Report No: ACS7005 Republic of Senegal: Social Safety Net Assessment December 10, 2013 AFTSW AFRICA Standard Disclaimer: This volume is a product of the staff of the International Bank for Reconstruction and Development/ The World Bank. The findings, interpretations, and conclusions expressed in this paper do not necessarily reflect the views of the Executive Directors of The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Copyright Statement: The material in this publication is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applicable law. The International Bank for Reconstruction and Development/ The World Bank encourages dissemination of its work and will normally grant permission to reproduce portions of the work promptly. For permission to photocopy or reprint any part of this work, please send a request with complete information to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA, telephone 978-750-8400, fax 978-750- 4470, http://www.copyright.com/. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA, fax 202-522-2422, e-mail pubrights@worldbank.org. GOVERNMENT FISCAL YEAR January 1 to December 31 CURRENCY EQUIVALENTS Currency Unit : CFA Franc (FCFA) (as of October 2013) US$1 : FCFA 488 WEIGHTS AND MEASURES Metric System ABBREVIATIONS AND ACRONYMS APR Annual Progress Report APDC Acteurs porteurs de dynamiques communautaires CAPSU Caisse autonome pour la protection sociale universelle CEA Community Executing Agency CEL Cadre d’exécution local CCD Cadre de concertation départemental CCR Cadre de concertation régional CI Coordination de l’initiative CNCAS Caisse nationale du crédit agricole du Sénégal CLM Cellule de lutte contre la malnutrition CSA Food Security Commissariat (Commissariat à la sécurité alimentaire) CSS Social Security Agency (Caisse de sécurité sociale) DCaS National School Lunch Program (Programme d'alimentation scolaire) DHS/MICS Demographic Health Survey/Multiple Indicator Cluster Survey ESPS Senegal Poverty Survey (Enquête de suivi de la pauvreté au Sénégal) ESSD Sustainable Development Network FC Fonds de calamité FCFA West African Franc FG Fonds de Garantie FSN National Solidarity Fund (Fond de solidarité nationale) GDP Gross Domestic Product HIV-AIDS Human Immuno Deficiency Virus/Acquired Immune Deficiency Syndrome IFAN Institut fondamental d’Afrique noire IMF International Monetary Fund INPS National Social Protection Initiative (Initiative Nationale de la Protection Sociale “Squali Jaboot�) IPRES Retirement Institute of Senegal (Institut de prévoyance retraite du Sénégal) IPSEV Social Protection Initiative for Vulnerable Children (Initiative de protection sociale des enfants vulnérables) MDGs Millenium Development Goals MIS Management Information System NETS Cash Transfers for Child Nutrition Program (Nutrition ciblée sur l'enfant et transferts sociaux) NGOs Non-Governmental Organizations OECD Organization for Economic Cooperation and Development OEV Educational Support for Vulnerable Children (Bourses d'étude pour les orphelins et autres enfants vulnérables) PAPA Old Age Support Program (Projet d'appui à la promotion des aînés) PBSF Programme de bourse de sécurité familial PMT Proxy Means Test PREM Poverty Reduction and Economic Management Network PRBC Community-Based Re-adaptation Program (Programme de réadaptation à base communautaire) PRN Nutritional Reinforcement Program (Programme de renforcement nutritionnel) PRP Poverty Reduction Program (Programme d'appui à la mise en œuvre de la Stratégie de Réduction de la Pauvreté) PSNP Productive Safety Net Program (Ethiopia) SENELEC National Electricity Company UNICEF United Nations Children’ Fund WB World Bank WFP World Food Program School Lunch Program (PAM Cantines Scolaires) Vice President : Makhtar Diop Country Director : Vera Songwe Sector Director : Tawhid Nawaz Sector Manager : Stefano Paternostro Task Team Leader : Aline Coudouel / Phillippe George Leite ACKNOWLEDGEMENTS This report benefitted from the inputs of many. The report’s editors are Aline Coudouel, Phillippe George Leite, and Azedine Ouerghi (AFTSW). The report draws from background reports prepared by Julie Van Domelen, Ludovic Subaran, Damien Echevin, Aissata Fall, and Mame Marame Diop (consultants), with logistical and administrative support from Boury Ndiaye and Aminata Ndiaye Bob (WB Senegal Country Office) and Josiane Luchmun (AFTSW). Comments were received on the various background papers from members of the WB Senegal country team including Linda English (lead human development specialist/sector leader), Philip English (country economist), Mamadou Ndione (economist), Djibril Ndoye (economist), Atou Seck (education specialist) and Christophe Lemiere (senior health specialist). Additional comments were received from Theresa Jones (LACSH). Peer reviewers for this report were: Ruslan G. Yemtsov (Lead Economist, HDNSP) and John Blomquist (Lead Economist, SASSP). The technical assistance program and final report would not have been possible without the engagement of many Government officials in Senegal. Ousmane Ka, the head of the Poverty Reduction Coordination Unit in the Ministry of Family, provided leadership to the overall effort. Other agencies involved in the technical assistance program include the PRSP Monitoring Unit and the Direction of Forecasting and Economic Studies in the Ministry of Finance, the National Statistics Agency. The Social Protection Donor Working Group, chaired by UNICEF, participated in reviews and presentations of the various technical reports. This report benefitted from the financial support from:  the Rapid Social Response (RSR) Program, which has been generously supported by the Russian Federation, Norway, The United Kingdom, Sweden and Australia;  Belgium Poverty Reduction Partnership Program. Table of Contents Executive Summary .......................................................................................................................................................1 Chapter 1: Context: Poverty, Vulnerability and Social Protection as an Economic Driver ...........................................8 I. Introduction ...............................................................................................................................................................8 II. Overview of Poverty and Vulnerability in Senegal ...................................................................................................11 III. Social Protection and Safety Nets as a Contributor to Economic Growth ..........................................................16 Chapter 2: Previous Government Response to Shocks and Existing Safety Net Programs .........................................21 I. Government response to shocks in the last 10 years ..............................................................................................21 II. Review of Safety Net Programs in 2011 ...................................................................................................................26 Chapter 3: Looking Forward: Building a National Safety Net System in Senegal ........................................................44 I. Targeting Options for Safety Nets ............................................................................................................................44 II. Institutional Arrangements for a National Social Safety Net System.......................................................................50 III. Towards a National Safety Net System ...............................................................................................................52 IV. A Financial Framework for a Safety Nets in Senegal ...........................................................................................56 V. Conclusions and Next Steps .....................................................................................................................................61 Tables Table 1 : Distribution of disabled individuals by poverty level. ...................................................................................12 Table 2: Frequency of Problems Satisfying Food Needs over the last 12 months .......................................................13 Table 3: Incidence of Self-Reported Shocks (%) ..........................................................................................................13 Table 4: Household-Reported Responses to Shocks (%) .............................................................................................14 Table 5: Price Increases of Selected Energy Products, January 2005–December 2007 ..............................................15 Table 6: Senegal: Subsidies on Basic Goods and Utilities, 2005-2011 .........................................................................22 Table 7: Distribution of Utility Beneficiaries by Consumption Quintile, Poverty Status and Urban-Rural .................24 Table 8: Distribution of utility benefits by consumption quintile, poverty status and urban-rural ............................25 Table 9: Objectives and Type of Benefit for Each Safety Net Program .......................................................................28 Table 10: Safety Net Program Coordination Mechanisms...........................................................................................29 Table 11: Safety Net Beneficiaries by Program and Year, 2009 - 2011 .......................................................................30 Table 12: Geographic Distribution of Safety Net Programs.........................................................................................31 Table 13: Distribution of Beneficiaries of WFP School Lunches, NETS, CSA and OEV by Region .................................32 Table 14: Targeting Methods, Criteria and Information Sources by Program .............................................................33 Table 15: Distribution of Beneficiaries of Social/Safety Net Programs Queried in the ESPS .......................................36 Table 16: Transfer Mechanisms, Actors and Frequency ..............................................................................................37 Table 17: Value of Transfers and Cost per Beneficiary by Program (FCFA) .................................................................38 Table 18: Evaluation Components ...............................................................................................................................39 Table 19: Safety Net Programs Annual Budgets 2009-2011 (FCFA) ............................................................................40 Table 20 and Figure 7: Sources of Safety Net Program Funding, 2009 - 2011 ...........................................................42 Table 21: Typology of Targeting Mechanisms .............................................................................................................45 Table 22: The Principal Determinants of Urban Poverty Based on Observable Household Characteristics ...............46 Table 23: The Principal Determinants of Rural Poverty Based on Observable Household Characteristics .................46 Table 24: Distribution of Safety Net Program Beneficiaries Based on Alternative Targeting Scenarios .....................48 Table 25: Examples of Institutional Arrangements in NETS Cash Transfers and INPS – “Squali Jaboot� ....................51 Table 27: Financial Reference Parameters for Financial Dimensions of a National Safety Net ...................................58 Table 26: Financial Parameters of Safety Net Programs in Senegal and Reference Programs in Africa .....................59 Table 28: Fiscal Space Projections 2012-2016, in FCFA billions ...................................................................................60 Figures Figure 1: Social Protection, Economic Growth and Poverty Reduction – A Schematic View ......................................18 Figure 2: Access to Water and Electricity by Consumption Decile ..............................................................................24 Figure 3: Institutional Framework for Safety Net Programs ........................................................................................29 Figure 4: Targeting Vulnerability within Categorical Targeting, by Program ...............................................................33 Figure 5: Distribution of safety net programs by target age groups ...........................................................................35 Figure 6: Distribution of Program Costs by Component ..............................................................................................41 Table 20 and Figure 7: Sources of Safety Net Program Funding, 2009 - 2011 ...........................................................42 Boxes Box 1: Analytical Underpinnings of the Social Safety Net Assessment........................................................................11 Box 2: Recommendations to Move from Subsidies to a Targeted Safety Net .............................................................25 Box 3: Summary of Targeting Outcomes of Social Programs ......................................................................................36 Box 4: Evaluating the Impact of the nutrition cash transfer program NETS ................................................................39 Box 5: Common Pitfalls in Reforming Safety Net Systems ..........................................................................................54 Executive Summary Objective of the Safety Net Assessment i. In late 2010, the World Bank initiated a technical assistance program in social protection with the Government of Senegal, with a focus on safety nets. The World Bank and Government identified a number of just-in-time technical notes and training sessions tailored to the immediate needs of Government. This Senegal Safety Net Assessment integrates and summarizes the content of these technical notes. Country Context: Poverty, Vulnerability and Social Protection ii. Poverty has remained elevated in Senegal, and has stagnated over the past few years. Poverty rates in Senegal came down from 55.2 percent to 48.3 percent between 2001 and 2005 but barely reduced during the five years after that, and reached 46.7 percent in 2011. A series of shocks affected Senegal in 2006-2011, including poor rains in 2006 and 2007, global food and fuel price shocks in 2008 and floods in 2009. iii. A significant percentage of households remain vulnerable. Household characteristics that are most correlated with poverty include low educational achievement, family size and male-headed households. Poverty is concentrated in rural areas. Extreme poverty, defined as the inability to satisfy nutritional needs, went from 17.2 percent to 15.9 percent between 2001 and 2005, before reaching 14.8 percent in 2012. Vulnerable children, the disabled and elderly without family support tend to be in highly precarious situations. Formal social security coverage only reaches 13 percent of the population, including 6.2 percent covered by a formal pension program, 3 percent receiving social security administration benefits and 3 percent having some form of health insurance. In particular the poor and informal sector workers have little access. iv. At a household level, shocks occur more frequently to poor households. One-third of households report experiencing, at least, one shock over the last twelve months, rising to 40 percent for the poorest two quintiles. The most frequent shock was serious illness or accident, which occurred twice as often in rural areas. Over half of households have no specific coping strategies in response to shocks. Very few report receiving help from the Government or NGOs, relying on family members, savings or going into debt. v. At the country level, significant shocks frequently affect the Senegalese economy with effects on economic growth. Exogenous shocks, such as rising prices of imported goods or the effects of global economic recession, strike the Senegal particularly hard because of its small, open economy, as demonstrated in the financial, food and fuel crises of 2008-2009. Natural disasters also affect certain population groups. At least 5 million Senegalese are exposed to drought risk. Flooding has also affected several regions, with effects on production and infrastructure, as well as loss of household assets. vi. The general conclusion of the Government and key international partners has been to build a better targeted safety net system as a better option than continued reliance on general subsidies. The basis for this argument can be grounded not only in equity concerns but also in direct linkages to economic growth. Social protection contributes to economic growth in multiple ways. Specifically, safety nets can contribute to human capital formation, for example through cash transfers to poor households, as well as the creation of infrastructure, for example through labor intensive public works. Such 1 programs help households build assets. The demand-stimulating effects can help develop local markets. Such interventions can also build social and political cohesion which forms a basis for a more productive society. The Government’s Previous Response to Shocks and Existing Safety Net Programs vii. The Government has had to respond to multiple covariate shocks over the last decade, including, of note, the drought in 2002-2003 and the economic crisis of 2008-2009. In terms of natural disaster, historically the Government of Senegal has historically responded directly to drought with financial support to farmers as well as general assistance to the rural population. These rural support actions reflect a willingness on the part of government to spend resources to respond to critical needs. However using interest rate subsidies and debt forgiveness as a response to weather-related shocks suffers from several drawbacks. Support is often poorly targeted, with subsidies and write-offs benefitting larger rural producer and those able to participate in the formal credit system. Decisions like forgiving all producer debts do not match the differential nature of the shocks which rarely hit all producers equally. viii. The Government responded to the fuel and food price hikes with a series of fiscal measures, including subsidies on basic foodstuffs (rice, wheat, and milk), butane/natural gas and electricity. This ensured a quick and visible action by the Government to respond to growing social unrest and immediate needs, but proved very expensive, absorbing 2.4 percent of GDP, or one-tenth of all spending in 2008. The use of subsidies during the food and fuel crises came with administrative difficulties and economic disincentive effects. More importantly, the bulk of benefits went to the non-poor. For example, only one-third of water subsidy beneficiaries were poor and only 8 percent were in the poorest quintile (poorest 20 percent). Similarly, 31 percent of electricity subsidy beneficiaries were poor and about 7 percent were in the poorest quintile. The strong majority of beneficiaries of both subsidies were urban dwellers. ix. A review of safety net programs identified 12 programs under implementation by the Government in 20111: (1) Food Security Commissariat (Commissariat à la Securité Alimentaire - CSA) provides food aid assistance to vulnerable populations either in response to catastrophes or through rice distribution at public rallies and religious festivals; (2) National Solidarity Fund (Fonds de Solidarité Nationale – FSN) is responsible for providing immediate responses to crisis and emergency situations, including financial, medical and material support; (3) Community-Based Re-adaptation Program (Programme de réadaptation à base communautaire – PRBC) provides social, economic and cultural integration for disabled persons via material support and funding of income generation activities; 1 In addition, a temporary job creation program (public works) called "Food for the Creation of Sustainable Assets" was implemented in partnership with WFP. This program provides food assistance to vulnerable households, with a view to promote community activities contributing to the improvement of their food security. However, this review did not find public work programs that could be characterized as a safety net in terms of targeting poor communities and beneficiaries with wage below the prevailing rates, and transfers timed to address the lean season in rural areas. The Government had implemented some public works programs in rural areas during the 90s and early 2000s through the Agence d’Exécution des Travaux Intérêt Public (AGETIP) in partnership with the WFP. 2 (4) Old Age Support Program (Projet d'appui à la promotion des aînés – PAPA) aims to address the vulnerable elderly (over 60 years) via capacity strengthening, grants and subsidized loans for income generating activities to groups of elderly; (5) National School Lunch Program (Programme d'alimentation scolaire - DCaS) provides school lunches funded through the national budget; (6) WFP School Lunch Program (PAM Cantines Scolaires) supports the national school lunch program by providing hot meals in pre-schools and primary schools located in rural and peri-urban vulnerable areas; (7) Educational Support for Vulnerable Children (Bourses d'étude pour les orphelins et autres enfants vulnérables – OEV) a program through the National HIV-AIDS Council to provide for schooling or professional training to children orphaned or affected by HIV-AIDS and other vulnerable children; (8) Sesame Plan (Plan Sésame) waives health service fees for all persons over 60 years; (9) Poverty Reduction Program (Programme d'appui à la mise en œuvre de la Stratégie de Réduction de la Pauvreté – PRP) supports grants for income generating activities for vulnerable groups, primarily women, the disabled and HIV-AIDS affected populations; (10) a pilot Cash Transfers for Child Nutrition Program (Nutrition ciblée sur l'enfant et transferts sociaux- NETS) entailing cash grants to mothers of vulnerable children under 5 years old to mitigate the negative impacts of food price increases;2 (11) WFP Vouchers for Food Pilot Program (Bons d’Achat – PAM CV) to address food insecurity among vulnerable households due to rising food prices. (12)The Social Protection Initiative for Vulnerable Children (Initiative de protection sociale des enfants vulnérables – IPSEV) Cash grants to households to help them maintain vulnerable children and ensure access to health and education services. x. The twelve safety net programs seek to achieve their objectives using a variety of types of support. There are three main types of benefits of these safety net programs: support to daily existence, nutritional support, and improving access to basic services. This is carried out through monetary transfers (cash grants and loans), food aid and fee waivers for health services. xi. These safety net programs are spread across a number of ministries and agencies. There has been a history of institutional instability which has hampered developing coherent programming, particularly within the social welfare ministries. For at least the last 10 years, there have been frequent reconfigurations of the main social welfare ministries and shifting of departments and programs between ministries. This weakened the ability to effectively design and implement safety net programs. Various institutional coordination mechanisms are in place, though they are often project-specific. xii. Current safety nets afford little effective coverage. About 4 million people are estimated to receive some type of safety net assistance each year. This is equivalent to a little under one-quarter of the national population. However, this grossly overestimates the number of people covered by an effective targeted safety net. Indeed, CSA accounts for about 80 percent of these beneficiaries and school lunches for an additional 17 percent. Neither the CSA nor school lunches screen beneficiaries based on need. Considering only those programs which target and screen vulnerable beneficiaries and for which there is data on number of beneficiaries, (excluding CSA and school lunches), about 100,000 people benefitted. This is mainly from the two cash transfer pilots, the NETS and PAM CV. 2 In selected areas, a partnership between WFP, CLM and the Ministry of Health and Social Action provides nutritional and fortified food to fight moderate acute malnutrition for children aged 6 to 59 months and pregnant and lactating women, and to prevent acute and chronic malnutrition among children aged 6 to 23 months. 