Report No. 33435-SN Senegal Managing Risks in Rural Senegal A Multi-Sectoral Review of Efforts to Reduce Vulnerability March 30, 2006 Human Development II (AFTH2) Africa Region Document of the World Bank M D G s Millennium Development Goals M F I s ' Micro-Finance Institutions MICS Multi-Indicator Cluster Survey MOA Ministry o f Agriculture MOE Ministryo fEducation MOH Ministry o f Health MFFDS Ministry o f Women, Family and Social Development NAC National Agriculture Caisse PAREP Programme d'appui ci la re'duction de la Pauvrete' N G O s Non-Governmental Organizations PASEC Programme d 'Analyse des SystBmes Educatf PDEF Programme De'cennal de 1'Education et de la Formation PIV Pe'rimBtres Irrigue's Villageois PLCP Project to Fight against Poverty PNDS NationalDevelopment HealthPlan PRSC Poverty Reduction Support Credit PSIA Poverty and Social Impact Assessment AFDS Agence de Fond de De'veloppement SONACOS Socie'te'Nationale des Ole'agineux du Se'ne'gal UEMOA Union Economique et Monktaire de I'Afrique de I'Ouest UNACOIS Union Nationale des Commerqants et Industriels du Se'ne'gal UNCAS Union Nationale des Coope'rativesAgricoles du Se'ne'gal UNDP United Nations Development Program UNICEF UnitedNation Fundfor Children WFP World FoodProgram This report was prepared by a team consisting o f Julie Van Domelen (TTL, Lead Social Protection Specialist, AFTH2), El Hadj Adama Toure (Agriculture and Rural Development Specialist, AFTS4), Moukim Temourov (Human Development Economist, AFTH2), Atou Seck (Education Economist, AFTH2), Leopold Remi Sarr (Education Specialist, SASHD), Julien Bandiaky (Economist, AFTP4), Maguette Niang (Consultant), Djibril Ndoye (consultant), Gisele Munge (Intern) and Astou Diaw-Ba (Program Assistant, AFCSN). Peer reviewers were Lant Pritchett (SASHD), Valerie Kozel (HDNSP) and Daniel Sellen (SASAR). The views expressed herein are those of the authors and do not necessarily reflect the opinion or policies o f the World Bank or its affiliated institutions ... 111 Table of Contents EXECUTIVE SUMMARY ............................................................................................. vi CHAPTER I:Objectives and Context of the Study ..................................................... 1 CHAPTER 11: Risks and VulnerableGroupsin the RuralAreas ............................. 5 A. 5 Access to Critical Infrastructure and Services.............................................................................. Characteristics o f the Rural Population and Determinants o f Rural Poverty ............................... B . 9 C. Estimation of Frequency and Severity o f Risks and Shocks ..................................................... 10 C.1. NaturalRisks and Shocks............ ................................. ............................. 10 C.2. Economic Risks and Shocks ...................................................................................................... 16 C.3. Health Risks and Shocks ............................................................................................................ 18 C.4. Risks and Shocks to Educational Accumulation........................................................................ 25 C.5. Overview of Risks. Multiple Shocks. Vicious Cycles. and Seasonal Concentrations................28 Social Risks................................................................................................................................ D . CHAPTER 111: Informaland PrivateRiskManagementMechanisms Overview o f Householdand Community Social Risk Management Strategies ......................... ...................34 29 A. 34 B. C . Membership inSocial Networks and Solidarity Groups ............................................................ 35 D. Labor Migration and Remittances .............................................................................................. 37 Insurance.................................................................................................................................... Savings and Credit ..................................................................................................................... 40 E . 43 CHAPTER IV: Review of PublicPolicies and Expenditureson Social Risk Managementin Rural Areas .......................................................................................... 46 A . Overview o f InstitutionalFramework and Taxonomy of Programs........................................... 46 B. C . Managing Critical Agricultural Risks ........................................................................................ 53 Social Assistance. Safety Nets and Disaster Management ......................................................... 47 D. Managing Critical HealthRisks................................................................................................. 59 E. Managing Critical Education Risks............................................................................................ 67 F 72 Targeting o f Infrastructure and Service Programs inRural Areas ............................................. Employment Generation ............................................................................................................ G .. 76 H. Credit and Savings ..................................................................................................................... CHAPTER V: Key PolicyReformOptionsand ExpenditurePriorities ..................82 78 List o f Boxes B o x 1 : The Most Vulnerable Groups..................................................................... 9 B o x 2 : Senegal's Ecological Zones........................................................................ 10 B o x 3 : Micro - Finance inRural Areas ................................................................... B o x 4 : Extending Health Insurance to Rural Areas - Evidence o f Impact........................... 43 45 52 B o x 6 : Managing Risks Associated with Desert Locust Swarms .................................... B o x 5 : Who Receive Assistance., ........................................................................ Box 7: Conditional Cash Transfers to Boost School Enrollment and Health Utilization............56 Box 8: Index-Based Agricultural Insurance.............................................................. 83 Box 9: International Experience inWaivers and Exemptions for Health Services..................87 89 Figure 1 :Rural-Urban Distribution among poverty quintiles (%) ................................................................. List o fFigures 6 Figure 2 :Per-Capita Daily Expenditures by quintiles, rural/urban (CFA)..................................................... 7 Figure 3 :Average Annual Rainfall 1978 - 2003 (inmillimeters)................................................................ Figure 4 : Primary Sector Growth Rate, 1998-2004...................................................................................... 11 16 Figure 5 : HIV Prevalence among Pregnant Women, by Region, 2002 and 2003 (average and confidence Figure 6 : Seasonal Concentration o f Rural Risks......................................................................................... interval)................................................................................................................................................ 23 31 Figure 7: Severity and Frequency of Risks inRural Areas ........................................................................... 33 Figure 8 : Share o f Households that Belong to Group by Poverty Quintile .................................................. 36 Figure 9 : Poverty Targeting o f Rural Investments ....................................................................................... 77 iv Listof tables Table 1: Examples of Social Risk Management Instruments.......................................................................... 2 Table 2 : Senegal Population and Number o f Rural Poor, 2002 ..................................................................... 5 Table 3 : Poverty and inequality inSenegal, 1994 - 2002................. ................................................ Table 4 : Rural Household Characteristics ........................................ ................................................ 86 Table 5 : Production Effects of Drought....................................................................................................... 11 Table 7 : EstimatedLosses in Grain ProductionDue to Locusts, 2004 ........................................................ Table 6 : Annual Groundnut Harvest (inkilos) per Household, by poverty level....................................... 12 14 Table 8 : Evolution of Animal Stocks 1999 15 Table 9 : Frequency and cause of livestock depletion, 2002 in %................................................................. - 2003 (in `000 heads).............................................................. Table 10: Evolutionof Purchasing Power (PP) and Price Indexes, 1990-2002 ............................................ 15 18 Table 11......................................................................................................................................................... 19 Table 12 : Selected Maternal Health Indicators .............................. ........................................................ 20 24 Table 14 : Difference Between School-Aged Population and Enrollments, by region ................................. Table 13 :Place of consultation if sick and ifconsulted, rural all ages (%).................................................. 26 26 Table 16 :Determinants o f Rural Primary Schooling Outcomes .................................................................. Table 15 : School Attendance by Rural Children 7-12 year olds .................................................................. 27 Table 17 : % Children Working inGroundnut Production, by gender and poverty level ............................. 28 Table 19 : Share o f Community Members Affected by Shocks, by type ..................................................... Table 18 : Effect o f Multiple Negative Shocks ............................................................................................. 30 30 Table 21 : Presence o f Associations inRural Areas, by type ........................................................................ Table 20 ;Summary Estimates o f Rural Populations at Risk........................................................................ 32 35 Table 22 : International Remittances to Senegal, 1994-2004........................................................................ Table 23 : Sources o f Loans for Rural Households....................................................................................... 41 38 Table 25 : Comparative Safety Net and Social Insurance Expenditures as % o f GDP ................................ Table 24 : Estimated Annual Expenditures on Social Assistance and Safety Nets ...................................... 48 49 Table 26 : Estimated Annual Expenditures on Disasters and Agricultural Risks.......................................... 55 Table 27 :Financing o f Funds for Security o f Agricultural Credit 1998-2004, inCFA .............................. 57 Table 28 : Expenditures from the Funds for Calamities, Bonification and Guarantee 2004 Table 29 : Health Sector Financing, million CFA......................................................................................... ......................... 57 Table 30 : Health infrastructure and personnel 1998-2002 .......................................................................... 60 61 Table 3 1 : Net Change inHealth Infrastructure and Personnel, 1998-2002................................................. 61 Table 32 : Dissatisfaction with health consultations and reasons................................................................. 62 Table 34 : Financing ofthe Strategic Plan Against HIV-AIDS 2003-2004 ................................................. Table 33: Distribution o f ITBNs,by region.................................................................................................. 63 64 Table 35 : Interventions supported by CNLS 2004, by Activity and Urban-Rural ...................................... 65 Table 36 : Financing o f the Nutrition Enhancement Project, 2003-2004 ..................................................... 66 68 Table 38 : Cost per Student o f Primary Education, rural and urban.............................................................. Table 37 : Estimated Direct Annual Expenditures on Rural Primary Education(1)..................................... Table 39 : Opening o f New Classrooms inRuralAreas, by Sponsor and Poverty Level ............................ 69 69 Table 41 : Share o f rural students who said they received school meals in2003......................................... Table 40 :Rural: Problems at school, just for those currently attending (age 7 to 12 only), ........................ 70 71 Table 42 : Estimated Annual Expenditures on Selected Employment CreationPrograms .......................... 73 Table 43 : Synopsis o f Expenditures onRural Social Risk Management, billion CFA ............................. 81 93 ANNEX B: Statistical Annex................................................................ ANNEX A: Description of MainData Sources ............................................ 94 B~LIOGRAPHY ................................................................................ 117 V EXECUTIVESUMMARY The main objective of the study is to provide analyticalsupport to the Government of Senegalin the design and implementationof policiesto reducethe vulnerability of the rural poor. Despite relatively robust economic growth since the mid-199OYs,the development o f Senegal's rural areas continues to lag further behind, hindered by structural vulnerabilities, unequal access to basic public and private services, and a limitedrange o f strategies to prevent, mitigate and cope with risks. Although economic growth in recent years has reduced poverty in Senegal, these gains have been less for rural populations, home to 6 million out of the total population of 10 million. The rural economy remains largely agrarian and sixty-five percent o f the rural population i s poor. Better-off rural households are either less dependent on agriculture inthat they are more likely to holdpublic sector employment or be engaged in services and commerce, or, if engaged in agriculture, have greater concentrations o f land and/or animals. Despite the greater poverty in rural areas some households have found effective strategies for economic growth within this environment. Natureand Severity of Risks Naturalrisks are intertwinedwith the agriculturalnatureof the ruraleconomy and its Sahelian setting. Cyclical droughts reduce agricultural production and decimate herds. However, the impacts o f these shocks are highly variable depending on the agricultural season, the agro-ecological zone, the type o f crop and whether land i s under irrigation. Over the last 25 years, Senegal experienced six years that could be characterized as major precipitation shocks with late onset o f rains, irregularity in spatial distribution, and an early end o f the rainy season, Indrought years, loss o f production o f groundnuts has beenvalued between US$35-135 million and for millet/sorghum between US$24-60 million. Despite the overall fragility o f precipitation, certain localities are affected by floods (Saint Louis, Matam, Louga, Tambacounda and Kolda) causing deaths, displacement, destruction o fproperty and important loss o f crops and animals. Other natural risks involve pests, plant and animal diseases. Senegal periodically experiences locust swarms, most recently in 1988 and 2004. Unlike general drought, locust damage i s geographically sporadic. An estimated 20 percent of rural households were affected in 2004 with grain losses estimated at less than 50,000 tons (about US$14 million), plus extensive damage to livestock pastures in certain regions. Other endemic infestations affecting agriculture and livestock are less spectacular but can cause significant production losses, including grain-eating birds, aphids, white fly and various animal diseases. Prevention and treatment o fparasitic infestations remains limited. Economicrisksin the rural areas are relatedto those of an open economy, changes in the agricultural production and commercialization systems, and income and purchasingpower risks. While employment is a major economic risk in urban areas, this is not as relevant for rural Senegal where most o fthe labor force is self-employed or engaged in informal or part-time activities to diversify income. Groundnuts and vi groundnut products, the main agricultural export crop, have seen exports decrease over time partly because of poor political management o f sub-sector reforms since the mid- 9Os, as well as depletion in the quality o f seeds and soil and downward trends in world prices over the last decade. The transition to a more liberalized market structure has caused dislocations, including lack o f access to critical inputs like seeds and fertilizers, The 1994 devaluation of the CFA spurred economic growth overall, but the growth was not strictly `pro-poor' as gains were less apparent for rural producers due to a drop inthe real value o f crops after the devaluation and an erosion o f rural purchasingpower. Rural populationsare more exposedto healthshocks and face greater constraintsin accessing health services. Malaria is the most widespread health issue facing Senegal. Nationally, malaria accounts for 31percent o f morbidity and 32 percent o f mortality, and is more likely to strike rural populations. It i s estimated that the direct and indirect effects o f malaria reduce annual GDP growth by 0.35 percent in Senegal, Infant mortality, maternal mortality and child malnutrition are also much higher in rural areas. Contributing the high maternal mortality rates in rural area, 45 percent o f rural women give birth either alone or with a family member or friend, while only 12 percent o f urban births took place in such risky settings. Although Senegal has one o f the lowest HIV prevalence rates in Sub-Saharan Africa, at an estimated 1.5 percent o f the adult population, risks to rural populations come from transmission through labor migrants, a lower level of knowledge about prevention and treatment options, and less access to testing and life-saving treatment. In the face of a health shock, rural populationstend to use health services less than urban dwellers. When sick, two-thirds o f rural households sought out health services versus three-quarters for urban residents, declining by poverty level. More importantly, rural residents access lower quality health services, including greater use o f traditional healers, and less likelihood o f using public and private hospitals and clinics. Financial access appears to be a significant factor, along with physical access and perceived quality of care, For those rural residents who said they had been sick, fully 40 percent did not seek treatment because it was too seen as too costly, although this i s interrelated with both distance and quality. Risks to educationalaccumulationincludenever enrolling, droppingout duringthe course of the year or between years, or not learning the sufficient skill base. The estimated gross primary enrollment rate i s 95 percent in urban areas and 70 percent in rural areas. This translates into over 300,000 rural children o f primary school age that are not inschool, one-third of whom are inthe Diourbel region and other significant numbers in Kaolack and Louga. There is little difference in attendance rates between poverty quintiles (not true of urban areas). As constraints to physical access have begun to ease through extensive investmentsinrural classrooms, demand-sidefactors and issues related to quality and cost are coming more to the forefront. Factors that appear to most impede rural primary educational attainment are illness o f the child, lack o f perceived relevance (due to a combination of low returns to primary education in rural areas, low levels o f parental schooling, and parental preferences for religious and Arabic education), and child labor and economic decisions made both intimes o f good and badharvests. vii Regionalconflict inthe Casamance and communityand householdlevel occurrenceof theft, property disputes and violence are the principalsources of social risks in rural Senegal. Conflict in the Casamance since the early 1980s has displaced populations, left people disabled from land mines and destroyed infrastructure and resulted in a loss o f tourism receipts and private investments and a slow down in public investments. Nationwide, livestock theft i s a growing problem and heightens rural vulnerability given the use o f livestock as a store o f wealth and a hedge against shocks. Most villages experience multiple negative shocks. The greater the number of shocks, the greater was the likelihood that villagers perceive a worsening situation. Nonetheless, even within the same village, these shocks hit households differently. With its close ties to agricultural production and weather patterns, there i s an important seasonal aspect to rural vulnerability. This critical period, known as the `hungry season', exacerbates vicious cycles and is a time o f heightened vulnerabilities. Peak labor demands occur at a time o f highest malnutrition and malaria. The malaria peak also coincides with the rainy season when transport for health emergencies i s most difficult. Lack o f available income in the pre-harvest period coincides with heightenedneed for cash to pay for health treatment and may delay the entrance o f childreninto school as fees can only bepaid after the harvest. InformalandMarket-BasedRiskManagementMechanisms Households and communities seek to decrease their vulnerability through long- standingpracticesof risk managementinthe face of the vagaries of lifein the Sahel. Informal strategies to prevent shocks from manifesting take many forms including vaccinating children and livestock, irrigation, clearing to prevent bush fires and so on. Strategies to mitigate the impact o f a shock center on diversification o f production and revenue, Within rain-fed agriculture, throughout the 1990s there has been diversification away from groundnuts and other cultivations that have been most affected by adverse weather patterns and into such crops as watermelon and nie`be`. Coping strategies include reduction in quantity and quality o f food eaten, sale o f animals, use o f traditional medicines, non-enrollment of children inschool and recourse to temporary migration. Social networks and local organizations are the frontline of defense in the face of shocks to households. Religious organizations, women's groups, informal savings clubs (tontines) and culture and sport organizations are present in the vast majority o f rural communities and include all income categories equally. Associations around economic advancement, education water and health management are less prevalent and may be more exclusionary in their membership. Rural households contribute financially to the maintenance of these networks, across all income groups. Households also seek to diversify against rural risks by out-migration. Even though marriages and other family events are the main reason for rural migration, a significant amount of migrant households cite economic reasons (e.g. lack o f jobs, a new job in a different location, and access to land). International migration has grown in importance, with an estimated half a million Senegalese residing abroad. Statistically, migration to other African nations accounts for the largest number o f Senegalese emigres. From this V l l l ... migration, international remittances to Senegalhave grown steadily and account for about 6 percent o f Senegal's GDP. Anecdotal evidence suggests that there are regions where this can be an important, and even predominant, income source. Informal systems to transfer remittances tend to be lower cost than using formal financial systems, though risks are higher. Formal transfers are largely affected through the Post Office with its extensive rural presence, and increasingly micro-finance institutions (MFIs). Among the market-based risk management mechanisms, modern financial instruments, including savings and credit, are essential tools for consumption smoothing and growth. Accumulation o f savings in financial stores allows for more efficient consumption smoothing in the face o f shocks, self-insurance against risk, safekeeping o f wealth accumulation and increased credit worthiness. In addition to financing investment and growth, credit i s crucial to smoothing consumption inthe face o f emergencies. Most rural savings are still `on the hoof`, although formal and informal financial savings are growing in importance. The micro-finance industry in Senegal has flourished in recent years. The rural areas are an important segment o f the market for several o f the larger MFIs. However, most funding o f agricultural credit i s provided through the semi-public commercial bank CNCAS, Insurance helps households mitigate the adverse effects on income and consumption of health crises, death, disability, old-age, crop loss, business failure and losses through theft, fire and accident. Formal insurance mechanisms are extremely limitedin Senegal, and virtually non-existent in rural areas. The most active segment of the insurance market involves mutual health associations, which have increased significantly since pilot efforts in the mid 1980s and now count over 90 nationwide with 66,000 paying membersand 220,000 total beneficiaries, researchers estimate that well under 10percent o f total beneficiaries are in rural areas, mostly involved in community-based health mutuals. There appears to be potential to extend health insurance through the larger-scale producer cooperatives and federations as well as pilot efforts through MFIs. However, health insurance only works for groups o f people with economic resources to pay the premia and where the availability o f quality services incites the member to pre-purchase some future stream o f services. PublicSector PoliciesandExpendituresto ManageRisk Risk management activities are distributed across several ministries and agencies, including: (a) Ministry of Family, Women and Social Development, and the Ministry o f National Solidarity (MFFDSISN); (b) Ministry o f Health and Medical Prevention; (c) Ministry of Education; (d) Ministry of Labor and Public Administration; (e) Ministry of Small and Medium EnterpriseDevelopment, Micro-Finance and Women Entrepreneurs; (0 Ministries in charge o f agriculture, rural water and livestock; (g) Ministry o f Interior; and (h)MinistryofYouth as well as selectedprograms underthe Prime Minister's Office. The Government spent US$43 million in 2004 on social assistance, safety nets and disaster management, including community development programs of the MFFDS. The focal point for disaster management is the Civil Protection Department o f the Ministry of the Interior, supported by the MFFDS/SN and NGOs like the Red Cross to provide ix emergency supplies, cash and rebuilding materials. The Ministry o f Agriculture responds to the after-effects on production capacity. Responses to floods in 2003-2004 were quickened with the entry into disaster relief o f the recently created National Solidarity Fund. But, the parallel approaches of community committees and top-down distribution through local officials can reduce efficiency and create overlap. For social assistance and safety nets, the Ministry of Women, Family and Social Development (MFFDS) and the Ministryof NationalSolidarity (MSN) have the basic mandate to protect the most vulnerable families and individuals. MFFDS has a long list o f social assistance and community development programs targeting women, disabled, children at risk, elderly, and vulnerable rural communities. At the MSN, the National Solidarity Fund finances disaster relief and intends to expand into infrastructure, micro- credit and social assistance. The Commissariat for Food Security (CSA) ensures food security through food distribution and monitoring o f country conditions, including extensive rice distribution largely through religious leaders. Programs use different approaches and criteria, with little coordination and monitoring. To manage agricultural risks many projects in the investment budget support the generaldevelopment of agriculture, contributingin various ways to the preventionof shocks. The bulk o f investments in agriculture and livestock are to increase productivity through irrigation, improvement o f technologies, and support to the development o f producer organizations. Inigation-related investment projects alone account for between 40- 60 percent of the Ministryo f Agriculture investment budget. In addition, from the late 1990s a series of financial mechanismswere put in place to mitigate and cope with risks to agriculture and to seek to ensure an adequate flow of credit to the sector. Collectively known as the Funds for Security o f Agricultural Credit (Fonds de Sbcurisation du Cre'dit Agricole FSCA), they include a guarantee hnd, and interest rate subsidy on agricultural credit through CNCAS and a fund to respond to agricultural calamities. From 1998-2004, 28 billion CFA (US$56 million) was allocated through these funds to mitigate and cope with shocks to rural producers. It is difficult to assess the overall impact of the FSCR mechanisms. The various funds have kept some measure o f credit flowing at a time o f,stress, but, they do not properly address longer-term sector creditworthiness and solvency issues. The public sector, includingcentral government,donors and local government,spent US$176 million on health in 2004, equivalent to 2 percent of GDP. Despite growth in the number o fhealth centers and healthposts, health services have had difficulty inkeeping pace with population growth in recent years. Rural residents perceive an overall lower quality o f service, centered on high cost and ineffective treatment. These drawbacks affect management o f health risks like control o f maternal mortality. In terms of malaria prevention and treatment, incremental funding has been mobilized, but delivery of bednets and revisions to treatment protocols are limited by weaknesses in the community health system inrural areas and inability for the rural poor to pay for services. The Government has scaled up efforts to prevent and treat HIV-AIDS, though rural populations were less likely to benefit. Malnutrition shows signs o f being effectively addressed through the X Nutrition Enhancement Project, though scalability inrural areas i s constrained by cost and presence o f technically capable NGOs. In 2004, the Government and donors spent about US52 million on rural primary education.The Government has made significant efforts to increase the supply of primary education in rural areas. Despite this investment, incomplete schools are still more frequent inrural than urban areas and the M O E continues to have difficulty retaining rural teachers and rural children are less likely to receive textbooks. School lunches provide an incentive for families to send their children to school, but their coverage i s limited, costs can be significant depending on level o f benefit and there i s evidence that the poorest households are less likely to receive them. The M O E i s also attempting to address demand-side factors by diversifying the educational offerings, including instruction in Arabic incertain regions. Given itstransversalnature, many ministriesand agencies are involvedin activitiesto generate employment, but programsthat specifically seek to create jobs amounted to about US$ 4 million in 2004. These are all fully domestically-financed and most have been created within the last few years as the pressure to address employment, particularly among youth, have increased. There is little systematic attention to the creation o f temporary employment through public works and infrastructure. However, a review o f the targeting o f several mutli-sectoral local development projects found that programs like AFDS and PNIR have been effective at channeling resources to small-scale infrastructure inthe poorest rural communities. Interms of developingsavings and credit mechanisms, more Governmentefforts have sought to use existingMFIs to channelsubsidizedcredit to specific groups. This may work at cross purposes with the objective o fbuildingsustainable financial services for rural areas. The Government invests little in direct support to the development o f the industry, thoughdonors have made substantial direct investments. Study Recommendations The PRSPand sectoralpoliciesidentifymain risks facing rural Senegal. But, there has been less attention to whether programs and expenditures line up with these risks, who benefits and the impact on reducing risks. It is important for policymakers to identify the risk being addressed as well as the justification for intervention. Does the mechanism proposed adequately address the risk? Who i s covered or excluded? H o w much prevention, mitigation or coping is `bought' with this public spending? Recommendationsto enhance the efficiency and effectiveness of response to natural disasters center on integrating disaster responses and better use of early warning systems. Responses to risks tied to the person needto be integrated with response to risks tied to production to improve overall coherence. Reaction time and cost can be reduced through early warning systems, including clear guidelines on responsibilities. Methodologies for estimating damages should be systematically applied to better identify those affected and structure a response. And, the high cost o f coping with natural xi calamities (for example locusts in 2004) justifies a greater investment in the prevention of shocks. The core challengefor extendingand improvingsafety nets and social assistance is to create mechanisms tied to programmatic objectives to reduce vulnerability. This involves the consolidation o f the various transfers to vulnerable groups (women, elderly, disabled etc.) into a coherent set o f interventions and a transition from the current `unconditional transfer' mode to a tighter focus on the truly vulnerable, use o f improved targeting tools and more systematic monitoring o f outputs and impacts. The central role of the State as ultimate risk management mechanism for rural production poses several problems, includingfiscal sustainability, moral hazard and overall efficiency. As a Sahelian country with 95% o f land under rain-fed agriculture, there are three main strategies for mitigating the climatic risk: diversification, irrigation, and non-farm employment. Better management o f agricultural risks can be achieved through (a) the prevention o f shocks through developing irrigation, including small scale groundwater irrigation in the Groundnut Basin, and facilitation o f access to appropriate technologies; (b) recourse to private sector services to manage the credit securization funds; (c) investigation o f the applicability o f agricultural insurance instruments, like weather-indexed crop insurance; (d) reconsideration o f the policy o f blanket rural debt forgiveness; and (e) development o f better informational tools to underpin agricultural risk management. For education, addressing rural risk requires demand-side strategies as well as continuing the good momentumin building and staffing rural schools. Expansion o f school lunches may be an effective strategy, as well as the diversification o f educational programs relevant for rural populations, including language o f instruction and local curriculum concerns. Options to address the trade-offs parents face between farm labor, and agricultural cycles ingeneral, and schooling will need to be explored. And, the school- based healthprogram requires betterrural coverage oriented to critical rural healthrisks. Health policies can boost prevention of critical rural health shocks through: (a) an expansion in the presence o f midwives at the health post level; (b) investing in better community-level services, including outreach equipment and materials to reinforce the referral system; (c) ensuring that proposed incentive system for workers i s effectively implemented to create more stable staff in rural areas; (d) international experience with fee waiver programs should be consulted to structure the proposed exoneration o f fees; (e) to prevent malaria through greater use o f bed nets, the full force o f rural associations should be enlisted, including associations who have more village presence, the social marketing strategy reassessed, and an early warning system installed inthe North; (f) effective models of interventions to prevent malnutrition should be mainstreamed and scaled-up; (g) HIV- AIDS efforts should be targeted to specific risks and characteristics o f rural populations and treatment models diversified to allow greater access by rural residents; and (h) community-basedhealth insurance schemes expanded. xii The biggest gaps in government policies and expenditures are for rural income diversification and expanded financial services. While most attention to date has been on improvements within agricultural production, there has been relatively less attention to diversifying to non-agricultural sources o f rural income. To enhance income diversification, a greater employment focus on rural public works and infrastructure development i s needed to help the poor cope with shocks and agricultural cycles. In addition, an evaluation o f the lessons learned from income-generating projects in rural areas is urgently needed to better orient the vast array o f projects and expenditures to this end. Expansion of savings and credit mechanisms i s crucial. Existing practices o f subsidized credit lines need reconsideration since they make MFIs more vulnerable to delinquency, erode the credit culture, and distort the market for sustainable interest rates. Options to promote expansion o f MFIs into rural areas, drawing on international best practice and country experience, should be analyzed. The Government should also explore avenues to optimize the impact o f remittances. The Post Office and MFIs will be key points of service to pay out remittances, although Western Union and others also pursue links to retail networks (shops, pharmacies, gas stations, etc) and this could be explored further. Competition will help reduce the fees for money transfers. Looking at risks from a multi-sectoral perspective helps identify vicious cycles that will require collaboration between sectors. The revision o f its Poverty Reduction Strategy Paper and development o f a National Social Protection Strategy in the coming months affords the Government an opportunity for making these linkages more explicit. This is also an opportunity for redressinggeographical imbalances through the scaling up of public investments and expenditures in rural areas. The transition to more sophisticated risk management instruments is an integral part o f the development process. Enhancing rural households' ability to prevent, mitigate the impact o f or recover from the frequent shocks they face will be a key determinant inSenegal's attainment o fthe MDGs. ... X l l l CHAPTERI:Objectives and Context of the Study ConceptualBackground 1. Despite relatively robust economic growth since the mid-l990's, the development of Senegal's rural areas continues to lag further behind, hindered by structural vulnerabilities, unequal access to basic public and private services, and a limited range o f strategies to prevent, mitigate and cope with the risks faced by rural households. Reducing vulnerabilities o f the rural population i s a key development challenge in the medium-termand centralto Senegal's ability to meet the MDGs. 2. Rural areas are exposed to a variety of shocks that impede households from accumulating assets and human capital to help them rise out of poverty. Covariate shocks that affect the whole community or country, such as drought and commodity price shocks increase income volatility. Idiosyncratic (household-level) shocks include health crises, life-cycle events associated with birth, old age and death, social risks associated with crime, and economic risks o f business failure or loss o f migrant remittances. Vicious cycles between these risks heighten vulnerability. The multi-sectoral nature o f the issue requires an integratedperspective to better manage risks inrural Senegal. Definitionof Vulnerability,SocialRiskManagementand Social Protection 3. Vulnerability is defined in terms of particular risks and exposed populations. Vulnerability analysis seeks to identify the likelihood that someone, whether poor or not, will experience a significant loss in welfare as the result o f a change in circumstance (or shock). This looks both at the nature o f the forces acting upon a person's welfare and the underlyingability to protect themselves from the risks they face. 4. Poverty andvulnerability are intertwined. Beingpoor increases vulnerability by the very nature o f limited assets, less recourse to savings andborrowing intimes o f stress, reduced human capital, and less income to pay for essential services. Vulnerability in turn reinforces poverty. Householdsmore likely to experience shocks to income are more likely to be poor. Poverty and risk both impact on households' abilities to improve their well- being inthe future. 