through existing cash transfer programs. Taking into account the context of Mali, the following selected country case examples could provide valuable lessons in understanding how programs are implemented in other African countries. Ethiopia's Productive Safety Net Program (PSNP) In Ethiopia, over 40 percent of the population lives below the national poverty line and over 20 percent of the population is extremely poor (below 1,650 kilocalories per person per day). Since the variability in rainfall is among the highest in the world, and fluctuations in rainfall are inversely related to mean incomes, every year for more than two decades the government of Ethiopia has launched an international emergency appeal for food aid. This annual emergency assistance was designed to meet the consumption needs of both chronically and transitorily food-insecure households. Despite a substantial amount of humanitarian assistance, evaluations have shown that emergency assistance was unpredictable for both planners and households, often arriving late relative to need. As a result of the delays and uncertainties, the emergency aid could not be used effectively and did little to protect livelihoods, prevent environmental degradation, generate community assets, or preserve household assets (physical or human capital). Characteristics of the Program Given these shortcomings of the emergency aid regime, in 2005 the Ethiopian government started implementation of a new program, the Productive Safety Net Program (PSNP). The PSNP replaced the emergency humanitarian appeal system as the chief instrument in the country's safety net. The program is currently operational in 234 chronically food- insecure districts (of a total of 692 districts), and targeted about 7 million people in 2006. The PSNP provides resources to chronically food-insecure households in two ways: (i) through payments to the able-bodied for participation in labor intensive public works activities; and (ii) through direct grants to households composed of the elderly or those who cannot work for other reasons. Impact of the PSNP A 2005 beneficiary survey found that the PSNP had a significant positive effect on beneficiaries' well-being as calculated by both subjective and objective indicators. The survey found that three in five beneficiaries avoided having to sell assets to buy food in 2005, and according to 90 percent of the households, this was a result of their participation in the PSNP. Moreover, almost half the beneficiaries surveyed stated that they had used health care facilities more and 76 percent of these households credited the PSNP with this enhanced access. More than one-third of surveyed households enroled more of their children in school and 80 percent of them attributed this to participation in the PSNP. 113 Ongoing Reforms Significant work is planned to further improve implementation capacity and bring systems to a level of functioning not previously possible with fragmented and temporary programs. Work is also beginning on a contingent grant mechanism (conditional cash transfer) to provide resources in the same districts to help transient food-insecure households during periods of drought. The mechanism will use a rainfall-based index that uses 30 years of rainfall data to trigger funding. Moreover, the PSNP is complemented by a larger food security program that tries to help households raise incomes by means of resettlement grants, household income-generating packages, and water harvesting. Households that benefit from the PSNP are also entitled to assistance under other parts of the food security program. Food security interventions financed by donors that fall outside the PSNP are, however, rarely coordinated at local levels, and their links to basic rural services are also weak. Lessons Learned The PSNP illustrates many of the issues that surround safety nets in very low income countries, namely: The program is moving in a clearly beneficial direction by means of a basic design that not only seeks to use resources in ways that save lives, but also assist in livelihoods. The progress in implementation to date suggests that this is possible even in a very low income setting. The design process and implementation planning have undergone a fairly harsh triage. Even when fully realized, the program will only provide a safety net in about a third of the country. The districts selected are appropriately the poorest, but many poor people also live in the unserved districts. Moreover, the program has phased its implementation. It is focusing first on consolidating the basic PSNP. It hopes to enrich it eventually in a number of dimensions, but program managers and donors have realized that everything could not be accomplished right away. Thus, for example, the contingent fund for droughts was not implemented until the third year of the PSNP. Good implementation requires diligent and sustained effort. By 2007, the program had many positive outcomes, and early qualitative assessments of its targeting and impacts are positive, but more remains to be done to consolidate implementation. Good implementation also requires