3 xiii. Safety net programs use different transfer mechanisms. School lunches and the CSA distribute food directly to the final beneficiaries. In terms of logistics, CSA maintains a network of warehouses and trucks throughout the country. The school lunch program either contracts with the CSA or through private suppliers for delivery to schools depending on the region. In terms of cash transfers, there are several distribution channels. Several programs distribute cash directly from the administration – either from the central Dakar office or through a local office of the program. The other cash transfer programs work through a financial partner, be they commercial banks, micro-finance institutions or postal banks. For food coupons through the PAM CV program, beneficiaries purchase food using their coupons and the intermediaries are reimbursed upon presentation of the coupons at partner banks. xiv. Safety nets in Senegal use a variety of targeting mechanisms, with a predominance of categorical targeting. Categorical is often reinforced by prioritization of certain geographical areas and confirmed through community–based mechanisms. The performance of these targeting systems is mixed. The 2011 ESPS Household Survey included questions on coverage of a range of social programs, including some of the safety net programs in this analysis. Programs cited include: nutritional reinforcement (PRN), youth employment programs (e.g. Offiice Banlieue); agricultural development (e.g. GOANA); elderly health care (Plan Sésame), food aid, educational support (scholarships, etc.), and housing assistance. Some programs were very effective at concentrating on the poorest households, like the PRN and agricultural support programs, while others revealed significant leakage to the non- poor, including educational assistance (like scholarships) and food aid. The elderly health care program benefits more the better off 40 percent of households with beneficiaries concentrated in urban households. xv. In terms of safety net spending, these safety net programs have averaged about 17 billion FCA per year over the last three years, equivalent to 0.27 percent of GDP. The school lunch programs account for over 70 percent of safety net expenditures, reflecting the large coverage. The CSA, despite reaching many beneficiaries, only has an annual budget of about 1 million FCFA because the value per transfer is low. xvi. Safety net funding remains largely dependent on development partner financing. Out of the nine programs with funding information, 62 percent of costs are financed by donors. Local governments account for 7 percent, largely through their participation in the school lunch program. The national budget accounts for 27 percent with the remaining 4 percent from community contributions. The actual spending per beneficiary shows wide variations between programs. The smallest costs per beneficiary are the school lunch program and the food distribution through the CSA. The transfer programs NETS and PAM CV seek a meaningful impact within the tension of an affordable program able to scale-up. xvii. In terms of effectiveness, there is no standardized monitoring of program implementation across safety net interventions. Because each program establishes its monitoring and evaluation plan, some more rigorous than others, this makes it difficult to get a national perspective on coverage and impact of safety net programs. Only four of the programs reported that they engaged in a formal impact evaluation, including the two agencies involved in the national school lunch program as well as the WFP CV and NETS cash transfer. At the time of this report, only the NETS evaluation was available. Impacts identified included a positive impact at the household level in terms of the number and variety of meals reported and a reduction in negative coping strategies. For participants, the cash transfers resulted in increased participation in educational sessions and improved overall reproductive health care. Children experienced better eating patterns, reductions in morbidity and improved vaccination coverage. 4 Looking Forward: Building a National Safety Net System in Senegal xviii. A national safety net system should include a common system that allows each program to target its populations and to register its beneficiaries; an institutional platform that integrates coordination mechanisms at the central regional and local levels; and a packet of interventions adapted to the needs. Reforms of this nature to establish a more effective national safety net will need to be developed with a financially sustainable fiscal framework over the medium-term. xix. In terms of targeting, existing safety nets in Senegal use categorical, geographical and/or community-based targeting. All programs use categorical targeting to identify certain demographic profiles of intended beneficiaries. Furthermore, seven out of twelve restrict program operations to geographical regions deemed poorer. And three programs add a layer of community-based targeting to involve local populations in the identification of beneficiaries. xx. There are opportunities to improve targeting mechanisms among safety net programs. In particular, screening individual household poverty levels could improve targeting. To date, there has been no utilization of proxy means testing, a mechanism popular in many safety net programs around the world. As part of the social protection technical assistance program with the Government, the World Bank provided training in PMT methodologies using several household survey data sets. xxi. Applying a PMT score to existing safety net programs could improve targeting considerably. Simulations of nine safety net programs under different targeting scenarios found that the ‘worst’ targeting outcomes, i.e. the most leakages of program coverage to the non-poor and better-off, are those that only use categorical targeting. Geographical targeting improves outcomes over the base case in all of the simulations. Using a proxy means test further improves targeting outcomes in almost all cases (PAPA urban, IPSEV urban and rural, WFP CV, NETS rural, DCas urban and rural, OEV urban, PRP, PRBC urban, WFP cantines urban and rural). The only cases where PMT does not provide better targeting over geographical targeting were PAPA rural, NETS urban (only slightly), OEV rural, PRBC rural. xxii. One caveat with focusing on improving targeting mechanisms is that it should not distract from the greater issue of the need to increase coverage of safety net programs in general and the need to reform the larger expenditures going to untargeted subsidies at present. Indeed, to reach a satisfactory coverage level, the main constraints are fiscal and institutional capacity to deliver assistance to more people, in addition to the limitations observed in terms of targeting. Institutional arrangements for a national social safety net system xxiii. An adequate institutional framework supporting the creation of a national safety net system will be critical to its success. Senegal does not have a fully developed institutional framework of a national social protection system or an institutional platform sufficient to support a scaled-up, integrated national safety net system, although several key components are in place. Despite institutional instability, dispersion of programs and weak coordination mechanisms, there is much strength in the institutional framework for safety nets in Senegal which the country can build on to create a more comprehensive safety net system. The national social protection strategy provides a unifying view of social protection and a common language to prioritize interventions. There is an important base of existing safety net programs, which has allowed for experimentation with implementation procedures and program-level coordination mechanisms. The Government recently established an Inter-ministerial Committee on Social Protection to promote coordination at the highest level of Government. And, the 5 new government has created the General Delegation for of Social Protection and National Solidarity (Délégation Générale à la Protection Sociale et à la Solidarité Nationale) to provide a focal point in the President’s Office signaling the importance of social protection in the priorities of the Government. xxiv. The core elements of a national safety net system might include development of a common set of tools to be shared across programs. This would include components of a national registry (which will form the basis for the targeting of programs), coordinated budget mechanisms within an overall medium-term expenditure framework, coordinated development of indicators and MIS so that they can be communicated consistently across programs and aggregated at the national level, , and common local coordination mechanisms. The national safety net system should also have the flexibility to respond quickly to the short-term needs resulting from shocks and to scale back down after the period of transitory crisis passes. This includes the contingency planning across all main program areas. A national safety net system would identify overall needs for capacity building in transversal topics, like needs identification among vulnerable groups, targeting, management of transfers, monitoring and evaluation to be shared across programs. Moving from a project or individual program approach to a national safety system will require developing medium-term expenditure framework for the sector that prioritizes expenditures and builds a sustainable funding basis for safety nets. Where development partners provide a significant level of safety net resources, bringing donors into a harmonized, Government-led framework is essential to creating a coherent safety net system. A systems approach reduces fragmentation, promotes harmonization and can enhance both the performance of individual programs and the overall protective, preventive, and promotive functions of safety nets in Senegal. A Financial Framework for Safety Nets in Senegal xxv. Public spending on safety nets and social assistance programs (not including general price subsidies) has historically been low in Senegal. Spending on social assistance and safety nets amounted to about 0.15 percent of GDP in 2004. During drought years of 2002-2003, this rose to 0.5 percent of GDP reflecting emergency food and materials to support rural populations. Moreover, the bulk of that spending went to programs like school lunches or CSA food distributions which have other objectives and/or are not necessarily targeted to the most vulnerable through a safety nets perspective. The programs that have large coverage levels also tend to distribute limited benefit as a share of household consumption, which can dilute potential impact on household well-being. xxvi. Scaling up safety nets will need to be based on a realistic cost framework. The overall costs of a national safety net will depend on the type of program, the level of coverage, the size of the transfer and overall administrative efficiency. Financing a robust safety net would not cost more than typically annual subsidies to basic food stuffs and utilities, and yet would have a far less distortionary and better impact on the poor (for a given cost). xxvii. There are several possible components to funding a national safety net. Fiscal space for a safety net can be created through increasing efficiency and effectiveness of existing social protection expenditures through reallocation from lesser to higher priorities and from less to more effective programs (e.g. from food subsidies to cash transfers). The most significant opportunity is shifting from the use of general price subsidies, particularly in response to economic shocks, to more targeted social protection interventions. Other financing could come from increasing the allocation of domestic resources to social protection. Economic growth has generated some increase in fiscal space in Senegal over the last decade. Nonetheless, development partner funding remains a core part of social protection 6 financing in Senegal. While additional external resources will be important, over time, the Government of Senegal should take responsibility for a growing share of national safety net funding. Conclusions and Next Steps xxviii. Senegal will certainly continue to face internal and external shocks, and extreme poverty remains an important factor. Safety net reforms should allow for increased coverage, greater impact and improved efficiency of safety net expenditures. A national safety net system should include an institutional platform that integrates coordination mechanisms at the central regional and local levels, a common system for targeting and registration of beneficiaries, and a packet of interventions that place an accent on the productive nature of safety nets and reinforce their contribution to economic growth. Reforms of this nature to establish a more effective national safety net will be better able to develop a financially sustainable medium -term fiscal framework. A national safety net of this nature could cost between 1 and 2 percent of GDP depending on scope and funding. A safety net of this magnitude would have a direct impact on poverty reduction and achievement of the Millennium Development Goals. Increased access to schooling and health services, protection and accumulations of assets and avoidance of negative coping strategies will affect the well-being of current Senegalese and generations to come. xxix. Suggested future analytical work to bolster these recommendations include:  Developing a more detailed view of possible implementation mechanisms and approaches to targeting  Analyzing in details the mechanisms for the delivery of safety net services from the cash transfer pilots, and in particular such operational issues as the feasibility of modern fund transfer mechanisms like cell phones;  An in-depth look at the potential phase out of subsidies;  An institutional review of social protection coordination mechanisms with a view of creating an integrated technical and institutional platform rather than the present series of isolated and ad- hoc arrangements; and  Development of a full medium-term expenditure framework for social protection in Senegal once priority programs and coverage targets have been established by the Government. This will require development partners’ buy-in given the share of external resources in social protection. xxx. The Government has recently made steps to expand safety nets in the country, informed by the technical notes from this World Bank TA program. Of note, the President appointed a Délégué General for Social Protection in November 2012 who reports directly to him. His role and responsibilities include setting strategies, coordinating programs, establishing the Caisse Autonome pour la Protection Sociale Universelle (CAPSU) – a social protection funding mechanism- , and launching the Programme National de Bourses de Securité Familiale (PNBSF), a cash transfer program expected to reach 250,000 in 2017. xxxi. In response to this heightened priority on safety nets by the Government, additional efforts will be needed to ensure that Senegal develop an effective institutional framework for a national social safety nets system. It will be critical to move from scattered programs to a system approach. Coordination of policies and actions on the social protection agenda should seek the further harmonization and integration between programs and to support the development of common operational mechanisms underpinning a national safety net system (in particular as they relate to targeting and registering) and of consistent or compatible mechanisms (in particular in terms of the information systems of different programs). 7 Chapter 1: Context: Poverty, Vulnerability and Social Protection as an Economic Driver I. Introduction Senegal has been buffeted by shocks in recent years 1. After a decade of strong economic performance, Senegal suffered from a rapid succession of shocks staring in the late 2000s, some external and some internal, which exposed the country’s vulnerabilities and the Government’s limited capacity to respond effectively to these shocks, in particular in protecting the welfare of the poor. From mid-90s until 2005, Senegal’s GDP grew on average by 5% annually in real terms. The run up in international oil prices, starting in 2007, and the surge in food prices weighed down on Senegal’s open economy. At the same time, Senegal experienced two successive years of agricultural production shortfalls resulting in reduced food availability at a time when prices for imported cereals reached record highs. Further, fiscal slippages in 2008 hurt private growth, especially in the construction and public works sector and the onset of the global recession further dampened the ability to rebound from previous shocks. Floods in the Dakar area constituted an additional shock, albeit one with a fairly narrow impact in time and geographic area. Each of these successive crises had its own time frame, channels of transmission, economic scale and social or regional targets, but taken together, they largely account for Senegal’s weaker performance in the second half of the decade. 2. Poverty had almost stagnated these past few years, from 48.3 percent in 2005 to 46.7 percent in 2011, and a significant percentage of households remain vulnerable. The country has made considerable progress on MDG related targets and saw a 23 percent reduction in malnutrition rates since 2000 after a decade of stagnation. Similarly, child mortality was reduced by 30% and maternal mortality by 21 percent since 1990. However, it is estimated that a 30 percent increase in the price of rice, cooking oil, sugar, wheat, millet, and milk products alone pushed poverty up by six percentage points raising the poverty to 52.7 percent in 2011. Gaps in access to social services and social protection further undermined the ability of poor households to cope with these shocks. 3. In response to the rising costs of fuel and food, the government introduced in 2007 general tax breaks and subsidies on rice and other commodities. These measures proved very expensive (between 3 and 4 percent of GDP) and poorly targeted to the poor. Towards the end of 2008, the government, under severe budget constraints, had no other choice than to lift most of these general subsidies. This experience highlighted to the Government to redouble its efforts to provide effective programs to buffer the most vulnerable from shocks and destitution, with particular interest in developing a national safety net system. The Government had limited policies and programs in place to mount an effective response 4. The social protection system has been under development in Senegal, but progress was insufficient to respond to recent shocks. The country’s first National Social Protection Strategy was developed in 2005 (the National Social Protection Strategy, NSPS, 2005-2015). The National Social Protection Strategy has as its principle objective the adoption of an integrated global vision of social protection that promotes access to risk management by vulnerable groups. The social protection strategy foresaw the diversification and expansion of social protection instruments, including safety net support for vulnerable groups. The social protection strategy was less specific on the exact nature of safety nets to be expanded, as there was little experience at that time in the country. There was little in 8 the way of guidance on priority interventions, implementation structures, program harmonization or institutional arrangements around safety nets. 5. The Senegal Second Poverty Reduction Strategy Paper 2006-2010 (PRSP-II) was adopted by the Government in mid-2006 and made a strong case for strengthening Senegal’s social protection system. The Second pillar of the PRSP-II3 aimed at promoting access to basic social services by a growing share of the population, the third pillar of the PRSP-II emphasized the need to improve the lives of vulnerable groups through specifically targeted interventions, and prescribed actions to ensure that these groups benefit from wealth creation and gain better access to social services. The PRSP-II identified a series of actions towards a more effective system, summarized as follows: (i) to extend the coverage of social protection – including health insurance - to both formal and informal sector workers; (ii) to reform the formal social security system - including pension systems; (iii) to increase the capacity to respond to shocks and risks affecting vulnerable groups; and (iv) to improve targeting, monitoring and evaluation, and payment mechanisms. 6. The PRSP Annual Progress Report (APR) recognized that the performance of the third pillar- protection of vulnerable groups- had been weak with low progress in weaving an effective social safety net, and ambivalence in constituency building efforts. The current social protection system is fragmented and inefficient and does not respond to the specific needs of the various vulnerable groups. The highly expensive and untargeted subsidies proven impossible to maintain during soaring food prices episodes in 2007-8 and humanitarian assistance in times of emergencies is expensive both for Senegal and the international community. Safety net and social assistance programs reach very few, despite efforts at introducing some new pilot approaches, including cash transfers, in response to the crisis. The APR recommended that efforts be stepped up to strengthen safety nets in Senegal, with more appropriate targeting mechanisms in the face of repeated shocks and limited resources.4 Despite challenges in meeting implementation goals, the objectives of social protection seem to be well incorporated in the national discourse in Senegal. World Bank Technical Support Program in Social Protection 7. Creating an effective social safety net is a major challenge in advancing the country’s goal of poverty alleviation. Defining the optimum contour and parameters of such an instrument required further analytical work. The recent crises revealed the sharp exposure of the most vulnerable to income shocks and the need for the Government to embark in sound counter-cyclical policies to protect their consumption—including food and basic social services— and promote their investing in human capital. A paradigm shift from large ineffective subsidies to well-targeted social safety net programs had been initiated and needed to be further brought forward. Further, the revitalization of social safety nets as anti-poverty programs was growing in international acknowledgement. 8. The Government expressed its intent to implement policy actions to improve the effectiveness of social safety nets and to enhance the fiscal sustainability of the Government’s social assistance 3 The four pillars are: (i) wealth creation; (ii) capacity building and the provision of basic social services; (iii) the protection of vulnerable groups; and (iv) the promotion of good governance and participation 4 The APR also emphasizes the need to improve the identification of vulnerable groups: currently, the definition of vulnerability is based on a multiplicity of diverse criteria that include age (children, youth and elderly), gender (women), regions, and other categories (handicapped and refugees). Such a wide-net definition does not allow for effective targeting, nor does it support the development of alternatives to interventions such as broad- based subsidies and tax exemptions that are ineffective, inefficient and financially unsustainable. 9 programs. The authorities sought advisory and technical assistance from the World Bank to ensure that design parameters—targeting, generosity, operational institutions and practices, information systems for operations and impact monitoring, benefit payment arrangements and fiduciary safeguards—are based on best practices. The Government identified an urgent need to enhance the knowledge base and further build constituency by reviewing in detail the existing safety nets programs (including, food and nutrition support programs, public works, school feeding, cash transfers, etc.) and providing suggestions for creating an effective and affordable social safety net system in Senegal. 9. In late 2010, the World Bank initiated a technical assistance program in social protection with the Government of Senegal. The technical assistance program sought to foster evidence-based decision-making in the social safety nets arena and ensure a quality jump in the effectiveness of existing and planned schemes. Its main objective is to develop recommendations on policies and technical and institutional platforms to develop safety net mechanisms to address the main risks and vulnerabilities in Senegal as a component of the country’s economic growth and poverty reduction strategies. 10. World Bank support aimed to provide policy advice as needed, including a series of just-in- time technical notes, and workshops and training in safety net tools, summarized in this report. The reason for this approach was that the Bank sought to capture relevant policy moments, for example preparation of the PRSP and providing relevant policy advice to the new Government elected in early 2012, as well as take advantage of new information as data became available from the ESPS 2011 (with data sets available in early 2012). The objectives were twofold: (a) to bring international experience on safety net policies, practices and analytical tools to the consideration of the Senegalese Government; and (ii) to assist the Government in assessing the current situation in Senegal and options for improving safety nets in the future. 