5. The poor are most exposed to a broad range of risks and have the fewest instrumentsto deal with these risks. With precarious underlyinghealthconditions, marginal land and less physical and financial access to social and economic services, the rural poor often experience irreversible effects from even small shocks to their consumption and well-being. While the patterns are different across countries, the poorest households strugglemost with shocks, adopting coping strategies which are costly in terms of both short-term and long- term well-being. Shocks may create poverty traps that militate against the accumulation o f human and physical assets by the poor and result inhousehold decisions about consumption and investment that underminethe foundations o f future economic growth. 1 6. A basiccategorizationof riskincludes: (a) Natural risks: drought, floods, land degradation etc. (b) Risksto humancapital (health and education accumulation): illness, injury, accidents, disability, epidemics (e.g., malaria), school drop-out etc (c) Life-cvcle risks: birth, maternity, old-age, family break-up, death, etc. (d) Social risks: crime, violence, theft, social upheaval, etc. (e) Economic risks: unemployment, harvest failure, business failure, resettlement, output collapse, balance o fpayments shock, financial crisis, currency crisis, etc. 7. Social risk managementpresents a framework for identifyingthe appropriate risk prevention, mitigation and coping strategies to minimize the adverse impact of shocks.' Looking at poverty reduction not only as a process o f accumulating stocks - o f wealth, education, health, savings - but o f addressing key risks to the process o f moving out o f (or avoiding moving back into) poverty, leads to an analysis o f strategies to prevent, mitigate or cope with those risks. Prevention strategies reduce the probability o f a down- side risk occurring. Mitigation strategies decrease the potential impact o f a future down- side risk. Coping strategies seek to relieve the impact o fthe risk once it has occurred. Risk management instruments may be informal (family or social group based), market-based (likeprivate insurance), or public arrangements. Table 1providesa categorizationof social risk management approaches along the continuum by type o f provider. Informal Market-Based Public RiskPreventior Drought-resistant crops Purchase o f bednets Disease and pest preventior Hygienefor diseaseprevention Private investment in Floodcontrol Proper feeding and weaning irrigation Policing Flood control Investmentsinfencing tc Preventive health programs Condom uselabstinencelfidelity prevent theft Expansion o f access to education Sound macroeconomic policies RiskMitigation Investment insocial networks Crop insurance Social security (reciprocal gift giving, Microfinance Support for extending membershipinsocial andreligiou Health & life insurance financial markets to the poo: organizations) crop diversification ; Support for permanent job Employment diversification Diversificationo f creation Holding o f real assets (livestock, - financial assets Disability policies gold) Financial savings Protection o f property rights RiskCoping Sale o f assets Selling o f financial asset Labor intensive public work Reducedfood intake Social assistance Withdrawal from school I I Disaster relief Curative health services I Source: Builds from Holzmann 2001 8. Social risk managementmeasures are publicinterventionsto assist individuals, households and communitiesbetter smooth incomeand consumptioncycles andlower the likelihood of devastatingloss of welfare and to provide support to the critically ' (Holzmann and Jorgensen, 1999 and 2001, Holzmann2003). 2 poor. This is a broad definition and encompasses programs, like flood control and malaria prevention. Such interventions are not commonly thought o f under the narrower categorization of social protection programs as pensions, social security, insurance and risk pooling, labor market interventions, social assistance and safety nets. This paper adopts the broader definition since issues such as access to preventive health care or decreasing vulnerabilities o f agricultural production are critical inrural areas. Social Risk Management within Senegal's PRSP 9. Senegal's PRSP 2003-2005 acknowledges the essential vulnerabilities of rural populations. Of the three main pillars o f the PRSP, the `wealth creation' pillar calls for reducing vulnerabilities in agriculture through rural water management, broader access to land for vulnerable groups, legal reforms and diversification o f income sources, as well as better animal health control and access to financing for livestock producers. The PRSP also seeks to expand rural savings and credit mechanisms and support employment through labor-intensive public works. The `capacity building and promotion o f basic social services' pillar emphasizes basic education for the most underprivileged strata o f society and reduced disparities betweenurban and rural areas, while improving quality overall, In health, provision of sufficient infrastructure, reduced expenditures for the poor, increased recourse to health insurance and increased prevention through community-based approaches are envisioned. Investmentsin water, transport and environmental quality will further improve health and human capital outcomes. Within `protection o f vulnerable groups', safety net and social assistance measures are framed in terms o f specific vulnerable groups, including children at risk, women, the disabled, elderly, displaced populations, unemployed youth, slum dwellers and the rural poor. Better targeting and greater emphasis on community-level delivery mechanisms i s a priority. 10. Senegal's first annual PRSP Progress Report for 2003 reported broad progress in its PRSP implementation, but also a need to reinforce actions for vulnerable groups, where little had been accomplished. The Government has created wealth with sound macroeconomic policies, improvements in the investment climate, and progress in long overdue structural reforms inthe groundnut and electricity sectors. Human capital has also received substantial attention through the implementation o f reforms aimed at strengthening public expenditure management and implementation capacity for delivery o f core social services. However, improvements inthe living conditions o f vulnerable groups have been less evident, with few o f the plannedpriority investments inthis area carried out. Objectives of the Study 11. The main objective of the study is to provide analytical support to the Government of Senegal in the design and implementation of policies to reduce the vulnerability of the rural poor. The study is designed to provide inputs into the development of the national social protection strategy and o f the country's second Poverty Reduction Strategy for 2006-2009, and more generally, to inform policies and programs in the rural areas. The particular focus on rural areas and the poorest segments o f society will help redress some o f the existing imbalances observed in sectoral strategies and public expendituresand it is inthe countryside that it will bemost difficult to meet the MDGs. The study seeks to answer the followingkey questions: What are the main risks and shocks facing rural households? What are the principal characteristics o fkey vulnerable groups? What are the existing risk management strategies used by rural households to prevent, mitigate and/or cope with vulnerability and income variability? What is the current strategy and level o f Government effort to reduce rural vulnerabilities, including scope, coverage, targeting and effectiveness o fprograms? And what policy reforms or public expenditures should be prioritized to best increase resilience to risk for rural populations Regarding Government interventions, the study will review public policies and programs considered important for social risk management in rural areas. These are: (i) social assistance and safety nets to vulnerable groups, including food aid, child labor and other programs for children at risk, programs for the elderly, the disabled, youth, women and other targeted groups; (ii)disaster management, including drought, flood and pest control; (iii)investments inmanaging critical a ~ c u l t u r arisks to prevent shocks as well as l support to farmers affected by natural and economic risks; (iv) efforts to manage critical health risks, including malaria, maternal health, malnutrition and HW-AIDS prevention; (iv) investments in reducing critical educational risks, including provision o f classrooms, teachers and school lunches; (v) labor market interventions, such as job creation via credit, training and public works; (vi) savings and credit, including expansion o f financial services to rural areas; (vii) social insurance, including health insurance and other forms o f risk pooling; and (viii) expansion o f service and infrastructure coverage to remote areas, including the targeting o f investments to the most vulnerable communities. Data andMethodology 14. Several existing data bases were exploitedto measurevulnerability, frequency of shocks, and overall profile of the conditions facing rural populations (Annex A). The maindrawbacks o fthe available data are: (a) the absence o f national household survey panel data to permit better understanding o f the changes invulnerability over time; and (b) none o f the data sets were developed specifically to assess risks and shocks. Nonetheless, they provide enough insights into the magnitude o f the main risks in rural Senegal to develop a reasonably sound profile. Expenditure information was collected directly from ministries and program managers. Although beyond the scope o f the study, some supplemental information was gathered fi-omNGOs. Outlineof the Report 15. Following this introductory section, Chapter 2 develops a profile o f rural vulnerability. Chapter 3 reviews informal and market-based risk management instruments used, by rural households. Chapter 4 reviews the public sector response, including an analysis o f key programs and expenditures to assist rural households. Chapter 5 concludes with a set ofrecommendations to decrease and better manage rural vulnerability and hence buildmore sustainable poverty reduction. 4 CHAPTER11: Risks andVulnerableGroupsinthe RuralAreas Core Questions: What are theprincipal characteristics of Senegal's ruralpopulation? FVho are the key vulnerable groups? What are the main risks and shocksfacing rural households? How do they interact across sectors and communities to heighten vulnerability? A. Characteristicsof the RuralPopulationand Determinantsof RuralPoverty Overview of RuralPopulation 16. Out of a nationalpopulation of 10 million, about 6 million Senegalese live in rural areas. Over two million people live in the greater Dakar urban area. Thies, Ziguinchor and Saint Louis also have a sizeable urban population (Table 2). The rural population makes up more than 80 percent o f the inhabitants in the rest o f the country. Diourbel and Kaolack, inthe GroundnutBasin have the greatest numbers of rural poor. Table 2 :SenegalPopulation and Number of RuralPoor,2002 Source: Senegal PopulationCensus 2002, Direction of Statistics, Ministry o f Finance Profile of RuralPoverty 17. Although economic growth in recent years has reduced poverty in Senegal, these gains have been less for rural populations. In 2002, 57 percent o f the population lived below the poverty line.2 Poverty i s more widespread inrural areas, where two-thirds are poor (Table 3). Poverty in rural areas declined by six percentage points since 1994, compared to a 14 percentage point decline for Dakar residents over the same period. The poverty gap is larger in rural areas, meaning the rural poor tend to be further from the poverty line, and poverty is more severe. The Gini coefficient3 for inequality indicates that income i s distributedrelatively more evenly inrural areas (0.30) than inurban areas (Dakar The poverty lines are defined according to the cost o f a basket o f basic needs. Three different poverty lines were usedto account for the differential cost o f living inDakar, other urban areas and rural areas. Gini coefficient of inequality. The coefficient varies between 0, which reflects complete equality, and 1, which indicates complete inequality. 5 -0.37andotherurban-0.35). Incomeinequalityforindividualsworsenedbetween1994- 2001 nationally and inboth urban andrural areas. Household level Individual level National Dakar lother urban( Rural national DakarI otherurban I rural 1994-95 Incidence o f poverty (%I 61.4 49.7 62.6 65.9 67.9 56.4 70.7 71.0 Depth o f poverty (%) 20.5 15.4 21.4 22.3 23.6 17.7 24.4 25.3 Poverty severity (%) 9.1 6.4 9.5 10.0 10.6 7.4 10.8 11.7 GiniCoefficient 38.6 45.8 39.7 31.7 32.6 36.7 34.0 29.4 2001-02 Incidence of poverty IGiniCoefficient 1 37.4 I 41.9 I 38.3 1 29.9I 34.2 1 37.3 1 35.2 I 30.1 Source; Senegal Poverty Assessment, World Bank 2004 based on E S A M I(1994-95) and E S A M I1 (2001/2002) 18. Poverty differs across rural regions in Senegal. Rural poverty is most widespread inZiguinchor, Kolda, Diourbel and Kaolack, at above 70 percent, and lowest in rural Louga (46%), Fatick (55%), Saint Louis (56%) and Dakar (58%). Injust one region, Ziguinchor, the poverty rate i s slightly higher inurban than inrural areas (Statistical Annex Table 1I.a). Despite the greater poverty in rural areas, rural households account for 16 percent o f the wealthiest quintile nationwide and these households enjoy consumption levels greater than 60 percent o f all urban households (Figures 1 and 2). On the other end of the income spectrum, rural households make up 91 percent o f the poorest quintile. These variations in rural poverty are important, since not all rural households are poor and some have found effective strategies for economic growth within this environment. Figure 1 :Rural-Urban Distribution among povertyquintiles(YO) richest quintile 4 quintile 3 qulntile 2 p o o r e s t 6 Figure 2 :Per-CapitaDailyExpendituresby quintiles, ruralhrban(CFA) 43 44 Q5 Characteristics of the Rural Poor 19. The rural economy remains largely agrarian, with about 70 percent of rural households engaged in agriculture. As a subset of this, livestock i s important both as the basis o f some regional economies and as a household risk management instrument. An estimated 300,000 households are engaged in the sub-sector (cattle and small ruminants). Remittances are an important source o f income both for rural and urban households. 20. There are key differences between the lowest and highest expenditure quintiles in rural areas. Per Table 4 (next page), poorer households are larger than the better-off ones. Better-off rural households are either less dependent on agriculture in that they are more likely to holdpublic sector employment or be engaged in services and commerce, or, if engaged in agriculture, have greater concentrations of land and/or animal^.^ Moreover, the rural non-poor have higher quality jobs, with 50 percent reporting permanent employment versus 27 percent o f the poorest households. The poor are more dependent on seasonal activities. Polygamous households are more likely to be poor, and households led by a single or widow head less likely. There are no large differences in educational outcomes o f heads o f households between poor and well-off households, except if they have more than primary level. The percentage o f poor households with disabled heads i s almost three times higher among poor than better-off households (see Box 1 on most vulnerable groups). In addition, auto-consumption i s significantly higher for the poorest quintiles.5 Wealthiest quintile households engaged inagriculture have 35 % o f holdings over 6 hectares and 37 percent o fthose between 4-6 hectares. The poorest quintile holds less than 5 percent o f the larger holdings. The PSIA survey o f producers inthe Groundnut Basin confirms this, with 64 percent o f revenues o f poor households from agriculture versus 40 percent for the non-poor and 28 percent for the wealthiest quintile. World Bank preliminary estimates (Senegal Poverty Assessment, forthcoming 2005) confirm that rural households with a head who reported owning land had higher levels o f consumption (by 68 percentage points for every extra hectare o f land) than other rural households. Gains to education were less visible in rural areas, where having a literate household head was associated with a gain in consumption o f 16-17 7 21. The rate of female headed Table 4 :Rural HouseholdCharacteristics by quintiles (i % except in :ated) households is quite high at over 50 Bottomtwc Top two percent. This i s above the Sub-Saharan quintiles quintiles average of 31 percent and reflects the 14.0 10.8 significant effect o f labor migrations, among others. However, rural households 358.6 1,326.3 headed by women are not more likely to be 14.0 26.4 0.6 1.6 poor and nationally female households on 3.4 3.1 average actually tend to be less poor than 3.5 4.3 male-headed households. Although the 77.9 64.3 link with poverty is not strong, there are ivestock None 14.4 18.3 10.8 8.2 specific issues that face female-headed Small animals 14.9 17.2 households in rural areas. For example, 59.6 56.3 with less male labor available, agricultural 56.4 59.1 output risks declining and there may be an 36.6 29.2 increased reliance on child labor. Women 4.3 6.5 face greater restrictions on effective access 1.1 1.6 51.7 51.9 to land and credit. Moreover, lower 26.3 30.6 literacy rates than male headed households 88.0 84.3 put female-headed households at a 9.0 8.1 disadvantage in terms of access to 2.2 1.o 5.5 information on diverse issues. 2.8 78.3 77.6 5.9 7.4 0.9 5.2 3.3 2.0 Own-account 95.3 92.0 ctivity: Agriculture 79.1 65.9 7.7 13.2 2.8 2.2 2.3 3.7 Public administratior 0.5 4.1 Transport- commun. 1.8 2.6 Job: permanentfull-til: 27.2 50.1 permanentpart-time 10.7 13.6 59.0 31.9 1.5 2.7 1.5 1.6 11.4 6.7 percent in urban areas but with no gain inrural areas. Inaddition, inrural areas, monogamous households had higher levels o f expected consumption (by 14 to 17 percentage points) than polygamous households. 8 Box 1:The Most Vulnerable Groups Beyond general poverty levels, there are several critically vulnerable populations. Often these groups come from rural areas or are at greater riskby virtue of living inrural areas: Orphans: Nine percentof Senegalese children aged 0-14 are orphans and 10percent of 5-14 yr old are fostered (i.e. not orphans but reside in households separate from parents). Children that have lost both parents are six percentagepoints less likely to attend school full-time and single orphans are 3 percentagepoints less likely. Femaleorphans face the highestriskofloss of schooling. Children in Dangerous or Exploitative Situations: According to various surveys and governmenl estimatesabout 400,000 children aged 6-18 are exposedto dangerous or exploitative situations: . 100,000 inonerous family work inagriculture, fishing andlivestock, N 30,000 boys involved inartisanandinformal sector activities, 9 .. 50,000 girls (6-18) employed as domestic workers, mostly coming from poor rural area. 100,000 children (talibds) living away from family mostly involved with coranic schools, 01 daara. A subset ofthese i s made to beginthe streets for longhoursandprovided little education. 100,000 inextreme situations- street children, sexually exploited, inconflict with the law. Disabled: There are an estimated 120,000 people with disabilities in Senegal. Disabled in rura areas are more likely to bepoor, andhaveless access to specializedequipment or training. Young Women Giving Birth: Rural women are more likely to marry andbear children at a young age. 9 percent ofruralbirthsare among girls 13 - 19versus 4 percentinurbanareas. sources: "Study on the Practice o f Trafficking in Person in Senegal" by B. Moens et al., for USAID .2004; "Orphanhood and Child Vulnerability: Senegal". L.Guarcello, S. Lyon, F. Rosati, July 2004; Senegal ESAM 11;MFSSD Letter of Sectoral Policy2004. B. Access to Critical Infrastructure and Services 22. In addition to consumption-based measures of poverty, access to basic services i s overall lower for rural residents. Underlying conditions o f vulnerability are increased when access to critical basic services i s limited. Health shocks are more frequent in the absence o f adequate drinking water and sanitation. Long distances to schools and health centers decrease the likelihood o f utilization and improvement o f human capital stocks. Distance to markets, roads and public transportation reduces economic integration and hampers access to services. For access to roads and markets: 40 percent of rural households live more than an hour-distance from food markets, and a third o f rural inhabitants have to walk more than an hour to reach main roads and public transportation. For water, 93 percent o f urban and 84 percent o f rural households live within a 15 minute walking distance from the nearest water source. However, interms o f quality, over half o f rural households get their water from open or unprotected sources, versus 6 percent o f urban households. In education, two-thirds of rural children have access to a primary school within reasonable walking distance (under one kilometer), 16 percent walk more than 3 kilometers. And for health services, more thanhalf of rural households live over 3 kilometers from a healthpost. 9 23. However, data suggests that within rural areas, poor households do not necessarily live farther away from infrastructure and services than do better-off households. The national household survey finds that it is the wealthiest rural households that tend to live the furthest distance from water sources, schools, and health infrastructure (Statistical Annex Table 1I.b). The PASEC household survey reinforces this in its finding that the wealthiest rural households also lived hrther from financial services (Statistical Annex Table 1I.c). This finding must be linked to the spatial structure o f landholdings and village settlement patterns, butbears further investigation. C. Estimationof Frequencyand Severity of Risksand Shocks 24. The vulnerability of ruralpopulationsis relatedto the risks they face andthe underlyingcapacity to address those risks. This section assesses the main risks and shocks experienced by rural populations and identifies those that, due to their frequency or magnitude and inter-relationship with other risk factors, pose the greatest challenge for poverty reduction. This includes naturalY6economic, health, education, and social risks. C.1. NaturalRisksand Shocks 25. Natural risks are intertwined with the agricultural nature of the rural economy and its Sahelian setting. Cyclical droughts reduce agricultural production and decimate herds. However, the impacts o f these shocks vary depending on the agricultural season, the agro-ecological zone (Box 2), the type of crop, and availability o f irrigation. 26. Senegal has about 3.8 million hectares of arable land. About 2.1 million hectares i s under cultivation, o f which 95 percent i s currently rain-fed agriculture. Half o f the land under cultivation i s for millet/sorghum, 38 percent for groundnuts, 4 percent in rice paddy, 4 percent in corn, and 1percent invegetable production. Irrigation potential exists for 350,000 hectares (less than 10 percent o f all arable land), with current levels o f irrigation reaching 105,000 hectares (less than 5 percent o f cultivated land). Therefore, Box 2: Senegal's Ecological Zones The Senegal River Valley includingthe regions of Saint Louis and Matam which has 75 percent of the country's irrigated hectares; .. .. The sylvo-pastoral zone which regroups the region ofLouga andthe department ofRanerou where livestock is the main activity; The Niayes zone along the Atlantic coast from Dakar to Saint Louis is oriented to vegetable production (10 % of irrigated hectares); .. The Groundnut Basin of Fatick, Kaolack Diourbel and Thies (and Kebemer) where groundnuts, millethorgum and cornpredominate; Western Senegal and the upper Casamance with cotton and irrigated agriculture; The lower and mid-Casamance with rice development (15 percent o firrigatedhectares) Risks of long-termenvironmental degradationandclimate change are outside the scope of this paper. 10 most agriculture i s subject to the vagaries o f rainfall in a drought-prone region. Rainfall and Drought 27. Rainfall is characterized by wide fluctuations between years. Over the last 25 years, rainfall in Senegal has fluctuated between a minimumof 394 mmand a maximum of 849 mm per year (Figure 3). Only one year out of five i s considered a good year for rain, at over 725 mm. Over the last 25 years, apart from five good years, the remainder had low rainfall and/or a bad repartition o f rain over time and space resulting in lowered levels of agricultural production. This rainfall irregularity is observed equally at regional levels and varies by climactic zone. In the Southern Casamance, the region with the highest average rainfall at 1100mm, annual precipitation can vary from minus 33 percent to plus 70 percent o f the mean (see Statistical Annex Table II.d, Figures 1I.a and 1I.b). Figure3 :AverageAnnualRainfall1978-2003 (inmillimeters) Source : National Meteorological Department 28. Over the last 25 years, Senegal has experience six years that could be characterized as major precipitation shocks. These were characterized by late onset o f rains, irregularity in spatial distribution, and an early end o f the rainy season. These shocks caused significant reductions in agricultural yields (Table 5). Ground nut yields fell to between 700-900 k g h a versus 1,100 kg/ha during good harvest years. Production shortfalls varied between 145,000 tons and 570,000 tons (between CFA17.4 billion and CFA 68.4 billion revenue losses at CFA 120kg average price) depending on the year. The pattern for millet/sorghum is much the same, production decreases o f 80,000 to 200,000 tons (CFA12 to 30 billion revenue losses at CFA15Okg average price). An 11 improvement o f average annual precipitation o f 1 mm increases by 1 k g h a groundnut yield and 0.4/kg for millet. However, the rainfall levels only explain 39 percent o f the variability in groundnut yields. Additional factors include degradation o f soil quality, timing of rains, use o f inputs and other shocks correlated with irregular rains such as parasites and locust invasions. 29. The most recent rainfall shock, in 2002, was a difficult year both for the amount and the timing of rainfall. This led to a drastic decrease in groundnut, cotton and nie'be' yields. Coping crops used traditionally as diversification suffered losses as well, including watermelons down 7.5 percent and sesame down 26 percent. An overall decrease o f 18 percent inbasic grain production aggravated the structural food deficit inthe country and caused the Government to carry out extensive food assistance and safety nets for farmers (reviewed inChapter IV). Infact, total groundnut production, at 265,000 tons, was the lowest recorded since independence. This shock also affected livestock, with a 5 percent reduction instocks. Overall, the economic losses caused by the out-season rains o f 2002 are estimated at 31billion CFA (about $62 million dollars). 30. Rainfall variations affect most rural households, and the poorest suffer the largest welfare losses. 85 percent o f rural households responded that they had experienced one or more bad harvests over the last 10 years (Statistical Annex Table ILe).' Table 7 shows production changes for producers in the Groundnut Basin 2000-2004 by household poverty level and consumption quintile. In2002, all poverty levels experienced production decreases. However, as the poorest households are more dependent on agricultural production alone for income, and in fact they produce a greater amount o f groundnuts, the absolute economic impact o fthe 2002 shock was much greater on the poorest households. Table 6 : Annual Groundnut Harvest (in kilos) per Household,by poverty level AgriculturalYea NonPo0 Poor Q1 I 42 I Q3 I 44 I Q5 2003-2004 1 I 948.6 2,932.5 I 3,731.91 2,112.5 I I 1,224.61 804.6 838.4 2002-2003 915.9 II 1,968.71 2,068.71 1,856.1 1,324.51 731.3 II 632.5 2001-2002 1,613.8 2,961.7 3,607.1 2,264.8 1,651.1 1,325.3 2,124.1 2000-2001 II 1,509.0 3,024.8 4,040.9 1,876.8 1,797.3 1,526.5 1,233.0 Floods 31. Despite the overall fragility of precipitation in Senegal, certain localities are adversely affected by floods. Over the last ten years, several areas suffered from floods caused both by localized intense rains and river overflows, concentrated in Saint Louis, Matam, Louga, Tambacounda and Kolda. These floods harm hundreds o f thousands o f While residents o f Dakar were not affected by bad rain years, 37 percent o f households in other urban said they were affected by at least one bad harvest, showing the urban linkages o f the agricultural economy, 12 people, causing deaths, displacement, destruction o f houses and other infrastructure, and loss o f crops and animals. Characteristics by region include: InSaint Louis and Matam floods are recurrent along the Senegal River. Since the early 1990s, there have been important floods in 1994, 1999 and 2003, with particularly devastating effects in 1994 and 1999. The floods in 1999 lasted until mid-October in many areas, killing people and destroying roads, houses, crops, and other assets, afrecting an estimated 100,000 residents. 70 percent o f the irrigated plots (pBrim8tres irrigu&s villageois Ply) in Podor were adversely affected. In addition to crop loss and destruction o f houses, the flooding promotes the infestation o f aquatic plants (Dia 2004 and Fall 2000). Flooding in Kolda varies by department. In 2003, heavy rains combined with lack of drainage infrastructure affected populations who have increasingly settled in potential flood zones along the Gambia, Casamance and the Kayanga rivers. In Tambacounda, during July-October 2003, strong rains exacerbated overflow of natural storage areas and the resulting flooding caused extensive damage to banana plantations. Seventy percent o f the 1,000 hectares were destroyed. The floods affected an estimated 22,000 hectares o f cultures under production by 3,200 producers, for an estimated economic cost of 3.6 billion CFA. In the Saloum Delta region, two forms of local flooding are common, one from rains and the other from penetrationfrom the sea, resulting in erosion o f sandy river banks, isolation o fpopulations, and the salinisation o f arable lands. Locusts, Pests and Parasites 32. Senegal periodically experiences locust swarms, most recently in 1988 and 2004. Locust invasions, which have beenpresent intermittently in the Sahel for thousands of years, present a highrisk to agricultural production and revenue. Duringthe most recent episode, according to an FA0 evaluation (FA0 2004), out o f 11 regions in the country, seven were adversely affected by the locusts - Saint Louis, Diourbel, Thies, Matam, Fatick, Louga and Dakar. The loss of production inbasic grains averaged 22 percent, including 34 percent for millet (compared to average o f previous five year) and 30 percent for sorghum. Overall loss of grain production was over 45,000 tons. The region o f Saint-Louis in the north accounted for half o f total losses. Halfo f grazing pastures inthe north (Louga, Saint- Louis, Matam) and 80 percent inDiourbel were devastated. Locust damage in2004 didnot significantly affect the main agricultural land inthe southwest. 33. An estimated 124,000 households were affected by the 2004 invasion, or 20 percent of all rural households. Unlike general drought, locust damage i s geographically sporadic, even within regions affected. Where the swarms land and feed, crop loss can be total. Households affected were inthe regions o f Matam, Thies, Diourbel, Saint Louis and Louga (Table 7). Production losses are estimated less than 50,000 tons (about CFA 7Bn at CFAl5O/kg average farm gate price). The damage seems to have been mitigated somewhat thanks to the early engagement and resources committed to the locust fight campaign. 13 Table 7 :EstimatedLosses in GrainProductionDueto Locusts,2004 34. Other endemic infestations affecting agriculture and livestock are less spectacular but neverthelesscause significant production losses. For crops, the main risks include: (a) grain-eating birds (quelea-quelea andpasser luteus) that damage harvests inthe Senegal River Valley; (b) grasshoppers that attach to young sprouts at the beginning of the rainy season (oedaleus senegalensis notably), especially in the regions o f Kaolack, Tambacounda, Fatick, Diourbel and (OEdaleus nigeriensis and Zonocerus variegates) in Kolda and Ziguinchor; (c) other insect swarms (amsacta moloneyi) in the Kaolack, Louga and Diourbel regions that attack niebe, groundnuts and maize; (d) floricoles insects (cantharides) that eat the young buds on millet in the Groundnut Basin area; (e) Aphides (aphis craccivora) which transmit leaf curl virus to groundnut and nikbk; and (f) white fly (bemisia tabaci) that affects cotton production. 35. In general, preventionand treatment of parasiticinfestationsremains limited. Not counting 2004 which was an exceptional year due to the mobilization around emergency locust operations, an average o f 428,000 hectares are inspected each year, or about 20 percent o f land under cultivation. Over the last six years, 63 percent o f fields inspected were infected and 53 percent o f those fields found to be infected were treated (Statistical Annex Table 1I.f). Threatsto Livestock 36. Animal production is threatened by drought and other climatic and pest/parasitic effects on pastures. Loss of grazing pastures translates into the sale of animals at lower prices and changes intranshumance patterns that can cause weight loss to animals o f up to 30 percent. It i s difficult to estimate the overall losses as herders may move animals over national boundaries, sell o f f at lower than expected prices, etc.8 Per Table 8, with the 2002 drought there was a reduction in the number of cattle, sheep and goats, but family poultry increased, consistent with their role in household coping strategies. However, it i s worth noting that the effect o f drought on livestock production i s much smaller than on agricultural production, hence the important role o f livestock as a hedgeagainst consumptionloss inthe face o f drought. ~~ * Moving further afieldinsearchof pastures also causes herder families to pulltheir children out of school. 14 Table 8 :Evolution of Animal Stocks 1999 2003 (in '000 heads) - 2002 2 996 937 14 540 380 3 899 972 291 450 496 095 399 547 4 000 5 174 255 20 207 048 2003 3 017513 14613 508 3 968737 303 368 501225 399547 4008 5 261866 20813 260 37. Major diseases include plague, foot and mouth disease, Newcastle disease, horse plague, and botulism (Statistical Annex Table 1I.g). The Ministry in charge of Livestock tracks animal diseases and promotes vaccinations. In 2002 (most recent data), MAE reported about 1,300 animal deaths, but many losses go unreported. Inthe PASEC survey, three-quarters o f rural households cited livestock loss over the last 10 years, with epidemics as the most frequent cause, followed by theft and then drought. (Table 9). Table 9 :Frequency and cause of livestock depletion, 2002 inYO ~~ ~~ Source: PASEC data set Effect of Agricultural Shocks on the NationalEconomy 38. Since the devaluation of the CFA in 1994, the Senegalese economy has regained dynamism, registering an annual growth rate of over 5 percent, although weather- based shocks create a drag on potential growth. GDP growth was interrupted in 2002, when the economy experienced a slowing down o f growth to 1.1 percent. This downturn was explained in part by the drop in agricultural production in large part due to poor precipitation, off-season rainfall and flooding along the Gambia River. Figure 4 traces the evolution o f economic growth inthe primary sector from 1998 to 2004. The restoration of economic growth to 6.3 percent in 2003 and 6.1 percent in 2004 was supported by growth inthe primary sector o f20 percent in2003 and9 percent in2004.9 'The tertiary sector of the economy accounts for over halfo f GDP. The primary sector accounts for under 20 percent o f GDP and this percentage has fallen slightly over time. However, there are strong linkages between agriculture and activities inthe secondary (processing) and tertiary (commercialization) sectors. A 1%increase inagricultural growth results in0.81% and 0.92% increase inlivestock and edible oils sub- 15 Figure 4 :Primary Sector Growth Rate, 1998-2004 40.0% 30.0% 20.0% 10.0% 0.0% -10.0% -20.0% + \ w/-7Qc6!? -30.0% \ / ~ -32.2% -40.0% -~ - Source: Ministry o f Finance, Bank estimates C.2. EconomicRisks andShocks 39. Economicrisks in the ruralareas are relatedto risks of an open economy, risks tied to productionand commercializationsystems, and income and purchasingpower risks, While unemployment is a major economic risk inurban areas, this is not as relevant for rural Senegal where most o f the labor force i s self-employed or engaged in informal or part-time activities to diversify income. Underemployment and low revenues from productive activities are more important than outright job loss. OpenEconomyRisks 40. Senegal is an open economy and its rural producersface risks from exchange rates and world price fluctuations. Agriculture accounts for 20 percent of total exports, comprised principally o f groundnut products (crude oil and cakes, accounting for half of total agricultural exports), cotton (raw and woven products) and to a lesser but growing extent horticulture. Groundnut production i s stagnant and exports have decreased over time partly because of poor political management of sub-sector reforms since the mid-90s, as well as depletion in the quality o f seeds and soil. The tariff structure that provided incentives for local industry to import and process cheaper edible oil categories for domestic markets, combined with downward trends in world prices over the last decade also played an important role in the decline. Untilrecently, government pricing policy did not transmit world market signals and producers were offered stable revenue, although guaranteed prices were well below the market price. Reforms undertaken since mid 90s promotedmarket-led prices negotiatedbetweenproducers andprocessing firms. sectors respectively inthe same year, 0.5 1% and 0.69% positive effects the following years in food processing and trade sub-sectors respectively. 16 41. In terms of exchange rate risks, the main shock was the 1994 devaluation of the CFA. A recent World Bank study on the effects o f the 1994 devaluation found that virtually all segments of the population, including the rural poor, benefited from improvements in standard o f living following devaluation, but the growth was not strictly 'pro-poor' as gains in the upper half o f the income distribution were greater. The slower economic growth o f the rural economy following the devaluation was due to (a) bad weather; (b) the devaluation-led public investment boom that primarily benefited the manufacturing and service sectors; and (c) a drop in the real value o f crops after the devaluation (Azam, Dia, Tsimpo and Wodon 2005)." RisksTiedto ProductionandCommercializationSystems 42. The system for financing and commercializing agricultural production, particularlyof groundnuts,has undergonestructuralreformsin recent years. Supply o f inputs and commercialization o f agricultural products have been increasingly liberalized over the last 20 years. For the groundnut sector, the ongoing reforms accelerated in 1997 resulted in the suppression o f SONAGAINES in 2001 and the establishment o f a direct delivery system, without any intermediaries or brokers between the producer and the milling company. Even though this new system is generally more efficient than direct Government involvement, the transition has caused dislocations as markets adjust. 43. Lack of access to seeds, and to lesser extent fertilizer, is a criticalissue and the by far the most significant reason for reductionsin land area under cultivation. In the PSIA survey o f the Groundnut Basinproducers, 72 percent o fnon-poor households and 68 percent o f poor households that reported reducing the area o f land under cultivation cited lack o f seeds as the principal reason (Statistical Annex Table 1I.h). For the poorest quintile, lack o f fertilizer was also a significant reason (12 percent o f responses). Risks associated with climate were a lesser reason for withdrawing acreage from production. The lack o f access to critical inputs stems from difficulties with rural credit markets and changes inthe structure o fproduction following the privatization o frural industries. PurchasingPower and PriceRisks 44. Over the last ten years (1994-2003) agricultural producers of cotton, groundnuts and millet have seen their purchasing power eroded. The general consumer price index has risen faster than the evolution o f prices for main agricultural products (Table 10). Notably, groundnuts and millet have experienced absolute price decreases inrecent years. Purchasing power for rice producers has maintained, thoughthey are confronted by strong competition from imported rice o f lower quality. Corn producers had strong gains, both inabsolute price and purchasingpower. lo Aside from movements in the exchange rate, macro-economic risks in the form o f increasing world oil prices, will ripple through to affect agriculture both in terms o f petro-chemical input costs, as well as processing and transportation costs. World Bank staff estimates indicate that a US$5 per barrel increase in oil prices above the current WE0 projections would cut average annual GDP growth by about 0.6 percentage points during 2005-2010 overall. 