flexibility and innovation. For example, the government was initially having problems with the program's monitoring system, but in the interim, it deployed so-called rapid response teams to visit districts to identify and solve implementation problems. This gave managers a sense of what was going well and what was not and whether adjustments were needed in individual districts or at a more systemic level. Meanwhile, the design of the monitoring system was simplified and a pilot to computerize it is under way. An important part of the reform is the shift to a multidonor, multiyear framework rather than an annual emergency appeal system with each donor running a separate initiative. This is complemented by the decision to deliver the program through regular government systems rather than special implementation units common in donor-funded programs. The multiyear framework and the reduction in fragmentation should permit the development of much more effective administrative systems. The multidonor 114 framework should also aid in resilience, in that withdrawal or a reduced commitment by a single donor will have a less deleterious effect. Kenya Cash Transfer Program for Orphans and Vulnerable Children (CT-OVC) This transfer programs began as a pre-pilot in 2004. It has since gone through a five-year pilot project and scaled up from a very small budget to a projected US$26m budget for fiscal year 2010 (WB 2009c). Extensively documented, the program has provided valuable experience in advocacy, design, and implementation of conditional cash transfers in SSA settings. It is a key component of Kenya's broader social protection strategy, as it addresses risks to children in communities where large number of OVCs, exasperated by adult deaths from AIDS, have begun to overwhelm informal safety net systems. In addition to donor interest, the CT-OVC initiatives have received strong domestic political support, including pressures to quickly scale up of the program. Objectives of the Pre-Pilot Program The goal of the pre-pilot was to generate evidence regarding the applicability of a cash transfer program to support OVCs in Kenya. The pre-pilot phase began in December of 2004, initially reaching 500 children. It was later expanded to reach at least 5,000 children. The pre-pilot was supported through UNICEF and SIDA and administered from the Department of Children Services (WB 2009c). The program's initial districts - Nairobi, Kwale, and Garissa - were selected because they were areas where UNICEF and SIDA already had ground-level knowledge and experience. The pre-pilot targeted poor households and households with OVCs that did not receive other formal support. Beneficiaries received Ksh 500/US$6.25 monthly per child (SCUK et al. 2005). Technically, the pre-pilot transfers had conditions attached, but there were no consequences for non-compliance (WB 2009c). Concerns that children would be separated from their households in order to meet program requirements led the pre-pilot to drop enforcement of conditions (WB 2005). However, communities and some donors requested that the transfers be conditioned, particularly as the program expanded to the west in areas with higher HIV levels. CT-OVC Redesigned for Full Pilot Drawing on pre-pilot experiences, the official pilot of the CT-OVC program began in 2005 and ran through mid-2009. The program specifically focuses on households with OVCs, with the goal of keeping children within families and encouraging investment in their human capital. The specific program goals are very similar to those seen in other well- known CCT programs in Latin America, including improving health, nutrition, education, and awareness of these issues. The pilot program planning envisaged reaching seven districts, with support from the Government of Kenya, DFID, UNICEF, and SIDA (WB 2009c). Funds from development partners in the pilot reached 17,500 households that are still being covered by benefits. Between 1,000 and 4,600 beneficiary households are covered in each of the districts. By the end of Phase 2 (June 2009), benefits reached 70,000 households. Targeting Takes a Complex Five-Step Approach including Community Committees Targeting in the pilot was refined from pre-pilot methods. The targeting is completed through five steps (WB 2009c). Geographic targeting selects program districts based on poverty and HIV/AIDS levels. The districts are ranked based on the number of extremely poor OVC households in the district. Within the districts, the number of households with 115 OVCs is calculated. Communities are selected to belong to the program provided there are more than 5,000 community members, of which at least 60 percent have to live below the poverty line (Hussein 2006). Community committees (Location OVC Committees) were created to select eligible households. The households must not be able to meet all of their basic needs, and they must have a permanent OVC member under 17 years old in the household who is not receiving benefits from another cash transfer program (GoK 2006). Within this group of eligible households, Location OVC Committees decide which households