AGRICULTURE GLOBAL PRACTICE TECHNICAL ASSISTANCE PAPER 96296 SENEGAL AGRICULTURAL SECTOR RISK ASSESSMENT Stephen D’Alessandro, Amadou Abdoulaye Fall, George Grey, Simon Simpkin, and Abdrahmane Wane WORLD BANK GROUP REPORT NUMBER 96296-SN AUGUST 2015 AGRICULTURE GLOBAL PRACTICE TECHNICAL ASSISTANCE PAPER SENEGAL Agricultural Sector Risk Assessment Stephen D’Alessandro, Amadou Abdoulaye Fall, George Grey, Simon Simpkin, and Abdrahmane Wane © 2015 World Bank Group 1818 H Street NW Washington, DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org Email: feedback@worldbank.org All rights reserved This volume is a product of the staff of the World Bank Group. The findings, interpretations, and conclusions expressed in this paper do not necessarily reflect the views of the Executive Directors of the World Bank Group or the governments they represent. The World Bank Group does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of the World Bank Group concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this publication is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applicable law. World Bank Group encourages dissemination of its work and will normally grant permission to reproduce portions of the work promptly. For permission to photocopy or reprint any part of this work, please send a request with complete information to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA, telephone: 978-750-8400, fax: 978-750-4470, http://www.copyright.com/. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, World Bank Group, 1818 H Street NW, Washington, DC 20433, USA, fax: 202-522-2422, e-mail: pubrights@worldbank.org. Cover photos from left to right: 1. A young Senegalese herder. IFPRI/Milo Mitchell 2. Climate smart farming practices in Senegal. M. Tall/CCAFS 3. Irrigation pipe. S. Kilungu (CCAFS) CONTENTS Acronyms and Abbreviations vii Acknowledgments ix Executive Summary xi Chapter One: Introduction 1 Methodology 2 Rationale 3 Chapter Two: Agricultural Systems in Senegal 5 Agriculture Sector in Senegal 5 Agro-Climatic Conditions 6 Land and Water Resources 6 Agro-Ecological Zones 8 Crop Production Systems 8 Livestock Production Systems 12 Agricultural Markets and Trade 13 National Agricultural Policy 14 Food Security 15 Key Growth Constraints and Trends 16 Chapter Three: Agriculture Sector Risks 17 Production Risks 17 Market Risks 25 Enabling Environment Risks 31 Chapter Four: Adverse Impacts of Agricultural Risks 35 Conceptual and Methodological Basis for Analysis 35 Production Risks 36 Impacts of Production Risks 39 Summary of Impacts 40 Chapter Five: Vulnerability Analysis 41 Stakeholder and Livelihood Risk Profiles 42 Income Levels 44 Chapter Six: Risk Prioritization and Management 45 Risk Prioritization 45 Agricultural Risk Management 46 Risk Management Solutions 49 Prioritization of Risk Management Measures 54 Conclusion 56 References 57 Agricultural Sector Risk Assessment iii Appendix A: Climate Change Impacts Analysis 61 Appendix B: Senegal Vulnerability Analysis 67 Appendix C: Agricultural Insurance in Senegal 71 Appendix D: Crop Yield Loss Analyses 77 Appendix E: Crop Pest Analysis 81 Appendix F: Rationale for Risk Assessment Methodology 85 Appendix G: Analysis of Weather Risk Events 89 Appendix H: Crop Production and Yields 91 BOX BOX 3.1: Case Study: Contagious Bovine Pleuropneumonia (CBPP/PPCB) in Senegal 24 FIGURES Figure ES.1: Timeline of Major Shocks to Agricultural Production in Senegal (2004–06 = 100), 1980–2012 xii Figure ES.2: Impact and Frequency of Major Agricultural Risks in Senegal, 1980–2012 xii Figure 1.1: Agricultural, Value Added (annual % growth), 1980–2013 2 Figure 1.2: Agricultural Sector Risk Management Process Flow 3 Figure 2.1: Climatic Zones of Senegal 7 Figure 2.2: Average Annual Rainfall by Region, 1980–2013 7 Figure 2.3: Agro-Ecological Zones in Senegal 8 Figure 2.4: Trends in Crop Production Area Harvested (thousand ha), 1980–2012 9 Figure 2.5: Share of Livestock Units in Senegal, 1961–2012 12 Figure 2.6: Growth in Poultry Production (thousands), 1997–2011 13 Figure 2.7: Trade in Cash Crops (in MT), 2002–11 14 Figure 2.8: Trade in Staple Crops (in MT), 2002–11 15 Figure 2.9: Retail Prices for Key Staple Crops (CFA/kg), 2001–13 16 Figure 3.1: Variations in Temperature and Rainfall, 1900–2009 18 Figure 3.2: Precipitation Patterns of Major Regions, 1978/79–2008/09 18 Figure 3.3: Climate Variability Map for Senegal 20 Figure 3.4: Nominal Producer Prices for Key Staple Crops (CFA/kg), 2000–13 26 Figure 3.5: International vs. Domestic Groundnut Oil Prices, 1984–2013 27 Figure 3.6: International vs. Domestic Cotton Prices, 1984–2013 27 Figure 3.7: International vs. Domestic Maize Prices, 1984–2013 28 Figure 3.8: Dahra Market Livestock Prices (CFA/head), 2005–10 29 Figure 3.9: Goats vs. Cereals Terms of Trade, 2005–12 30 Figure 3.10: Historical Exchange Rates, 2001–12 30 Figure 4.1: Timeline of Major Shocks to Agricultural Production in Senegal (2004–06 = 100), 1980–2012 36 Figure 4.2: Indicative Losses from Risk Events to Groundnut Production, 1980–2012 38 iv Senegal Figure 4.3: Proportional Impact of Various Adverse Risk Events by Crop, 1980–2012 39 Figure 4.4: Frequency and Cumulative Impact of Various Adverse Risk Events, 1980–2012 40 Figure 6.1: Integrated Risk-Layering Solutions 47 Figure A.1: Rain-Fed Maize Yield Change under Four Climate Models, 2010–50 (IPCC A1B Scenario) 63 Figure A.2: Rain-Fed Groundnuts Yield Change under Four Climate Models, 2010–50 (IPCC A1B Scenario) 64 Figure A.3: Rain-Fed Rice Yield Change under Four Climate Models, 2010–50 (IPCC A1B Scenario) 65 Figure B.1: Senegal Livelihood Zone Map 69 Figure B.2: Food Price Shocks by District 70 Figure D.1: Groundnuts, 1980–2012 77 Figure D.2: Cotton, 1980–2012 77 Figure D.3: Maize, 1980–2012 78 Figure D.4: Rice, 1980–2012 78 Figure D.5: Cowpea, 1980–2012 78 Figure D.6: Potato, 1980–2012 79 Figure D.7: Tomato, 1980–2012 79 Figure D.8: Onion, 1980–2012 79 Figure D.9: Green Bean, 1989–2012 80 Figure D.10: Mango, 1989–2012 80 Figure H.1: Millet Production, 2003–12 91 Figure H.2: Sorghum Production, 2003–12 91 Figure H.3: Cowpea Production, 2003–12 92 Figure H.4: Rice Production, 2003–12 92 Figure H.5: Maize Production, 2003–12 92 Figure H.6: Groundnut Production, 2003–12 92 Figure H.7: Cotton Production, 2003–12 93 Figure H.8: Onion Production, 2002–11 93 Figure H.9: Tomato Production, 2002–11 93 Figure H.10: Potato Production, 2003–12 93 Figure H.11: Mango Production, 2002–11 94 Figure H.12: Green Bean Production, 2002–11 94 TABLES Table 2.1: Senegal National and Agricultural Statistics, 2012 6 Table 2.2: Trends in Crop Production, 1980–2012 9 Table 3.1: Variability of Rainfall by Region 18 Table 3.2: Frequency and Impact of Rainfall Events by Region, 1980–2013 21 Table 3.3: Major Livestock Diseases 24 Agricultural Sector Risk Assessment v Table 3.4: Inter-Annual Crop Price Variability, 1991–2011 25 Table 3.5: Vaccination Coverage in Senegal, 2009 33 Table 4.1: Cost of Adverse Events for Crop Production, 1980–2012 37 Table 4.2: Coefficients of Variation for Crop Production, 1980–2012 37 Table 4.3: Indicative Losses for Major Crops, 1980–2012 38 Table 4.4: Dates and Frequencies of Agricultural Risk Events 39 Table 5.1: Frequency of Risk Events and Percentage Affected 42 Table 6.1: Risk Prioritization Matrix 46 Table 6.2: Listing of Priority Risks by Commodity 47 Table 6.3: Proposed Risk Management Mechanisms 48 Table 6.4: Filtering Criteria for Risk Management Solutions 55 Table B.1: Prevalence of Food Insecurity by District 68 Table B.2: Vulnerable Groups 70 Table C.1: Insurance Premium Components 74 Table E.1: Rice: Preharvest 81 Table E.2: Rice: Postharvest 82 Table E.3: Sorghum: Preharvest 82 Table E.4: Sorghum: Postharvest 83 Table E.5: Millet: Preharvest 83 Table E.6: Millet: Postharvest 83 Table E.7: Cowpea: Preharvest 84 Table E.8: Cowpea: Postharvest 84 Table G.1: Frequency of Low Rainfall Events by Region, 1981–2010 89 Table G.2: Frequency of High Rainfall Events by Region, 1981–2010 90 vi Senegal ACRONYMS AND ABBREVIATIONS Acronym Definition Acronym Definition ARMT Agricultural Risk Management Team Ha Hectare CAADP Comprehensive Africa Agriculture IFPRI International Food Policy Research Institute Development Programme IPM Integrated Pest Management CBPP Contagious Bovine Pleuropneumonia ISRA Institute Sénégalais de Recherches Agricoles CA Conservation Agriculture LSD Lumpy Skin Disease CFA CFA franc LGP Length of Growing Period CSE Centre de Suivi Ecologique MEF Ministry of Economy and Finance CNASS Compagnie Nationale d’Assurances Agricole MT Metric tons du Sénégal NAPA National adaptation programme of action DAPS Direction de l’Analyse de la Prévision et des NCD Newcastle Disease Statistiques NDVI Normalized Difference Vegetation Index DNCB Dermatose Nodulaire Contagieuse Bovine OIE Organisation International des Epizooties. DPC Directorate of civil protection PPCB Pleuropneumonie Contagieuse Bovine DRR Disaster risk reduction PPP Public-private Partnership FAO Food and Agriculture Organization of the UN PPR Peste de Petits Ruminants FAOSTAT FAO Statistics SODEFITEX Société de Développement et des Fibres FMD Foot and Mouth Disease Textiles GCM General Circulation Model TLU Tropical Livestock Units GDP Gross Domestic Product USAID U.S. Agency for International Development G-8 Group of Eight WFP World Food Programme GOS Government of Senegal XAF CFA franc Agricultural Sector Risk Assessment vii ACKNOWLEDGMENTS To better understand dynamics of agricultural risks and identify appropriate responses, incorporate agricultural risk perspective into decision making, and build capacity of local stakeholders in risk assessment and management, the Agricultural Risk Manage- ment Team (ARMT) of the Agriculture and Environment Services Department of the World Bank is conducting an agricultural sector risk assessment in Senegal. The cur- rent report was developed by a team led by Stephen D’Alessandro and comprising Amadou Abdoulaye Fall, George Grey, Kersten Hell, Traci Johnson, Simon Simpkin, Kilara Suit, and Abdrahmane Wane. The team is especially grateful to Aifa Fatimata Ndoye Niane and Vikas Choudhary of the World Bank for their support and inputs throughout the activity’s planning, fieldwork mission, and report preparation. The team would also like to extend its appreciation to Maguette Diop Ndiaye of the Ministry of Economics and Finance, the Ministry of Agriculture and Rural Equipment, other government representatives, farmers, market traders, and all those who shared their perspective and insights, which provided the basis for this study and its findings. This activity would not have been possible without the generous contributions from USAID, Ministry of Foreign Affairs of the Government of the Netherlands and State Secretariat for Economic Affairs (SECO) of the Government of Switzerland. Agricultural Sector Risk Assessment ix EXECUTIVE SUMMARY Senegal’s agricultural economy today accounts for roughly one-sixth of national gross domestic product (GDP), down from nearly one-quarter in the mid-1980s. Although sector output has expanded by 70 percent over the past 30 years, population growth has quadrupled. During this period, successive government policies have promoted intensification of crop and livestock production via supportive policies and public investments.1 And yet, growth has been lackluster amid limited take-up of improved seeds and fertilizer consumption that remains among the lowest in the region. A major limiting factor has been widespread reluctance among the millions of smallholder farmers in Senegal who dominate production to assume the risks associated with increased productivity. With only limited capacity to manage these risks, highly vulner- able farmers choose to limit their exposure by limiting their outlays. Moreover, unman- aged risks have a profound impact on sector performance. A sound understanding of the risks faced by farmers and other agricultural sector stakeholders enables develop- ment of risk management systems that can at once support new productivity invest- ments, strengthen resilience, reduce losses, and drive sector growth. This agricultural risk assessment study was undertaken to provide a review of produc- tion, market, and enabling environment risks facing farmers and other stakeholders across Senegal’s agriculture sector. The report has been compiled with extensive anal- ysis of crop and livestock production, price, and meteorological data records over the period 1980–2012. It includes a review of key documentary evidence of yield and risk events together with input from interviews held with farmers, traders, processors and others in rural Senegal as well as with government and agricultural research staff between March and May 2014. The results of the analysis are considered in the light of the vulnerability of the different stakeholders to the effects of ex post shock events and the resulting ex ante impact upon investment. The most salient issues and results of this analysis are outlined in the text of the report. A considerable volume of sup- porting data is supplied in the appendixes, including (1) an analysis of cumulative rainfall during 1980–2013; (2) an assessment of levels of vulnerability among key 1 GOS expenditures on agriculture (as a percentage of total expenditure) have exceeded 12 percent on average during the 10-year period 2000–10, well above the 10 percent commitment under NEPAD’s CAADP framework. Agricultural Sector Risk Assessment xi FIGURE ES.1. TIMELINE OF MAJOR SHOCKS TO AGRICULTURAL PRODUCTION IN SENEGAL (2004–06 = 100), 1980–2012 180 Crop production index Food production index Livestock production index 160 140 120 100 Late/erratic 80 Locusts, rainfall; 2004 locusts, Erratic Locusts, 2011 60 Locusts, Late rains, rainfall; 1988 1992 regional Severe birds, 40 Severe droughts, drought; 2007 Severe 1996-98 cold rains; drought; drought, locusts, 20 locusts, 1980 2002 1983-84 0 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 Source: World Development Indicators livelihood groups; (3) a review of current agricultural FIGURE ES.2. IMPACT AND FREQUENCY OF insurance initiatives and market development options; MAJOR AGRICULTURAL RISKS and (4) an analysis of probable climate change impacts on IN SENEGAL, 1980–2012 crop production systems. The report’s principal findings, 1000 conclusions, and recommendations are summarized below. 800 in US$ millions 600 Figure ES.1 depicts a historical timeline of the most nota- 400 ble risk events that adversely affected sector performance 200 during the period under review. At the national level, the 0 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 analysis highlights agricultural production and livelihood –200 Frequency systems that are highly vulnerable to downside risks. Locusts Birds Erratic rainfall/drought Flooding Armyworm These most notably include erratic rainfall and drought as a more extreme but less frequent expression of the same phenomenon. Severe drought, especially in northern regions, emerges as the biggest risk in terms of estimated review period. The results of trend analyses indicate that aggregate losses to crop and livestock—a one in every four for the 12 crops analyzed, the loss of production over the years event on average over the review period. The analy- period was approximately 4.82 million metric tons (MT), sis also suggests a corresponding increase in the frequency with an estimated value of US$1.40 billion, or 3.9 percent of floods over the same period, but with limited aggregate of agricultural GDP on an average annual basis (figure impact on agricultural supply chains. After drought, locust ES.2). It is worth noting that this reflects only the ex post outbreaks are the second most common and costly risk impact. The ex ante impact may be of equal magnitude affecting agricultural production. Other notable risks or even larger. Although the average annual impact of include price volatility and other crop pests. shocks on GDP is relatively modest (less than 4 percent), actual impacts when they occur can result in losses of the Since 1980, the agricultural sector has been subject to at order of 10 to 20 percent of sector GDP. According to the least 11 major production shocks, with a frequency of analysis, Senegalese agriculture is subject to losses exceed- every three to four years on average across the 33-year ing 10 percent of gross production value in one out of xii Senegal every five or six years on average due to unmanaged risks. diseases are one of the main risks to agricultural produc- Erratic rainfall and/or drought account for approximately tion, most farmers do not have the knowledge or the 50 percent of crop yield reductions. Pests and diseases, financial means to adequately tackle crop pests. especially locusts, account for a further 25 percent. For livestock production, the increasing unpredictability of rainfall is a notable risk. However, erratic rainfall or PRODUCTION RISKS even geographically limited drought is a risk among pas- toralists only in the event that they are unable to migrate The most important factor contributing to crop and to more favorable pastureland. It is widespread drought, livestock production risk in Senegal is weather (figure severely limiting the carrying capacity of the entire ES.2). The key aspect of weather risk is that due to regional grazing area, which constitutes the biggest risk moisture stress caused either by erratic rainfall, early faced by livestock producers. Cold rainfall can have an cessation of rains, delayed onset of rains, or extended equally devastating impact, although both occur relatively drought. Even in the absence of these specific condi- infrequently. Another noteworthy risk for livestock pro- tions, it has been shown that more than 40 percent of ducers are bushfires, which according to some estimates the variation in national crop yields can be ascribed sim- damage as much as 6 percent of the potential dry season ply to the variation in annual rainfall amounts. From an grazing area and destroy on average 3.8 million MT of agricultural perspective, the geographic extent of other biomass each year. aspects of weather (wind, floods, and hail) is so limited that they have no discernible impact on aggregate, Livestock disease in general was highlighted as a risk in national-level yields. fewer than one of five meetings conducted, but individual diseases were mentioned more frequently. In particular, The impact of historical and future climate change on producers noted losses due to Rift Valley fever, highly rainfall amounts and intensities in Senegal is uncertain. pathogenic avian influenza, and Newcastle disease. These Historically, national rainfall data suggest that rainfall three diseases are considered to be among the three most amounts were decreasing until 1990, but since then, important livestock production risks. Avian influenza and annual cumulative levels show an increasing trend. Rift Valley fever are also considered as market risks as they Although aspects of climate contribute substantially to both can have a large impact and influence on both local the risk faced by producers, the anticipated impact of cli- and international trade. mate change upon that risk appears to be uncertain, the most consistently predicted trend being an increase in the Notably, other diseases such as foot and mouth disease variability and intensity of rainfall amounts. (FMD) and contagious bovine pleuropneumonia (CBPP) could be considered as major risks if the response mecha- The three most significant crop pests are the Senegalese nisms required for control and eradication were actually grasshopper or “sauteuriaux” (Oedaleus senegalensis), locusts put in place. This would require the quarantining and (Locusta migratoria), and birds (mainly quelea finch). The slaughter of whole herds, which would have a major first two are non–crop specific whereas the third is con- impact on the whole livestock industry. Currently, how- fined mainly to sorghum and millet (although maize can ever, the government implements less extreme control also be affected). During the 33-year review period, there measures and although there are some losses in productiv- have been recurrent locust invasions in Senegal, with sig- ity, the impact is relatively limited. nificant impact on both cash and food crops, and also affecting livestock production through loss of grazing. Damage can be highly localized, but large swarms can affect vast tracts of land. Damages due to locusts in 2004 MARKET RISKS were estimated at 2 million tons of crops, equivalent to 20 Among market risks, price volatility for both food and percent of the population’s food needs in the Sahel region. cash crops and livestock was assessed through the statisti- Although all of the stakeholders confirmed that pests and cal analysis of both domestic and international time-series Agricultural Sector Risk Assessment xiii data. The analysis found considerable variability of components, notably corn and soya, which contribute domestic food crop prices together with limited variability 80 percent of poultry feed, is considered a major source of domestic cash crop prices. International prices of rice, of risk. maize, groundnut oil, and cotton were more variable, with coefficients of variation exceeding 40 percent in some cases. The analysis suggested that companies that process ENABLING ENVIRONMENT locally purchased commodities for export (that is, cotton, RISKS groundnuts) face a significant price risk because the When ranked in terms of impact and frequency, a key risk domestic purchasing price may vary independently of the noted within the livestock sector is derived from uncer- export price, as a result of both the local price setting tainty over land tenure and access. As noted above, access mechanism. Although exchange rate fluctuations can also and mobility are critical to pastoral livelihoods. The anal- contribute to price risk for exporters of locally purchased ysis indicates that inconsistent policy making and imple- products, the exchange rate of the CFA franc (XAF) to the mentation of regulations can weaken traditional coping U.S. dollar has shown only modest variability over the mechanisms and increase vulnerability levels among past 12 years. extensive pastoralist communities, particularly in the north where land use pressures are increasing. Similar The impact of price risk varies substantially according to uncertainty is derived from the inconsistent delivery of the crop and its importance to the rural economy. The animal health services, including the enforcement of poli- price of staple crops at this time is critical to household cies on vaccination, quarantine, and movement. food security and the risk that increased food prices might reduce the accessibility of food has a substantial Where conflict occurs, unrestrained by the rule of law, then impact on household resource management. It is the the impact of risk, both ex post and ex ante, is considerable. profoundly negative ex post impacts of these fluctuations Such risk has been widespread in Casamance, where grow- upon nutrition, health, and survival that result in staple ers limit both the area and the level of investment applied food price shocks as being listed as the highest priority to crop production. Impacts on wealthier livestock owners risk faced by rural households. Because few producers who own larger herds can be substantial, with more than grow cash crops without first securing their own supply 60 percent losses being reported in some cases. Conflict and through staple crop production, the ex post risks to nutri- tensions between herders and crop growers, particularly in tion, health, and survival caused by fluctuations in cash the north, were highlighted as risks by several interlocutors crop prices tend to be less pronounced. As a result, ex during the course of the study. In addition, anecdotal evi- ante risk impacts are also reduced. This may be less the dence would suggest that the conflict in northern Mali has case for producers of horticultural crops who are exclu- destabilized Senegalese livestock markets in recent years sively oriented toward the market and who are often and contributed to higher levels of price volatility. exposed to higher levels of price volatility. More gener- ally, price risk for cash crops is visited mainly upon the buyers of commodities, although price fluctuations can VULNERABILITY contribute significantly to the risks faced by all stakehold- Understanding levels of exposure to different risks and of ers in cash crop subsectors. mitigation and coping capacity among the various liveli- hood groups can help decision makers better target inter- Traditionally, the limited reliance of pastoralists upon ventions. Among livelihood groups, nomadic pastoralists markets implied a limited impact of price risk upon pas- manage weather risks by continually moving to fresh graz- toral livestock production, but this situation is changing. ing grounds. However, growing land pressures place Livestock prices often plummet while food prices strains on their mobility and their access to sufficient graz- increase; this is now a common shock-induced pattern ing and water, and thus their capacity to cope. Agro- in dry lands and a major risk for livestock owners. Within pastoralists tend to be among the poorest households, the poultry sector, volatility in price of imported feed which typically lack the resources needed to absorb shocks, xiv Senegal and exhibit the highest levels of vulnerability. The risk of (AGIR), a regional response to chronic food and nutri- inadequate moisture renders dry land smallholders more tional insecurity across the Sahel, and is an active member vulnerable to production risk than their irrigated counter- of the Comité Permanent Inter Etats de lutte contre la parts and this is reflected in the lower levels of inputs Sécheresse dans le Sahel (CILSS). In 2006, Senegal final- applied to dry land crops. The extent to which intensive ized its National Adaptation Programme of Action livestock producers are more or less vulnerable to risk (NAPA) for climate change adaption. Following 2011’s than their extensive counterparts is debatable. On bal- severe drought affecting northern pastoralists zones, GOS ance, it would appear that although the impacts of risk set up emergency feed stocks under its Operation sauveguard events upon intensive livestock production may be greater du betail. Such GOS initiatives are already helping to safe- than those experienced by extensive producers, most guard livelihoods, promote climate adaptation, and intensive producers have a greater capacity both to pre- strengthen household resilience. And yet, as highlighted vent such events and to withstand their impacts. by this report, agricultural supply chains in Senegal remain highly vulnerable to a wide range of risks that Commercial farmers face many of the same risks as do jeopardize rural livelihoods. The current study highlights smallholders, but their levels of vulnerability differ. Com- the need for a more targeted and systematic approach to mercial producers may be able to absorb more production agricultural risk management in Senegal. risk, but face greater price risk. Processors are vulnerable to market risk because of increased local prices and/or Based on an analysis of key agricultural risks, an evalua- reduced costs of competing imports. Cotton and ground- tion of levels of vulnerability among various stakeholders, nut processors are also vulnerable to the risk of aflatoxin and the filtering of potential risk management measures, contamination, which cannot be detected in the unpro- this assessment makes the following recommendations for cessed materials but which can render the final products GOS’s consideration. The proposed focus areas of inter- unmarketable if subsequently detected. Processors also vention encompass a broad range of interrelated invest- face the risk of inadequate supplies as a result either of ments, which together hold strong scope to improve poor production, or of a redirection of inputs toward food agricultural risk management and strengthen the resil- crops for own consumption, especially after a poor har- ience of agricultural systems in Senegal. vest. Traders are primarily vulnerable to risks caused by 1. Strengthening extension delivery systems (for market uncertainty. In particular, they can have poor example, face-to-face, farmer-driven, ICT-based knowledge of market volumes or of the extent of produc- [Information and Communication Technology]) tion. As a result of this vulnerability, few traders are will- for improved farmer access to technology and ing to accumulate large positions with the intention of agronomic advice on improved soil, water, and selling at higher prices. pest management practices (for example, conser- vation agriculture, integrated pest management [IPM]). RISK MANAGEMENT 2. Promoting improved water management mea- The government of Senegal (GOS) understands the sures (for example, water pans, roof and rock importance of putting in place effective agricultural risk catchment systems, subsurface dams) and microir- mitigation systems. It has adopted in recent years a range rigation technology (for example, drip irrigation) of capacity-building measures toward reducing Senegal’s via community-led initiatives (for example, cash/ exposure to natural disasters and impacts from a changing food for work programs). climate. These measures include the creation of the 3. To further reduce rainfall dependency and better Directorate of Civil Protection (DPC), the development exploit existing water and land resources, promot- of a National Platform for Disaster Risk Reduction ing expansion of irrigation infrastructure. (DRR), and the elaboration of a National Action Plan on 4. Promoting use of contour erosion and fire barriers, DRR (2010–15). Senegal also participates in the recently cisterns for storing rainfall and runoff water, launched, EU-led Global Alliance for Resilience Initiative controlled/rotational grazing, grazing banks, Agricultural Sector Risk Assessment xv homestead enclosures, residue/forage conserva- generated insight into which sources of risks are most tion, and other Sustainable Land Management likely to affect the sector and dependent livelihoods in the (SLM) practices to reverse degradation of water, future. By prioritizing risks, the study can help GOS focus soil and vegetation cover ensure sustainable access attention and resources on a smaller set of key risks that to grazing land. are having the most adverse impacts on production yields, 5. Establishing and improving regional and incomes, and livelihoods. The study suggests a framework national normalized difference vegetation index for the development of a more comprehensive, integrated (NDVI) and early warning systems and farmer risk management strategy to strengthen and broaden training linked to an effective and early emer- existing mitigation, transfer, and coping measures in gency response system for drought and locust Senegal. Finally, it provides a filtering mechanism to aid in outbreaks. the selection of a set of strategic interventions for improved 6. To improve decision making among farmers and agricultural risk management. pastoralists and attenuate price volatility, strength- ening the quality and access to needed agro The assessment recognizes that many of the proposed information, including weather forecasting, exten- strategies may already be covered to varying degrees sion advice and innovations (that is, seeds, water under existing risk management programs. Others may management), input/output prices, and so on for currently be in the process of implementation, either by improved decision making. government agencies or by donors. Moving forward, the 7. Strengthening seed distribution systems, vaccina- Phase II Solutions Assessment will analyze the effective- tion programs, and animal health services through ness of existing programs, identify and assess challenges improved monitoring and enforcement of existing impeding their effectiveness, and outline strategies for quality control regulations governing product and scaling up effective interventions to reach a larger number service delivery, institutional capacity-building, of beneficiaries. This follow-up activity will place strong reform measures, and so on. emphasis on ensuring a more coordinated, integrated 8. Building resiliency in northern pastoralist zones approach to risk management in Senegal to ensure more via more broadly inclusive policy making around effective and meaningful risk reduction and resilience land administration for improved mobility and building across the sector. access, and development of community-driven feed/fodder production and storage centers. It is hoped that the findings and conclusions of this assess- ment will help to contribute to the existing knowledge base regarding the agricultural risk landscape in Senegal. It is also hoped that the study will help to inform a dia- CONCLUSION