17 Table 10: Evolutionof PurchasingPower (PP) and PriceIndexes, 1990-2002 Source: Gestion de Risques Agricoles, Niang, workingpaper 2005 C.3. Health Risks and Shocks 45. Health shocks have effects beyond quality of life, impacting on income, productivity, schooling and ultimately, economic growth. Loss of days on the job translates into less wage earnings and/or domestic production, as well as loss o f service output (e.g. teacher days and other essential services). Health crises increase household expenditures and deplete savings, including transportation, medical fees, drugs and in the event o f death expenses associated with funerals. The cost effects o f health crises are particularly acute inthe absence o f any risk pooling or insurance mechanisms. The loss of savings inturn increases overall household vulnerability to other adverse events. Effects of health shocks on education attainment include reduced learning and may lead to drop out and repetition. Death of a parent exposes orphans to a significant risk o f not enrolling in school. To assess these heath risks and shocks, this section first looks at the general morbidity and mortality profile, then reviews the critical areas o f malnutrition, maternal mortality, malaria and other endemic diseases and HIV-AIDS. Overall Morbidity and Mortality Profile 46. Senegal has made important strides in several health outcomes, but lags in several critical areas for its income class. Life expectancy, at 54 years, is above the SSA average o f 49 (1999) reflecting Senegal's higher per capita income, increasing urbanization, low HIV prevalence and improved coverage o f primary education over regional comparators. However, maternal mortality remains high at 510 per 100,000 live births (DHS I1 1993), and total fertility at 5.7 children per woman is above regional averages. For its income level, Senegal has among the highest under five mortality rates. Neonatal mortality accounts for half o f the infant mortality rate and i s associated with pre- maturity and low birth weight, neonatal tetanus, pneumonia and birth trauma. Infant mortality rates are 40 percent higher in rural areas, and child mortality (age 1-5) i s 162 percent higher. 18 47. Malaria is the leading cause of morbidity and mortality in Senegal. From Table 11 the ESAM I1household survey, malaridfever Leading Causes of Child and Maternal was reported to be the most common disease in rural areas (half o f all self-reported causes o f illness), followed by diarrhea and respiratory 1. Malariaifever 26 % problems. Children and the elderly had 2. Malnutrition 22% malaria (or at least reported fever) more 3. Diarrhea 21% frequently than other groups (Statistical Annex 4. Tetanus 14% Figure1I.c and Table 1I.i). Malaria is estimated 5. Acute respiratory 7% to cause 26 percent of child deaths (Table 11). MaternalMortality 1. Hemorrhages 41 % 48. Public health service administrative data confirm this general morbidity profile. 3. Infections 20 Yo 4. Eclampsia/high blood Based on an average of 1996-2001, The pressure 7 OL I /u Government reports that about 900,000 cases 5. Other causes 9 % o f malaria presented for treatment annually, with 220,000 cases o f skin ailments, 180,000 Country Status Report on Health in Senegal, cases o f acute respiratory illness, 170,000 cases World Bank 2005 forthcoming. of diarrhea and dysentery, and 60,000 cases o f parasitosis as the leading causes o f morbidity nationwide. Malaria i s also the leading cause of death reported, averaging 1,500 reported deaths per year. Diarrhea claims 280 deaths reported and respiratory illness 250." Although HIV-AIDS i s not cited as a major cause o f morbidity, it accounts for an increasing share o f patient beds at the main infectious disease hospital in Dakar, from 10 percent in 1991to 60 percent in 1999. Official morbidity figures do not include information on underlying malnutrition as a cause and complicating factor o f illness, but estimates o f child deaths rank it as the second leading cause. Statistics are not disaggregated between urban and rural areas. However, when sick, 18 percent o f rural households missed more than 7 days o f worWschoo1 due to illness compared with 14 percent o f urban households. Malnutrition 49. One-quarter of rural households report that they often or always face difficulties with the provision of food. This is fairly constant across poverty quintiles, with the slight difference that 46 percent o f the highest quintile households said they never or rarely faced difficulties providing food, versus 39 percent o f the lowest quintile. And, 20 percent o f urban households also report persistent difficulties with provision o f food. However, it i s worth noting that this self-perception may differ sharply from what i s actually consumed due to differences inperception and expectation as to what constitutes a meal. (Statistical Annex Table 1I.j). 50. Differentsources reportdifferent figures in terms of child malnutrition,but all agree that rates are much higher in rural areas. According to the ESAM 11, in rural "Due to limitedutilization o f health services, the administrative figures o n morbidity and mortality greatly understate the actual number o f people experiencing health shocks. Although there are only 1,500 reported deaths from malaria annually in Senegal, the MOH estimates actual annual deaths at closer to 8,000. 19 areas more than 43 percent o f children under five show signs o f stunting (an indicator of chronic poor health and diet), 24 percent are underweight and 10 percent show signs of wasting, compared to 21 percent, 12 percent and 8 percent respectively in urban areas Malnutrition indicators are highest for Diourbel, Kaolack, Fatick and Tambacounda. According to the MICS 2000 survey, about 21 percent o f rural children under five had moderate or severe malnutrition, as measured by weight for age and height for age compared with about 14 percent o f urban children (Statistical Annex, Tables 1I.kand 11.1). Maternal Mortality Table 12 :Selected Maternal Health Indicators Indicator Urban (%) Rural (%) % o f all women 15-49 that use modemcontraception 12 3 % o f married women 15-49 that use modem contraception 18 3 Nurse-Midwife 83 62 N o one 11 27 Doctor 5 4 Matrone 1 5 Friend / Relative 4 31 Nurse-Midwife 80 30 N o one 8 14 Traditional birthattendant 1 11 Matrone 5 11 Doctor 2 2 Total qualified personnel 87 43 51. Health shocks to women of childbearing age are a critical risk in rural areas. Maternal mortality inSenegal i s alarmingly high, and insome rural areas exceeds 1,200 per 100,000 live births. Maternal mortality i s linked with other adverse outcomes. Studies in other Sahelian settings find that maternal mortality raises the likelihood o f child mortality.'2 It also raises the likelihood o f children not enrolling in school in Senegal by 3 percentage points (Guarcello et. a1 2004). The higher maternal mortality rate for rural , women i s caused by a number o f factors, including earlier age o f initial childbearing, higher fertility rate, complicating factors from malaria and poor nutrition, and lack o f care before and during birth. The higher fertility rate reflects lower average age at marriage, less access to contraceptives, lower female education level, as well as rural preferences for larger households. Only 3 percent o f rural women report using contraception versus between 12-18 percent o f urban women; and 71 percent o f rural women report receiving pre-natal care by a qualified health personnel versus 89 percent o f urban women. This I2Froma survival analysis o f 39 villages inBurkina Faso, an estimation o f hazard rate ratios showed death to mother and being a twin as strongest factors for mortality child mortality. Ifmother dies in first year of life, 7.4 fold increase in infant mortality risk, if mother dies within second year, a 3.9 fold increase in mortality risk (Becher, et. al. 2002). 20 coverage falls when it comes to the actual birth, where risks are highest. Forty-five percent of rural women report giving birth either alone or with a family member or friend. Only 12 percent o f urban births took place in such risky settings. The presence o f nurse-midwives to attend urban birthswas 2.5 times more likely inurban than inrural settings (Table 12). Malaria and Other Endemic Diseases 52. Malaria is the most widespread health issue facing Senegal. Nationally, malaria accounts for 31 percent o f morbidity and 32 percent o f mortality. Incidence varies over the year, with a pronounced spike during the rainy season (particularly September and October). For example, in one area under long-term demographical surveillance (Niakhar), 89 percent o f all malaria deaths were concentrated betweenAugust and December, with the peak in October (Ndiaye et a1 2001). In the north, malaria is mostly seasonal with mortality and morbidity concentrated within a period of a few m o n t h ~ . ' ~ the southern In areas o f the country, malaria i s endemic, with less pronounced peaks. At highest risk o f mortality are pregnant women and children from six months to five years old. Malaria is associated with severe anemia inpregnant women, low birthweight and consequent infant mortality, Malaria also hits hardest at those exposed to other health risks, including malnutrition and HIV AIDS, and maternalmortality is complicated by malaria risk.I4 With resistance to chloroquine, the burden o f malaria has increased, as evidenced in a two- to three-fold increase in hospital admissions and deaths, and a six-fold increase in pediatric malaria mortality from data collected since chloroquine resistance emerged in Senegal in the late 1980s (Trape et a12003). 53. The chances of contracting malaria are many times higher in rural areas than in the cities. The mean annual entomologic inoculation rate (ER) is 7.1 in the urban centers, 45.8 in periurban areas and 167.7 in rural areas (Robert et. al. 2003).15 This lower likelihood of being bitten by an infected mosquito in urban areas is due a less hospitable environment (e.g. due to pollution, lower vector capacity), mosquito avoidance behavior by households (e.g. screens, insecticides, bed nets), and higher human population density. MICS 2000 reports that 13 percent o f children 0-59 months slept under some form o f mosquito net versus 23 percent for urban children. 54. Malaria affects income, production and learning. A cross-country analysis o f the overall effect o f malaria on economic growth rates estimates that malaria reduces GDP growth in Senegal by 0.35 percent year. (McCarthy et. al. 2000). Economic impacts include loss of labor and agricultural productivity and foregone growth in industries like tourism. In addition, malaria causes loss o f school days and may negatively impact long l3Malaria has become problem in the Senegal River Valley area where the construction o f two dams has resulted in ecological changes. Severe flooding during the rainy seasons in late 1990s resulted in a change from malariabeing seasonal and relatively rare to malaria being transmitted throughout the year. l4Ina study o fNiakhar inthe Fatick Region, overall mortality inwomen 15-49 years o f age didnot differ by season, but maternal and direct obstetric deaths were significantly more frequent during the rainyimalaria season than during the rest o f the year (with a peak in September), even after adjusting for place o f delivery (Etard et. al. 2003). The ambulatory center for treating AIDS patients in Dakar reports spikes inhospitalizations o f AIDS patients duringOctober (CTA 2003). Is The EIR is expressed as the number of infected bites per person per year. 21 term learning capacity.I6 Average infection in school children i s 30-50 percent, but i s most common inthe South at 62 percent o f school children. 55. Rural populations are also exposed to other endemic diseases, including concentrationsof schistosomiasis in certain regions.Construction o f dams and irrigation along the Senegal River Valley has increased schistosomiasis. Prevalence appears to have increased among school children in recent years, from 19.5 percent to 22 percent in endemic zones in 1996-2002. This affects the health of school-aged children, causing an increase o f liver-related complications that will be costly to manage in the coming years, Senegal has successfully addressed onchocerciasis (river blindness)inrecent years through annual mass treatment of infected villages, reaching 606 villages in2003, with a decline in community prevalence from 9.5 percent to 0.2 percent in 1996-2003. HIV-AIDS 56. Senegal has one of the lowest HIV prevalence rates in Sub-Saharan Africa, at an estimated 1.5 percent of the adult population." This translates into an estimated 75,000 adults and 5,000 children under 15 currently infected (CNLS 2004). Factors that have contributed to keeping the prevalence rate low consist o f favorable geographical location (lower rates overall in the West African region), initial prominence o f the less virulent HIV-2 strain, social mores that proscribe certain risks, an early and aggressive policy by Government and civil society to educate the population about the risks o f HIV- AIDS, and aprogram to screen for sexually-transmitteddiseases amonghigh-riskgroups. 57. Prevalencerates vary by region,though data does not permit firm conclusions about rural areas. As presentedinFigure 5, the highestrates are observed inthe frontier regions o f western and southern Senegal, including Kolda (2.9%), Tambacounda (2.6%), and Ziguinchor (2.3%).'* Kolda and Tambacounda are two o f the most rural regions o f Senegal with over 80% o f the population living in rural areas. By contrast, Dakar has an HIVprevalence rate o f 1.7%. %ne study of Dakar estimated that during the period o f maximum transmission, clinical malaria prevalence was 1.36% and malaria was responsible for 36% o f school absences due to medical reasons. At the end o f the period o f minimum transmission, clinical malaria prevalence was 0.15% and malaria was responsible for 3% o f school absences due to illness (Trape 2003). In the Senegal River Valley, absenteeism due to malaria attacks was 70%. "Basedonprevalence rates amongpregnant womenat sentinelsites. DHSis currentlybeingcompleted and includes a sampling o f HIV prevalence rates inthe population at large. ''The Government's proposal to the Global Fund for Malaria HIV-AIDS and Tuberculosis posits that the Southern region of the country i s at greater risk inpart due to the conflicts inthat area which weaken social control and encourage the development o f sexually risky behavior. Cross-border relations with higher prevalence countries like Guinea and Gambia may play a role as well. 22 Figure 5: HIV Prevalence among Pregnant Women, by Region, 2002 and 2003 (average and confidence interval) Prevalence rate 5.00% 4.50% t 4.00% 3.50% t A 3.00% 2.50% 2.00% 1.50% 1.OO% 0.50% 0.00% t a v e r a g f max region 58. Rural populations are vulnerable to HIV-AIDS because they have less pertinent knowledge of the disease and less access to health services. Most rural residents say they have heard o f HIV-AIDS, which testifies to the awareness campaigns o f Government and civil society. But, rural women know less about specific transmission risks and prevention measures than urban women. Although three-quarters o f rural women hadheard o f AIDS, only 29 percent could identify ways to prevent its transmission, versus 47 percent o f urban women (Statistical Annex Table 1I.m). Sixty-two percent o f girls could not identify a single transmission risk (40 percent for urban girls), and fewer rural women know where to get an HIV test. There are large regional variations, with the largest lack of awareness inKolda and Tambacounda, the regions with the highestprevalencerate. 59. Labor migrations, which are critical for rural survival, also appear to increase HIV risks. About half of all AIDS cases being treated at the country's main out-patient clinic report extended periods abroad, most frequently (64 percent) in other African countries. Relatively fewer AIDS patients report travel to Europe (6 percent) or the U S (1 percent) (CTA 2003). A study of villages around Matam noted that 27 percent o f returned male international labor migrants were infected compared to less than one percent o f non- migrants (Kane et. al. 1993). A survey in rural Ziguinchor found that seropositivity was associated with history of blood transfisions, injections, sexually-transmitted diseases and seasonal migration (Pison et. al. 1993). In rural areas, HIV appears to be mainly transmitted first to adult men through sexual contacts with infected women met during 23 seasonal migrations and second to their wives or partners once back home.Ig Given the high rate o f female-headed households, in large part due to labor migrations, there are a substantial portiono f rural women for whom this risk i s present. 60. Moreover, economic and power relationships between genders put women at risk. In one study of southern Senegal, seasonal migrants and divorced or widowed women were more likely to declare casual sex. Overall, 28 percent o f sexually active men and 27 percent of the sexually active women declared at least one casual sex partner inthe last 12 months. Casual sex was motivated by material needs for 66 percent o f the women who reported it (Lagarde et a1 1996). Health Shocks and Access to Health Services Place of consultation 4 1 42 43 4 4 45 Total HealthPost 12 8 11 9 7 10 Public Hospitalldispensary 17 20 21 24 29 20 Public health center 43 47 43 42 47 44 Private doctoridentist 0 0 1 1 0 1 Traditional healeriMarabout 18 15 15 14 9 16 Midwifelnurse 0 0 0 1 2 0 Religious hospitalidispensary 1 2 2 3 2 2 Case de sante - health hut 7 6 5 4 3 5 Reasons for not consulting health services Not necessary 33 31 40 44 43 36 Too costly 45 45 34 35 24 40 61. In the face of a health shock, rural populationstend to use health services less than urban dwellers. Two-thirds o f the rural sick used health services versus three- quarters for urban residents, declining by poverty level. (Statistical Annex Table 1I.n). More importantly, rural residents access lower quality health services. Rural inhabitants were far more likely to go to traditional healers than urbanpopulations (16 percent versus 6 percent), were twice as likely to go to a public health centers, and used public and private hospitals and clinics less (36 percent versus 67 percent). The poorest households were the most likely to go to traditional healers. Moreover, utilization of the health huts (cases de sante? was very limited as a curative service (Table 13). 62. Financial access and distance are critical factors in use of health services in rural areas For those rural residents who said they had been sick but did not seek ~ ''Women migrants do not appear to present the same level o f risk. A study o f female Serere migrants found that they maintain traditional Serere marriage and sexual practices with moderate risk for HIV infection and do not create a linkbetween l o w risk and highrisk groups (Velyvis 2003). 24 treatment, 40 percent cited high cost as the reason (Table 13). This was even more pronounced for poorer rural households. Costs may include not only consultations and medicines, but also the costs in terms o f lost time for work and costs o f transport to the facility. These costs discourage populations from seeking prompt medical attention, which is time-critical in the case o f onsets o f malaria, birth complications and other health shocks.20 Overall, rural household spend about 2.1 percent o f their annual household expenditures on health versus 0.6 for urban households (Statistical Annex Table 11.0). Moreover, the relationship between cost, quality, distance and use i s inseparable. Rural and urban households alike make assessments about the quality o f care they will receive for the price, including transportation and other costs o f seeking treatment. C.4. Risks and Shocks to EducationalAccumulation 63, Risks to educational accumulation include never enrolling, dropping out during the course of the year or between years, or not learning the sufficient skill base. These shocks may translate into lower lifetime earnings, less potential for diversification of income sources, and increased exposure to health risks especially for uneducated women. International studies have shown that children are more at risk of never starting school if their parents had no schooling, thus transmitting vulnerability across generations. Levels o f female education are directly linked to health outcomes, including knowledge o f HIV/AIDS, fertility and utilization o f health services. Literacy levels remain low inrural areas. Literacy i s higher among men (52%) than women (30%), and for both genders each new generation has a higher literacy level than the prior one, with no difference across poverty levels. (Statistical Annex Tables 1I.pand1I.q). GeneralEnrollmentPatterns 64. The estimatedgross primary enrollment rate is 95 percent in urban areas and 70 percent in rural areas. This translates into over 300,000 rural children o f primary school age that are not in school, in contrast to only 36,000 children out o f primary school in urban areas (Table 14). For rural areas, a full one-third o f these are in the Diourbel region, with other significant numbers inKaolack and Louga. It i s not possible to say how many o f these do not go because there i s no school in the area, but extensive investments from 2000-2004 increased the number o f primary schools from 4,751 to 6,060 and the numbero fclassrooms from 21,530 to 30,483 reducingthis as a general factor. "In Touba, 64% o fhouseholds used informal health care for malaria due to cost reasons (Faye et. al. 1996) 25 Table 14 :DifferenceBetween School-AgedPopulation andEnrollments, by region 65. Within the rural areas, there is Table 15 :SchoolAttendanceby Rural little to no difference in attendance rates between poverty quintiles (Table 15). While economic constraints may keep poorer families from sending their children Quintile 2 I 57.7 I 38.0 I 2.3 to school, it is not clear what the constraints Quintile 3 I 60.2 I 34.6 I 3.9 are among the richest rural quintile. A Quintile 4 number of factors are possible. First, if 5 (wealthiest children's attendance i s strongly influenced Total 58.7 by parental education levels, then the fact Source: ESAM-I1Note: These figures differ from that adult illiteracy i s relatively flat across MOE enrollment figures as they are limitedto rural poverty quintiles would help explain school-aee children (net enrollment) someof this. &other ossible factor i s the role o f religious education and coranic schools among the rural elite!' Other possibilities include the perception by better-off rural families that education i s not that valuable, which may be particularly prevalent among certain trader groups where children follow inthe family trading practice. And finally, the opportunity cost o f child labor i s higher in those households with more access to land or grazing areas, who typically are wealthier. 66. The most common reason rural households themselves cite for their children leaving school is lack of interest or that school i s not useful, followed by health problems, too costly, and need to work at home. One-third o f those children who had dropped out said school was not useful or they had no interest (Statistical Annex Table 1I.r). Twenty percent cited healthproblems, 13 percent highcosts and 12percent that they had to work at home. Distance in fact was perceived as a lesser issue, cited only by 5 2'However, removing Diourbel from the data does not change the general pattern. 26 percent o f children who had dropped out.22 In terms o f the costs o f schooling, although primary education in theory i s free, various fees are levied usually through parents committees, as well as the cost to the families o f books, supplies and other inputs. On average, rural families spend about 1 percent o f their annual consumption on education expenses. Determinantsof EducationalAttainmentinRuralAreas 67. Analyzing household and community-level determinants of rural schooling supports several of these self-perceptions. A multivariate model was developedfrom the PASEC data set to explore determinants o f primary enrollment, primary completion, primary dropout and years o f schooling attained for rural children(Table 16). Table 16 :Determinantsof RuralPrimarySchoolingOutcomes Characteristics e communi Adj. R-squared 0.4246 0.577 0.3308 0.2307 Note: Bold denotes statistically significant at 5% confidence interval; years o f schooling i s for children aged 6-25. Source: PASEC Dataset, Bank estimates 68. The factorsthat most impederural primaryeducationalattainment are: (a) The older the child, the less likely they are to enroll inprimary school and the more likely to drop out. Further, years o f school and likelihood o f graduating from primary school are positively correlated with age, as one would expect. (b) Interms of gender, boys are more likely to graduate, have more years of schooling and to avoid dropping out. The gender difference disappears when it comes to enrolling, demonstrating the increases ingirls enteringprimary school. 22The E S A M survey question only asked for those children who dropped out. Distance may be more o f a factor for those who never enrolled. 27 Socio-economic status i s negatively correlated with graduating from primary and the number of years o f schooling and positively associated with drop-out. While the better-off rural households are as likely to initially enroll their children in school, they are less likely to continue. (Possible factors were discussed inpara, 65). The educational attainment regression shows, for both rural and urban samples, that when the highest level reached bv the head o f the household i s limited to primary, the child is likely to complete fewer years o f schooling. Distance i s only a factor in drop-outs, but not initial enrollment (but this only applies for behavior within communities with schools given the sample frame). Positive community interventions, typically local development projects and small- scale infrastructure, were associated with more years o f schooling. This may be picking up social capital aspects associated with more dynamic communities, proximity or more physical investmentin completing classrooms. Health shocks to the child are negatively correlated with the number o f years o f school completed and higher drop-out rates. Health shocks to the parents do not appear to significantly affect rural children's schooling.23 Negative crop shocks that adversely affected households over the past 10 years have been detrimental to the number o f years o f schooling that rural children are able to complete and their likelihood o f graduating from primary education. However, good harvest years also appear to decrease the likelihood o f enrollment, possibly due to demands for family labor inyears of good rains. There is some additional evidence to support the effect of child labor on the lower rural primary enrollment rates among the better-off households. In the survey of the Groundnut Basin, 9 percent o f children (12 percent o f boys) were involved in groundnut production (Table 17). More importantly, this share rose with householdwealth, to 16 percent of boys from the wealthiest quintile. The opportunity cost o f schooling i s viewed as greater among the better-off households. NonPoor Poor Q1 42 43 Q4 Q5 YOchildren who work 11.6% 7.6% 7.5% 7.2% 12.9% 10.1% 12.0% Yoboys who work 14.8% 9.1% 8.3% 10.1% 17.0% 11.9% 16.3% % girls who work 6.8% 5.7% 6.5% 3.7% 7.5% 7.3% 4.0% C.5. SocialRisks 70. Regional conflict in the Casamance, theft of livestock and property disputes are the principalsources of social risks in rural Senegal. A separatist movement inthe Casamance has caused loss of welfare in the southern region o f Senegal since the early 1980s, including armed conflict, land mines and increasing insecurity resulting in livestock theft and acts o f banditry. A peace agreement was signed in March 2001 and the situation has greatly improved in 2004-2005. It i s estimated that up to 60,000 people have been 23However, parental health shocks were associated with less educational attainment inurban areas. 28 displaced, 500 wounded by land mines, and 95 killed. Some 230 villages were abandoned and 4,000 students displaced due to destruction o f schools.24 Several rural health centers closed due to difficulties to recruit personnel. The impact on HIV prevalence i s unclear, but higher rates are often associated with conflicts, and Ziguinchor and Kolda are among the highest rates in the country. Insecurity has resulted in loss of tourism receipts and private investments and a slow down inpublic investments. 71. An important and growing social risk for rural residents is the theft of livestock which affects many regions. As previously mentioned, theft accounts for 22 percent o f the responses o f rural residents as to the cause o f livestock loss. This general observation i s echoed by the finding o f a social assessment in the poorest rural villages in Senegal that found that "although it can be said that the team anticipated problems o f rural theft o f livestock, the researchers never imagined the scale of theft as a problem inthe daily lives of rural villages today as was reported during the interviews. It was a significant factor affecting the prioritization o f village needs in every case" (Sey and Wilson 2001). The Ministry o f Livestock estimates annual losses due to theft at 2 billion CFA (US$4 million). Given the importance o f livestock as a store o f wealth and a hedge against shocks to consumption, livestock theft significantly heightens rural vulnerability. 72. There is also increasing tension and potential conflict over rural property rights. The juxtaposition o f traditional property rights and those conferred through the modem legal framework, the absence or clear territorial boundaries between local governments as well as rural and urban administrative areas, conflicts over traditional usufruct rights o f pastoralist and lack o f regulatory instruments are at the origin o f these problems. These have manifested as conflicts between agriculturalists and herders in agro- sylo-pastoral zones. Additional conflicts have occurred between hereditary land users and those conferred rights through recent irrigation investments along the Senegal River Valley (White 2000), as well as resistance by the Peuhl to expansion o f agricultural lands in Kolda. Moreover, there are a growing number o f potential conflicts resulting from return migrants seeking to procure land through local governments. The exact number o f people affected or economic impacts o f these property right conflicts i s not known, but they demonstrate a rural space and economic systems in flux and in need of effective and transparent governance mechanisms to reduce the risks to investments and livelihoods. D. Overview of Risks, Multiple Shocks, Vicious Cycles, and Seasonal Concentrations 73. Most villages experience multiple negative shocks. In the PASEC survey o f 32 rural villages, only four experienced no village-level shocks in the last 10 years. These villagers were about as likely to say the overall situation had gotten worse as to say it had improved. The greater the number o f shocks, the greater was the likelihood of a worsening situation perceived by villagers. For the villages that had experienced at least four shocks, 80 percent felt their overall situation had worsened (Table 18). 24Despitethis dislocation, children continued to attend school, even inif inprovisional buildings. Ziguinchor Region, including its rural areas, has some o f the highest primaryenrollment rates inthe country. 29 Table 18 :Effect of Multiple Negative Shocks Likelihood of Community # villages Shocks affected deterioration 1shock 2 shocks 12 70% 3 shocks 6 73% livestock loss, insect infestation, fire, epidemic, famine 74. Even within the same village, these shocks hit householddifferently. From the PASEC survey, villagers estimated that drought and agricultural shocks were the most widespread. But even in the case o f drought, only an estimated 84 percent o f households suffered, and only 70 percent were subject to famine (Table 19). This i s very similar to a recent WFP food security rural household survey which found that about one-third o f households within villages were not at risk for food insecurity (WFP 2004). Flooding was perceived to only affect about half o f villagers. And closing o f a health center was seen as impacting 30 percent of the population, perhaps reflecting limiteduse o f health services. 75. With its close ties to agriculturalproductionand weather patterns,there is an important seasonal aspect to rural vulnerability. Mutually reinforcing patterns of risk overlap intime to make parts of the year extremely difficult. This critical period, known as the `hungry season', exacerbates vicious cycles and is a time of heightened vulnerabilities (Figure 6). Peak labor demands occur at a time o f highest risk o f malnutrition and malaria. Malaria peak coincides with the rainy season when transport for health emergencies i s most difficult. Lack o f available income in the pre-harvest period i s concentrated at the same time as heightened need for cash to pay for health treatment for malaria and to pay schooling costs associated with the beginningo f the academic year in October. This timing often delays the entrance of children into school as fees can only be paid after the harvest, Interventions timed to address critical risk periods will be more effective if they take into account the cyclical realities o f rural areas. 76. Certain negative shocks increase the likelihood of other adverse events occurring. A drought mayprompt temporary labor migration, which increases exposure to HIV for both the migrant and the partner who stays inthe village. A bad harvest and the resulting income shock may reduce the likelihood that a woman with a pregnancy complication seeks out adequate health care or increase the risk that a child drops out o f school. Not having attended school as a girl reduces likelihood o f neonatal tetanus shots as a mother. These vicious cycles o f vulnerability make it difficult for rural households to move out o fpoverty, or drive those that have improved their situation back into difficulties. 30 Figure6 :SeasonalConcentration of Rural Risks Agricultural Cycle [ Dry season -Preparationfor planting][Planting)[ Ifson I,........,......,.Rainy ...............I Culture / I -- # - - - FoodllncomeII II Availability -11 4 -I / ---I* 77. Table 20 summarizes the leading risks and vulnerable groups in rural Senegal. These risks create vulnerable groups that range in size from about 85 percent o f rural residents at risk o f being adversely affected by cyclical drought, to smaller groups, like those who are HIV positive or wounded by land mines inthe Casamance conflict. Groups exposed to general vulnerability, like the 1.8 million rural residents in the lowest consumption quintiles or the 1.4 million o f the poorest households who are engaged in seasonal agricultural labor, are less likely to be able to cope with any type o f shock. 31 Table 20 ;Summary Estimates o f Rural Populations at Risk I I# o fpregnant rural women not receiving pre-natal care I 61,400 # o f children 0-5 likelyto die before 5'hbirthday I 36,300 Ofthose, from malaridfever 9,500 Estimatedrural adult population 15-50 yrs infected with HIV/AIDS 41,0OO2' Rural AIDS orphans 12,000 Education Risks Illiterate adults 4,111,600 Rural children inschool but not receiving school lunches 511,200 Rural childreno fprimary age who are not inschool 308,000 Difficult physical access to schools -7-12 yr olds who live+ 3 kms 188,000 Employment Bottom 2 quintile households with principal occupation as seasonal 1,400,000 Risks labor Social Risks A t risk o f livestock loss from theft over 10 yrs 944,000 # displaced from Casamance conflict 60,000 # wounded by landmines 500 Other Disabled 70,800 Orphans (0-14 lost at least one parent) 238,000 Inaccessibility # living over 1 hour away from public transportation 1,939,262 78. Figure 7 highlights where these risks and vulnerabilities fall along a spectrum of extensiveness and severity. For risks that have a minor inimpact, like routine illnesses, small brushfires, etc., self-insurance via personal savings i s typically adequate to cope with associated costs and welfare losses. In such situations, recourse to negative coping strategies which impact on longer-term ability o f households to improve their socio- economic status are unlikely for all but the most destitute families. As risks move up the severity scale, the justification for risk-pooling and public intervention increases. In addition, for co-variate risks that affect a wide range o f the population and/or multiple risks concentrated in particular populations or at times o f the year, the ability to risk pool i s curtailed and recourse to public interventions is important. 2sBased on 2001 householdsurvey data, does not take into account investments made since that date. 26Assumes nationalprevalence rate applied to ruralpopulation, inabsence o f specific data on rural prevalence 32 Figure 7: Severity and Frequency of Risksin RuralAreas Catastrophic - Localiz I I I f Malariasufferers S e V e r i t Y non-enrollment 0 f I m P a C 1 1 t Minor - Localized Minor -Widespread 0 Number of people at risk 33 CHAPTER 111: Informaland PrivateRiskManagementMechanisms As afirst line of social risk management, how do rural households prevent, mitigate and cope with risk? What are the main responses at the household and village level? What are the main informal and market-based mechanisms? A. Overview ofHouseholdand CommunitySocial RiskManagementStrategies 79. Households and communities seek to decrease their vulnerability through long-standingpractices of risk management in the face of the vagaries of life in the Sahel. Strategies to prevent shocks from manifesting take many forms including vaccinating children and livestock, irrigation, clearing to prevent bush fires and so on. Strategies to mitigate the impact o f a shock include diversification o f production and revenue (e.g., crop diversification, sale o f garden produce and firewood, petty commerce) as well as self-insurance and risk pooling through social networks and informal savings clubs (tontines). Coping strategies include reduction in quantity and quality o f food eaten, sale of animals, use o f traditional medicines, non-enrollment o f children in school and recourse to temporary migration. Some short-term coping strategies can have long-term negative consequences, particularly when they lead to asset depletion, erosion o f human capital and environmental degradation. Mitigation instruments like holding livestock as hedge against hardship are ill-suited against covariate shocks like drought. 80. Diversification of sources of income and types of assets held is practiced by almost all rural households. Diversification reduces risk exposure, stabilizes income flows and increases potential liquidity in times o f crisis. From the PSIA survey, poorer households in the Groundnut Basin were more likely to have their household income concentrated in agricultural production, versus non-poor who were more diversified into livestock and other economic activities. Within rain-fed agriculture, throughout the 1990s there has been diversification away from groundnuts and other cultivations that have been most affected by adverse weather patterns and into such crops as watermelon and nikbe'. 81. Risk managementstrategiesvary by place and may change over time. A 2004 vulnerability assessment found that rural households innorthern Senegal responded to food insecurity primarily through sale o f animals, migration and income diversification, whereas adjustments to the daily diet, with high short-term impacts, were more prevalent in Tambacounda, Kaolack and Fatick (WFP 2004). Oral histories taken from the region o f Ndam Mor Fademba found that'inresponse to major catastrophic events over the last 60 years, such as the locust invasion o f 1950, a village fire in 1967, the drought o f 1973, a rat infestation in 1976, and locusts in 1988, changes ineating patterns and recourse to food aid had declined in importance as coping strategies (IIED 1993) in preference to rural-urban and international migration and the selling of large animals. The sale o f small animals like chickens remained a core coping strategy throughout this period.27 27A study o frural vulnerability inKenya found that small ruminants are usually more liquid, and inthe face of idiosyncratic shocks, display better market integration (Christiansen and Subbarao 2004). 34 B. Membership in Social Networks and Solidarity Groups 82. Social networks and local organizations are the frontline of defense in the face of shocks to households. Local-level networks and organizations exist to organize social relations, mobilize and regulate collective action particularly around production systems and management of natural resources, and to reduce and manage local risks. Mutual assistance through labor exchanges at critical agricultural periods i s common. These groups and associations may be important not only to foster village solidarity, but also to create links with the State or other external actors like NGOs inorder to attract resources. 83. Patterns and forms of association are heterogeneous in rural Senegal. The main membership associations consist o f producer's cooperatives, economic interest groups (GIEs), producers' cooperatives and workers' unions, water and irrigation user associations, village development associations, women's groups, youth and sport associations and religious organizations. These organizations can be village-based or federated in larger regional or national networks. They can be inclusive o f all members o f a certain village, typical o f community organizations, or more exclusive with membership limited as in the case o f producer organizations. Moreover, these may be induced through Government efforts, as in development o f producer cooperatives in the peanut basin and rural women's groups (groupements depromotionfeminine) inthe last twenty years. 