11. The World Bank and Government identified a number of just-in-time technical notes and training sessions to be delivered over the 18 month period (end 2010 to mid 2012) tailored to the immediate needs of Government. The notes were intended to help suggest, based on international experience and country analysis, concrete steps for improving current interventions (e.g. coverage, efficiency, relevance, linkage to productivity and human development, etc.) and the entire system (targeting approaches, payment mechanisms, monitoring and evaluation systems, financial sustainability, institutional arrangement and coordination, etc.). Key outputs which for the analytical underpinnings of this Safety Net Assessment include (see Box 1): a. An analysis of the Government’s response to shocks in recent years to identify the general ability to protect the poor during downturns and adverse events; b. Relatedly, an incidence analysis of subsidies for certain currently key goods ( water, electricity) based on households survey data from both 2006 and 2011. c. To support the PRSP III foundation for the social protection axis, the working groups within the Ministry of Finance expressed interest in a just-in-time knowledge note on the linkages between social protection and economic growth and advice on how to structure social protection interventions so that they are consistent with an economic growth strategy rather than detract from it; d. Review of social protection aspects in the ESPS II survey results. Based on a request from the National Statistics Agency, the Bank provided guidance on developing a chapter on social 10 protection for the final report of the ESPS II and supported a statistical review of vulnerability and social protection using the new survey data; e. Technical guidance on best practice and technical options for targeting mechanisms for safety net programs, including the development of a proxy means test instrument for Senegal. This was initially based on the household survey data available from ESAM II 2005. Working in conjunction with the Ministry of Finance and the National Statistics Agency, the Bank provided technical assistance in the installation of and training in ADEPT software. Once the 2011 household survey data was available, the proxy means test information was updated; and f. A review of the institutional framework of safety nets in Senegal, including the identification of essential institutional components of a national safety net system; Box 1: Analytical Underpinnings of the Social Safety Net Assessment The following inputs serve as the basis for the Senegal Social Safety Net Assessment:  “The Government of Senegal’s Response to Shocks over the Past Decade�, Julie Van Domelen, November 20, 2011  “The Contribution of Social Protection to Economic Growth – Perspectives on Senegal�, Julie Van Domelen, January 2011.  “Review of Safety Net Programs in Senegal�, Aissata Fall, November 2011.  “Issues and Options in Targeting and Social Transfers in Senegal » Damien Echevin, October 2011.  “Institutional Arrangements in Safety Nets in Senegal�, Julie Van Domelen, April 2012.  “Targeting Performance of Water and Electricity Subsides in Senegal�, Damien Echevin, September 2012.  “Profile of Social Protection in Senegal: Analysis of the ESPS 2011�, Mame Marame Diop, September, 2012. 12. The products of the technical support program has three sets of audience: (i) policy makers – particularly Ministries of Finance, Prime Minister’s Office, Social Ministries (women and family, health, education, etc.), and others; (ii) development partners and NGOs; and (iii) the Senegal World Bank Country Team. The goal is to identify areas that need support to facilitate the implementation of reforms (including capacity building) through both investment and budget support. II. Overview of Poverty and Vulnerability in Senegal Poverty reduction has slowed 13. Poverty has remained elevated, with only slight progress since 2005. The per capita annual GDP growth rate slowed from 5 percent between 1995-2005 to 0.8 percent in 2006-2011. Poverty rates in Senegal came down from 55.2 percent to 48.3 percent between 2001 and 2005 but barely reduced during the five years after that, and reached 46.7 percent in 2011. A series of shocks affected welfare during 2006-2011, including poor rains in 2006 and 2007, global food and fuel price shocks in 2008 and floods in 2009. Growth in GDP per capita has been less than 1 percent per annum during the past five years, well below the average for sub-Saharan Africa. 14. During the same period, extreme poverty has only been reduced very slowly. Extreme poverty – defined as the proportion of the population whose total consumption is less than the costs of a food basket that provides minimum calorie requirements – increased from 17.2 percent in 2001 to 15.9 percent in 2005, before reaching 14.8 percent in 2011. 11 15. Poverty remains concentrated in rural areas. In rural areas, 57 percent of the population were poor in 2011, more than twice the rate in urban Dakar at 26 percent. This is largely driven by low productivity agriculture. About 62 percent of people living in households with a head whose main occupation was agriculture were poor, compared to 33 percent for other occupations. Between 2001 and 2011 poverty came down fastest in the capital of Senegal where it decreased by 12 percentage points, compared to 8 and 4 percentage points in rural areas and ‘other urban centers’ respectively. 16. Household characteristics associated with poverty are related to education, place of residence, family size and male-headed households. In 2005, 54 percent of people living in a household whose head had no formal education lived in poverty. In 2011 this proportion had not changed. In contrast, poverty among those living in households whose head had primary education came down from 43 percent in 2005 to 34 percent in 2011. About 83 percent of the poor live in households with a non- educated head, a figure that has not changed over the past 10 years. A larger household size is strongly associated with higher poverty. While 15 percent of households with less than 5 members are poor, this number is 77 percent for households with 20 members or more. About a quarter of Senegalese live in a household that is headed by a woman. These households tend to be smaller and thus less poor. Certain vulnerable households and individuals face even greater challenges 17. Formal social security coverage only reaches 13 percent of the population. This includes 6.2 percent covered by a formal pension, 3 percent receiving social security administration benefits and 3 percent having some form of health insurance. In particular the poor and informal sector workers have little access. Over 80 percent of IPRES and social security (Caisse de Sécurité Sociale)_beneficiaries are non-poor. Even health mutuals overwhelmingly serve the non-poor, who account for three-quarters of the self-reported membership. 18. Vulnerable children, the disabled and elderly without family support tend to be in highly precarious situations. Beyond poverty, certain groups face greater risks than the general population. Table 1 presents estimates of this vulnerable population. Of note, there are an estimated 238,000 disabled persons in the country (defined as those with a disability that prevents them from working), of whom nearly half are under the national poverty line. Almost half of the elderly are poor and tend to be highly vulnerable on family support. Vulnerable children, defined as those with a disability, in early marriage, uneducated and poor account to about 3.4 million children (about 18,000 with disability, 39,000 in early marriage, 1 million out of school, and 2.9 million in poverty). Table 1 : Distribution of disabled individuals by poverty level. Non poor Poor Total Poverty incidence Children 10,139 7,759 17,898 43.4 Youth 29,822 26,700 56,522 47.2 Adults 24,171 27,638 51,809 53.3 Elderly 53,796 58,075 111,871 51.9 Total 117,928 120,172 238,100 50.5 Source : own calculations, based on ESPS 2011. 19. Food insecurity is another important factor in household vulnerability. According to the DHS/MICS 2011 survey, 27 percent of children under five suffer from chronic malnutrition and 10 percent from severe malnutrition. Chronic malnutrition is higher in the rural areas (31 percent) than in urban areas (19 percent). In terms of self-reported difficulties obtaining food, the 2011 ESPS found that 12 poor households respond more often as having ‘always’ or ‘often’ difficulties satisfying household food needs. The highest rate was among the urban poor at 32.7 percent (Table 2). Table 2: Frequency of Problems Satisfying Food Needs over the last 12 months (%) Non-Poor Poor Freqency Urban Rural Total Urban Rural Total Never 42,6 27,1 36,2 18,6 18,1 18,3 Rarely 25,8 25,0 25,5 19,0 20,2 19,8 Sometimes 19,3 26,9 22,4 29,7 32,6 31,6 Often 11,1 18,3 14,1 24,2 24,1 24,1 Always 1,2 2,8 1,8 8,5 5,1 6,2 Total 100,0 100,0 100,0 100,0 100,0 100,0 Source : Diop (2012) based on ESPS data 2011. Shocks occur more frequently to poor households 20. One-third of households report experiencing at least one shock over the last twelve months (Table 3). However, poorer households experienced higher rates of shocks, about 40 percent for the poorest two quintiles. This drops to 16.5 percent for the wealthiest quintile. Shocks occur more frequently to rural households where 43 percent of households report experiencing at least one shock in the last year compared with 20 percent for urban households. Overall, the most frequent shock was a serious illness or accident affecting 11 percent of households. This type of shock occurred almost twice as often in rural areas than urban areas. For the very poorest quintile, the greatest incidence of shocks were loss of livestock and loss of harvest, Other frequent types of shocks include floods, insects and theft, all of which affect poorer households more frequently. Table 3: Incidence of Self-Reported Shocks (%) Quintile Zone of Residence Type of shock Total 1 2 3 4 5 Urban Rural Death of family support 5.0 6.3 6.5 5.0 4.2 3.1 4.0 6.0 Serious illness/accident 11.5 11.9 15.0 14.4 10.1 6.2 8.1 14.8 Loss of employment 3.1 1.7 4.1 3.3 1.9 4.3 4.1 2.0 Failure of family enterprise 0.7 0.7 0.4 0.8 1.8 0.0 1.0 0.5 Loss of harvest (to fire, drought, flood, etc.) 7.4 13.8 9.7 8.2 4.1 1.3 1.0 13.6 Loss of livestock (to fire, disease, theft, etc.) 9.2 14.9 13.9 10 5.6 1.5 0.9 17.2 Significant loss of income (temporary layoff etc.) 2.0 1.6 2.2 3.2 1.4 1.8 2.0 2.1 Partial or full loss of housing from fire, floods etc. 2.8 4.0 3.4 2.9 2.4 1.1 1.6 3.9 Loss of main means of production 0.9 1.1 1.1 0.9 0.8 0.4 0.4 1.4 At least one of these shocks 31.9 40.2 40.4 36 26.5 16.5 20.1 43.3 13 Source : Echevin (2012) based on ESPS data 2011. Households adopt various coping strategies in response to these shocks 21. Over half of households have no specific coping strategies in response to shocks . Per Table 4, almost one-quarter of households report tapping into savings in response to a shock, specifically in cases of health shocks (illness or death) and business failure. Another important response is sale of assets, which can lock households into long-term poverty through depletion of assets. Family support, whether from within the country or abroad, is present in 27 percent of shock responses. Other sources of support are minimally present. NGO support was provided in 2 percent of the cases and Government aide was reported by only 1 percent of households. This profile of response to shocks highlights the essential vulnerability of households to shocks where most have no formal coping strategies and those that do tend to rely on assets and savings, which are less available to the poor. Given that loss of livestock and harvest were the most frequently cited type of shock to the poorest households and they are also the types of shocks for which there is the least amount of coping strategies and assistance, this can have serious repercussions for short-term negative coping strategies and the perpetuation of long-term poverty traps. Table 4: Household-Reported Responses to Shocks (%) Aid Aid NGO, Sale from Aid Govt Utilization from No Type of shock CBO of Borrowing Family from Aid of Savings Family Strategy Aid Assets in the Friends Abroad country Death of family support 0.9 2.3 12.5 24.9 8.8 31.2 14.0 17.1 38.7 Serious illness/accident 1.2 3.4 25.9 36.3 12.5 27.7 16.4 15.8 18.9 Loss of employment 0.1 0.5 12.2 19.2 7.7 13.9 4.7 11.1 60.4 Failure of a family enterprise 0.0 0.0 26.8 30.7 27.5 1.5 0.4 6.2 36.3 Loss of harvest from fire, flood, insects, etc. 0.7 0.8 7.0 6.3 7.5 5.2 2.3 2.2 77.6 Loss of livestock from a fire, flood, pests, theft etc. 0.3 0.5 7.5 6.1 2.2 1.2 0.3 0.9 82.0 3.4 0.0 3.5 11.3 9.1 14.8 3.9 12.0 62.3 Significant loss of income (temporary layoff etc.) 2.1 3.0 8.3 12.5 2.9 3.7 1.7 2.7 66.4 Partial or full loss of housing from fire, floods etc. 0.0 1.2 0.4 16.3 1.7 0.0 0.0 8.4 57.2 Loss of main means of production Source : Echevin (2012) based on ESPS data 2011. The Senegalese economy and households are vulnerable to exogenous shocks 22. Significant shocks frequently affect the Senegalese economy with effects on its economic growth trajectory. Exogenous shocks, such as rising prices of imported goods or the effects of global economic recession, strike the Senegal particularly hard because of its small, open economy, as demonstrated in the financial, food and fuel crises of 2008-2009. Senegal imports all its oil (which powers most of its electricity), and 80 and 100 percent of its rice and wheat consumption, respectively. In 2007-2008, the price of rice in local markets tripled while grain prices increased by 50 percent and prices of other staples like sugar wheat and milk products rose an average of 30 percent. Increases in the price of fuels, as presented in Table 5, were significant as well, with higher increases on those fuels upon which poor households rely, like butane gas. 14 Table 5: Price Increases of Selected Energy Products, January 2005–December 2007 Product Percent Increase Lamp oil (kerosene) 70.9% Gasoline and diesel 42.7% Butane 66.8% Electricity 21.9% Source: Senegal: Selected Issues, IMF Country Report No. 08/221, July 2008. 23. Macroeconomic effects of these price increases were substantial. The onset of the global recession in 2008 and its deepening in 2009 further prevented a rapid rebound from previous shocks. A widening current account deficit and fiscal slippages in 2008 led to a slowdown in private growth, especially in the construction and public works sector. Real GDP growth fell to 2.2 percent in 2009. These price increases affected businesses both directly through increased outlays on fuel as well as indirectly through the general inflationary effects of these price increases. With Senegal’s dependence on petroleum products for electricity generation, these input price hikes caused financial strain on the national electricity company, SENELEC. Senegal’s GDP growth was hindered by frequent electricity outages which caused a slowdown of economic and manufacturing activities. According to local reports, the outages contributed to the closure of many small and medium-sized enterprises in the food processing, textile and tourism sectors. Larger companies reported declines in output averaging 30 percent (US Government 2009). 24. In terms of endogenous shocks, environmental conditions are precarious given that Senegal is a Sahelian county in which 60 percent of the population is engaged in agriculture, with groundnuts as the principle product. Rural regions are highly vulnerable to variations in rainfall from one year to the next. Significant rainfall shocks arrive on average every four to five years. Rainfall shortages cause significant reductions in agricultural harvests and, thus, rural revenues. Rainfall variability creates substantial short-term risk to farmers and has significant impacts on rural incomes, with often devastating results for the poorest rural households. For example, as a result of drought in 2002, annual ground nut production went from 1 million metric tons in 2001 to 300,000 metric tons (IMF 2005). The average cash flow to groundnut farmers fell from 200,000 FCFA in 2001 to 60,000 FCFA during that year. Overall, at least 5 million Senegalese are exposed to drought risk. Flooding has also affected several regions, with effects on production and infrastructure, as well as loss of household assets. 25. In addition to these historical short-term swings in rainfall, there has been a downward trend on rainfall in Senegal over time. From 1960 to 2000 average annual rainfall decreased on a steady trend line from 650 to 600 millimeters while average annual groundnut production declined from 1 million metric tons to 750,000metric tons. The long-term effects of climate change are expected to exacerbate these rainfall trends. The risk of drought will continue to be one of the main sources of vulnerability for rural households for the foreseeable future as well as of the biggest internal risks. Safety nets are needed for both the chronic as well as the transient poor 26. Senegal’s poverty and vulnerability profile points to the need for safety nets both to address the chronic poor as well as those affected by transient shocks. Senegalese households are exposed to a wide range of shocks on an individual basis as well as covariate shocks from natural disasters and economic events that affect large groups of populations. Safety nets should be scalable to respond to these transient needs while ensuring a minimum social protection for chronic poor and vulnerable populations, which represent a significant portion of the Senegalese population. 15 III. Social Protection and Safety Nets as a Contributor to Economic Growth Do other social protection measures offer better alternatives to general subsidies? 27. Social protection measures are an essential part of life, given the inevitable risk and vulnerabilities of human existence. Social protection can be defined as private (both formal and informal) and public initiatives that connect men and women to labor markets, reduce people’s exposure to risks, and enhance their capacity to protect themselves against hazards and loss of income that threaten their present and future wellbeing. Social protection measures typically include labor market policies and programs, social insurance (pensions, unemployment and health insurance, etc.), social assistance and safety nets, as well as targeted interventions to promote access to services by poor communities and households. 28. The importance of the protective role of social protection interventions is widely recognized. Family networks, community and religious associations, private insurance companies and public sector social protection policies and programs all play a role in helping households manage risk and respond to the inevitable and often random shocks of life. Because of its role in protecting vulnerable groups, social protection is frequently viewed as primarily a redistributive instrument. The effect on societal equity, whether in poverty or income or better access to services and jobs, is seen as a fundamental contribution of social protection. However, there is some question of whether this comes at the expense of economic growth, in the age-old debate of economic growth versus equity. Resources that could have been used to promote economic growth, like investments in infrastructure or market development, are instead used for what may be viewed as ‘hand-outs’ and hence are viewed as an opportunity cost. There is a growing recognition that social protection is a direct contributor to economic growth 29. The evidence is mounting that social protection can act as a direct contributor to economic growth. Social protection includes a broad variety of programs and policies. Each may contribute to economic growth in multiple ways. The general avenues whereby social protection can contribute to economic growth are provided in the schematic in Figure 1. With specific reference to safety nets, there are several links to economic growth: a) Increased human capital and productivity. In economic growth models, human capital is generally considered a factor of production in its own right and therefore makes a direct contribution to economic growth (Mankiw et al 1996). Conditional cash transfers have been found to increase school enrollment and expand use of health services like vaccinations and prenatal care in almost all countries where they have been introduced. Multiple evaluations, including using experimental design methods, have shown the impact of school feeding programs on improved enrolment. Safety nets can also protect households from erosion to their human capital by providing resources that mean the households will not have to turn to negative coping strategies, like pulling children out of school or eating less nutritious food. b) Improved functioning of the labor market, including better labor productivity and avoidance of exploitative conditions. For example, if not addressed, child labor leads to lower education attainment and, for girls, an earlier age of marriage, all of which create risks for future poverty and a drag on economic participation rates (Beegle et al 2009). Conditional cash transfers have been shown to reduce the occurrence and intensity (number of hours) of child labor. 16 c) Increasing entrepreneurial activities (less risk aversion). Studies from India and Africa show that because poor households deploy their assets more conservatively than wealthy households, their return on assets is 25–50 percent lower (Alderman and Hoddinott 2008). The extra financial support and sense of surety from predictable safety nets as well as the risk pooling features of social insurance can help the poor take on more investment risk, usually producing higher returns. Risk mitigation schemes such as insurance programs or safety nets, can raise growth rates and reduce the volatility of growth. For example, evidence from the Employment Guarantee Scheme in Maharashtra, India shows that the program has encouraged farmers to take greater risks by planting higher-yielding crop varieties. d) Accumulation (and protection) of assets. Asset accumulation is an important basis for economic growth. A minimum stock of assets is necessary to accumulate wealth over time. To the extent that social protection interventions can help the poor accumulate tangible assets and not just cover today’s consumption, economic growth prospects are enhanced. Not only do assets become the basis for future investments, they serve as a household hedge against shocks. Livestock has played this role traditionally in many parts of Africa. Cash transfers and income from public works programs can help households build their asset base. For instance, Ethiopia’s Productive Safety Net Program significantly lowered the distressed sale of livestock among beneficiaries affected by drought compared with control groups, and public works participant households experienced a net growth in livestock holdings of 0.28 Tropical Livestock Units. 17 Figure 1: Social Protection, Economic Growth and Poverty Reduction – A Schematic View Increases human capital and productivity (e.g. increased school enrollment, reduced malnutrition via CCTs) Improves functioning of the labor market (e.g. unemployment insurance facilitates job changes and active labor market programs increase ‘employability’ of target groups) Social Increases entrepreneurial activities Protection: (e.g. by reducing the cost of downside risk) Economi c Growth Labor Accumulates and protects assets Poverty Markets (e.g. through avoidance of distressed sales of assets) Reduction Risk, vulner- Social Deepens capital markets ability insurance (e.g. pension funds provide capital to stock and bond markets) and Social Stimulates aggregate demand and local markets shocks assistance (e.g. safety nets provide countercyclical spending during downturns) and safety nets Promotes social and political cohesion, enables reform (e.g. by addressing worker dislocation through retraining) Targeted support to Creates economic growth-enhancing infrastructure access (e.g. public works create roads, social and water conservation… services Makes society more equitable (e.g. by avoiding destitution and longer-term poverty traps) Redistrib -ution/ equity 18 e) Stimulates demand and develops local markets. Developed countries have long used counter- cyclical public spending on infrastructure and services as an effective tool to provide the foundation for rapid recovery and job creation. Spending on safety nets is often increased during economic downturns not only to protect the poor but to bolster aggregate demand and protect economic recovery and growth. A recent IMF study shows that on average, for all economies, a discretionary stimulus package equivalent to 1 percent of country's GDP is associated with GDP increases of about 0.1 to 0.2 percent above the amount of the initial investment (IMF 2008). Global experience shows that countries with effective safety nets, which strategically target the poorest, are most successful in responding to crisis. Local effects of social protection spending can be observed as a result of cash transfers or cash-based public works, This is done both by injecting cash into local markets and by strengthening the purchasing power of poor households, which tend to buy local, domestically manufactured goods (OECD 2009). f) Facilitates infrastructure development. There is a well-documented literature on the relationship between investment in infrastructure and growth (World Bank 1994). Infrastructure boosts the economy in two ways – by providing direct employment and purchase of goods and services during the construction phase and a second round of impacts resulting in increased economic activities (transport, use of services etc.) which further boost employment and output. According to one study, the difference in effective use of infrastructure resources explains almost 25 percent of the growth differential between Africa and East Asia, and more than 40 percent of the growth differential between low-growth and high-growth countries (Hulten 1996). Safety nets that include labor-intensive public works create infrastructure as a mechanism for generating temporary employment with productive benefits with the added benefit of building community assets and services among underserved populations. Ethiopia’s Productive Safety Net Program financed the rehabilitation of over 167,000 hectares of land through area closures, 275,000 kilometers of stone and soil bund embankments, and the planting of almost 900 million seedlings, all of which will contribute to mitigating the effects of future droughts. g) Builds social and political cohesion. Social protection measures help underpin society’s social and political structures. Social protection provides greater equity and opportunity to vulnerable and marginalized elements of society and so reinforces the basic underlying social contract that provides greater stability to the system. In times of economic shocks, safety nets provide visible support to affected populations, an important element in preserving political and social stability. In Indonesia, a series of difficult economic reforms was supported through provision of social safety nets, which helped build legitimacy for the reforms (Sumarto 2007). h) Make other sectors more efficient and effective. To the extent that education, health and agriculture contribute to economic growth, safety measures can enhance their outcomes. For example, cash transfers facilitate households’ accessing health and education. In Ethiopia, there were important interaction effects between a public works program and an agricultural inputs program. The income provided through the public works employment did not directly impact farmers’ output, but neither did the agricultural inputs program. However, when a household received both, productivity increased by more than 200 kg/hectare (Gilligan et al 2009). i) Social protection can also contribute to growth indirectly through its effect on equity. There have long been debates on whether inequality is either good or bad for economic growth. A World Bank study concluded that ‘there is no intrinsic trade-off between long-run aggregate 19 growth and societal equity� (Bruno et al 1996). Indeed, inequality may negatively impact on economic growth if it leads to political instability, capital flight or increased economic volatility. 