logue moving forward between the GOS, the World Bank, This Phase I assessment assesses agricultural risks and and GOS’s other development partners that will lead to impacts during the period 1980–2012. By documenting concrete interventions toward improved agricultural risk and analyzing how Senegal’s agricultural economy has management and stronger resilience among stakeholders been affected in the past by risk events, the study has in the years ahead. xvi Senegal CHAPTER ONE INTRODUCTION Risks are a pervasive and permanent fixture of the agricultural landscape. They are also costly. Unchecked, they breed uncertainty and stifle investments. For a given rate of return, the higher the risk associated with an agricultural enterprise, the lower the level of investment that it can attract. On the aggregate, this can have a debilitating impact on sector growth. This is especially true when risk is amplified by a limited capacity to absorb shocks. When risks do manifest, they can cause substantial losses to income and assets—especially among the most vulnerable communities—placing livelihoods, and in extreme cases, sector growth, in jeopardy. Failing to address agricultural risk can severely hamper long-term economic growth and poverty reduction efforts. The performance of Senegal’s agricultural performance exemplifies the impact of unmanaged risk on productivity among vulnerable smallholder crop producers and pastoralists. Despite the fact that well over half (57.1 percent in 2012) of the popula- tion lives in rural areas and derives some portion of its livelihood from agriculture, the sector itself contributes less than one-fifth (16.7 percent in 2012) to GDP, according to the World Bank. Despite several years of strong performance, sector growth has averaged 2.3 percent since 1980, amid notable volatility in year-on-year performance (figure 1.1). A succession of agricultural strategies designed to increase productivity has largely failed to intensify production beyond a subsistence level, and much of the country, although suitable for agriculture, remains underdeveloped. Keeping risks in check, shielding the most vulnerable, and building resilience among all agricultural stake- holders to better withstand and recover from inevitable shocks requires moving from ad hoc interventions to proactive, systematic and sustained risk management. The government of Senegal has historically responded to drought and other shocks with direct financial support to farmers as well as general assistance to the rural popu- lation. More recently, GOS put in place a series of emergency response and financial mechanisms to help affected communities better cope with shocks and enhance flows of rural credit. These include the Fonds de Bonification, Fonds de Garantie, and the Agricultural Sector Risk Assessment 1 FIGURE 1.1. AGRICULTURAL, VALUE ADDED (ANNUAL % GROWTH), 1980–2013 25 20 15 10 5 0 –5 –10 –15 –20 –25 –30 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 Source: World Development Indicators. Fonds de Calamité. The newly launched Operation de chains. The methodology was designed by the Agricul- Sauvegarde du Betail organizes distribution of feed sup- tural Risk Management Team of the World Bank. It plements to protect at-risk, breeding livestock (for exam- offers a conceptual framework and set of detailed ple, lactating females, calves) when access to sufficient guidelines for conducting a more systemwide assess- grazing is constrained. In 2014, Senegal was one of five ment of risk, risk management, and vulnerability countries to subscribe to a new pan-African drought index within agricultural supply chains. The methodology insurance facility under the Agricultural Risk Capacity contains logical steps within four consecutive phases (ARC) initiative. Important as these and other initiatives (figure 1.2). Phase I, for which this study is the primary are, GOS recognizes that these efforts alone are insuffi- deliverable, focuses on identifying, quantifying, and cient to insulate agricultural supply chains and the liveli- prioritizing the major risks that cause adverse shocks to hoods they support from adverse shocks. the sector. It is within this context that the World Bank, with support Following in-depth analysis of historical, time series rain- from the G-8 and the USAID and in collaboration with fall, crop and livestock production, pricing, and other the Ministry of Agriculture and Rural Equipment baseline data, the Assessment Team conducted broad- (MARE), commissioned the present study. The objective based, in-country consultations with stakeholders during of this assessment was to assist the government of Senegal March 2014. These included individual farmers, farmer to (1) identify, analyze, quantify, and prioritize principal groupings, input suppliers, market traders, food proces- risks (that is, production, market, and enabling environ- sors, and representatives of the government and research ment risks) facing the agricultural sector; (2) analyze the and academic institutes in and around Dakar and in key impact of these risks; and (3) identify and prioritize appro- agricultural production zones across Thies, Fatick, priate risk management (that is, mitigation, transfer, cop- Diourbel, Kaolack, Kaffrine, Louga, and St. Louis. The ing) interventions that might contribute to improved mission team organized a wrap-up roundtable consulta- stability, reduced vulnerability, and increased resilience of tion hosted by the Ministry of the Economy and Finance agricultural supply chains in Senegal. This report presents (MEF) on March 21 to share preliminary results and a summary of the assessment’s key findings. solicit feedback. Participants were asked to prioritize pos- sible future interventions by ranking a long list of risk mitigation, transfer, and coping interventions. Their METHODOLOGY input provided valuable insights into GOS priorities and The analysis presented in this report is based on a all feedback has been incorporated into the study’s analy- methodology for assessing risks in agricultural supply sis and findings. 2 Senegal FIGURE 1.2. AGRICULTURAL SECTOR RISK MANAGEMENT PROCESS FLOW PHASE I PHASE 2 PHASE 3 PHASE 4 Client demand Risk Solutions Development of risk Implementation and assessment assessment management plan risk monitoring RM plan development Desk review Desk review Implementation by stakeholders Stakeholder In-country Monitoring risks consultations assessment mission Incorporation into existing govt. programs and Stakeholder development plans Finalize analysis Refining RM strategy workshop Source: World Bank. The results of this assessment will provide the concep- investments because perceived risks associated with the tual basis for Phase II, which will focus on identifying production and marketing of a specific crop or animal. priority solution areas and related risk management Such ex ante impacts of risk might be quantified as poten- interventions best suited to managing the priority risks tial losses, but their attribution and measurement are identified. By the end of this activity, the World Bank— extremely complex. in close collaboration with GOS, its donor partners, and other sector stakeholders—will develop and validate a This analysis focuses mainly upon the ex post impacts of matrix of priority interventions related to risk mitiga- adverse events associated with risk. Ex ante impacts of tion, transfer, and coping within a comprehensive and risk upon investment decisions are largely ignored. The systematic risk management framework. It is hoped that measurement of perceived risk and associated impacts the outcome of this assessment will serve to inform upon investment decision making is a complicated task ongoing and future GOS sector policy and planning, that goes beyond the resources available in this prelimi- which will secure improved sustainability of agricultural nary assessment. Elbers, Gunning, and Kinsey (2007) pro- investments and enhance long-term agricultural resil- posed that ex ante impacts of risk upon agricultural GDP ience and growth. (in terms of foregone production) are potentially as great if not greater than ex post impacts from risk events. This assessment is based on the premise that ex ante impacts can reasonably be expected to be roughly proportional to RATIONALE ex post losses. Thus, after taking into account the qualita- The rationale behind the risk prioritization exercise is tive input from interviews and focus groups, the priority based upon the nature of risk in agriculture. There is no components of risk can be readily identified and responses standardized procedure for the quantification and mea- recommended. A more detailed justification of the surement of agricultural risk. Although it is possible to methodology is given in appendix F. measure the ex post impacts of events that contribute to risk in terms of the loss of yield or income resulting from Chapter 2 of the report provides an overview of the agri- those events, it is far more difficult to estimate income cultural sector in Senegal and a discussion of key growth foregone by producers, traders, and others who limit their constraints. This is followed by an assessment of the main Agricultural Sector Risk Assessment 3 agricultural risks (that is, production, market, enabling livelihood groups. The study concludes in chapter 6 with environment) in chapter 3. Chapter 4 analyzes the fre- an assessment of priorities for risk management and a quency and severity of highlighted risks and assesses their broad discussion of possible risk management measures impact. Chapter 5 presents some stakeholder perceptions that could help to strengthen the resiliency of agricultural of risks and evaluates levels of vulnerability among various supply chains and the livelihoods they support. 4 Senegal CHAPTER TWO AGRICULTURAL SYSTEMS IN SENEGAL To inform the analysis and discussion of agricultural risk in Senegal, this chapter presents an overview of the country’s agriculture sector. The most pertinent sector characteristics related to risk are given particular attention. The analysis primarily covers the 33-year period from 1980 to 2012 to assess the frequency and severity of the most important risks. AGRICULTURE SECTOR IN SENEGAL Table 2.1 shows the key economic indicators for Senegal with notable import to agri- culture. Of particular relevance is the fact that while approximately 57.1 percent2 of the population lives in rural areas and is largely dependent upon agriculture, over the five years up to 2012, agriculture and associated activities generated only 16.7 per- cent of national GDP, according to the World Bank’s World Development Indicators. Considering the current GDP of US$14.05 billion and a population of 13.73 mil- lion, it is evident that there is considerable discrepancy in rural vs. urban incomes. Whereas per capita GDP may be US$1,023, the average per capita value added for the population dependent upon agriculture is US$300; the rest of the largely urban population is 6.7 times higher (US$2,000). This is due largely to the relatively low output of the country’s agricultural production systems, which generate less than US$1.00 per capita per day. This suggests that the bulk of agricultural activity is of a subsistence nature. In fact, the level of production is inadequate to meet national demand and Senegal imports significant volumes of food. Food imports in 2012 were worth US$1,546 mil- lion, or 11 percent of GDP and up to 26 percent of total imports by value, according to the World Bank. Senegal’s exports are valued at slightly over 50 percent of total imports and the current account is balanced largely through a combination of remit- tances and development assistance. 2 According to the World Bank, Senegal has one of the highest levels of urbanization in Africa, estimated at 42.9 percent in 2012 and growing at a rate of 3.6 percent per year. Agricultural Sector Risk Assessment 5 TABLE 2.1. SENEGAL NATIONAL AND AGRICULTURAL STATISTICS, 2012 National Agricultural GDP (Current US$ million) 14,050 Total land area (ha) 19.25 million Population (million) 13.73 Total agricultural area (ha) 9.51 million Per capita GDP (Current US$) 1,023 Arable area (ha) 1.27 million Per capita GDP (US$ PPP) 1,906 Cereal crop area (ha) 3.85 million Population growth rate (%) 2.9 Permanent cropped area (ha) 58,000 GDP growth rate (%) 3.5 Arable land per person (ha) 0.29 Forest area (ha) 8.43 million Contribution to GDP: Avg. cereal yield (kg/ha) 1,310 Agriculture 17% Avg. fertilizer use (kg/ha) 7.6 Manufacturing 14% Cereal production (MT) 1.66 million Other industry 10% Cereal demand (MT) 2.48 million Services 59% Imports (Current US$ million) 5,901 Exports (Current US$ million) 3,372 Shoats (head) 10.93 million Remittances (million) 1,478 (2005) Pigs (head) 375,000 Net ODA (Current US$ million) 1,084 Camels (head) 5,000 Foreign Direct Investment 338 (US$ million) Inflation (yr on yr CPI basis) 1.4% Poverty headcount (national 46.7% (2011) poverty line) Gini Coefficient (income) 40.3 Source: World Bank, FAOSTAT. stations for which consistent and reliable information was AGRO-CLIMATIC CONDITIONS available for the period 1980–2013. Senegal has four climatic zones (figure 2.1). They are characterized by varying levels of rainfall and tempera- ture with conditions that gradually become increasingly dry moving north from Senegal’s high rainfall southern LAND AND WATER regions to its northern arid zones. RESOURCES Senegal is a flat country within the Senegal-Mauritanian All zones have unimodal rainfall. The length of the rainy Basin. Elevations above 330 feet (100 meters) are found season differs from one year to the next and from one only on the Cape Verde Peninsula and in the southeast. region to the other, being longer in the south (figure 2.2). The country is drained by the Sénégal, Saloum, Gambia With less than 1 percent of agricultural land under irriga- (Gambie), and Casamance rivers. Water resources are tion, the growing season in Senegal strongly correlates to estimated at over 35 billion cubic meters, of which 31 bil- the rainy season. This strong dependence of crop produc- lion are renewable surface water and 4 billion cubic tion on rainfall results in highly variable production, as meters are groundwater. However, the flat topography is both rainfall amounts and the onset and cessation of the for the most part unsuitable for the impoundment of rains are subject to marked space-time variability and water, limiting the potential for irrigation in many regions. temporal changes. Tables G.1 and G.2 in appendix G The Senegal River Valley alone accounts for an estimated compare cumulative rainfall amounts across 25 weather 240,000 ha of irrigable land, of which about 110,000 is 6 Senegal FIGURE 2.1. CLIMATIC ZONES OF SENEGAL Podor Saint-Louis Louga LEGEND Sahelian zone (250–500 mm) Thiès Sudan-Sahelian (500–900 mm) Dakar Diourbel Fatick Sudan zone (900–1100 mm) Kaolack Guinee zone (>1100 mm) Nioro Tambacounda Kolda Ziguinchor Kédougou Source: ISRA, adapted. FIGURE 2.2. AVERAGE ANNUAL RAINFALL BY REGION, 1980–2013 1,267 1,199 1,203 1,163 1,400 1,042 984 1,200 853 1,000 741 708 707 628 603 602 569 800 532 522 500 485 452 408 390 380 600 295 273 235 400 200 0 Cap Nioro Thies Bakel Kolda Podor Saint-Louis Dakar Fatick Louga Mbour Matam Kaffrine Goudiry Kaolack Sedhiou Bambey Diourbel Linguere Velingara Oussouye Kédougou Ziguinchor Koungheul Tambacounda Saint-Louis Louga Mat. Dakar Thies Diourbel Fatick Kaolack Kaffrine Tamba Kéd. Kolda Sed. Ziguinchor North North Central South Central SE South West Source: ANACIM. either currently under irrigation or is in development. one-tenth of this area receives average annual rainfall val- Some irrigation has also been developed in Casamance in ues below 500 mm, effectively limiting production. the south. The soils of Senegal are highly diversified. They include Although Senegal has over 19 million ha of land, over half fertile valley soils near the Senegal and Saloum rivers, of this is undeveloped bush and arid land used for livestock sands suitable for groundnuts, and sandy clays that can grazing; the total agricultural area is 9.5 million ha of support other crops in the western and eastern areas. In which 3.9 million hectares is suitable for arable crops. Of the south and center of the country, poor lateritic soils pre- this, 40 percent is regularly cultivated (that is, 20 percent dominate, whereas in the Casamance region, crops can be of the total agricultural land area is used for seasonal crop grown on the more fertile clay soils. In almost all cases, production). Though much of the arable area receives however, the soils are vulnerable to degradation and fertil- rainfall that is sufficient to produce average yields, roughly ity levels are declining as cultivation pressure increases. Agricultural Sector Risk Assessment 7 » The Eastern Senegal zone of 51,958 km2, is AGRO-ECOLOGICAL ZONES subject to rampant rural poverty due to heavy Senegal has six agro-ecological zones, based on biophysical population pressure on natural resources, despite and socioeconomic criteria: (1) the Niayes; (2) the Senegal its strong agro-pastoral potential. River Valley; (3) the Sylvo-pastoral Zone; (4) Groundnut » The Casamance can be divided into three Basin; (5) Eastern Senegal; and (6) Casamance (figure 2.3). zones—the lower, middle, and upper. With a total Each zone is a natural region, with its own potential and surface area of 28,324 km2, the region is charac- vulnerability to ecological and weather-related hazards: terized by lowland soil acidification, water erosion, » Niayes is a 5 to 10 km strip covering 2,759 km2. It loss of forest diversity, increased salinity, acidity, is the major commercial vegetable-producing area in iron toxicity, and acute mangrove degradation Senegal. It is a densely populated area and faces chal- within the Casamance estuary. lenges of soil and water salinity and coastal erosion. » The Senegal River Valley covers a surface area of 9,658 km2 in the north of the country border- CROP PRODUCTION ing Mauritania. This zone is characterized by allu- SYSTEMS vial plains and sandy uplands. Rain-fed farming is Although Senegal encompasses more than 19 million ha, almost nonexistent in the delta, and most agricul- the area available for agriculture is limited by poor soils tural production is derived from irrigation. Some and climate to less than 10 million ha (table 2.1). Forty- areas are subject to salinity, but much of the mid- three percent of the land area remains as undeveloped river area has a high level of fertility due to regular bush available for grazing, whereas a significant propor- flooding. tion of the remainder receives less than 500 mm of rain- » The Sylvo-pastoral zone covering 55,561 km2 is fall so that yields are severely constrained and much of the Senegal’s major cattle-breeding area and is mainly agriculture that is undertaken is inadequate even for sub- populated by nomadic Fulani ethnic groups. sistence. Shifting cultivation is commonly practiced and » The Groundnut Basin of 46,367 km2 is highly substantially less than 50 percent of the arable area is populated and subject to ecosystem degradation cropped at any one time. Crop composition has varied and depletion of land resources (soil fertility and little over the past 30 years (figure 2.4). Where there is timber resources). In addition, soil regeneration adequate moisture, the main crops cultivated are ground- has slowed as a result of upland soil acidification nuts and millet, which together account for almost 75 per- and lowland salinity. cent of the planted area. Maize, rice sorghum, cowpeas, FIGURE 2.3. AGRO-ECOLOGICAL ZONES IN SENEGAL Niayes River valley Sylvo-pastoral zones Groundnut basin Eastern senegal Casamance Source: Adapted from Directorate of Water, Forests and Hunting Conservation. 8 Senegal FIGURE 2.4. TRENDS IN CROP PRODUCTION AREA HARVESTED (thousand ha), 1980–2012 1200 Millet 1000 800 Groundnut 600 400 Maize 200 Sorghum Rice 0 Cowpea Mango 1980 1982 1984 Cassava 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 Source: FAOSTAT. and cotton make up about 25 percent and less than TABLE 2.2. TRENDS IN CROP PRODUCTION, 1 percent is sown to other crops, including vegetables. 1980–2012* Area Production Yield Senegal experiences a variable climate with low levels Change (%) Change (%) Change (%) of rainfall (< 600 mm per year) over much of the Food Crops northern part of the country. Its soils are for the most Maize 108.9 230.2 54.2 part sandy and acid so that levels of agricultural pro- Rice 39.6 99.9 359.9 duction are generally low. The agricultural sector has Millet 2.6 44.3 42.8 been traditionally dominated by two cash crops Sorghum 65.6 52.4 8.3 (groundnuts and cotton) produced for the export mar- Cowpeas 227.4 271.1 7.7 ket, although many producers also focus on the pro- Cash Crops duction of staple crops for their own subsistence. Tomatoes 486.6 647.5 23.9 Nevertheless, food crop production does not meet Potatoes –7.5 38.9 43.1 national demand, and the country is regularly obliged Cotton –14.7 7.5 26.3 to import substantial volumes of rice (1–1.2 million Groundnuts –10.0 23.1 36.2 MT in recent years) and wheat. Onions 235.7 438.3 66.2 Source: FAOSTAT. Crop production in Senegal essentially comprises three * Five-year average, 1980–84 vs. 2008–12; for onions and tomatoes, 2012 was categories of producers, though there is increasing diver- not available, so 2007–11 averages were used. sity across both food and cash cropping systems: » Subsistence smallholders who produce occasional Table 2.2 shows levels of variation in cropping area, yield, commercial surpluses for sale, but undertake other and total output for each of the main food and cash crops income-generating activities to sustain their liveli- during the period 1980–2012. In terms of output, tomato hoods. and onion production have grown the most during the » Commercial smallholders whose livelihood review period, followed by cowpeas, maize, and rice. depends upon the sale of cash crops, but who often Much of this growth has come from the expansion of produce some crops for their own consumption. cropping area, with the exception of rice production, » Pure commercial producers whose livelihood is which has benefited from a near quadrupling of yields based upon the sale of cash crops and horticulture. (359.9 percent) during the 33-year period. Production of Agricultural Sector Risk Assessment 9 tomatoes and onion, and to a lesser extent, cowpeas and monoculture throughout most of the country, with the maize, has seen the most marked growth. exception of the Groundnut Basin, where it is often replaced in the rotation by groundnut. Cowpeas require few inputs, and as a legume fixes nitrogen, enhancing the FOOD CROPS fertility of the soil for subsequent crops of millet or sor- Millet and Sorghum ghum. The short growing period of most cowpea varieties Millet and sorghum are the traditional cereal crops of is a key factor in their capacity to avoid moisture stress, Senegal. Millet (chiefly pearl millet) is the most widely cul- but it also results in lower yield. Senegal contains many tivated of the two and a large number of landraces local landraces of cowpeas and modern, higher-yielding adapted to different conditions are grown throughout the varieties are also available. Nevertheless, yields in 2012 country. The crop is grown almost exclusively in mono- were only 425 kg/ha and production at 55,000 MT culture in rotation with groundnuts or cowpeas. Produc- remained substantially below demand. tion covered over 800,000 ha in 2012, or roughly 36.1 percent of total harvested area. The coarse grain is Rice well adapted to moisture stress and is grown on a low Rice was not traditionally a staple in Senegal. However, input basis that is well suited to meeting smallholders’ inadequate production of millet and sorghum has led to subsistence needs, for which the bulk of the crop is pro- increased consumption of imported rice and a substantial duced. Higher yielding varieties have been produced by increase in domestic production to meet the growing Institute Sénégalais de Recherches Agricoles (ISRA) and demand. Rice is produced under both irrigated and rain- are being increasingly used by farmers, but production is fed conditions. Irrigated production, primarily in the still inadequate to meet demand. northern River Valley, occurs twice a year and accounts for roughly 45 percent of harvested area. Rain-fed rice is Sorghum is also drought tolerant, and although it yields produced either in small lowland basins or as upland rice, more heavily than millet (960 kg/ha in 2012), the crop which constitute 42 percent and 7 percent of area har- requires higher levels of fertility and deeper soils than mil- vested, respectively (the balance being mangrove produc- let, and it is indicative of the conditions in much of Sen- tion). The crop is produced mainly in the Senegal River egal, that the area and production of sorghum at 143,000 Valley and Casamance regions. Rain-fed rice is produced ha and 137,000 MT respectively are less than 20 percent almost exclusively by smallholders, whereas dry land rice of millet. Sorghum tends to be grown in wetter areas of in particular is produced mainly under slash-and-burn the country to the south where it tends to replace millet in conditions with minimal inputs. Lowland rice is produced the rotation. Traditional landraces adapted to a range of more intensively. A significant proportion of the irrigated conditions exist, but improved varieties are available and rice is produced under commercial conditions. A number are used by approximately 45 percent of smallholders. of improved varieties are available for irrigated, basin, and Sorghum is also grown on a low input basis as a subsist- upland cultivation and have contributed to the 360 per- ence crop, most of which will be consumed by the house- cent increase in yield observed over the past few decades. holds producing it. Although millet and sorghum are well adapted to the edaphic and climatic conditions of Sene- gal, both are regularly subject to depredation by birds CASH CROPS (quelea finch) and parasitism by Striga. Maize Like rice, maize is not a traditional Senegalese staple. Today, Cowpeas it is produced as much as a cash crop as it is for household Cowpeas exhibit more drought tolerance than do ground- consumption. Production occurs in most agro-ecological nuts. The crop also has a particular advantage in Senegal; zones, with the exception of the Silvipastoral zone. It is it can be eaten early as near-mature green pods as well as especially grown in the Groundnut Basin (where it is rotated harvested dry. Cowpeas can thus provide food during the with groundnuts), Casamance, and increasingly in Eastern traditional “hungry gap period.” The crop is grown in Senegal. Both smallholders (1 to 2 ha, 90 percent) and larger 10 Senegal commercial growers (20 to 50 ha, 10 percent) cultivate textile production subsector at the end of the last maize. The crop is entirely rain fed. As much as 97 percent century. Nevertheless, production expanded in 2012 as a of production involves improved varieties (ASTI 2009) and result of a substantial increase in the price of cotton lint most are short-season varieties (90 days to maturity). on the world market in 2011, although prices have since reverted to normal levels. The crop requires little fertil- Although the area planted has varied moderately between izer, but is nevertheless expensive to grow, because as 100,000 ha and 200,000 ha over the past 10 years, yields many as five applications of insecticide may be required. have fluctuated considerably and total production has Although cotton is less sensitive to moisture stress than thus varied between 400,000 MT in 2005 and 110,000 many crops, yields have trended downward over the past MT in 2011 (see figure D.3 in appendix D). There is a 10 years. growing demand for maize for both livestock feed (65 per- cent of production) and as maize flour (approximately Onions 100,000 MT). The shortfall is currently met by imports Onions are produced as a cash crop mainly by smallhold- that have averaged 100,000 MT in recent years. To ers in the Niayes agro-ecological zone. The crop is grown reduce reliance on imports, the government has subsi- with supplementary irrigation and average yields are dized the costs of maize crop inputs and promoted the among the highest in the region. The crop requires fertil- provision of finance for their purchase. Average maize izer but few chemical inputs, and given adequate irriga- yields in 2010–12 are somewhat lower than those in tion, is subject to little production risk. The crop is 2003–05, suggesting that although producers have produced between March and July, with the majority of increased the area planted to maize, they have not the crop being harvested in May and June. Because of a increased the intensity of production. lack of drying and storage capacity, the market is typically saturated with onions during these months. Groundnuts Groundnuts are grown throughout Senegal, except in the Silvipastoral zone, and especially in the Groundnut Basin. Tomatoes They are grown both as a staple for household consump- The country has a good climate for horticultural produc- tion and, more important, as a cash crop that can be sold tion throughout the year. About 70 percent of Senegal’s on the domestic market, mostly for processing into oil for exports to the EU are green beans, cherry tomatoes, man- export. The crop is grown almost exclusively by small- goes, and melons. The labor-intensive vegetable and fruit holders. The high cost of seed is offset by a government industry employs more than 17,000 families in rural Sen- subsidy and by input credit available through local coop- egal. Mangoes, green beans, and industrial tomatoes are eratives. However, the availability of good-quality seed is among Senegal’s major horticultural crops. Tomatoes are still inadequate. Yield are variable and range from 550 to produced under irrigation in peri-urban areas within the 1,200 kg/ha. Production reached a peak in 2010 of over Niayes zone. The crop is predominantly cherry tomatoes, 1.2 million MT, but has declined by about 40 percent grown in greenhouses or under shade netting for the Euro- since that time. The processed groundnut oil is Senegal’s pean market, a requirement that places it beyond the main agricultural export. Groundnuts are potentially sus- capacity of most smallholders. Production primarily takes ceptible to aflatoxin contamination, but the frequency of place during the months March–May. The crop is highly this in Senegal is low. The production of groundnuts is perishable and the domestic market can be saturated with considered politically sensitive and the crop is well sup- second grade produce during these months. Average yields ported by ISRA plant breeding programs and public fluctuate considerably from one year to the next, ranging extension. Nevertheless, groundnut production remains from 52 MT/ha in 2003 to 18 MT/ha the following year, particularly susceptible to erratic rainfall. largely the result of insect pests and market demand. Cotton Potatoes Cotton is produced mainly in Eastern Senegal. The crop Potatoes are also predominantly grown in the Niayes zone has been in decline since the collapse of the country’s by smallholders. The crop is grown between March and Agricultural Sector Risk Assessment 11 July and may be provided with supplementary irrigation when necessary. Yields generally vary about 15–25 MT/ha, LIVESTOCK PRODUCTION depending mainly upon the incidence of disease (blight and SYSTEMS virus) and, to a lesser extent insect pests. Production is Much of Senegal’s livestock sector, especially ruminants, grossly insufficient to meet demand and prices are consis- remains under a traditional extensive or mixed farming tently close to import parity. system. Pastoralists produce animal products that are sup- plemented, in the case of agro-pastoralists, by crops. The Mangoes annual production is primarily self-consumed but some Senegal produces 0.4 percent of global mango exports, portion is marketed. Although Senegal’s livestock sector is which are destined almost exclusively for European substantial, the country is dependent upon imports to supermarkets. Production is mainly from the Niayes meet its growing demand for meat. region, although some portion is now being exported from Casamance. Almost all production is rain fed and According to FAO data, the main species are cattle, sheep, organic. Fruit is produced by smallholders from indi- goats, pigs, poultry, equines, and camels. Livestock are kept vidual trees and small orchards and marketed either largely for meat, and to a lesser extent, for dairy and other through associations or directly to end buyers who products. They are also important for draft power. Small export by air to Europe. Although Senegal has always ruminants dominate the livestock sector, as shown in figure produced mangoes for the local market, the export seg- 2.5. Sheep are particularly important during the annual reli- ment has rapidly increased (by 15-fold in five years). In gious feasts of Tabaski and at baptisms of newborns. Pigs are 2005, Senegal experienced a glut of mangoes and of limited significance, being consumed only by the small prices fell considerably. Improved and expanded mar- non-Muslim population. According to FAOSTAT, pork meat keting arrangements have now reduced the probability represents roughly 6 percent of the total meat consumed of this recurring. [data taken from FAO’s online database at faostat.org]. Green Beans The poultry population in 2012 was 44 million birds. Green beans (also called bobby beans) are primarily Modern or intensive systems are usually practiced in grown in Senegal by smallholders and medium-size farm- urban and suburban areas and associated enterprises vary ers on contract to wholesale companies who airfreight the in terms of levels of technical sophistication. These range produce to Europe. Ninety percent of Senegal’s green from highly developed and bio-secure enterprises to bean exports are produced in the Niayes region through backyard, village production. Today, poultry production is vertically integrated supply chains. The Senegalese indus- split roughly equally between intensive commercial farms try is able to fill the out-of-season niche that exists from and backyard or village production, with growth in the December through to March before European producers former segment eclipsing noncommercial production begin production. (FAO 2014). The intensive sector is dependent upon six FIGURE 2.5. SHARE OF LIVESTOCK UNITS IN SENEGAL, 1961–2012 Cattle Goats Sheep Asses Horses 100% 75% 50% 25% 0% 1961 1971 1981 1991 2010 2011 2012 Source: FAOSTAT. 