84. The associative tissue is very dense in rural Senegal, creating webs of solidarity that are critical in such a risk- prone environment. As presented in Table I s organization 21,religious organizations, women's groups, Type of Association present in informal savings clubs (tontines) and culture village? and sport organizations are widespread, Yes present in the vast majority o f rural Religious Organization 80 communities. Parent-student committees are Women's PromotionGroup 69 prevalent (at 58%) and can be expected to Informal Savings club (Tontine) 67 increase over time as education opportunities Culture and S ort association expand in the rural areas. Associations Parent-Student Association around economic advancement, water and Economic Interest Grou (GIE) Village Development Association 35 health management are available to only about one-third o f rural households. For Cooperative 33 Water Management Group 31 water and health, this reflects the lower coverage o f these infrastructure services. Health Committee 28 85. This associative density appears to increase over time, with more formal and production-oriented organizations emerging. A 2002 survey o f villages in the Senegal RiverValley, the Groundnut Basin and the Niayes, found that 78 percent o f villages had at least one association, up from 22 percent reporting they had such organizations twenty years ago. Producer's organizations had grown at the fastest rate. Analysis o f the determinants o f organizations at the village level found that village size, presence o f non- agricultural activities (like fishing villages), larger holdings and closer proximity to other villages were associated with greater numbers o f associations (de Janvry et a12003). 35 86. Membership varies by type of organization, with some much more inclusive than others. As shown in Figure 8, when present, the vast majority o f community members in a village belong to religious organizations, women's groups, village development associations, informal savings clubs, culture and sports organizations and economic interest groups - even the poorest households. Cooperatives have broad membership, though participation i s lower for the very poorest households. Three-quarters of households are members in communities with a parents-student association. Membership i s lowest and least inclusive o f the poorest among health and water committees. Figure 8 : Share of Householdsthat Belongto Group by Poverty Quintile I -Village development 80.0 association -A- Women's group --sCReligious organlzation 70.0 -Culture - and Sport -0- Informal SavingsTontine +Economic Interest Group 60.0 -Cooperative -Water management group 50.0]-a Parent-student association . +Health committee 40.0 ~ 1 2 3 4 5 Household Poverty Quintile 87. There is some evidence that the pattern of distribution of benefits among members varies by type of organization. From the survey o f 280 villages cited previously, the more inclusive and more informal community organizations tended to be more selective of distribution of private or semi-public goods among members (79% receive benefits). The more exclusionary formal producers' organizations were restrictive of membership,but distributed benefits more widely (to 92% of their members (de Janvry et a12003). 88. Rural households contribute financially to the maintenance of these networks, across all income groups. Overall, 68 percent o f rural households report making financial contributions to associations (e.g. religious, social, cooperatives). In absolute amounts, rural households give an annual average o f 18,500 CFA to associations. Even the poorest rural households contribute 2 percent of their total expenditures (roughly equivalent to health spending). (Statistical Annex Table 1II.a). 36 89. These patterns of density and membership are important in designing programs that seek to build on existing social networks. First, despite the overall density o f associations, there are a critical number o f villages, typically the smallest and more remote, that do not have associations and tend to rely on family-based networks, with fewer links to `outside'. These are the most vulnerable and most difficult to reach with programs that work through existing associations. Second, existing informal associations may be extensive, but not be as inclusive in their distribution ex-post o f benefits. And lastly, organizations which have more established conduits to the state, namely school, health and water committees, may not be the best channels for broader public interventions because they are less present, have limited membership and the poorest may not be as represented. 90. Relying on family and informalredistributionand support alone has its limits. Gift giving is an important part o fsocial traditions. Inbothrural andurban areas, the value of gifts given i s spread fairly evenly between immediate family, friends and other people/associations. Most giving stays locally. Among rural households, 63 percent o f gifts remain in the village and 28 percent within rural areas, and about 8 percent goes to urban areas. The pattern i s similar inurban areas. This gift giving is consistent with Zakat, one of the five compulsory duties in the Muslim religion, which consists o f giving a fixed part o f wealth to needy people. Eligible recipients are as close as possible to the giver within the family, the village and thenoutside the village. For rain-fed crops, 10percent o f the harvest should be donated to vulnerable groups such as economically destitute people, disabled people, orphans, refuges and prisoners etc. For irrigated crops only 5 percent i s required, However, community solidarity networks are limited in that they are not effective for covariant shocks when everyone needs help at the same time. C. Labor MigrationandRemittances Migration 91. Households seek to diversify against rural risks by anchoring part of their revenue outside of the risk zone, through out-migration. Migration has become an important risk coping mechanism for many rural households. With little access to services, unpredictable crop production and greater exposure to shocks, family survival strategies often focus on having one or several family members go to urban centers, and ideally abroad, to look for work. The importance o f migration i s not so much that it relieves pressure in the sending communities since it i s often the `ablest' that move. Rather, migration establishes linkages with localities from which remittances are sent back to villages and `toe-holds' to which fbture migrants may move usingfamily contacts. 92. Even though marriages and other family events remain the main reason for migration,a significant amount of migrant households cite economic reasons, such as a lack of jobs, a new job in a different location, and access to land. The data also indicates that the rich are more likely to move in search o f better employment opportunities, while the poor are more likely to be forced to move due to conflicts, diseases and other insecurities. About 3 percent o f rural households report having moved due to insecurity, disaster and conflict. Data from the PSIA Groundnut Basin survey report that 37 one-third o f adults had departed at least once, mostly due to employment opportunities elsewhere. Men cited lack o f seeds and women cited marriage as well. Upper income households were more likely to have members depart for educational opportunities (6 percent) (Statistical Annex Tables 1I.b and1I.c). 93. International migration has grown in importance. There are an estimated 400,000-500,000 Senegalese residing abroad, and about 70 percent of households reported a member abroad in 2001. Although historically strong to France, particularly from the Senegal River Valley, in more recent times migration has diversified, including growing migration towards other European countries, particularly Italy, and the United States. Statistically, migration to other African nations accounts for the largest number o f Senegalese CmigrCs. The largest recipient countries have historically been Cote d'Ivoire, though the recent conflict there may change this pattern, Gabon, Mali and Guinea. Urban households were most likely to have a migrant in Europe, whereas the majority o f rural migrants went to other African countries (Statistical Annex Table 1II.d). Remittances 94. Remittances from abroad provide a cushion against internal shocks and may promote economic growth. Some studies have found that remittances have an impact on reducing poverty in receiving countries (Adams and Page, 2003). Remittances tend to be stable over time and may even rise intimes o f economic difficulty inthe recipient country (Ratha 2003). However, some argue that remittances dampen growth over the long term by reducing motivation to work and invest or by creating pressures to currency overvaluation (Chami and others 2003). There i s concern that remittances may worsen income inequality, including remittance-driven price increases for land and housing. 95. International remittances to Senegal have grown steadily and account for about 6 percent of Senegal's GDP (Table 22). This compares to an average o f 1.5 percent o f GDP for all developingcountries in 2002, and 2.9 percent o f GDP for the low- income countries. These levels understate the overall amount o f resource transfers as they capture only formal transfers through banks, money transfer operators like Western Union, currency bureaus, and post office-based financial services. Informal transfers incash or in- kindthrough family members, friends, traders and informal networks are substantial andby some estimates may be as large as or larger than recorded transfers (Global Development Finance 2004). Table 22 :InternationalRemittancesto Senegal, 1994-2004 Source: IMF Balance of Payments, Bank estimates. 96. While there is little systematic data on the share of remittances in rural household income, anecdotal evidence suggests that there are regions where this can 38 be an important, and even predominant, income source. ** For example, the Soninke and Pular ethnic groups fiom the Senegal River Valley have strong traditions o f rural to international migration and studies report that remittances may account for as high as 90 percent o f their household income (Cotoula and Toulmin 2004).29 The PSIA survey o f the groundnut basin confirms the widespread nature o f these transfers. Thirty percent o f poor and 25 percent o f non-poor households reported receiving non-pension transfers (national and international), equivalent to 27 and 36 percent of household income, respectively. (Statistical Annex Table 1II.e). From ESAM 11, 72 percent o f rural households with international migrant membersreportedthey send money back. 97. The priority use of remittances appears to be to smooth consumption and address basic needs. Recent studies estimate that current consumption accounts for about 75 percent o f remittance use (Sander and Barro 2003, C. Diop 2003, A.S. Fall 2002). Higher amounts may be transferred at critical moments, like to pay school entrance fees or to pay for labor at peak agricultural periods. About ten percent goes into savings accounts, often with the idea o f future investments in real estate. The remainder goes directly into investments to purchase o f land, buildings, or start-up capital to finance commerce or services (e.g. taxis). Very little appears to finance agricultural investments, though there are anecdotal reports o f vegetable production and livestock purchases. On a community level, ethnic groups with strong migrant traditions have formed associations o f Cmigrks abroad to raise funds and finance larger-scale projects, like schools, health centers and mosques, intheir villages o f origin. 98. Systems for informal transfer of remittances are varied and most likely account for the lion's share of transfers to rural areas. These informal systems encompass cash transfers through friends, relatives and traders, as well as in-kindtransfers that are either hand-carried to rural residents, or financed through private shopkeepers that then distribute goods through their kiosks. As they rely on social and family networks, they are easier to use for rural populations with little schooling and limited knowledge o f formal banking procedures. And, CmigrCs that lack legal status may hesitate to use formal channels as they may require providing identification or opening a bank account. Moreover, informal transfers are not subject to any ceilings, like formal transfers. 99. The Post Office and, increasingly, micro-finance institutions, transfer remittancesto rural households through formal channels. The formal banking system has limitedpresence inrural areas.30 The Post Office (laPoste) has the largest reach o f any financial service provider in rural Senegal, with 137 branch offices. All accept financial transactions. The Post Office uses its own transfer systems for within country money %ome researchershave estimated that remittances may represent between 30-80 percent o frecipient householdincome (van Doom 2002) and inregions o f highmigration, like Louga, remittances may represent 90 percent o f household income (Tall 2001 inAS Fall, 2003). 29Inan ethnographic review of the rural community Moudery, whose inhabitants are mostly Soninke with long experience o f migrations, o f the 32 elected local councilors, 7 have dual nationality and 22 have been or are migrants (as reported in Cotoula and Toulmin 2004). Moudery scores high for rural areas interms o f composite access to infrastructure and has a full complement o f basic infrastructure except full road access. 30Inaddition to postaltransfers, formal internationaltransfers are also effectedthroughwire transfers between bank accounts andprivate transfer operators like Western Union and MoneyGram. 39 transfers, but has contracted with Western Union for international transfers. The Post Office channels most o f the domestic financial transfers in the country and accounts for almost the vast majority o f all transfers to rural locales. It i s well-known to villagers and trusted and offers an array o f services, including savings accounts. However, there are occasional liquidity problems at rural branches that requiremultiplevisits from recipients. 100. The rapid expansion of decentralized financial structures (micro-finance institutions-MFIs) is creatingalternativeconduitsfor rernittan~es.~~ Most MFIs focus on traditional functions o f savings and credit (see following section), but an increasing number are entering the market o f money transfers. Two MFIs are formally involved inthe remittances market - 1'UNACOIS (Union Nationale des Commerpznts et Industriels du Se`ne`gal)and the DjoloffMutuelle d'Epargne et de Cre`dit (DJOMEC) in collaboration with a private company, Money Express, to handle international transfers. Most expansion in remittance transfers to rural areas i s expected to come through micro-finance institutions, either as they diversify the services they offer or enter into partnership with private banks, specialized money transfer services and/or the Post Office due to the advantageous position they have in terms o f extensive outreach to clientele and their ability to offer a range o f services including credit. This expansion o f services through MFIs plus the growth o f private sector transfer services like Western Union and Money Gram should create competitive forces that will reduce the cost o f transferring remittances. 101. There are several policyconstraintson further or faster developmentof formal money transfer services. Financial sector liberalization has increased the presence o f and diversity o f formal money transfer mechanisms, driving down unit costs and providing for quicker transfer mechanisms. However, UEMOA regulations limit amounts that can be transferred both internally and internationally (below US$5,000, depending on zone). For rural areas, the greatest impediment to increasing formal remittance flows i s the lack o f basic financial service backbone. In addition, the potential benefits o f remittances are mitigated by the high fees o fwire services or lack o f security o f informaltransfers. D. Savings andCredit 102. Modern financial instruments,includingsavings and credit, are essentialtools for consumption smoothing and growth. The poor need appropriate and sustainable financial services. Accumulation o f savings in financial stores allows for more efficient consumption smoothing in the face o f shocks, self-insurance against risk, safekeeping o f wealth accumulation and increased credit worthiness. In addition to financing investment and growth, savings and credit are crucial to smoothing consumption in the face o f emergencies. For example, research from Tanzania found that households respond to transitory income shocks by increasing child labor, but that the use o f child labor as a buffer is lower when households have access to credit, even controlling for wealth and household characteristics (Beegle et al. 2003) In Indonesia, households that were situated incloser physical proximity to micro-finance institutions had significantly less decrease in consumption following a health shock, controlling for wealth affects, mainly due to the 31The term used inSenegal `structures decentralishsfinanciers ' includes micro-finance institutions, savings and credit unions, etc. This report uses `microfinance institutions' MFIs as its equivalent. 40 savings effect of having financial services in close proximity.32 Access to credit and savings mechanisms may also be important for education attainment and to permit economic diversification by capitalizing petty commerce activities, a critical aspect o f income diversification inrural areas. 103. Most rural savings are still `on the hoof, although formal and informal financialsavings are growingin importance. From the PASEC survey o frural villages, livestock and small animals were the most common savings mechanism, but, over half o f villages cited access to savings accounts and a significant percentage mentioned informal savings through tontines or holdings o f cash. The survey found that 12 percent of rural households responded that a member had a deposit at a bank, credit union (mutuelles d'kpargne et crkdit) or other formal financial institution, versus 25 percent o f urban households. Forty percent had savings in a tontine (versus 53 percent urban). Rural poor were far more likely to save in tontines than better-off rural households (57 percent o f lowest quintile verses 18 percent inhighest) (Statistical Annex Tables 1II.fand 1II.g). 104. Credit sources are also largely informal, with most people depending on friends and family for loans. For both urban and rural residents, family and friends or other individuals (probably private money lenders) account for most loans to households (23). Formal credit sources, including government agencies, agricultural development banks, commercial banks and credit unions provide only 9 percent o f loans (in number not value) that rural residence say they have gotten, versus 20 percent for urban households. One concern about the reliance on informal sources, particularly on private individuals, i s the usurious levels o f effective interest rates and, inthe ' case o f shocks, the development o f debt traps that may lead to further impoverishment. 105. The micro-financeindustry in Senegalhas flourished in recent years. Over the last ten years the number o f MFIs has grown from 18 to 724 organizations, including various types of savings and credit unions (mutuelles d 'kpargne et credit MEC, groupement d 'kpargne et credit GEC, etc.) and NGOs or poverty reduction projects that handle micro- finance. As o f 2003, total client base was estimated at over 500,000,33with over US$85 millionin savings deposits and US$115 millionincredits. The overall portfolio i s healthy, with a default rate below 3 percent among the mainnetworks. There are 7 large networks 32For communities with a micro-finance (BRI) branch, health shocks had no net effect on consumption; in contrast consumption lowered by 2 to 3 percent inareas without a BlU. Savings, even inthe relatively small amounts, were useful for consumption smoothing. (Gertler, Levine and Moretti 2002). 33As some `clients' are savings groups, some estimates put the total number of individuals served by decentralized financial system at over 700,000 (IBM, 2004). 41 o f savings and credit unions that regroup over 400 individual GECs and over 100 MEC. These networks account for about 85 percent o f the micro-finance market.34 Nominal interest rates currently average around 18 percent, but taking into account commissions, fees and obligatory savings, effective interest rates can reach 50 percent (IBM 2004). 106. Micro-finance institutions tend to serve a poorer clientele than commercial banks, though they too face constraintsserving the pooreststrata of society and have a far greater presence in urban areas. A 2001 survey o f MFIs to assess targeting and impact found mixed results depending on which poverty measures were used. On dimensions related to human capital, MFI clients were better o f f than the Senegalese average; though on housing and access to land quality, clients tended to be worse off. Overall, micro-finance institutions that had more rural outreach were found to do the best job at reaching the poor (BCEAOKGAP 2001). Constraints on expansion o f MFIs inrural areas include low levels o f literacy, dispersed population and difficult access, lack o f collateral, and the inherent riskiness o f agricultural production the Sahel. Nonetheless, some o f the MFIs orient a significant portion o f their portfolio to rural areas. For example, CMS estimates that about 25 percent o f its portfolio is inrural areas (see Box 3 on CMS). 107. The majority of rural households that obtain credit use it for consumption smoothing. For both urban and rural households, personal uses - especially family consumption - account for over three-quarter o f the loans taken (in number, not amounts). Households most often use credit to stabilize consumption and confront crises and to invest in social capital. Loans are taken more frequently to diversify sources o f income out of agriculture (14 percent) than to invest in agriculture (9 percent). About 7 percent o f loans went to finance health care and education (Statistical Annex Table 111.0. 108. Most funding of agricultural credit is provided through the semi-public commercial bank CNCAS, to finance seeds, fertilizers and other inputs for the agriculturalyear. The MFIs play only a marginal role, providing about 4 percent o f all agricultural credit. Over the agricultural seasons 1997/1998 - 2002/2003, the CNCAS injected a total o f 67.5 billion CFA (US$130 million). Most o f this (80-85 percent) was for short term loans which finance inputs to that year's agricultural campaign. Some medium- term credit i s available for equipment and commercialization credits for agricultural transporters. Except rice and horticultural produce which are cropped under irrigation conditions in the Senegal River Valley and the Niayes areas, the bulk o f CNCAS short- term credit goes to rain-fed crops like groundnuts and cotton. Reimbursement rates tend to be very low and as low as 28 percent in 2002 when the country experienced drought that severely impacted groundnutproduction (Statistical Annex Table 1II.i). 34These networks are: Alliance de Crkdit et Eparngepour la Production (ACEP) Crkdit Mutual de Sknigal (CMS), Uniondes Mutuelles du Partenariatpour la Mobilisation de 1Epargne et Crkdit du Senegal Rkseau des Caisses d'Epargneet Crkdit des Femmes de Dakar (RECECIFD), Reseau des Mutuelles d'Epargneet Crkdit de UNACOIS(REMECU), Union des Mutuelles d'Epargne et Credit (UMEC), Union des Mutuellesd'Epargneet Credit de UNACOIS(UMECU) 42 E. Insurance 109. Insurance helps households mitigate the adverse effects on income and consumptionof health crises, death, disability, old-age, crop loss, business failure and losses through theft, fire and accident. Without access to formal market-based insurance mechanisms which allow for risk pooling, households must `self-insure' against loss through savings and asset accumulation, use informal social-based arrangements, and/or face significant reductions in income and consumption in the face o f these adverse events (or hope for emergency support from public safety nets). The vast networks o f solidarity groups, significant expenditures on membership, gifts, ceremonies and the like can be viewed as one big informal insurance system, committing people to mutual obligation in times of need. The need to self-insure reduces opportunities for riskier and more profitable investments, and may lock assets into lower-return holdings. Inaddition, informal systems m a y be helpful for idiosyncratic shocks, but are less effective at covering covariate risks affecting the whole community, like drought and macro-economic shocks. 110. Formal insurance mechanisms are extremely limitedin Senegal, and virtually non-existent in rural areas. Formal social protection, under the guise o f social security and social insurance provides coverage almost exclusively to the public and formal private Box3: Micro-Finance inRuralAreas-the Experienceof CriditMutuel du Sinkgal (CMS) CMS i s one of the three largest MFIs in Senegal, and with PAMECAS and ACEP, accounts for about three-quarters o f market share. CMS was created in 1980 out of a program to support the creation o f credit and savings mutuals largely among agricultural communities inthe Groundnut Basin. Today, CMS has 71member caisses, o f which 42 are inrural communities. As o f 2004, CMS had U S 5 7 million in deposits, o f which 25 percent rural, and US$40 million in outstanding loans, o f which 30 percent rural. Savings have almost doubled and loans almost tripled over the last 3 years. CMS builds rural savings and credit unions from the bottomup, with local populations forming `caisses ' with a minimum o f 300 members, administered by a local Administrative Council. CMS clients have access to three types of credit: (a) personal credit for purposes like purchase of durable goods etc, (b) professional credit, which in rural areas i s mainly to finance the purchase o f seeds and equipment, but can also finance commerce; and (c) a `hungry season' credit initially designed to help finance consumption needs prior to harvest, although it has been rarely accessedfor limited debt capacity o f peasants. Most credit i s short-term (under one year) and credit amounts are accorded inrelation to savings capacity. Clients must save for at least six months before accessing credit, a maximum amount of 3 times savings. CMS provides transfer services for salaries to government and private company workers posted to rural areas, as well as government pension payments. For rural teachers this has reduced the cost and time lost o f going to the capital for monthly payment. Domestic and international money transfer services will begin offered inthe coming months. In addition, CMS has joined with other MFIsinWest Africa to form the recently created IMAO(InstitutMutuel de L'Afrique de I'Ouest), a commercial bank, to provide longer-term credit and attract capital not linked to member savings. In terms of expanding access to rural populations, CMS management notes that the most significant constraint i s the US$30,000 average upfront installation costs (building, motorcycle, etc). Moreover, there i s an on-going consolidation, or rather federation, o f smaller, rural caisses which are mergingwith larger ones nearby to enhance the sustainability o f rural finance. (salaried) sector. For formal sector workers, who fall under the Labor Code, social insurance including health, retirement and disability coverage i s provided by the autonomous Institutions de Pre`voyance Sociale (IPS). These include the Caisse de Se`curitk Sociale (CSS), the Institution de Pre`voyance Retraite du Se`nkgal (IPRS) and the Institutions de Pre`voyance Maladie (IPM). The CSS covers maternity, work-related accidents and illnesses. IPRES providespensions, disability and life insurance and the IPM manages health care. For public sector employees, health coverage and other benefits are provided directly by the Ministry o f Finance and pensions through the national retirement fund. Public andprivate social security in Senegal covers about 300,000 people, including about 5 percent of the rural population (ILO 2002).35 111. The most active segment of the insurance market involves mutual health associations. In Senegal, health mutuals are definedas non-profit associations that provide insurance to their membersand their families to help support the costs o f health care. This excludes private health care insurers by nature o f their non-profit status and membership- based structure. Payment o f monthly dues gives members the right to benefits defined by the particular association. Health mutuals are viewed as an alternative model to public financing o f health, transferring cost from the public to the private sector, as well as a form of social protection to pool risks and to mitigate the cost o f illness. 112. Health mutuals haveincreaseddramaticallysince pilot efforts in the mid-1980s and now count over 90 nationwide. A recent survey o f health mutuals undertaken by the Government o f Senegal and the Canadian Research and Development Institute estimates 66,000 payingmembers and 220,000 total beneficiaries o f existing health mutuals, or about 2.2 percent of the p ~ p u l a t i o n . ~These include those who join `complementary' health ~ mutuals formed by public sector workers' associations to provide coverage for co- payments, purchase of medicines, and other expenses not covered under the public system, community health mutuals typically based in neighborhoods or villages, and professional mutuals organized mainly by informal sector professional associations, like artisans, women's associations, etc. Despite the growth o f health mutuals, researchers estimate that well under 10percent o f total beneficiaries are inrural areas. 113. Several donor agencies, including USAID and ILO, work directly with private organizationsto develop healthinsuranceschemes. For example, the ILO is supporting health insurance mutuals as part o f its strategy to extend formal social protection systems to workers, Because income loss from health shocks undermines other social protection measures like long-term savings for old-age, Strategies and Tools Against Social Exclusion and Poverty (STEP) was initiated to develop appropriate health insurance models for the informal sector. With a Senegal country budget o f about US$ 400,000 per year, STEP supports the Union des Cooperatives Agricoles du Senegal (UNCAS) a farmer's organization of 800,000 members to establish a social protection system, including health insurance, An initial pilot test inthe Niayes region linked health posts, the health center of Tivaouane and the regional hospital in Thiks. A second pilot test inDiourbel was recently 35The PSIA survey confirms this low coverage of formal old-age insurance, with an average of 1.9percent of poor households and 1.5 percent ofnon-poor households receiving pensions. 36ILO estimates are slightly higher, at 120 healthmutuals covering about 324,000 beneficiaries (Protection 44 launched. The first scheme covers about 1,200 people who pay 300 CFA each inmonthly premiums for generic drugs, hospital visits, obstetric services, pre-natal check-ups and in- patient care hospitalization. STEP i s also supporting PAMECAS (a micro-finance network o f 130 000 members) to explore linking health insurance with micro-finance. After a pilot test 2003, PAMECAS plans to extend the experience more o f its members o fthe network. 114. The ultimate potential of community-based schemes is as yet untapped. Benefits include protecting against the financial hardships imposed by illness, the freeing up of other self-insurance means (savings and credit) to be put to greater productive use, better health outcomes by encouraging timely use o f services, and the potential to reach a wider client base than for-profit insurance providers. However, constraints include the ability to reach the poorest, exclusionary decision-making at the local level, limited scope of the risk pool for small community schemes exposing them to covariate shocks, and weak managerial and leadership capacity that threatens long-run sustainability. There appears to be potential to extend health insurance through the larger-scale producer cooperatives and federations due to the experience with collection of membership fees and administration. However, given membershippatterns of producer cooperatives, these may exclude some o f the poorest households. Micro-insurance offered through micro-finance institutions i s another option, but its reach into rural areas will depend on MFIrural expansion. I Box 4: ExtendingHealthInsuranceto RuralAreas Evidence ofImpact - Impact evaluation of mutual health insurance systems in rural Senegal shows that risk- pooling and pre-payment, no matter how small-scaled, can improve the financial protectionfor the poor. The Thies region i s home to one of the first community-basedhealth insurance schemes inthe country, initiated in 1990 around the servicesprovidedby anon-profit hospital, St. Jean de Dieu. Ninety percentofthe schemes operate inthe rural area. Tojoin costs 1,000CFA per householdandmonthly premiumsare between 100and 200 CFA, or an estimated 2 percent of annual household income. Only hospitalization costs are covered (50 percent reduction in hospital costs). The share of households in each village that are members ranges from about 30 to 90 percent. An impact evaluation basedonhouseholdsurveys of members and non-members found that membership had a strong effect on the probability of going to a hospital when sick and a strong negative effect on expenditures in the case o f hospitalization. Members stressed that healthinsuranceledto a reduction intheir worries. Inthe case of serious illness, they no longer depend on assistance from their social network or money from the local money-lender. However, there was little participation by the pooresthouseholds, who appeared not to be able to afford the premiums. 1 Source: JuttingandTine 2000 45 CHAPTER IV: Review of Public PoliciesandExpenditureson Social RiskManagementinRuralAreas What is the mainfocus ofpublic interventions to manage risk and shocks in rural areas? Which Ministries are involved? What is the level of expenditure, program coverage, targeting and geographical distribution of public social protection measuresfor rural populations? How aligned are these expenditures with the main rural risks? A. Overview of Institutional Framework and Taxonomy of Programs 115. Risk management ac.tivities are distributed across several ministries and agencies, including: (a) Ministry o f Family, Women and Social Development (MFFSD) and the Ministry o f National Solidarity (MSN); (b) Ministry o f Health and Medical Prevention; (c) Ministry o f Education; (d) Ministry o f Labor, Public Administration and Professional Organizations; (e) Ministry o f Small and Medium Enterprise Development, Micro-Finance and Women Entrepreneurs; (f) Ministries in charge o f agriculture, rural water and livestock; (g) Ministryo f Interior; and (h) Ministryo f Youth. These ministries account for 43 percent o f the 2005 national budget (investment and recurrent), though only part of this i s to manage rural risks.37The review covers 2003 actual expenditures and preliminary figures available at the time o f review for 2004. Final figures on 2004 mayvary depending on actual disbursementso f authorized credits. 116. This review covers expenditures specifically targeted to rural areas as well as national-scale programs that operate in the rural space. The rural area cannot be looked at in isolation from national policies and programs. One o f the challenges to the government o f Senegal i s not only to design interventions with relevance for rural areas but also to develop implementation mechanisms that ensure that rural populations receive fair coverage o fprograms that are meant to be national in scope. 117. Social risk management programs have been classified into broad categories that relate to the type of risks they seek to address. Programs with similar objectives are often divided among several ministries. The chapter i s organizedby type o f intervention to better assess overall coverage, gaps and coherence o finterventions: I Social Assistance, Safety Nets andDisaster Management # Managing Critical Agricultural Risks Managing Critical Health Risks # Managing Critical Education Risks I Employment Creation I TargetingofRural Infrastructure and Services I Expanding Savings and Credit Mechanisms to Rural Areas 37Alsoincluded are two programs under the Prime Minister's Office: the Nutrition Enhancement Project and the National HIV-AIDS Council and Technical Secretariat. Not included are: Ministry o f Justice: which protects the legal rights o f certain vulnerable groups and provide services for juvenile delinquents; Social Security System (INPRS. CSS),expenditures fall outside o f the review as they are less relevant in rural areas. Information i s not available o n social programs funded by the National Lottery (LONASE), nor on ad-hoc transfers directly from the President's Office or through the Ministryo f Local Government. 46 B. Social Assistance, Safety Nets and Disaster Management Overview of Institutions, Policies and Expenditures 118. Social assistance, safety nets and disaster management are meant to target individuals and households at heightened risk of extreme adverse outcomes. These are the groups who are least likely to rebound from shocks and for whom vicious cycles create long-term poverty traps as assets are depleted and long-term damage is done to human and physical capital accumulation. This includes individuals in extreme poverty, children at- risk,displaced persons, the disabled, orphans, elderly without support, andpeople impacted by disasters. The majority o fthese groups live inrural areas. 119. The focal point for disaster management is the Civil Protection Department of the Ministry of the Interior, supported by an ad-hoc institutional framework to provide emergency response. As a first response, Civil Protection intervenes to protect from loss o f life and stabilize the situation. In 2002, Civil Protection developed regional maps of main risks, including industrial accidents, transportation risks (auto, air, maritime, trains), fire, floods, auto and maritime safety, and crowd safety. Both the MFFDS/SN and NGOs like the Red Cross provide emergency supplies, cash and rebuilding materials. The Ministryof Agriculture responds to the after-effects on production capacity inthe case of rural natural disasters (covered under the following section on agricultural risk). 120. For social assistance and safety nets, the Ministry of Women, Family and Social Development (MFFDS) and the Ministry of National Solidarity (MSN) have the basic mandate to protect the most vulnerable families and individuals. This area has known a great deal of institutional instability in recent years. The ministry was recently separated into two ministries and both have issued new policy orientations. The MFFDS's Letter of Sectoral Policy 2005 prioritizes actions towards children at risk, women, vulnerable families, the disabled, the elderly, and vulnerable communities. Inaddition, the M S N developed a "New Orientation o f National Solidarity". Through its executing arms, the National Solidarity Fund (NSF) and the Commissariat o f Food Security (CSA), the M S N will support: (a) social infrastructure inpoor communities, (b) emergency assistance and food security to vulnerable groups, and (c) the creation o f a National Solidarity Bank, 121. The Government spent US$43 million in 2004 on social assistance, safety nets and disaster management, including US$22 million in poverty-focused community development programs under the MFFSD. The more traditional social assistance and disaster management totaled almost 20 million (Table 24). Recurrent expenditures o f MFFDS/SNincluding social assistance transfers to vulnerable groups were about US$12 million. Over $US20 million financed multi-sectoral community development projects supportedby externalpartners. These are included here because they are under the tutelage of the MFFDS and target vulnerable communities, even though the bulk o f investments are insmall-scale infrastructure or credit rather than social assistance. 47 Table 24 : Estimated AnnualExpenditures on Social Assistance and SafetyNets3' (ofwhich UNICEF) (219,000,000) ~ Small Credits to Women 1,000,000,000 1,200,000,000 (of which Taiwan) (700,000,000) - Agence de Fond deDCvCloppmentSocial (AFDS) 2,548,000,000 5,272,000,000 Source: DDI, donor agency reports. Note: All of the MFFSDiSN operating budget and transfers included. Does not include CFA 250 millionin204 through Ministry of Justice for a building for youth in conflict with the law. Also does not include an ILO-managed program against child labor in the Ministry of Labor and Public Administration as no expenditures were reported inthe BCI. 122. Spending on social assistance and safety nets, not including the community poverty investmentsand fire protection, amounted to about 0.15 percent of GDP in 2004. Duringrecent drought years (2002-2003), this rose to 0.5 percent of GDP reflecting emergency food and materials to support rural populations (see Box 4.40The Government 38Does not include the emergency distribution of rice to rural areas in2002-2003 under Programme d'Assistance au Monde Rural via extraordinary expenditure for US$22 million in2003 (see Box 5). 39Of which 380,338,000 indirect food purchaseand the remainder to support CSA functioning. 