30. In addition to these potential positive effects on economic growth, safety nets can reduce chronic poverty. This is done directly by providing households with much needed cash or in-kind resources or by facilitating their access to higher-paying, more productive jobs. Scaled up social protection programs can have an impact on national poverty indicators. Over the last decade, Brazil has experienced a remarkably rapid improvement in income distribution driven mostly by the reduction of extreme poverty. This improvement in income inequality can be explained in large part by expansions in social protection programs in Brazil. In particular, studies have found that Bolsa Familia, a conditional cash transfer program was responsible for twenty percent of the decline in inequality (Soares and Ribas 2006). The potential for economic growth-enhancing safety nets in Senegal 31. The challenge for Senegal will be to design and implement economic-growth enhancing safety net interventions. Given the empirical evidence globally on how and under which circumstances social protection can contribute to economic growth, the following elements can provide greater economic growth contribution from social protection policies and programs in Senegal: a) Using social protection interventions to promote human capital development. Tying cash transfers or food distribution to the use of health services and school enrollments can promote economic growth in the long-term and help Senegal achieve the MDGs in these sectors. b) Creating basic infrastructure and increased agricultural productivity for poor communities through public works. The rural areas still suffer from a combination of reliance on low productivity agriculture, heightened vulnerability during the lean season, and a dearth of basic infrastructure in the poorest regions and departments. Well-designed productive safety net- type operations, twinned with other service packages, can engage local authorities and communities in creating the necessary conditions to protect rural populations from destitution while building the basic household and community assets necessary for economic growth. 20 Chapter 2: Previous Government Response to Shocks and Existing Safety Net Programs I. Government response to shocks in the last 10 years5 32. The last decade has shown how frequently large-scale shocks occur to the Senegalese economy and the limited range of Government response available to effectively help households cope with the negative effects. The Government has had to respond to multiple covariate shocks over the last decade. This section reviews two experiences: the drought in 2002-2003 and the economic crisis of 2008-2009. Responding to natural disaster and drought: the example of 2002 33. The Government of Senegal has historically responded directly to drought with financial support to farmers as well as general assistance to the rural population. A series of financial mechanisms were put in place in the late 1990s to mitigate and cope with the risks to agriculture as well as ensure an adequate flow of credit.6 In 2001-2004, payments were made to producers affected by floods, pest infestations, and to clear the unpaid debt of producers for the agricultural season in 2002 and 2003 due to globally insufficient rainfall. The fiscal costs of these responses to agricultural shocks rose to 0.2 percent of GDP during this period (counting only the expenditures from these funds). It is not possible to discern how many farmers directly benefitted from the support of the agricultural security funds. 34. These rural support actions reflect a willingness on the part of government to spend resources to respond to critical needs, however using interest rate subsidies and debt forgiveness as a response to weather-related shocks suffers from several drawbacks: a. Support is often poorly targeted, with larger subsidies and write-offs to the larger rural producer and those able to participate in the formal credit system. The limited recourse to formal credit for many rural families limits effectiveness of this measure as it cannot be applied to informal credit arrangements. b. The rules for accessing the various funds are not very clear. c. Providing loan write-offs undermines the culture of repayment, creating the expectation that losses will be compensated for by the Government. d. Decisions like forgiving all producer debts do not match the differential nature of the shocks which rarely hit all producers equally. 5 Summary of Government response in this section drawn largely from World Bank “Managing Rural Risk in Senegal� 2006 unless otherwise noted. 6 The agricultural security funds are comprised of: (a) Fonds de Bonification was created to improve access of rural producers to credit by reducing financing charges. The Fonds de Bonification finances the difference between the interest charged by commercial banks and the interest rate the Government limits to farmers (7.5 percent). The Government pays the spread to the national agricultural bank, the Caisse Nationale du Crédit Agricole du Senegal (CNCAS), which has operating since 1984 and is the largest source of rural finance; (b) The Fonds de Garantie (FG) reimburses delinquent loans to CNCAS up to 75 percent for agriculture and 50 percent for livestock; and (c) Fonds de Calamité (FC) helps rural producers cope with natural disasters, allowing them to repay loans and continue their agricultural activities, either by restoring creditworthiness with CNCAS or to finance supplies necessary to respond to shocks. 21 e. Treatment across farmers may be uneven. For example, with elimination of rural producer debts following the 2002 drought, by the time payments were made, many producers had already reimbursed their debts often by liquidating assets. 7 f. Interest rate subsidies under-price rural credit which undermines the financial sustainability of the rural finance sector, particularly with the expansion of new entrants to the market, most notably micro-finance institutions starting in the mid-2000s, but also commercial banks, NGO revolving loan funds, etc.; and g. Multiplicity of instruments adds complexity to the system as a whole and reduces probability that the poorest producers benefit. Using subsidies to respond to the economic shocks of 2007-20098 35. The lead-up to the economic crises of 2007-2009 created underlying conditions of vulnerability, including internal shocks due to poor rains in 2006 and 2007 that exacerbated the impact of the rapid increase in food and fuel prices starting in 2007. Moreover, troubles with the internal efficiency of the national electricity utility were present in the years leading up to the fuel price increases, with significant fiscal resources used to support the company in 2006. Against the backdrop of these strains, the global economic recession in 2008-2009 led to a significant drop in household purchasing power and difficulties in affording basic consumer products and services. 36. The Government responded to the fuel and food price hikes with a series of fiscal measures, including subsidies on basic foodstuffs (rice, wheat, milk), butane/natural gas and electricity. This ensured a quick and visible action by the Government to respond to growing social unrest and immediate needs, but proved very expensive, absorbing 2.4 percent of GDP, or one-tenth of all spending in 2008. Many of these built on existing subsidy arrangements (Table 6). Total price subsidies on basic consumer goods were on the order of 162 billion in the crisis years of 2007-2008. Food subsidies were 21 billion FCFA in 2007 and 46 billion FCFA in 2008.9 Food subsidies were placed on the largest value imported products: rice, wheat and milk. For rice subsidies, the Government pays the subsidy to several big importers with administrative price controls applied to retailers. Table 6: Senegal: Subsidies on Basic Goods and Utilities, 2005-2011 In FCFA billions 2005 2006 2007 2008 2009 2010 2011 Transfers and subsidies 165 308 287 333 286 259 331 Of which - Société Africaine de Raffinage and other LGP producers 14 66 55 69 33 0 15 - SENELEC 12 86 0 30 30 0 124 - food subsidies 0 0 21 46 0 0 0 As % of GDP: Subsidies on basic consumer goods 0.6% 3.1% 1.4% 2.4% 1.0% 0.0% 2.1% Source: International Monetary Fund. 7 Credit Mutuel de Senegal estimated that 60% of agricultural loans had already been paid back at the time debt cancellation, as reported in World Bank 2005 Senegal: Managing Rural Risk. 8 In addition to direct subsidies to stabilize prices, the Government suspended VAT and customs duties on a number of consumer products in mid-2007 at a cost of 0.5 percent of GDP. 9 This does not include historic protection of the sugar and vegetable oil industry through administered prices. 22 37. However, it is difficult to control actual prices paid in the market. Retailers that already had stock bought at the higher price were reticent to absorb losses. The market was further destabilized with delays in payments to importers. In the case of milk subsidies, payments were only made to importers. This policy was met with complaints by local dairy suppliers, who supply a little under half of the domestic market. Benefits were concentrated in urban areas were the consumption of imported powdered milk is high. Wheat subsidies were also paid to importers, but there is no significant domestic wheat production. Retail prices were then fixed at the point of sale. 38. Subsidies on LPG gas were already in place but were significantly increased during the fuel price crisis, from FCFA 4 billion in 2004 and FCFA 14 billion in 2005 to FCFA 66 billion in 2006, FCFA 55 billion in 2007, and FCFA 69 billion in 2008, dropping to FCFA 33 billion in 2009 and none in 2010, before climbing back to 15 billion in 2011. Previous subsidies were largely justified on environmental terms as a measure to decrease deforestation and encourage the lowest income segments to shift from charcoal and wood. Broader social protection arguments were used to justify the increases in response to increasing world prices of fuel. The Government funds subsidies though the African Refining Company, a company with majority State ownership, through a system of administered prices through its system of distributors. 39. The Government has a long history of financial support to SENELEC, including recapitalization and debt restructuring most recently in mid-2000s.10 The government controls the administered price of electricity provided through this parastatal.11 There has typically been incomplete pass-through of fuel costs to consumers, which has caused financial difficulties for the company, leading to frequent outages and service provision problems. Subsidies to SENELEC were paid to tamper upward pressure on electricity prices from the global fuel price increases, amounting to FCFA 30 billion in 2008 and in 2009. The subsidies did not fully prevent a rise in electricity prices, which increased by 6 percent in late 2007 and 18 percent in mid-2008. Subsidies reached a record level of 124 billion FCFA in 2011. 40. The use of subsidies during the food and fuel crises came with administrative difficulties and economic disincentive effects. The use of subsidies is often viewed as a relatively easy to implement in the short-term method of buffering the population from shocks. However, in practice there are implementation challenges in using subsidies. Subsidies to importers and distributors, for example of food and fuel, only have a welfare effect if prices changes are fully passed through to consumers. This requires significant monitoring and enforcement. In addition, subsidies introduce market distortions that can reduce incentives for domestic supply responses or undercuts nascent markets for local suppliers when they are only applied to imports. Moreover, difficulties in timely payments caused market disturbances and uncertainty. Who benefits from the price subsidies? 41. The level of subsidy support is a function of the size of the subsidy and the share of household consumption of the subsidized item. In terms of electricity and water, access is a function of income, with lower access among lower income groups, though this has improved slightly over the last 10 years (Figure 2). This disparity in access has significant consequences for the performance of general price subsidies on utilities as a means of protecting the poor during shocks. 10 SENELEC’s 2006 losses alone required an injection of budgetary subsidies equivalent to 1.8 percent of GDP. 11 Prices are set by the Commission de Régulation using a price cap method, with a 5 year tariff planning horizon. 23 Figure 2: Access to Water and Electricity by Consumption Decile Source : Echevin (2012). 42. Overall, the analysis of the impact of current subsidies and other policy measures to offset energy price increases shows that they entail high leakage of benefits to the better-off. Analysis done in 2008 by the IMF – at the beginning of the crisis - found that, with butane the 40 poorest percent of the population gain only 19 percent of the total improvement in welfare while the richest 40 percent gains 61 percent. Performance is better with kerosene, but the top two quintiles still receive 33 percent of the benefit and since the weight of kerosene consumption in the household budget is very small, the impact, though better-targeted, is slight. Food subsidies were generally better targeted. Rice is an important component of household food consumption, especially for the bottom 40 percent of households and in rural areas. For powdered milk and bread, the richer groups of the population gained most.12 43. Data from the 2011 ESPS confirms this trend. In terms of distribution of beneficiaries, the poorest households have less access to water and electricity. As a result, only one-third of water subsidy beneficiaries were poor and only 8 percent were in the poorest quintile (poorest 20 percent). Similarly, 31 percent of electricity subsidy beneficiaries were poor and about 7 percent were in the poorest quintile (Table 7). The strong majority of beneficiaries of both subsidies were urban dwellers. Results were even more skewed for the amount of benefit received because the better off tend to consume more. As presented in Table 8, only 23 percent of water subsidy benefits and 14 percent of electricity benefits accrue to poor households. Only 5 percent of water benefits and 2 percent of electricity subsidy benefits went to the poorest quintile. Half of water subsidy benefits and two-thirds of electricity benefits were enjoyed by the wealthiest quintile. Table 7: Distribution of Utility Beneficiaries by Consumption Quintile, Poverty Status and Urban-Rural % Consumption Quintiles Poverty Status Zone of Residence Total Q1 Q2 Q3 Q4 Q5 Poor Non-Poor Urban Rural Water 100.0 8.3 14.2 18.1 25.9 33.4 33.6 66.4 68.3 31.7 Electricity 100.0 6.6 11.9 17.4 28.3 35.9 31.2 68.8 74.8 25.2 Source : Echevin (2012). 12 Incidence analysis drawn from IMF Country Report No. 08/221 “Senegal: Selected Issues�, July 2008, IMF. 24 Table 8: Distribution of utility benefits by consumption quintile, poverty status and urban-rural Consumption Quintiles Poverty Status Zone of Residence % Total Q1 Q2 Q3 Q4 Q5 Poor Non-Poor Urban Rural Water 100.0 4.9 9.0 13.0 23.4 49.7 22.5 77.5 77.2 22.8 Electricity 100.0 2.0 4.4 8.7 19.4 65.4 13.9 86.1 85.7 14.3 Source : Echevin (2012). The need for alternatives to subsidies 44. The recourse to subsidies does show a willingness to respond to immediate needs of the population, but limited tools in place that could be more impactful. The Government has committed significant resources to address the needs of vulnerable populations in response to shocks. Subsidies, for all of their drawbacks, are quick to put in place and provide a very visible action by the Government in addressing social needs. In the heat of a crisis, they are often viewed as one of the few viable short-term options. 45. The general conclusion of the Government and key international partners has been to build a better targeted safety net system as a better option than continued reliance on general subsidies (Box 2). Policy reforms to shift from general subsidies to targeted safety nets in response to short–term shocks can be an important part of developing a more efficient and effective national social protection system. However, this shift will require a nuanced approach and further reflection. For example, 60 percent of the population depends on agriculture, while crop and livestock loss are the most frequent shocks, and there have been significant price increases in fuel and staple commodities many of which are imported. Moving immediately away from subsidies to cash transfers may make sense if retail and import markets are developed, but if not some transition period may be needed. Box 2: Recommendations to Move from Subsidies to a Targeted Safety Net Over the last ten years, there is a growing consensus on the need to rethink recourse to general subsidies as a way of responding to shocks in Senegal:  In 2005, in analyzing the response to drought using the agricultural security funds, the IMF concluded that: “A safety net is needed to provide assistance to farmers in the event of a serious shock… A safety net based on price subsidization, although easy to administer, would have drawbacks…A more efficient social safety net program would explicitly target poor farmers for compensation in response to a severe shock. In the short run, the authorities should use budgetary transfers directed to the poorest farmers to dampen the impact of a price or weather-related shock. Senegal should develop the capacity to target assistance based on household characteristics that are good proxies of overall wealth. In the long-run, conditioning transfers to poor households on acquiring human capital to get access to alternative sources of income would reduce poor farmers’ dependence on groundnut farming.� (Senegal: Selected Issues; IMF Country Report No. 05/155, May 2005).  The IMF drew similar conclusions in their assessment of the response to the global economic crisis of 2007-2009: “A better long-term solution for an effective protection system to address both structural and cyclical poverty in Senegal would be to introduce a conditional cash transfer system. Such a program would be a more permanent way to protect the poor from welfare losses, including food and energy price increases.� (Senegal: Selected Issues, IMF Country Report No. 08/221, July 2008).  This echoes the World Bank recommendation that: “efforts be stepped up to strengthen safety nets and better monitor the impact of existing or future targeting policies. In the short term, now that the food and fuel subsidies have been eliminated, there is a need for an appropriate targeting mechanism, especially since food and fuel prices, despite their 25 recent decline, remain elevated relative to their historic levels. This could ultimately take the form of a conditional cash transfer program and could be based on indicators correlated with poverty and living conditions drawn from household survey data.�(“Senegal: Public Finance Support Credit Program Document�, Report No 48557-SN June 2009).  “A key lesson learnt from the recent wave of crises is to reprioritize prevention. Senegal has been particularly hit by the food and fuel crisis and continues to suffer from the global economic slowdown. In addition, natural disasters such as floods and droughts are exposing both the urban and rural poor to stress and poverty. Highly expensive and untargeted subsidies have proven impossible to maintain during soaring food prices episodes in 2007/8 and humanitarian assistance in times of emergencies is expensive both for Senegal and the international community. Shifting from mitigating the negative impacts of shocks on the population at large to preventing these shocks to erode both the income and human capital of the most vulnerable appears to be a critical step towards inclusive growth. In the past few years, a growing number of developing countries have successfully introduced large-scale cash transfer programs as effective social protection mechanisms to respond to high levels of poverty and vulnerability.� (World Bank, “Senegal: Poverty Reduction Support Credit IV Program Document�, Report No. 52784- SN, April 2010). II. Review of Safety Net Programs in 201113 Senegal has several types of safety net-related programs 46. A review of safety net programs identified 12 programs under implementation by the Government in 2011, with the following objectives: 14 (1) Food Security Commissariat (Commissariat à la Sécurité Alimentaire - CSA) provides food aid assistance to vulnerable populations either in response to catastrophes or through rice distribution at public rallies and religious festivals. The objective of the agency is to ensure the availability, management and monitoring of the country’s food security grain stocks. (2) The National Solidarity Fund (Fond de Solidarité Nationale – FSN) is responsible for providing immediate responses to crisis and emergency situations, including financial, medical and material support. (3) Community-Based Re-adaptation Program (Programme de réadaptation à base communautaire – PRBC) provides social, economic and cultural integration for disabled persons via material support and funding of income generation activities. (4) Old Age Support Program (Projet d'appui à la promotion des aînés – PAPA) aims to address the vulnerable elderly (over 60 years) via capacity strengthening, grants and subsidized loans for income generating activities to groups of elderly. 13 This chapter is derived primarily from “Review of Safety Net Programs in Senegal�, Aissata Fall, November 2011. 14 In addition, a temporary job creation program (public works) called "Food for the Creation of Sustainable Assets" was implemented in partnership with WFP. This program provides food assistance to vulnerable households, with a view to promote community activities contributing to the improvement of their food security. However, this review did not find public work programs that could be characterized as a safety net in terms of targeting poor communities and beneficiaries with wage below the prevailing rates, and transfers timed to address the lean season in rural areas. The Government had implemented some public works programs in rural areas during the 90s and early 2000s through the Agence d’Exécution des Travaux Intérêt Public (AGETIP) in partnership with the WFP. 26 (5) National School Lunch Program (Programme d'alimentation scolaire - DCaS) provides school lunches funded through the national budget. (6) WFP School Lunch Program (PAM Cantines Scolaires) supports the national school lunch program by providing hot meals in pre-schools and primary schools located in rural and peri- urban vulnerable areas. (7) Educational Support for Vulnerable Children (Bourses d'étude pour les orphelins et autres enfants vulnérables – OEV) a program through the National HIV-AIDS Council to provide for schooling or professional training to children orphaned or affected by HIV-AIDS and other vulnerable children. (8) Sesame Plan (Plan Sésame) waives health service fees for all persons over 60 years. (9) Poverty Reduction Program (Programme d'appui à la mise en œuvre de la Stratégie de Réduction de la Pauvreté – PRP) supports grants for income generating activities for vulnerable groups, primarily women, the disabled and HIV-AIDS affected populations. (10)Child Nutrition Program (Nutrition ciblée sur l'enfant et transferts sociaux- NETS) des cash transfers to mothers of vulnerable children under 5 years old to mitigate the negative impacts of food price increases. 