12 Senegal FIGURE 2.6. GROWTH IN POULTRY PRODUCTION (thousands), 1997–2011 30,000 Backyard or family poultry Intensive or industrial poultry 24,000 18,000 12,000 6,000 - 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 Source: DAPS (Direction de l’Analyse de la Prévision et des Statistiques). feed manufacturers and 10 hatcheries producing up to 14 noting that market dynamics in Senegal are highly influ- million day-old chicks per year. enced by supply and demand in two neighboring coun- tries: Mali and Mauritania. Poultry products are important to food consumption, par- ticularly among urban households. Chicken is the third AGRICULTURAL MARKETS most consumed meat after beef and mutton (FAO 2014). Average per capita consumption of poultry meat per year is AND TRADE 3 kg. According to the Ministry of Livestock (2011), Sene- Senegal exports substantial volumes of the high-value gal produces 499 million eggs per year, with average annual products such as groundnut oil and cotton (mainly to per capita consumption of approximately 20–25 eggs. China and to Europe) and cherry tomatoes, green beans, Government policy has focused upon achieving self-suffi- and mangoes (also to Europe). It also imports equally sub- ciency in poultry production and has put in place measures stantial volumes of rice, wheat, maize, onions, and pota- to reduce competition from imported poultry products as toes, some which are also produced domestically but in well as to support the production of maize to ensure an volumes insufficient to meet local demand. Prices for adequate supply of poultry feed. Intensive poultry produc- many of these commodities are thus strongly influenced tion has increased substantially as a result (figure 2.6). by international market prices. The export and import volumes of the main cash crops The significance of the livestock subsector is considera- are shown in figure 2.7. There is a marked contrast ble. Livestock production occupies 30 percent of the between the trade in cash crops, for which large volumes population and generates about 36 percent of agricul- are consistently traded in a given direction, and domestic tural GDP and 3.7 percent of total GDP (1994–2000). staple crops. For the latter, volumes are generally small Sixty-eight percent of Senegalese households, 90 per- and the trade is inconsistent4 (figure 2.8). Only rice is con- cent of rural households, and 52 percent of urban sistently imported in large volumes of 0.75–1.1 million households have herds. Livestock also provide significant MT per year. Hence, with the exception of rice, the inter- advantages: (1) a very wide and diversified range of national market is of little significance to these domestic products according to agro-ecological zones; (2) draft staples. Year-on-year exports of processed groundnut oil power for transport or cultivation;3 and (3) especially for are highly erratic but have exceeded 65,000 MT in the poultry, opportunities for export of animal products past, whereas exports of cotton lint have declined by made possible by a favorable animal health situation and roughly half since hitting their peak in 2003. a supportive trade policy environment. It is also worth 3 An estimated 90 percent of rain-fed agricultural land in Senegal is plowed by 4 FAO trade data for cowpeas are captured under the general heading “dry animals. beans,” of which Senegal intermittently exports small volumes. Agricultural Sector Risk Assessment 13 FIGURE 2.7. TRADE IN CASH CROPS (IN MT), 2002–11 Maize Groundnut Oil 70000 Imports Exports Imports Exports 20000 60000 0 50000 –20000 40000 –40000 30000 –60000 –80000 20000 –100000 10000 –120000 0 –140000 –10000 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 Cotton Lint Onions 20000 Imports Exports 30000 Imports Exports 0 25000 –20000 20000 15000 –40000 10000 –60000 5000 –80000 0 –100000 –5000 –120000 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 Tomatoes Potatoes 12000 25000 Imports Exports Imports Exports 10000 0 8000 –25000 6000 4000 –50000 2000 –75000 0 –2000 –100000 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 Mangoes Green Beans 10000 10000 Imports Exports Imports Exports 8000 8000 6000 6000 4000 4000 2000 2000 0 0 –2000 –2000 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 Source: FAOSTAT. administration and investments in the agriculture, forestry, NATIONAL AGRICULTURAL and pastoral sectors. However, implementation of the law POLICY since adoption has generally been slow. Successive programs have all stressed the importance of stimulating productivity, achieving food self-sufficiency, and Building on Senegal’s Poverty Reduction Strategy Papers I attracting new investments into the sector. Institutional (2003–05) and II (2006–10), the current Stratégie de Crois- reforms implemented since 1980 have generally facilitated sance Accélérée (SCA) adopted in January 2008 targets an the withdrawal of the state from agricultural production economic growth rate of 7 to 8 percent via the expansion and marketing, the restructuring and reorientation of public of five key sectors including agriculture and agro-industries. enterprises, and support favoring increased private and The Programme d’Accélération de la Cadence de cooperative sector participation. Adopted in 2004, the Loi l’Agriculture au Sénégal (PRACAs) provides the organiza- d’Orientation agro-sylvo-pastorale (LOASP) provides the tional and operational framework for GOS interventions in overall, long-term policy framework for public sector promoting sustainable agriculture, productivity, and farm- 14 Senegal FIGURE 2.8. TRADE IN STAPLE CROPS (IN MT), 2002–11 Sorghum Millet 5000 Imports Exports 5000 Imports Exports 0 0 –5000 –5000 –10000 –10000 –15000 –15000 –20000 –20000 –25000 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 Dry beans Rice 1000 Imports Exports 200000 Imports Exports 0 0 –200000 –1000 –400000 –2000 –600000 –800000 –3000 –1000000 –4000 –1200000 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 Source: FAOSTAT. ers’ resilience to external shocks. PRACAS is an integral pates in regional locust control activities and undertakes part of the Plan Sénégal Emergent (PSE), which defines the national livestock vaccination and emergency feed distri- country’s overall development goals. It places the agricul- bution campaigns, although the scope and reach of these ture sector at the heart of economic development. It high- programs can be variable. lights climate change as a major challenge and stresses the importance of finding effective and sustainable solutions to Despite signs of progress, the current situation leaves enable people to adapt and build their resilience against farmers, herders, and other sector stakeholders vulner- climate shocks and other hazards. able to a wide range of natural hazards and other shocks. It also in part explains why the sector has fallen All key elements of support to both crop and livestock sec- short of achieving sustained growth despite high rates tors are present in Senegal, including research into plant of public investment in the sector.5 A more targeted and breeding and disease control, multiplication and dissemi- systematic approach to risk management is needed to nation of government-bred seed stocks, veterinary ser- protect livelihoods and support sector growth and vices and livestock disease control programs, agricultural development. extension, provision of subsidized inputs (fertilizer and seed), and input loans. GOS has also supported programs promoting improved access to crop and livestock insur- FOOD SECURITY ance, although coverage is limited and the level of service At the national level, food security in Senegal, as measured provided remains low. by the Bonilla Index (Diaz-Bonilla et al. 2000) is among the lowest in Africa, ranging from 0.48 to 0.72 during the period Research in particular to produce disease-resistant and 1995–2010.6 The 2013 national food security and nutrition short-season crop varieties (for example, groundnut, maize, cowpea) is ongoing and substantial progress has been made in recent years in making these available to 5 In 2013, public expenditures in the agricultural sector reached 9.2% of Sen- farmers. However, the majority of farmers continue to egal’s national budget, just shy of the 10% commitment under the Framework of the Comprehensive Africa Agriculture Development Programme (CAADP). face difficulties in sourcing improved varieties and 6 The Bonilla Index is the ratio of the value of food imports to the value of total home-grown, recycled seeds are still widely used. In terms exports. As an index, it captures both domestic productivity and the cost of of direct risk mitigation efforts, the government partici- imported foodstuffs, that is, both the availability of and access to food. Agricultural Sector Risk Assessment 15 FIGURE 2.9. RETAIL PRICES FOR KEY STAPLE CROPS (CFA/KG), 2001–13 400 Millet Sorghum Maize Rice 320 240 160 80 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 Source: DAPS. survey found that 18.8 percent of households, correspond- insecurity in Senegal. Local production levels are always ing to some 245,000 households or 2.2 million people, are inadequate to meet demand. Thus, it is the price of food food insecure. The situation is especially accentuated in and a household’s capacity to pay that price that largely rural areas, where 25.1 percent of households are food inse- determines food security in Senegal. cure versus 15.1 percent reported in 2010. In recent years, food insecurity has dramatically worsened in the conflict- affected Casamance (Kolda, Sedhiou, and Ziguinchor KEY GROWTH CONSTRAINTS regions), as well as in the Kedougou and Matam regions AND TRENDS (WFP 2013) (see Senegal Comprehensive Food Security In Senegal, low levels of soil fertility and limited farmer use and Vulnerability Assessment 2013, SE/CNSA, SAP, of improved seeds, fertilizers, and agro-chemicals (for exam- WFP). National levels of stunting at age 5 are 26.5 percent ple, insecticides) limit productivity. In the absence of mecha- (this statistic comes from the UNICEF database at http:// nized equipment and services, there is a high reliance on www.unicef.org/infobycountry/senegal_statistics.html). family labor. Poor access to inputs and financial services fur- ther contributes to low adoption of productivity-enhancing Although demand exceeds local supply, imported food- technologies. Inadequate storage, roads, and other market- stuffs are available in markets throughout the country and ing infrastructure discourage farmers from investing in it is only in the most isolated areas that food insecurity upgrades. For the livestock subsector, low-performing could be described as an issue of food availability. House- breeds, poor husbandry management, insufficient feed/fod- hold economic analysis has shown that the majority of the der supply, high cost of poultry feed, and strong competition rural poor are dependent upon markets to meet their food from imports are among key growth constraints. These con- needs. Markets for staple crops are not strongly devel- straints hinder sector growth by limiting many producers’ oped, though they are both extensive and reasonably well ability to raise productivity and move beyond subsistence. integrated. Nevertheless, there is a high degree of uncer- These same constraints can also amplify the impacts of tainty in terms of domestic cereal prices which can fluctu- adverse shocks by increasing the scope of losses and weak- ate substantially both within and between seasons. ening the coping capacity of agricultural stakeholders. Figure 2.9 highlights the high degree of uncertainty that In addition to constraints, trends such as climate change, many households face as domestic cereal prices fluctuate soil erosion, and decreasing groundwater availability as a substantially both within and between seasons. It is this result of salinization and declining water tables are fluctuation in domestic market prices (together with global increasing the vulnerability of producers to climate risks price variations) that is the fundamental cause of food and other threats. 16 Senegal CHAPTER THREE AGRICULTURE SECTOR RISKS The main sources of risk in Senegal’s agricultural sector are reviewed in this chapter. These include production risks, market risks, and a general set of risks associated with the enabling environment for agriculture. The incidence and implications of multiple or successive shocks are also considered. PRODUCTION RISKS In terms of production, the most important factor contributing to agricultural risk is weather. Other factors include crop pests and diseases (both before and after harvest), windstorms, bush fires, and livestock diseases having weak response mechanisms. Each of these factors is considered in more detail below. WEATHER RISKS Weather-related risks manifest mainly through extremes of temperature and precipita- tion. In the former case, the actual impact of extreme temperatures upon crop yields in Senegal is uncertain. Over the past 40 years, mean annual temperatures across Senegal have increased rapidly (averaging 0.19°C per decade), but prior to that period, mean annual temperatures had decreased substantially, so that current mean annual temperatures have only recently regained the levels experienced at the beginning of the last century (figure 3.1). Extreme Temperatures Although it is evident that high temperatures that inhibit plant production (generally in excess of 35°C) will occasionally occur during the growing season, their impact is not easily determined, because it will often have been confounded by moisture stress. Similarly, the impact of extreme cold is not easily measured because it occurs so rarely in Senegal (although an important exception is the impact of cold rain upon weakened livestock, as happened in 2002). Extreme temperatures may well constrain production, but although changes in the frequency and impact of extreme temperature anomalies can be modeled (see appendix A), they are not empirically evident. Temperature per se is rarely considered by stakeholders as a contributory factor to crop production risk, and does not feature as an observed cause of ex post impacts. Agricultural Sector Risk Assessment 17 FIGURE 3.1. VARIATIONS IN TEMPERATURE AND RAINFALL, 1900–2009 2 +0.9 degrees Temperature-south celsius central 1 Standard deviations 0 Rainfall-north central Temperature-north –1 central Rainfall-south –14 percent central rainfall –2 00 55 09 19 19 20 Source: Funk et al. 2012. FIGURE 3.2. PRECIPITATION PATTERNS OF MAJOR REGIONS, 1978/79–2008/09 1600 Casamance Sénégal Oriental Sine Saloum Diourbel-Dakar Louga-Saint-Louis Average annual rainfall (mm) 1200 800 400 0 1978/79 1980/81 1982/83 1984/85 1986/87 1988/89 1990/91 1992/93 1994/95 1996/97 1998/99 2000/01 2002/03 2004/05 2006/07 2008/09 Source: ANACIM. TABLE 3.1. VARIABILITY OF RAINFALL BY REGION Sénégal Sine Diourbel- Louga- Casamance Oriental Saloum Dakar Saint-Louis Mean 1125 735 611 437 308 Standard deviation 189 149 149 123 80 Coefficient of variation 17% 20% 24% 28% 26% Source: ANACIM. Erratic Rainfall variation in national crop yields can be ascribed simply to The single most important factor affecting crop produc- the variation in annual rainfall amounts. (Kandji, Ver- tion is the availability of moisture. Moisture stress can chot, and Mackensen 2005). Annual rainfall tends to manifest itself through the delayed onset of rains, through decrease with latitude, being highest in the south of the erratic rain, through the early cessation of rain, or through country and decreasing to the north (figure 3.2). extended drought, all of which can result in reduced yield. Nevertheless, even in the absence of these specific condi- In general, the variability of rainfall tends to increase as tions it has been shown that more than 40 percent of the the absolute amount decreases (table 3.1), although the 18 Senegal high variability of rainfall in Louga-Saint-Louis appears Today, the majority of growers do not plant at the statisti- to be an exception to this pattern. More detailed rainfall cally optimal period, but wait until they are convinced analysis supports the general thesis that not only do yields that the rains have set in properly. This has the result of decrease moving from south to north, but the probability limiting potential production, because limited mechaniza- of an abnormally low yield due simply to reduced rainfall tion capacity causes the period of soil preparation and is also increased. sowing to be extended well past the optimal sowing date. Indeed, in the case of millet, research showed that there The various detrimental aspects of rainfall tend to act in was little to no impact of delayed rains upon yield. This specific ways. The impact of erratic rainfall is modified to occurred because growers generally waited for a tradi- a considerable extent by the growth stage of the crop tional date before sowing, a date at which the rains would under consideration. Thus, maize may experience erratic have almost certainly set in. This date may not have been rainfall during early growth yet still yield well, whereas the optimal for production, but had become accepted as part impact of the same rainfall regime during tasseling and of an effective risk mitigation strategy. silting, when evapotranspiration demand is much higher, can result in substantial loss of yield. Similarly, sorghum When the probabilities of late onset and early cessation and millet are prone to notable yield reduction if rainfall of rains are mapped, it is evident that the highest proba- amounts are reduced during inflorescence. By contrast, bility of both occurs in the middle latitudes of Senegal cotton has an indeterminate growth form that allows it to (figure 3.3). This area is very similar to that where the recover from a period of moisture deficit, although germi- probability of erratic rainfall is also highest. Conse- nation and early seedling growth can be much reduced by quently, even though the area may not experience the moisture stress (FAO 1971). Cowpeas and groundnuts are lowest levels of rainfall, it is the area in which the contri- particularly affected by dry conditions during pod bution of uncertain precipitation to risk is the highest. It development. is notable that this area includes the Groundnut Basin as well as part of the maize and cotton production areas. Drought The production of these crops is thus particularly exposed The early cessation of rain can cause loss of yield through to risk due to uncertain rainfall. reduced grain filling, especially in groundnuts and, to a lesser extent, maize and cowpeas. Its impact is less pro- An analysis of standardized cumulative rainfall data col- nounced for sorghum or millet, or an indeterminate crop lected from 25 weather stations over the period 1980– such as cotton but can have a major effect on horticultural 2013 is summarized in table 3.2. The analysis provides crops such as potatoes or tomatoes. Detrimental impacts insights into the frequency and severity of rainfall events can be reduced through the planting of short-season vari- during the 34-year period. For the purpose of this analy- eties, although these may be lower yielding than their sis, drought is defined as rainfall less than one standard longer-season counterparts. deviation (>–1) from the mean and extreme drought as rainfall less than two standard deviations (>–2) from the The risks to production posed by erratic rainfall or drought mean, whereas excess rainfall and severe flooding are are particularly significant because they occur once the defined as rainfall more than one (>1) and more than two farmer has invested the bulk of the required resources (>2) standard deviations from the mean, respectively. into the crop and there is effectively “no going back.” By contrast, the risk of the delayed onset of rains, which can During the review period, the country experienced result in poor germination and death of young seedlings, 10 years of drought, three of which were categorized as can be avoided by late planting. If planting is delayed severe drought (1980, 1983, 2002). This equates to a fre- beyond a certain extent, then yields will invariably suffer. quency of drought of roughly one out of every three to However, it may be possible to delay planting and benefit four years, on average. A subsequent analysis of crop from abnormally late rains to achieve normal yield levels. yield losses showed that these severe drought years coin- This response to the possibility of delayed onset of rains is cided with the largest losses in terms of annual crop pro- widespread across Senegal. duction value. The probability of drought was highest in Agricultural Sector Risk Assessment 19 Figure 3.3. Climate Variability Map for Senegal Source: ANACIM, WFP, IRI 2013. Tambacounda and Kolda. Regional droughts affected appendix G, table G.2). In 2010, abnormally high rainfall production in on or more province during other years (for was recorded in as many as 13 of Senegal’s 14 provinces, example, 1986, 2003), but these events were not classified when floods devastated many parts of the country. as drought due to their localized nature. It is worth noting that the analysis of cumulative rainfall fails to capture the It is worth noting that despite the observed increasing fre- poor rainfall events in 2007 and 2011 that severely quency of flood events in Senegal, the analysis suggests affected crop and livestock production during those years. that associated impacts relative to losses to crops and rural A more detailed table summarizing the analysis can be livelihoods are limited. Although individual farms may be found in appendix G. prone to flooding (especially some of the lowland rice areas), the majority of agricultural production is not at Floods risk from flooding and few respondents reported this as a The rainfall analysis also highlights an observed wetting risk to crops. Notwithstanding, impacts to urban house- trend, with a higher frequency of excess rainfall years in the holds and infrastructure can be substantial. To illustrate, most recent decade. It is worth noting that since the droughts flooding in 2009 in Dakar cost over US$100 million and of 1996–97, there has been only one episode of drought affected over 400,000 people (GFDRR 2012). Although (2002), an extreme event that affected half of Senegal’s 14 direct losses to crops and livestock from flooding may be provinces. According to the analysis, Senegal has experi- limited at the aggregate level, extreme rainfall events do enced nine excess rainfall years since 1980; all but one contribute to leaching and soil erosion, when organic mat- (1989) of these events occurred during the past 15 years. ter and other soluble plant nutrients in the topsoil are car- During the most recent decade, Senegal experienced excess ried away. This impedes the future capacity of soils to rainfall more than half the time, or six out of 10 years (see retain nutrients and moisture. Researchers have identified 20 Senegal TABLE 3.2. FREQUENCY AND IMPACT OF RAINFALL EVENTS BY REGION, 1980–2013 Inidicative Loss Value % of Agr Production Year Event Provinces Affected In US$ Value* 1980 xx Matam, Diourbel, Fatick, Kaolack, Tambacounda, –128.2 –19 Kolda, Sedhiou, Ziguinchor 1983 xx 12 of 14 provinces (exluding St. Louis and Matam) –128.5 –9.3 1984 x Saint-Louis, Matam, Dakar, Tambacounda –91.7 –13.8 1989 + Louga, Dakar, Thies, Diourbel, Zuiquinchor – – 1990 x Saint-Louis, Fatick, Kaolack, Kaffrine, Kedougu, – – Kolda 1991 x Matam, Kaolack, Kaffrine, Tambacounda, Kolda, –9.6 –1.4 Sedhiou 1992 x Saint-Louis, Louga, Matam, Dakar, Kedougou, –69.8 –10.5 Ziguinchor 1996 x Saint-Louis, Diourbel, Kaolack, Kolda –50.3 –7.5 1997 x Louga, Matam, Fatick, Kaffrine –68.0 –0.2 1999 ++ Matam, Kaolack, Kaffrine, Tambacounda, –33.5 –5.0 Koudougou, Kolda, Ziguinchor 2000 + Matam, Thies, Fatick, Kaolack, Kedougou –28.1 –4.2 2002 x Saint-Louis, Louga, Thies, Tambacounda, Kolda, –217.3 –32.6 Sedhiou, Ziguinchor 2003 + Saint-Louis, Matam, Tambacounda, Kedougou, –15.3 –2.3 Kolda, Sedhiou 2005 + Dakar Diourbel, Kaffrine, Kolda – – 2008 + Thies, Diourbel, Fatick, Tambacounda, Ziguinchor –9.5 –1.4 2009 ++ Saint-Louis, Louga, Dakar, Diourbel, Fatick, Thies, –10.7 –1.6 Tambacounda 2010 ++ 13 of 14 provinces (except Diourbel) –19.1 –2.9 2012 ++ 11 of 14 provinces (except Louga, Kedougou, Kolda) — — Key: x Drought xx Severe drought + Excess rainfall ++ Severe flooding Source: ANACIM 2014; FAOSTAT; Authors’ calculations and notes. *Average 2005–07 resulting low rates of fertilizer-use efficiency as a key rea- extent of such impact is so limited that hail has no dis- son why more farmers choose not to invest in fertilizer. cernible impact on yield from a national perspective. Hail and Windstorms The impact of windstorms appears to be similar to that of Other weather-related risks to agriculture in Senegal hail, albeit to a lesser extent, perhaps with once critical dis- include hail and windstorms. Hail is a common compo- tinction. Sustained windstorms contribute to soil erosion, nent of Intertropical Convergence Zone (ITCZ) precipi- particularly in flat areas with dry, sandy, and/or finely gran- tation and there is a high probability that in a given year ulated soils, which can be found in key production zones some crops will be damaged by hail. Nevertheless, across Senegal. Resulting wind erosion damages land and although the impact of a hailstorm can be devastating at natural vegetation by removing soil from one place and an individual smallholder level (resulting in 100 percent depositing it in another. It contributes to the deterioration loss of crop under the worst circumstances), the geographic of soil structure and causes nutrient and productivity losses. Agricultural Sector Risk Assessment 21 BUSHFIRES of climate change upon that risk appears to be uncertain, The risk to production posed by bushfires can be signifi- the most consistently predicted trend being an increase in cant at an individual smallholder level toward the end of the variability of rainfall amounts (see appendix B). This crop production when crops are mature and combustible, may be expected to increase the level of risk faced by farm- but at the aggregate level the impact of bushfires on ers, but parallel trends in farmers’ perception of that ele- national crop production is negligible. By contrast, bush- ment of risk, or in the frequency of weather-related risk fires can have a significant impact on livestock production, events that have affected yield, have not yet been recorded. both positive and negative. Early burning can result in regrowth from soil moisture reserves, which can help to CROP PESTS AND DISEASES extend the grazing season. Late burning tends to result in Crops in Senegal are subject to depredation by a range of fiercer fires that damage the bush upon which cattle and pests. By far the most significant are the Senegalese grass- goats graze in the dry season. Estimates suggest that as hopper, or “sauteuriaux” (Oedaleus senegalensis), locusts much as 6 percent of the potential dry season grazing area (Locusta migratoria), and birds. The first two are non–crop suffers from bushfire each year (Oceanium 2014). Data specific whereas the third is confined mainly to sorghum shared by the Centre de Suivi Ecologique (CSE) suggest and millet (although maize can also be affected). that on average 3.8 million MT of biomass are lost annu- ally through bushfires. This is equivalent to 100,000 MT Most farmers do not have the knowledge and much less of meat and can have a major impact upon production the financial means to adequately tackle crop pests; how- levels in those areas in which such burning has occurred. ever, all of the stakeholders confirmed that pests and dis- According to CSE, 762,921 hectares of land were ravaged eases are one of the main risks to agricultural production. by bushfires during the period October 2012 through There are different ways of dealing with pests and dis- May 2013, representing roughly 3.9 percent of national eases in cash and food crops; although in cash crops inputs territory (CSE 2013). The vast majority of this occurred in are mostly distributed on a credit basis, for food crops fer- the southern half of the country. The region of Tamba- tilizers and pesticides are procured on a cash-and-carry counda was the most affected, accounting for a third of the basis or diverted from cash to food crops. There is vast national total. Large swaths of pastureland in Sédhiou and number of pests that damage food crops in Senegal. A Kédougou were also damaged by bushfires, accounting for more detailed listing of pre- and postharvest pests that 11 percent and 10.16 percent, respectively, of total land. damage crops can be found in appendix F. CLIMATE CHANGE Since 1980, there have been six major locust invasions in The impact of historical and future climate change on Senegal, with significant impacts on both cash and food rainfall amounts and intensities in Senegal is uncertain. crops and livestock. For the period of analysis, the two Historically, national rainfall data suggest that cumulative worst infestations occurred in 1987–88 and 2004–05. rainfall amounts were decreasing until 1990, but since Locusts eat crops and other vegetation that is in their area then annual levels show an increasing trend. It is possible of infestation, often leading to total crop losses. They can that this is due more to decadal and/or multidecadal cli- also adversely affect livestock production through loss of matic cycles than to a longer-term trend (Kandji, Verchot, grazing. Damage is highly localized, but big swarms can and Mackensen 2005). It is worth noting that anecdotal cumulatively affect vast tracts of land. During the 1997– evidence collected for this study has indicated that the fre- 98 outbreak, GOS estimates at the time reported that quency and intensity of dry spells has increased over the locusts had infested almost 5 million acres of land and past 10–15 years. However, these results are based upon destroyed 10 percent of the year’s harvest.7 More than personal recollection, which has been shown to be biased 2 million acres were treated with pesticides to combat the toward a more favorable past (de Nicola and Gine 2011) invasion. Control of the locusts and grasshoppers is and should consequently be treated with caution. Overall, although aspects of climate contribute substan- 7 See “Senegal Fights Worst Locust Infestation in 30 Years” by Susan Katz tially to the risk faced by producers, the anticipated impact Miller, 25 November 1988. 