40Additional resources are channeleddirectly to farmers to respond to shocks to production, describedin the following sectionon agricultural risks. Farmer support programs are not typically includedinsocial assistance estimates, but ifthey were the figure for 2004 for social assistance and safety nets would rise to 0.58 percentof GDP. 48 spends an additional 0.85 percent o f GDP on social security for public sector workers, bringing the total of social security and social protection to about 1 percent o f GDP. As incomes rise, spending on Senegal 1.16 safety nets tends to increase as a share o f GDP, Sub-SaharanAfrica 1.44 although there are significant variations within South Asia 1.49 regions and between regions. In a 2003 East Asia and Pacific 2.373 benchmarking exercise which compared public Latin America and Caribbean 2.91 expenditures on social security and welfare (slightly Middle EastNorthAfrica 4.66 different definition to one used above) in over 63 Eastern and Central Eurooe 10.31 I NorthAmerica I 11.19 I countries from 1972-97, Senegal spent an estimated 1 1.16 percent o f GDP (Table 25). This i s slightly I Source; Besley,Burgess andRasul2003 I Western EuroDe 113.57 I lower than the Sub-Saharan average despite Senegal having a per capita income slightly above the SSA average. EmergencyResponseto NaturalDisasters 123. Emergency relief from local flooding has been the main disaster response in 2003-2004. Overall, the Government support was at least US$2 million with additional support from the Red Cross. Although information i s limited and depends on the specific emergency. A rough estimate o f the cost per person has averaged $30-$120, Responses by agency include: Civil Protection: The investment budget reserves 300,000,000 CFA (US$ 600,000) per year for expenses in response to disasters and spent an additional US$3 million annually on fire protection in 2003-2004 including buildings and equipment. NationalSolidarityFundwas created in2002 as an autonomous agency under the then-MFDSSN to provide support to the critically poor through a start-up budget o f 1billion CFA (US$2 million), and amandate to raise funds privately as well.4' The first actions were in response to the 2003 floods in the region o f Matam and Tambacounda, providing tents, blankets, mattresses, lamps, rice, and other basic needs, as well as financial compensation (Statistical Annex TableIV.a).42 Governors transmit overall needs and Sub-prefects identify individual beneficiaries and distribute goods locally. Commissions of local authorities and deconcentrated staff are in charge o f reconstruction o f destroyed houses, with the FSN providing equipment for reconstruction. FSNhas spent about US$1.3 million to date. MFFSD Social Action Department(DSA) andthe CSA. Inthe case of shortfalls, the DAS and CSA can intervene with emergency support based on their food and grain stocks. For example, in response to the late rains in 2003, the CSA sent part o f its cereal stock from 2002 to Matam, Saint Louis and Louga to feed livestock. The Red Cross of Senegal: The Red Cross has a community-based approach to disaster response, built on a structure o f local voluntary committees. The RedCross 41Local fundraising on the first Day o f National Solidarity March4 2005 collected CFA 43 1,000,000 million from government officials, embassies and other benefactors to be usedinhture programs of FSN. 42100,000 CFA for death and 5,000 CFA per injured. 49 intervened in December 2003 with the floods in Saint Louis (assisted 500 families), inearly 2004 inMatam(1,341 families), andinmid-2004 inKanel(1,126 families), with rice, tents and bedding, mosquito nets, emergency medicines and reconstruction materials. The Government provides seconded staff and some operating funds to the NGO, and the Red Cross raises funds internationally. Social Assistance to the Handicapped,Elderly,Orphans and Others in Need 124. Assistance to associations and individuals, totaling almost US$2 million in 2004, is treated on a case by case rather than programmatic basis. Requests for social assistance typically in the form o f letters from individuals or groups/institutions are presented either directly to the Minister's Office, to the main office o f the DAS or channeled through the regional services.43 Support can be in the form o f equipment or vocational training for the disabled, grants to organizations and individuals, cash or food as~istance,~~ or payment o f medical bills for the indigent through four hospitals in Dakar. In 2004, out o f 3,547 requests received by the DAS, 910 were approved. The average amount financed per request i s 550,000 CFA. In 2004, the D A S provided 200 disabled persons with equipment and financed 48 associations o f disabled people, supported 160 training-oriented structures for children (including to daara to improve conditions o f the talibis), supported 133 widows and orphans, placed 200 at-risk children into educational and training programs, and funded 97 income generating projects for elderly (Statistical Annex Table 1V.b). Assistance covers only about one-quarter o f the requests officially received an insignificant portion o f the target vulnerable groups and few rural residents. Children-at-Risk 125. Assistance to children-at-risk is provided in addition through two programs to fight child exploitation: (a) Proiect Against the Worst Forms o f Child Labor, through the MFFSD supported by UNICEF and the Italian government, seeks to improve the situation o f children in dangerous or exploitative work in 6 regions (Dakar, Thies, Saint-Louis, Louga, Fatick and Ziguinchor). The project has provided professional training for 480 girls in Fatick, literacy training for 1,370 girls at-risk, educational materials for 800 children, the construction o f 3 health posts, and a pilot to eradicate child begging. (b) Child Labor Promam: A child labor program supported by ILO with U S financing executed by a project unit in the Ministry o f Labor seeks to help 3,000 children under 18 get out o f dangerous work situations, including begging, domestic workers, and dangerous jobs in agriculture, fishing and livestock, as well as prevent 6,000 others from entering, supporting 1,000 families in the Fatick, Saint Louis, Thies, Diourbel, Kaolack and Dakar (average cost o f about $75 per childper year). The program supports the legal framework, training andassistance to families. 43The D A S has 42 Centers for Social Promotion and Reinsertion staffed by social workers as well as personnel seconded by local governments or recruitedby the Community Management Committee o f each center. These centers, o f which 12 are inDakar and 30 inother regions including two located inrural communities, provide outreach services to address social needs. 44The D A S channels donations (muttons and dates) from Saudi Arabia for distribution to the poor. 50 Supportto VulnerableWomen 126. The MFFSD assists women's groups through subsidies to women's associations,direct provisionof goods andequipment,andsmallcredits,including: Project to Support Vulnerable Groups, financed by the GOS ostensibly to help women reduce onerous work (e.g. by providing millet grinders, water pumps, drip irrigation). The MFFSD purchases directly the goods, valued at over US$1 million annually, and distributes to women's groups. In 2003-2004 the bulk o f purchases went for computers, office furniture, tee-shirts, refrigerators, etc. The Program o f Small Credits to Women's Groups, supported by the Government of Taiwan and GOS, helps women generate income through credit and training. During2001-2004, a credit line of 3.1 billion CFA financed small loans to women entrepreneurs at 5 percent interest and, via CNCAS, a guarantee fund for larger loans at 8 percent interest. In2003, 567 projects were financed for 1.6 billion CFA, with 40 percent inDakar. The needfor guarantees tends to exclude the poorest. Support to NGOs, as tutelle the MFFDS oversees NGOs interventions including women's associations. In 2004, the Community Development Department approved 24 projects (average cost US$25,000) presented by women's groups in health, education, water supply, and income generation and micro-credit. The bulk of MFFSD investments are executed through three externally- financed community development projects that work mainly in rural areas, combiningtraining, credit andinfrastructure: Project to Fight Poverty (Projet de Lutte Contre la Pauvretk-PLCP): a $21.5 million project financed by BAD and NFD to support community groups in Tambacounda, Thies, Kolda, Diourbel, and Dakar in literacy training, income generating activities via a credit line o f CFA 1.4 billion channeled through MFIs, and small-scale infrastructure (executed by AGETP). Different credit windows exist for individuals, associations and enterprises with different loan sizes and interest rates (12 to 14 percent). From March 2003 to February 2004, 18,500 loans went to 36,900 beneficiaries for 1.6 billion CFA, with a recovery rate o f 48 percent. Social Development Fund Agency (Agence de De`ve`loppment Social- AFDS): - US$30 million IDA-financed program in poor communities in Louga, Fatick, Kolda, Kaolack, and Dakar that supports small-scale infrastructure executed directly by communities, income generating activities through CBOs and institutional strengthening to MFIs to improve their outreach to rural areas. In2004, AFDS financed 226 infrastructure projects (CFA 3.3 billion), mainly in health, education and water supply and 332 income-generating projects (CFA 1.4 billion) inlivestock, agriculture andcommerce (see Figure 9 for targetingperformance). Program for Poverty Reduction (Programme d 'Appui d la Re`duction de Pauvretk - PAREP): a UNDP-financed $4.5 million project initiated inmid-2004, buildingon the Programme Elargi de Lutte contre la Pauvretk that had previously operated in Dakar, Diourbel, Saint Louis and Tambacounda to support local planning, provide training and credit to women (through MFIs) and build small-scale infrastructure. 51 GeneralFoodAssistance 128. The Commissariatfor FoodSecurity(CSA) ensures food securitythroughfood distribution and monitoringof countryconditions. The CSA used to stabilize prices by purchasing cereals at peak production times and selling grains in food deficit times particularly during the hungry season to prevent price spikes in periods o f scarcity. This program ended with the liberalization o f cereal markets in the early 1990s. CSA currently produces a monthly food security bulletin on regional grain prices, rainfall, agricultural production, grazing pastures, population migrations and remittances through the Post Office, and health, including malaria cases and nutritional status. Information i s gathered through local Food Security Committees headed by the S~b-Prefects.~~ 41 4 2 43 4 4 45 Poor Non-Poor Received `help' from Government (PASEC-all rural) 76% 82% 91% 84% 94% Received food assistance -all sources(PS1A groundnut basin) 90% 92% 90% 92% 86% 91% 90% Received assistance from Gov't- all kinds (PSIA) 85% 88% 87% 87% 85% 86% 87% 129. CSA handles most of the Government's direct distribution of food. There are four on-going food assistance programs: (a) ad-hoc assistance to populations affected by natural disasters or supported through requests to D A S (discussed above), (b) regular rice distribution through the CSA, (c) limited amounts of food aid used within targeted programs inhealth and community development (via U S Food for Peace and WFP), and (d) a school feeding program supported by WFP, Italy and GOS (discussed in the section on educational risks). Most direct food distribution o f CSA is for the AVCER (Food Assistance to Religious Ceremonies). In 2004, CSA distributed over 5,000 tons o f rice, 451n2003, Senegal joined the international Famine Early Warning System (FEWS), funded by USAID, which collects data from CSA, MOA, National Meteorological Service, Center for Ecological Monitoring (CSE), WFP, FAO. Informationis primarily usedto program support fromUSAID and WFP. 52 mostly at religious events (Statistical Annex Table IV.c). Allocations are based on requests through the regional Governor's Office or directly to the central level, but information is not available on number or location o fbeneficiaries. A small was usedfor emergencies. General Observations Observations on social assistance, safety nets and disaster management are: There is an active network of government and non-governmental organizations to respond to emergencies but responsibilities overlap. The entry into disaster relief of the NSF has improved availability and timeliness o f funds. But, at the ground level, the parallel approaches o f community committees and top-down distribution through local officials can reduce efficiency. Coping strategies in flood-prone areas should not overshadow the need for preventative investments like flood control and development o f mitigation mechanisms like insurance. Social assistance and safety net operations suffer from a dispersion of efforts and multiplicity of actors and approaches. There are several windows o f assistance for individuals and groups, with overlapping mandates, different criteria, mechanisms and program rules, and little coordination. Donor-financed programs remain in silos with few contacts between each other or with the transfer programs of the MFFSD, missingopportunities for synergies and institutionbuilding. Targeting criteria appear loosely defined and monitoring and evaluation is seriously under-resourced, particularly for domestically-financed programs, The identification o f all women, disabled, elderly and children as vulnerable dilutes program focus. Food distribution i s not viewed in a risk management focus (i.e. programmed in accordance with agricultural production cycles or distributed via criteria generated throughthe food security bulletin). Moreover, there i s little effort to evaluate the cost-effectiveness or impact o fthe assistance provided. The role of the FSNin unclear. The FSNis preparingto buildbasic infrastructure in underserved communities, to create a National Solidarity Bark to provide subsidized credit to the poor, and even provide support to needy individuals (through a fund for medical emergencies). This appears to duplicate on-going programs o f the MFFSD (AFDS, DAS). The division into two ministries creates inefficiencies and coordination challenges in social assistance and disaster management programs. Managing Critical Agricultural Risks Overview of Institutions, Policies and Expenditures 131. The public sector helps rural producers manage natural and economic risks. The main natural risks to agriculture are drought, floods, locust and other infestations, and disease. The majority of the rural population i s at risk for loss o f production through extreme drought on average about every five years, with below average rainfall levels occurring one year out of three. Selected populations suffer from the other risks. 53 132. In June 2004, the Government adopted the Loi d'Orientation Agro-Sylo- Pastorale. Its policy orientations are to: (i) reduce the impact o f economic, climatic and environmental risks; (ii)improve revenue and living conditions o f rural populations; and (iii)improve the quality of life of rural populations. The law introduces a greater attention to risk management instruments and calls for a social security system for agricultural workers as well as protection against natural calamities and development o f crop insurance. 133. The Ministries in charge of Agriculture and Rural Water and Livestockhave the primary responsibilityfor addressing agriculturalrisks. The central levels of these ministries are supported by regional Directorates o f Rural Development (DRDR, for agriculture) and regional veterinary services (Inspections re`gionales des services veterinaires, IRSV), as well as departmental rural development and veterinary services. Advice and technical assistance i s provided through the semi-autonomous National Agency for Rural and Agricultural Extension (ANCAR). The Ecological Monitoring Center (Centre de Suivi Ecologique - CSE) provides information for the management o f natural resources, including mapping and satellite imagery used to monitor and predict shocks. Local actors, including local government (Communautks Rurales, CR) and producer organizations through the Conseil National de Concertation et de Coope`ration des Ruraux (CNCR) play an increasingrole inmanagingagricultural risks. 134. Overall expenditures on agriculture, livestock and related rural production amount to US$152 million in 2004 (Table 26). In 2004, the rural sector, including agriculture, livestock, fishing, forestry and rural water, accounted for 15 percent o f the intemally-financed investment budget and 24 percent o f internal and external resource combined (see Statistical Annex Table Ned).Specific risk mitigation and coping programs account for about 10 percent of total expenditures. Preventionof Shocks to Agriculture 135. Many projects in the investment budget support the general development of agriculture, contributing in various ways to the prevention of shocks. The bulk of investments in agriculture and livestock are to increase productivity through irrigation, improvement o f technologies, and support to the development o f producer organizations. Irrigation-related investment projects alone account for between 40-60 percent o f the annual investmentbudget o f the Ministryof Agriculture (between $20-50 million per year). However, given the limited potential o f irrigated land in terms o f total hectares and population with access to irrigated land, these investments are trained on a fairly narrow range o f the existing agriculture sector. Smaller programs to prevent certain diseases and infestations, like veterinary services, control o f white fly (bemisia) and others account for under US$l million per year. 54 MinistryExpenditure Area/program (1) 2003 2004 Total Ministry of Livestock: 2,329,000,000 3,706,000,000 Of which veterinay sewices/control of animal diseases 480,000,000 434,000,000 Total Min. Natural Resources and Environment 2,675,000,000 10,708,000,000 Of which Combat Brush Fires 95,000,000 164,000,000 Total (CFA) 43,325,000,000 76,O 15,000,000 Total (US$) $86,650,000 $152,030,000 Specific RiskMitigation and Coping (CFA) (3) 3,536,000,000 9,142,000,000 Specific Risk Mitigation and Coping (US$) $7,072,000 $18,284,000 (2) Of which 3 billion CFA LFR (Mid-Year Budget Supplement) Locust Invasion. (3) Includes the three funds (calamity, bonificationand guarantee) andprograms to control fires, pest and parasites 136. In addition, the Ministry of Livestock recently announced the launch of a programto limit livestocktheft. With an estimated cost o f 1.7 billion CFA, the program would put inplace a systemfor identifying livestock. The Government carried out a test in two regions in 2001 and 2002, building on experience in other countries like Guinea. The system registers livestock by attaching an identifier to the ear with information on village, CRYregion, and owner, linked to an owner's card. AgriculturalRiskMitigationandCopingMechanisms 137. Following intensive and countrywide consultations on agricultural credit series of financialmechanismswere put in place in the late 1990s to mitigate and cope with risks to agriculture and to seek to ensure an adequateflow of credit to the sector: They are collectively known as the Funds for Security o f Agricultural Credit (Fonds de Se'curisationdu Cre'ditAgricole FSCA). (a) The Fonds de Bonification 0%) was created to improve access o frural producers to credit, reduce the financing charges they face and thereby increase profitability o f agriculture. The FB finances the difference between the interest charged by 55 commercial banks (13 percent for the CNCAS) and the interest rate the Government limitsto farmers o f 7.5 percent. The Governmentpaysthe spread to the CNCAS. The Fonds de Garantie (FG) covers part o f the credit risk o f agricultural and livestock loans under the justification that financial institutions would not lend to rural producers as the risk premium i s too high. Delinquent loans at CNCAS are reimbursed up to 75 percent for agriculture and 50 percent for livestock. Only CNCAS so far has benefited from the FB and the FG. The Fonds de Calamite' (FC) i s to assist rural producers cope with natural calamities, allowing them to repay loans and continue their agricultural activities. The FC i s managed directly by the Ministry o f Finance and can be accessed either to restore producers to creditworthiness at CNCAS or to finance supplies necessary to respond to shocks to agricultural production, such as the purchase o f pesticides in the recent locust invasion. There are no clear rules for the use o f these funds. Complex administrative procedures delay payments (for example compensation to banana producers for the 2002 floods has not been given yet). Over the last 4 years, payments have been made in 2000/2001 for cotton producers affected by white fly, for banana producers in Tambacounda affected by flooding in 2003/204, to clear the unpaid debt o f producers for the agricultural season 2001/2002 and 2002/2003 due to globally insufficient rainfall, and to mitigate the costs o f the locust infestation in 2004/2005 (see Box 5 on the 2004 locust swarm). Box 6: ManagingRisks Associated with Desert Locust Swarms Desert locust swarms in 2004 prompteda region-wide response to cope with the current infestation as well as set up regional early warning and prevention systems to reduce the impact of future infestations. In2004, IDA approved a region-wide Locust Emergency Project (US$60 million) covering seven countries inWest Africa. The Senegal portion would finance emergency assistance and long-term institutional capacity building, for a total program amount o f US$12.4 million. Its objectives are to reduce the hardships imposed on people and the environment by current and future locust invasions. 2004 contributions to Senegal's Desert Locust Program are estimated at US$39 million: GOS (US$9 million); other donors (US$12.5 million) and FA0 Trust Fund(US$25 million). The failure of the affected countries and their partners to maintain a minimal capacity duringthe long remission phases has resulted in a weakened system unable to respond ina timely fashion to prevent the current upsurge. To address this, a permanent facility would be put into place to provide early warning and prompt preventive actions in the future, and to ensure that control operations are triggered based an a realistic assessments of risks. For control activities, the evidence suggests that more targeted control programs focusing on areas planted to high-value crops and sprayed accurately may have a muchhigher cost-benefit ratio than large-scale campaigns. Source: TechnicalAnnex, Africa EmergencyLocustProject(T7645-AFR)World Bank 2004 56 Table 27: Financing of Funds for Securityof Agricultural Credit 1998-2004, in CFA GuaranteeFund BonificationFund (FG) (FB) FC) Fund Total Financing 1998 400,000,000 300,000,000 300,000,000 1,000,000,000 1999 1,925,000,000 700,000,000 225,000,000 2,850,000,000 12000 2,100,000,000 700,000,000 300,000,000 3,100,000,000 12001 3,100,000,000 900,000,000 II 3,500,000,000 II 7,500,000,000 12002 3,000,000,000 202,500,000 - I 3,202,500,000 /2003 IIII 500,000,000IIII - II 1,556,000,000 I 2,056,000,000 1III 2004- (prelim.) 3,100,000,000 900,000,000 4,500,000,000 8,500,000,000 TOTAL 14,125,000,000 3,702,500,000 10 381,000,000 28,208,500,000 Source: Ministry o f Finance 138. From 1998-2004, 28 billion CFA (US$56 million) was allocated through these funds to mitigate and cope with shocks to rural producers. This amount is about 64 percent o f the total programmed amount o f 43.5 billion CFA that was included inthe three year public investmentprogram (PTIP) for the 3 funds. Per Table 27, over halfwas usedto finance delinquent loans, one-third for calamities, and the remainder to subsidize interest rates. 139. 2004 was the most active year to date, amounting to 10.2 billion CFA (US$20.4 million). Per Table 28, 41 Destination Amount YO I Source percent was to respond to the locust Locust Response 4 200 000 000 41% FC infestation through the FC, 34 percent Out of season rains 1 100 000 000 11% FG III went to forgive debts o f rural producers, with the remainder to forgive debts of producers affected by the out-of-season C N C A S rains, to help banana producers cope with flood-induced losses and payments to CNCAS to finance the interest rate spread. At present, there are accumulated `deficits' in both the FG and FB. Full payment-to cover guarantees applied to delinquent loans and interest rate spreads has been less than what should have been transferred, with a shortfall of 32 percent inthe FG and 12percent inFB (Statistical Annex Table 1V.e). 140. I t is difficult to assess the overall impact of the FSCR mechanisms. Givenboth the natural shocks and the changes in the commercialization systems, agriculture has been under extreme stress and the various funds have kept some measure o f credit flowing. But, they do not properly address longer-term sector creditworthiness and solvency issues. They enable the Government to respond when natural disasters occur, but work less well when it comes to helping people recover their lost activitiedassets. The FB appears to have increased the profit o f farmers that benefit from lower interest rates. In assessing the impact on a subset of rice producers, profit margins were increased by between 1 and 5 percentage points (Statistical Annex, Table 1V.f). 57 141. There has been continued erosion in the overall portfolio of the CNCAS. The total capital lent to agriculture through CNCAS has decreased from a high o f almost 14 billion C F A in2000 to about 8 billion in2003 (Statistical Annex, Table IV.g.). Delinquent loans rose from about 2.5 billion CFA in 1997-99, peaking at an average o f 5 billion CFA in 1999-2001, and easing to 3.5 billion CFA in2001-03. The average repayment rate was 70 percent which may reflect the moral hazard risk o f anticipated debt forgiveness. Other Supportto AgriculturalProducers 142. Groundnut price and commercialization risks are mitigated through a producerpricesupportfacility createdin 1997. The facility (Compte de Soutien au Prix de 1'Arachide) was put inplace through an agreement signed between the Government and the EU. The facility was established to cover (i) the negative balance resulting o f the overall subsector deficit at the end o f the marketing campaign, (ii) difference between any the agreed producer price at the beginning o f the crop season and the resulting effective world price for groundnut oil; and (iii) any covariant losses resulting in natural disasters that hinder the viability o f the subsector. Initial financing o f 5.25 billion CFA came from the national budget (CFA 2.05 billion) plus 3.2 billion CFA from STABEX. The facility replenishment i s to come from a special tax on imported crude and refinedoil and from a positive balance o f the subsector at the end o f the marketing campaign. The facility, which reached about CFA 9 billion in 2002, has been used to compensate producers who have been left with unpaid bills by intermediaries (bons impaybs) in 2001/2002, to hnd advances to SONACOS and NOVASEN to cover deficits and to guarantee credit o f CNCAS. An external audit noted a lack o f clear criteria o f application o f the facility. 143. The State intervenes to reduce the cost of agricultural inputs. This includes subsidies o f groundnut seed prices, support to the re-activation o f corn and manioc production and the acquisition o f fertilizer and equipment. Over the last ten years, the Government has managed a stock o f about 40,000 tons o f groundnut seeds per year. In 2000/2001,the Government distributed 70,000 tons o f free groundnut seeds to producers in the face o f difficulties producers had o f repaying seed credit. The State again subsidized seeds in2003/2004 to producers, equal to about 2/3 o f the commercial seed price. GeneralObservations 144. Early warning systems permit a more planned and less costly response to mitigate or cope with shocks to agriculture and rural production. Public expenditures to cope with climate shocks on rural production are very high and the instruments fairly imprecise. In 2004, more than $US26 million was mobilized to address the locust infestation, a shock which in the end affected about 20 percent of rural households. Transaction costs of mobilizing emergency external assistance are not insignificant and delays in provision o f financing or goods mean that farmers have already had to exercise informal coping mechanisms long before assistance arrives. Compared to the annual budget o f the Department o f Plant Protection o f US$1.1 million, better early warning systems and prevention would increase efficiency. 58 145. Experimentation with new mechanisms to mitigate and cope with risks has yielded some lessons and cautions for future use. Delays inproviding loans to producers well as time lag in accessing relief and compensation are major shortcomings that reduce effectiveness in implementation. Furthermore, it i s not clear the extent to which these instruments actually reach the poorer fanners. Multiplicity o f instrumentsadds complexity to the system as a whole and reduces probability that the poorest producers benefit. Moreover, the lack o f clear criteria and operating procedures makes it difficult for farmers to count on this support, opens the door to inefficient uses and is likely to penalize the less educated and least connected, i.e. the poorest. Moreover, decisions like forgiving all producer debts do not match the differential nature o f the shocks which rarely hit all producer^.^^ Limited recourse to formal credit for many rural families limits effectiveness o f this measure as it cannot be applied to informal credit arrangements. 146. Senegal is in need of a sustainable financing system for agriculture. Lack of recovery o f outstanding loans i s high, in part due to calamities (drought, insects, floods etc.). But also, a culture o f non-repayment that has gradually installed itself in the rural areas, possibly in response to political promises o f debt relief to farmers. The erosion o f producer prices and low profitability o f traditional rainfed agriculture have further exacerbated repayments. And, the guarantee scheme itselfmay weaken CNCAS incentives to recover outstanding loans. D. ManagingCritical HealthRisks Overview of Institutions,Policiesand Expenditures 147. Senegal will have difficulty reaching the health-related MDGs, particularly because of challenges in the rural areas. Infant and maternal mortality, malnutrition and malaria are more prevalent in rural areas, compounded by household factors including greater exposure to natural and agricultural shocks, lower education levels, and less access to savings and credit mechanisms to mitigate health shocks. 148. The 1998-2007 National Health Development Plan (Plan National de D&veloppement Sanitaire 1998-2007-PNDS) identifies the main policies and strategies for improving health outcomes in Senegal. A recent evaluation o f PDIS found mixed results, with increased availability o f cost-effective interventions and essential drugs at health posts, an expansion in capacity and better management o f hospitals, but insufficient access to health services for rural populations, insufficient coordination with communities and private sector, over-centralization o f resources, and high financial barriers to health services. Only one-quarter o f the new health posts planned at the beginning had been realized. As a result, the Government issued the Plan National de De'veloppement Sanitaire (PNDS): Phase Deux 2004-2008 which seeks to address these issues by consideringboth supply and demand sides as well as by improving accessibility, efficiency, quality and sustainability o f health services, especially inremote areas. 46CMS estimates that 60% of agricultural loans had already beenpaid back at the time debt cancellation. 59 149. Other government agencies also execute health expenditures. The National HIV-AIDS Council's Technical Secretariat under the Prime Minister's Office channels funding to several government agencies and NGOs for HIV/AIDS prevention and treatment. The Prime Minister's Office also manages the Nutrition Enhancement Project. Inaddition, local governments, through own resources or through fiscal transfers, finance a good deal the drugs and equipment in health posts and health centers. And, a significant number o fNGOs provide health services. Overview of Public Spendingto ManageHealthRisks 150. The public sector, includingcentralgovernment,donors and localgovernment, spent about 88 billion CFA on in the health system in 2004 (US$176 million), equivalent to 2 percent of GDP (29).47 An additional $36 million came from the community level payments for drugs and services inpublic health clinics. Donor financed investments in health amounted to 18.9 billion CFA (US$36 million) in 2004 with at least ten major sources o f donor funding. In2004, central government spent about 7 percent o f its recurrent budget on health and 3 percent o f the domestically-financed investmentbudget (Statistical Annex Tables 1V.hand 1V.i). Table 29 :Health Sector Financing,millionCFA ~ ~ 2000 2002 2 004 Central Government 29 637 38 569 46 178 Local Governments 820 4 316 4 381 Fees collected from populations 7 822 17 869 18 137 External Partners 23 232 27 826 18 902 Total 61 511 88 580 87 598 Health expenditures as a % o f GDP 1,98% 2,55% 2,13% 151. The centrallevelofthe MOHabsorbs about halfofthe sector's resources. This centralization of resources usually works to the disadvantage o f local health districts and rural areas. The administrative level currently accounts for about one-quarter o f Ministry o f Health expenditures. This has shown a slight tendency to increase over time (Statistical Annex Table IV.j). Unfortunately, it i s difficult to isolate rural sending. First, statistics on health services are kept at the health district level and not disaggregated betweenurban and rural areas. Second, rural dwellers use urban services, for example by coming to the regional hospital in the case of major illnesses. And finally, expenditure data i s mostly managed centrally by various programmatic areas or by nature o f the expenditure (e.g. drug purchases), with no geographical disaggregation o f expenditures. Extendingthe HealthService Networkin RuralAreas 152. Despite growth in the number of health centers and health posts, health services have had difficulty in keepingpace with population growth in recent years. 47 This does not include healthexpenditures outside of the Ministryof Health, including an average of USS$25 millionper year (2003-2004 on multisectoral HIV-AIDS and nutrition interventions (para 152-57) 60 For 1998-2002, there was a expansion in primary health care infrastructure, including an increase in health posts from 753 to 848, rural matemities from 570 to 643 and health huts from 1,087 to 1,519 (Table 30). At the district level (Le. not counting hospitals), there was a 50 percent increase in doctors, 9 percent increase inmidwives, 12 percent for nurses and 30 percent for community health workers. But, health posts and health centers saw a decrease in coverage in terms o f infrastructure per inhabitant when population growth i s factored in. Nurses and health agents have not increased their presence per population. Table 30 : Health infrastructure and personnel1998-2002 Indicator 1998 1999 2000 2001 2002 # of functional health posts 753 762 785 817 848 # of rural maternities 570 I 581 I 599 619 643 # of cases de sante' - health huts 1,087 I 1,096 I 1,217 1,483 1,519 Source: CEFOREP 153. Between 1998-2002, districts that are mostly rural benefited from additional infrastructure and staff. Table 30 shows net change for mainly urban (districts with regional capitals) versus mainly rural districts (Statistical Annex Table 1V.k by district). Rural areas benefitedfrom more infrastructure, doctors and nurses; however, recruitment o f midwives was more likely to benefit urban areas. The biggest expansion occurred at the community-financed level o f health huts and community health workers. Despite this progress, poorer regions still have less infrastructure and personnel. The Dakar region, with 17 percent of the nation's poor, accounts for 35 percent o f public health personnel. Fatick, Louga, Tambacounda and Saint Louis have equal share o f the poor and health personnel and Diourbel, Kaolack, and Thies have less personnel than their share o f the poor. Moreover, as of 2002, most midwives were still inDakar (228 out o f 496 nationally). #health #rural # health #doctors #mid- #nurses #ICP # comm. Health districts posts materni. huts wives & agents personel Mostly Urban +37 +17 +70 +15 +I2 +2 1 +40 +353 Mostly Rural +46 +30 +266 +19 +9 +46 +63 +309 61 154. Rural residents perceive problems with quality of health service. For both urban and rural residents, cost was the biggest reason for dissatisfaction, followed by long waiting times (Table 32). Rural residents were slightly more sensitive to cost and urban residents to waiting times. Rural residents were more likely to cite high cost and ineffective treatment. Interms o f out-of-pocket expenses, rural households spend a larger portion o f their income on health (2.1 percent o f total consumption versus 1.6 for urban). Table 32 : Dissatisfactionwith health consultations and reasons Source : ESAMI1 155. In terms of specific services, the period 1996-2000 saw significant erosion in vaccination coverage in rural areas, but this appears to have rebounded in recent years, In 2000, rural children 12-23 months o f age were about half as likely to have complete vaccinations before the age o f one versus their urban counterparts (Statistical Annex Table IV.l). Administrative reports indicate that the national vaccination rates have improved inthe last few years.48 The Ministry o f Health reports that coverage o f DTC3 as of September 2002 was 65 percent versus 30 percent in 2001. Investments in cold chain equipment and other outreach materials enabled this improvement. Administrative data for 2004 report that out o f 57 health districts, nine had difficulty achieving the 85 percent goal for DTC3, six o fwhich were rural districts. 156. Regarding health services to prevent maternal mortality, about 71 percent of rural women report going to at least one pre-natal control compared with 88 percent in urban areas (MICS 2000). When it comes to the birthitself, one-third o frural women give birth without any assistance (compared with 3 percent in urban areas) and only 43 percent are attended by medical or para-medical personnel. Inpart this i s due to the limited availability o f health services and dispersed population, with bad roads, less forms of transportation and limited outreach ability by rural health posts. Preferences and ability to pay also influence this outcome. Malaria 157. The Government of Senegal has mobilized additional external resources specifically to address malaria. A Strategic Plan 2001-2005 was developed which estimated a cost of 26.2 billion CFA (US$52 million) to meet the national malaria objectives o f access to effective treatment within 24 hours and coverage o f insecticide treated bed nets (ITBN) to 60 percent o f pregnant women and children under 5, or about $US 10 million per year (Statistical Annex Table IV.m).49 The main elements o f the 48A more definitive view will be available from the Demographic and Health Survey IV currentlybeing carried out and when the HealthInformation System is fully function, expected in2005. 49ITBNshave proveneffective insignificantly reducing malaria transmission. Infield researchin southeastern Senegal (hyperendemic zone) bednets were distributedto the whole of the population o f 62 national strategy include coverage o f ITBNs, utilization o f bi-therapy (amodiaquine plus SP) as a first treatment instead o f chloroquine, prevention and treatment o f malaria during pregnancy, and control of malaria epidemics (in north) through sentinel sites. Almost half o f the Strategic Plan's estimated cost i s for insecticide treated bed nets, with a target o f 3 million to be distributed. A 2003 Roll Back Malaria mission estimated a financing gap o f 14 billion CFA, 8 billion CFA o f which was the shortfall in financing for ITBNs. 158. The Government'sstrategy for distributionof bed nets providesfor subsidized pricesto the end user distributedeither through the public health system or through NGOs and the privatesector throughsocial marketing. As o f 2000, coverage o f IBNs was still quite low, particularly inrural areas. Only 13 percent o f rural children 0-59 slept under a bed net (1 percent ITBN) versus 23 percent o f urban children (3 percent ITBN) (MICS 2000). Some pilot experiences have recently documented coverage rates o f over 40 percent incommunities inTambacounda, Kolda, Kkdougou and VClingara. 