meet three of a list of over ten items related to poverty (such as whether the household has access to a safe water source, members are in poor health, or members eat one or fewer meals per day). Households meeting at least three of the criteria are considered poor (WB 2009c). Post Office Functions Well for Transfers The transfer size was set at a level that was believed to cover enough of the needs of OVCs to help keep them within their households. Transfer values vary by the number of OVCs in the household. Ksh 1,000 (US$14) is given in households with one or two OVCs, Ksh 2,000 (US$28) is given to households with three or four OVCs, and Ksh 3,000 (US$42) is given to households with five or more OVCs (WB 2009c). Using Ksh 1,500 (US$20) as a reference transfer value, the transfer is sizable compared to the average of Ksh 1,800 per adult equivalent for consumption. The transfer therefore is approximately equal to 20 percent of poor Kenyan households' expenditures (WB 2009c). However, the transfers have not been indexed to inflation, so their value has eroded as food prices have grown. Transfers in the pilot districts are delivered using the Postal Cooperation of Kenya, which was found to function well. Payments are awarded once every two months (OVC 2007). The transfers are supplied along with a receipt outlining if the household received the full possible payment, and if not, why (GoK 2006). They are given to the household's mother or female head/caretaker whenever possible. Soft Enforcement of Conditions Similar to conditions in cash transfer programs in Latin America, Kenya's CT-OVC beneficiaries have responsibilities relating to child health and education: beneficiaries under one year old must attend a local clinic six times within their first year to receive immunizations, vitamin A supplements, and to have their growth monitored; beneficiaries between one and three years old must have a growth-monitoring check-up and receive vitamin A supplements twice per year; children between 6 and 17 years old must enrol in school and maintain attendance for 80 percent of all days; and caretakers must attend educational seminars at least once annually (WB 2009c). However, until very recently, these conditions have not been applied in the program; it has essentially been an unconditional transfer. Part of the program's design was to test a conditional versus an unconditional transfer, but this design component did not begin to be tested until late 2008. Thus far, there has been confusion over how to apply health conditions, and hence only education conditions have been applied. In areas where conditions are applied, the reduction in transfer, for non-compliance, is Ksh 400 per child or adult that does not comply with co-responsibilities (WB 2009c). Program exit occurs if there is no longer an OVC in the household under 17 years old or the household is reassessed and no longer deemed to be poor. Households that migrate from the program area, voluntarily withdraw, or are found to have falsified information are also no longer in the program (WB 2009c). Finally, after three consecutive periods of failing to fulfill co-responsibilities, households are supposed to exit the program. 116 Organization and Management System Requires Inter-Sectoral Coordination The pilot's Central Program Unit, comprised of units for operations, monitoring and evaluation, administration/finance, and information systems, was originally situated within the Department of Children's Services in Kenya's Ministry of Home Affairs (GoK 2006). The Vice-President holds ultimate control over the program (Hussein 2006). Enforcement of conditions requires close coordination by line ministries with the program, as the education objectives are to be executed by the Ministry of Education, and the health objectives are executed by the Ministry of Public Health and Sanitation. Other coordination with the Ministry of Medical Services and Ministry of Immigration and Registration of Persons are also being supported (WB 2009c). Analysis of results Awaited from Experimental Evaluation Design The pilot has taken significant measures to maintain adequate controls, including the use of an extensive management information system (MIS). The current MIS is centered at the national level but will later be decentralized to the districts (WB 2009c). Teachers and health care workers fill out forms reporting school attendance and health center visits. The central MIS tracks information by district. Conditions are monitored every two months for children ages zero to one, every six months for children ages one through five, every three months for the educational conditions, and once every year for the adult training sessions. Conditionality monitoring also is supposed to work through this system. Application of conditions is supposed to be spot checked, including through visits to beneficiary households to ensure program compliance. Appeals may be submitted to the District Children Office, who also accepts complaints concerning payment quantities and quality of supply side services. Impact Evaluation of the Program The pilot program