22 Senegal usually through chemical pesticides, which can harm a exporters. More important, due to reputational risks, the wide range of organisms. Since 2000, a more ecological confiscation of a single batch can ruin the efforts of a solution has been available with the entomopathogenic whole campaign. fungus Metarhizium anisopliae var. acridum, registered as Green Muscle. LIVESTOCK DISEASES In addition to crop and livestock losses, response measures There is a wide range of diseases that threaten and can also be costly. To control 2004’s invasion, GOS had adversely affect livestock production in Senegal. How- initially budgeted for 1.8 billion CFA francs. The state ever, many diseases can be considered as constraints finally had to approve 4 billion CFA francs to bring the rather than risks, as they rarely lead directly to animal menace under control. This does not include outlays in mortality and the majority of owners know how to cash and in kind by private citizens, nor bilateral and manage them. However, growing animal populations, multilateral inputs. Damages due to locusts for the same declining vaccination coverage, and erratic availability outbreak were estimated at 2 million tons of crops, of quality vaccinations and medicines are among fac- equivalent to 20 percent of the population’s food needs in tors that can weaken existing risk management capac- the Sahel region. ity. This report concentrates on diseases that are either of a trans-boundary nature or of major zoonotic con- Granivorous birds, mainly the red-billed quelea, have sub- cern, and considers them in terms of the risk to local sisted on cereal crops in Africa for centuries and regularly and international trade or potential risk to the human cause substantial damage to crops. A study in Senegal’s population. River Valley estimated that annual bird damage averaged approximately 13.2 percent of the potential rice produc- The World Organization for Animal Health (OIE) con- tion during the wet seasons of 2003–07. This translates ducted an in-depth review in 2010 of livestock diseases in into an average annual economic loss of 4.7 billion CFA Senegal. Table 3.3 provides a list of the most important francs (US$9.7 million). These results were consistent diseases threatening cattle, small ruminants, and poultry. with farmers’ perceived bird-inflicted crop losses, averag- During the present mission, interviews with senior gov- ing 15.2 percent. ernment and private vets revealed that many of the con- straints and recommendations made in the OIE report There are more than a dozen fruit fly species that attack still reflect the situation today. mangoes. Across West Africa, losses from damage caused by mango fruit flies (Tephritidae diptera) have been growing. Although some diseases constitute a risk themselves, the This is especially true since the arrival of Bactrocera invadens, major risk to the livestock sector remains the inability of a fly species from Sri Lanka, first discovered in West Africa the veterinary services to respond to needs. Whereas dis- in 2004 by the International Institute of Tropical Agricul- ease control policies and strategies exist on paper, it is ture (IITA) in Benin. The insect pest has the potential to widely recognized that because of chronic underresourc- jeopardize the recent commercial success of the region’s ing and resulting scarcities of material resources and qual- mango export sector. Fruits showing the slightest trace of ified personnel, the ministry is often unable to implement a fly bite must be identified, removed, and destroyed dur- them or respond as required. Recognizing this challenge, ing harvesting and in-station sorting. Because fruit flies are the veterinary department has recently reprioritized and classified as “quarantine insects,” if a single fruit is detected now covers fewer diseases in their vaccination campaigns, that is infested with larvae, the whole batch can be rejected concentrating on PPR, Newcastle disease, African horse by European phytosanitary services. According to the sickness (AHS) and lumpy skin disease, with episodic cam- ACP-EU Technical Centre for Agricultural and Rural paigns against other diseases as they occur (for example, Cooperation (CPA), whole containers of fruit from Africa CBPP, pox). Of particular note: the OIE report empha- are regularly intercepted and destroyed each year in incin- sizes the involvement of the communities in policy design erators in European harbors and airports because of and implementation; a recommendation that this study infestation. This results in substantial economic losses for strongly supports. Agricultural Sector Risk Assessment 23 TABLE 3.3. MAJOR LIVESTOCK DISEASES Species Diseases Vaccinated Against Diseases under Surveillance Cattle Lumpy skin disease (LSD/DNCB) Rinderpest (RP); Foot and mouth disease Contagious bovine pleuropneumonia Hemorrhagic septicemia/ Rift Valley fever Pasteurellosis (HS) Blackleg Anthrax Botulism Small ruminants Peste des petits ruminants (PPR) Rift Valley fever Pasteurellosis Poultry Newcastle disease (NCD) Highly pathogenic avian influenza (HPAI) Source: Gary et al. 2010. BOX 3.1. CASE STUDY: CONTAGIOUS ter of whole herds, which would have a major impact on BOVINE PLEUROPNEUMONIA the whole livestock industry not to mention the livelihoods of hundreds of thousands of households who depend on (CBPP/PPCB) IN SENEGAL it. However, in Senegal, export of locally produced ani- CBPP (PPCB) was eradicated in Senegal in 1978 and vac- mal products (with the exception of day-old chicks and cinations ended in 2005. But in November 2012, there eggs) is virtually nonexistent, and most animals are of the was an outbreak in the south after animals crossed into the local indigenous breeds and quite tolerant of the above country. Additional outbreaks have been reported in 2013. diseases. The government implements less extreme con- Impact: trol measures and although there are some losses in pro- • Losses in production, including some mortality. ductivity, the impact is relatively insignificant. • Restrictions on movement. • Higher veterinary (and vaccination) costs. Other diseases that occur in Senegal are largely known • Need for up to a further 10 years of vaccination and and to some extent predictable. Thus, these are consid- surveillance. ered more as production constraints (for example, trypanosomiasis,, helminthiasis). Most well-maintained animals recover naturally from disease even if left Among livestock diseases, this assessment considers Rift untreated; however, productivity is obviously affected and Valley fever, highly pathogenic avian influenza, and New- most owners will opt to treat the diseases. The risk is that castle disease to be among priority livestock production without proper prevention, when a large outbreak occurs, risks. Avian influenza and Rift Valley fever can also have a the losses can be considerable, as very few animals have significant impact and influence on both local and inter- been vaccinated. According to officials at the Ministry of national trade. As notifiable diseases, they can also affect Livestock and Animal Production, approximately 5 per- policy decisions, which have the potential to cause dam- cent of poultry is vaccinated against Newcastle disease age to livestock systems if they are inappropriate or are each year; 20 percent of smallstock; 63 percent of cattle; implemented quixotically. and 38 percent of horses. In addition, the quality and efficacy of available vaccines is questionable; it is esti- Paradoxically, other diseases such as foot and mouth dis- mated that as many as 50 percent of veterinary products ease and contagious bovine pleuropneumonia, could used could be counterfeit or of poor quality. Thus, it is potentially cause even greater harm if the response mech- the lack of adequate vaccination services and mostly the anisms required for control and eradication were actually new, emerging diseases that this study highlights as the put in place. Such measures would necessitate the slaugh- major risks. 24 Senegal MARKET RISKS TABLE 3.4. INTER-ANNUAL CROP PRICE Among the most common market risks presented in this VARIABILITY, 1991–2011 section are price variability for crops and livestock, Coefficients of Variation exchange rate and interest rate volatility, and counter- Cereal Crops Cash Crops party risk. Maize 0.29 Cotton* 0.15 Sorghum 0.31 Groundnuts 0.21 Millet 0.34 CROP PRICE VOLATILITY Rice (paddy) 0.34 Price fluctuations are to be expected in agricultural mar- Source: FAOSTAT. kets. This is partly due the unpredictable nature of supply *Price for cotton lint. and demand, weather patterns, and related yields. How- ever, extreme price volatility deters producers from mak- markets so that producers are directly affected by interna- ing productivity-enhancing investments and can tional prices and by exchange rate fluctuations. jeopardize household access to food among poorer seg- ments of the population. An analysis of producer price variability is based on inter- annual price variability for the period 1991–2011, meas- The impacts of the global price increases in food and fuel ured by coefficients of variation (CV). Nominal prices in experienced in 2007-2009 were substantial. The 30 per- US$/ton taken from FAOSTAT are used for the analysis cent increase in the price of household foodstuffs (for of domestic producer prices. Table 3.4 compares levels of example, rice, cooking oil, sugar, wheat, millet, milk prod- inter-annual price volatility across Senegal’s principal ucts) increased poverty levels by six percentage points, cereal and cash crops. During consultations in the field, from 51 percent in 2005/06 to 57 percent in 2008 (Del stakeholders repeatedly emphasized the volatility of Ninno and Mills 2015). In 2007 and 2008, the price of domestic food crop prices. Indeed, the analysis highlights rice in local markets tripled, whereas grain prices increased the extent to which price variation among domestic food by 50 percent. People took to the streets to protest these crops is considerably greater than that of domestic cash price increases, with riots destabilizing the political envi- crop prices. ronment. Domestic food prices were 74 percent higher at the end of 2012 than they were in early 2006, according Food Crops to FAO’s food price index. As a result, living conditions of Staple crops are grown throughout Senegal, but the level the poorest households continue to deteriorate, with of production achieved by the majority of households is reductions in the quality and frequency of meals and inadequate to provide 100 percent food security. Thus, higher incidences of food insecurity and malnutrition. most households are dependent upon purchased food for at least some months of each year. The dependence upon Although the risks associated with price can be important markets is greatest in August when prices tend to be high- for household well-being, they are of only marginal sig- est. The price of staple crops at this time is therefore criti- nificance to crop producers. This is especially true of sta- cal to household food security and the risk that increased ple crops such as millet, sorghum, and cowpeas. Although food prices might reduce the accessibility of food has a high prices of these commodities can have a drastic substantial impact upon household resource manage- impact on food security, they do not have a major impact ment. It is difficult to quantify the impacts of risks because on the finances of producers, most of whom will consume of price volatility, but respondents frequently reported the almost all that they produce. This is less true of maize and diversion of crop inputs intended for cash crops such as rice, which are produced more as cash crops that must groundnuts or cotton to crops such as sorghum and millet compete with imported commodities so that low interna- to maximize food availability. They also cited retention of tional prices can result in reduced profitability. The same grain at the household level, this despite inadequate stor- is true for export fruit and vegetables, including mangoes, age conditions and consequent high levels of loss due to tomatoes, and beans, all of which must compete on world storage pests and other risks. Agricultural Sector Risk Assessment 25 FIGURE 3.4. NOMINAL PRODUCER PRICES FOR KEY STAPLE CROPS (CFA/KG), 2000–13 400 Maize Millet Rice Sorghum 350 300 250 200 150 100 50 0 00 01 02 03 04 05 06 07 08 09 10 11 12 13 n- n- n- n- n- n- n- n- n- n- n- n- n- n- Ja Ja Ja Ja Ja Ja Ja Ja Ja Ja Ja Ja Ja Ja Source: DAPS. In terms of impacts, the high level of market dependency contractual relationships with producers, but instead buy of the poorest households would suggest that an increase on an opportunistic basis so that prices are subject only to in the price of staple crops during the lean months would market forces, and when volumes are thin they can fluctu- increase the proportion of households experiencing food ate considerably. insecurity, with consequent effects upon levels of malnu- trition and associated morbidity. In practice, much greater Groundnuts fluctuations in prices are regularly observed (figure 3.4) Although growers of groundnuts may be confident that, and it is the profoundly negative ex post impacts of these because of the high political profile of the groundnut fluctuations upon nutrition, health, and survival that result crop, the price negotiated at harvest time will be adequate in staple food price shocks as being listed as the most to sustain their livelihoods, there is no guarantee that the important risk faced by rural households (see table 5.1). negotiated price will be that which they ultimately receive. If the processors’ margin between the agreed-on price for Cash Crops groundnuts and the international price for groundnut oil Because few producers grow cash crops without first is inadequate, then processors may restrict or delay pur- securing their own supply through staple crop production, chases. This has two effects. Intermediaries who have pur- the risks to nutrition, health, and survival caused by fluc- chased groundnuts from farmers find themselves holding tuations in cash crop prices tend to be less pronounced. large stocks, either in warehouses or more commonly on Horticultural producers are a general exception to this trucks, for which they have no immediate market. The because they typically produce fruits and vegetables exclu- intermediaries’ liquidity is thus dramatically reduced and sively for the market. Nevertheless, price fluctuations can their business effectively halted unless they can find alter- contribute significantly to the risks faced by all stakehold- native markets. ers in cash crop subsectors. Although the production of export crops is vulnerable to price risk, the production of Producers selling directly to processors similarly find domestically marketed cash crops is subject to even greater themselves unable to raise cash to meet their immediate price volatility. In the case of potatoes and onions, the sea- needs. Because they generally lack storage facilities, they sonality of production combined with a lack of suitable become vulnerable to postharvest losses unless they can storage infrastructure can result in a glut of these two otherwise dispose of their produce. The net effect is for commodities on the market leading to reduced prices and both intermediaries and farmers to dispose of their significant losses to both growers and traders. The impact groundnuts on the parallel market for domestic con- of seasonality is exacerbated by the poor articulation of sumption at a reduced price. Although it might appear value chains whereby traders do not develop regular or that the fixed price for groundnuts exposes the processor 26 Senegal FIGURE 3.5. INTERNATIONAL VS. DOMESTIC GROUNDNUT OIL PRICES, 1984–2013 3000 200 Groundnut oil price Delivered groundnut price International price (USD/MT) 2400 Domestic price (XAF/Kg) 150 1800 100 1200 50 600 0 0 1984 1987 1990 1993 1996 1999 2002 2005 2008 2011 Source: DAPS; FAOSTAT. FIGURE 3.6. INTERNATIONAL VS. DOMESTIC COTTON PRICES, 1984–2013 200.0 325 Lint price Seed cotton price International price (US cents/l b) 160.0 260 Domestic price (XAF/Kg) 120.0 195 80.0 130 40.0 65 0.0 0 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 Source: DAPS; Cotton Outlook Index A. to price risk, the existence of a parallel domestic market price risk is visited mainly upon the processing company. and the effective oligopsony of three large groundnut In practice, SODEFITEX is mandated by GOS to pur- processors, allows much of that risk to be passed back to chase all the cotton that is produced. Thus, it is unable to the producers. respond to anticipated price risk, and therefore bears the full brunt of unfavorable price fluctuations. The recent Cotton decline in the international price for cotton has led to the The price for seed cotton is fixed at the beginning of the erosion of liquidity accumulated in the past when prices growing season by Société de Développement et des of cotton lint were significantly higher (figure 3.6). It is Fibres Textiles (SODEFITEX), which is the only buyer of worth noting that unless prices rebound or financial sup- the crop. Consequently, it is exposed to the risk of interna- port can be obtained, current dynamics may well render tional price fluctuations for cotton lint until the ginned the company insolvent. cotton has been sold. Moreover, to guarantee a minimum level of throughput, SODEFITEX is obliged to offer a Since the mid-1980s, the international price index for cot- price that is competitive with that for groundnuts. The ton lint has generally varied between US$0.50–US$0.90 company is thus limited in the extent to which it can fac- per pound (figure 3.4), although a major spike occurred in tor the price risk into its buying price. Thus, for cotton, prices throughout 2011. The variability of international Agricultural Sector Risk Assessment 27 FIGURE 3.7. INTERNATIONAL VS. DOMESTIC MAIZE PRICES, 1984–2013 375.0 240 Imported maize FarmGate maize price International price (USD/MT) 300.0 Domestic price (XAF/kg) 180 225.0 120 150.0 60 75.0 0.0 0 1984 1987 1990 1993 1996 1999 2002 2005 2008 2011 Source: DAPS; FAOSTAT. cotton prices over the past 30 years has been relatively high, marked than those of either cotton or groundnuts. Never- with a coefficient of variation of 33 percent. By contrast, the theless, buyers of maize face only the risks of normal price international prices for both groundnut oil and for maize fluctuations in either market. have been more variable with coefficients of variation of 44 percent and 45 percent, respectively (figures 3.5 and 3.7). LIVESTOCK PRICE VOLATILITY The limited reliance of pastoralists upon markets implies Maize a limited impact of price risk upon pastoral livestock pro- Maize prices in Senegal are determined primarily by domes- duction. However, this situation is changing. In very tradi- tic supply and demand. Although the country is not self- tional low-input, low-output pastoral systems, market sufficient in maize (significant volumes are imported each dynamics were not a major concern and market risks were year, mainly for poultry feed), domestic prices to farmers are mostly of concern in commercial, intensive livestock pro- generally lower than import parity, attributable mainly to duction systems. However, the majority of livestock own- the cost and difficulty of aggregating substantial volumes of ers even in remote extensive systems are dependent on consistent quality with which to manufacture animal feed. markets to some extent. These same factors prevent the competitive export of maize. Producers of maize face only limited price risk (figure 3.7). The relatively recent trend of growing involvement and The most common experience over the past 30 years has dependence of pastoralists on markets is not without risks. been for maize prices to spike upward rather than down- These include market quarantines on animal sales because ward. The main risk due to maize price fluctuations is thus of disease outbreaks and food and feed animal price insta- visited on livestock rather than maize producers. bility. During market shocks, livestock prices often plum- met while food prices increase, which has now become a The price risk faced by the companies that process locally common shock-induced pattern in dry lands. The major purchased commodities for subsequent export (that is, reason that pastoralists of the Senegalese Sahel use live- cotton and groundnuts) is increased by the fact that the stock markets is to satisfy their own consumption needs, domestic purchasing price may vary independently of the which usually are aggravated and increased during dry export price, as a result of both the local price setting seasons and droughts (Wane et al. 2010). mechanism and of fluctuations in the exchange rate. By contrast, the export market for maize is negligible. Figure 3.8 shows seasonally adjusted prices for animals in Both local production and imported maize are sold within Dahra, Senegal’s primary livestock market. The cattle Senegal and although domestic prices may not track prices are stable in comparison with those of female goats international prices completely, the discrepancies between and sheep, which tend to decrease while those from rams the domestic and international prices are much less and billy goats are increasing. The trends in smallstock are 28 Senegal FIGURE 3.8. DAHRA MARKET LIVESTOCK PRICES (CFA/HEAD), 2005–10 Male cattle gross prices Male seasonally adjusted prices 300000 calf young bull bull 250000 calf young bull bull 250000 200000 200000 150000 150000 100000 100000 50000 50000 0 0 Jan-05 Jan-06 Jan-07 Jan-08 Jan-09 Jan-10 Jan-05 Jan-06 Jan-07 Jan-08 Jan-09 Sep-05 Sep-06 Sep-07 Sep-08 Sep-09 Sep-05 Sep-06 Sep-07 Sep-08 Sep-09 May-05 May-06 May-07 May-08 May-09 May-05 May-06 May-07 May-08 May-09 Female cattle gross prices Female cattle seasonally adjusted prices 200000 veal heifer cow 200000 150000 veal heifer cow 150000 100000 100000 50000 50000 0 0 Jan-05 Jan-06 Jan-07 Jan-08 Jan-09 Jan-05 Jan-06 Jan-07 Jan-08 Jan-09 Sep-05 Sep-06 Sep-07 Sep-08 Sep-09 Sep-05 Sep-06 Sep-07 Sep-08 Sep-09 May-05 May-06 May-07 May-08 May-09 May-05 May-06 May-07 May-08 May-09 Sheep gross prices Sheep seasonally adjusted prices 80000 ram sheep ram sheep 70000 60000 40000 50000 40000 30000 30000 20000 20000 10000 10000 0 0 Jan-05 Jan-06 Jan-07 Jan-08 Jan-09 Jan-05 Jan-06 Jan-07 Jan-08 Jan-09 Sep-05 Sep-06 Sep-07 Sep-08 Sep-09 Sep-05 Sep-06 Sep-07 Sep-08 Sep-09 May-05 May-06 May-07 May-08 May-09 May-05 May-06 May-07 May-08 May-09 Goat gross prices Goat seasonally adjusted prices 100000 goat female goat goat female goat 80000 25000 60000 20000 15000 40000 10000 20000 5000 0 0 Jan-05 Jan-06 Jan-07 Jan-08 Jan-09 Jan-05 Jan-06 Jan-07 Jan-08 Jan-09 Sep-05 Sep-06 Sep-07 Sep-08 Sep-09 Sep-05 Sep-06 Sep-07 Sep-08 Sep-09 May-05 May-06 May-07 May-08 May-09 May-05 May-06 May-07 May-08 May-09 Source: Ministry of Livestock. significant because of the number of animals being sold long-standing livestock flows and trade dynamics between and their importance at the level of household income. the two countries, with observed decreases in the availa- Anecdotal evidence collected during this study would sug- bility of animals for sale in Dahra in recent years. gest that market parameters in Dahra depend to some extent on the dynamics of neighboring Malian and The most striking characteristic of the graphs in figure Mauritanian markets. Anecdotal evidence would suggest 3.8 is the notably high intra-annual variation in price. that the ongoing conflict in northern Mali has disrupted This is likely due to seasonality in demand (religious Agricultural Sector Risk Assessment 29 FIGURE 3.9. GOATS VS. CEREALS TERMS OF TRADE, 2005–12 Local rice : Goat Local millet : Goat Local sorghum: Goat 160 Units of goat to units of cereal 140 120 100 80 60 40 20 0 2005 2006 2007 2008 2009 2010 2011 2012 Source: Ministry of Livestock; DAPS. FIGURE 3.10. HISTORICAL EXCHANGE RATES, 2001–12 800 600 XAF / US$ 1 400 200 0 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 Source: World Bank. celebrations; buying male animals to plow land) and sup- gos imposed in response to avian influenza constitute a ply (willingness to sell to cover needs in dry seasons, and major threat, with the possibility of cheaper imported unwillingness to sell when conditions are good and herd poultry products from Brazil and the United States, in rebuilding is taking place). It could also be due to disease particular, undermining the economic sustainability of outbreaks. Senegal’s emerging poultry industry. No national price data series for livestock is available. EXCHANGE RATE VOLATILITY Nevertheless, figure 3.9 compares the annual average Exchange rate fluctuations can also contribute to price price for goats in the Saint-Louis region to average cereal risk for exporters of locally purchased products. The prices for rice, millet, and sorghum. It illustrates that a exchange rate of the XAF to the U.S. dollar has shown goat bought more units of cereal in 2006 and 2010, but little erratic variation over the past 12 years, declining less substantially less in 2008 and 2012. Although 2006 from XAF 750 in 2001 to XAF 425 in 2008 and remain- was a drought year in Louga and Matam, 2008 and 2010 ing relatively stable thereafter (figure 3.10). Comparing saw food and oil price shocks and the global financial cri- the variation in international prices over the same period, sis. Despite these observations, there are insufficient data it is evident that the risk due to unexpected variations in to draw any firm conclusions. exchange rate has been relatively limited. Price volatility affecting imported feed components, nota- bly corn and soya, which together make up 80 percent of INTEREST RATE VOLATILITY poultry feed ingredients, can be considered the major risk As elsewhere, poor access to credit is one of the principal facing Senegal’s poultry industry. In the longer term, constraints to agricultural growth in Senegal. This is espe- changes in international trade policies and trade embar- cially true among the country’s smallholder farmers. 30 Senegal Agricultural credit as a share of total bank credit (3 percent) of risk when public involvement in sector activities has in Senegal is among the lowest in the region. Much of this unexpected, adverse consequences. Other risks include is used for purposes other than farm production, such as general insecurity as a result of domestic unrest or regional agro-food processing or storage. For example, a number of conflict that can also disrupt agricultural production sys- commercial banks finance the cotton and groundnut sec- tems and livelihoods. tors in consortium with the agricultural bank Caisse Natio- nale de Crédit Agricole du Sénégal. For groundnuts, the POLICY UNCERTAINTY bank’s direct partners are processors, warehouses, and seed Within the context of an enabling environment, including suppliers. For cotton, it is the processors and the national the promulgation of policy and development and imple- cotton producers’ federation, the Federation Nationale des mentation of regulations, it is not the development of that Producteurs de Coton. A relatively small number of small- environment per se, but the nature of stakeholders’ per- holder farmers, mostly involved in commercial tomato and ception of risk within it that is most critical to production. sesame production, are able to access seasonal credit via To that extent, any inconsistencies in policy or process can contract-financing arrangements. enhance uncertainty. Overall, the challenge to decision makers is that agricultural policy is obliged to reconcile For borrowers, high variability of interest rates can pose a the dilemma that although the development of the sector risk to their enterprise when sudden spikes in lending rates would benefit from higher commodity prices, the majority can adversely affect resources and operations, and in of rural households depend upon access to cheap food. extreme cases, cause them to default on their loans. How- This has led to inconsistencies in program implementa- ever, in Senegal, such risks are minimal as the number of tion, which have been a consistent criticism of agricul- borrowers is small and interest rates have been relatively tural policy in Senegal (Resnick 2013). Indeed, the analysis stable during the review period. highlighted a widespread perception among smallholders, pastoralists, and others that government interventions can COUNTERPARTY RISK increase the level of uncertainty associated with both crop In addition to price risk, producers, traders, processors, and livestock production and marketing. and others all face the risk of nonperformance by the other party in the course of a trade. Nonperformance in Over the past 20 years, GOS has consistently reduced its trade includes such occurrences as goods supplied under- involvement in the agricultural sector. Exceptions to this weight or below specifications, partial or delayed pay- have generally been positive for agriculture. Rice, maize, ment, or even complete failure to supply or make payment. and cassava, in particular, have been the subject of special The risk of nonperformance is generally higher when value chain development programs aiming to streamline markets are poorly regulated. Although the level of mar- interventions in these sectors and to intensify production. ket regulation in rural Senegal appears minimal, traders indicated that the frequency of nonperformance was low. For groundnuts, GOS policy has consistently been to set a They ascribed this to their own behavior in trading only processing price that will allow smallholders to make a with those whom they knew and trusted. Such a limitation profit. However, the single cotton ginning company of trading partners can reduce the efficiency of markets SODEFITEX is also obliged to offer prices that are com- and increase transaction costs. parable in terms of ultimate earnings to discourage farm- ers from switching crops. As noted elsewhere, this policy exposes SODEFITEX to price risk when international ENABLING ENVIRONMENT cotton prices fall. For onions and tomatoes, GOS has RISKS introduced occasional import bans to support domestic Other sector risks arise from changes in the broader polit- prices during times of surplus. Nevertheless, the timing of ical and economic environment in which agriculture the imposition and removal of such bans does not appear operates. These changes can be both internal and exter- to be understood by stakeholders. There are no clear cri- nal. Agriculture sector policy and regulation are a source teria for the changes in access, and the delay between the Agricultural Sector Risk Assessment 31 announcement and implementation of such bans or their sector if GOS were to be compelled to lift the ban. In addi- removal, which should be at least 120 days to allow grow- tion, uncertainty over how the issue will be resolved can ers to react, is often much less than this. adversely affect decision making and dampen investments. The analysis also highlights major concerns about land Similarly, the analysis highlights weaknesses within the access, land tenure, and user rights of livestock owners. animal health service delivery system that merit added As in many arid countries where nomadic pastoralism attention. Following structural reforms in the 1980s, live- and transhumance is important, the land tenure systems stock vaccination has largely been handed over to the pri- are pluralistic and complicated, and with increasing glo- vate sector. Overall, livestock vaccination coverage has balization and settlement often livestock owners’ rights dropped to approximately 20 percent of smallstock and are not only not recognized but often not even under- 63 percent of cattle, significantly short of the 80 percent stood. As such, large-scale land acquisitions, expanding coverage required. This gap substantially increases the agriculture and irrigation often take place without suffi- risk of heavy losses when disease outbreaks occur. Quality cient consideration of risks posed to livestock owners. control on the import, sale, and use of veterinary drugs is There are very few cases in which such developments poorly monitored, contributing to added uncertainty over adequately address the pastoralists’ needs. Such develop- the efficacy of treatment and prevention programs at the ments should not be outlawed, as they often increase pri- farmer level. mary production and can provide multiple opportunities for livestock owners; however, their design and implemen- Although private sector involvement is common in almost tation must be done in conjunction with all the stakehold- African countries, there remain questions about the eco- ers including pastoralists. nomic viability of veterinarians operating in remote extensive pastoralist systems. Indeed, many of these pri- Livestock producers who provided input for this study vate sector operators may not be operating within the noted, in particular, their concerns over access to land nationally and internationally required norms. One pri- alienated under the Grande Offensive Agricole pour la vate vet interviewed estimated more than 50 percent of Nourriture et l’Abondance for large agricultural projects, drugs sold are counterfeit and 40 percent of chickens which had previously supplied dry season grazing and eaten by humans in Ndiaye show signs of antibiotic resi- access to riverine pastures. The speed and lack of consul- dues in meat. Table 3.5 illustrates the