159. Despite delays in mobilizing resources and procuring bed nets, the Table 33: Distributionof ITBNs,by region bigger constraint appears to be on Pop per net developing a delivery system that Region # distributed distributed provides for rapid scale-up. The Government has purchased roughly 600,000 ITBNs to date (200,000 with IDA financing, 100,000 with Japanese financing and 300,000 financed by the national budget), or about 20 percent ofthe strategic plan goal. For the Japanese and IDA- financed nets, the MOH Malaria program reports distribution o f only 184,000 to date (Table 33). Distribution was fairly widely spread among regions, with the highest concentrations per capita in the Tambacounda, Zinguinchor and Saint Louis. Dakar received comparatively fewer, befitting o f its lower risk profile. N o information is collected on distribution to rural areas. The MOH has recently reduced the price to 1,000 CFA for pregnant women sold through health centers to increase uptake. Even at this price, cost will remain a barrier inrural areas and among the poor. 160. The shift from home treatment with chloroquine to bi-therapy requires managementby communityhealth system, as opposed to self-care. This may be more difficult in rural areas. Each community will need to have a community health worker trained inbednets and treatment. Villages without a community-managed health hut (case de santk) will need to find an individual to be trained as focal point. The program to prevent and treat pregnant women with traitement prkventif intermittent (TPI) A la Wassadou in 1995. Compared to baseline data taken in 1992-93, bednets led to a sharp decrease inthe density o f the vector population and malaria transmission. The number ofbites by infected mosquitoes decreased by 69 percent and their density inside dwellings by over 90 percent. The daily rate o f entomological inoculation decreased by 88 percent (Faye et a1 1998). 63 sulfadoxine-pyrimkthamine (SP) is to be integrated into pre-natal services. However, initial experience with the change in treatment protocols supported by community health works through major NGOs operating health services suggest that the community model may be effective and that local level actors are better able to determine who can and who cannot pay and devise repayment schedules according to household ability to pay, both for treatment and prevention. 161. In the North, early warning systems can reduce the impact of malaria. The North is exposed to epidemics with an explosion o f cases in one point in time, whereas in the South malaria is endemic throughout year with lighter peaks. This creates in the North the possibility of a malaria early warning system. A predictive factors study analyzed historical data from the Senegal RiverValley to identify epidemic thresholds bywater level in the river and also out of season rains. An early warning system would focus on three main groups o f predictors: vulnerability risk indicators (drug resistance, HIV, malnutrition), transmission risk indicators (e.g. increases in rainfall, river flow or temperatures) and early case detection (e.g. malaria morbidity data using critical epidemiological thresholds (Thomas and Connor 2001). The French Cooperation i s supporting the establishment o f sentinel sites to provide advance warning o f epidemics. HIV-AIDS 162. Senegal has been successful in mobilizing financing to fund its 2002-2006 Strategic Plan for the Fight Against HIV-AIDS. The Strategic Plan seeks to hold HIV prevalence below 3 percent, extend treatment through free anti-retrovirals, provide psychosocial support to AIDS patients and their families, increase public awareness and behavioral change through prevention outreach, increase voluntary testing and reduce discrimination and stigmatization. In 2002-2004, 19 billion CFA (US$38 million) was provided by Government and donor agencies (Statistical Annex Table IV.n). The Ministry of Finance estimates that 13.6 billion CFA (US$27 million)) was spent in2004, all sources combined, a sharp increase from 2003 as disbursements from donors increased (Table 34). Table 34 : Financingofthe Strategic PlanAgainst HIV-AIDS 2003-2004 External Government Total 2003 2,362,944,920 -- 2,362,944,920 2004 13,351,535,000 200,000,000 1 3 31,535,000 163. The approach of the CNLS is to mobilize both public sector agencies and civil society to scale-up prevention and treatment across the country. The multisectoral response incorporates action plans in six areadministries: Health, Education, Social Development, Family and National Solidarity, Youth, Security Forces (Army, Customs, Fire Protection, Police Prisons). For the IDA-financed portion, regional allocations have been loosely correlated with risk, with high prevalence zones like Ziguinchor and Matam receiving the highest per capita amount, though spending in Kolda was below the average (Statistical Annex Table 1V.o). 64 164. The civilsociety responseremainsmost activeinDakar andurbanareas. Data on CNLS fundingto civil society show that o f the 573 projects that entered execution inthe first round o f selection (executed mainly in 2004), 54 percent were oriented to urban areas, 27 percent to both urban and rural and 19 to rural areas. CNLS supported about 7,000 `interventions' (training and outreach sessions, counseling and testing services, psycho- social support, etc.), 80 percent of which were in urban areas (Table 35 and by region in Statistical Annex Table IV.p). Few activities in rural areas focused on promoting testing and treatment as well as little outreach in psychosocial support. The strongest rural response was ingeneral awareness raising through community-based organizations. Table 35 : Interventionssupportedby CNLS2004, by Activity and Urban-Rural Region Prevention Prev. Mother- Testing Care Treat- Psycho- Care Total sexual Blood Child for ment social Orphans transmission Trans. Trans. STD: support Total Urban 4945 76 85 268 30 50 132 29 5615 Total Rural 1340 28 23 20 0 0 19 6 1436 165. There are few HIV testingservices outsideof the main cities. As o f end 20004, only Thies (7 out o f 8 districts), Ziuginchor (3 out o f 3 districts) and Diourbel (Bambey) offered public HIV testing at the district level. However, promoting testing in rural areas can boost the productivity o f urban testing sites, which up until now have low utilization rates. One assessment found that existing CDVs in Kaolack, Thies and Ziguinchor had served 1.8, 1.7 and 3.3 percent o f the adult population, respectively (Woynar 2004). The higher uptake in Ziguinchor is due partly to a 50 percent increase in testing rates after a rural outreach campaign by the N G O ASVIE, including ambulatory testing services. 166. Care for rural populationslivingwith HIV-AIDS poses particular challenges. Through the ISAARV (Initiative Sknkgalaise d`Acc2s aux Antiretroviraux) program, free ARV treatment has been extended to all regional hospitals as o f end 2004, with almost 2,000 patients under treatment. However, as shown inthe statistics o f health care coverage, rural people rarely use hospitals. Lower disposable income reduces financial access which, although the ARVs are free o f charge, includes expensive diagnostic and monitoring test and treatment o f opportunistic infections. In addition, rural populations must come to regional hospitals, resulting in transport fees and lost days o f work. In 2004, a pilot experience was carried out in Ziguinchor to decentralize treatment to 3 health centers (Ziguinchor, Bignona and Oussouye). As o f late 2004, the decentralized health centers accounted for 60 percent o f the patients under treatment inthat region (Woynar 2004). TargetedSupportto Address Malnutrition 167. The Nutritional Enhancement Program under the Prime Minister's Office is the country's mainprogramto reducemalnutrition(Table 33). Senegal has recognized the benefits o f improving the nutritional status o f the population and is investing an important amount o f resources into community nutrition programs. Initiated in 2002, the Nutrition Enhancement Program (Projet de Renforcement de la Nutrition - PRN) i s a 10 year program to improve the nutritional status o f children under three and pregnant and 65 lactating women. The program contracts out to NGOs that to deliver a nutrition package, including cost-effective interventions such as child growth promotion at the community- level, vitamin A distribution, breastfeeding promotion among others. The PRN complements general health programs to prevent and treat malnutrition and supports equipping and training o f healthposts. The project has also begun to give small grants for income generation. The project, currently active in 60 percent o f health districts, reaches an average o f 165,000 children under three and their mothers every month, including 98,520 children in urban areas (25 percent o f urban under 3 year olds) and 67,230 in rural areas (12 percent o f rural under threes). The median annual cost per beneficiary o f the community-level services i s $4 per year, with full program cost o f $17. From June to December 2004, the incidence o f underweight children in the project areas declined fi-om 23 percent to 18 percent. Table 36 : Financingof the NutritionEnhancementProject,2003-2004 External Government Total CFA Total US$ 2003 2,112,73 1,374 176,000,000 2,288,731,374 $4.6 million 2004 5,473,904,564 176,000,000 5,649,904,564 $11.3 million HealthInsurance 168. Government supports the creation of health insurance mechanisms. While much o f the direct investments inestablishing health mutuals are channeled from donors to NGOs and local federations, the Ministry of Health plays a critical role in developing the regulatory framework and inpreparing the necessary instruments and capacities within the pubic health system to allow it to interface with these insurance mechanisms. There i s a specific department o f health mutuals within the Ministry o f Health to oversee development o f these services. In 2003 a Law on Mutualism to regulate the operations o f health insurance mutuals was passed, but no application decree has yet been emitted by the MOH. In 2004, the Government provided 200,000,000 CFA in the investment budget to MEFPOP support the creation o f a health mutual for public sector workers to smooth income shocks of health crisis. Although civil servants benefit from coverage o f health services, required co-payments as well as the price o f prescription medicines represent significant private costs. GeneralObservations 169. There has been some success in mobilizing funds for critical health risks in malaria,HIV-AIDS and malnutrition,but the populationmost at risk remains under- served, Ifinterventions were targeted first to those groups at greatest health risk, the map of interventions would look different. Malaria and maternal health programs would be overwhelmingly rural, including a greater concentration of midwives and preferential distribution o f bed nets to rural areas. Malnutrition interventions would be more prevalent inruralthanurban areas, whereas the opposite isnowthe case. 170. The bulk of critical health services, like pre-natal control, vaccinations, malariaandHIV preventionare carriedout at the communitylevel andin rural areas 66 the role of the community health agents is key. Inthe rainy season when malaria i s at its peak, villagers are isolated from health posts, making the role o f community health workers all the more important. For bed net distribution, expanding beyond the health system appears to hold the most promise for scaling up use, working through the extensive network o f community based organizations in rural areas instead o f relying on health comrnittee~.~~Efforts to address malnutrition rely on the presence o f NGOs with technical capacity, which i s weaker in rural areas. All o f these programs suffer from the general centralization o fresources, personnel and decision-making within the MOH. 171. Rural health districts vary greatly by the level and coverage of services, cultural factors, economic base, access to remittances to finance health fees and frequency of health risks (HIV prevalence, malaria morbidity etc.). For rural populations, focus should be on developing appropriate materials and outreach strategies (for example via rural radio or using the widespread women's groups or village development associations) to increase their general understanding o f health risks and prevention strategies. Programs need to be tailored to the specificity o f rural populations, like targeting menwho migrate for work (and their wives) for HIV messages. 172. Financial accessibility is a critical complicating factor. The rural health system i s largely based on community payment for services. But, almost half o f the rural poor who are sick say they did not seek treatment because they could not afford it. Efforts to increase the use of qualified personnel at births, access to life-saving HIV-AIDS treatments, timely recourse to anti-malaria drugs in the case o f onset, and other critical actions to prevent and cope with health shocks will be contingent on affordability o f services. The extension o f health mutuals to provide some level o f health insurance appears promising for the better-off rural populations. Other policies currently being piloted, like free deliveries for pregnant women inthe poorest regions should be monitored for their potential to benefit rural women. The basic policy o f charging for bednets should bereconsidered for the ruralpoor. E. Managing CriticalEducation Risks Overview of Institutions, Policies and Expenditures 173. The Ministry of Education is the main sector agency responsible for promoting educational attainment in Senegal. As presented inChapter 2, the challenge o f achieving basic education for all i s greatest inrural areas. A significant share o f rural children are not currently in school and for every 100 children in rural areas who enroll only 58 will complete the primary cycle. Barriers in terms o f cost, distance and perceived relevance combine with household factors such as health and agricultural shocks and the use o f child labor create a set of conditions that must be address if Senegal i s to reach its objectives. "The MOHrecently startedusing OCBs as agents for ITBNdistributionin 15 districts. OCB sells nets for 1,000 CFA (200 stays with OCB and 800 to districts. 67 174. General educationpolicy was framed in the Ten Year Education and Training Programme (Programme De'cennal de 1'Education et de la Formation, PDEF) 2000- 2010, and updatedin a 2005 Letter of SectoralPolicy. The policy identifies strategies to reach the MDGs in terms o f gender equity and complete coverage o f primary education, including (i) priority to the rural zones for construction o f classrooms; (ii) development o f bilingual education and introduction o f Arabic into certain zones where populations perceive this as fundamental to educational value; (iii) promotion of access for all including disabled, low-income and other children at risk, (iv) use o f double shifting and multigrade classrooms; (v) recruitment o f up to 3,000 contractual teachers and redeployment o f teachers to classroom purposes, (vi) reduction in repetition and improvements in internal efficiency; (vii) better involvement o f communities in school management; (viii) respect for the norm o f 900 teaching hours per year over 30 teaching weeks; and (ix) attainment o f a ratio of five textbooks per student inprimary. 175. In 2004, the Government and donors spent about U S 5 2 million on rural primary education (Table 37). The central government budget accounted for about 78 percent o f education expenditures in 2004. Households through contributions via parent- student associations account for 15 percent, donors for 6 percent and local governments 2 percent. This amounts to 6.2 percent o f GDP (including private education) and has risen over time. Education accounts for one-third o f the national recurrent budget and primary education claims 42 percent o f education expenditures (Statistical Annex Tables 1V.p and 1V.q). MinistryDZxpenditure Area/program 2004 Budget (CFA) Ministry of Education Personnel and Non-wage Recurrent (rural teachers and textbooks) 19,000,000,000 School Lunches: (rural) WFP : 1,246,000,000 : State and other partners (rural andurban*) 413,000,000 Investment (BCI) Rural classroom construction 5,442,720,000 Total CFA 26,101,720,000 Total US$ $52,203,440 Percent of 2004 EducationExpenditures 12% 176. Resources are inequitably distributed between rural and urban areas. Resources are heavily tilted towards urban primary students (Table 38). The cost per student in rural areas is about 28,000 CFA per year, versus 47,000 CFA spent by the government per urban student. This cuts against the general impression that it more expensive for the government to invest in rural education because o f low population density and smaller class sizes. On the contrary, for every additional urban student, the state could add 1.7 new rural students under the prevailing cost structure. The differences are largely due to lower cost teachers inrural areas as well as possible overstaffing in some urban schools. In addition, the significant overhead and administrative expenses in the systemact to the detriment o f expenditures at the service delivery level. 68 Table38 :Cost per Student of PrimaryEducation,ruralandurban TotalExpenditures YO CFA Per Student (public) Totalexpenditures 82,354,000,000 100% 67,046 Ruralexpenditures 19,223,000,000 23% 27,870 Urbanexpenditures 25,188,000,000 31% 46,764 Administration andother 38,865,000,000 47% School Constructionand Openingof New Classrooms 177. The Governmenthas made significantefforts to increasethe supply of primary education in rural areas. Between 200-2004, the Government built 7,109 new primary school classrooms and rehabilitated 930 (Statistical Annex Table 1V.n). O f these, domestic resources financed 53 percent, and donors working through the MOE an additional 42 percent, with the remaining 5 percent financed through donor-financed multi-sectoral projects working directly through local governments and communities. Despite this investment, incomplete schools are still more frequent in rural than urban areas. In rural areas, three-quarters o f primary schools do not have all o f the requisite grades, increasing rural children's risk o f not finishingprimary education. Table 39 : Openingof New ClassroomsinRuralAreas, by Sponsor andPovertyLevel 300 55.3 68.9 75.0 62.7 60.6 350 67.6 79.2 83.8 74.0 74.1 400 79.6 86.6 91.5 82.8 83.4 450 92.5 89.8 94.4 87.6 87.7 500 100.0 100.0 100.0 100.0 100.0 178. The bulk of the new classrooms opened were in rural areas, with strong support from parentsassociations. Interms o f opening o f new classes, which includes adding a teacher and may or may not require construction, in the 2003/2004 school year 81 percent were in rural areas and 19 percent in urban areas, demonstrating the strong commitment to expand access to primary education in rural areas. From 2001-03, 43 percent of newly opened classrooms were sponsored by parent's associations, even though these were often in provisional locations. Community mobilization to start a school or expand a classroom appears to have played a central role in reaching poorer rural communities, with the State following after. Table 39 shows the distribution o f classroom openings in rural areas by sponsoring organization and poverty level o f the 69 rural community. Classrooms supported by parents associations tend to be more heavily concentrated inthe poorer rural communities, with the 12 percent o f the rural population living inthe most under-served communities accounting for 27 percent of the classrooms sponsored by parents associations, versus about 20 percent by the government, other donors and NGOs. Provisionof Teachers,Textbooksand School Lunches 179. Despite progress on building and opening classrooms, the MOE continues to have difficulty retaining rural teachers. In 2004, Senegal hired 2,800 contractual teachers, but the total number o f teachers in the classroom only increased by 1,407. Teachers leave classrooms for other non-teaching functions within the education system, resulting in the lower net gain innumber o f teachers inthe classroom. It is not uncommon for recruitment o f contractual teachers in rural areas to replace civil service teachers who then transfer into urban areas or the administration. Teacher transfers have exacerbated differences between better-off and worse-off zones, and have led to the closing o f classrooms, particularly in difficult rural zones (or `freezing' until the next academic year). Transfers in 2004 favored Dakar, with a net increase o f 230, while Tambacounda lost 74 teachers, Louga 66 and Diourbel 47 (Statistical Annex Table 1V.s). Moreover, teacher absence is the most frequently cited reason by children for missing school, accounting for 77 percent o f responses in rural areas (PASEC dataset). The main reasons for teachers' absenteeism were problems with health, need to travel to get their salary and to attend family ceremonies. 180. Rural areas are also less equippedwith basic textbooks, Despite a national goal o f 5 textbooks per student, rural students average just 1.8 versus 2.4 for urban students (Statistical Annex Table IV.6). This lack o f materials i s reflected in household survey results where lack o f books and supplies was by far the most likely complaint about problems at school for rural children o f primary school age (Table 40). This problem was more pronounced among children from poorer families. Teacher availability and poor school condition were also noted more often among poorer rural households. Table 40 :Rural:Problems at school, just for those currently attending(age 7 to 12 only), There is no informationfor the kids who never attended schools or for drop-outs Source: ESAM-I1 181. The Government of Senegal supports school lunches to attract students and improvelearning. Twenty-eightpercent o f all primary schools offer school lunches. The situation i s slightly better for rural schools at 34 percent. O f the total o f 1,589 schools with school lunches, WFP support covers over 1,000 schools and 120,000 students inrural areas 70 o f Tambacounda, Fatick and Kaolack. Criteria for WFP support include: location ina high vulnerability zone according to WFP's map o f vulnerability to food insecurity, minimum number o f students, must be a school where children walk long distances and parents are willing to pay contributions in cash (300 CFA per student per month) andor in-kind (donation of fire wood, labor for cooking etc.). Two meals a day are provided. WFP estimates the average cost per student per year o f their school lunchprogram to be $24, o f which WFP contributes 88 percent and communities 13 percent. The WFP rations are more generous and more frequent than school lunches financed by the MOE. An expansion in school lunches to poor zones was anticipated in the priority program o f the PRSP, amounting to 600 million CFA, but to date has not received financing. 182. School lunches appear to have a positiveeffect on enrollment and educational attainment. A 2004 impact evaluationof the WFP school lunch program inSenegal found that the number of children enrolled in schools that received school lunches grew by an annual 12 percent versus 8 percent for without school lunches, with slightly higher effect on girls. Retention during the first 3 years i s 90 percent in assisted schools versus 71 percent in non-assisted. This difference falls to 57 versus 52 percent by sixth grade, undoubtedly influenced by existence o f incomplete schools. Schools participating in school lunch program also report slightly lower repetition rates and higher rates o f passing standardized exams (Akakpo 2004). 183. School lunches are distributed unequally in rural areas and inequitably betweenpoor andnon-poorhouseholds. Within rural areas, coverage ranges from a high o f 88 percent in Tambacounda and 54 percent in Kaolack to only 2, 3 and 10 percent in Matam, Louga and Diourbel respectively (Statistical Annex Table IV.u).'l The latter are also the regions with the largest number o f rural children not currently attending primary school. Within regions it serves, the WFP positively discriminates towards the poorer rural communities (Statistical Annex Figure 1V.a). But, within schools, students from the poorest families are less likely to receive meals, which may indicate exclusionary effects o f contributions (Table 41). Q1 Q2 Q3 44 QS Male Female Share of rural students 8% 26% 31% 32% 39% 26% 29% Demand-SideFactorsAffectingRuralPrimaryEducation 184. For rural populations, the long-term benefits of sending their children to school may notbe too apparent. Lack o fperspectives for continuity (incomplete schools, few secondary schools) gives rural residents little opportunity to see success through schooling. There i s the tendency to seek success through commercial trading or emigration Matam, Louga nd Diorbel fall outside o f the most vulnerable zone o n WFP's Vulnerability Map because many households have access to remittances to help smooth consumption intimes o f food insecurity. 71 abroad, both of which are done through family connections. The zones of highest emigration (Louga, Diourbel, Kaolack) also have the lowest enrollment rates. 185. The out-of-pocket costs of schooling affect parents' decisions on whether to send their child to school. Education fees reported by households in 2003 were 1,275 CFA per child in rural areas, with an additional 4,649 CFA for materials. Overall, rural households spend about 1 percent o f annual consumption on education. Schools make some exceptions based on hardship, with 43 percent o f primary schools reporting parents payingless based on revenueand 21 percent sayingreducedfees were possible for orphans. There i s no formal policy to provide support to those that cannot pay. 186. The Groundnut Basin, with strong demand for religious or arab-islamic schooling, is a challenge to the model proposed by the State, with its basis in French. One reason for the low performance o f primary enrollment o f rural children in Diourbel is from the daara and parental resistance to secular schooling. Insuch zones, the Ministry o f Education is diversifying its educational offerings through the establishment o f Franco- Arab schools where bi-lingualteachers follow the official curriculum. Observations 187. Among rural regions, there are large variations in enrollment in first grade and retention through primary cycle. The most critical region i s Diourbel with the lowest enrollment rate for first grade and lowest retention rates. While Tambacounda has a relatively high enrollment rate for first grade (gross rate about loo%,) only a little more than half of those complete six years. Other regions like Ziguinchor and Kolda out- perform the national averages. 188. Despite recent progress, infrastructure expansion and redoubling efforts to post teachers in rural areas will be needed. Population growth calls for continued infrastructure investments. Moreover, expanding access to secondary school, for example through locating them in areas more accessible by rural residents or exploring support for students to come infrom the countryside, may increase perceivedrelevance o f education, 189. Demand-side factors are clearly an issue, but there is less policy and program development in this area. Rural children do not attend or they drop out o f school because they suffer from health shocks, are affected by the agricultural economy, do not have the economic resources, or their parents do not see it as a worthwhile investment. Policies and programs will need to be developed to address these demand-side issues. F. Employment Generation Overview of Institutions, Policies and Expenditures 190. Opportunities to diversify income through off-farm employment are an important risk management tool. Creating alternative sources o f income and employment is a means o f escaping from low productivity agriculture. Given agricultural 72 cycles, the need for cash income in certain periods o f the year means that even if a household's base i s agricultural, supplemental income and employment outside o f agriculture is sought duringcertain times o f the year. 191. The most recent National Employment Policy (1997) has six strategic axes: (i) to maximize creation o fjobs in the modern sector; (ii) support the modernization o f the to informal sector through productivity increases, better access to social protection, technological innovations, and increased access to credit through micro-finance; (iii) to support initiatives for local employment including rural employment opportunities; (iv) to adapt training programs to labor market needs; (v) to better integrate employment issues into national policies, including favoring high labor-intensity in technical choices and developing national social protection systems to extend coverage o f health and unemployment insurance and retirement benefits; and (vi) to better monitor performance o f programs to support employment creation, and the labor market ingeneral. 192. Given its transversal nature, many ministries and agencies are involved in activities to generate employment. Inaddition to the National Employment Action Fund under the labor ministry (Ministire de 1'Emploi, Fonction Publique et Organisations Professionnelles - MEFPOP), the majority are carried out through programs that target specific groups, for example youth or women. However, there are many programs that in the end have similar objectives, such as the income generating activities supported through poverty reduction programs or credit lines to women as part o f social assistance programs through the MFFSSD.The division between employment generation projects and those that seek to raise incomes is not clear. 193. Programs that specifically seek to create jobs amounted to about US$4 million in 2004 (Table 42). These are all fblly domestically-financed and most have been created within the last few years as the pressure to address employment, particularly among youth, have increased and become an explosive political and social issue. However, this i s a fraction o f the funding that i s channeled through poverty reduction programs that finance income generating projects. Table 42 :EstimatedAnnual Expenditures on Selected Employment CreationPrograms Source: DDI 73 EmploymentPrograms 194. The National Employment Action Fund is the main direct support for employment creation by the MFPEOP. Created in 2000 to reduce unemployment and under-employment, the FNAE provides financing to the Convention State-Employer (for on-the-job training) and channels credit through two MFIs to finance business start-ups. In 2003 (most recent data), 358 credits were granted for a total o f 305,184,967, or an average o f 850,000 CFA per credit. The MFIs receive 8 percent in interest and 10 percent insupport for non-financial services. The program estimates creation o f 11,000 permanent jobs to date (cost of 27,744 CFA per job).52 Few serve rural populations, but there have beenprojects for agricultural processing, likemilk distribution. The MinistryofYouthhas three programsto support youth employment: The National Youth Emplovment Agency (Aaence National pour 1'Emploi des Jeunes ANEJ) created in2001 to assist youth inobtaining jobs through training and advice as well as links to potential employers. ANEJ estimates it reaches 50,000 mostly urban youth. With support from Taiwan, ANEJ i s establishing 34 Tool Houses (Maison d 'Outils)to help artisans, mainly from villages, beginproduction. The National Fund for Youth Promotion (Fond National pour la Promotion des Jeunes FNPJ). Also created in2001, FNPJ finances lines of credit through MFIs to support young entrepreneurs. 6.2 billion CFA inloans has financed 2,378 projects, creating an estimated 12,000 jobs.53 Rural areas have benefited from 33 percent o f the financing and 40 percent o f the employment created. One-third o f funding supports agriculture activities, 28 percent commerce, 21 percent services, and 12 percent manufacturing. Repayment rate on these credits currently averages only 24 percent, largely due to the impression by early recipients that these were politically motivated. FNPJ has tightened selection procedures and i s actively working with MFIs to improve performance. The FNPJand ANEJ have ajoint venture to finance Youth in Agricultural Farms, to test a private sector approach, working with 100 farms, 10 youth per farm and a financing o f 10,000,000 CFA per farm. Voluntary National Civil Service program: Initiated in 1999, youth are recruitedfor a two-year commitment to assist communities to undertake development efforts. Volunteers are paid a stipend o f 30,000 CFA per month. In 2003, 192 volunteers were recruited to reinforce capacity o f health structures, 300 for the program o f reconstruction o f the Casamance and 100to prevent floods inMatam. Temporary IncomeandEmploymentGenerationthrough PublicWorks / 196. While employment policies support the concept of labor-intensive public works, there is very little strategic attention to this issue as part of public works investments. Few explicit public works programs appear to be operating in rural areas.54 52This calculation appears optimistic as it assumes every credit created 31jobs. 53Based on these figures, each project (credit) generates about 5 jobs at a cost o f 502,000 CFA per job. 54For example, inthe PASEC survey, when villagers were asked if there had been public works programs inthe village that employedpeople 12.5% answeredyes and 87.5% answered no. 74 There are no overall guidelines for labor-intensive approaches to be used inpublic works, Infrastructure investments in rural areas are not planned taking into account agricultural cycles in order to inject income through temporary employment in moments o f critical income need. There is little monitoring of the jobs created through public works nor consistent methodology usedbetween the few programs that do reportjob creation. 197. In general, temporary employment is seen as a by-product of infrastructure investments, not as an end in itself. As its name denotes- Agence d'Exkcution des Travaux d 'IntkrEt Public contre le Sous-Emploi- AGETIP was created as an agency with its core mandate to create jobs through the efficient contracting o f public works. Over time, its focus has evolved toward managing construction contracts to overcome capacity constraints, particularly as a conduit for donor financing.55 In the period 200-2004, AGETIP realized 73,790,935,447 CFA in investments in classrooms, health centers, administrative buildings, canals and roads (Statistical Annex Table 1V.v). Overall, AGETIP estimates a labor intensity of 25 percent, which is low for public works programs with a labor objective but average for the type of civil works constructed. AGETIP created 16,000 person-years o f employment from 2001-2004 at an average cost o f 4.6 million CFA per person year (about US$9,000). Rural-urban distribution is not available, though an estimate o f no more than 25 percent in rural areas would put the annual rural employment at 1,000 person-years, or about 1percent o f the male rural labor force ages 20-50.56 Observations 198. The critical employment risks of rural populations are not yet well addressed in public programs, despite the existence of clear policy directives to promote rural employment. Most efforts are appropriately oriented to youth, but delivery mechanisms that rely on credit exclude the most vulnerable. Rural employment training and credit focuses mainly on agriculture, but there are some examples, like diversification into artisan production, to decrease vulnerability on agricultural production. At present, public works are not really used as an extension o f employment policies. For rural populations, timing infrastructure investments to the agricultural cycle would inject much needed cash income into the rural economy at critical moments and reduce the need for seasonmigrations to the cities to look for work. Overall, there i s little monitoring and evaluation and inconsistent methodologies for measuring employment impacts. Geographical targeting systems to ensure needy areas benefit are the exception (see following section). 199. Other proposalsto support the extension of social protection to rural workers are beingcontemplated. The Orientation Law on Agriculture calls for the development o f social protection systems for agricultural workers in order to improve employment conditions o f rural producers. The modalities have yet to be developed, but the general s5AGETIP reports 2004 sources o f financing as 62 percent IDA (Pac, PDIS, PDEQ), 8 percentlocal government, 8 percent Government o f Senegal (counterpart funding o f donor projects), 7 percent KfW (PROCR) 4 percent BAD (PLCP, PDIS,) 4 percent France (PAC), 3 percent Canada (PLP), and other small contracts with PAM, EU, Norway etc. s6I t is worth noting that even when the infrastructure itself is located inrural areas, the contractors hiredby AGETIP usually come from Dakar, or to a lesser extent regional capitals. This reduces the potential income multipliers o fpublic works inthe rural areas. 75 intention i s to work through cooperatives and producer organizations to develop insurance mechanisms for risks tied to the person (heath, death, disability). G. TargetingofInfrastructureandService ProgramsinRuralAreas PolicyandInstitutionalFramework 200. Historically,the most vulnerable rural villages have not benefited from much public investment. For example, in one study o f 250 rural villages carried out in 2002 (prior to implementation o f many o f the current decentralized infrastructure programs), it was found that counting back as far as 1982, 40 percent o f these villages had never benefited from an outside intervention (government or non-governmental) of any type (de Janvry 2003). Small size, remoteness and difficult access exacerbate their vulnerability. The most vulnerable rural communities tend to have the least stock o f basic infrastructure and tend to experience multiple negative shocks, making it difficult to accumulate assets and move out o fpoverty. 201. The Government is the main provider of basic infrastructure and services to rural areas and uses a wide array of agencies and institutionalmechanisms. To reach the MDGs, the Government will need to increase investments in rural communities, particularly water supply, education, health, roads, markets and other economic infrastructure. The relative responsibilities between central and decentralized levels as well as between sectoral and multisectoral programs often overlap. In theory, the 1995 decentralization law transferred responsibilities for most basic infrastructure to local governments (communes in cities and communuutt5 rurules inrural areas). Inpractice, the bulk o f all infrastructure investments continue to be carried out by the central sectoral ministries. This is due to the failure to fully implement fiscal decentralization, a desire by the central level to retain control over resources and the perception by sectoral agencies that local governments lack capacity. Over the last several years, a growing number o f programs have experimented with delegated contract management, decentralized and community-based implementation mechanisms in order to reinforce local implementation capacity. In addition, departing from top-down sectoral planning, there has been much effort in developing local planning processes. However, linkages remain weak between these local investment plans and investment programmedby the central sectoral ministries. 202. A general accounting of the main multi-sectoral programs estimates over US$25 million per year mainly to rural infrastructure, not including the sectoral programs in transport, education, health and water. Most o f the decentralization and community development multi-sectoral programs combine infrastructure with training and insome casesmicro-credit andincome generating activities (StatisticalAnnex Table N.w.) They are carried out under a wide range of ministries andwith different geographical focus and operating procedures. All finance small-scale infrastructure intended to meet critical community needs in poverty reduction, rather than the more large-scale regional infrastructure, like the main road system or major irrigation systems. Local governments manage resources in some o f the programs, community and village committees in others, and delegated contract management through AGETIP in others. All are based on donor 76 funding run out o f special-purpose project implementation units with project-specific manuals and procedures. Figure9 : Poverty Targeting of RuralInvestments 100 90 80 70 .. s 5 60 -: v) c e 50 E e - m - m = 40 5 0 30 20 10 0 0 10 20 30 40 50 60 70 80 90 100 Deciles of Rural Population by CR Ranking of Index of Access to Basic SeMces (poorest to richest) tPNIR +AFDS -A- PSAOP - PBA (ex-PROCR) -o- PDlS Bank estimates based o n CIBLAGE Index o f Access to Basic Services in Rural Communities, DPS . The 45- degree line (line o f equality) indicates a neutral distribution. Any curve above that line represents a pro-poor distribution o f resources between CR, any curve below it indicates a regressive distribution. The more the curve is inthe upper left quadrant, the greater the share o f its resources go the most vulnerable CR. Targeting Performance 203. The programs with the most positive discrimination towards the more vulnerable rural communities are AFDS, by a large margin, and PNIR. Figure 9 presents the geographical targetingresults for five programs that deliver rural infrastructure and services: Social Development Fund Agency (AFDS), National Rural Infrastructure Program (PNIR), Groundnut Basin Program (PSA-ex-PROCER) and Projet des Services 77 Agricoles et Organisations des Producteurs - PSAOP, as well as the health sector program PDIS. AFDS allocates 93 percent o f its resources to the 20 percent o f the rural population living in the poorest CRs, PNIR 46 percent and the other programs 20 to 26 percent. PSAOP finances producer organizations, which are usually less present in the poorest villages. PBA (ex-PROCER) concentrates most o f its investments across the poverty range, though there i s underpresentation o f the poorest CRs due to program rules that require a certain level o f local government capacity. In terms of overall amounts, cumulative resources vary between US$40 per capita to the poorest CRs, down to US$1.50 in the better-off rural communities, helping to redress inequities (Statistical Annex Table IV.x). Sectoral investment programs tend to have a more neutral distribution. The health investment program (Figure above) and the opening o f primary classrooms sponsored by the Government inrural areas (Table 39) both tend to be appliedequally across the poverty spectrum, with little positive discrimination o finvestmentsinthe least served areas. Observations 204. The development of these decentralized delivery systems for rural infrastructureis an important advance in recent years. Rural populations will benefit from moving control over decisions and resources closer to them. The experience with implementation through local government and communities, even in the most remote CR and villages, has been quite positive, both in terms o f cost effectiveness (generally lower unit costs) andtransparency, as well as building capacity at the local level. This mayhave spill-over benefits on increasing local government and community capacity more broadly to manage risks as they gain experience in implementing development projects, managing money and delivering services. Other benefits to the rural economy include greater recourse to local contractors, with important multiplier effects on rural incomes, and the promotion o f local financial services as resources are transferred locally to be managed. H. Credit and Savings Overview of Institutions,Policiesand Expenditures 205. The 2004 Letter of Sectoral Policy on Microfinance acknowledges the remarkable growth of MFIs over the last decade, but also underlines their unequal distribution across the country. The policy sets forth a vision to "create a professional microfinance sector that i s viable and sustainable, diversified and innovative, integrated into the financial sector, assuring satisfactory coverage o f national demand, and operating within an appropriate legal, regulatory, institutional and fiscal framework. To accomplish this, four strategic axes are identified: (i) improve the legal and regulatory environment, (ii)promote appropriate and diversified services, notably inzones not currently covered by MFIs, (iii)better integration with the financial sector including stronger articulation between MFIs and commercial banks; and (iv) a strengthened institutional framework. Expanding MFI services in rural areas will require the development o f institutional capacities, the diversification o f products and services adapted to rural clients, and building a stable savings base. Support could come in the form o f technical assistance, capitalization, guarantees and lines o f credit at market interest rates. 