15 (11)WFP Vouchers for Food Pilot Program (Bons d’Achat – PAM CV) addresses food insecurity among vulnerable households due to rising food prices. (12)Social Protection Initiative for Vulnerable Children (Initiative de protection sociale des enfants vulnérables – IPSEV) Cash grants to households to help them maintain vulnerable children and ensure access to health and education services. 47. These safety net programs cover a wide range of objectives, such as: Increasing school attainment (enrollment and years completed) for vulnerable children; improving access to health services; maintaining children within families; providing response and resistance to shocks; reducing poverty for instance through income generation efforts; and ensuring social and economic integration for excluded groups. 48. These safety net programs seek to achieve their objectives using a variety of types of support. There are three main types of benefits: support to daily existence, nutritional support, and improving access to basic services. This is carried out through transfers (cash grants and loans), food aid, and fee waivers for health services. Objectives, methods and types of benefits are summarized below in Table 9. Although several public works programs promote the use of labor intensive methods on certain types of infrastructure, this review could find no public works programs that could be characterized as a safety net in terms of targeting poor communities and beneficiaries with below prevailing wage transfers often timed to address the lean season in the rural areas. 15 In selected areas, a partnership between WFP, CLM and the Ministry of Health and Social Action provides nutritional and fortified food to fight moderate acute malnutrition for children aged 6 to 59 months and pregnant and lactating women, and to prevent acute and chronic malnutrition among children aged 6 to 23 months. 27 Table 9: Objectives and Type of Benefit for Each Safety Net Program Program Objective Type of Benefit 1 Commissariat à la Sécurité Alimentaire (CSA) Resistance to shocks Food 2 Fonds de Solidarité Nationale (FSN) Resistance to shocks Cash, materials 3 Programme de Réadaptation à Base Communautaire (PRBC) Social integration Grant, materials 4 Projet d'Appui à la Promotion des Aînés (PAPA) Social integration Loan 5 Direction Cantines Scolaires (DCaS) Access to education Food 6 Programme de Cantines Scolaires du PAM Access to education Food 7 Programme de Bourses d'étude pour les Orphelins et autres Enfants Access to education Cash Vulnérables (OEV) 8 Plan Sésame Access to health services Fee waiver 9 Programme d'appui à la mise en œuvre de la Stratégie de Réduction Poverty reduction Loan de la Pauvreté (PRP) 10 Programme Nutrition ciblée sur l'Enfant et Transferts Sociaux (NETS) Resistance to shocks Cash 11 Programme de bons d'achat du PAM (PAM CV) Resistance to shocks Cash 12 Initiative de Protection Sociale des Enfants Vulnérables (IPSEV) Family integration Cash Safety net programs are spread across a number of ministries and agencies 49. Half of the programs are under the two main social welfare ministries. Programs oriented to vulnerable groups like the elderly or vulnerable children, tend to be concentrated in the Ministry of Family and the Ministry of Social Action and National Solidarity. Two multi-sectoral programs, the nutrition program NETS and the HIV-AIDS orphans and vulnerable children program are under the Prime Minister’s Office. For since sector programs, like school lunches and health fee waivers, the respective social sector ministries are responsible (see Figure 3). Institutional anchorage appears linked mainly to a combination of objectives and target groups combined with institutional mandates rather than any coordinated strategy of interventions. To a certain extent the institutional dispersion reflects the transversal nature of social protection. 50. There has been a history of institutional instability which has hampered developing coherent programming, particularly within the social welfare ministries. For at least the last 10 years, there have been frequent reconfigurations of the main social welfare ministries and shifting of departments and programs between ministries. This weakened the ability to effectively design and implement safety net programs. Moreover, there remains a certain level of institutional coherence. For example, the CSA now is under the Ministry of Family but the Prime Minister’s Office makes decisions regarding when and where the CSA distributes food. 28 Figure 3: Institutional Framework for Safety Net Programs Coordination Interinstitutionnelle SESAME Min. Min. Santé Famille Alimentation & Scolaire Prévention PRP Cantines Min. Eco. Fin. scolaires Min. Culture CSA Min. Education Min. PAM CV Jeunesse PAPA Min. Petite Enfance NETS IPSEV PRCB OEV FSN Min. Primature Action Sociale 51. Various institutional coordination mechanisms are in place, though they are often ad-hoc or project-specific (Table 10). Inter-ministerial coordination mechanisms have been put in place for five of the twelve programs, either at the central or local levels. These inter-ministerial mechanisms are used mainly for implementation and monitoring issues, but in some cases they have a role in programming decisions. Table 10: Safety Net Program Coordination Mechanisms Program Coordination Mechanism 1 Commissariat à la Sécurité Alimentaire (CSA) No formal mechanism 2 Fonds de Solidarité Nationale (FSN) Internal orientation committee 3 Programme de Réadaptation à Base Communautaire (PRBC) Internal within Social Action Dept. 4 Projet d'Appui à la Promotion des Aînés (PAPA) National committee 5 Direction Cantines Scolaires (DCaS) Interministerial multisectoral technical group 29 6 Programme de Cantines Scolaires du PAM Multisectoral technical group 7 Programme de Bourses d'étude pour les Orphelins et autres Executive secretariat and multi-sectoral council Enfants Vulnérables (OEV) 8 Plan Sésame None 9 Programme d'appui à la mise en œuvre de la Stratégie de Inter-ministerial steering committee Réduction de la Pauvreté (PRP) 10 Programme Nutrition ciblée sur l'Enfant et Transferts Sociaux Inter-ministerial steering committee (NETS) 11 Programme de bons d'achat du PAM (PAM CV) Single ministry steering committee 12 Initiative de Protection Sociale des Enfants Vulnérables Technical monitoring committee (IPSEV) Current safety nets afford little effective coverage 52. Data on coverage is limited in some cases, but Senegal’s safety nets include both programs with limited coverage as well as programs with impressive national coverage, in particular the school lunch program. Data is uneven on access to Senegal’s safety net programs. Several programs do not keep beneficiary registries, such as the Social Action Department’s support to the elderly through PAPA, as they tend to work with groups which range widely in number of members. Others, like the CSA distribute rice widely during religious ceremonies, with only rough estimates of beneficiary numbers. Moreover, data is not disaggregated between support to those affected by disasters and food distributed through public events (i.e. no screening of beneficiaries). School lunches are becoming universal, with no screening of beneficiaries within schools (it is assumed all receive). For Plan Sésame, the Ministry of Health was unable to provide the actual number of elderly utilizing the services on an annual basis. With these limitations and caveats, Table 11 presents annual figures on beneficiaries provided by the safety net programs. 53. Annually, about 4 million people are estimated to receive some type of safety net assistance, mostly through CSA and school lunches. This is equivalent to a little under one-quarter of the national population. CSA accounts for about 80 percent of these beneficiaries and school lunches for an additional 17 percent. However, neither the CSA nor school lunches screen beneficiaries based on need. For the CSA, there are political factors which guide food distribution. For school lunches, this is increasingly seen as part of a universal packet of educational services. Even when targeted to regions of food insecurity, for instance through the WFP assistance, not all students would necessarily be vulnerable. Table 11: Safety Net Beneficiaries by Program and Year, 2009 - 2011 Number of Beneficiaries Total 2009 – Program 2009 2010 2011 2011 1 Commissariat à la Sécurité Alimentaire (CSA) 2 760 000 3 000 000 3 600 000 9 360 000 2 Fonds de Solidarité Nationale (FSN) 32 000 NR NR 32 000 Programme de Réadaptation à Base Communautaire 3 1 500 1 900 NR 3 400 (PRBC) 4 Projet d'Appui à la Promotion des Aînés (PAPA) NR NR NR 0 5 Direction Cantines Scolaires (DCaS) 700 414 761 439 780 000 2 241 853 6 Of which Programme de Cantines Scolaires du PAM 567 185 565 560 596 253 1 728 998 Programme de Bourses d'étude pour les Orphelins et 7 3 290 5 060 4 956 13 306 autres Enfants Vulnérables (OEV) 8 Plan Sésame NR NR NR 0 9 Programme d'appui à la mise en œuvre de la Stratégie de 1 274 1 440 700 3 414 30 Réduction de la Pauvreté (PRP) Programme Nutrition ciblée sur l'Enfant et Transferts 10 2 982 21 986 26 294 51 262 Sociaux (NETS) 11 Programme de bons d'achat du PAM (PAM CV) NC 97 000 55 000 152 000 Initiative de Protection Sociale des Enfants Vulnérables 12 NC NC 900 900 (IPSEV) 4 467 85 Total 3 501 460 3 888 825 11 858 135 0 NR: Non renseigné /NC: Non concerné. 54. Considering only those programs which target and screen vulnerable beneficiaries and for which there is data on number of beneficiaries, i.e. excluding CSA and school lunches, about 100,000 people benefitted last year. This is mainly from the two cash transfer prorgrams, the nutrition program NETS and the cash voucher program of the WFP (PAM CV). Between 2009 and 2011, the NETS had an average of 17,000 beneficiaries, while the PAM CV served an average of 50,000 beneficiaries each year. 55. Many of the programs aim to have national coverage. Geographic coverage ranges from the national level to targeted regions. Eight programs have national or quasi-national coverage, though three of them concentrate or rural or peri-urban areas. Three of the newer safety net programs intervene in more restricted areas: NETs in poor rural districts, WFP CV in selected urban areas and IPSEV in rural and peri-urban areas (Table 12). The PRP is only implemented in the rural areas of three regions. Table 12: Geographic Distribution of Safety Net Programs Program Geographical area 1 Commissariat à la Sécurité Alimentaire (CSA) National / All the territory 2 Fonds de Solidarité Nationale (FSN) National/ Rural and peri-urban Programme de Réadaptation à Base Communautaire 3 National / All the territory (PRBC) 4 Projet d'Appui à la Promotion des Aînés (PAPA) National / All the territory 5 Direction Cantines Scolaires (DCaS) National/ Rural and peri-urban 6 Programme de Cantines Scolaires du PAM All regions except St Louis & Dakar / Rural and peri-urban Programme de Bourses d'étude pour les Orphelins et 7 National / All the territory autres Enfants Vulnérables (OEV) 8 Plan Sésame National / All the territory Programme d'appui à la mise en œuvre de la Stratégie de 9 3 regions: Matam, St Louis, Louga / Rural Réduction de la Pauvreté (PRP) Programme Nutrition ciblée sur l'Enfant et Transferts (pilot) 6 Regions (64 rural communities): Matam, Louga, 10 Sociaux (NETS) Kaolack, Tambacounda, Sédhiou, Kédgougou / Rural 11 Programme de bons d'achat du PAM (PAM CV) (pilot) 2 regions (10 cities): Pikine, Ziguinchor / Urban Initiative de Protection Sociale des Enfants Vulnérables (pilot) Kolda region, 2 cities: Coumbacara, Kolda ; 35 rural 12 (IPSEV) and peri-urban communities 31 56. Only four of the safety net programs were able to provide beneficiary numbers to the program review disaggregated by region. As presented in Table 13, the predominance of beneficiaries is in the Diourbel region, laying testament to the concentration of CSA activities during religious events. Table 13: Distribution of Beneficiaries of WFP School Lunches, NETS, CSA and OEV by Region WFP School Lunch NETS CSA OEV 2010 2009 2010 2011 2009 2010 2011 2009/2010 Dakar 2 576 NA 0 0 515 605 589 331 519 660 1 784 Diourbel 32 170 NA 0 0 1 108 607 1 332 012 1 977 348 172 Fatick 78 695 NA 0 0 25 239 31 110 27 655 307 Kaffrine 34 784 NA 0 0 14 175 21 578 26 630 254 Kaolack 72 412 NA 2 874 2 010 270 709 240 229 229 120 136 Kolda 46 865 NA 0 0 42 212 40 214 36 659 361 Louga 35 860 NA 6 351 2 010 102 326 88 952 77 981 285 Matam 31 823 NA 9 779 8 414 27 229 24 946 23 155 282 Sédhiou 21 797 2 982 2 982 5 238 0 0 0 95 Tamba 61 093 NA 0 10 843 38 601 45 959 46 468 313 Kédougou 24 566 NA 0 1 799 0 0 0 Thiès 31 699 NA 0 0 510 561 510 724 569 725 489 Ziguinchor 61 400 NA 0 0 52 233 45 800 40 356 416 Saint Louis 0 NA 0 0 52 503 29 145 25 243 323 TOTAL 535 740 2 982 21 986 30 314 2 760 000 3 000 000 3 600 000 5 217 NA – not applicable. Operational Aspects Safety nets in Senegal use mostly categorical targeting 57. Categorical is the main type of targeting, often combined with prioritization of certain geographical areas and confirmed through community–based mechanisms (Table 15 and Figure 5). Overall, the 12 programs reviewed here use categorical targeting. Among these twelve, five programs use only categorical targeting, with the following categories: people over 60 (Plan Sésame, PAPA), disabled (PRBC) and those affected by disasters (FSN, CSA). Four other programs combine categorical targeting with geographic targeting: the IPSEV, the OEC, the DCas and the WFP school feeding (cantines scolaires du PAM) target specific regions. Finally, The three remaining programs combine categorical with geographical and community targeting mechanisms: the PRP, as well as the cash transfer (NETS) and food voucher programs (PAM CV), which delegate the responsibility for the selection of beneficiaries to community-based organizations. 32 Figure 4: Targeting Vulnerability within Categorical Targeting, by Program Table 14: Targeting Methods, Criteria and Information Sources by Program Program Method Criteria Sources Categorical targeting only Plan Sésame Categorical Age ID card Programme de Réadaptation à Base Categorical Disabled Candidate dossier Communautaire (PRBC) Projet d'Appui à la Promotion des Aînés Candidate dossier Categorical Age & Vulnerability (PAPA) Fonds de Solidarité Nationale (FSN) Categorical Victims of disasters Candidate dossier Commissariat à la Sécurité Alimentaire Candidate dossier Categorical Food insecure (CSA) Categorical and geographical targeting Vulnerable children at risk of family Initiative de Protection Sociale des Geographic Reports separation Enfants Vulnérables (IPSEV) Categorical Age and vulnerability Social worker survey Programme de Bourses d'étude pour Geographic Epidemiological situation Epidemiological data/surveys les Orphelins et autres Enfants Vulnérables (OEV) Categorical Orphans and vulnerable children Social worker survey Geographic Rural food insecurity Poverty surveys Direction Cantines Scolaires (DCaS) Categorical School enrollment lists IA, IDEN Programme de Cantines Scolaires du Geographic Food insecure rural areas Poverty surveys PAM Categorical School enrollment lists IA, IDEN Categorical, geographic and community-based targeting Programme d'appui à la mise en œuvre Geographic Poverty Local development plans 33 de la Stratégie de Réduction de la Categorical Women, disabled, HIV-AIDS Neighborhood reports Pauvreté (PRP) Community- Prioritized at community level Community information based Geographic Zones with high malnutrition Nutritional surveys Programme Nutrition ciblée sur l'Enfant Categorical Vulnerable children Reports et Transferts Sociaux (NETS) Community- Prioritized at community level Community information based Geographic Vulnerable areas Poverty surveys Programme de bons d'achat du PAM Categorical Food insecure Reports (PAM CV) Community- Prioritized at community level Community information based 58. Age is the most significant basis of categorical targeting of beneficiaries. Access to health and education and family integration programs all target specific age groups, mainly children and elderly as being vulnerable segments of society (Figure 4). For children, this is linked to the overall objective of investments in human capital, targeting not only the vulnerable but also aiming to break the intergenerational transmission of poverty due to poor current education and health outcomes of tomorrow’s adults. Programs oriented to responding to shocks, social integration and poverty reduction tend to target the household rather than individual beneficiary categories within the household. 59. Targeting of vulnerable groups is based on varying notions of vulnerability, rather than a standard poverty measure, for instance. Each program uses its own concept of vulnerability, mainly in function of the mandate of the responsible institution. The elderly and disabled are seen as a vulnerable group in and of themselves, without further definition of vulnerability within those groups. Women are targeted, often with the added qualifiers of those women heading households or responsible for young children. Vulnerable households are cited by several programs, basically in terms of their capacity to maintain acceptable conditions of the household but without reference to a poverty line or other such objective indicator. Figure 5 shows the overlap between the category of group identified as eligible and whether a program seeks the vulnerable within that category. 34 Figure 5: Distribution of safety net programs by target age groups 60. There are two main methods to recruit beneficiaries. To benefit from some programs, individuals or families have to formulate a request for assistance, which is then considered by the program (the beneficiaries come to the program for assistance): SESAME, PRBC, PAPA, PRP, FSN, and CSA. Other programs reach out to the potential beneficiaries (according to their target population groups), and offer then services (the programs go to the beneficiaries): IPSEV, NETS, OEV, WFP CV, DCaS, and PAM Cantines. For those that respond to demand, information is disseminated to populations through associations’ networks, the local administration, and other support organization. Widespread communication strategies do not seem to be well developed. What eligible populations understand about their right to or how to access program resources may be unequal, as many people often do not access information through official networks. 61. In terms of verification of targeting, programs focus mainly on errors of inclusion . IPSEV, OEV and CLM include the verification of selection criteria though surveys of a sample of pre-selected beneficiaries. If the sample is verified, the full list is entered into the program. In the case of the WFP food coupon program a comprehensive survey of pre-selected households, with a vulnerability score attached to each household, including ownership of durable goods and a food consumption score. For the PRP, the deconcentrated services validate the choice of beneficiaries in collaboration with neighborhood committees, who created the pre-selection list. For all programs, there is no verification of errors of exclusion or a system to register complaints if an eligible beneficiary feels they have been excluded. 62. The performance of these targeting systems is mixed. As summarized in Box 3, some programs were very effective at concentrating on the poorest households, like the PRN and agricultural support programs, while others revealed significant leakage to the non-poor, including educational assistance (like scholarships) and food aid. When looking at the performance in terms of absolute share of the poor that benefit, none of the programs have significant coverage of beneficiaries (as self-reported). Even for programs that are universal, like the Plan Sesame for elderly health services, limited reported utilization of the program can in part be explained by the high number of responses that indicate households are not aware of the programs. 35 Box 3: Summary of Targeting Outcomes of Social Programs The 2011 ESPS Household Survey included questions on coverage of a range of social programs, including some of the safety net programs in this analysis. Programs cited include: nutritional reinforcement (PRN), elderly health care (Plan Sésame), food aid, and educational support (scholarships, etc.). The surveys highlight the low coverage of these programs (Table 14). Poor households report the highest coverage rate for agricultural development, food aid and the Plan Sésame, but this does not surpass 12 percent. Furthermore, these programs tend to have similar levels of non-poor coverage. Table 15: Distribution of Beneficiaries of Social/Safety Net Programs Queried in the ESPS Program Non poor Poor Total Non- Non- Non- beneficiary Beneficiary beneficiary Beneficiary beneficiary Beneficiary Nutritional reinforcement (PRN) 94,6% 5,4% 90,2% 9,8% 92,6% 7,4% Health care for elderly (Sesame) 89,2% 10,8% 90,0% 10,0% 89,5% 10,5% Food aid 88,5% 11,5% 88,2% 11,8% 88,4% 11,6% Education (scholarships) 82,9% 17,1% 93,5% 6,5% 86,9% 13,1% Source: Diop 2012. The nutritional program is the only one that is successful in targeting poor households (i.e. where more than 40 percent of beneficiaries are in the poorest 2 quintiles). Food aid, which ought to be concentrated on the poor uniquely instead has about half of beneficiaries from the non-poor and is underrepresented in the poorest households. Educational supports are relatively ineffective at reaching the poor, with two-third of leakage to the non-poor. The elderly health care program benefits more the better off 40 percent of households with beneficiaries concentrated in urban households. Distribution of Self-Declared Beneficiaries (%) Poverty Quintiles (based on household Poor or Non- Geographic Zone consumption) Poor Total Q1 Q2 Q3 Q4 Q5 P NP Urban Rural Nutritional reinforcement (PRN) 100.0 28.4 37.7 23.9 9.2 1.0 65.5 34.5 5.5 94.5 Health care for elderly (Sesame) 100.0 14.9 13.4 17.7 29.7 24.3 49.4 50.6 75.9 24.1 Food aid 100.0 8.3 14.6 28.5 28.6 19.9 49.3 50.7 84.4 15.6 Education (scholarships, etc.) 100.0 8.4 8.4 24.2 30.1 29.0 25.7 74.3 76.2 23.8 Source: Echevin (2012). Under 20 percent of households are eligible for these programs. Overall, only a small percentage of eligible beneficiaries actually report benefiting from these interventions. The most common reason cited for all of the programs was that the household did not know about the program. The second most often cited reason was not having sufficient documents. Other reasons were lack of access, Government lack of compliance and that the benefit was too little for the effort required. Among Eligible Non-BNeneficiaries Impossible Amount of % to Get to Benefit too Did not % Beneficiarie Government Place to Little for Know about Eligibl s among Lack of Has Not Collect Effort the e Eligible Documents Paid Benefits Required Program Nutritional (PRN) 19.5 7.5 27.5 17.8 13.4 20.0 53.5 Health care for elderly (Sesame) 19.3 10.4 30.2 20.7 10.0 15.4 44.8 Food aid 15.9 11.7 27.6 26.9 14.2 26.0 60.7 Education 12.6 13.0 37.4 21.0 9.0 23.7 49.9 Source : Echevin (2012). 