challenges of ani- tation associated with the process have engendered a per- mal health service provision in Senegal, bearing in mind ceived risk among pastoralists that their livelihood may be that to achieve effective disease control requires 80 per- less sustainable than was originally understood. The lack cent vaccination coverage. There is a general recognition of consistency in the application of regulations regarding that implementation of the law is weak and chronically animal movement in the event of disease, or even in the underfunded. The Ministry of Livestock requires a budget extent to which vaccination might be effectively carried of CFA 3 billion (US$600 million) per year. It receives out, also contributes to uncertainty that will constrain the one-third of this amount, of which 60 percent is spent on extent to which livestock producers are willing to invest in vaccination. The Ministry also recognizes the noncom- production. petitiveness of local vaccine production. Senegal’s existing trade embargo on imported poultry products protects the country’s emerging poultry industry CONFLICTS, THEFT, AND INSECURITY from the threat of avian influenza while keeping competi- Where conflicts occur, unrestrained by the rule of law, then tion from lower-price imports at bay. However, there is the impact of risk is considerable. Since 1982, the region strong international pressure from major exporters such as of Casamance has been affected by internal conflict and Brazil and the United States who charge the ban is unlaw- tensions. An estimated 30,000–60,000 people have been ful and unnecessary. Although the legal procedures may displaced and agricultural production in Casamance take a long time to play out, there is a major risk to the reduced by 50 percent since 1985 (World Bank 2013). Sen- whole viability of the Senegalese commercial poultry egal’s richest agricultural region, Casamance is plagued 32 Senegal TABLE 3.5. VACCINATION COVERAGE IN SENEGAL, 2009 Population Number Animals Animal of Animals Vaccinated Population Aim/Objective Targeted 2008–09 Vaccinated (%) Lumpy skin disease/dermatose nodulaire contagieuse 3,136,500 934,057 29.8 bovine (DNCB), nationwide Foot and mouth diseases/fièvre aphteuse 3,136,500 35,863 1.1 Specifically targeted exotic breeds Pasteurellosis bovine 3,136,500 70,456 2.2 Not nationwide, but many regions affected Horse sickness—nationwide 517,634 134,362 26.0 Peste des petits ruminants—nationwide 9,259,450 1,644,254 17.8 Newcastle disease (backyard and village level) 22,077,800 166,319 0.8 Botulisme—East and North Sénégal 465,600 48,531 10.4 Pasteurellosis in smallstock—paid for by livestock owners 9,259,450 43,454 0.5 Clavelée—paid for by livestock owners 9,259,450 35,708 0.4 Source: Gary et al. 2010. with chronic food insecurity. In 2014, the region had the clashes over natural resources. Their frequent mention in highest levels of hunger, with 37 percent of households, Senegal is a notable concern. representing some 1.8 million people or 14 percent of Sen- egal’s population, facing food shortages. Ten percent of The effects of conflict include reduced mobility by trad- households faced severe food insecurity. Food shortages in ers, veterinary auxiliaries, and private vets who are unable Casamance are often aggravated by the low-intensity to reach livestock owners in time; seasonal labor move- rebellion by the Casamance Democratic Forces Movement ment and some agricultural activities are curtailed (for (MFDC) that began in 1982 and, albeit infrequently, ham- example, transplanting of rice, harvesting); access to mar- pers agricultural transport, trade, and other socioeconomic kets is reduced and markets become less efficient. More- activities. It has also restricted access to farms because of over, government officials, vehicles and services are unable the proliferation of landmines. The most affected areas are to reach these insecure areas, and vaccination and veteri- in Sindian in northern Casamance and in the south near nary services are discontinued. For example, despite a the border with Guinea-Bissau. major outbreak of LSD in 2008, few animals were vacci- nated, leading to higher mortality rates. In addition, con- Conflict and tensions between herders and crop growers flict leads to increased larceny and cattle rustling and a were highlighted as risks by several interlocutors during sense of impunity among criminals, increased violence the field mission and are commonly stated in the literature and even death of livestock owners, as well as increased as an increasing problem in the extensive Sahel pastoralist “unofficial” taxation by armed forces and others. It is systems. Indeed, much of the conflict in the region (Cen- worth noting that theft is not restricted to the Casamance tral African Republic, Niger, Mali, and Darfur) can be region alone. Throughout the survey, the problem of theft traced back to divisions between farmers and herders and was repeatedly cited by producers and traders. Agricultural Sector Risk Assessment 33 CHAPTER FOUR ADVERSE IMPACTS OF AGRICULTURAL RISKS The frequency, severity, and costs of adverse events are analyzed in this chapter as the basis for prioritizing the various sources of risk. The conceptual and methodological basis described below is then applied to production, market, and enabling environ- ment risks. The various sources of risk are then reviewed to discern the most critical. CONCEPTUAL AND METHODOLOGICAL BASIS FOR ANALYSIS For the purposes of this study, risk is defined as an exposure to a significant financial loss or other adverse outcome whose occurrence and severity is unpredictable. Risk thus implies exposure to substantive losses, over and above the normal costs of doing business. In agriculture, farmers incur moderate losses each year as the result of unex- pected events such as suboptimal climatic conditions at different times in the produc- tion cycle and/or modest departures from expected output or input prices. Risk refers to the more severe and unpredictable adverse events that occur beyond these smaller events. This concept differs from the common perception of “risk” by farmers and traders, based on the year-to-year variability of production and prices. It should also be distin- guished from constraints, which are predictable and constant limitations to productiv- ity and growth and which contribute to inefficiencies in production and marketing systems. LOSS THRESHOLDS As agricultural production is inherently variable, the immediate step for analysis is to define loss thresholds, which distinguish adverse events from smaller, inter-annual varia- tions in output. This is achieved by first estimating a time trend of “expected” produc- tion in any given year, based on actual production, and treating the downside difference between actual and expected production as a measure of loss. A loss threshold of 0.33 standard deviation from trend is then set to distinguish between losses resulting from Agricultural Sector Risk Assessment 35 FIGURE 4.1. TIMELINE OF MAJOR SHOCKS TO AGRICULTURAL PRODUCTION IN SENEGAL (2004–06 = 100), 1980–2012 180 Crop production index Food production index Livestock production index 160 140 120 100 Late/erratic 80 Locusts, rainfall; 2004 Erratic locusts, Locusts, Locusts, Late rains, rainfall; 2011 60 1988 1992 regional Severe birds, 40 Severe droughts, drought; 2007 Severe cold rains; drought; 1996–98 drought, locusts, 20 1980 locusts, 1983–84 2002 0 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 Source: World Development Indicators; Authors’ notes. adverse events and those that reflect the normal costs of considered the most suitable. These data allow the analy- doing business. Those below threshold deviations from sis of risk over a 33-year period. trend allow estimation of the frequency, severity, and cost of loss for a given time period (see appendix D for illustra- tions of indicative crop loss estimates). The frequency and PRODUCTION RISKS severity of losses derived in this manner were also checked Based on analysis of available quantitative and qualitative against historical records to ensure consistency with actual data, the most common risks to agricultural production in adverse events. Senegal are drought, locust outbreaks, and flooding. The incidence of these and other adverse events is indicated in figure 4.1, largely based on reports of adverse events for THE INDICATIVE VALUE OF LOSSES the period 1980–2012. During the 33-year period 1980– Available data on actual losses resulting from adverse 2012, Senegal’s agricultural sector has been subjected to events are not always accurate or consistent enough to at least 10 major shocks. Erratic rainfall and drought facilitate comparison and ranking of the costs of adverse emerge as the most common sources of production events. Analysis was thus based on estimates of the “indic- shocks, followed by locusts. Related risk events may occur ative” value of losses, which provide a more effective basis in isolation, but can also present as multiple, overlapping for comparison. Indicative loss values are also compared shocks—as was the case in 1984, 2002, 2007, and 2011— with the value of agricultural GDP in the relevant year to with far greater impacts and higher associated losses. provide a relative measure of the magnitude of loss. Although these estimates draw on actual data as much as Measured in terms of gross agricultural value,8 crop pro- possible, it is emphasized that they represent indicative, duction in Senegal was significantly reduced 11 times by not actual losses. adverse events during the period 1980–2012, for an over- all frequency of one in three to four years on average DATA SOURCES (table 4.1). All but three of these events resulted in a drop Analysis of this nature requires a consistent set of data on in aggregate production value of 10 to 30 percent. It is both production and prices for an extended time period. worth noting that three (1980, 1983, 2002) of the four Of the various sources of data available, FAOSTAT’s data series on the value of gross agricultural production 8 Gross aggregate value is the total value of volume of production for each crop (1980–2012) and crop production (1980–2012) was multiplied by the producer price. 36 Senegal TABLE 4.1. COST OF ADVERSE EVENTS FOR CROP PRODUCTION,* 1980–2012 Indicative Loss Value US$ % of Gross Year Description (in millions) Prod. Value* 1980 Severe drought affecting 8 of 14 regions, incl. Kaolack, Kaffrine, Fatick; –128.2 –19 –333 K MT of gnuts 1983 12 of 14 provinces affected by severe drought; locusts; 340 K MT of –128.5 –19.3 groundnuts lost 1984 Regional droughts in Matam, Tambacounda, –91.7 –13.8 1992 Regional droughts; locusts infestation; estimated 205 K MT groundnuts lost –69.8 –10.5 1996 Regional droughts in Diourbel, Kaolack, Kolda –50.3 –7.5 1998 Late, erratic rains; more than 172 K MT of losses in maize, millet, –47.9 –7.2 sorghum, cotton 1999 Delayed start to the season; erratic rainfall –33.5 –5.0 2002 Severe drought affecting 50% of country; locust infestation; 420 K MT of –217.3 –32.6 groundnuts lost 2004 Locust infestation; over 133 K and 49 K MT of millet and cowpea lost –68.6 –10.3 2007 Erratic rainfall; birds; substantial losses in millet, sorghum, maize, and –110.8 –16.6 groundnuts 2011 Erratic rainfall and locust outbreak affects maize, millet, and groundnut –97.7 –14.7 production Sources: FAOSTAT. Note: Cowpea losses were included from 1989 forward because of data availability. Potato losses were calculated from 1980 to 2004 because of inconsistencies in data thereafter. *Average 2005–07. years with the highest crop losses coincided with extreme TABLE 4.2. COEFFICIENTS OF VARIATION FOR drought events, whereas poor rainfall affecting key crop CROP PRODUCTION, 1980–2012 production areas coupled with locust and bird infestation Production Area Yield contributed to high crop losses in 2007 and 2011. Maize 0.70 0.36 0.38 Although figure 4.1 highlights the frequency of major Groundnuts 0.32 0.20 0.24 risk events, it does not show the extent of the risk asso- Onions 0.72 0.54 0.25 ciated with individual crops in different provinces, nor Rice 0.62 0.27 0.30 the severity of the losses that have occurred. The Millet 0.25 0.12 0.18 degree of risk can be partially estimated from the vari- Sorghum 0.30 0.25 0.17 ability of yield, as indicated by the coefficient of varia- Cowpeas 0.68 0.51 0.30 tion (table 4.2). This shows that maize yields are the Tomatoes 0.88 0.75 0.35 Cotton 0.36 0.25 0.24 most variable, closely followed by tomatoes, rice, and Potatoes 0.33 0.32 0.26 cowpeas. The variability of tomato yields is unexpected given the widespread use of drip irrigation to grow the Source: FAOSTAT. Note: For onions and tomatoes, 2012 was not available. Instead, data through crop and most probably reflects the impact of pests and 2011 were used. diseases. Yield variation for millet and sorghum are the lowest among the crops assessed, as might be expected Because it is based upon deviations from a trend observed for crops that show a high degree of tolerance to mois- at a national level, a national-level analysis may well ture stress, whereas cotton, groundnuts, potatoes, and underestimate the impact of risk events at the departmen- onions exhibit moderate levels of variability. tal level wherein the observed extent of variation can be Agricultural Sector Risk Assessment 37 expected to be much higher. Moreover, the quantification suggest that over the 33-year period, shocks reduced pro- of shock impacts is in fact a proxy measure of associated duction by 2.15 million metric tons (that is, an average of risk. The ex post impacts are but one aspect of risk that 6.4 percent of agricultural GDP in loss years). It is worth also affects production through its ex ante influence upon noting that the highest annual losses occurred in 2002 investment. Hence, absolute accuracy is less relevant as when the industry underwent a major restructuring with long as the relative impacts of different shocks are cor- the privatization of Sonacos and the closure of its sub- rectly assessed. sidiary, Sonagraines. This restructuring coincided with one of the worst droughts in 20 years and more than a 75 As an example, figure 4.2 illustrates the analysis of indi- percent drop from peak output of more than 1 million cate losses for groundnut production at the national level MT two years earlier. It took another seven years (2009) during the period of review (1980–2012). The results before output would return to precrisis levels. The results FIGURE 4.2. INDICATIVE LOSSES FROM RISK EVENTS TO GROUNDNUT PRODUCTION, 1980–2012 1.20 Yield (tonnes/ha) Trend .33 Trend 1.00 0.80 0.60 Drought Drought, Erratic locusts Erratic rainfall, 0.40 Severe Drought, rianfall, locusts drought locusts locusts Extreme 0.20 drought 0.00 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 Source: FAOSTAT. TABLE 4.3. INDICATIVE LOSSES FOR MAJOR CROPS, 1980–2012 Crop Frequency Production Loss (MT) Value (US$, millions) Maize 0.36 –388,459 –94.3 Groundnuts 0.30 –2,152,174 –635.2 Cotton 0.33 –102,795 –37.4 Onion 0.24 –131,867 –72.5 Tomato 0.27 –184,969 –33.0 Mango 0.42 –120,355 –47.8 Bean, green* 0.28 –16,173 –8.3 Millet 0.30 –998,878 –219.7 Sorghum 0.33 –250,622 –55.1 Rice 0.24 –335,423 –106.6 Cowpea 0.30 –136,503 –89.7 Total –4,818,218 –1,402.4 Source: FAOSTAT. Note: Price are averaged producer prices during 2005–07. *Covers the period 1988–2012 only. 38 Senegal of similar analyses conducted for other staple food and TABLE 4.4. DATES AND FREQUENCIES OF cash crops are shown in table 4.2. The crop-specific anal- AGRICULTURAL RISK EVENTS yses are summarized in appendix D. Frequency Risk Event Year (33 years) The results of the trend analyses indicate that for the 11 target crops analyzed, the total cumulative loss of produc- Locusts (migratory 1983–84; 1987–88; 0.181 locusts, 1992; 2002; tion over the 33-year period was approximately 4.82 mil- grasshoppers) 2004–05; lion MT, with an estimated value of US$1.40 billion, or 2011–12 3.9 percent of agricultural GDP on an average annual Birds 1994; 2007 0.061 basis. Among crops, maize exhibited the highest level of Drought; erratic 1980; 1983–84; 0.30 vulnerability in terms of frequency, whereas groundnuts rainfall 1990; 1996–98; incurred the highest losses, accounting for nearly 45 per- 2002; 2007; cent of aggregate losses. 2011 Flooding 1989; 1999–2000; 0.27 Though average annual impact of shocks on agricultural 2003; 2005; 2008–09; 2010; GDP is relatively limited (less than 4 percent on average), 2012 it is the relative impacts between crops that are of greatest Armyworm 2010–11 0.061 significance. Thus, of the 11 crops assessed, shocks to groundnut production account for nearly half (46 per- cent) of the total ex post impact, reflecting the importance of that crop to national GDP. Despite its low coefficient of FIGURE 4.3. PROPORTIONAL IMPACT OF variation of yield, millet shows the greatest impact of VARIOUS ADVERSE RISK shock after groundnuts. This again reflects the large area EVENTS BY CROP, 1980–2012 under cultivation, but it also reflects the higher variability Pests Drought Other of rainfall in the lower rainfall areas where millet produc- 1.00 tion predominates. Among the other crops, maize and Proportion of loss 0.75 cowpeas show similar impacts in terms of the size of loss. 0.50 The absolute volume of loss of cowpeas is relatively small, 0.25 but the higher value of the crop results in greater losses. Relatively limited aggregate losses for cotton reflect both 0.00 Millet Maize Sorghum Groundnuts Cotton Cowpeas the relative tolerance of the top to erratic rainfall as well as the area under cultivation. and other sources for each crop to indicate the propor- tional impact of different shocks on each crop. This analy- IMPACTS OF sis was done for the six main field crops (millet, maize, PRODUCTION RISKS groundnuts, sorghum, cotton, and cowpeas) with results The attribution of yield loss to specific shocks is inevitably shown in figure 4.3. an approximation, but it is nevertheless useful to compare the losses experienced during different years and thereby The data show the relatively minor impact of pests (mainly to determine the relative impact of different risk events. locusts) and the dominant impact of drought (or more Table 4.4 indicates the years when specific shocks occurred properly, drought and erratic rainfall). Significantly, and their frequency over the period 1980–2012. although cowpeas experienced loss of yield in some years, such losses occurred when neither drought nor locusts The frequencies calculated in table 4.4 can be combined were prevalent so that all losses appeared to have been with the yield loss data calculated from the trend analyses caused by other factors. Agricultural Sector Risk Assessment 39 FIGURE 4.4. FREQUENCY AND CUMULATIVE SUMMARY OF IMPACTS IMPACT OF VARIOUS ADVERSE There are insufficient data available to separate the differ- RISK EVENTS, 1980–2012 ent impacts of specific risk events/shocks with a high 1000 degree of accuracy, or to develop an accurate assessment 800 of actual losses incurred because of these events at a local in US$ millions 600 level. Nevertheless, it is possible to draw some broad con- 400 200 clusions from this analysis, namely: 0 » Adverse impacts on Senegal’s agricultural produc- –200 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 tion from risk events are equivalent to at least 3.9 Frequency percent of agricultural GDP on average. Locusts Birds Erratic rainfall/drought Flooding Armyworm » Senegalese agriculture is subject to losses exceed- ing 10 percent of gross production value in one out of every five or six years on average due to unman- Results of the analysis can also be summarized in terms of aged risks. the frequencies and expected losses associated with the » The most significant cause of loss is drought/ main risks to crop production (figure 4.4). The frequency erratic rainfall, which accounts for approximately of each risk is based on its occurrence during the period 50 percent of crop yield reductions, particularly 1980–2012. The associated loss is an estimate of the for groundnuts and cereal crops. Pests and diseases, indicative costs for each type of risk during the period of especially locusts, account for a further 25 percent. analysis. The graph clearly shows drought as the major » Variation in price can be a significant risk to crop source of risk, causing the highest losses. However, it is farmers, particularly horticultural producers, even worth noting that all these events and associated losses when prices have been set by institutional buyers. occurred prior to 2003; there have been no incidences of » For livestock producers, production risk is largely widespread drought over the past decade. Locusts emerge related to disease and to occasional devastating as the next most important source of crop production risk events such as cold rain. after drought, whereas the high incidences of floods, par- » Market shocks can occasionally occur, but livestock ticularly in the last decade, causes limited damage to crop producers are under less pressure to sell their produce production. than are crop producers, so the overall risk is less. 40 Senegal CHAPTER FIVE VULNERABILITY ANALYSIS All stakeholders in agricultural value chains assume some level of risk. Vulnerability presents when the potential impact of risk events is greater than a stakeholder’s capac- ity to absorb adverse impacts. This may be because knowledge of the frequencies or impacts of risk events is inadequate (as a result, for example, of unforeseen changes in climate) or because actions that might otherwise limit the level of risk assumed are constrained by circumstances so that involuntary exposure to risk is increased. (This is especially true where capacity for adaptation lags behind changing circumstances.) A more comprehensive analysis of the factors contributing to vulnerability is presented in appendix C. Its key conclusions and relevance to different types of stakeholders are listed below. Table 5.1 provides a simple ranking of risks reported by smallholders (Wane and Galandou 2012). The most oft-cited risk, reported by almost 50 percent of households, highlights their perceived vulnerability to food price shocks. This is indicative of high levels of market dependence because of insufficient productive capacity together with low levels of off-farm income. It is perhaps notable that the most important vulnerabil- ity is not to a risk inherent to agricultural production but to rural households as con- sumers. Initiatives designed to increase the prices paid for staple crops may well expose households to higher levels of vulnerability if not accompanied by other measures. The second most oft-cited risk also relates to cash dependence and the limited finan- cial resources of producers that constrain their capacity to deal with increases in the costs of inputs. It is again significant that although this shock does affect agricultural production, it is mediated through local input markets rather than through a physical event such as drought or disease. The same is true for the fourth most important area of vulnerability, that is, the loss of productive capacity through death, disease, or dis- ability, which altogether affect almost 25 percent of households. These shocks may affect agricultural production, but the effect is not upon the agricultural production system itself. It is only the third most frequently cited risk event (poor rains), reported by nearly one- quarter (23.8 percent) of households, that has a direct effect on agricultural production Agricultural Sector Risk Assessment 41 TABLE 5.1. FREQUENCY OF RISK EVENTS AND PERCENTAGE AFFECTED Type of Shock No. of People Affected Share of Total (%) Rising prices of food 2,878 48.9 Increase in the price of inputs/farm equipment 1,891 32.1 Poor rains 1,402 23.8 Illness/accident of household member 1,221 20.7 Decline in the price of products sold by the household 560 9.5 Theft of property or animals 533 9.0 Animal disease/death of animals 468 7.9 Insecurity 455 7.7 Animal damage 439 7.5 Animal disease/death of animals (cattle) 407 6.9 Invasion of pests/granivorous birds 351 6.0 Diseases of plants 207 3.5 Loss of employment or unemployment 199 3.4 Floods 134 2.3 Fire/bushfires 74 1/3 Conflicts 72 1.2 Source: Wane and Galandou 2012; Authors’ notes. systems. Moreover, although the impacts of most of the PASTORALISTS subsequently listed shocks, including pests and diseases, It is widely recognized that the least vulnerable of liveli- theft, floods, and fires, also directly affect agricultural pro- hood groups are the nomadic pastoralists, who although duction, three of the four most important areas of vulner- living in arid areas have adopted a lifestyle that enables ability lie beyond agriculture and reflect inadequate them to mitigate the risk of drought by continually household income on the one hand and limited health moving to areas of fresh grazing. Studies have shown service capacity on the other. that as long as the mobility of pastoralists is not con- strained, their vulnerability to weather risks such as Nevertheless, table 5.1 shows that rural households are drought is low. However, if pastoralist herds are unable vulnerable to a variety of agricultural shocks. In practice, to move freely to new grazing areas, their chances of that vulnerability has been reduced through a variety of survival are substantially reduced. Ready access to ani- practices that limit their exposure to risk. In an agricul- mal protein contributes to a balanced diet, and the tural context, vulnerability to risk can be considered from characteristic of livestock to maintain value allows them the perspective of those practices and the extent to which to be used as a source of finance that is both mobile and they can be employed by different types of agricultural little prone to decay (in marked contrast to the produc- supply chain actors. tion of crop farmers). Nevertheless, the pastoralist livelihood is particularly vul- STAKEHOLDER AND nerable to four risks: namely, widespread (regional) drought, locust infestations, severe animal disease out- LIVELIHOOD RISK PROFILES breaks, and constraints upon movement. The frequency This section profiles various types of agricultural stake- of the first three risks is low, though ex post impacts of holders in terms of some of the key risks they face and related shocks can be high. However, growing land pres- their capacity to mitigate such risks and recover from sures and other factors are increasingly inhibiting tran- related shocks. shumance, and with that, the ability of pastoralist 42 Senegal households to mitigate risk and rebound from shocks of animals either for draft, for home consumption, or for when they manifest. cash. These smallholders make a significant contribution to the national livestock herd, but their vulnerability is related more to their cropping activities. Indeed, for these OTHER LIVESTOCK PRODUCERS households, livestock may help to reduce vulnerability Livestock are raised by at least three other types of pro- because they can be sold as a source of cash in the event ducers. Agro-pastoralists who are transitioning out of pas- of crop failure. toralism are among the most vulnerable households. Such households generally possess few livestock and occupy lands on the margins of pastoral areas—lands that are inherently low in rainfall. These households are thus DRY LAND SMALLHOLDER FARMERS exposed to low and variable rainfall, but their sedentary The risk of inadequate moisture renders dry land small- lifestyle denies them the advantages of mobile pastoralism holders more vulnerable to production risk than their so that capacity for risk avoidance is limited. Few house- irrigated counterparts. This is reflected in the lower lev- holds adopt an agro-pastoralist livelihood by choice. els of inputs applied to dry land crops. Not only does Instead, they are obliged to do so by changing circum- the lower anticipated average yield not justify the same stances, especially reduced access to grazing. As a result, level of investment, but also the increased probability agro-pastoralists tend to be the poorest households, lack- of risk events that could result in little or no return acts ing the resources needed to absorb the impact of shocks, as an additional hindrance to investment. From this which further enhances their vulnerability. perspective, the advantage of irrigation is not only that it allows greater yields to be achieved, but also that by Intensive livestock producers are constrained by the avail- enhancing certainty that such yields are attainable, it ability of land to support their livestock. They are obliged justifies a much higher level of investment into the crop. to use supplementary feed to fatten their animals, which Thus, the impact of irrigation upon average yields is are often kept in close confinement. The extent to which usually much greater than the yield increase that might intensive producers are more or less vulnerable to risk be ascribed to the better availability of adequate than their extensive counterparts is debatable. On the moisture alone. one hand, the purchase of feed allows intensive produc- ers to be largely independent of weather and grazing, whereas the confinement of livestock can reduce expo- IRRIGATED SMALLHOLDER FARMERS sure to disease. On the other hand, a higher level of man- Generally, irrigated smallholder producers face far less agement is required to ensure consistent rates of growth risk of inadequate moisture than their rain-fed counter- and disease outbreaks, when they do occur, can be more parts. They therefore can be considered to be less vulner- debilitating than those experienced under extensive pro- able. This is correct insofar as a given level of production duction. The greater level of investment into intensive is concerned, but it is not inevitably so. Growers of irri- livestock production systems also increases the potential gated crops are often obliged to adopt a more intensive vulnerability of producers who have more to lose in the approach to production to generate the revenues needed event of a disaster. On balance, it would appear that to cover the cost of the irrigation systems. Whereas the whereas the impacts of risk events upon intensive live- greatest component of agricultural risk (insufficient mois- stock production may be greater than those experienced ture) may be controlled, the level of investment under irri- by extensive producers, most intensive producers have a gation is generally increased so that the overall risk may greater capacity both to prevent such events and to with- not be reduced unless all other potential risk factors can stand their impacts and are hence less vulnerable to the also be controlled. This implies the need for a high level of impacts of risk events. technical competence in pest and disease control and in overall crop management, as well as a greater capacity to Finally there are those smallholders whose main liveli- manage losses when they do occur. From this perspective, hood is derived from crops but who raise a small number the risk faced by a smallholder is not reduced through the Agricultural Sector Risk Assessment 43 introduction of irrigation systems unless that system has TRADERS low fixed costs and/or the smallholder possesses the nec- Traders are primarily vulnerable to risks caused by mar- essary agricultural acumen to avoid losses caused by other ket uncertainty. In particular, whereas mobile phones risk factors such as pests, diseases, mismanagement, and allow traders good access to immediate price information, market prices. they have poor knowledge of market volumes or of the extent of production. As a result of this vulnerability, few COMMERCIAL FARMERS traders are willing to accumulate large positions with the Commercial farmers face the same risks as smallholders, intention of selling at higher prices. The risk that they but their levels of vulnerability may differ according to the might not be able to sell their stock owing to declining type of risk. On the one hand, access to paid labor and prices or to the presence of other traders in the same mar- machinery can reduce exposure to some risks, whereas ket limits most traders to short-term back-to-back trades. higher levels of savings can allow commercial producers Similarly, whereas traders might discover a higher price of to absorb the impacts of risk events. On the other hand, a commodity in a remote area, they are unlikely to pur- commercial producers face a greater level of market risk chase and transport large volumes to take advantage of than their subsistence counterparts. In general, commer- that market because they are uncertain that the potential cial production systems have an increased reliance upon demand will be great enough to justify the expenditure. As management practices that reduce the risk of loss (includ- a result, most markets are both temporally and spatially ing the use of insecticides and fungicides), but the overall fragmented so that both seasonal fluctuations and geo- increased intensity of production can increase vulnerabil- graphic disparities in prices can be significant. ity in the event of an unforeseen loss or a breakdown in crop protection practices. INCOME LEVELS Income level is a variable characteristic of many liveli- PROCESSORS hoods that has a dramatic effect upon vulnerability to risk Processors in Senegal include groundnut processors, mill- and is therefore worthy of consideration in its own right. ers, and cotton ginners. All of these are vulnerable to mar- Studies have shown that risk is perceived relative to one’s ket risk because of increased local prices and/or reduced capacity to absorb that risk so that for households of low costs of competing imports. Cotton and groundnut pro- income, the perception of a given risk is much greater cessors are also vulnerable to the risk of aflatoxin con- than it is for those of higher income. This is particularly tamination, which cannot be detected in the unprocessed evident for agricultural households of limited resources, materials but which can render the final products unsale- who might be obliged to pledge vital assets (such as oxen, able if detected at a later date. Processors also face the risk donkey carts, or savings) as collateral for a loan to increase of inadequate supplies because of either poor production food production through intensification. The impact of a or a redirection of inputs toward food crops for household low yield that for such a household could result in a failure consumption, especially after a poor harvest. In the case to repay and consequent loss of assets can be so debilitat- of groundnuts, in cases in which the buying price is prede- ing that the loans are often refused. This behavior has led termined, processors may also be vulnerable to political to the general observation that smallholder farmers are expediency that results in a price that is higher than the risk averse. This is not necessarily the case, a more accu- market can bear. rate observation being that poorer farmers are risk averse. 