in the original seven districts is subject to an impact evaluation, conducted by Oxford Policy Management, with qualitative and quantitative components. The evaluation design of the program is experimental (although there were significant differences across the treatment and control groups), in which two treatment locations and two control locations are randomly selected within each of the seven districts. The unconditional/conditional design was also randomly assigned (Hurrell, Ward, and Merttens 2008). The baseline sample includes 2,759 households. Its analysis of targeting revealed that most selected households did have an OVC (98 percent), and most of these households were poor. However, the extremely poor were underrepresented in the program (Hurrell, Ward, and Merttens 2008). Strong Government Ownership and Funding and Expansion included in Medium-Term Plan The CT for OVCs is included in Kenya's Medium-Term Plan and Vision 2030. The Kenyan government funded the CT-OVC program in 2005-2006 using US$675,000, or KShs. 48,000,000 (Hussein 2006). Due to its expansion, the program is expected to cost US$26m in FY10. This figure is 0.08 percent of nominal GDP and 0.31 percent of government expenditures. When the program reaches 100,000 households, it is expected to cost between US$32m and US$35m, or approximately 0.07 percent of nominal GDP and 0.28 percent of government expenditures (WB 2009c). Administrative costs in the program are expected to be approximately 25 percent by 2012, and they are expected to continue to drop. This percentage is much lower than the 40 percent administrative costs in the pre-pilot. 117 CT-OVC Adaptations for Phase Three Scaling-Up The political pressures for more rapid expansion of the CT-OVC pilot into additional districts have resulted in two parallel programs running alongside each other. A new (third) program phase now seeks to harmonize the programs and build the capacity for their effective implementation. The goal of the Government of Kenya is to cover 100,000 poor households with OVCs by 2012 (approximately 2,000 households per district), in order to cover approximately half of the 600,000 extremely poor OVCs in the country (WB 2009c). During the third phase the following measures will be introduced: The targeting mechanism will be adjusted based on results from evaluations, Kenya's Integrated Household Budget Survey, the MIS, and baseline data. Improvements will be made to the standardized program based on lessons learned in the second phase. The MIS will be upgraded to enable it to efficiently handle the greatly increased system demands from the rapid scale-up, and an organization will be contracted to provide external monitoring. That external monitoring will spot check the program, conduct community censuses to evaluate the quality of Local OVC Committees, and conduct so- called "citizen report cards" that will determine beneficiary and non-beneficiary opinions and satisfaction with the program. This improved accountability is particularly important in light of concerns over governance and corruption in Kenya (WB 2009c). Extensive effort will be made to improve communication about the program to both beneficiaries and non-beneficiaries. Implementation and monitoring of the co- responsibilities is expected to improve. By mid- to late 2010, testing regarding the use of penalties resulting from non- compliance with co-responsibilities should be complete (WB 2009c). Evaluations of supply side capacity will also be conducted. This capacity building is crucial; program officials have made notable achievements in implementing the program and improving capacity already, but more must be done to meet the challenges of continued scaling up (WB 2009c). Malawi Cash Transfer Programs The Social Cash Transfer program began with UNICEF support as a pilot in Mchinji District in 2006, with goals to scale up eventually to a national program. Its objective is to decrease poverty, hunger, and starvation of the extremely poor and those without an eligible member able to participate in the labor force. This includes many households with Orphans and Vulnerable children (OVCs). Characteristics of the Program The objective of the Mchinji Social Cash Transfer Pilot was to address extreme poverty. Schubert and Huijbregts (2006) report that around 10 percent of all Malawian households (250,000) are extremely poor and incapable to work (i.e., labor constrained or labor incapacitated). It was suggested that if that 10 percent of households all received social cash transfers, the country's extreme poverty rate would decrease from 22 percent to 12 percent, at a cost of US$41m per year. This analysis contributed to the decision to target 10 percent of extremely poor households in the targeted pilot area of Mchinji, equal to approximately 3,000 households/15,000 individuals (Chipeta and Mwamlima 2007). In 118 addition to its objective related to poverty reduction, the program sought to improve beneficiary children's enrolment and attendance at schools, to provide information about how well a cash transfer program could fit into Malawi's social protection