36 Only a few of the programs include conditionality in seeking to influence behavioral choices 63. Most safety net programs in Senegal provide unconditional support. Two types of conditions are placed on behaviors: (i) School attendance is a condition of further participation in the programs of OEV and IPSEV; and (ii) Financial transparency is a condition for further participation in the grants and loans distributed in the programs PRP, PAPA, PRBC. The transfer programs, PAM CV and NETS, exercise ‘soft conditionalities’ in terms of encouraging mothers to register their children at birth and access social services, but there is no withdrawal of assistance if these are not followed. Safety net programs use different mechanisms, including monetary transfers 64. Different transfer mechanisms are used depending on the nature of the program (Table 16). School lunches and the CSA distribute food directly to the final beneficiaries. In terms of logistics, CSA maintains a network of warehouses and trucks throughout the country. The school lunch program either contracts with the CSA or through private suppliers for delivery to schools depending on the region. In terms of cash transfers, there are several distribution channels. The FSN, PRBC and PAPA distribute cash directly from the administration – either from the central Dakar office or through a local office of the program. PAPA does this through a commercial project account. The other programs use financial partners, be they commercial banks, micro-finance institutions or postal banks. For food vouchers through the PAM CV program, beneficiaries purchase food using their coupons and the intermediaries are reimbursed upon presentation of the coupons at partner banks. In certain very isolated zones, transfer services are not available, so for example NETS was obliged to deliver cash directly to the beneficiary households. Table 16: Transfer Mechanisms, Actors and Frequency Transfer Mechanism Actors Frequency and Period 1 Commissariat à la Sécurité Alimentaire (CSA) Food Direct Admin Variable 2 Cash, Fonds de Solidarité Nationale (FSN) Direct Admin Variable Material 3 Programme de Réadaptation à Base Cash, Variable – one Communautaire (PRBC) Direct Admin Material disbursement/group 4 Projet d'Appui à la Promotion des Variable - one Cash Direct Admin Aînés (PAPA) disbursement/group 5 Direction Cantines Scolaires (DCaS) Admin, Assoc., Food Direct Daily during school year providers 6 Programme de Cantines Scolaires du Food Direct Prog, Admin, Assoc. Daily during school year PAM 7 Programme de Bourses d'étude pour les Orphelins et autres Enfants Cash Bank Admin , Bank Daily during school year Vulnérables (OEV) 8 Plan Sésame Fee waiver Indirect Admin. Permanent 9 Programme d'appui à la mise en œuvre de la Stratégie de Réduction Cash Bank Prog, Bank Variable de la Pauvreté (PRP) 10 Programme Nutrition ciblée sur Twice monthly for 6 months of l'Enfant et Transferts Sociaux (NETS) Cash Bank Prog, Bank the year 11 Programme de bons d'achat du PAM Prog, NGO, Assoc, (PAM CV) Vouchers Vouchers Monthly for 6 months of the year Bank 12 Initiative de Protection Sociale des Cash Bank Admin, NGO, Bank Daily during school year Enfants Vulnérables (IPSEV) 37 65. The size of transfer depends in large part on the objective of the program. In most cases the size of the transfer is fixed a priori as a function of program objectives. Six programs (WFP school lunches, DCaS, WFP CV, NETS, OEV, IPSEV) have a specific base for the calculation of transfer values (Table 17). In some, such as the disaster response programs, the values are determined based on a needs assessment. 66. The actual costs per beneficiary show wide variations between programs. For ten programs, a cost per beneficiary has been estimated based on annual budgets and total beneficiaries. The largest cost per beneficiary are programs with specialized needs, like the program to support disabled people or the schooling costs of children affected by HIV-AIDS which extends to secondary school or professional training. The smallest costs per beneficiary are the school lunch program and the food distribution through the CSA. In the case of CSA, although millions are reached each year, at a cost per beneficiary of 353 FCFA per year it is unlikely that the program has any real impact on the well-being of households. Table 17: Value of Transfers and Cost per Beneficiary by Program (FCFA) Annual Cost Program Transfer Amount Basis of Calculation per Beneficiary 1 Commissariat de la Sécurité Alimentaire (CSA) Food Variable According to demand 353 2 Fonds de Solidarité Nationale (FSN) Cash and Estimate of damages and Variable 9 375 materials needs 3 Programme de Réadaptation à Base Grant, Variable According to size of project 177 158 Communautaire (PRBC) materials 4 Projet d'Appui à la Promotion des 500,000 to 4,000,000 FCFA Grant/loan According to size of project 353 Aînés (PAPA) /group 5 1 meal estimated at 125 Direction Cantines Scolaires (DCaS) Food 35% of daily food needs 8 238 FCFA 6 Programme de Cantines Scolaires du 1 meal estimated at 125 PAM Food 35% of daily food needs 5 916 FCFA 7 Programme de Bourses d'étude pour 149 000 per school year les Orphelins et autres Enfants Cash 308 000 per year of prof. Actual costs of schooling 129 495 Vulnérables (OEV) training 8 Plan Sésame Fee waiver Free care – value variable NC n.a 9 Programme d'appui à la mise en 25,000 to 150,000 FCFA œuvre de la Stratégie de Réduction Loan According to size of project 76 406 de la Pauvreté (PRP) /group 10 Programme Nutrition ciblée sur l'Enfant et Transferts Sociaux (NETS) Cash 14 000/ 2 months 30% of poverty line 41 920 11 Programme de bons d'achat du PAM 3000/pers0n, maxi 6 pers (PAM CV) Cash 50% of daily food needs 13 188 /ménage 12 Initiative de Protection Sociale des 7500/month for one child Estimate of household Cash 15,000/month per n.a Enfants Vulnérables (IPSEV) needs additional child 67. Monitoring and evaluation is uneven and difficult to aggregate across safety net programs. There is no standardized monitoring of program implementation across safety net interventions. Because each program establishes its monitoring and evaluation plan, some more rigorous than others, this makes it difficult to get a national perspective on coverage and impact of safety net programs. Monitoring information is often collected through program-specific management information systems, 38 with reporting responsibilities at the local, regional and national levels. While performance indicators and a monitoring of beneficiaries exists to a certain degree in all the programs surveyed, there is a more robust collection of information on program performance among NETS, WFP and DCas school lunches, IPSEV, WFP CV, OEV et PRP (Table 18). Table 18: Evaluation Components WFP/DCaS School IPSEV WFP CV PRP NETS PAPA PRBC SESAME OEV CSA FSN Lunches Information system X X X Performance evaluation X X Process evaluation X Impact evaluation X X X X Rapid assessments X X X X X X 68. Only 4 of the programs reported that they engaged in a formal impact evaluation, including the two agencies involved in the national school lunch program as well as the WFP CV and NETS cash transfer pilots. The two cash transfer pilots planned for impact evaluations are part of program design since the pilots allows for a first appreciation of the relevance and appropriateness of the cash transfer model for Senegal. At the time of this report, only the NETS evaluation was available and is summarized in Box 4. Box 4: Evaluating the Impact of the nutrition cash transfer program NETS The NETS was initiated in 2009 as a temporary measure to respond to food price shocks. The intervention was built on a long- standing nutrition program that identifies and works with families with children vulnerable to malnutrition using decentralized services delivery through NGOs and local Governments. The main objective was to increase household consumption (or prevent decreases) in response to shocks and to reinforce human capital through better access to education and health services. An impact evaluation was designed to capture short term program effects using double difference methodology of treatment and control households and before and after measurements in both. The evaluation found: At the household level:  The cash transfer was used for a variety of expenditures, but the predominant one was for food  A positive impact on the number of meals reported  A reduction in negative coping strategies At the level of the mother:  Increased participation of the mothers in information, education and communication sessions  Improved overall reproductive health care of women as measured by a composite score of services At the level of the child:  More diverse diets for children 0 to 23 months old  Adoption of good eating habits for children 0-5  Reception of the minimum recommended number of meals for children 6 – 23 months old  Greater likelihood of having a growth monitoring card than non-beneficiaries  Reduction in infant morbidity, particularly diarrheal disease  Improved vaccination coverage of the children in the program There were no discernible effects on parents signing their children up for identity cards, the consumption of iodonized salt, or pre- and post-natal visits. There could be supply side factors to this given the predominance of remote rural locations. In terms of beneficiary view of the program, 63 percent were satisfied with the timing of the bi-monthly transfers but only 43 percent felt the amount was ‘sufficient’. In addition, beneficiaries often noted the financial burden of additional transpor t costs to collect the transfers. 39 Source: Institut Fondamental d’Afrique Noire (IFAN). “Impact Evaluation of Cash Transfers in the NETS Program of Social Transfers Targeting Children� University of Dakar, Laboratory of Research on Social and Economic Transformations (LARTES- IFAN), 2012. 69. Where outcome information is available, it points to safety net effects at the level of households and in the immediate environment of the beneficiaries. Outcomes cited by program include: a. Educational grants to HIV-AIDS affected vulnerable children were found to have a positive effect on school enrollment. For instance, during the initial stage of identification of beneficiaries, 3.290 were enrolled which rose to 5,229 by the end of the project. There were also increases observed in children with identity cards. Evaluations also observed a strengthening of capacity for associations of people living with HIV-AIDS. b. Evaluation of the school lunch program found positive effects in a number of areas. Enrollment increased by 4 percentage points compared with before the school lunches, with a larger impact on girls. Teachers noted a quality improvement in the learning environment and less grade repetition. School lunches led to more regular eating for the children, contributing toward the family’s food security. Children also received health services such as treatment for parasites. In the Casamance region, it was felt that the school lunches contributed to social cohesion, creating an incentive for displaced and conflict-affected populations to send their children to school. c. The impacts of the cash transfer pilot NETS (summarized in Box 4) point to significant positive impacts on households. Costs and Financing 70. The safety net programs have averaged about 17 billion FCA per year over the last three years, equivalent to 0.27 percent of GDP (Table 19). The school lunch programs account for over 70 percent of safety net expenditures, reflecting the large coverage. The CSA, despite reaching many beneficiaries, only has an annual budget of about 1 million FCFA because the value per transfer is very low. Table 19: Safety Net Programs Annual Budgets 2009-2011 (FCFA) Program Budget 2009 Budget 2010 Budget 2011 Total 1 Commissariat à la Sécurité Alimentaire (CSA) 3 600 000 000 1 000 000 000 1 600 000 000 1 000 000 000 2 Fonds de Solidarité Nationale (FSN) 310 000 000 NR NR 310 000 000 3 Programme de Réadaptation à Base Communautaire (PRBC) 900 000 000 300 000 000 300 000 000 300 000 000 4 Projet d'Appui à la Promotion des Aînés 101 000 000 (PAPA) NR NR 101 000 000 5 Direction Cantines Scolaires (DCaS) 4 000 000 000 6 000 000 000 10 280 000 000 20 280 000 000 6 Programme de Cantines Scolaires du PAM 3 454 411 000 3 888 512 000 8 319 499 000 15 662 422 000 7 Programme de Bourses d'étude pour les Orphelins et autres Enfants Vulnérables 1 790 475 690 (OEV) 468 657 820 816 257 870 505 560 000 40 8 Plan Sésame NR 3 000 000 000 NR 3 000 000 000 9 Programme d'appui à la mise en œuvre de la Stratégie de Réduction de la Pauvreté 262 500 000 (PRP) 93 500 000 99 000 000 70 000 000 10 Programme Nutrition ciblée sur l'Enfant et Transferts Sociaux (NETS) 2 639 746 905 144 404 570 932 725 835 1 562 616 500 11 Programme de bons d'achat du PAM (PAM 2 928 573 000 CV) NC 1 874 286 720 1 054 286 280 12 Initiative de Protection Sociale des Enfants NR Vulnérables (IPSEV) NC NC NR Total 9,770,973,390 18,510,782,425 23,192,961,780 51,474,717,595 71. Of these annual expenditures, between 64 and 99 percent are dedicated to the cost of the transfers themselves, with the remainder dedicated to administration and monitoring and evaluation. As shown in Figure 6, the programs with the highest non-transfer costs are donor-financed programs and/or those that tend to invest more in monitoring and evaluation. Figure 6: Distribution of Program Costs by Component PAM Cantines 74% 21% 5% DCaS 92% 4% 3% PRP 89% 5% 6% DAS PAPA 88% 12% DAS PRBC 99% 1% CSA 90% 5% 5% FSN 97% 3% PAM CV 64% 26% 10% CLM 82% 3% 15% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% % Coût Prestation % Coût Mise en œuvre % M&E 72. Safety net funding remains largely dependent on donor financing. Out of the nine programs with funding information, 62 percent of costs are financed by donors. Local governments account for 7 percent, largely through their participation in the school lunch program. The national budget accounts for 27 percent with the remaining 4 percent form community contributions (Table 20 and Figure 5). The share of the national budget in safety nets has declined over the three years under review as donors have financed pilot cash transfer programs, including the World Bank and the World Food Program. 41 Table 20 and Figure 7: Sources of Safety Net Program Funding, 2009 - 2011 Government Donors Municipalities Communities/beneficiaries Financing (FCFA) 8 987 866 095 21 207 929 300 2 419 000 000 1 471 918 000 4% Etat 7% 27% PTF Municipalités Communautés / Bénéficiaire 62% Key Issues 73. The limited coverage of safety nets creates both the downside risk of ineffective protection as well as the lost opportunity of the upside potential for safety nets to contribute to human capital development and more inclusive economic growth. The World Bank and development partners have been working closely with the Government for the last two years in assessing safety nets and testing pilots with a view to creating the conditions for scaling up. 74. Institutional dispersion and lack of a systems approach impedes effective scaling up.. The Safety Net Review found a large dispersion among many small programs across several ministries and agencies. This makes coordination at the national and local levels more difficult. At a strategic level, these programs are not conceived as being part of a global policy or national safety net system. At an operational level, there is little sharing of tools, information and other potential synergies between programs. While there may be multiple instruments in any country’s safety net response, these programs should be seen within an overall coherent strategy and with coordination mechanisms that reinforce national strategy. 75. Coordination mechanisms range from the non-existent and ad-hoc arrangements to formalized project structures. Only the PRP, through its steering committee, has inter-ministerial coordination to determine its annual programming. The other programs use these coordination mechanisms to oversee and monitor implementation. The frequency of meetings also varies greatly. Of note, even though there are overlapping areas of interest, be they by the type of safety net or the group targeted by the intervention, there is almost no direct coordination between these programs. 76. Targeting criteria are not applied systematically across programs . The existing safety net programs use a variety of targeting mechanisms and definitions of vulnerable groups. Most rely on categorical targeting, with some search for equity/prioritization of poorer areas in the distribution of program resources across regions. However, categorical targeting, for example the elderly or women or children, typically require some further eligibility screening to ensure that the truly needy among those vulnerable groups are those that effectively benefit. To date, no proxy means test methodology has been used to screen beneficiaries at the household level. Moreover, geographical targeting may use different maps and none of them provide for systematic identification of poorer communities below the regional or departmental level. 77. However, there is updated poverty targeting data available for Senegal. Senegal has a strong tradition of poverty statistics and surveys, including the Demographic and Health Survey (EDS) in 2010, and the national household survey (ESPS) carried out in 2011. Geographical targeting has been 42 supported by village level infrastructure surveys as well as a more disaggregated national household survey in the ESPS. At the household level, the ESPS will permit the development of proxy means test indicators to better screen for eligibility for safety net programs. 78. There is a need to better link an eventual scaling up of priority expenditures with clear outcome goals. There is little information on the impacts of the existing safety net program expenditures. In some cases, the number of beneficiaries is not available. This makes linking expenditures to expected outcomes difficult and needs to be addressed with more systematic evaluation of safety net programs. For example, recent impact evaluations have been done on the cash transfer pilots which found that monthly stipends to at risk households resulted in improvements in food security and a reduction in child illnesses, and better utilization of health services like vaccinations, among other impacts. This is important for establishing the justification for spending on safety nets as not just a ‘hand out’ but rather a ’hand up’. This will be a prerequisite for increasing funding to safety net programs. 79. These constraints will need to be addressed in order to create a coherent safety net and develop a more integrated national social protection system. A better targeted, more impactful and scaled up national system of safety nets will contribute directly to poverty reduction among the vulnerable populations. 43 Chapter 3: Looking Forward: Building a National Safety Net System in Senegal 80. A national safety net system should include a central system for the registration of households, which forms the basis for the targeting and registration of beneficiaries; an institutional platform that integrates coordination mechanisms at the central regional and local levels; and a packet of interventions that place an accent on the productive nature of safety nets and reinforce their contribution to economic growth. Reforms of this nature to establish a more effective national safety net will need to be developed with a financially sustainable fiscal framework over the medium-term. I. Targeting Options for Safety Nets16 Targeting Methodologies 81. The appropriate choice of targeting mechanisms is crucial in the contexts where needs are great and resources constrained. It is therefore necessary to concentrate limited resources on the most vulnerable populations with the aid of effective targeting mechanisms. There are several advantages to effective targeting mechanisms, including the reduction of errors of exclusion (eligible beneficiaries who do not benefit) and inclusion (non-eligible beneficiaries who do benefit). Targeting helps promote pro- poor public expenditures. One of the main challenges is to identify the target groups in a context where more than half of the population lives below the poverty threshold, and where differences among poor households are often minimal. 82. Several types of targeting mechanisms exist that can be used to help identify the poor and vulnerable populations. Table 21 presents a typology of targeting mechanisms. Many of these mechanisms are not mutually exclusive. For example, a labor-intensive public works project may use self-targeting by adopting less than market wages to encourage only the poor to apply combined with geographical targeting to ensure that the infrastructure created benefits the poorer areas of a county. A safety net program may target a certain category of person, like orphans, but program participants go through a community-based selection process to confirm they are in fact the neediest. 83. Most existing safety nets in Senegal use categorical, geographical and/or community-based targeting. As presented in the review of safety net programs in Part II of this report, all use categorical targeting to identify certain demographic profiles of intended beneficiaries. Seven out of twelve restrict program operations to geographical regions deemed poorer and three add a layer of community-based targeting to involve local populations in the identification of beneficiaries. 84. There are opportunities to expand and improve targeting mechanisms among safety net programs. For example, screening individual household poverty levels could improve targeting. To date, there has been no utilization of proxy means testing, a mechanism popular in many safety net programs around the world. As part of the social protection technical assistance program with the Government, the World Bank provided training in PMT methodologies using several household survey data sets. This section summarizes the findings of some of the simulations from that effort. 16 This chapter is largely drawn from “Targeting Performance of Water and Electricity Subsides in Senegal�, Damien Echevin, September 2012. 44 Table 21: Typology of Targeting Mechanisms Targeting Method Description Advantages Disadvantages Mechanism Selection based on a direct Costly and requires significant Means test evaluation of household Precision administrative capacity and reliable means (income) income data Selection based on a Better at identifying chronic poor Household/ Proxy means statistical correlation of Less costly, shorter and remains than those affected by transitory test (PMT) observable characteristics of transparent shocks ; requires significant and up- Individual household well-being to-date data Community- Community members select Less costly as selection handled by Can be manipulated; can based beneficiaries based on local community members. Locally- perpetuate social exclusion of targeting knowledge tailored selection criteria. certain groups and individuals Sex, age, Based on observable Administratively simple and less High risk of errors of inclusion ; can Categorical occupation, categorical characteristics costly be politically controversial location correlated with poverty Selection based on the Can have high errors of exclusion Self-selection willingness to participate in a Low administrative costs particularly if stigma of participation program is perceived Source : Coady et al 2004. Proxy means tests can guide beneficiary selection based on poverty characteristics 85. The formula for targeting can be established based on observable characteristics of the population. A characteristic is deemed ‘observable’ when precise data can be collected and not hidden by the household. This method of selecting proxy means test (PMT) indicators is based on an analysis of the characteristics of the poor population. A PMT score consists of a linear combination of characteristics (x1i,x2i,…,xki) such that: SCOREi = a1 + a2 . x2i + a3 . x3i + … + ak . xki 86. Using recent household survey data, the determinants of poverty in Senegal can be identified for observable characteristics. The ESPS 2011 survey provides a wide range of indicators that vary with poverty status. Certain indicators can be falsified by households if they know that their answers could render them eligible, and others are difficult to observe or verify in the field (e.g., employment status, education, marriage status). Other characteristics, such as size and composition of the household as well as housing characteristics, are more easily verifiable. These types of more observable characteristics are preferable when constructing proxy means scores of household poverty. 87. Regression analysis using only observable characteristics of poverty yield the results for urban and rural areas found in Tables 22 and 23. The power of each variable to reflects poverty is measured by its R-squared, with the contribution of each variable ranked from largest contributor to lowest. In both urban and rural households, having a household of 14 members or more is the biggest determinant of poverty. 45 Table 22: The Principal Determinants of Urban Poverty Based on Observable Household Characteristics Coefficient P-Value Cumulative R2 Contribution to R2 14 members or more -0.386 0.000 0.133 0.133 Water from a well -0.252 0.000 0.196 0.063 Latrines -0.233 0.000 0.249 0.054 Rudimentary lighting -0.472 0.000 0.271 0.022 Children (6-14) in the household -0.263 0.000 0.292 0.020 Lighting with kerosene lamps -0.129 0.000 0.297 0.006 Children (0-5) in the household -0.111 0.000 0.302 0.005 Rudimentary roof -0.066 0.000 0.303 0.001 Rudimentary kitchen/cooking -0.142 0.000 0.304 0.001 Unfinished walls -0.062 0.001 0.305 0.000 Disability -0.050 0.001 0.305 0.000 Table 23: The Principal Determinants of Rural Poverty Based on Observable Household Characteristics Coefficient P-Value Cumulative R2 Contribution to R2 14 members or more -0.319 0.000 0.075 0.075 Rudimentary roof -0.213 0.000 0.131 0.057 Water from a well -0.229 0.000 0.172 0.041 Lighting with kerosene lamps -0.149 0.000 0.181 0.009 Children (0-5) in the household -0.147 0.000 0.187 0.006 Rudimentary lighting -0.145 0.000 0.192 0.005 Rudimentary kitchen/cooking -0.299 0.000 0.196 0.004 Children (6-14) in the household -0.149 0.000 0.199 0.003 Latrines -0.073 0.000 0.201 0.002 Piped water -0.057 0.000 0.201 0.001 Disability -0.051 0.002 0.202 0.000 Source : Echevin (2012). 