44 Senegal CHAPTER SIX RISK PRIORITIZATION AND MANAGEMENT Previous sections have highlighted sources of risks that are pervasive across the Senega- lese agriculture landscape. These risks are both numerous and complex. They manifest with varying levels of frequency and severity, and can cause substantial losses to crops and livestock, with profound short-term and long-term impacts on income and livelihoods. Putting in place effective risk management measures can help mitigate adverse impacts on agricultural supply chains and the livelihoods they support. However, it is virtually impossible to address all risks at once. Thus, it is necessary to prioritize interventions based on which risks occur most frequently and which cause the greatest financial losses. RISK PRIORITIZATION Using quantitative measures and anecdotal evidence collected directly from stakehold- ers, this analysis has evaluated risks for the crop and livestock subsectors. Owing to the lack of reliable data, some of the risks could not be quantified. In such cases, the assessment team relied more on qualitative measures. Based on the team’s combined quantitative and qualitative assessment, table 6.1 prioritizes the most important risks. This prioritization was presented during a roundtable at MEF in Dakar on March 21, 2014. It provides a basic ranking of agricultural risks on the basis of the probability of the event occurring and the anticipated impacts in terms of financial losses. The iden- tified risks located in the grayer areas represent the most significant risks. Overall, this prioritization identified (1) erratic rainfall, punctuated by drought; (2) locusts; (3) price volatility; and (4) crop pest and diseases as the most important risks facing Senegal’s agricultural sector. Parasitic weeds such as Striga, aflatoxin contami- nation (maize, groundnuts) and other postharvest threats, and livestock diseases were also deemed important, but to a lesser extent. It is worth noting that incidences of drought have declined significantly within the most recent decade, but it remains unclear whether this change is temporary or rather the result of a shift in long-term weather patterns. Although an observed wetting over roughly the same period has led to recurrent flooding in many parts of the country, affecting urban and coastal areas, impacts on crop and livestock production have been limited at the aggregate level. Agricultural Sector Risk Assessment 45 TABLE 6.1. RISK PRIORITIZATION MATRIX Negligible Moderate Considerable Critical Catastrophic Highly probable Tifa (R-I) Pests and diseases Perceived Striga (M,S) inconsistency of Price volatility (dom policy (L) crops.) Newcastle (L) Probable Birds (M, S) Pests and Diseases Erratic access to Erratic rainfall Bushfires (L) (P, T, C, Ma) quality inputs (S, M, R-R, Mz,C, Erratic access and services G) to quality (M, S, G, Ct) Locusts medicines/ Dom/Int’l price Price Volatility (T, O, vaccines (L) variability G, C) (R, Mz) Pest/disease (R) Birds (R-I) Aflatoxin (Mz, G) Occasional Termites (All crops) Windstorms (M, S, Price volatility for Localized Severe drought Theft (M, R-R, MZ, L) R-R) poultry feed Drought Absence of regional Temperature Cold season rains (L) standards (R-I) variability (R-I) Contract default/ Floods (All crops) counterparty risk Regional conflict (All crops) Release of Food Fires (M) stocks (R-I, R-R) Rare Source: World Bank. Key: Sorghum, Millet, Rice–Irrigated, Rice–Rain-fed, Maize, Tomato, Potato, Mango, Cowpeas, Groundnuts, Cotton, Livestock, Onion. The above assignment of risk priorities is in part based This prioritization exercise is based upon the direct upon the responses of stakeholders collected through impacts of risk events. The indirect ex ante impacts are interviews over the course of the assessment. From that less easily quantified, but can for the most part be consid- exercise, it is possible to list the priority risks faced by each ered proportional to the ex post loss of profitability. Thus, subsector (table 6.2). It is notable that only cash crop pro- from the perspective of prioritization, an analysis of ex ducers highlighted severe drought as the most important ante responses to risk is unnecessary. risk they faced. The listing of priority risks clearly shows the major impor- tance of erratic rainfall as the main risk and of drought as AGRICULTURAL RISK a more extreme but less frequent expression of the same MANAGEMENT phenomenon. After this, the other priorities vary among The various risk events that give rise to both ex post losses crops and it is hard to identify a consistent theme. How- and reduced ex ante investment can be addressed at three ever, the reporting of locusts as a primary, secondary, and levels, according to impact and frequency as shown in tertiary threat would make this phenomenon the second figure 6.1. aggregate priority for the agricultural sector. Beyond this point, aggregation of crop-specific risk loses meaning and Risk mitigation—Responding to risk events of high fre- the next level of priority could well be accorded to price quency and low impact, the adoption of techniques or volatility, pests, and diseases, and perceived inconsisten- behaviors that reduce the impact of the event on produc- cies in policy implementation in equal measure. tion or profitability. Risk mitigation would include the 46 Senegal TABLE 6.2. LISTING OF PRIORITY RISKS BY COMMODITY Risk Commodity Priority #1 Priority #2 Priority #3 Sorghum Erratic rainfall Locusts Timely access to quality inputs Millet Erratic rainfall Pests, diseases, and Striga Timely access to quality inputs Rice (Irrigated) Bird damage Price volatility Water weed (Tifa) Rice (Rain fed) Erratic rainfall Price volatility Windstorms Tomatoes Price volatility Inconsistent policy Diseases and pests Potatoes Price volatility Locusts Disease/quality of inputs Beans Locusts Diseases Access to inputs Mango Drought Locusts Price volatility Cowpeas Erratic rainfall Drought Locusts Groundnuts Drought Logistics breakdown Disease Cotton Drought Logistics breakdown Locusts Livestock Erratic rainfall Inconsistent policy Poor policy implementation Aggregate for Sector Erratic rainfall/drought Locusts Source: World Bank FIGURE 6.1. INTEGRATED RISK-LAYERING impact of risk, but to absorb it. For example, food-insecure SOLUTIONS households can be provided with food aid or cash, whereas others that may have taken out loans can in the event of dis- Layer 3 aster be afforded debt relief or debt restructuring. Coping Probability Very low frequency, Layer 2 very high losses Risk mitigation strategies may be applied on an ex ante or ex post basis. Low frequency, + Risk transfer medium losses + Risk coping Layer 1 Risk mitigation Potential responses to key risks are recommended on the + Risk transfer High frequency, basis of the above-referenced risk-layering approach. low losses Risk mitigation Table 6.3 lists the proposed risk management mechanisms that are appropriate to the most significant risk events as Severity defined by the combined impact of their frequency and Source: World Bank. extent of losses incurred. The majority of these are aimed at risk mitigation, although some coping strategies may development of irrigation systems to reduce the impact of also be required. The main risk transfer mechanism rec- drought, or the breeding and distribution of disease or ommended is insurance, which not only can transfer risk drought-resistant crop varieties or animal breeds. impacts, but also can be used in conjunction with the cop- ing strategy of debt relief. Moreover, it is appropriate that Risk transfer—Risk events of low frequency and medium the recent introduction of parametric crop insurance in impact can benefit from both risk mitigation and risk Senegal is closely linked to the provision of credit. A vari- transfer, that is, the transferring of risk to third parties, ety of coping strategies could also be leveraged. Many of either through insurance or through other financial mech- these rely upon social protection mechanisms that may not anisms such as hedging (against currency risk) or the pur- be fully in place and will require development assistance to chase or sale of futures contracts and/or options. implement rapidly. It is notable however that almost all of the population of Senegal is dependent upon access to Coping Strategies—Risk of very low frequency and very high markets for some part of their food security so that rapidly losses, strategies that are designed not to reduce or transfer the implementable market-based solutions can be put in place Agricultural Sector Risk Assessment 47 TABLE 6.3. PROPOSED RISK MANAGEMENT MECHANISMS Mitigation Transfer Coping Drought Promoting development of small-, medium-, Macro- Use of weather index to trigger early and large-scale irrigation; water harvesting level crop warning and response insurance Establishing and improving regional and Farm-level Facilitating temporary migration and national NDVI and early warning systems crop and transhumance (EWS) linked to an effective and early livestock emergency response system. insurance Improve or establish sufficient livestock related Contingent financing and other infrastructure: borehole, fodder reserves, instruments to support coping measures roads, sale yards, abattoirs. Improve livestock feed and forage supply Promoting development of social safety through either local provision or through net programs (for example, food aid, subsidies, vouchers, and so on and Food-for-Work, Cash-for-Work) developing community-level food and fodder/forage (that is, livestock) banks Commercial destocking Livestock supplementary feed programs Debt restructuring/relief Erratic rainfall All the above drought mitigation measures, Farm-level Promoting household/community plus improved access to weather forecasts crop and savings to inform and advise farmers on adequate livestock time for cropping operations insurance Promotion of reforestation Improved access to finance and micro- finance. Promotion of conservation farming technique Debt restructuring/relief Improved access to drought tolerant and short Facilitating temporary migration and season crops and varieties. transhumance Locusts Strengthening early outbreak detection/ Crop Contingent financing and other response systems Insurance instruments to support coping measures Promoting development of social safety net programs (for example, food aid, Food-for-Work, CFW) Debt restructuring/relief Price volatility Adoption of mixed farming and crop rotation Hedging of Social safety net programs currencies (for example, food aid, Food-for-Work) Promote the development of private sector Use of Direct cash payments to affected aggregation points, warehouse capacity commodity households and cold storage facilities as well as the futures and development of warehouse receipt and options inventory credit systems markets Enhanced domestic market capacity, Promoting household savings including strengthening of market linkages and improved access to finance Access to credit during commercialization to Substitutions and/or reductions in avoid any sale household diet Spreading production over time using appropriated varieties Market regulation (import ban) Improving harvesting quality and processing for longer conservation 48 Senegal TABLE 6.3. PROPOSED RISK MANAGEMENT MECHANISMS (Continued) Mitigation Transfer Coping Pests/diseases Strengthening early outbreak detection/ Crop Developing social protection programs (crops) response systems insurance (yield-based) Promoting crop rotation and transition to Use of savings and borrowing more pest/disease resistant crops Promoting IPM techniques Direct compensation to affected farmers Diversifying seeds varietals within crops Debt restructuring/relief Strengthening of P/D-tolerant seed development and distribution systems Improving farmer access to agrochemicals Pests/diseases Biosecurity including active surveillance, Livestock Emergency vaccination and treatment (livestock) vaccination and quarantine insurance programs Improved application of existing veterinary Direct subsidies, vouchers standards, laws and policies. Training and capacity building of privatized Quarantine and decentralized animal health services including vet vouchers LEGSa programming in planning LEGS programming in responses Inconsistent Promote the development of business Establish clear emergency animal livestock associations and advocacy groups to enable health response and declaration of sector policy stakeholder and private sector participation emergencies formulation and in agricultural policy formulation implementation Disengagement of government from market Capacity building, training intervention Diversification of agricultural entreprises Enforcement of rules and regulations Increased vertical integration within value Linking emergency response to modern chains development policies. More transparent policy and better enforcement of existing policies and laws Broader community involvement in land use plans to ensure access (forage, water) Source: World Bank. a The Livestock Emergency Guidelines and Standards (LEGS) provide a set of international guidelines and standards for the design, implementation, and assessment of livestock interventions to assist people affected by humanitarian crises. Established in 2005, the LEGS Project is overseen by a Steering Group of individuals from the AU, FAO, the International Committee of the Red Cross, the Feinstein International Center at Tufts University, the World Society for the Protection of Animals, and Vetwork UK. in the event of an unforeseen shortage in domestic produc- IMPROVED WATER AND SOIL tion. Similarly, cash-based debt relief and restructuring MANAGEMENT initiatives can relieve the burden of risk events on house- There is one intervention that can truly mitigate the impact holds that have borrowed to invest in production. of drought, namely, total or supplementary irrigation, according to the circumstances. The introduction of irriga- tion to those areas where it is economically feasible and revi- RISK MANAGEMENT talization and maintenance of existing infrastructure should SOLUTIONS be viewed as a major priority, although earlier studies have The following discusses some of the broad intervention suggested that the economics of irrigated production areas that the government could consider in responding to systems should be carefully scrutinized before development risks prioritized by this assessment. proceeds (Franzel 1979). In particular, irrigation per se is of Agricultural Sector Risk Assessment 49 limited value if it is not accompanied by an increased inten- extract nutrients from a larger volume of soil; and (2) pro- sity of production that allows the fixed costs of irrigation moting timely planting so that the crop can take maximum systems (especially their maintenance and management) to advantage of rains without becoming susceptible to mois- be met fully. This requires access to improved inputs, espe- ture stress. The CA system can substantially improve a cially pesticides and fungicides, and the technical compe- soil’s capacity to absorb water and a crop’s ability to extract tence to ensure optimal yields at a much higher level of it resulting in yield increases of 50 percent and more within production than before. the first year of implementation. It is also well suited for use in fragile soils that might otherwise be subject to erosion From this perspective, the construction of a new irrigation and degradation, since it minimizes cultivation so that soils scheme is less critical than the development (in the case of are not left exposed to wind or rain erosion. communal small-scale irrigation) of an effective water users’ association and irrigation management structure, as Almost all new technical interventions require the small- well as ensuring both the supply of improved inputs and holder to apply additional inputs or labor and increase the providing the necessary training to growers. Finally, not cost of production and the risk associated with it. CA does only must growers be able to access inputs and produce, neither. It allows for reduced labor and requires no but they should be able to market their increased produc- increase in inputs (USAID 2013). From this perspective tion without difficulty as well. The new public-private alone, it reduces risk to the smallholder, but the benefits of partnership (PPP) program for staple crop-processing the technique are such that the risk associated with erratic zones being undertaken by the Nigerian government is an rainfall is also substantially reduced. It is for these reasons example of such integrated market-led development. that CF has achieved high rates (>90 percent) of sustain- able adoption by smallholders (Kabamba & Muimba- From a policy perspective, the high priority of drought as a Kankolongo 2009). GOS together with USAID have risk facing both crop and livestock producers requires con- been promoting the establishment of CA since 2009. It is scious decisions to be made to achieve the best use of a recommended that this initiative should receive greater limited resource (the irrigable land beside perennial rivers). support including the training of agricultural extension On the one hand, it can be rendered extremely productive workers in CA and the promotion of its use among all for crop production, albeit at considerable cost, whereas on smallholders that lack access to irrigation. the other, it can be critical to the survival of many pastoral- ists. It is difficult to reconcile these two priorities, but it is To maximize limited resources, farmers need good and important to develop a land use strategy for such areas that timely information. For example, enhancing farmer access will allow both crop and livestock producers to benefit from to weather forecasting information coupled with technical the limited resource. Unless such a strategy can be not only advice can greatly aid farmers in making better decisions, developed, but also agreed and adhered to by all stakehold- such as the most opportune time to plant. Similarly, as ers, there is a further risk of conflict and suffering. weather becomes more unpredictable, technical advice with regard to crop and variety selection can be custom- ized based on anticipated rainfall patterns and amounts, The impacts of erratic rainfall can be mitigated not only by helping farmers mitigate impacts. irrigation, but also by the small-scale adoption of water conservation practices. Nevertheless, it is the broad-scale adoption of two interventions that have the potential for STRENGTHENING SEED DEVELOPMENT the greatest impact, namely, conservation agriculture (CA) AND DISTRIBUTION SYSTEMS and the development of crop varieties bred for reduced Demand for and use of improved seed varieties remains drought susceptibility or drought avoidance. CA, or con- limited in Senegal. Enhancing farmer access to improved servation farming, as practiced elsewhere in Zambia and planting materials can greatly strengthen their ability to Zimbabwe, has been shown to increase yields by (1) increas- manage production risks. GOS and donors have invested ing soil water infiltration rates, and hence reducing runoff; substantial resources over the years in variety development, (2) increasing root development so that each plant can and seed multiplication and distribution projects. As a 50 Senegal result, seven seed production and processing centers and of the individual farmer who is obliged to rely upon nine seed laboratories have been established across the national and international institutions to control this peri- country. Since 1989, the Senegalese government has grad- odic pest. Nevertheless, it is a regrettable fact that out- ually shifted to a private mode of seed multiplication and breaks of locusts continue to occur and that a significant distribution. Today, groundnut seed is produced and dis- proportion of these remain—for diverse reasons—uncon- tributed by a combination of parastatal and private entities, trolled at an early stage (Lecoq 2010). The cost of control whereas millet seed is multiplied and distributed exclusively measures increases substantially as a locust swarm devel- by individual farmers and farmers’ groups or associations. ops and it is therefore in the interest of all parties that Under the West Africa Agricultural Productivity Program initial identification and response should both be as rapid (WAAPP), approximately 12,000 tons of certified seeds as possible. Continued support to the national locust con- were produced over the past year, 2,000 tons of which were trol center in strengthening outbreak identification and distributed to farmers in August 2014 when the rains failed. response systems is essential, and programs such as the Italian Institute of Biometeorology’s (IBIMET) collabora- ISRA has developed a number of improved seed varieties tion with FAO/EMPRES to provide a meteorological covering the country’s main crops (that is, millet, ground- information service to assist in the prediction of outbreaks nuts, rice, cowpeas, sorghum), including seven new varieties will in the long term provide greater benefit than the (nec- of millet and sorghum, which are either less susceptible to essary) emergency responses that still remain typical of moisture stress or mature in a shorter period of time, reduc- many donor interventions. ing the probability that moisture stress could affect growth. Thirteen groundnut varieties have been introduced. Never- As a frontline country for locust control in Sahel West theless, their availability among smallholders remains lim- Africa, Senegal has already established autonomous ited and adoption rates across many crops remains limited. national locust control units (CNLA) responsible for all Numerous instances of poor-quality or mislabeled seed desert locus control activities. With support from neigh- were reported to the study team and it is evident that the boring countries and from the African Development seed multiplication and distribution system does not enjoy Bank, USAID, the World Bank, France, and FAO, GOS the full confidence of producers. In addition, lack of infor- has equipped CNLAs to both prevent and respond to mation and poor farmer access to certified seeds further SGR outbreaks by building needed infrastructure and constrain uptake. As a result, many smallholders use home- training technical staff. Such efforts merit additional saved seed and do not benefit from ISRA research. support. There is a clear need to improve the seed multiplication Over the past 10 years it has been recognized that the and distribution system with particular regard to the testing probability of locust swarms developing and their initial of seed to meet minimum germination and purity stand- development can both be controlled through the use of ards. The current system is not meeting the needs of farm- pheromones and mycopesticides. There is also stronger ers who noted that as a result, the purchase of seed has recognition that locust control is not simply a matter of become in itself a risk for farmers. Support for increased observation and reaction, but can be achieved through testing capacity and stronger enforcement of the regula- proactive IPM. This will require government support for tions regarding seed standards will encourage broader the registration and production of biopesticides, as well as adoption of improved varieties by smallholders and help training in their effective use. The aim of a locust IPM is reduce farmer vulnerability to erratic rainfall and drought. to manage locust populations so that uncontrolled out- breaks no longer occur. This requires continuous inter- FASTER, MORE TARGETED LOCUST vention, but may well be less costly in terms of control CONTROL measures and will certainly be both more environmentally The impact of locust swarms can be visited upon all forms friendly, and more effective in terms of reduced impacts, of agriculture, including field and horticultural crops as than the more reactive responses that characterized the well as livestock. Risk management lies beyond the scope last locust invasions in Senegal. Agricultural Sector Risk Assessment 51 UPGRADE CROP AND LIVESTOCK higher likelihood of disease outbreak but also a significantly SERVICES greater risk of higher losses when outbreaks occur. Such Both crop and livestock producers are faced with the risk low coverage levels means that diseases will not be eradi- of losses caused by pests and diseases. These are suffi- cated and, indeed, in recent years Senegal has seen a ciently frequent for them to be considered more as con- resurgence of CBPP, which once had been eradicated straints than as risks were it not for the fact that they can within Senegal’s borders. be avoided. In practice, outbreaks of livestock and crop diseases are widespread and common, due in the case of Improved service delivery could be achieved in a num- livestock to inadequate institutional disease control sys- ber of ways, including the use of PPPs and/or of para- tems (veterinary services, vaccination programs, and con- veterinarians, but in all cases it will be critical that trols over animal movement) and in the case of crops to government should provide a level of oversight that can the slow rate of dissemination of disease-resistant varie- ensure performance. Currently, this is not the case. The ties and to the high cost and limited availability of study heard that those tasked with providing services imported agrochemicals. Among both crop and livestock frequently fail to do so, but continue to receive remu- producers there is limited knowledge of disease control neration, suggesting that the strengthening of agricul- procedures as well as limited capacity to identify and react tural and livestock extension management systems could to insect pests in a timely fashion. significantly enhance performance and reduce risks to stakeholders. Although public spending on the agricultural and live- stock sectors exceeds 10 percent of the national budget, IMPROVED MARKET EFFICIENCY the current impact of agricultural extension and veteri- A key area of uncertainty is that of price, which varies nary services does not reflect this expenditure. Instead, substantially both within and between seasons. Those producers face a level of risk from pests and diseases that institutions (such as the groundnut processors and SOD- could be significantly reduced through improved agricul- EFITEX) that buy on the local market and sell interna- tural extension and veterinary services. Improper animal tionally could reduce price risk through the use of disease control, in particular, has a substantial impact on mechanisms such as hedging, futures contracts, and the production and productivity and public health and purchase of options. The greater risk, however, is to the contributes to increased pandemic risks. smallholders and livestock producers as well as traders who sell and buy on the domestic market in which no Mitigation of disease risks is largely through raising such mechanisms exist. The uncertainties of the market awareness, good public hygiene, effective quarantines and reduce the extent to which traders in particular are will- vaccination programs, promoting good agricultural prac- ing to take a position and thereby limit selling opportuni- tices, and capacity building at the farmer level. In the case ties for producers as well as contributing to increased of poultry, the intensive or industrial production system market fluctuations. Such uncertainties could be reduced also mitigates risk through good biosecurity and hygiene. through enhanced market regulation, including stronger The Senegalese government’s decision to ban poultry enforcement of existing performance requirements, imports in 2005 following the avian influenza outbreak, more rapid dispute resolution, and strong adherence to although supporting the growth of its domestic poultry competition law. industry, has remained the country’s principal mitigation strategy against the disease. There are a number of mechanisms that can enhance market efficiency, including the development of warehouse The privatization of veterinary services has taken place receipt systems and improved dissemination of produc- without fully understanding the economic viability of tion information (price information travels freely and rap- such an approach. This has led to very low levels of cover- idly through the mobile phone network). In the livestock age in the extensive livestock system, creating not only a sector, support for commercial destocking exercises 52 Senegal through the development of market linkages has proved ally had user rights to the land. Land zoning in the Sen- effective in stabilizing prices, although such exercises must egal River Valley is supposed to protect such loss of be carefully managed to avoid market disruption. Overall, land. However, recently, further use of grazing land for however, it is the general performance of the market as a cropland expansion purposes can be observed. If live- result of its inadequate regulation that is the key cause of stock owners and other stakeholders are all fully involved uncertainty. Regional, departmental, and arrondissement in the process, such changes may not necessarily consti- authorities all have a role to play in effective market regu- tute risk, but there are many examples in which only lation, and strengthening at all levels will help to enhance elites or community representatives are involved in plan- market performance and reduce price and other market ning and decision making, leaving the vast majority out risks facing stakeholders. of the process. Additionally, there is a need to strengthen existing value Knowing that farming and livestock keeping in non- chains within the livestock sector. This requires increased equilibrium Sahelian environments entail high risk, to commercialization and market orientation as well as create an enabling environment there must be policies capacity building at all levels, including support for and structures that build resilience, include early warn- improved market access and strengthening of market ing and early response, and can link and integrate long- information systems. term development approaches and actors to short-term emergency responses. IMPROVING THE ENABLING ENVIRONMENT SOCIAL SAFETY NET PROGRAMS This study noted the perception of uncertainty among Social protection programs can be an integral part of stakeholders in terms of policy and/or policy implemen- effective agricultural risk management. They often focus tation. Although the cause could not be accurately deter- on speeding up postshock recovery but can also aid in mined, the existence of the uncertainty was quite evident. supporting mitigation strategies. Such programs can Stronger