agenda (Chipeta and Mwamlima 2007), and to test whether District Assemblies could implement cash transfer programs that were both cost-effective and able to reach targeted household groups (Schubert and Huijbregts 2006). Mchinji was chosen for the pilot due to its strong District Team, average poverty levels, and relatively close location to the capital of Lilongwe. Targeting of the Mchinji Pilot includes Elected Village Committees Targeting criteria classified the extremely poor as those who reside in the bottom expenditure quintile and below the national extreme poverty line. Based on this definition, beneficiary households should be unable to purchase needed non-food goods. Labor constrained households are those with a dependency ratio of over three (Schubert and Huijbregts 2006). To select these households, local committees known as Community Social Protection Committees first create a list of all households they think may fulfill the program's requirement that they be "ultra poor" or "labor constrained." These committees are selected through community elections during the initial program meeting (Schubert 2007b). Village headmen are not allowed to be on the committees. The committees must then call on and interview all potential beneficiary households; the village headman must verify this information, and the committees rank identified households according to their level of neediness. The ranking is discussed and approved or changed in a community meeting. The information is passed to the Secretariat and a Social Protection Sub- Committee, who must approve or disapprove of the list. The lists are supposed to contain the 10 percent of households in the community agreed to be most needy. Design and Delivery of Transfers Monthly transfers in Mchinji, all unconditional, were graduated by household size and number of children in school. One-person households received MK 600 (about US$4), two-person households received MK 1,000 (US$6.67), three-person households received MK 1,400 (US$9.33), and four person households or larger received MK 1,800 (US$12) (Schubert and Huijbregts 2006). Households with children in primary school received MK 200 (US$1.33) additionally per child, and households with children in secondary school earned an additional MK 400 (US$2.67) per child. This bonus was not tied to school attendance. It was simply given when school aged children were in the household. The average transfer value was MK 1,700 (US$11.33)57 per household monthly, which was deemed large enough to fill the extreme poverty gap in targeted households (Schubert and Huijbregts 2006). There were 3,000 household beneficiaries by the beginning of 2008, and expenditures were US$43,000 monthly (Miller et al. 2008). The pilot scale up was postponed due to funding delays; however, it was able to reach seven districts by the end of 2008 (Horvath et al. 2008). As of April 2009, the scheme reached 92,786 beneficiaries in 23,561 households in seven districts (UNICEF/GOM 2009). 57 Standardizing the exchange rate to that reported previously; this is slightly different from the 2006 report by Schubert and Huijbregts. 119 Implementation Mechanisms and Financing The Social Cash Transfer program was implemented locally. The Ministry of Women and Child Development and the Department of Poverty and Disaster Management coordinated the pilot with help from UNICEF (Chipeta and Mwamlima 2007). The Mchinji pilot was implemented by the Local Assembly, whose District Executive Committee had a Sub- Committee on Social Protection with line ministry representatives. This sub-committee approved applications to the program. The Malawi district structure has officers that come from various departments and are able to support the program. Capacity is limited at the district level, but not as constrained as in some other countries (e.g., Zambia) implementing a similar program (Schubert and Huijbregts 2006). Below the Sub- Committee is the Social Cash Transfer Scheme Secretariat with personnel who implement the program, control the budget, and perform periodic monitoring. Below this, the Village Development Committee is in charge of the Community Social Protection Committee, which both targets and tracks beneficiaries (Schubert and Huijbregts 2006). The Community Social Protection Committee teams receive remuneration to compensate them for some activities performed (Schubert 2007b). For the pilot, UNICEF provided technical assistance, supported program setup, funded the transfers until December of 2006, and supported advocacy and capacity building in Malawi. This included funding visits of government representatives to Brazil and Zambia, holding workshops, and conducting field trips to Mchinji. Additional funding to scale up the program in 2008 and 2009 came from the National AIDS Commission through The Global Fund to Fight AIDS, Tuberculosis and Malaria. The Global Fund's contributions to the scale up were around US$8.8 million, and National AIDS Commission funds were used since approximately 70 percent of beneficiary households have been affected by HIV/AIDS (Schubert 2007a). The EU planned to fund external