88. Using these results, a targeting methodology can be constructed. The goal is to create a series of indicators based on the fewest determinants of poverty while achieving the largest coverage of the poor with the least amount of leakage to the non-poor. By adding up the coefficients applicable to a given household, a composite score of household wellbeing can be constructed. The higher the score, the better-off the household and vice versa. For example, in rural areas, a household with 14 members that get their water from a well and only have rudimentary lighting would score -.697. An identical household but having access to electricity would score -.548. The PMT can be built separately for rural and urban areas. For example, in an urban area a household that has more than 14 members, with children 6 – 14 years and 0 - 5 years old, has access to a source of lighting and a toilet as well as rudimentary energy sources has housing made of rudimentary materials and water obtained from a well, with no disabled household members would have an overall proxy means test score of -0.89, which 46 represents the consumption per adult equivalent of 0.41 times the poverty line. A rural household with more than 14 members, including children 0-5 and 6-14 years old but no disabled members, with access to lighting and a latrine that uses charcoal wood or gas and gets water from a well and lives in a house made of rudimentary materials including the roofing would have a PMT score of -0.57 which is equivalent to a household consumption per adult equivalent of 0.56 times the national rural poverty line. Thus, without measuring household consumption or poverty level directly, it is possible to classify households according to their score and to help identify the poorest households within areas of generalized poverty. Applying a PMT score to existing safety net programs could improve targeting considerably 89. Applying different targeting scenarios to existing safety net programs reveals potential for improving targeting outcomes. Table 24 presents simulations of nine safety net programs under different targeting scenarios. The ‘base targeting’ scenario uses categorical targeting based on the general beneficiary characteristics of each program. This is typical of safety net programs that identify categories like the elderly, children under 5, school children etc. in screening beneficiaries. The second scenario estimates targeting outcomes using the proxy means test mechanism developed in the previous section. The third simulation is geographical targeting based on concentrating resources in the poorest areas in the regions where programs operate, differentiated between urban and rural. 90. Applying a PMT could significantly improve the targeting of existing social safety net programs. Simulations of the different targeting methodologies for 9 safety net programs showed that categorical targeting (used alone) was the least performing method, that is the method that resulted in the most leakages of program coverage to the non-poor (baseline scenario). Geographical targeting improves outcomes compared to the baseline scenario in all of the simulations. Using a proxy means test further improves targeting outcomes in almost all cases (PAPA urban, IPSEV urban and rural, WFP CV, NETS rural, DCas urban and rural, OEV urban, PRP, PRBC urban, PAM cantines urban and rural). The only cases where PMT does not provide better targeting than the combination of geographical and categorical targeting were PAPA rural, NETS urban (only slightly), OEV rural, PRBC rural. 47 Table 24: Distribution of Safety Net Program Beneficiaries Based on Alternative Targeting Scenarios Total Urban Rural Consumption Quintiles Consumption Quintiles Non- Q1 Q2 Q3 Q4 Q5 Poor Q1 Q2 Q3 Q4 Q5 Poor Non-poor poor PAPA Base targeting 100.0 23.6 22.3 20.1 19.7 14.3 38.3 61.7 21.5 21.3 21.5 20.7 14.9 61.3 38.7 PMT 100.0 32.6 30.1 17.6 17.2 2.5 51.8 48.2 30.5 25.8 19.6 17.1 7.0 72.7 27.3 Geographical 100.0 40.7 20.2 15.2 13.6 10.3 45.2 54.8 35.4 25.8 17.9 13.7 7.2 76.9 23.1 IPSEV Base targeting 100.0 21.2 20.7 20.9 19.9 17.1 34.6 65.4 20.1 20.1 20.1 20.0 19.7 57.1 42.9 PMT 100.0 35.7 30.6 17.0 13.2 3.5 55.6 44.4 30.8 25.5 20.3 14.9 8.5 73.3 26.7 Geographical 100.0 46.3 21.5 15.5 11.4 5.3 52.1 47.9 30.8 24.6 20.2 14.1 10.3 72.6 27.4 PAM CV* Base 100.0 52.8 18.9 11.5 9.4 7.4 57.7 42.3 n.a. n.a. n.a. n.a. n.a. n.a. n.a. targeting PMT 100.0 66.8 24.1 1.7 3.6 3.8 73.3 26.7 n.a. n.a. n.a. n.a. n.a. n.a. n.a. Geographical 100.0 52.8 18.9 11.5 9.4 7.4 57.7 42.3 n.a. n.a. n.a. n.a. n.a. n.a. n.a. NETS Base targeting 100.0 22.4 22.5 21.0 18.3 15.8 37.3 62.7 20.7 20.9 20.4 19.6 18.4 58.8 41.2 PMT 100.0 34.9 29.4 17.4 14.3 3.9 54.1 45.9 31.6 25.9 20.3 14.4 7.8 74.3 25.7 Geographical 100.0 48.5 21.9 15.1 9.6 4.9 54.3 45.7 31.6 25.2 20.3 13.3 9.5 74.2 25.8 DCaS Base targeting 100.0 20.7 20.5 20.5 19.1 19.3 33.8 66.2 21.0 21.1 19.0 19.3 19.6 58.2 41.8 PMT 100.0 45.1 26.8 14.0 12.9 1.2 60.4 39.6 31.5 24.9 20.3 15.2 8.1 74.1 25.9 Geographical 100.0 47.7 21.3 14.2 11.3 5.6 53.4 46.6 31.0 25.7 19.2 13.4 10.7 73.4 26.6 OEV Base targeting 100.0 12.0 19.3 20.1 21.1 27.5 25.9 74.1 18.2 16.7 18.1 21.1 25.8 50.8 49.2 PMT 100.0 32.6 30.1 17.6 17.2 2.5 51.8 48.2 29.3 20.5 23.0 18.3 8.9 69.3 30.7 Geographical 100.0 40.7 20.2 15.2 13.6 10.3 45.2 54.8 32.5 24.6 20.2 12.7 10.0 74.9 25.1 PRP Base targeting n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 20.0 20.0 20.1 19.9 20.0 56.9 43.1 PMT n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 30.8 25.5 20.3 14.8 8.5 73.3 26.7 Geographical n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 30.7 24.4 20.2 14.1 10.6 72.3 27.7 PRBC Base targeting 100.0 25.6 26.0 18.7 20.0 9.7 40.9 59.1 25.2 22.6 20.2 17.2 14.9 64.3 35.7 PMT 100.0 43.2 32.9 12.2 11.6 0.0 64.5 35.5 33.0 27.1 18.5 13.2 8.3 75.7 24.3 Geographical 100.0 51.1 23.4 15.8 7.4 2.3 56.0 44.0 38.7 29.0 16.1 9.7 6.5 81.7 18.3 PAM Cantines Base targeting 100.0 20.7 20.5 20.5 19.1 19.3 33.8 66.2 21.0 21.1 19.0 19.3 19.6 58.2 41.8 PMT 100.0 45.1 26.8 14.0 12.9 1.2 60.4 39.6 31.5 24.9 20.3 15.2 8.1 74.1 25.9 Geographical 100.0 47.7 21.3 14.2 11.3 5.6 53.4 46.6 24.7 23.4 19.5 17.5 14.8 64.9 35.1 Source : Echevin (2012) Note : *only Ziguinchor urban. 48 Recommendations for improving targeting 91. Different ways of targeting (PMT, geographic, community based, etc.) should be further analyzed in order to better understand their potential uses. For many safety net programs, categorical targeting alone will be insufficient to ensure that the programs result in pro-poor expenditures. Categorical can be combined with one or more other types of targeting. Geographical targeting can be particularly relevant in rural areas. Otherwise, there is little statistical information on targeting outcomes of those programs that use community-based targeting, but it has been employed successfully in many African contexts. Finally, simulations show that PMT might be considered in the future in Senegal. The only caveat with focusing on improving targeting mechanisms is that it should not distract from the greater issue of the need to increase coverage of safety net programs in general and the need to reform the larger expenditures going to untargeted subsidies at present and reallocate them to targeted programs. 92. The PMT mechanism could be useful for all poverty-oriented programs that target the household level. A unified targeting mechanism, to which each program would add specific filters depending on its target groups, can be used across the range of targeted programs in education, health, agriculture, water, and other key sectors. To create this unified targeting mechanism, next steps include: a. Update the targeting methods based on poverty maps that are more decentralized than the present ones using the new data available from the ESPS. b. Adapt the targeting formula to specific contexts like natural disasters. c. Implement pilots to test the actual performance of the PMT targeting tool. 93. A common targeting tool like PMT should be supported by other elements of a common targeting system. The PMT system should work towards a common registry for vulnerable households and program beneficiaries, which would be based on a common intake questionnaire that can be used to apply the PMT and that includes additional information relevant for the targeting of each program. In addition, using a more harmonized approach to community-based targeting at the local level could be a strong complement to the PMT scoring system to reach a common system. 94. A national targeting system should be supported by an institutional framework that identifies one operational agency in charge of the operationalization and monitoring of the targeting system and unique household registry. This agency would coordinate system implementation with the range of other agencies implementing targeted programs. A clear institutional framework will improve the management of targeting and provide greater stability to the social protection system in Senegal. 49 II. Institutional Arrangements for a National Social Safety Net System 95. An adequate institutional framework supporting the creation of a national safety net system will be critical to its success. Senegal does not currently have an institutional framework of a national social protection system or an institutional platform sufficient to support a scaled-up, integrated national safety net system, although several key components are in place. 96. Despite institutional instability, dispersion of programs and weak coordination mechanisms, there are many strengths in the institutional framework for safety nets in Senegal which the country can build on to create a more comprehensive safety net system: a. While dated, the national social protection strategy provides a unifying view of social protection and a common language to prioritize interventions. b. Compared to 2005 when the strategy was adopted, there is an important base of existing safety net programs. From an institutional perspective, the existing program base has allowed for experimentation with implementation procedures and program-level coordination mechanisms. c. The Government recently established an Inter-ministerial Committee on Social Protection to promote coordination at the highest level of Government. As social protection is a national priority and it cuts across several sectors, it is necessary to have a point of institutional articulation between the main actors. Although the inter-ministerial committee is new and had not yet become fully operational, the aim is to provide overall policy guidance and coordination across the full range of social protection measures. The composition is broad-based, with representation of all key ministries as well as donors and civil society. d. The new government has created the General Delegation for of Social Protection and National Solidarity (Délégation Générale à la Protection Sociale et à la Solidarité Nationale) to provide a high profile focal point in the President’s Office signaling the importance of an effective social protection system to the priorities of the Government. 50 Table 25: Examples of Institutional Arrangements in NETS Cash Transfers and INPS – “Squali Jaboot� NETS Cash Transfer INPS Institutional Design Builds on existing Projet de Renforcement Nutritionel (PRN) Is embedded within line agency structures Features arrangements Central level Cellule de Lutte Contre la Malnutrition (CLM) is the La coordination de l’initiative (CI) au niveau national sera assurée par un comité de pilotage placé sous governance structure overseeing inter-ministerial coordination committee under la tutelle du Ministère de la Famille de la Sécurité Alimentaire, de l’Entreprenariat Féminin, de la Micro the Prime Finance et de la Petite Enfance et composé de : Cabinet et conseillers techniques, directions et autres Minister’s Office services, projets et programmes, les organisations faitières de la société civile, les partenaires techniques et financiers du secteur, les élu Technical secretariat National Executive Bureau as its technical secretariat Directions du Ministère de la Famille de la Sécurité Alimentaire, de l’Entreprenariat Féminin, de la Micro Finance et de la Petite Enfance Institutional Prime Minister’s Office Le Ministère de la Famille, de la Sécurité Alimentaire, de l’Entreprenariat Fémin in, de la Micro Finance anchorage et de la Petite Enfance Regional coordination Three Regional Executive Bureaus Cadre de Concertation Régional (CCR) : L’autorité administrative, les services régionaux du ministère mechanisms (développement communautaire, action sociale, Cases des tout-petits, CSA) ou les autres services techniques de l’Etat impliqués dans la mise en œuvre, les projets et programmes du Ministère, les élus. Cadre de Concertation Départemental (CCD): L’autorité administrative, les services départementaux du ministère (développement communautaire, action sociale, CEDAF, Cases des tout-petits,CSA) ou les autres services techniques de l’Etat impliqués dans la mise en œuvre, les projets et programmes du Ministère, les élus. Local coordination Local Governments monitor and report on activities and are Cadre d’Exécution Local (CEL) : L’autorité administrative, le chef de village/quartier, les élus locaux, mechanisms expected to start incorporating nutrition and nutrition l’APDC, les conseils de village ou de quartier, les représentants des bénéficiaires ou Organisations indicators in their Local Development Plans. Communautaires de Base, les personnes ressources. Community LG’s choose one Community Executing Agency (CEA), an Les agents relais communautaires appelés Acteurs Porteurs de Dynamiques Communautaires « APDC » participation aspects NGO or CBO, through a competitive process with assistance constitués d’acteurs locaux initiés et formés aux techniques d’organisation et de mobilisation s ociale, from the CLM jouent le rôle d’interface entre les pouvoirs publics et les populations. Beneficiary selection Utilizes the PRN community organization structure to Identified locally though APDC in consultation with Cadre d’Exécution Local et les Services Déconcentrés identify eligible beneficiaries (mothers of young children in de l’Administration (SDA) vulnerable families) Transfer mechanisms Subproject funds are directly deposited into a separate Transfers are direclty processed through micro-finance institutions, the Poste Finance, Banks or other account of the CEA who contract local payment service transfer agencies. providers 51 97. The existing experiences include both autonomous projects as well as programs embedded within line ministry structures. Table 25 provides more detail on the institutional arrangements of these two differing approaches. The institutional arrangements to channel cash transfers under the NETS have been developed building on the institutional structure of the Nutritional Reinforcement Project which was reaching 320,000 children under five years old, or about 22 percent of children in Senegal, in 2009 when the economic shocks occurred. The Government launched a cash transfer as part of its safety net response, using the existing project’s institutional arrangements and existing interface with poor households. In addition, the Ministère de la Famille, de la Sécurité Alimentaire, de l’Entreprenariat Féminin, de la Micro Finance et de la Petite Enfance designed an integrated safety net program building on line ministry structures that combines cash transfers, access to basic services for the poor via small- scale infrastructure and micro-finance/income generating investments, named the National Social Protection Initiative –“Squali Jaboot� (INPS). Subsequently, the Government invested resources in establishing and training the community level agents which form the basis of this community-driven approach, though additional resources for the transfers have not yet been forthcoming. 98. These two programs present distinct institutional differences. As shown, in times of crisis, a safety net can be grafted onto an existing program. The PRN already had an institutional structure as a multisectoral approach to addressing malnutrition, based in the Prime minister’s Office operating with a Project Implementation Unit structure. The PRN’s team, supported by international donors, was highly effective at reaching communities using a contracting approach through NGOs. The main institutional coordination was through local governments. This provided a ready platform for launching the country’s first cash transfer pilot. The INPS represents an approach more embedded in a line ministry, allowing for an articulation with other efforts of the Ministry and a reliance on budgetary systems rather than project financing mechanisms. Both approaches rely on effective local structures and direct payment mechanisms for transfers. III. Towards a National Safety Net System 99. The overall objective of the institutional arrangements around safety net interventions should be to ensure that they reach their objectives and that they contribute to coherence of an integrated national safety net system. In other words, not only should institutional arrangements provide sufficient support for implementation of an individual program, they should contribute to the creation of a coherent national system. A national safety net system aggregates and integrates these efforts based on a clearly specified legal and programmatic framework which clarifies roles and responsibilities. The system should focus on a minimum packet of interventions that protects vulnerable Senegalese, both the chronic and the transitory poor, helping them to manage risks. 100. The core elements of a national safety net system might include: a) Development of a common set of tools to be shared across programs, including: i) Targeting mechanism. The new data available from the ESPS will permit significant advances in geographical and household level targeting and development of eligibility criteria. The survey data will allow for the development of proxy means test indicators that help identify household characteristics correlated with poverty and vulnerability which can be used to develop screening criteria for access to various programs. 52 ii) Coordinated budget mechanisms within an overall medium-term expenditure framework for safety nets. iii) Coordinated development of indicators and MIS so that they can be communicated consistently across programs and aggregated at the national level. iv) Beneficiary registries. Inter-operable beneficiary registries across programs, and a central registry, to improve coordination and efficient use of resources. v) Common local coordination mechanisms. Typically, each agency or program develops its own local coordination mechanisms, with multiple committees and consultation mechanisms. A common institutional platform would foster synergies between social protection programs. b) Linkages between programs. Rather than an isolated set of a dozen or so safety net interventions with little interaction between each other, a national safety net strategy seeks to build explicit synergies between programs. This can be achieved by for example by building links between programs that deliver different safety net functions in order to better enable beneficiaries to graduate from poverty. Households enrolled in cash transfer programs could be automatically eligible for health or education fee waivers or households participating in public works could be linked with income generating projects, microfinance or other asset building programs to help them graduate from safety net programs. c) Intentional scalability. The national safety net system should have the flexibility to respond quickly to the short-term needs resulting from shocks and to scale back down after the period of transitory crisis passes. This includes the contingency planning across all main program areas. d) Capacity building. A national safety net system would identify overall needs for capacity building in transversal topics, like needs identification among vulnerable groups, targeting, management of transfers, monitoring and evaluation. These could be shared across programs. e) Medium-term expenditure framework. Moving from a project or individual program approach to a national safety system will require developing an overall financial framework for the sector that prioritizes expenditures and builds a sustainable funding basis for safety nets. The medium- term expenditure framework used for other sectors is a well-known tool that would help translate a national safety strategy into public spending priorities within a coherent multi-year macroeconomic and fiscal framework and integrated into the budget process. f) Donor harmonization. Where international partners provide a significant level of safety net resources, bringing donors into a harmonized, Government-led framework is essential to creating a coherent safety net system. Arrangements like pooled financing, and joint program supervision and monitoring helps avoid excessive transaction costs for the Government. 101. International experience points out important elements of institutional design of effective national safety nets. Box 5 summarizes some of the findings of the World Bank’s review of safety nets. Designing a national safety net system requires a pragmatic approach that builds on a focus on core vulnerable groups, with reasonable targets for coverage and a rationalization of programs. The institutional framework should include sufficient attention to building administrative capacity for effectiveness and strong monitoring and evaluation systems that will allow for measuring whether results are achieved and reporting on the effectiveness of the resources used, both as a way to establish impact as well as to justify future funding. 53 Recommendations for an institutional platform for a national safety net system in Senegal 102. For Senegal to advance in developing a national safety net system, the institutional framework will need to be developed that integrates the current dispersed efforts and creates the technical and institutional platform able to support national scale programs. Recommendations for the main elements of this nascent national safety net system include: 103. Strengthening coordination at the central level. At a minimum this should seek overall policy coherence and coordination of programming between elements of the national safety net within the overall poverty reduction strategy institutional framework. Collaboration with other relevant sectors is also essential, for example health and education services, decentralization and natural disaster management. Central level coordination on budget formulation and developing a sustainable medium- term financial framework will help underpin the sustainability of the system. Harmonization with external partners, including international agencies and NGOs can be improved through central level coordination mechanisms. The newly formed Inter-ministerial Committee on Social Protection and the General Delegation for of Social Protection and National Solidarity (Délégation Générale à la Protection Sociale et à la Solidarité Nationale) can serve as the basis for this. Box 5: Common Pitfalls in Reforming Safety Net Systems Based on a review of international experience in safety nets, with a focus on developing countries, when developing reform plans, avoiding common pitfalls such as the following is important: • Having unrealistic expectations. Safety net programs can never fully compensate for macroeconomic instability or eliminate the causes of poverty, although they can be helpful when used in conjunction with policies that address the root problems of these. • Avoiding conflicts between policies and programs. Safety net programs cannot be expected to, for example, solve an unemployment problem caused by excessively restrictive labor market regulation or rural poverty caused by distortions in agriculture markets. • Avoiding having too many programs. International experience is rife with countries that have too many programs, each with low coverage, low benefits, inadequate administrative systems, and high overheads. Having fewer, larger programs would allow them to achieve economies of scale. In countries with too many programs, they often overlap and are not sufficiently coordinated to achieve the best possible synergies. • Preventing an imbalance in target groups. Programs may be excessively based in the formal sector or favor “virtuous groups� such as children or the elderly while failing to cover other groups, such as minorities or those with disabilities. • Rationing entry into a program by budget rather than by eligibility threshold. When the funding for a program is insufficient to allow all those who meet the defined eligibility criteria to be included, horizontal inequity is created, transparency declines, and opportunities for rent seeking are created whereby eligibility intake officers may demand bribes or give favorable treatment to those with whom they share an interest or affiliation. • Having insufficient administrative effort, monitoring, and evaluation. Programs are often set up in a hurry with only rudimentary systems. They may die altogether or fade away after a few years, especially if a change in government occurs. Developing and fine-tuning the most effective safety net systems takes time. Source: Grosh, M., C. del Ninno, E. Tesliuc, and A. Ouerghi. For Protection & Promotion: The Design & Implementation of Effective Safety Nets, World Bank, 2008, page 397. 