stakeholder participation in the development take many forms, and can be especially transformative and implementation of market interventions and of other when resources are channeled into strengthening ex legislation and regulations affecting the agricultural sector ante resiliency. They can include Food-for-Work pro- would help to reduce this uncertainty. It is recommended grams that provide relief while facilitating the recovery that greater support be given to the development of pro- of affected communities and enhancing their future ducer and marketing associations and advocacy groups. resiliency. They can include programs that promote Although such institutions are indeed involved in the household savings to direct ex post cash payments and negotiation of groundnut prices, there is a wider role to be food aid delivery to affected communities. The World played in determining the nature, extent, and timing of Food Programme’s Food for Assets programs (2005–10) other interventions. This will help to develop a stronger supported 37,000–209,000 beneficiaries a year in 14 partnership between private sector stakeholders and gov- departments and seven regions via food security analysis ernment so that the perception of inconsistency in policy and community-level targeting. Participants received a can be effectively reduced. combination of food and other incentives, such as train- ing and seedlings, for asset construction during the lean For the livestock sector, in particular, policy initiatives in season (WFP 2014). Other social safety net initiatives the past and uncertainty over current GOS policy as can include construction of community-level food and well as perceived weakness in the capacity of govern- fodder banks that can greatly enhance access to food ment to implement regulations continue to have adverse and livestock feed in times of emergency while speeding impacts on livestock husbandry. Policies that led to up food distribution and relief efforts. GOS is currently changes in land use in the 1960s and 1970s alienated developing a social safety net system as part of a national key grazing areas from livestock owners who tradition- social protection framework, and a National Cash Agricultural Sector Risk Assessment 53 Transfer Program (PNBSF) targeting poor and vulnera- ble households. PRIORITIZATION OF RISK MANAGEMENT MEASURES Most all of the measures outlined above are complemen- DEVELOPING INSURANCE MARKETS tary in nature and have potential to contribute to improve The government of Senegal has long recognized the agricultural risk management systems in Senegal in the importance of agriculture insurance to protect rural short, medium, and long term. However, decision mak- households from disasters. In 2008, it created the Com- ers in often resource-constrained environments are com- pagnie Nationale d’Assurances Agricole du Sénégal pelled to find the quickest, cheapest, and most effective (CNAAS) as a PPP together with the country’s insurance measures among myriad policy options. Ideally, a industry. Today, CNAAS offers a wide range of products detailed, objective, and exhaustive cost-benefit analysis to farmers and herders, but will need substantial growth will help in selecting the most appropriate intervention and product evolution to have a significant impact in options. But conducting a cost-benefit analysis of so addressing smallholder vulnerability and protecting their many different options can often be a costly and time- assets. Ongoing government support such as premium consuming process. subsidies (GOS contributes 50 percent of the premium) and tax exemptions are powerful stimulants but leave room for further strengthening. The use of decision filters is an alternative approach to evalu- ate and prioritize among a lengthy list of potential interven- The experience of CNAAS in serving smallholders has tions. This can aid decision makers in making appropriate benefited from a number of private sector pilot projects resource allocation decisions more expediently and more centering on index insurance, supported by the World Food cost effectively. The following decision filters were developed Programme, World Bank Group, USAID, and other and used by the World Bank team. The study team applied donors. They provide important proof of concept for inno- these filters to facilitate a rapid assessment to obtain first vative approaches to agriculture insurance, but cannot order of approximation, based on its assessment of the situa- aspire to reach substantial numbers of farmers or herders tion in the field. Whatever the filtering process and criteria in the absence of a more coordinated and strategic adopted to evaluate decision options, it is important to ensure approach. their clarity and consistency. Public-private partnerships allow the private sector to pro- Table 6.4 describes the basic filtering criteria the assess- vide services that a government might wish to offer, but in ment team used to rate each intervention, based on a scale which it lacks the necessary expertise to undertake effi- of 1 to 5 (1—No; 2—marginally; 3—somewhat; 4—yes; ciently. Agricultural insurance is well suited to be under- 5—absolutely). taken by a PPP, which can maintain commercial objectivity while allowing for subsidization as necessary. If GOS has a long track record of investing in risk reduc- well implemented, such a program could help support, tion. In recent years, Senegal has adopted a broad array when implemented in complementarity with other risk of measures toward increasing capacity around Disaster management measures, substantial increases in produc- Risk Reduction (DRR). Such measures include the crea- tivity. The challenge is to provide coverage to smallholders tion of the Directorate of Civil Protection (DPC), the across wide areas without incurring excessive costs of development of a National Platform for DRR, and the administration. Parametric insurance programs are elaboration of a National Action Plan on DRR (2010– designed to achieve this, but require a level of geographic 15). Senegal also participates in the recently launched, resolution commensurate with local variations in soil type EU-led Global Alliance for Resilience Initiative (AGIR), a and climate if the system is to be effective. This may regional response to chronic food and nutritional insecu- become increasingly possible as costs of technology rity across the Sahel and is a member of the Comité Per- decrease, but beneficiary responses suggest that the pro- manent Inter-Etats de Lutte contre la Sécheresse dans le cess remains in development. Sahel (CILSS). These and other initiatives are primarily 54 Senegal TABLE 6.4. FILTERING CRITERIA FOR RISK MANAGEMENT SOLUTIONS Criteria Description Applicability to current agricultural policy/programming Public sector: Is the proposed solution in line with or business objectives current/existing agricultural policy/programs/priorities and so on? Private sector: Is the proposed solution in line with current/existing business objectives, and so on? Feasibility of implementation Is the proposed solution “easy” to implement in the short to medium term? Affordability of implementation Is the proposed solution affordable to put into action/implement? Scalability of implementation Is the proposed solution easy to scale up/make available to an increased number of beneficiaries? Long-term sustainability Is the proposed solution sustainable in the long term? Source: World Bank. focused on emergency rescue and response support to “animal feed” banks. Launched in July 2008, the Fonds victims of disasters, rather than actual prevention, prepar- d’Appui à la Stabulation (FONSTAB) contributes to the edness, and mitigation measures. modernization and intensification of animal production by promoting infrastructural and process upgrade invest- In 2006, Senegal finalized its National Adaptation Pro- ments. These include the acquisition of equipment for the gramme of Action (NAPA) for climate change adap- production, processing, packaging, and marketing of ani- tion. Under its NAPA, Senegal identified saltwater mal products; the provision of fodder crops; and the intrusion, coastal zone inundation, drought, storm installation of artisanal, semi-industrial, and industrial surges, and extreme temperatures as urgent climate- units for the modernization and intensification of live- related hazards that called for immediate action. In stock production. Launched in February 2014, GOS’s looking at areas of vulnerability and possible adapta- newest initiative, the National Agro-sylvo-pastoral Devel- tion options, Senegal’s NAPA focused on the water opment Fund (FNDASP), aims to promote broader dis- resources sector, the agriculture sector, and coastal semination of technological innovations through value zones. In line with these principal hazards and areas of chain approaches, producer training, institutional sup- concern, Senegal’s NAPA prioritizes adaption projects port, and funding for research programs. related to the development of agro-forestry, programs to promote the rational use of water, protection of the These and other GOS and donor initiatives are already coastline, and programs to raise awareness and educate helping to address vulnerabilities and strengthen the resil- the public on related issues. iency of the agricultural sector. And yet, as highlighted by this report, agricultural supply chains in Senegal remain Other initiatives are not new and have been adapted over highly vulnerable to a wide range of risks that jeopardize the years to meet the shifting risk landscape. For example, rural livelihoods. The current study highlights the need to help pastoralists cope through drought periods, the for a more targeted and systematic approach to agricul- GOS’s Opération Suvegarde du Bétail (OSB) protects the tural risk management in Senegal. most sensitive categories of livestock species (for example, lactating females, calves, and animal traction animals) by Based on an analysis of key agricultural risks, an evalua- distributing subsidized animal feed during emergencies to tion of levels of vulnerability among various stakeholders, vulnerable areas. In March 2012, the GOS bought over and the filtering of potential risk management measures, CFA 3.5 billion of animal feed from national industrial this assessment makes the following recommendations for mills and distributed it at a 50 percent subsidy, thereby GOS’s consideration. The proposed focus areas of inter- encouraging the establishment of community-managed vention encompass a broad range of interrelated Agricultural Sector Risk Assessment 55 investments, which together hold strong scope strengthen access, and development of community-driven agricultural risk management systems and improve agri- feed/fodder production and storage centers. cultural resilience in Senegal. 1. Strengthening extension delivery systems (for CONCLUSION example, face-to-face, farmer-driven, ICT-based) This Phase I assessment assesses agricultural risks and for improved farmer access to technology and impacts during the period 1980–2012. By documenting agronomic advice on improved soil, water, and and analyzing how Senegal’s agricultural economy has pest management practices (for example, Conser- been affected in the past by risk events, the study has gen- vation Agriculture, IPM). erated insight into which sources of risks are most likely to 2. Promoting improved water management mea- affect the sector and dependent livelihoods in the future. sures (for example, water pans, roof and rock By prioritizing risks, the study can help GOS focus atten- catchment systems, subsurface dams) and micro- tion and resources on a smaller set of key risks that are irrigation technology (for example, drip irrigation) having the most adverse impacts on production yields, via community-led initiatives (for example, cash/ incomes, and livelihoods. The study suggests a framework Food-for-Work programs). for the development of a more comprehensive, integrated 3. To further reduce rainfall dependency and better risk management strategy to strengthen existing mitiga- exploit existing water and land resources, promot- tion, transfer, and coping measures in Senegal. Finally, it ing expansion of irrigation infrastructure. provides a filtering mechanism to aid in the selection of a 4. Promoting use of contour erosion and fire barriers, set of strategic interventions for improved agricultural risk cisterns for storing rainfall and runoff water, con- management. trolled/rotational grazing, grazing banks, homestead enclosures, residue/forage conservation, and other The assessment recognizes that many of the proposed Sustainable Land Management (SLM) practices to strategies may already be covered to varying degrees under reverse degradation of water, soil, and vegetation existing risk management programs. Others may currently cover ensure sustainable access to grazing land. be in the process of implementation, either by government 5. Establishing and improving regional and national agencies or by donors. Moving forward, the Phase II Solu- NDVI and early warning systems and farmer tions Assessment will analyze the effectiveness of existing training linked to an effective and early emergency programs, identify and assess challenges impeding their response system for drought and locust outbreaks. effectiveness, and outline strategies for scaling up effective 6. To improve decision making among farmers and interventions to reach a larger number of beneficiaries. pastoralists and attenuate price volatility, strength- This follow-up activity will place strong emphasis on ening the quality and access to needed agricultural ensuring a more coordinated, integrated approach to risk information, including weather forecasting, exten- management in Senegal to ensure more effective and sion advice and innovations (that is, seeds, water meaningful risk reduction and resilience building across management), input/output prices, and so on for the sector. improved decision making. 7. Strengthening seed distribution systems, vaccina- It is hoped that the findings and conclusions of this assess- tion programs, and animal health services through ment will help to contribute to the existing knowledge improved monitoring and enforcement of existing base regarding the agricultural risk landscape in Senegal. quality control regulations governing product and It is also hoped that the study will help to inform a dia- service delivery, institutional capacity building, logue moving forward between the GOS, the World Bank, reform measures, and so on. and GOS’s other development partners that will lead to 8. Building resiliency in northern pastoralist zones concrete interventions toward improved agricultural risk via more broadly inclusive policy making based management and stronger resilience among stakeholders on land administration for improved mobility and in the years ahead. 56 Senegal REFERENCES Ahmed, M. M., J. H. Sanders, and W. T. Nell. 2000. “New Sorghum and Millet Cultivar Introduction in Sub-Saharan Africa: Impacts and Research Agenda.” ­ Agricultural Systems 64: 55–65. Ainsworth, E., and J. M. McGrath. 2010. “Direct Effects of Rising Atmospheric Carbon Dioxide and Ozone of Crop Yields.” Global Change Research, 37, 109–130. ­ DOI: 10.1007/978-90-481-2953-9_7. ANACIM (National Agency for Civil Aviation and Meteorology of Senegal), WFP (World Food ­ Programme), IRI (Columbia University’s International Research Institute for Climate and Society). 2013. “Climate Risk and Food Security in Sen- egal: Analysis of Climate Impacts on Food Security and ­Livelihoods.” Rome/ Dakar. Online: http://documents.wfp.org/stellent/groups/public/documents/ newsroom/wfp269381.pdf Bot, A. 2003. “The Importance of Soil Organic Matter.” FAO Bulletin 80. Caprio, G., and D. Klingebiel. 1996. “Bank Insolvencies: Cross Country Experience,” Working Paper 1620, World Bank, Washington, DC. CSE (Centre de Suivi Ecologique). 2013. Rapport sur les feux détectés par les capteurs MODIS. Dakar. Del Ninno, Carlos, and B. Mull. 2015. “Safety Nets in Africa: Effective Mecha- nisms to Reach the Poor and Most Vulnerable.” Washington, DC: World Bank ­ Publications. de Nicola, F., and X. Gińe. 2011. “How Accurate Is Recall Data? Evidence from Coastal India.” IFPRI and World Bank, DECRG. Diaz-Bonilla, E., M. Thomas, S. Robinson, and A. Cattaneo. 2000. “Food security and Trade Negotiations in the WTO: A Cluster Analysis of Country Groups.” TMD Discussion Paper No. 59, IFPRI, Washington, DC. Diop, A., and O. Faye. 2012. Senegal Scenario Analysis 2011–2012: Two Rural Liveli- hood Zones Assessed Using the Household Economy Approach (HEA). London: Save the ­Children UK. Eilerts, G., J. Verdin, and L. White. 2012. “A Climate Trend Analysis of Senegal.” U.S. Geological Survey Fact Sheet 2012–3123, p. 4. Elbers, C., J. W. Gunning, and B. Kinsey. 2007. “Growth and Risk: Methodology and Micro Evidence.” World Bank Economic Review 21 (1): 1–20 . Ericksen, P., et al. 2011. “Mapping Hotspots of Climate Change and Food Insecurity in the Global Tropics.” CCAFS Report. 5. CGIAR Research Program on Climate Change, Agriculture and Food Security (CCAFS). Copenhagen, Denmark. FAO (Food and Agriculture Organization of the UN). 1971. “Irrigation Practice and Water Management.” Irrigation & Drainage Paper 1. ———. 2014. “Secteur Avicole Senegal”. Revues nationales de l‘élevage de la division de la production et de la santé animales de la FAO. No. 7. Rome. FAOSTAT. 2009. “Food Balance Sheet.” Agricultural Sector Risk Assessment 57 Franzel, S. 1979. “An Interim Evaluation of Two Agricultural Production Projects in Senegal: The Economics of Rainfed and Irrigated Agriculture.” Working Paper 28, USAID African Rural Economy Program. Funk, C., J. Rowland, A. Adoum, G. Eilerts, J. Verdin, and E. White. 2012. A Climate Trend Analysis of Senegal, USGS Famine Early Warning Systems Network— Informing Climate Change Adaptation Series. Gary F., P. Facelli, and S. Hutter. 2010. “Analyse OIE des écarts du PVS au Sénégal.” Report prepared for OIE PVS Gap Analysis. GFDRR (Global Facility for Disaster Reduction and Recovery). 2012. Senegal: Disaster Risk Financing and Insurance Country Note. World Bank, November. Hathie, I., et al. 2012. “Economic Impacts of Climate Change on Farmers in Nioro du Rip, Senegal: An Integrated Assessment.” Iglesias, A., and C. Rosensweig. 2009. “Effects of Climate Change on Global Food Production from SRES Emissions and Socioeconomic Scenarios.” Palisades, NY: NASA Socioeconomic Data and Applications Center (SEDAC). http://dx.doi .org/10.7927/H4JM27JZ. International Monetary Fund. December 2010. IMF Country Report No. 10/368. Senegal: Poverty Reduction Strategy Paper Annual Progress Report. Washington, DC. IPCC (International Panel on Climate Change). 2007. “Climate Change 2007: Impacts, Adaptation and Vulnerability.” Contribution of Working Group II to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. M. L. Parry, O. F. Canziani, J. P. Palutikof, P. J. van der Linden, and C. E. Hanson, eds. Cambridge University Press, Cambridge, UK, 976 pp. Jalloh, A., G. C. Nelson, T. S. Thomas, R. Zougmoré, and H. Roy-Macauley. “West African Agriculture and Climate Change: A Comprehensive Analysis.” Washing- ton, DC: International Food Policy Research Institute. Kabamba, H., and A. Muimba-Kankolongo. 2009. “Adoption and Impact of Con- servation Farming on Crop Productivity among Smallholder Farmers in Kapiri Mposhi District of Zambia.” Journal of Animal & Plant Sciences 3 (2): 205–214. Kandji S. T., L. Verchot, and L. Mackensen. 2005. “Climate Change and Variability in the Sahel Region: Impacts and Adaptation Strategies in the Agricultural Sector.” Khouma, M., et al. (2013). Senegal. IFPRI book chapters, 291–322. Lecoq, M. 2010. “Integrated Pest Management for Locusts and Grasshoppers: Are Alternatives to Chemical Pesticides Credible?” Journal of Orthoptera Research 19 (1): 131–32. Mahul, O., and C. Stutley. 2010. Government Support to Agricultural Insurance. Washington, DC: World Bank. McSweeney, C., M. New, and G. Lizcano. 2010. “UNDP Climate Change Country Profiles: Senegal.” http://country-profiles.geog.ox.ac.uk/. Muilenburg, V. L., and D. A. Herms. 2013. “Responses of Insect Pests to Climate Change: Effects and Interactions of Temperature, CO2, and Soil Quality.” In Climate Change and Food Security in West Asia and North Africa, 115–130. Amsterdam, the Netherlands: Springer. Oceanium. 2014. “Bushfires.” http://www.oceaniumdakar.org/Les-feux-de-brousse. html?lang=en. 58 Senegal Resnick, D. 2013. “The Political Economy of Food Price Policy in Senegal.” Research Brief No. 7, March 2013. Cornell University, New York. http://dyson .cornell.edu/faculty_sites/pinstrup/pdfs/ResearchBriefs/senegal-2.pdf. Sene, I. M., et al. (2006). “Impacts of Climate Change on the Revenues and Adapta- tion of Farmers in Senegal.” CEEPA DP20, University of Pretoria, South Africa. Sultan, B., et al. 2013. “Assessing Climate Change Impacts on Sorghum and Millet Yields in the Sudanian and Sahelian Savannas of West Africa.” Environmental Research Letters 8 (1): 14–40. USAID (U.S. Agency for International Development). 2013a. “Low Cost Farming Techniques Defy Drought in Senegal.” http:// www.usaid.gov/results-data/success -stories/Low-cost-farming-techniques-defy-drought-in-senegal. ———. 2013b. “Global Food Security Response: Senegal Rice Study. October 2013. microREPORT #160. https://www.microlinks.org/library/global-food-security -response-senegal-rice-study. Wane, A., V. Ancey, and I. Toure. 2010. “Pastoralisme et recours aux marchés: cas du Sahel sénégalais (Ferlo).” Cahiers Agricultures 19: 14–20. Wane, A., and Diouf Galandou. 2012. Vulnérabilité économique et résilience au Ferlo., Déterminants contextuels en dynamique. Weeks, J. 1970. “Uncertainty, Risk, and Wealth and Income Distribution in Peasant Agriculture.” Journal of Development Studies 7 (1): 28–36. WFP (World Food Programme). 2011. “Comprehensive Food Security and Vulner- ability Analysis (CFSVA): Republic of Senegal.” New York. ———. 2014. Summary Report of the Evaluation for the Impact of Food for Assets on Livelihood Resilience in Senegal, Evaluation Reports, Rome, February 10–11. Wheeler, D. 2011. “Quantifying Vulnerability to Climate Change: Implications for Adaptations Assistance,” CGD Working Paper 240. Washington, DC: Center for Global Development. (and accompanying data set and maps). World Bank. 2013. Project Information Document (PID) Concept Stage. Public Dis- closure Copy Report No.: PIDC491. Project Name SN: Casamance Regional Development Pole (P125506). Region: Africa. Country: Senegal. Agricultural Sector Risk Assessment 59 APPENDIX A CLIMATE CHANGE IMPACTS ANALYSIS INTRODUCTION Climate change is a long-term trend that will exacerbate natural resource constraints on agricultural production in Senegal by making weather patterns more variable, and increasing the frequency and intensity of severe weather events. As result, climate change will directly affect the incidence of some agricultural risk events and indirectly affect the incidence of others. Understanding how climate change trends affect farm productivity is essential to formulating an agricultural risk management plan that maximizes the use of scarce resources. Regardless of the future extent of global warming, identifying and implementing risk management strategies that address agricultural risks, including those exacerbated by climate change, can reduce volatility and improve sustainability in the sector. VULNERABILITY TO CLIMATE CHANGE In the Mapping the Impacts of Climate Change index under “Agricultural Productivity Loss,” the Center for Global Development ranks Senegal sixth out of 233 countries globally for “direct risks” due to “physical climate impacts” and 23rd out of 233 for “overall vulner- ability” due to “physical impacts adjusted for coping ability” (Wheeler 2011). Like most of the countries in the Sahel region, Senegal’s agricultural sector is highly vulnerable to the effects of climate change. The country’s climate is already charac- terized by high temperatures and low, highly variable annual precipitation, factors that negatively affect the productivity of heat-sensitive crops. More than 95 percent of the total cropped area depends on rain-fed systems, and most farmers and herd- ers practice traditional forms of agriculture (Khouma et al. 2013, 319). Because of these factors and the importance of the sector in Senegal’s national economy, cli- mate change impacts on crop yields and land suitability will have far-reaching effects. The agricultural sector accounts for approximately 15 percent of GDP and is an important source of foreign exchange earnings. Agriculture also plays a key role in Agricultural Sector Risk Assessment 61 poverty reduction and food security through its contribu- frequency of extremely hot temperatures will increase tion to livelihoods. The sector employs 77 percent of the (McSweeney, New, and Lizcano 2010). workforce and generates 55 percent of national grain requirements (IFPRI 2012). Although Senegal is a net PRECIPITATION food importer, domestic production is an essential source In contrast to the general consensus on rising and extreme of household food consumption, especially in rural areas. temperatures, models of future rainfall conditions in Sene- In 2009, the average Food Self-Sufficiency Rate was 86 gal predict varied outcomes. There is no consensus on the percent for millet/sorghum but just 39 percent for rice magnitude of change in precipitation, and little agreement (IMF 2010), which constitutes a larger share of total cereal on the direction of change in different regions, although consumption in urban households (54 percent) compared model projections trend toward decreases in mean annual with rural households (24 percent) (USAID 2013b). rainfall. According to a multimodel United Nations Development Programme (UNDP) analysis, projected PATHWAYS OF IMPACT annual change in rainfall ranges from –38 to +21 percent Climate change affects agriculture through temperature by the 2090s, and the mean annual change predicted ranges increases, changes in precipitation, and increases in the from –18 to +7%. In sum, different models predict a wide frequency and severity of extreme weather events. There range of scenarios for large parts of Senegal. are direct impacts, such as changes in land suitability for There is broad agreement across models that precipita- crops related to temperature changes, and indirect tion extremes will be more frequent, increasing the inci- impacts, such as changes in food prices that ultimately dence of droughts and floods (IPCC 2007). Most models affect food demand and well-being. Models predicting the also predict an increase in the intensity of high-rainfall effects of climate change on agriculture vary across events, and greater variability in the onset and cessation regions and crop/livestock sectors, and depend heavily on of the rainy season (Sene et al. 2006). These factors would the underlying assumptions. The projected effects of negatively affect agricultural production, especially in changes in precipitation are particularly difficult to recon- regions where precipitation is already highly variable. cile given the vast regional variation in annual rainfall and limited district-level data. Rising temperatures are also expected to increase evapotranspiration, offsetting pro- LENGTH OF GROWING PERIOD ductivity gains. Although there is a large degree of uncer- The length of growing period (LGP) is a key determinant tainty about the magnitude of impact, this appendix of land suitability for agricultural production, and is defined synthesizes existing climate projections and crop forecasts, by the average number of days per year when average air highlights areas of consensus between different studies, temperature and evapotranspiration rates are conducive to and identifies areas of disagreement. crop growth. In central Senegal, a large area is expected to flip from an LGP greater than 120 days in the 2000s to an LGP less than 120 days by 2050 (Erickssen et al. 2011). The PRINCIPAL FINDINGS 120-day threshold is significant because cultivating crops like maize is considered very difficult below this threshold. TEMPERATURE Temperature changes over the past 40 years have been faster in Senegal than temperature changes globally. Since AREAS OF UNCERTAINTY 1975, temperatures have increased by approximately » Extent of crop yield increases due to CO2 fertilization 0.9°Celsius (Funk et al. 2012), and downscaled general cir- » Extent of increase in pest and disease incidence due culation models (GCMs) predict future increases in the to CO2 fertilization (Muilenburg and Herms 2013) average daily maximum temperature during the warmest » Impact on total crop production and postharvest month of at least 1°C to 1.5°C ( Jalloh et al. 2012). The rate losses caused by the evolution of pests and diseases of warming is higher in the interior regions than in the » Impact of ozone damage on crop yields (Ainsworth coastal region, and all projections indicate that the and McGrath 2010; Iglesias et al. 2009) 62 Senegal The study simulated the effects of climate change METHODOLOGIES with and without adaptation on crop yields using Data analyses from the literature reviewed in this appen- multiyear baseline (1980–2009) and future (2040– dix draw from downscaled GCMs. The studies use multi- 69) climate projections. The analysis was limited to ple GCMs, simulate between one and five greenhouse gas Nioro du Rip, a district in southeastern Senegal. emissions scenarios, and incorporate crop prediction » A multicountry study published in Environmental models. As a result, the conclusions vary depending on the Research Letters (Sultan et al. 2013) projected future underlying model assumptions. crop yields in Senegal based on several climate mod- » A country-level International Food Policy Research els from CMIP3 (the World Climate Research Pro- Institute (IFPRI) study (Khouma et al. 2013, 291– gramme’s Coupled Model Intercomparison Project), 322) compares yield projections for 2050 across three IPCC emissions scenarios (A2, A1B, and B1), four different GCMs (CNRM-CM3, ECHAM 5, and the crop model SARRA-H v.32. The analysis CSIRO, and MIROC), using 2000 climate data incorporated multiple climate datasets for the 1961– as the baseline. Each model is simulated with the 90 baseline period and simulated the effects of cli- IPCC A1B emissions scenario, which assumes fast mate change on crop yields through 2100. economic growth, a midcentury global population peak, new and efficient technologies, and a mix of fossil and nonfossil energy sources. CROP PROJECTIONS » AgMIP (Hathie et al. 2012) projected future crop yields using two crop simulation models calibrated MAIZE with household survey data and future climate Maize is grown in most parts of the country, except in the data under five GCMs (CCSM4, GFDL-ESM2M, Sahel and other very dry regions (figure A.1). The most Had GEM2-ES, MIROC5, and MPI-ESM-MR). important areas of production areas are located in the FIGURE A.1. RAIN-FED MAIZE YIELD CHANGE UNDER FOUR CLIMATE MODELS, 2010–50 (IPCC A1B SCENARIO) CNRM-CM3 GCM CSIRO Mark 3 GCM 2000 old area lost Yield loss > 25% of 2000 Yield loss 5–25% Yield change within 5% Yield gain 5–25% Yield gain > 25% ECHAM 5 GCM MIR MIROC medium-resolution OC 3.2 medium resolution GC GCMM 2050 new area gained Source: Khouma et al. 2013, 311. Agricultural Sector Risk Assessment 63 southern Groundnut Basin and the Casamance region. experience declines in crop yields, net farm revenue, and Maize is also grown under irrigation in the Senegal River per capita income. The simulated models estimated Basin in the north, with higher yields (about 2 tons per hect- greater economic losses for maize-based farms compared are) than rain-fed maize (about 1 metric ton per hectare) with nonmaize farms. (Khouma et al. 2013). After rice and wheat, maize is the most-consumed grain in the country, accounting for approx- GROUNDNUTS imately 10 percent of the food supply (FAOSTAT 2009). Groundnuts are the most widely grown crop in Senegal, ranking first in terms of area harvested and second in According to the IFPRI study, all models predict that terms of production volume (figure A.2). Groundnut oil is some areas of Senegal will experience maize yield gains the country’s top agricultural export and an important of between 5 and 25 percent. The models display consid- component of the domestic food supply. Typically, erable agreement on the location of the yield increases, but groundnuts are intercropped with millet. disagree on the magnitude. All models also predict that some areas will experience yield losses in places where All of the models in the IFPRI study predict a general maize currently grows. The CNRM and ECHAM models 5 to 25 percent decline in groundnut yields. Model pre- predict a larger decline in maize yields than do the CSIRO dictions diverge at the northern edge of the Groundnut and MIROC models. The ECHAM model also predicts a Basin: Two of the four models predict yield losses of greater loss in total harvested area than do the CSIRO more than 25 percent in this area. Two of the models and MIROC models. The IFPRI study concludes that predict small areas of yield gains closer to the coast and maize will be less negatively affected by the impacts of in the southeast. climate change compared with millet and groundnuts. All of the model simulations in the AgMIP analysis pre- The AgMIP study, which was limited to a single district in dict lower groundnut yields in Nioro district in the 2040– southeastern Senegal, concluded that farmers would 69 period compared with the baseline period. FIGURE A.2. RAIN-FED GROUNDNUTS YIELD CHANGE UNDER FOUR CLIMATE MODELS, 2010–50 (IPCC A1B SCENARIO) CNRM-CM3 GCM CSIRO Mark 3 GCM 2000 old area lost Yield loss > 25% of 2000 Yield loss 5–25% Yield change within 5% Yield gain 5–25% Yield gain > 25% ECHAM 5 GCM MIROC 3.2 medium-resolution GCM 2050 new area gained Source: Khouma et al. 2013, 310. 