78 206. Historically, the supervisionof MFIs has been the responsibility of the Micro- Finance Unit (Cellule d'Assistance Technique au Caisses Populaires d'Epargne et Crkditt) in the Ministry of Finance. Responsibility for the promotion o f MFIs was devolved to the Ministry o f Small and Medium-Enterprise and Micro-Finance created in 2003. The ministry has a relatively small budget and does not yet manage significant levels o f investment. In 2003-2004, recurrent budget averaged about US$300,000, with an additional transfer o f less than US$500,000 to support the Agence de De`veloppement et Encadrement de PME, which assists small business. 207. The main donor funding for the micro-finance sector has been channeled directly to MFIs as part of support to private sector development. There has been extensive and on-going support from a number o f bi-lateral agencies (USAID, French Agency for Development, Canada, Germany, Belgium, etc.) to improve management and technical capacity, expand outreach o f services, and build a sustainable micro-finance industryinSenegal. Inrecent years, one of the main supports to the micro-finance industry has been through USAID-financed Dyna Enterprises, amounting to US$26.5 million 1999- 2004 in technical assistance to microfinance institutions (IBM 2004). The program also provided support to the Micro-Finance Supervision Unitwithin the Ministry o f Finance. Observations 208. Significantamounts of public resources havebeen channeledthrough MFIs for other objectives such as employment creation and income generation. Almost US$10 million in Government credit lines oriented to youth, women, the unemployed, and so forth were channeled last year through MFIs. The ultimate impact on MFIs is not clear. Some of these programs may have helpedto bring poor clients into the service orbit o f the MFIs through outreach and other development activities (literacy training etc.). The general performance o f these credit lines in terms o f repayment has been much lower than the average portfolio. Part o f this may be that they are viewed as politically motivated. Moreover, interest rates are often set below prevailing market rates. Anecdotal evidence suggests that effects may be negative in the sense that messages to credit recipients are incoherent overall and these programs do little to help capitalize MFIs, and come with no training or institutional support above expenses.57 209. The Government's efforts may be working at cross purposes when it comes to supporting the development of sustainable financial services to rural areas. The success o f the Senegalese micro-finance industry i s notable, with some o f the highest penetration rates in West Africa and an increasingly diversified set o f financial services available. Although coverage in rural areas is still low, those microfinance institutions present are providing significant benefits to rural dwellers, including more secure and liquid savings opportunities, some level o f agricultural credit, opportunities to finance economic diversification into services, and spill over benefits like payment o f teachers salaries so they do not have to travel and, expected in the near future, a more secure channel for remittance transfers. However, over the long term, the Government policies o f 57There are some exceptions to this, for example AFDS has a pilot program with selected MFIs contracted on a competitive basis to extend outreach to poor rural communities. 79 forgiving rural debts and its preference for subsidized credit lines to ostensibly support poverty reduction may undermine the sustainability o f MFIs. Moreover, there are several poverty reduction-oriented programs that seek to create small local 'caisses', but a proliferation o f small, dispersed savings and credit mechanisms i s difficult to supervise and they are not likely to grow and survive without links to larger federations. Summary of ProgramCoverage, Gaps and OverallExpenditurePatterns 210. In summary, public expenditureson social risk management are spread across many sectors, coverage is limited but expanding, and monitoring of coverage and impact needs improvement. Table 43 provides a general; picture o f amounts, coverage, gaps and key issues. There is a critical lack o f information from many programs on numbersand geographical location o fbeneficiaries, making it difficult to assess the overall cost effectiveness and impact o fprograms. 211. Although Senegal's PRSP includes most of the above areas of risk management, execution of planned expenditures has been less than hoped for in support of vulnerable groups. A recently completed review o f PRSP expenditures carried out by the World Bank (Public Expenditure Review, lOOS), as well as the first two annual reports on PRSP implementation produced by the Government show that progress was uneven inreaching the goals set forth inthe PRSP inthis area: (a) Within the component to improve the living conditions o f vulnerable groups, investments were realized in the areas o f activities to protect children from dangerous work and exploitation through MFFSD,prevent malnutrition through the PRN, and support women via small credit and poverty reduction programs o f MFFSD. There were no incremental investments made for the elderly or disabled nor expansion in many areas to assist children, including funding an increase in school lunches. The problem does not appear to be lack o f financing, particularly since at the same time there has been a significant expansion o f investment on non- PRSP areas, such as administrative buildings. Dispersion o f programs, institutional instability, weak capacity, as well as a lack o f policital clout and visibility of vulnerable groups help explain this under-performance. (b) Within the creation o f wealth pillar, agriculture and rural production received significant investments, including expansion o f irrigation, investments in crop diversification, mobilization o f the bonification, guarantee and calamity funds, and general rural development programs. Employment creation through public works and diversification of rural income bases received less attention. (c) In expanding access to social services, there have been extensive investments through the education and health sector programs (PDEF and PDIS) which form the core part o f the PRSP. However, in health investments at the community level fell far sort o f the level envisioned. As mentioned before, there was significant investment inexpanding access to education, which benefited ruralpopulations. 212. As the Government begins to revise its PRSP and set strategies for the next PRSP period, it will be important to better capture these elements of risk and vulnerabilities. Since many o f these expenditures, including social assistance transfers, food aid, school lunches, staffing o f school and health centers, and routine services o f the 80 agriculture and rural extension agents are not covered inthe investment budget, but instead appear as recurrent costs, it will be important to expand the universe of expenditures captured under the PRSP to the broader national budget. Table 43 :SynopsisofExpendituresonRuralSocialRiskManagement, billionCFA Area Coverage, Issues and Gaps Social Assistance, 1 Disaster management has supported several thousand households affected, for example, by Safety Nets and flooding. Cost per person assisted from US$30-120. Coordinationneeded to avoid multiple and Natural Disasters overlapping response I Transfers to vulnerable groups are about US$1,100 per request. Total beneficiaries not available, (Net of poverty but small share (less than 1%) o ftotal vulnerable and few rural reduction investments 1N o data on number o fbeneficiaries or amount per beneficiary o f rice distributionbut household to communities) surveyspoint to widespread distribution 1 ILO and UNICEF supported children at-risk programs reachunder 15,000 children, or less than 6 percent o f target population, at about US$75 per child 1 Significant support for women's groups, no data o n number or location o f recipients Agriculture and 1Most rural households depend o n agriculturalAivestock Livestock 1 Large subsidies to groundnuts via credit, inputs and price supports, scant information on impacts (Total spending) INeed for more expenditures on prevention (of parasites, cattle theft, etc.) Of which spec@ risk 1N o data on number o f producers helped by guarantee, calamity or bonification funds mitigation and coping 1 Potentially negative impact o n financial system Health 1 L o w presence o f trained staff inrural areas, difficulty retaining staff (Total spending - not 1 Affordability i s main issue - 25% o f rural sick do not seek treatment due to cost dissagregated for 1 Half o frural population lives more than 3 k m s from health center rural) Specijk HIV-AIDS 1 Rural communities have increasing access to financing -recent scale up o f activities inrural areas, interventions though overall urban areas receive most 1 Only Ziguinchor has pilot outreach for rural treatment 1 Rural highrisk groups (e.g. migrants and their families) should be targeted Specijic spending on 5.6 1 Coverage o f 12% o f rural children under 3yrs at 4$/benef.; $17 for total program cost malnutrition (PRN) 1 Initial impact evaluation shows drops inmalnutrition,rates 1 Scale-up potential, with rural areas as priority to target largest malnutritionrisks Specific Malaria n.a 1 Bednets are cost effective prevention strategy (about US$5 per ITBN). Scale-up will require interventions expanding outside o f health systemto increase coverage I Rural areas have highest riskand lower access to curative services - 1 Epidemic early warning system is possible for the north Education 26.1 1 300,000 rural children out o f school (total spending on 1 Recurrent cost per rural student $58 per year rural primary 1 School lunches $21 for recipients o f WFP lunches, but problems reaching poorest students - education) 5 11,000 rural students not receiving school lunches - 1 Demand-side factors a key challenge Employment 1.8 1 Overall coverage figures for employment programs (via credit and training) low and urban- Generation centered. At most optimistic, 25,000 jobs created or 1% of 20-40 yr olds 1N o consolidated information o nj o b creation through public works nor linkages to rural seasons 1 Poverty-focused rural infrastructure programs shown to reachpoorer rural communities, - decentralized management creates local economic multipliers Savings and Credit 9.5 1N o t including the agricultural h n d s (FB, FG) which are channeled through CNCAS, credit lines (total o f various credit account for 7 percent o f total MFI credits lines - includes credit 1 Lower repayment rates on Government credit lines through MFIs and CNCAS) programs founding 1 Beneficiary selection disfavors access o f rural and poorer households-information o n total above sectoral beneficiaries not available for many programs categorie~)~' 1 Inconsistencies with strategies for sustainable expansion o f microfinance '':FB, FG, AFDS, PCLP, FNPJ, FNAE, Program of Credits to Women MFFDSITaiwan 81 CHAPTERV: Key PolicyReformOptions and ExpenditurePriorities BringingRisk into Policy Focus 213. The rationale for policy intervention arises when individuals fail to attain optimal levels of risk pooling, saving, and prevention. Theoretical justifications for Government interventions inrisk management include (from Gill andIliahi 2000): to provide risk-pooling instruments in the cases o f market failure in information (disasters, pre-existing health events with catastrophic results like malnutrition); to provide superior risk management measures (like safe and reliable forms o f liquid savings) when individuals would otherwise only have access to inferior mechanisms (usinglivestock as precautionarysavings); to build and protect human capital as a risk prevention strategy (better educated individuals are more likely to invest in preventative activities like pre-natal care, healthier children are more likely to attend school and adults to work); to make up for capital market failures where individuals without access to credit or savings under-investinrisk management instruments; and to invest in less costly prevention measures for which either individuals may not have the resources or there i s a public good nature and coping costs to the government are high, such as vaccinations, flood and locust control. 214. The PRSP and sectoral policies identify main risks. But, there has been less attention to whether programs and expenditures line up with these risks, who benefits and the impact o f policies and expenditures on reducing risks. It i s important for policymakers to identify the risk being addressed as well as the justification for intervention. Does the mechanism proposed adequately address the risk? Who i s covered or excluded? H o w much prevention, mitigation or coping i s `bought' with this public spending? Improving the Targeting, Relevance and Impact of the Nascent Social Safety Net 215. As income levels rise, governments tend to spend more, not less, of their resources on safety nets and other social assistance to vulnerable groups. Senegal has a patchwork o f social assistance, most internally financed, but it suffers from a dispersion of effort and a lack o f clarity on what is to be achieved. The core challenge i s to create mechanisms tied to programmatic objectives to reduce vulnerability. 216. To achieve this end, specific recommendations involve: (a) The report points out the multiplicity o f institutions, with problems o f lack o f coherence, dispersion o f efforts and duplication. Moreover, there i s no obvious institutional home for risk and vulnerability as a transversal issue. The institutional framework has suffered from shiftingresponsibilities and ministerial dismemberments. The formation o f the National Social Protection Strategy Steering Committee may help to improve strategy setting across ministries, but it needs to ensure that it i s operational. Implementation o f a joint annual review mechanism like the one used in 82 health and education to bring together the diverse actors and partners and monitor progress annually could be an effective mechanism. (b) Consolidation o f the various transfers to vulnerable groups (women. elderly, disabled etc.) into a coherent set o f interventions. Efforts are diluted across ministries and dispersed in too many small programs leading to duplication. The Government will need to prioritize programs that are substantial enough in size to capture administrative economies o f scale and to reach a greater share o f the target populations. (c) Transition from the current `unconditional transfer' mode to a programmatic focus on assisting targeted groups reduce their vulnerabilities. This includes moving to program-based budgeting by objective with indicators o f outputs and goals for geographical coverage. Inthe area o f disability, the Community-Based Re-adaptation Program, which holds promise as a cost-effective approach, has been approved and staff trained, but investment resources have not yet been allocated. In other areas like grants to organizations working with vulnerable groups, there needs to be greater focus on what services the Government i s purchasing with these transfers (e.g. number of children reached, cost per beneficiary, % o f disabled finding employment etc.). Several countries have initiated conditional cash transfer programs (see Box 7). The Government o f Senegal's PRSP calls for design and piloting o f such an incentive system to increase human development outcomes among poor households, however, p o x no progress to date has been made to define such a program. 7: Conditional Cash Transfers to Boost School Enrollment and HealthUtilization Several developing countries have recently introduced conditional cash transfer programs which provide money to poor families contingent on certain behavior, usually investments inhuman capital, such as sendingchildren to school or bringingthemto health centers, as an alternative to traditional social assistance programs and a demand-side complement to the supply of health and education services. Rigorous impact evaluations for programs launched in Colombia, Honduras, Jamaica, Mexico, Nicaragua, and Turkey reveal successes in addressing many of the failures in delivering social assistance. These conditional cash transfer programs demonstrated a positive effect on enrollment rates for both boys and girls, improvements in child health and nutrition, including nutrition monitoring and immunizations, and increases in household consumption., for example as measured by median caloric intake per person. Monthly subsidies vary between program and country, ranging between US$4 and $13 per beneficiary (family or individual) per month. The flagship program PROGRESA inMexico currently serves 20 million people or one-fifth ofMexico's population. FromRawlingsandRubio (2005) (d) A tighter focus on the truly vulnerable and adoption o f targeti,ng too the risk rather than the general group which include individuals that may not be at risk. Categories like `women' or `vulnerable communities' are too broad to ensure appropriate targeting and create opportunities for elite capture. Ensuring that resources flow to those at greatest risk can be furthered through better analytical tools (poverty maps, identification o f household characteristics, community-based targeting mechanisms). Moreover, in the absence o f easily identifiable household or village characteristics o f poverty and limited data and administrative capacity to target benefits to specific 83 households, the Government should explore targeting to the risk rather than the person, This could include grouping transfers to needy individuals, distribution of rice and expenditures on public works to critical times o f the year insynch with the agricultural cycle, or providing school grants to orphans as they are at greater risk o f not attending, or identifying villages experiencing multiple negative shocks and prioritizing public investments to them. Improving the legal framework. Reducing the vulnerability certain groups (e.g. disabled, children being exploited, divorced and widowed women, rural populations in conflict over land titles) can be addressed inpart by changing, or better enforcing, laws and better informing the population o f their rights. An IntegratedApproachto DisasterManagement 217. Disaster management is one of the most visible and appropriate areas of Government intervention. Effective response requires a pre-established institutional framework, accessible resources and simplified criteria and procedures. For civil emergencies, like fire, accidents etc., the basic institutional elements are in place, but lack clearer division o f responsibilities inassisting affected populations. 218. From experience in recent years, the most frequent disasters are linked to natural shocks in rural areas. The Civil Protection system is limited in the case o f agricultural calamities and the main agricultural response systems, like the FC, is not linked with the other mechanisms, like the CSA, the NSF and others. Recommendations to enhance the efficiency and effectiveness o f the system include: (a) Integrate natural disaster responses: Responses to risks tied to the person (loss o f life, needs o f temporary assistance in food, cash, materials) need to be integrated with response to risks tied to production (loss o f crops, seeds, livestock). At present, the responses runinparallel, reducingoverall effectiveness. (b) Reduce reaction time and cost o f response through early warning systems. There are several parts o f a disaster early warning system in place or in process.59Increased predictive power through vulnerability modeling and hazard analysis would help to identify those zones and people at greatest risk and develop lead indicators based on analysis o f previous disasters, particularly since these disasters are not random, but concentrate in certain times o f the year, among specific sub-zones and are linked with certain threshold occurrences. The system should be targeted to the most frequent and economically damaging shocks and based on existing national scientific and technical capacity as well as international systems (CLCPRO and EMPRESS for locusts). Clear guidelines on responsibilities, including the centralization o f information and its dissemination, are needed. (c) Better identify those affected and the extent o f damages. The most difficult phase in disaster management i s the evaluation o f damages and identification o f affected 59CSA's FoodSecurity Bulletin, FEWS, the National MeteorologicalDepartment, the EcologicalMonitoring Center, Climate Change and Drought Program(Institut des Sciences de la Terre - UCAD),Water Resources Monitoring System (Dept. of Water Resources and Management) Ministry of Agriculture and Hydraulics, Locust Early Warning System (EMPRESS, FA0 and the Ministry of Agriculture and Hydraulics' Dept. of Plant Protection). 84 households. Methodologies exist to develop loss profiles for the principal types o f agricultural risks and should be systematically applied (for example how to estimate loss o f crops or livestock in the case o f drought, floods or locusts). The same standardization o f methodology should be applied to create financial models o f predicted resource needs inthe case o f other disasters. The level of resources committed in the face of natural calamities (for example locusts in2004) justifies a greater investment in the prevention o f shocks, including flood protection and plant and animal disease control. Better ManagingAgriculturalRisks 219. Traditionalshocks of drought,insects, andplant and animal disease are joined with modern shocks associated with those of an open economy in the process of changing economic production systems as well as growing social ills like livestock theft. Responses have been put into place for prevention as well as coping with the after effects o f these shocks, at significant financial cost. But the impact o f these measures has been limited by a lack o f clarity in rules o f the game and a dispersion o f responsibilities. Modern instruments to mange agricultural risks have not yet been sufficiently explored. 220. In the absence of the development of a private insurance market and limitations of the agricultural credit market, the Government has become both the lender and insurer of last resort. The central role o f the State as ultimate risk management mechanism for rural production poses several problems, including fiscal sustainability, moral hazard and overall efficiency. Even if the nature o f the problem will require public financing for the medium-term, it does not necessarily follow that the government i s best placed to manage this system. Key recommendations include: Given the importance o f development of sustainable rural credit systems and the need for effective mechanisms to manage risks, the Government should carry out an evaluation o f the experience over the last few years o f the various funds created, including the potential distortionary effects on credit behavior, and spill over effects on the broader rural finance situation including MFIs. If retained, the State should favor the management o f the credit securization funds, and notably the guarantee fund, by contractingprivate sector services. Investigate the applicability o f amicultural insurance instruments, like weather- indexed-based agricultural insurance, currently being implemented in several developing countries (Morocco, Mongolia, to cite a few) are applicable to both rainfed agriculture and extensive livestock systems o f Senegal (see Box 8). Prevent shocks through developing irrigation, including small-scale groundwater irrigation in the Groundnut Basin, and facilitation o f access to appropriate technologies, research, dissemination and training o f producer organizations. Reconsider the policy o fblanket rural debt forgiveness. It i s too blunt an instrument, subject to political pressures, does not necessarily help the poorest rural producers, and may adversely affect the development o f sustainable agricultural credit. Develop better informational tools to underpin the amicultural risk management system, including a Livestock Census and a Professional Electronic Card (Carte 85 professionnelle klectronique) linked to the national identity card that would include basic information (land ownership, areas under production, units of livestock etc.). 86 Box 8: Index-BasedAgriculturalInsurance' The inherent instability o f farm income has led governments to devise programs and policies that have a stabilizing effect on agricultural income. Public subsidized agriculture insurance has been implemented in more than fifty countries as a mechanism to stabilize farm incomes by reducing exposure to production risks, mainly under a private-public partnership. Inthe past however the experience with subsidized insurance programs has been disappointing. A new framework for the financing o f agricultural risks, supported by the World Bank Contractual Savings and Insurance Unit,highlightsthe objectives andmodalities ofindex-based insurance: Differentiating market-based insurance and social insurance. The lack o f clarity regarding the objectives of public interventions in agricultural insurance has contributed to its inefficiencies. Social insurance with a safety-net function aims at assuring a minimum economic security to farmers involved in low return activities. These social objectives rely on redistributive subsidies. Market-based insurance i s oriented toward viable business activities that generate enough profit to afford insurance premium. These instruments, based on sound actuarial principles, should apply only to viable farms whose survival may bejeopardized by the occurrence o f an insurable event. Assessing agricultural production risks. The existence o f reliable and accurate long-term loss data series i s a precondition for the development o f any market-based product as they are used to assess the hture probable losses. Individual farm data are almost always missing or unreliable. Consequently, loss assessment i s performed usingaggregate data to provide an objective estimate o fpotential losses and capture the spatial correlation o f losses caused by widespread events. Selecting an insurance index. Index-based insurance makes payments based on an index, irrespective of the individual losses. It transfers covariate production losses caused by widespread weather events (e.g., floods, droughts) or epidemics to financial and reinsurance markets, Index- based contracts offer advantages over traditional individual insurance (reduced moral hazard and adverse selection, low administrative costs, standardized product) but it exposes the contract buyer to the possibility that the payout i s different to the individual loss, and relies on good data. Experiences in index-based insurance. Area yield crop insurance, where the index i s based on the average yield in a given geographical area, has been offered inMorocco, India, Brazil, Canada and the USA. Parametric insurance (e.g., rainfall insurance) has been proposed in Canada, India and Mexico. A livestock mortality index has been recently considered as a basis for indemnifying herders in Mongolia. The mortality index-based insurance pays indemnities whenever the adult mortality rate exceeds a specific threshold for a localizedregion. 0 Financing agricultural production risks. Riskfinancing strategies deal with the remaining part of the risks that cannot be mitigated with cost-effective preventive measures. They are financed through farmers' self-retention, private financial markets, and governments and international donors through an appropriate layering of risks. The bottom layer of risk includes high frequency (e.g., occurring once every five years or more frequently) but low consequence risks that affect farmers from a variety o f mainly independent risks. The mezzanine layer of risk includes less frequent (e.g., occurring once to six times every 30 years) but more severe risks that affect several farmers at the same time (e.g., hail, frost). The top layer of risk includes low frequency (e.g., occurring once in 30 years or less frequently) but high severity risks. These catastrophic risks are by definitionnot well documented andthe probable maximum loss canbe very large. From Olivier Mahul, The Financing of Agricultural Production Risks: Revisiting the role of agriculture insurance. Contractual Savings andInsuranceUnit,World Bank. Draft January 2005 87 ImprovingRuralHealthand EducationOutcomes 221. The greatest challenge to improving Senegal's human capital is in the rural areas. The ability o f the sectoral ministries to ensure that adequate facilities, staff, materials and other support are effectively allocated to rural areas has been a continuing issue. Conditions in rural areas have made it difficult to recruit and retain qualified staff. Centralized budget systems have limited access to and control over resources and decisions to a handful o f central actors. To boost performance o f rural health and educational. systems, systemic changes inhow and where decisions are made over allocation o f funding and personnel may be necessary. Initial experience in education o f delegating budgets to IDEMs appears encouraging and similar actions areproposed for healthinthe PNDS Phase I1policy document. Inaddition, specific sectoral recommendations include: 222. For education, addressing rural risk requires demand-side strategies as well as continuingthe good momentumin buildingand staffing rural schools: (a) Development o f an expanded program for school lunches in rural areas that i s affordable within fiscal constraints and addresses issues o f access by the poorest. Using the WFP's estimate o f per student cost per year o f US$24, 13 percent o f which i s paid by parents, expanding coverage o f school lunches to all rural primary students in Diourbel, Louga and Kaolack for example (242,000 students) - the regions with the lowest rural enrollment rates - would cost on the order o f US$5.1 million in 2006, rising to US$6 million over the next two years as enrollments expand. This would represent a 10 percent increase in rural education spending, or about 1.2 percent o f total education spending (b) Diversification o f educational programs relevant for rural populations, including language of instruction and local curriculum concerns; (c) Identification o f options to address the trade-offs parents face between farm labor, and agricultural cycles in general, and schooling. Some countries have adjusted the rural academic year to better avoid conflicts with peak family labor times. Other options include revisions to the schedule for payment o f school fees. (d) As part o f social assistance programs, making school enrollment conditional on receiving support that i s provided by MFFSSD for orphans, children at risk, disabled etc. and including provisions for school expenses within this support. (e) Expansion o f the school-based health proaam to deepen rural coverage and address critical rural healthrisks that cause childrento drop-out. Health policies can boost preventionof critical ruralhealthshocks through: (a) PNDS I1calls for expansion o f midwives at the health post level, which should be allocated by priority to rural areas where maternal mortality i s highest. (b) Investing in better community-level services, including outreach equipment and materials, can help to link remote health posts and hutswith the referral system. (c) PNDS Phase I1proposes incentives for health workers in `difficult' areas and a rural premium. However, an extensive list o f other incentives (e.g, for performance, responsibilities) is administratively unwieldy, subjective and may dilute impact. (d) PNDS I1would address affordability through expansion o f community-based health insurance and exoneration o f fees for the indigent and vulnerable groups. No 88 specific mechanism i s yet designed for the latter, but given the abuses inherent in such a system, Senegal could benefit from international experience in designing effective fee waiver options (see Box 9). (e) Inpreventing malariathroughbednetdistribution, the existing network ofNGOs and rural health facilities do not reach the majority o f rural residents. Giventhe impacts o f malaria, the full force o f rural associations should be enlisted, including women's groups and village development associations who have more village presence. While social marketing i s an important element, it excludes families too poor to afford prevention. Given the greater cost o f treating (or leaving untreated) malaria, increased prevention expenditures can reduce overall costs. For the North, an early warning system can better address epidemics. (f) Effective models o f interventions to prevent malnutrition should be mainstreamed and scaled-up. To limit costs, more refined geographical targeting will be necessary, including a priority on rural areas. (g) HIV-AIDS efforts should be targeted to specific risks and characteristics o f rural populations and treatment models diversified to allow greater access by rural residents. (h) Community-based health insurance schemes can increase utilization o f services and mitigate the effect o f health shocks on families, though their applicability will be limited inreaching the poorest families. Box 9: International Experience inWaivers and Exemptionsfor Health Services In response to shortages in public budgets for government health services, many developing countries around the world have adopted formal or informal systems o f user fees for health care. The problem with user fees i s that the lack o f provisions to confer partial or full waivers to the poor often results in inequity in access to medical care. The dilemma, then, i s how to make a much needed system o f user fees compatible with the goal o f preserving equitable access to services. Different countries have tried different approaches. Those which have carefully designed and implemented waiver systems (e.g., Thailand and Indonesia) have had much greater success interms of benefits incidence than countries that have improvised such systems (Ghana, Kenya, Zimbabwe), Key to the success o f a waiver system i s its financing. Systems that compensate providers for the revenue forgone from granting exemptions (Thailand, Indonesia, and Cambodia) have been more successful than those who expect the provider to absorb the cost o f exemptions (Kenya). Where waiver systems exist, performance will improve with the timeliness o f the reimbursement, Other success factors include the widespread dissemination o f information among potential beneficiaries about waiver availability and procedures; the awarding o f financial support to poor patients for non- fee costs o f care, such as food and transportation (as in Cambodia); and the existence of clear criteria for the granting o f waivers, thereby reducing confusion and ambiguity among those responsible for managing the system and among potentialrecipients. FromRicardo Bitran and Ursula Giedion, Waivers and Exemptions for Health Services inDeveloping Countries, Social ProtectionDiscussion Paper No. 0308; PublicationDate: 3/03, World Bank 89 The Biggest Gaps: Rural Income Diversification and Expanded Financial Services 223. In their private risk management strategies, rural households often seek to diversify out of agriculture. Household surveys show that in general the more successful rural households are those less reliant on agricultural income. Priority actions to enhance income diversificationinclude: Rural public works program (or improving the employment focus o f existing rural infrastructure program_S1: Lessons from extensive international experience o f public workfare from countries in Sub-Saharan Africa and Asia particularly confirm the considerable potential to help the poor cope with covariate risks associated with climatic and systemic shocks (Subbarao 2001). Advantages include: (i)income transfers to poor at critical times; (ii)consumption smoothing benefits to counter shortfalls during slack seasons; (iii)minimization o f the trade-off between public spending on income transfer versus developmental activities via construction o f infrastructure; (iv) can be targeted to communities most at risk. Evaluation o f the lessons learned from income-generating proiects in rural areas. There is a vast array o f projects and millions o f dollars spent each year to support income generating activities across many ministries and programs, as well as NGOs, each using different approaches. To date, there has been little review o f the impact o f these on incomes and no systematization o f lessons learned. Expansion o f savings and credit mechanisms. Beyond generating capital for rural economic growth, the broader benefits o f protection o f assets, enhanced ability to smooth consumption, and fostering remittance flows to rural households place this as priority for rural risk management. While existing practice o f subsidized credit lines need reconsideration, options to promote expansion o f MFIs into rural areas, drawing on international best practice and country experience, should be analyzed. Possible approaches include developing guarantees better suited to rural areas, removing distortions from subsidization, developing matching grants to encourage MFIs to expand their networks, and developing linkages betweenMFIs and banks. These efforts must be accompanied by capacity building through training, management information systems and technical assistance so that they can adequately manage expansion. Moreover, currently most supervision by Government regulatory agencies i s focused on urban MFIs. Increased budgets for these efforts are needed better monitor rural institutions since they manage considerable volumes o f savings and money transfers that would be at risk should the institutions fail. To optimize the impact o f remittances, many governments aim to provide their migrant communities with better information on transfer services and costs (e.g. Mexico, Bangladesh), and some attempt to attract remittances also into investment vehicles or programs they have launched, such as a pension plan or business ventures based on matching or joint inputs. Other countries, such as Haiti, have negotiated to bring down the price o f international transfer services. 90 Towards a More EquitableDistributionof Public Goods to RuralAreas 224. Overall, public investments and expenditures are tilted to urban areas. In addition to expanded private sector opportunities, the likelihood o f benefiting from public spending i s greater for city dwellers. While it was outside the scope o f this review to analyze overall geographic expenditure patterns, most ministry expenditures (with the exception o f agriculture) show a general concentration o f resources in urban areas, notably Dakar. This bias may help explain part o f why rural areas have lagged behind their urban counterparts interms o f poverty reduction. 