104. Similar safety net programs should be rationalized, harmonized, and managed under a common institutional structure. This is particularly important for cash transfers where there the risks of multiple cash transfer programs in terms of overlap and dispersion is significant. Moreover, administrative efficiencies of scale will not be achieved through multiple small programs. The pilots were essential to test delivery mechanisms and evaluate impact sin order to justify a more ambitious national program. However, neither of the pilots was part of a long-term strategy, but rather an ad-hoc 54 response to economic crises. The lessons learned from these pilots and the institutional experience, including operational manuals and beneficiary screening systems etc., should be reviewed in order to salvage the capacity gained in service of developing a national program better anchored in the institutional structure of country, including greater use of national systems and line agencies. If the country were to consider developing labor-intensive public works as part of a national safety net, the institutional structure of this should work closely through the decentralized structure of local governments and village-level associations given their mandate for provision of small-scale infrastructure. 105. At the decentralized level, rationalization of the multiple coordinating structures of individual projects and programs is needed. Setting up local-level safety net committees could address some of the present dispersion in local structures and allow for better coordination of efforts. A common local committee structure would also be able to support the identification and screening of beneficiaries and direct them to various appropriate programs. Such local committees would provide an integrated view of needs locally, while different programs could retain their specific operating procedures. Common beneficiary registries could be shared among programs to further promote more harmonized coverage across beneficiary groups. 106. Common tools and systems should be developed to be shared among programs. Developing a national proxy means test is possible with the new household survey. From international experience, the proxy means test parameters can be developed with the technical resources of the National Statistical Agency but the implementation of the system should be operationalized by an autonomous actor in the sector. There are several reasons for this, including to put in place a targeting system that includes activities like house visits, involvement of community organizations, outreach, etc. as well as avoid potential misinformation from households on future surveys if they perceive it affects whether they receive benefits. Geographical targeting should also be shared across programs using similar if the same set of targeting maps. Inter-operable information systems and the creation of a unique central registry of vulnerable households would improve coordination and efficient use of resources and these can be developed from the bottom–up, with information verified at the level of the local committees. 107. Finally, a national monitoring and evaluation system for safety nets should be established under the authority of the institution in charge of the sector coordination. A concerted approach in terms of M&E would allow for a more coherent reporting on coverage and impacts of all the country’s safety net system through a system of key performance indicators common to all safety net programs. Moreover, this would allow for learning across programs to improve performance. 108. A systems approach reduces fragmentation, promotes harmonization and can enhance both the performance of individual programs and the overall protective, preventive, and promotive functions of safety nets in Senegal. The national social protection system should have the flexibility to respond quickly to the short-term needs resulting from shocks. A national safety net systems approach attached to concrete coverage objectives and sufficient funding, built on a strong institutional platform, will provide Senegal with an important component to address poverty reduction and exclusion, avoiding destitution and building assets for the poor. 55 IV. A Financial Framework for a Safety Nets in Senegal 109. Overall spending on social protection in Africa (including social insurance systems such as pensions or health insurance) is low by international comparison, but safety net expenditures are relatively similar (Box 6). Box 6 : Costs and funding sources for safety nets in Africa Dépense en programmes de filets sociaux, excluant les subventions universelles Dépense Total subventions Year % total de la expenditure % donor universelles (M.=Average) Country % du PIB dépense du (% GDP) financed (% du PIB) Gouv. Benin 0,3 1,1 65 0,5 0,9 M. 2005-10 a a Botswana 1,5 9,5 0 0 1,5 M. 2005/06-11/12 Burkina Faso 0,6 <1,0 80 0,7 1,3 M. 2005-09 Cameroun 0,2 1,5 77 1,4 1,6 M. 2008-10 b b Ethiopia 1,2 , 100 0 1,2 2009 Kenya 0,8 1 71 0 0,8 2010 Lesotho 4,6 8 0 4,6 2010/11 Liberia 1,5 4,4 94 0 1,5 M. 2008-11 Madagascar 1,1 5 , 0 1,1 2010 Mali 0,5 , 60 0,1 0,5 M. 2006-09 Mauritania 1,3 4,6 38 1,35 2,9 M. 2008-13 Ile Maurice 4,4 9 , 0,8 5,2 2008/09 Mozambique 1,7 , 62 1,4 3,1 2010 Niger , 1,0-5,0 67 , , M. 2001-06 Rwanda 1,1 , , 0 1,1 2010/11 Sierra Leone 3,5 13,1 85 2,1 5,6 2011 South Africa 3,5 , , , , 2010 Swaziland 2,1 , , 0 2,1 2010/11 Tanzania 0,3 1 , 0 0,3 2011 Togo 0,5 1,8 75 0,8 1,3 M. 2008-10 Zambia 0,2 , 75 1,9 2,1 2010/11 Regional average Africa 1,6 4,4 70 0,5 2,1 . Low-income 1,1 3,6 75,9 0,5 1,6 . countries* Middle-income 2,4 7 50,7 0,7 2,9 . countries** Other regions Europe and Central c c Most recent 2008- 1,8 - 1,8 Asia 10 Latin America and d d 1,1 - 1,1 2010 Caribbean Middle East and e 0,7 5,7 6,4 most recent North Africa *PFR: Benin, Burkina Faso, Ethiopia, Kenya, Liberia, Madagascar, Mali, Mozambique, Niger, Rwanda, Sierra Leone, Tanzania, Togo ** PRI: Botswana, Cameroun, Lesotho, Mauritanie, île Maurice, Afrique du Sud, Swaziland, Zambie, Sources: Table compiled and presented in Monchuk, 2013 56 110. There are a handful of large social protection programs operating at a national scale with significant coverage in Africa. All of these programs which operate at the national level cover a substantial portion of their target population, whether that is the entire population as in the case of Rwanda’s national health insurance program or targeted to a specific population group like Lesotho’s old age social pensions or Ethiopia’s rural food insecure. In lower income countries, development partners’ funding is heavily present in these expenditures. In middle income countries like South Africa, these programs are fully domestically-funded. Current safety net expenditures in Senegal are insufficient 111. Public spending on safety nets and social assistance programs (not including general price subsidies) has historically been low in Senegal. Spending on social assistance and safety nets amounted to about 0.15 percent of GDP in 2004. During drought years of 2002-2003, this rose to 0.5 percent of GDP reflecting emergency food and materials to support rural populations (World Bank 2006). 112. Current safety net spending in Senegal is about 0.2 percent of GDP. Moreover, the bulk of that spending is on programs like school lunches or CSA food distributions which have other objectives and/or are not necessarily targeted through a safety nets perspective. The programs that have large coverage levels also tend to distribute a very low level of benefit as a share of household consumption, which can dilute potential impact on household well-being. Estimating scale-up expenses of effective safety nets 113. The goal of scaling up safety nets will need to be based on a realistic cost framework. The overall costs of a national safety net will depend on the following factors: a. Type of program, for example labor intensive public works programs are more expensive per unit of transfer because of the additional non-transfer resources required to construct infrastructure, which increase the cost of a program by 20 – 40 percent depending on the level of labor intensity. However, the infrastructure created provides additional benefits through the investments created in poor communities. b. Level of coverage. The level of coverage is often a function of resources available. However, one of the goals in Senegal should be to reach a level of coverage that extends beyond the pilots and small-scale efforts seen thus far to create a national safety net with a meaningful reach. c. Size of the transfer. The level of resources required depends on the size of the transfers. There is a tension between retaining a level of benefit that serves the objectives and being able to afford scaling up. Determining the appropriate level of benefit must optimize between adequacy, affordability and acceptability. In terms of adequacy, impact evaluations, like those done for Ethiopia’s Productive Safety Net Program confirm that the size of the transfer received matters in terms of household level impacts. In terms of acceptability, the transfer cannot be so large that it appears unfair to those that do not receive the transfer. The intension of the transfer is not that it, in and of itself, lift a family out of poverty but rather that it help avoid destitution, malnutrition and negative coping strategies, for example. An international rule of thumb is that the cash transfer amount should be between 20 and 40 percent the poverty line in order to be meaningful to the beneficiary. 57 d. Administrative efficiency. Program operating costs and overhead typically range between 10 to 20 percent of total costs. Larger programs can have efficiency gains in this regard, in light of the relative size of fixed costs. 114. Information from on-going pilots, simulation exercises and insights from other larger programs in Africa can give an indication of the scope of financial requirements. Table 26 presents costing scenarios for scaling up critical safety net programs in order to create a minimum social floor that would provide responsiveness to shocks. In terms of cash transfer programs, the existing pilots reach a small proportion of the population and represent a very small draw on financial resources. Several simulations of larger cash transfer programs have been done, using an assumed transfer amount of 5,300FCFA per month. Depending on the coverage, the financial costs of a larger cash transfer program could vary from 0.33 percent to 1.7 percent of GDP. Building a viable financing base for safety nets 115. Expanding resources for safety nets in Senegal should be based on a vision of safety nets as a productive investment, helping the country to increase human capital, build assets for the poor and contribute to broad-based economic growth. Budgets for vulnerable groups have typically been viewed as an opportunity cost for more impactful investments, making it difficult to claim adequate fiscal space. 116. If a national safety net were dimensioned at an annual cost of 68 billion FCFA it would be equivalent to about 1 percent of GDP and 3 percent of the total national budget. Table 27 presents some comparative financial parameters to put this amount in perspective. Financing a robust safety net would not cost more than typically annual subsidies to basic food stuffs and utilities, and yet would have a far less distortionary and better targeted impact. Table 26: Financial Reference Parameters for Financial Dimensions of a National Safety Net Total National Budget 2011 2 032 billion FCFA (29.7% GDP) 2011 National Budget Current Expenditures 1 159 billion FCFA 2011 National Budget Capital Expenditures 823 billion FCFA Food subsidies 2008 46 billion FCFA Electricity subsidies to SENELEC 2011 76 billion FCFA Total transfers and subsidies 2011 283 billion FCFA If national safety net = 1% of GDP 68 billion FCFA 58 Table 27: Financial Parameters of Safety Net Programs in Senegal and Reference Programs in Africa Pilot Programs Total Population Coverage % of Target Population % National Annual Cost % GDP Covered Population NETS 26 294 households with children <5 n.a <1% 1 562 616 500 (2011) 0.02% (158,000 beneficiaries) ($3.0 million) WFP voucher 97 000 beneficiaries (max 2010) n.a <1% 1 874 286 720 (2010) 0.025% ($3.6 million) Simulations Universal child allowance 2,448,000 children < 5 years 100% poor children < 5 20% 1.7% (Samson et Cherrier 2009) UNICEF (2011) 4% $78.3 million 0.55% Recommended 534,000 children < 5 years 37% poor children < 5 Scenario Alternative 1 311,000 children < 5 years 23% “ “ 2% $45.6 million 0.33% Alternative 2 802,000 children < 5 years 50% 6% $117.4 million 0.83% Alternative 3 1,411,000 children < 5 years 76% 11% $206.7 million 1.46% Other programs in Africa Ethiopia Productive Safety Net 7 million beneficiaries 30 % of rural food insecure 10% $500 million 1.2% program (5,6 million HIMO et 1,4 million (80% public works; 20% transfer) unconditional transfers) South Africa social grants 15 million beneficiaries n.a 30% 3.5% Kenya OEV 130,000 households 3% US$31 million 59 117. There are several possible components to funding a national safety net, including: a. Increasing efficiency and effectiveness of existing social protection expenditures. Fiscal space for a safety net can be created through reallocation from lesser to higher priorities and from less to more effective programs (e.g. from food subsidies to cash transfers). Given overall low level of social protection spending not counting subsidies, making existing expenditures more efficient or re-orienting within the existing resource envelope offers only limited opportunities to increasing program coverage. Nonetheless, efficiencies are possible:  Better targeting can heighten poverty impact.  Linking transfers with more productive public works or increased utilization of health and education services can better help lift communities and households out of poverty over the medium and long-term even within existing expenditure levels.  Harmonizing across programs and integrating an often disparate set of small programs can streamline administrative costs, allowing existing resources to reach more people. b. The most significant opportunity is shifting from the use of general price subsidies, particularly in response to economic shocks, to more targeted social protection interventions. Such subsidies routinely account for between 2-4% of GDP, largely benefit better-off households more, and create economic distortions. c. Increasing the allocation of domestic resources to social protection. A key challenge for African countries is to raise the share of domestic revenue financing and to increase the share going towards social protection. Economic growth has generated some increase in fiscal space in Senegal over the last decade. In looking forward, some real fiscal space is expected to increase in the national budget (Table 28). If half of this could be claimed by a national safety net it would create a sufficient funding base. However, this estimate of fiscal space may be optimistic given the deficits run in the national budget in recent years. d. Mobilizing external assistance will continue to be important, particularly as a vehicle for introducing reforms and testing approaches. Donor funding remains a core part of social protection financing in Africa, particularly in countries were aid dependency is high. The ability to attract additional external resources to fund a national safety net will depend on (i) whether Government prioritizes these expenditures; (ii) the ability of donors to harmonize around a consolidated national safety net instead of multiple individual programs, and (iii) perspectives for the integration of safety net funding into budget support operations. While additional external resources will be important, over time, the Government of Senegal should take responsibility for a growing share of national safety net funding. Table 28: Fiscal Space Projections 2012-2016, in FCFA billions 2012 2013 2014 2015 2016 Total Revenue and Grants 1714 1838 1978 2130 2262 Projected Real Increase in Fiscal Resources 126 128 137 149 129 Source: IMF Statistics. 60 V. Conclusions and Next Steps 118. Senegal will certainly continue to face internal and external shocks. With much of the rural area dependent on variable rainfall in a drought prone zone, rural shocks will continue to occur with predictable frequency. Senegal’s economy is increasing open, which has helped drive significant improvements in poverty reduction but also makes Senegal vulnerable to global economic shocks, as seen over recent years. These shocks will hit the chronic poor and also create transient poor. These populations will require safety net measures more effective, less distortionary and more efficient fiscally than the general subsidy mechanisms adopted over the last ten years in Senegal. 119. Safety net reforms should allow for increased coverage, greater impact and improved efficiency of safety net expenditures. A national safety net system should include a institutional platform that integrates coordination mechanisms at the central regional and local levels, a common system for targeting and registration of beneficiaries, and a packet of interventions that place an accent on the productive nature of safety nets and reinforce their contribution to economic growth. This packet of interventions should consider the following basic components: (i) cash transfers to promote human capital and targeted to the most vulnerable households, (ii) generation of temporary employment and creation of local infrastructure in poor communities through a public works safety net program that uses labor intensive methods timed to the moments in the year of highest vulnerability and targeted to low- skill laborers, and (iii) more integrated social assistance services will-targeted to high risk groups that require an added level of assistance (orphans and high risk children, disabled etc.). 120. Reforms of this nature to establish a more effective national safety net will be better able to develop a financially sustainable fiscal framework over the medium-term. A national safety net of this nature could cost between 1 and 2 percent of GDP depending on scope and funding. Unlike many countries in Africa, there is some element of fiscal space possible in Senegal if the Government prioritizes safety nets and reforms are carried out to transfer some areas of current spending like general price subsidies, to a more productive safety net. 121. A safety net of this magnitude would have a direct impact on poverty reduction and achievement of the MDGs. Increased access to schooling and health services, protection and accumulations of assets, and avoidance of negative coping strategies will affect the well-being of current Senegalese and the generation to come. 122. While these conclusions point to several likely safety net components, the main objective of this report was to synthetize the main messages of the technical notes . It was not expected to include a detailed discussion of specific program options. As part of the ongoing social protection policy dialogue, and based on the finding of this report, the World Bank team has already started much more in-depth discussion on policy reforms under consideration by the Government. 123. Suggested future analytical work to bolster these recommendation include:  Developing a more detailed view of possible implementation mechanisms and approaches to targeting  Further study the delivery mechanisms for safety net programs on the basis of past and ongoing cash transfer programs, and in particular such operational issues as the feasibility of modern fund transfer mechanisms like cell phones;  An in-depth look at the potential phase out of general subsidies; 61  An institutional review of social protection coordination mechanisms with a view of creating an integrated technical and institutional platform rather than the present series of isolated and ad- hoc arrangements; and  Development of a full medium-term expenditure framework for social protection in Senegal once priority programs and coverage targets have been established by the Government. This will require donor buy-in given the large share of external resources in social protection. 124. The Government has recently made steps to expand safety nets in the country, informed by the technical notes from this World Bank TA program. Of note, the President appointed a Délégué General for Social Protection in November 2012 who reports directly to him. His role and responsibilities include setting strategies, coordinating programs, establishing the Caisse Autonome pour la Protection Sociale Universelle (CAPSU) – a social protection funding mechanism- , and launching the Programme National de Bourses de Securité Familiale (PNBSF), a cash transfer program expected to reach 250,000 in 2017. 125. In response to this heightened priority on safety nets by the Government, additional efforts will be needed to ensure that Senegal develop an effective institutional framework for a national social safety nets system. It will be critical to move from scattered programs to a system approach. Coordination of policies and actions on the social protection agenda should seek the further harmonization and integration between programs and to support the development of common operational mechanisms underpinning a national safety net system (in particular as they relate to targeting and registering) and of consistent or compatible mechanisms (in particular in terms of the information systems of different programs). 62 BIBLIOGRAPHY Alderman, H. and J. Hoddinott (2008). “Growth-Promoting Social Safety Nets� 2020 Focus Brief on the World’s Poor and Hungry People. World Bank. Beegle, K., R. Dehejia, R. Gatti, S. Krutikova (2008). “The Consequences of Child Labor: Evidence from Longitudinal Data in Rural Tanzania�. World Bank Policy Research Working Paper 4677, World Bank. Bruno, M. M. Ravallion, and L. Squire (1996). “Equity and Growth in Developing Countries� World Bank Policy and Research Paper no 1563, World Bank. F.V. Soares, Ribas, R.P. “Evaluating the impact of Brazil’s Bolsa Familia: Cash transfer programmes in comparative perspective�, International Policy Centre for Inclusive Growth, 2006. Gilligan, D., Hoddinott, J., Kumar, N. and A. Taffesse. “An Impact Evaluation of Ethiopia’s Productive Safety Nets Program�, International Food Policy Research Institute, March 16, 2009. Hulten, C., 1996. "Infrastructure Capital and Economic Growth: How Well You Use It May Be More Important than How Much You Have." NBER Working Paper 5847, Washington, D.C. International Monetary Fund 2005. “Senegal: Selected Issues and Statistical Appendix� Country Report No. 05/155, IMF, Washington, D.C. International Monetary Fund 2008. World Economic Outlook 2008, “Financial Stress, Downturns and Recoveries�, IMF, Washington, D.C. International Monetary Fund 2010. “IMF Statistics�, IMF, Washington, D.C. Ivanic M, and W. Martin, 2008. Implications of Higher Global Food Prices for Poverty in Low-Income Countries. World Bank, Washington, D.C. Mankiw, G., D. Romer, and D. Weil. “A Contribution to the Empirics of Economic Growth� The Quarterly Journal of Economics, Vol. 107, No. 2. (May, 1992), pp. 407-437. OECD.,�Social Protection, Poverty Reduction and Pro-poor Growth� OECD DAC 2009. World Bank. “World Development Report: Infrastructure for Development�, Washington, DC, 1994. Sall, M., G. Dramani, Y.-M. Prevel, Y. Kameli, S. Fall, 2012. “Évaluation de l’Impact du Transfert en Espèces dans le cadre du projet nutrition ciblée sur l’enfant et transferts sociaux (NETS) », Laboratoire de Recherche sur les Transformations Economiques et Sociales, Institut Fondamental d’Afrique Noire (LARTES-IFAN), University of Dakar. 63 Soares, F.V and R.P. Ribas, R.P. “Evaluating the impact of Brazil’s Bolsa Familia: Cash transfer programmes in comparative perspective�, International Policy Centre for Inclusive Growth, 2006. Sumarto, M. “Social safety nets and economic transition in Indonesia: Paradox of social services�, JOAAG,Vol. 2. No. 1, 2007. United States Government , 2009. “Senegal: Country Commercial Guide 2009�, United States Mission to Senegal. World Bank, 1994. “World Development Report: Infrastructure for Development�, Washington, DC. World Bank 2006. “Managing Risks in Rural Senegal�. Human Development Department II, Africa Region. Report No. 33435-SN. March, 2006. World Bank, 2008. Sénégal: Diagnostique de la pauvreté. Rapport préliminaire. World Bank, Washington. World Bank, 2009., “ Senegal: Rapid Response Child-Focused Social Cash Transfer and Nutrition Security Project�, Report No. 47740-SN, Human Development II, Country Department AFCFl, Africa Region World Bank Social Safety Net Assessment Background Papers: “The Government of Senegal’s Response to Shocks over the Past Decade�, Julie Van Domelen, November 20, 2011. “The Contribution of Social Protection to Economic Growth – Perspectives on Senegal�, Julie Van Domelen, January 2011. “ Review of Safety Net Programs in Senegal�, Aissata Fall, November 2011. “Issues and Options in Targeting and Social Transfers in Senegal » Damien Echevin, August, 2012. “Institutional Arrangements in Safety Nets in Senegal�, Julie Van Domelen, April 2012. “Targeting Performance of Water and Electricity Subsides in Senegal�, Damien Echevin, September, 2012. “Profile of Social Protection in Senegal: Analysis of the ESPS 2011�, Mame Marame Diop, October, 2012. 64