monitoring and evaluation. The country has expressed interest in obtaining further financing from development partners through a basket fund after the Social Cash Transfer has been incorporated into the National Social Protection Strategy and received full Cabinet support (Schubert and Huijbregts 2006). Other donors in a pool fund were expected to be the World Bank, DFID, CIDA, and NORAD (Horvath et al. 2008). Challenges in Evaluation Design Internal monitoring is completed through the production of monthly reports on costs, activities, outputs, and more. UNICEF and USAID support a joint external program evaluation conducted by Boston University and the Center for Social Research in Malawi (Miller et al. 2008). Targeting evaluations were completed in March and June of 2007, and a systems evaluation was conducted in October of 2007. The baseline household survey was conducted in treatment and comparison village groups in March of 2007 before treatment households received a grant. Follow-up surveys were carried out in August- September 2007 and March 2008, and qualitative data was collected from October- November 2007. However, it appears that experimental methods were compromised in the evaluation. The Mchinji District Secretariat chose which village groups were treatment and comparison groups; both treatment and comparison households were selected using the community targeting methods, and comparison households did not understand that the research was unrelated to their grant receipt. 120 Lessons Learned The targeting evaluation of the Mchinji Program found much need for improvement. Almost one-third of community members in program areas thought targeting was not fair. The evaluation suggested that less subjective indicators be used to determine program beneficiaries: targeting should be more objective, standardized, and transparent (Miller et al. 2008). Depending on their definition of eligibility, the exclusion errors in communities ranged from 37 percent through 68 percent. Beneficiaries' food consumption and diversity had improved over that of the comparison group. In addition, the health of both children and adult has improved, and the self-reported school attendance and capacity to study of children has increased. Child labor had also decreased significantly in the treatment group, while the comparison group's labor did not change. The evaluation also concluded that household productivity had increased since they received the transfers. The expected cost of scaling up the program nationally to 273,000 households (1.2 million individuals, of which 60 percent are expected to be OVCs), is around US$55 million annually, or 1.4 percent of GDP (Schubert 2009). In June of 2007, delivery of the transfers cost less than 2.5 percent of program costs, and administrative costs were less than 15 percent of the program costs (Horvath et al. 2008). The program has faced significant challenges, including: (i) the need for more, better trained district level staff; (ii) ongoing concerns over household dependency and corruption in the program (Chipeta and Mwamlima 2007); (iii) dealing with high turnover of government employees; (iv) the need for improved financial mechanisms to transfer funds at high levels, and improved MIS system that connected district and national level data; (v) the need to put a complaints/appeals procedure in place (UNICEF/GOM 2009); and (vi) scaling up the program necessitates increased government commitment, particularly from the Ministry of Finance, and additional capacity building at all levels of government. 121 4°W 2°W 0° 2°E This map was produced by the Map Design Unit of The World Bank. The boundaries, colors, denominations and any other information shown on this map do not imply, on OUDALAN the part of The World Bank Group, any judgment on the MALI legal status of any territory, or any endorsement or acceptance of such BURKINA FASO Gorom boundaries. SOUM Gorom LOROUM Djibo NIGER Dori 14°N To To Mopti a SÉNO Niamey YATENGA g ig N er n Titao te Ouahigouya BAM SANMATENGA a Sebba NAMENTENGA To Y Kongoussi YAGHA To Dosso San SOUROU Gourcy ZONDOMA GNAGNA KOSSI Tougan Kaya Sirba To Yako Ségou PASSORÉ KOURE- Bogandé KOMONDJARI Nouna Toma WÉOGO bi NAYALA Boussé OUBRITENGA Gayéri ou Boulsa or Ziniaré G Dédougou OUAGADOUGOU Réo Koudougou GANZOURGOU BANWA MOUHOUN KADIOGO r m a G o u Vol Solenzo Zorgo Koupéla Fada BOULKIEMDÉ ta N'Gourma Bla Kombissiri KOURITENGA Diapaga SANGUIÉ TAPOA 12°N n 12°N Vo GOURMA ch lta BAZÉGA e e Ro ir No Boromo ug ZO Tenkodogo ZIRO e BALÉ UN Manga ta BOULGOU KOULPÉLOGO Vol KÉNÉDOUGOU Sapouy DW Houndé ÉO Ouargaye KOMPIENGA HOUET G TUI O NAOURI Pama To Bougouni Dano SISSILI Orodara Bobo- IOBA Léo Po BENIN Dioulasso Diébougou To 2°E To Téna Kourou Djougou Parakou (747 m) Sindou BOUGOURIBA Banfora LÉRABA KOMOÉ PONI Gaoua To Tamale To Sokidé BURK INA FASO Komoé Volta Noire PROVINCE CAPITALS 10°N NOUMBIEL NATIONAL CAPITAL Batié GHANA To Bole TOGO RIVERS MAIN ROADS 0 20 40 60 80 100 Kilometers CÔTE D'IVOIRE RAILROADS SEPTEMBER 2004 IBRD 33379 PROVINCE BOUNDARIES 0 20 40 60 80 Miles To To Bouaké Koutouba INTERNATIONAL BOUNDARIES 4°W 2°W 0°