225. At the same time, this review has confirmed a growing number of effective delivery mechanisms for investments in rural areas. Decentralization and the use of community-driven investment programs have shown that even the most remote communities can be reached in an efficient manner. NGOs are increasingly being contracted by the State to provide outreach services in HIV-AIDS, malnutrition, and general community development. The panoply o f mechanisms scattered across ministries requires harmonization and rationalization, but the basic conditions o f scaling up are present. This scale-up can have positive spill-over effects on rural economies and institutions, including creating a stable technical staff in rural local governments, NGOs, and other decentralized programs as well as local economic multiplier effects (and cost savings) through contracting with local construction firms and suppliers. Puttingthe Pieces Together - a Multi-SectoralApproach to Risk Reduction 226. Lookingat risks from a multi-sectoralperspectivehelps identify vicious cycles that will require collaboration between sectors. This review has identified a number of vicious or virtuous cycles o f interactions between sectors. For example, the health sector targets its anti-malaria efforts on pregnant women and children under 5 years o f age who are at greatest mortal risk. However, malaria is also an educational risk, with illness o f children - o f which malaria i s the leading cause - a key determinant o f not achieving primary education. Thus, to prevent education shocks, malaria prevention and treatment policies will need to include a focus on school-age children. Rural infrastructure investments are important for expanding access to services, but they also can serve the interests o f income and.employment policies. Social assistance programs and food aid can be linked to achieving outcomes in terms o f meeting the MDGs in nutrition, education, vaccinations, and reduction o f extreme poverty if appropriately designed. 227. This multi-sectoralapproach would benefit from additional diagnosticwork in several key areas. The dearth o f information on program coverage and characteristics o f beneficiaries limits understanding o f the reach and targeting o f current policies and programs. At a minimum, programs should begin to report number o f beneficiaries and general benefit levels to complement the existing emphasis within the PRSP on expenditure amounts. Inaddition to improvements in administrative information on coverage, national households surveys, especially those oriented to monitoring the poverty situation, should be expanded to include a module on coverage o f public programs. In addition, creation o f a national panel household survey, which several countries have done, will provide a far better diagnostic tool intracking the dynamic interactions o f poverty and risk. While there 91 are several programs that have included rigorous impact evaluations, such as PRN and PNIR, these remain the exception, with too little information generated on the performance and effectiveness o f programs. As part o f PRSP monitoring and evaluation, the Government should identify a set o f critical programs that should have impact evaluations and enlist international experience and donor financing to review experience. This will make it easier to attract incremental fundingto scale-up programs that have an impact. 228. The revision of its Poverty Reduction Strategy Paper and development of a National Social ProtectionStrategy in the coming months affords the Government an opportunity for highlighting the importance of risk and vulnerability to poverty reduction and for making the multisectoral linkages more explicit. The process has benefited from broad stakeholder discussions and significant input from Senegal's external partners, creating conditions for donor alignment on these themes. Moreover, the transition to more sophisticated risk management instruments, including better targeting o f programs to at-risk groups as well as the development o f market-based mechanisms, like insurance or expanded financial services, i s an integral part o f the broader development process. Enhancing rural households' ability to prevent, mitigate the impact o f or recover from the frequent shocks they face will be a key determinant o f Senegal's ability to meet its development objectives. 92 ANNEX A: Descriptionof Main DataSources 0EnquCte SCnCgalaise Aupr6s des MCnages (ESAM 11) - This is the most recent national household survey implemented in 2001-02 by the National Statistics Department which includes modules on household consumption and expenditures, labor and migration, health and nutrition, education, and household assets (based on QUIDmodules);As part of the ESAMI1a qualitative 0 Perceptions of Poverty survey was carried out the gauged respondents impressions o f poverty and identification o f priority needs. The survey is representative nationally, urban-rural andby region; 0 EnquCte sur les Objectifs de la Fin de DCcennie sur 1'Enfance (MICS - 11, 2000) - Carried out by the Government of Senegal and UNICEF, the household survey contains data on well-being o f rural (and urban) children in the areas o f health, education, water and sanitation, child labor and orphanhood. Representative nationally and rural-urban; 0 PASEC Data set -- The data set consists o fpanel data from the PASEC (Programme d 'Analyse des Systimes Educatifs de la CONFEMEN) standardized achievement test (1996 base year through 2003), together with community, school and household modules applied in 2003 to identify contributing factors to student performance as well as a household schooling decisions. Sample was based on 118 primary schools, of which 58 rural. Village questionnaire applied to 32 rural communities; household survey to 938 rural households. The data is representative nationally and rural-urban for school children (sample bias against communities without schools). Since this is the only source for information on such issues as access to rural credit, perceptions o f shocks, etc. we have chosen to report findings from the survey that are not linked to education, despite the slight sample bias. 0 National Agricultural Census 1998-99 - Has both a village and household questionnaire which includes agricultural assets and production, access to infrastructure and farm labor; 0 Groundnut Poverty and Social Impact Assessment (PSIA)-Household survey carried out in 2003 in the Groundnut Basin to estimate the effects o f changes in the structure o f production and commercialization inthe groundnut sector 93 Annex B: Statistical Annex Chapter I1 Table 1I.a: Senegal poverty incidence by region (YO) Source: Senegal ESAM-I1 3 24.93 4.25 53.8 I 0.5 I 100 J 94 4 23.29 5.01 13.44 58.02 0.23 100 5 22.54 3.82 13.6 59.91 0.13 100 Total 24.63 4.03 14.32 56.55 0.48 100 Source: PASEC data set, Bankestimates. Figure I I a :Evolution of Rainfall in South, Center and North of Senegal, 1971-2000 I I evolution pluviom etrie 1600 500 1400 450 400 1200 350 1000 300 800 250 600 200 150 400 100 200 50 0 0 +Casamance -+--normale 71-2000 +Louga-St-Louisf--normale71-2000 I 95 Figure1I.b :EvolutionofRainfallinDakar-Diourbeland Sine Saloum, 1971-200 700 Evolution pluviometrie evolution pluvio 1000 600 900 800 500 700 400 600 300 500 400 200 300 200 100 100 0 0 I -+- Diourbel-Dakar -R- normale 71-2000 1 1 --eSine Saloum -t-normale 71-2000 I Source: National Meteorological Department i Rural Dakar I Other UrbanCenters I Affected en % Affected en % Affected en % Affected en % YES 1223 87,42 14 3,76 203 36,64 1440 61,94 NO 176 12,58 358 96,24 351 63,36 885 38,06 Total 1399 100 372 100 554 100 2325 100 Year Prospected Treated YOinfected/ YOtreated/ Areas Infected Areas Areas prospected prospected 96 Table 1I.g :Incidence of LivestockDisease Disease # of households I # of illnesses # of deaths Plague from small ruminants ClavelCe 3 78 7 Foot-and-mouth disease 12 250 4 Newcastle diseases 2 6 6 Equine plague 5 24 12 Nodular dermatosis 12 712 0 Bovine Pasteurellose 65 230 125 Pateurellose from small ruminants 41 975 588 Botulism 17 192 115 Bacterial coal 14 92 85 Symptomatic coal 30 154 76 Source : Direction de 1'BlevageiMAE annee 2000 NonPoor Poor Q1 42 43 44 Q5 Lack o f seeds 71.7% 68.2% 66.6% 69.8% 71.6% 76.9% 64.0% Lack of fertilizers 4.6% 9.5% 11.6% 7.6% 7.2% 2.6% 3.0% Lack o f labor 3.1% I 2.5% I 3.2% I 1.6% I 2.2% I 2.8% 15.4% Lack o f equipment 2.6% 3.8% 3.6% I 3.9% 3.0% 2.1% I2.9% I Lack of draft animals I 0.2% II 0.4% I1 0.0% I 0.9% II 0.0% II 0.0% I 0.6% I Source: PSIA Groundnut Basin Survey 97 Urban rural total Number o f days absent from worWschoo1 None 53 43 47 less than 7 days 28 33 31 7 to 14 days 9 13 11 more than 14 days 8 10 9 Figure1I.c: Causesof reportedillness duringlast month, rural households malaria 49% e - diarrhea dental 7 0 / Source: Senegal E S A M I1 Table 1I.i:Percentage of sick or injuredpeople over the last4 weeks by type of illnesshjury, 98 Fever or Dental Skin Ears, nose Malaria Diarrhea Accident problem problem Eyes Or throat Other 50-64 9.8 1.1 0.3 1.1 1.5 3.7 2.8 18.4 65+ 10.6 2.0 1.2 0.9 1.7 4.4 3.5 22.1 Source: Senegal ESAM-I1 Participation innutrition Projected Stunting (%) Wasting (%) Underweight (%) program growth Location Rural 42.7 10.1 23.6 14.4 30.7 Urbain 20.8 7.7 11.7 31.2 57.2 Table 11.1:ChildMalnutritionIndicators,by ruraland urban Source: MICS 2000 99 Source: MICS 2000 Table 1I.n:Percentage of households who consult health services if sick, all ages (YO) Rural 1 2 3 4 5 Total Yes 62 61 63 61 67 62 No 38 39 37 39 33 38 ND 0 0 0 0 0 0 Total 100 100 100 100 100 100 Urban 1 2 3 4 5 Total Yes 66 68 70 75 77 74 N o n 34 32 30 25 23 26 ND 0 0 0 0 0 0 Total 100 100 100 100 100 100 Table itures Source: Senegal EsamI1 100 Poor 4 2 4 3 4 4 Rich Total Literate 22.8 23.0 23.8 26.0 23.3 23.5 Illiterate 74.3 75.2 74.1 72.4 74.5 74.3 ND 2.9 1.8 2.1 1.7 2.1 2.2 Total 100.0 100.0 100.0 100.0 100.0 100.0 age 7 to 12 only Reason % o f total responses agelfinished school 4.7% school too far 4.9% too costly 12.9% work at home 11.8% no interesthot usehl 33.8% health problems 19.7% failed exams 9.6% Marriage 2.6% Source: ESAM-I1 100.0% 101 Chapter I11 Table 1II.a: Share of HouseholdsContributingto Associations. bv auintile and amount urbanby quintiles 1 Q1 42 I Q3 1 44 1 Q 5 1 total I ves I 51.64 I 54.18 I 56.95 I 58.74 I 57.22 I 57.04 I average contributions, annual, CFA (if cont>O) 15133.0 28647.0 29188.0 31752.0 55141.0 39421.0 rural only, by quintiles Q1 2 3 4 5 total no 36.83 30.51 29.53 26.81 30.99 3 1.89 yes 63.17 69.49 70.47 73.19 69.01 68.11 average contributions, annual, CFA (if cont>O) 16,668.0 14,957.0 21,800.0 21,904.0 25,834.0 18,453.0 Source : Senegal E S A M I1 102 1I Table 1II.d: Foreign Country Where Household Migrant Resides destination #urban #rural #total %urban %rural %total Benin 11 463 II 347 II 810 II 1 II 0 II 0 1I Source: Senegal ESAMI1 per eq adult for those receiving these transfers 100,546.6 68,776.4 69,361.9 65,804.3 74,897.7 86,779.5 159,080.8 % Revenue Transfers per eq adult over total o f households 36.3% 26.8% 26.6% 28.2% 27.9% 40.3% 37.0% 103 Source; PASEC data set, village questionnaire o f 32 rural villages Table 1II.g : Access to Formal and Informal Financial Services % of householdswith a deposit in a bankor other formalinstitution, by asset index Rural households I 1 (poorest) I 2 1 3 4 5 I Total YES 11.83 14.84 I 5.91 I 12.64 I 16.39 I 12.3 Rural households 1 2 3 4 5 Total YES 56.99 57.46 43.55 25.97 18.38 40.48 NO 43.O 1 42.54 56.45 74.03 81.62 59.52 Urban households 1 2 3 4 5 Total YES 41.62 51.46 56 63.37 51.76 52.85 NO 58.38 48.54 44 36.63 48.24 47.15 104 Table 1II.h :Frequencyof PurposeCited for Taking a Loan, by number of loans Other related costs to the activity 8% I 10% Purpose of Loan Urban PERSONNEL II 78% I 77% A ricultural Activit Agricultural Inputs Purchase1housing upgrade Purchase o f heavy e ui ment Scholarly expenses Purchase of other equipment 1 %" 1%" Purchase of Animals Religious Ceremonies, marriages, I I 1% 0% Other Agricultural Reasons 2% II 1O hnerals h tL Durable consumption of goods 1% 4% Non-Agricultural Activity I 14% I 19% IPurchase of raw material and capital Other 3% II 5% mobility 4% 1 4% III100%1 100% Purchase or improvement of Source: PASEC data set, Bank estimates landbuilding>equipment 1II 2% I 5% I Table 1II.i: CNCASloansin 2003 (inCFA) Source: CNCAS 105 Chapter IV Table 1V.a :F S NFlood Assistance 2004 Source: DDI Table IVb: CSA Rice Distribution 2004 (tons) Destination Volume intons MCdina Gounass 40 Sindia 40 Koussanar 5 Kedoudou 40 Kaffrine 40 Religious Ceremonies 5414 Total 5579 Table 1V.c :Situation des demandes de secours et des satisfaction en 2004 Source: D A S 106 Table 1V.d: Financingto Agriculture and RuralProduction, 2003 2004 - PublicFinancial InvestmentsfromPrivateFunds 2003 2004 Unit Budgeted Realized YO Budgeted Total PublicInvestments 169400 148347 0,8757202 202250 Public Investment inthe agricultural sub-sector MFCFA 16931 10996 0,6494596 18423 In WhichIrrigation MFCFA 4993 5382 1,0779091 5372 Public Investment inlivestock sub-sector MFCFA 1057 492 0,4654683 1636 Public Investment infishing sub-sector MFCFA 4968 1892 0,3808374 3238 Public Investments inwater sub-sector MFCFA 4987 4561 0,9145779 4870 Public Investments inother rural sub-sectors (forests) MFCFA 1685 1476 0,8759644 1340 PublicInvestmentinthe rural sector MFCFA 29628 19417 0,6553598 29507 Public Investment in the rural sector % 0,1748996 0,1308891 0,1458937 PublicInvestments :All sources of financing combined Unit Budgeted Realized YO Budgeted TotalPublicInvestments 449903 298000 0,662365 46865 1 Public Investment inthe agricultural sub-sector MFCFA 62199 26624 0,4280455 63426 I n Which Irrigation MFCFA 18263 14558 0,7971308 24991 Public Investment in livestock sub-sector MFCFA 5659 3770 0,6661954 5663 Public Investment in fishing sub-sector MFCFA 11218 5278 0,4704938 9238 Public Investments inwater sub-sector MFCFA 18577 10166 0,5472358 19948 Public Investments in other rural sub-sectors (forests) MFCFA 13955 6554 0,4696525 12368 PublicInvestment inthe rural sector MFCFA 111608 66950 0,5998674 110643 Public Investment in the rural sector 0,2480713 0,2246644 0,2360883 FromPrivateFunds Unit Budgeted Realized YO Budgeted TotalOperationalBudget MFCFA 639,l 634,98 0,99 728,72 Operational Budget o f the agricultural sub-sector MFCFA 636 6,51 0,9863636 7,847 In Which Irrigation MFCFA 0,217 0,2 17 1 0,217 Operational Budget o f the livestock sub-sector MFCFA 1 4 1,77 0,9833333 1,234 Operational Budget o f the fishing sub-sector MFCFA 0,725 0,63 0,8689655 0,851 Operational Budget o f the water sub-sector MFCFA 0,61 0,59 0,9672131 0,63 Operational Budget o f the other rural sub-sectors (forests) MFCFA 394 3,3 0,9705882 5,758 OperationalBudget of the rural sector MFCFA 13,135 12,8 0,9744956 16,32 Operational Budget of the rural sector YO 0,0205523 0,0201581 0,022395L Source :DB/MEF Loi des Finances de 1998 a 2004 et Situation d'exkcution dubudget de 1998 a 2003 ; DCEF/MEF PTIP et Biland'exkcution projets et programmes de 1998a 2004 Guaranties Allowances 75% Guaranties Realized Allowance Allowances Years Unpaid unpaid Engagement gap credits 5,5% Engagements gap 199711998 2450 1837 400 -1437 12278 675 300 -375 107 Table 1V.f: Incidenceof revenue allowances for the 2001/2002 campaign GIE GIE GIED. GIEK. GIE Ndelle Mboltogne Diakhk Mari Dinguiraye Products 16 185 14375 16 890 7 230 3 850 6 665 I Variationon marein I +170 1 +577 I +228 1 +85 1 +54 I +95 1 % I +1.4% +5.1% +2.0% +l.6% +2.1% I +2.2% Period 1997198 1998199 199912000 2000101 2001/02 2002102 Total Credits Loans 12280 12426 13891 9201 11306 8440 67544 Required 12969 13125 14673 9719 11943 - 8915 71344 Unpaid 2450 2573 5135 4568 3718 3322 21766 Financingrate 83% 82% 57% 53% 63% 63% 70% UE 1 003 2 486 - 1337 I Beleium I 413 1 BAD 1500 IDA 3 560 I IDA I 6 153 I USAID Total 18 902 Source: DCEFlDDI 108 Table 1V.i :Evolutionof InternalFinancingof Health Table 1V.j:Distributionof Expensesinbillionsof CFA Regional 15 197 645 20 330 91C 23 742 380 31 110 565 41 880 71! 47 237 926 179 500 14: 51% Central 16464 116 17 032 031 28 471 290 30 401 9 7 ~34 266 035 45 385 458 172 020 901 49% TOTAL 31661761 37362941 5221367C 6151254: 76146755 92623 384 35152105: 100% Source: Ministryof Health Table 1V.k:Net Change inHealthInfrastructureand Staff (1998-2002) I #health I #health 1 # I #mid I #nurses/ #ICP I I# c o r n . I 109 Source: PLNP 110 Table 1V.n: Financingof NationalStrategicPlanAgainst HIV-AIDS 2002-2006 111 Table 1V.o: Distributionof CNIS IDA Financing2003-2004 Pubic Sector Civil Society Total Per Capita (CFA) Central 1,287,619,313 1,097,081,565 Regional 1,239,504,082 2,626,003,249 Total 2,527,123,395 3,723,084,8 14 627 Dakar 147,421,095 432,038,680 256 Source: CNLS 112 2003 2004 YOen 2004 Government 160 000 199 766 78% Local Communities 3713 I 3899 2% Households 35838 I 37630 15% Total 209 385 I 256733 100% I Total Expenses for Education as a % of GDP I 5,4% I 6,2% I 2003 2004 Operational Expenditures o f the Government 483 000 520 724 Public Operational Expenditures allocated to Education 149 100 182 498 Capital Expenditures o f Govt over internal financing 211800 236 000 Public Investment Expenditures allocated to Education 10 900 17 265 Total Public Expenditures allocated to Education 160 000 199 763 . Total Public Expenditures of Govt over internal financing 694 800 756 724 Education as a % o f Operational Expenditures o f the Government 30,9% 35,0% O f Which Investment Expenditures o f Govt 5% 7% Public Expenditures on Education as a % o f GDP 4,1% 4,9% B C I Total 518 495 560 556 148 356 383 453 591 169 4229 BIDTotal 65 120 170 145 500 JICA IV Total 48 107 55 58 55 323 OPEP I11 Total 30 209 239 PAC Total 163 9 0 8 7 31 218 PAOES Total 180 165 345 PEQT Total 465 269 301 565 1600 PNIR Total 7 6 17 4 10 13 4 5 8 74 Grand Total 794 647 1170 764 785 597 1079 1029 852 322 8039 113 Source: MOE, Bank estimates urban areas 2,42 0,26 1 0,88 Source: MOE Diourbel I 296 28 93% Fatick 456 188 41,2% I Kaolack I 503 I 272 I 54.1yo I Kolda 683 148 21,7% Louga 555 15 2,7% 1 Matam I 180 I 3 I 1.7% I Saint Louis I 366 145 39,6% Tambacounda I 560 490 87,5YO Ziguinchor I 261 139 53,3?'0 Total 4377 1489 34,0% Source: MOE 114 Figure1V.a: Concentration Curve of WFP SchoolLunches,by PovertyLevelof CR analyse de la concentrationdes bbneficiairesdes cantines scolaires du PAMselon le sexe et la zone de pauvretd 120.00 100.00 80.00 t G A F i C O N S cuml 60.00 cFILLESencum1 TOTAL cum1 --__ equalny 40.00 20.00 0.00 0.00 10.00 20.00 30.00 40.00 50.00 60.00 70.00 80.00 90.00 100.00 deciie de population Ministryheadquarters Unit 2 Rural town halls Unit 100 Urbantown halls Unit 95 Community centers Unit 70 Culverts, bridges Unit 47 Canals Km 150 Drains,culverts (70% rural) Km 207 Lighting Km 43 115 Program Institution - Total Budget ($US Zones millions National Rural Infrastructure Program Min.Agric. - $48.9 million All regions except (PNIR) (IDA, FIDA, AFD, Dakar Gov., beneficiaries) Groundnut BasinProgramme Kaolack and Fatick (Programme du Basin Arachidier PSA - KfW ex-PROCER) Program to Support Decentralization MinistryInt.and - $7.5 million 38 CR inLouga and- (Programme Appui Decent - PADMIR) LocalGovt. II (FENU,PNUD) Kaolack Local Development FundFonds IIMinister Charge - $4.0million Tambacounda and Dtveloppement Local FDL) o f Planning (FENU/PNUD) Kedougou Program to Support Local Development Min.Finance - $10.5 million 59 CR Initiatives (Programme Soutien Initiatives Ministry 1nt.and (EU) Kolda, Matam, Devel. Local - PSIDEL) Local Govt. Ziguinchor Project to Fight Against Poverty (Projet MFFDS - $21.5 million Tambacounda, Thies, de Lutte Contre la Pauvrete -PLCP) (BAD,FND) Kolda, Diorbel, Dakar Social Development FundAgency MFFDS - $30 million Louga, Fatick, Kolda, (Agence de Fond de Developpement - (IDA, Government, Kaolack, Dakar - Social - AFDS) beneficiaries) poorest CR PAREP MFDSSN - $4.5 million (PNUD) Table 1V.x :Per Capita Expenditures by Rural Poverty Zones of SelectedPrograms CR Index PNIR AFDS PSAOP PBA( PROCR) GLOBAL 0 6,05 1 12,524 206 0 18,781 I 5n I 2.374 I 3.431 I 57 I 846 I 6.708 I 116 Bibliography Akakpo, K.Mesure d'Impact du Programme d'Alimentation Scolaire du P A M dans les Ecoles Publiques Rurales au Senegal, May 2004, IJNESCO/BREDA for PAM. Anderson, Lykke, E .2005. "The Impact o f Aide on Recipient Behavior: A micro-level dynamic Analysis of Remittance, school, work, consumption, investment and social mobility in Nicaragua" Sixth Annual Global Development conference. Andvig, Jens, Christopher,200 1 `Family- Controlled child labor in sub-Saharan Africa- A survey o f research' World Bank SP Discussion paper No.0122 Beegle, K, R, Dehejia, and R, Gatti ,2003. "Child labor, income shocks and access to credit" World Bank policy researchworking paper 3075. Babacar 18, W o l f Angelika, Dr. Michael Niechzial? "Health insurance, a possible solution to reduce vulnerability o f the poor?" UniversitC Gaston Berger; St.Louis (Senegal), institute for Health Sciences and Management in Medicine, University o f Bayreuth. Barrett, C., and Readon T.2000. "Asset, Activity, and Income Diversification among African Agriculturalists: Some Practical Issues' USAID (United States Agency for International Development). Becher, H., 0.Muller, A. Jahn, A. Gbangou, G. Kynast-Wolf, B. Kouyate. 2002. "Risk Factors for Child Mortality: Survival Analysis o f 1,3 12 Children in Rural Burkina Faso" Control of Tropicla Infestious Diseases, SFB 544, Discussion Paper 02/2002. Biya, M Marius, 2003.'Atelier sur L'amClioration de la Connaissance sur les donnCes et les mCthodes de recherche sur la VulnCrabilitC et des Enfants' Ministkre de 1'Economie et des Fiances du SCnCgal, Banque Mondiale, Fond des Nations Unies pour L'enfance, Cellule de suivi du programme de lutte contre la PauvretC. Boccanfuso, Dorothee and Luc Savar, 2004 ` Analyse d'impact de la liberation de la filikre Arachide au sCnCgal. Un modkle d'dquilibre gCnCral calculable multi- mCnages' Banque Mondiale. Carrin, Guy.2002. Social Health Insurance in Developing countries: A continuing challenge' World Health Organization, International Social Security Review, Vo1.55. Christiansen, J. Luc and Subbarao, K. 2004 "Toward and Understanding o f Household Vulnerability in Rural Kenya" World Bank Policy Research Working paper 3326 .CNLS, 20004. "Bulletin Sero-EpidCmiologique No. 11 de Surveillance du VIH D, CNLS, Minisryo fHealth Senegal Cotula, Lorenzo and Toulmin, Camilla "Till and Tiller: International Migration, remittances and landrightinWest Africa. Cruces Guillermo and Quentin, Wodon `Risk-Adjusted Poverty in Argentina: Measurement and Determinants. World Bank: Discussion Paper Distributional Analysis Research Program No, DARP 72 the Toyota Center Cisse, Ibrahima, 2004. Mutuality, cultural Diversity and Development' GM Sosar Senegal. D e Janvry, Sadoulet et a1 `Organisations paysannes et dkveloppement rural au SCnCgal : Etude d'impact des CLCOP. Projet des services agricoles et Organisations de Producteurs. Octobre 2003. 117 Dia, Mamadou Aliou. 2004. "Earth observation and GIs-based flood monitoring inthe Senegal River Estuary and Valley" Floods and Erosions Working Papers IRD. Diaw, BA, Sey-nabou and Sarr Cheikh Tidiane.1997 ` Dynamiques locales et lutte contre la pauvretk. Guinaw Rails- Pikine : Programme DCveloppement Social. Diagne, Abdoulaye, and Gaye Daffe.2005.' Profil et caractiristiques de la croissance au SCnCgal' Confkrence Annuelle de la Banques Mondiale sur 1'Economie du dCveloppement. Dror, David, M. and Alexander, S. Preker? Social Reinsurance: a new approach to sustainable community health financing. International Labor Office. World Bank. Edwards, Cox, Alexandra and Ureta, Manuelita. 2003. "International Migration, Remittances and Schooling: Evidence from El Salvador' National Bureau o f Economic Research Inc. (NBER) working papers 9766. Etard, J.F., Kodio, B. and Ronsmans, C. "Seasonal Variation in Direct Obstetric Mortality in Rural Senegal: Role o f Malaria? ?"" Fall, Toure & Kane `les inondations dans la VallCe du fleuve Senegal: analyse de leur Cvolution et des moyens mis en oeuvre pour les contr6ler'. OXFAM-GB,2000. FAO, 2005. Rapport evaluation des recoltes et des disponibilitks alimentaires PAM/FAO/Gouvernement du SCnCgal. Faye, C. 2000. "Protection Sociale et PauvretC Rurale au SCnCgal" ILO, Dakar Faye, 0,N'Dao, 0,Camara, B, Soumara, M, Dieng, T, Bah, IB, Dieng, Y, Gaye, 0.N'Dir, 0, Diallo, S. Senegal 1993. "Management o f severe Malaria in Children in developing countries. A protocole for economic evaluation' Departement de Parasitologie, UniversitC Cheikh Anta Diop, FacultC de MCdicine et de Phamarcie. AM J Trop M e d Hygiene 48(6) 748-56. Faye, 0.Faye B, Dieng B, Faye C, N'Dir 0,Gaye 0,Diallo S 1996. "Informal Care o f Malaria: Determinants o f Demand and Supply Inventory: A Study Conducted in Touba City" Bulletin de Sociologic Pathologiaue Exotique 89( 1):35-40. Faye, 0.Konate L, Gaye 0,Fontenille D, Sy N,Diop A, Diagne M. 1998. "Impact o f the use o f permethnn pre-impregnated mosquito nets on malaria transmission in a hyperendemic village in Senegal" MCdCcine Tropicale 58(4): 355-60. Gertler, Paul, David I,Levine, and Enrico Moretti.2002. `Micro finance programs help families insure consumption against illness. Gour, Michel and Elitte Dupre, 2003 mises en place d'un rCgime de protection sociale agricole et rural au SCnCgal. Guarcello, LS,Lyon, F Rosati, 2004. "Orphanhood and Child Vulnerability: Senegal". UnderstandingCildren's Work (UCW), July 2004. Heitzmann, Karin, R. Sudharshan, Canagarajah and Paul, B, Siege1.2001. `Les sources de VulnCrabilitC: expose et ligne de conduite pour l'kvaluation des risques et les rCponses au risques' Banque mondiale, DCpartement de Politique Sociale, UniversitC d'Economie et D'entreprise de Vienne, DCpartement d'Economie et de Gestion, Tel Hai Academic collage, Israel. Hoddinott, John, and Agnes, Quisumbing.2003 `Methods for Micro econometric risk and Vulnerability assessments' World Bank, International Food Policy ResearchInstitute. IIED 1993 Natural Resource Management inNdamMor Fademba Senegal. Monograph 118 Jutting, J., and Tine, J. 2000. Micro insurance schemes and health care provision in developing countries: The impact of mutual health Insurance scheme in rural Senegal' Center for Development, researchpaper (PDF) Kakwari, Nanak, Anita Schwarz and Kalarihi, Subbarao, K.? `Living Conditions o f elderly in Africa and the Role of Social Protection. Kane, F. 1993. "Temporary Expatriationis Related to HIV-1 Infection inrural Senegal" ,AIDS no. 9( 1993):1261-65. Largard, E, Pison G and Enel, C.1966. "Knowledge, Attitudes, and Perception of ADIS in rural Senegal: Relationship to Behavior and Behavior Change. 10(3):32-34 Lagard, E, Pison, G, Enel, C, L e Guenno, and Seck, C 1993. "Seasonal Migration: A risk factor for HIV Infection in Senegal )) Journal of Acquired Immune DefSyndrome: 196-200. Lentz, Erin et Christopher, B, Barret, .2004 `Food Aid Targeting, shocks and private transfers among East African Pastoralist' Ligon Ethan and Schechter Laura. 2003, `Measuring Vulnerability, World Bank Marques, JosC, 2003. Social Safety Net Assessments from Central America: Cross-Country Reviewof Principal Findings, Social ProtectionDiscussion Series. World Bank. McCarthy, F., H.Wolf, and Y. Wu. 2000. Malaria and Growth. Policy Research Working Paper Number 2303. Washington, DC: The World Bank. Mills, Bradford, Carlo, Del Ninno and Harivelo, Rajemison.2003. `Commune Shock and HouseholdEconomies well-being inMadagascar'? Moen, Bruno, Veronica, Zeitilin, Codou Bop, Rokhaya Gaye.2004. `Study on the practice o f trafficking in person in Senegal' USAID (United States Agency for International Development) Mintech International, 2004. `Suivi de l'execution physique et financier des investissements public de 1'annCe2003. Rapport final Lot No. 2, projet No.2004 Moneti, Francesca? `Enabling women to address their priority health concerns. The role of community based systems and social protection. ILO (International Labor Office) Global campaign on social security and coverage for all working paper Morduch, Jonathan and Manohar, Sharma.2002. `Strengthening Public safety nets from the bottom up' World Bank. Social Protection discussion paper series No.0227. Ndiaye, A. ` Etude d'impact de la micro finance dans la pauvretC au Stntgal' BCEAO (Banque Centrale des Etats de 1'Afrique de l'ouest). Banque Mondiale, (CGAP) Group Consultatif Assistance aux plus Pauvres. Ndaiye 0,Hesran JY, Etard JF, Diallo A, Simondon F, Ward MN. 2001. ((Climate Variability and number o f deaths attributable to malaria in the Niakhar area Senegal from 1984- 1996", w,Jan-Feb; 11(1):25-33. PRAESC (Programme de Relance des ActivitCs Economique et Sociales en Casamance) Preker, S. Alexander, Guy, Carrin, David, Dror, Melitta, Jakab, William, Hsiao, and Dyna Arhim, 2001.' Role of Communities in Resource Mobilization and risk sharing HNP Discussion Paper, a Synthesis report. 119 Preker, S. Alexander, Guy, Carrin, David, Dror, Melitta, Jakab, William, Hsiao and Dyna Arhin- Tenkorang, 2002"`Effectiveness o f community health financing in meeting the cost o f illness' Bulletin o f the world Health Organization Ratha D., 2003. "Worker's Remittances: An Important and Stable Source o f External Development Finance", inWorld Bank Global Development Finance, 2003 Rawlings, Laura and Gloria Rubio 2005. "Evaluating the Impact o f Conditional Cash Transfer Programs" World Bank Research Observer, Volume 20, Number 1 Spring 2005 pp 29- 55. Robert, Vincent, Macintyre, Kate, Keating, Joseph, Trape, Jean Francois, Duchemin Jean Bernard, Mcwilson, Warren and Beier John C. 2003. "Malaria Transmission in Urban Sub-Saharan Africa" American Journal o f Tropical Medicine and Hygiene, 68(2), 2003, pp. 169-176. Rucci, Graciana, 2004. `Macro Shocks and schooling decision: The case o f Argentina" University o f California- Los Angeles, JEL classification 130,120,128 Sander, Cerstin and Issa Barro` Etude sur le transport d'argent des Cmigres au Senegal et les services de transfert en micro finance' Social finance programme. Document de travail No.40. She, Sock, Vilane, Cheikh, Anne CirC Ibrahima. 2002. `Etude relative a l'extension de la protection sociale a l'agriculture ))(ILO) version I. SCnegal, Ministere de la Jeunesse, Fonds National de Promotion de la jeunesse. Convention- Cadre d'agrkment, en qualit6 de structure financiere dCcentralisie au dispositif du Fonds National de Promotion de lajeunesse. SCnCgal, Ministere de la SantC, Direction de la Santt, Service National des Grandes EndCmies. Programme National de Lutte contre le Paludisme. StnCgal, 1997, Ministere du Travail et d'Emploi. Politique National de L'emploi, Plan de strategique pour faire reculer l e paludisme au SCnCgal. SCnCgal, 1998. Ministere du Travail et de L'emploi. Plan D'action pour L'emploi. Document Valide par le seminaire de la comite de suivi de la politique de l'emploi tenu Dakar et Soumis l'adoption du Conseil Interministhielsur L'emploi. StnCgal, 2000.Rapport de l'enquete sur les objectifs de la fin de dtcennie sur l'enfance. UNICEF. SCnCgal, 2001a. SociCtC National de la Croix-Rouge SCnCgalaise. Deuxieme plan quinquennal de developpement. SCnCgal, 2001b. Ministhe de la Jeunesse, Agence National pour l'emploi des jeunes. Plan stratigique de l'agence National pour L'emploi des Jeunes. SCnCgal, 2002a. Agence du Fond de DCveloppement Social Rapport de synthkse du sCminaire d'Hannonisation de MCthodologie des EPP. SCnCgal, 2002b. ` Convention d'appui a la cellule MEF pour la mise en place d'un dispositif de suivi de la stratCgie de rCduction de la pauvretC' AFDS (Agence du Fonds de DCveloppement Social) Ministhe de 1'Economie et des Finances. SCntgal, 2002c. Ministere de 1'IntCrieur Direction de la Protection civile `Carte rCgionales des risques et mesures de Prevention' Doc No.2DPC. 120 SCnCgal, 2002d.APAPS (Agence Pour la Promotion des ActivitCs de Population) Rapport gCnCral des commissions sur le vieillissement. SCnCgal, 2002e. Document de stratkgie de rCduction de la PauvretC. SCnCgal, 2002f. Ministere du DCveloppement Social et de la solidarite Nationale. L a Direction de 1'Action sociale et les centres de Promotion et de Reinsertion Sociale. SCnCgal, 2002g. Ministere de 1'Economie et des Finances, Direction de la CoopCration Economique et Financikre. Budget Consolide D'investissement Gestion 2003. SCnCgal, 2003a. Collectif des Organisation de la sociCtC civile pour le suivi de DSRP `Rapport final de l'atelier national de la sociCtC civile sur les filets de sCcuritC sociale et la poursuit de programme de suivi duDSRP. SCnCgal, 2003b. Ministhe de la famille et de la solidaritk National, service de l'administration gCnCrale et de 1'Cquipement. Assistance au sinistre suite aux inondations SCnegal, 2003c .Ministkre de l'hgriculture et de L'hydraulique. Programme National d'hfrastructures rurales. Aide mCmoire : Etude d'impact du PNIR sur la RCduction de la PauvretC en milieurural Senegal, 2003d. Participatory Poverty assessment, Niger Office o f the president o f the republic, Office o f the Prime Minister, Permanent Secretariat o f the PRSP, cabinet National d'expertise en sciences sociales SCnCgal, 2003e. Ministkre de la Famille, du DCveloppement Social et de la SolidaritC National. Evaluationfinale duplan d'action national de lafemme' SCnCgal, 2003f. Ministkre de 1'Economie et de Finances, Direction de la Cooperation Economique et Financiere. Suivi physique et financier des projets du programme triennal d'investissements public. Secteur Quaternaire B.C.I. SenCgal, 2003g. Ministere de la Fonction Public du Travaille, 1'Emploi et organisation professionnel. Rkcapitulation gCnCrale des crCdits des titres par section, Projet de L o i des finances pour 2003. SCnCgal, 2003h. .Ministkre de la Jeunesse, Agence National pour L'emploi des Jeunes. Bilan des rkalisations 2003 et 2004 SCnegal, 2003i. Ministere de la Jeunesse, RCcapitulation gCnCrale des crCdits des titres par section. Projet de loi de finances pour 2003 SCnCgal, 2003j. .Ministere de 1'Agriculture et de 1'Elevage. RCcapitulation gCnCrale des crCdits des titres par section. Projet de L o i des finances pour 2003. . Senegal, 2003k. .Ministkre de 1'Environnement et de 1'Assainissement, Recapitulation gCnCrale des crCdits des titres par section, Projet de l o i de finances pour 2004. SCnCgal, 20031. .Ministkre de 1'Education. Programme DCcennal de 1'Education et de la formation. Evaluation des rendements des Clkves au CP et au CE2 SCnCgal, 2003m. .Ministhe de la famille, du DCveloppement Social et de la SolidaritC Nationale. Budget 2004, Rapport de prksentation. SCnCgal, 200311. Ministere de l'environnement et la protection Nature, RCcapitulation gCnCrale - des crtdits des titres par section. Projet de L o ides finances pour 2003 SCnCgal, 2004a..Ministre de 1'Cconomie et des finances `Rapport d'avancement de la mise en aeuvre du DSRP, cellule de suivi du programme de lutte contre la pauvretC 121 Senegal, 2004b. Ministry o f health and Medical Prevention, CquitC et MutualitC au Senegal, University o f Montreal, HYGEA cabinet d'Ctudee, IDRC SCnCgal, 2004c. Ministkre de la famille du DCveloppement social et de la SolidaritC Nationale SenCgal, 2004d.Ministkre du developpement Social et de la SolidaritC National: Rapport Village de Sorom phase pilote, Version final. SCntgal, 2004e. Ministhe de 1'Economie et des finances, L a PauvretC au SCnCgal : de la dkvaluation de 1994-02' Direction de la PrCvention et de la Statistique version prkliminaire. StnCgal, 2004f. Ministere de 1'Cconomie et des Finances- Protection des populations du SCnCgal. Directionde la prCvision et de la statistique. SCnCgal, 2004g.Ministere de la santC et de la PrCvention. EquitC et MutualitC au SCnCgal Etude qualitative, donnCe de base. SCnCgal, 2004h. Ministkre de 1'Economie et des finances, Direction de la CoopCration Economique et Financikre. Suivie de I'exCcution Physique et financikre des Investissements Public de l'annee2003. Rapport final secteur primaire. SCnCgal, 2004i. .Ministhe de la Famille, du DCveloppement Social et de la SolidaritC Nationale. Rapport sur 1'Etat d'avancement des activitCs du PLCP. Projet de lutte contre la pauvretC. SCnCgal, 2004j. Ministere de la famille, du DCveloppement social et de la SolidaritC Nationale, Direction de L'action Sociale et de la SolidaritC Nationale. ElCment de stratkgie de la politique de Protection Sociale du Ministhe de la Famille du DCveloppement Sociale et de la SolidaritC Nationale. Senegal, 2004k. .Ministkre de la famille, du DCveloppement social et de la SolidaritC Nationale, Direction de L'action Sociale et de la SolidaritC Nationale. Plan d'action de la Direction de 1'Action Sociale et de la solidaritk Nationale (2005-2009). SCnCgal, 20041.Ministkre de la famille, du DCveloppement social et de la SolidaritC Nationale, Direction de L'action Sociale et de la Solidaritt Nationale. Politique d'action sociale en Faveur des groupes sociaux vulnkrables. Senegal, 2004m.Ministkre de la famille, du DCveloppement Social et de la SolidaritC Nationale. Cellule de suivi opkrationnel des projets et Programmes de lutte contre la PauvretC. Contribution du Ministhe de la famille, du DCveloppement Social et de la SolidaritC Nationale a la lutte contre la PauvretC et la protection des personnes VulnCrables. SCnCgal, 2004n .Ministkre de la famille, du DCveloppement social et de la SolidaritC Nationale, Direction de L'action Sociale et de la SolidaritC Nationale. Projet de crCation de la Maison des Solidarites SCnCgal, 20040. Ministhe de la Jeunesse. DCpense du Personnel, Projet de loi de finances pour 2004. SCnCgal, 2004p. Ministere de 1'Entreprenariat FCminin. RCcapitulation gknerale des credits des titres par section, RCcapitulation gCnCrale des crCdits des titres par section. Projet de L o i des finances pour 2004. Sey, Haddy Jatou, and Wendy Wilson. 2001 `Social Assessment: perception o f poverty, Local Definition o f Decline, Growth and Development. The World Bank AFDS program. Skoufias, Emmanuel, 2002 `Rural Poverty Alleviation and Household Consumption smoothing evidence from PROGESA inMexico' International Food Policy Research Institute. 122 Squire, Chris and Samuel, Braima, J Social Protection Study: Risk and Vulnerability Assessment' World Bank Subbarao, K.2001. `Systemic Shocks and Social Protection: Role and Effectiveness o f Public Work Programs. World Bank Region Human Development working Paper Series Subbaroa, K.Angel Matttimore, Kathrin, Plangermann `Social Protection o f Africa's Orphan and other Vulnerable Children' World Bank- Africa region Human development paper. Subbaroa, K. `Systemic shock and Social protection; Role and effectiveness o f public works program' World Bank. Tasliuc, Emil, D and Kathy Lindert.2002.' Vulnerability: A Quantitative and qualitative Assessment' LCSHD. Tanguy, Bernard, Maguelonne, Chanron and Mouhamadou Lamine Dia1.2002. `Analyses de l'impact du renforcement des capacitCs des organisation paysannes au SCnCgal Rapport des Ctudes de cas. Thiam, Macoumba, Perry Rebecca, PichC.2003. "Migration and HNinNorthern Senegal. Thomas, Madeleine and Stephen Connor. 2001. "The Development o f Malaria early Warning Systems for Africa' M A L S A T Research Group, Liverpool School fo Tropical Medicine, Liverpool UK1. Trape, JF, E Lefebvre-Zante. F Legros, P. Druilhe, C. Rogier, H. Bouganali, D. Salam. Senegal 2003 ((Malaria Morbidity among children exposed to low seasonal transmission in Dakar, Senegal and its implication for malaria control in tropical Africa" Laboratoire de Paludologie et Programme Urbanization et Sunti, ORSTOM. Am .J. Trop. Med. Hyg., pp.169-176. Velyvia. "The Sexual Networks o f Migrant Serere Women and the Spread o f HN in Senegal" CDE Working Paper # 98-26. Warning, Mathew and Elisabeth, Sadoulet. 1998. "The performance o f Village Intermediaries in Rural Credit Delivery under Changing Penalty Regime: Evidence from Senegal" Journal of development Studies, 35(1): 115-138 Wampfler, Betty, 2000. `The financing o f family farming in the context o f liberalization: what can be the contribution o f Micro finance?' Article published in Techniques Financiers et DCveloppement No.59 White, Jeffrey.2000. "Rural Transition: Agricultural Development and Tenure Right a Case Study inthe Senegal River Valley ))African Studies ProgramPaper, Spring 2000. World Bank, 1999. Rural development note.1999. Policy operation for employment in the rural non-farm sector. World Bank, 2001. Social Safety Net assessment: A toolkit for Latin America and the Caribbean. Human Development department, Social Protection Group Latin America and the Caribbean Region Office. World Bank, 2003a. Costa Rica: Social Spending and the Poor. Report 24300-CR, Humand Development Unit, Latin America and Caribbean Region. World Bank, 2003b. `Ethiopia; Riskand Vulnerability' Assessment report. World Bank.2003~.Burkina Faso risk and vulnerability assessment: Human Development unit, AFTH2. 123 World Bank.2003d. .Poverty and Human Development. Report No.26189-Sn World Bank, 2004a. Senegal Public Expenditure Review. Report No. 29357- Sn World Bank, 2004b.' International Development Association Program Document for a proposed private sector adjustment credit. ReportNo.25943. World Bank, 2005.' Beyond survival; Protecting household from the impoverishing effect of health shocks. World Food Programme, 2004. Analyse de la Vulnkrabilitk 1'Inskcuritk Alimentaire dans les zones affectkes par le Pkril Acridien,VAM UnitDakar. December 2004. Worralla, Eva, Basub, Suprotik and Hanson Kara. London 2002. `Ensuring that malaria control interventions reach the poor' London school o f hygiene and Tropical Medicine, World Bank Washington D.C. Woynar, Skbastien, 2004. Dkcentralisation du dkpistage, du traitement et du suivi psychosocial et politique pharmaceutique des ARV. Banque Mondiale. Yamano, Takashi, Horold Alderman, Luc Christiaesen.2003 `child growth, shocks and food aid inrural Ethiopia.WorldBankPolicyresearchpaper 3128 Zeller, Manfred and Manohar, Sharma.2000.' Many borrow; more save and all insure implications for food and microfinance policy' (IFPRI) International Food Policy Research Institute. Zimmermann, J. Frederick and Michael, R. Carter. 2002. `Asset smoothing, consumption smoothing and reproduction o f inequality under risk and subsistence constraints' Journal of Development Economics. 124