64 Senegal FIGURE A.3. RAIN-FED RICE YIELD CHANGE UNDER FOUR CLIMATE MODELS, 2010–50 (IPCC A1B SCENARIO) CNRM-CM3 GCM k 3 GCM CSIRO Mark 2000 old area lost Yield loss > 25% of 2000 Yield loss 5–25% Yield change within 5% Yield gain 5–25% Yield gain > 25% 2050 new area gained ECHAM 5 GCM MIROC MIR 3 2 medium-resolution GC OC 3.2 M GCM Source: Khouma et al. 2013, 313. RICE MILLET Rice is the most widely consumed food in Senegal, After groundnuts, millet is the most widely grown crop in accounting for approximately 29 percent of the food Senegal, accounting for approximately 7 percent of the supply (FAOSTAT 2009). Between 2001 and 2005, 80 food supply (FAOSTAT 2009). Millet’s contribution to food percent of domestic rice consumption depended on security is especially important in the Sahel region, where it imports. Since 2005, domestic rice production has the most widely consumed and widely cultivated crop. more than doubled, but the country still relies heavily on imports.9 Rice accounts for approximately 4 per- A study published in Environmental Research Letters concluded cent of the total area harvested in Senegal, behind mil- that millet yields in Senegal would be negatively affected by let, groundnuts, sorghum, and maize. Rain-fed rice is climate change (Sultan et al. 2013). Out of 35 scenarios, 31 grown in most areas in the southern half of the coun- showed a negative impact on millet yields, with yield up to try, and irrigated rice is grown in the Senegal River 41 percent. Yield reductions were predicted to be greater in Basin. the Sudanian region (southern Senegal), compared with the Sahelian region. According to the multicountry study, According to the IFPRI study, predicted yield increases traditional cultivars were more resilient than modern high- for rain-fed rice are similar to those for maize, but yielding varieties. However, owing to the large difference in relatively greater in magnitude, with larger areas of mean yields, modern varieties would still outperform tradi- yield gains between 5 and 25 percent (figure A.3). tional ones (under optimal fertility conditions), even if they Overall, rice yields are expected to be less negatively are more affected by climate change. affected than groundnuts. The AgMIP study in Nioro district also concluded that millet yields would be negatively affected by climate change. All model simulations predicted a decrease in 9 OECD, 2008. future yields compared with the baseline period. Agricultural Sector Risk Assessment 65 changing incidence of pests and diseases will affect farm CONCLUSION productivity. Despite these uncertainties, most studies Uncertainties surrounding the degree of temperature agree that Senegal’s agricultural sector is highly vulnerable increases and the direction of change in precipitation lev- to the effects of climate change, and that the overall impacts els make it difficult to determine the precise impact of of climate change will be detrimental to national food future climate change on crop yields. Predictions of crop security. Absent climate adaptation measures, the effects of yields vary depending on the underlying assumptions, and climate change can be expected to exacerbate the impact most models do not account for the way in which the of risk events on farm productivity and food security. 66 Senegal APPENDIX B SENEGAL VULNERABILITY ANALYSIS INTRODUCTION Agricultural shocks are one important factor driving chronic poverty and food insecu- rity in Senegal. Shocks affect household well-being in a variety of ways, by limiting food availability, weakening food access, and negatively affecting monetary well-being through the depletion of productive assets. Chronically vulnerable groups with high exposure to hazards experience a disproportionate impact from adverse events and lack coping mechanisms available to other groups. In this context, vulnerability is a useful lens through which to examine agricultural shocks because it allows policy mak- ers to determine which groups are most affected and to target risk management solu- tions accordingly. GENERAL TRENDS Rural households are more likely to be poor and food insecure than are urban house- holds in Senegal. In rural districts, 15.1 percent of households are severely or moder- ately food insecure, compared with 8.6 percent in urban districts. Malnutrition, a direct cause of food insecurity, is significantly higher in rural areas: almost 20 percent of rural children under age five are stunted, compared with 9 percent of children in urban zones. Important sources of vulnerability in the rural environment include (1) poor access to markets, (2) low levels of educational attainment and poor quality edu- cation, and (3) lack of diversity in income opportunities. All three factors increase the vulnerability of households to food insecurity and affect rural communities more acutely than urban ones. Agricultural Sector Risk Assessment 67 TABLE B.1. PREVALENCE OF FOOD Location is a key determinant of poverty and food secu- INSECURITY BY DISTRICT rity status for Senegalese households. In fact, national averages of food security and poverty mask stark regional Severe Moderate differences. The prevalence of severe food insecurity Food Food ranges from less than 1 percent in Dakar and Diourbel to Insecurity Insecurity District (%) (%) 35 percent in Oussouye (see table B.1). Dakar <1 4.8 Food price shocks are by far the most important source of Diourbel <1 1.9 vulnerability for both rural and urban households. More Mattam 1.5 11.1 than 34 percent of rural households and 29 percent of Fatick 1.7 3.5 urban households reported experiencing a spike in food Bounkiling 1.9 11.1 prices between April 2009 and April 2010. Households Goudiry 2.2 14.4 experience food price volatility more frequently than any Saint-Louis 2.3 10.7 other risk event. Tambacounda 2.5 15.9 Louga 2.6 7.9 AGRO-CLIMATIC Koumpentoum 2.6 17.8 CONDITIONS, LIVELIHOODS, Bakel 3 9 AND VULNERABILITY Kaolack 3.2 23.9 In rural areas, the prevailing agro-climatic conditions, par- Thiès 3.2 12.6 ticularly rainfall and water resources, are strong predictors Goudomp 4.3 18.8 of livelihood activities, which in turn influence household Kanel 4.6 14.1 food consumption and vulnerability (figure B.1). In areas Kaffrine 4.8 12.1 rich in water resources, producers are more likely to grow cash crops, employ farm laborers, sell market garden Salémata 5.2 21.9 products, and engage in fisheries. In the Sahel region Kédougou 5.8 22.7 where rainfall is highly variable, communities rely heavily Vélingara 5.9 16.5 on pastoralist activities and drought-tolerant food crop Médina Y.F 6.4 18.7 production. Because three-quarters of rural Senegalese Saraya 7.9 28.4 households do not produce enough food to meet their Ranérou 9 8.8 minimum food requirements, diversification of revenue Kolda 9.2 23.4 sources and access to markets are important determinants Sédhiou 9.6 10.9 of both poverty and food security. Households that lack multiple sources of income and market opportunities are Bignona 19.7 22.7 more vulnerable to a range of production and market risk Ziguinchor 20 30.4 events. These types of households exist in all livelihood Oussouye 35 23.7 zones, but are most prevalent in the agro-pastoral and Rural Districts 3.7 11.4 sylvo-pastoral zones. Several underlying factors increase Guédiawaye 0.6 5.5 the vulnerability of pastoral and agro-pastoralist commu- Kédougou 1.9 5.8 nities, including land fragmentation, population growth, Tambacounda 2.3 5.3 low literacy and education provision, and poor infrastruc- Oussouye/Bignona 12.4 21.3 ture. These chronic weaknesses undermine the capacity of Urban Districts 2 6.6 communities to respond to shocks. In turn, the increasing frequency and simultaneous occurrence of multiple shocks Source: WFP 2011. 68 Senegal FIGURE B.1. SENEGAL LIVELIHOOD ZONE MAP Source: FEWSNET http://www.fews.net/west-africa/senegal/remote-monitoring-report/may-2014. erode the effectiveness of traditional coping mechanisms, creating a vicious cycle of crisis and underdevelopment. FOOD PRICE VOLATILITY AND VULNERABILITY LIVESTOCK OWNERSHIP AND As mentioned previously, the vast majority of both urban and rural households purchase at least some of their food FOOD SECURITY on the market, especially during the annual lean season Although pastoralists and agro-pastoralist zones have when prices peak. Dependence on the market is com- above-average rates of food insecurity, livestock owner- pounded by Senegal’s reliance on food imports and expo- ship is actually positively correlated with food security. sure to international price fluctuations, which directly In rural areas of Senegal, food-secure households have affect household purchasing power. As a result, Senega- an average of 8.4 tropical livestock units (TLU), lese households are extremely vulnerable to food price compared with 5.1 TLU for food-insecure households. shocks. Figure B.2 depicts the geographic distribution of The ability to sell productive assets during the annual food price shocks, measured by the percentage of house- lean season to obtain food and basic necessities is an holds that reported experiencing one or more food price important coping strategy for agricultural households. increases in the previous 12 months. Households in the Households with large herd sizes are better equipped to northern Sahel region were more likely to experience employ this strategy, without compromising their future food price shocks than any other part of the country income, than households with only a few animals or (except the Kaolack region, where households were simi- stockless households. larly affected). Agricultural Sector Risk Assessment 69 FIGURE B.2. FOOD PRICE SHOCKS BY DISTRICT Source: WFP 2011. TABLE B.2. VULNERABLE GROUPS Pastoralists and agro-pastoralists • 37% of households that depend on livestock belong at the bottom income quintile. More than 60% are considered poor or very poor. • Illiteracy rate is 59% among head-of-households (HH). Households with an illiterate HH are more likely to be food insecure than are those with similar characteristics and a literate HH. • Low savings rate (between 11% and 13%). Unskilled wage laborers • 29% of households that depend on wage labor belong to the bottom income quintile. More than 60% are considered poor or very poor. • Low savings rate (between 11% and 13%). • Second-lowest monthly expenditure (US$1.13 per person per day). Food crop farmers • 28% of these households belong to the bottom income quintile. More than 50% are considered poor or very poor. Cash crop farmers • Similar to food crop farmers, 29% of these households belong to the bottom income quintile. More than 50% are considered poor or very poor. Households that rely primarily • More than 50% of these households belong to the bottom income quintile. More than on forest resources 75% are considered poor or very poor. • Low savings rate (between 11% and 13%). • Lower monthly expenditure than any other livelihood group (US$0.94/person/day). • 25% of households do not eat three meals a day. Source: WFP 2011. 70 Senegal APPENDIX C AGRICULTURAL INSURANCE IN SENEGAL Agriculture is important for Senegal. It contributes 20 percent to GDP and provides employment to 60 percent of its workforce. It is key to food security and to protect the country against the increasing fluctuations of international food prices, which saw rice prices triple in 2007–08. The increase of agriculture productivity is a paramount objective of the government of Senegal. The Programme de Relance et d’Accélération de la Cadence de l’Agriculture au Sénégal aims at implementing the objectives of the Plan Senegal Emergent in the agriculture sector and focuses on four strategic crops: rice, onions, groundnuts, and off-season fruits and vegetables. The underlying goals that explain the selection of these crops are: (1) coverage of the whole of Senegal with these products; (2) gradual suppression of food dependence (rice and onions); (3) development of exports (off-season fruits and vegetables); and (4) jobs creation and additional income generation. Agriculture is risky in Senegal. Ninety-eight percent of cultivated land is rain fed, and as Senegal belongs to the Sahel, rain is scarce in parts of the country and unreliable across the country, causing damage by excess as well as scarcity. The World Bank’s Agriculture Risk Management Team estimates that during 1980–2012, total losses from production risks affecting maize, rice, millet, sorghum, peanuts, and cowpeas alone totaled US$1.38 billion, that is, US$41.7 million per year. A detailed report was shared with the government of Senegal in the second half of 2014; its findings have informed the work of the AIDP and this report. World Bank analysis using the Modèle d’Analyse des Risques de Cultures du Sénégal modeled an average annual crop loss cost of 10 percent of national crop value and indicates a 1 in 100 year loss of 44 percent of the national average crop value. The 2002–03 drought cost an estimated 35 percent of national crop production, almost US$50 million. Annual groundnut production decreased by 70 percent, likewise the average cash flow to groundnut farmers. The 2004–05 locust infestation reduced mil- let yields by 23 percent and sorghum yields by 14 percent. The livestock sector is also Agricultural Sector Risk Assessment 71 affected: the 2011 drought required government to spend ments for risk transfer; it has contributed to prosperity of CFA 4 billion on 21,000 tons of forage to control livestock developed countries around the world, and is increas- mortality. ingly reducing vulnerability and protecting productive investment also in developing countries. The formaliza- Climate change is likely to aggravate rural households’ tion of organized solidarity for the redistribution of exposure to risk. Mean annual temperature has increased money between individuals and in time under the con- by almost 1°C between 1960 and 2006 and may increase cept of insurance has been refined for centuries. The by up to 3°C by midcentury. Variability of rainfall pat- mathematical law of large numbers explains how the terns has also increased, causing both droughts and floods. outcome of a collection of comparable units (such as The model of the World Food Programme and the Agence agriculture hectares or livestock units) exposed to risk is Nationale de l’Aviation Civile et de la Météorologie esti- more predictable than the outcome of each one of the mates that climate variability explains half of agriculture units, thus moving from gambling to the statistical man- yield variations in Senegal (ANACIM 2012). Repeated agement of risk that allows commercial companies to droughts, along with population growth and agriculture sustainably offer insurance in exchange for a commensu- expansion, have already led to severe environmental deg- rate premium. In this way, insurance transforms future radation in wide parts of the country. uncertainty into predictability, for government budgets as well as for households. Agricultural risk limits access to finance, and that restrains agriculture productivity. To increase productivity, crop Although insurance is a mature and well-developed instru- and livestock producers have to invest, and in most cases ment, agriculture insurance is one of the later lines to they depend on credit to do that. The supply of credit emerge, and it has not been available to small-scale farm- usually does not meet their needs, and although logistical ers until very recently. The principal obstacle that has challenges are one reason for this mismatch, risk is another. excluded small-scale farmers from agriculture insurance is In Senegal, the 1988–91 banking crisis shows how drought the cost of claims assessment, which traditionally requires precipitated the closure of seven banks (Caprio and farm visits of highly specialized experts but is not viable Klingebiel 1996). for small farmers. Fortunately, this obstacle is being over- come with the spread of index-based agriculture insur- One common risk management approach of farmers in ance during the past 15 years. Senegal as elsewhere is diversification of income sources. Forty-one percent of rural households have Agriculture insurance serves two main purposes. On the two main sources of income and 44 percent depend on one hand, it reduces vulnerability, in that it compensates three or more sources. In contrast, 43 percent of urban producers for the economic losses suffered from insured households depend on one livelihood activity. The agri- events, thus preventing them from falling into poverty culture production of rural households is itself charac- and/or using suboptimal strategies to cope with the losses, terized by diversification, as different crops are planted such as reducing food consumption, selling productive on the often-fragmented small plots. This is a very sen- assets, or taking children out of school. On the other hand, sible approach to addressing risk in the absence of it increases productivity through increased investment by other instruments to manage risk. But it limits their securing credit in case of loan default due to insurable productive capacity as they forgo higher profits in risk- events; this encourages lenders to offer credit to the other- ier activities. wise risky client group of farmers and herders. Subsistence farming may work without access to credit, but the market Insurance has been recognized as one important instru- oriented productivity growth that Senegal wants from its ment to address risk by pooling and transferring it. Risk agriculture sector will require considerable credit. This is transfer complements risk mitigation and risk coping as unlikely to become available without insurance solutions the third fundamental pillar to address agriculture risk. that remove the agriculture risks that lenders are not pre- Insurance is one of the oldest and best-developed instru- pared and equipped to bear themselves. 72 Senegal Although insurance is a powerful ex ante instrument to learned on products and processes, Senegal’s private sec- address risk before it materializes in the form of adverse tor insurance companies were invited to jointly own the events, it coexists with ex post mechanisms, typically gov- majority of the CNAAS. This has resulted in a well-run ernment or donor handouts after the event. Setting up and equipped national champion of agriculture insurance disaster funds with more or less elaborated rules of fund- whose development has benefited from ongoing govern- ing and disbursing to target populations in case of calam- ment support. ities is more straightforward than developing insurance solutions in countries where the corresponding insurance International research shows that agriculture insur- markets are not well developed. Such funds give govern- ance programs conducted by either the public sector or ments control and flexibility, may attract donor funding, the private sector alone struggle considerably to suc- and can serve other political objectives. If not based on ceed; moreover, public-private partnerships tend to be clear and generally known rules, however, they keep the a necessary condition for successful scale of agriculture target beneficiaries in uncertainty about what to expect, insurance. On a global basis, only 7 percent of agricul- limiting investment on the one hand and individual initi- tural insurance transaction volume is purely private. ative to mitigate risk on the other. And if not based on Market and regulatory impediments are often invoked solid governance, such funds may not be available to justify public intervention; they include the systemic as needed. nature of agriculture risk, information asymmetries, and the low motivation of any private sector company Agriculture risk is a continuum, and different instruments to invest in the public good of the Agricultural Risk are best suited to address different severities. Just as instru- Market Infrastructure necessary for functioning agri- ments for risk mitigation, risk transfer, and risk coping culture insurance. Private insurance providers oper- each have their role to address different degrees of sever- ated in 54 percent of the 65 countries surveyed by the ity, different instruments for risk transfer such as contin- World Bank in 2008, and public-private partnerships gent budgets or funds, contingent credits, and insurance were implemented in 37 percent of them (Mahul and solutions can be combined for best results. Stutley 2010). The government of Senegal has implemented a number CNAAS has experienced good growth and offers a wide of measures aimed at reducing vulnerability of agricul- range of products for farmers and herders, but will need ture and livestock producers and at increasing their substantial additional growth and product evolution to access to finance. The Fonds de Securisation, Fonds serve significant proportions of rural households, reduce d’Appui à la Stabulation, and Fonds de Garantie des their vulnerability, and secure their investments in Investissements Prioritaires all aim to promote lending increased productivity. In March 2014, the CNAAS esti- to farmers and herders and include guarantees that mated that it had insured approximately 5,000 producers operate like insurance by compensating lenders for loan with about 8,000 hectares in 2013, and 1,500 livestock defaults. The Opération de Sauvegarde du Bétail and producers with about 200,000 animals (three-quarters of the African Risk Capacity are insurance or insurance- them chickens). This is a substantial growth since the like mechanisms that reduce the vulnerability of Sene- company’s first year of operation, and healthy growth is gal’s herders. projected going forward. But the outreach to wider rural populations is still limited, and considerably more scale Acknowledging the importance of insurance, the govern- will be required for the CNAAS to noticeably reduce ment of Senegal created the Compagnie Nationale Senegalese vulnerability to shocks and secure the required d’Assurances Agricole du Sénégal (CNAAS) in 2008 large amounts of credit. With the current level of devel- together with the country’s insurance industry. Before opment of agricultural insurance, Senegal’s small and then, agriculture insurance was limited to farm equip- marginal crop and livestock producers are still too reliant ment and had little outreach. To crowd in the private on ex post disaster relief interventions by the government sector and let the insurance industry profit from lessons and donor partners. Agricultural Sector Risk Assessment 73 TABLE C.1. INSURANCE PREMIUM COMPONENTS Premium Component Function Potential Reduction Actuarial premium Quantifies the statistically expected payments Focus on infrequent severe events Loading for deviations Honor payment liabilities in years when claims Larger pool size to reduce statistical variance exceed premiums Loading for reinsurance Sharing risk with larger pools to protect insurer Larger pool size to reduce statistical variance against statistical outliers and the need for reinsurance, reduce Reduce capital requirements statistical variance of cessions to reinsurers, Benefit from specialized expertise of reinsurers and offer more attractive business proposition to reinsurers Government financing of reinsurance layers Loading for amortization of Recover initial investment in product and Donor funding of start-up expenses up-front cost process development, IT, and so on Loading for data cost Fund payment to providers of recurrent data Public good data access needs Other administration expense Fund ongoing variable operation cost Outsource administrative tasks to lower cost loading providers Distribution expense loading Fund ongoing distribution cost Partner with lower cost distribution channels, such as government rural outreach infrastructure Profit margin Justify the effort and build reserves Increase business volume and outlook for future market growth Agriculture insurance is expensive. Depending on circum- promising; they may come from the private sector—such stances, it can reach 10 percent or more of the expected as rural financial institutions—as well as from the public payout amounts, compared with less that 1 percent for life sector—such as Regional Office for Rural Development insurance, for example. This is not necessarily an obstacle for (DRDR), Departmental Office for Rural Development prosperous producers (which are few) but usually is a chal- (SDDR), Service Regional de l'élevage (SREL) Regional lenge for low-income farmers and herders (which constitute Office for Livestock Service, Service Départemental de the majority), and contributes to explain the slow growth of l'élevage (SDEL) Departmental Office for Livestock Ser- agriculture insurance in most developing countries. vices, and Agence Nationale du Conseil Agricole et Rural. The level of the insurance premium—and hence the Ongoing government support such as premium subsidies affordability of insurance—is driven by various factors, and tax exemptions are powerful stimulants but leave and each one of them can be addressed to make premium room for further strengthening of agriculture insurance in more affordable. Senegal. Examples of possible additional support include the following: The main components of any insurance premium are 1. Reduction of administration and distribution ex- contained in table C.1. penses by access to government rural infrastruc- The various components of an insurance premium corre- ture and staff for insurance-related tasks such as spond to various roles and interventions that the public and veterinary services or awareness creation the private sector can assume. Distribution and servicing 2. Reduction of statistical variance, reinsurance cost, of insurance to rural populations, for example, is costly, and per-policy administration and distribution cost and initial volumes of agriculture and livestock insurance through growth of the insurance pool by manda- may not look sufficient to justify the up-front investment in tory inclusion of insurance in rural development developing the necessary structures. Partnering with organ- projects related to agriculture, livestock, and izations that already have that infrastructure in place is fishing 74 Senegal 3. Reduction of statistical variance, reinsurance withdrawn too rapidly—and thus reverse the ben- cost, and per-policy administration and distribu- efits intended by the subsidy. tion cost through growth of the insurance pool by mandatory inclusion of insurance in government- In addition to the government-led interventions aiming to supported agriculture lending reduce farmer and herder vulnerability and increase 4. Reduction of up-front cost and administration access to finance, Senegal is characterized by a dynamic expenses through free access to meteorological landscape of donor-driven agriculture insurance initia- and crop yield data necessary to design, price, and tives. CNAAS is testing novel approaches to agriculture service index insurance insurance in pilots with partners including the Global 5. Reduction of administration expenses through Index Insurance Facility, the World Food Programme, and the establishment of and free access to a livestock USAID, and another program funded by the West Afri- registry can Development Bank is expected to start soon. A com- 6. Reduction of administration and distribution prehensive government strategy on agriculture insurance expenses if cross-selling of other lines of insur- should take into consideration the lessons learned and ance to rural populations is allowed in addition to expected outcomes of these projects, and be informed of agriculture insurance the gaps not yet covered by any of the existing govern- ment and nongovernment insurance and noninsurance Governments often provide support for the financing of interventions, to avoid duplications. risk through direct premium subsidies, but there are risks with this approach. Governments hope that subsidizing Any agriculture insurance strategy should also take into premiums will incentivize insurers to enter the market and consideration existing social protection mechanisms will increase the take-up of insurance products and there- and objectives to explore synergies. Subsidized agricul- fore outreach. However, experience documented from ture insurance premiums amount to self-targeted trans- elsewhere points to potential risks with this approach, fer payments. They are aimed at reducing vulnerabilities including the following: of specific target groups, but only farmers and herders 1. Sustainability: Direct premium subsidy can who buy insurance benefit from them; hence these become quite costly and is subject to changing transfers often benefit the rich more than they do the political priorities. poor. Links between insurance and social protection 2. Market distortion: Premium subsidy can lead to mechanisms are varied and can be complex. Increas- market distortion. ingly, the concept of payments triggered by readily 3. Poor incentives: Premium subsidy can motivate available indexes that correlate with hardship is trans- undesired behavior by insurers and reinsurers (lead- ferred from agriculture and disaster index insurance to ing to overpricing) and insured policyholders (by social safety nets, where benefits are increased or pro- crowding out alternative risk mitigation strategies). vided to larger populations based on suitable indexes 4. Poor targeting: It is difficult to target subsidies to that allow such dynamic response to be very fast and those who need it. well targeted. If based on insurance quality indexes, 5. Effects of withdrawal: Withdrawal of subsidy may such increases in government payments can be trans- to lead to an increase in price, which can severely ferred to international reinsurance markets, reducing affect future take-up—particularly if the subsidy is budget uncertainty. Agricultural Sector Risk Assessment 75 APPENDIX D CROP YIELD LOSS ANALYSES FIGURE D.1. GROUNDNUTS, 1.2 1980–2012 Yield (MT/Ha) Trend 0.3 trend Linear (Yield (MT/Ha)) Loss Loss Years (in MT) (US$, millions) 0.9 1980 333,101 –117.2 1983 –339,662 –$119.5 Yield (MT/Ha) 1984 –197,985 –$69.6 0.6 1992 –204,762 –$72.0 1996 –113,385 –39.9 1997 –107,586 –37.8 2002 –420,376 –147.8 0.3 2006 –41,281 –14.5 2007 –182,587 –64.2 2011 –211,477 –74.4 0 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 Total: –2,152,202 –757.0 Source: DAPS. 0.60 FIGURE D.2. COTTON, Yield (MT lint/Ha) 0.3 trend Linear (Yield (MT lint/Ha)) 1980–2012 Loss Loss Years (in MT) (US$, millions) 0.45 1980 –2,443 –0.400 Yield (MT lint/Ha) 1982 –1,988 –0.492 1996 –2,516 –0.444 0.30 1997 –3,250 –0.563 1998 –17,301 –2.45 1999 –1,766 –0.204 2002 –1,715 –0.179 0.15 2009 –1,913 –0.272 2011 –1,350 –0.441 2012 –1,347 –0.261 0.00 Total: –35,589 –5.7 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 Source: FAOSTAT. Agricultural Sector Risk Assessment 77 FIGURE D.3. MAIZE, 1980–2012 3 Loss Loss Yield (MT/Ha) Trend 0.3 trend Linear (Yield (MT/Ha)) Years (in MT) (US$, millions) 1980 –16,677 –5.3 1994 –25,517 –8.0 2.25 1995 –18,238 –5.7 1996 –21,469 –6.7 Yield (MT/Ha) 1997 –21,723 –6.8 1.5 1998 –27,694 –8.7 1999 –29,902 –9.4 2000 –19,385 –6.1 2001 –15,899 –5.0 0.75 2002 –74,229 –23.4 2007 –63,302 –19.9 2011 –54,419 –17.1 Total: –388,454 –122.3 0 1980 1983 1986 1989 1992 1995 1998 2001 2004 2007 2010 5.00 Yield (MT/Ha) Trend 0.3 trend Linear (Yield (MT/Ha)) FIGURE D. 4. RICE, 1980–2012 Loss Loss Years (in MT) (US$, millions) 3.75 1980 –34,165 –10.8 1996 –31,944 –10.1 Yield (MT/Ha) 2000 –29,967 –9.5 2001 –33,236 –10.5 2.50 2002 –41,616 –13.2 2004 –37,596 –11.9 2006 –69,667 –22.1 1.25 2007 –57,233 –18.2 Total: –335,423 –106.6 Source: FAOSTAT. 0.00 1980 1985 1990 1995 2000 2005 2010 FIGURE D.5. COWPEA, 1.38 Yield (MT/Ha) Trend 0.3 trend 1980–2012 Loss Loss 1.10 Years (in MT) (US$, millions) 1982 –5,792 –3.8 Yield (MT/Ha) 1983 –1,414 –0.93 0.83 1984 –3,121 –2.0 1990 –3,657 –2.4 0.55 1992 –8,761 –5.7 1996 –9,601 –6.3 2002 –31,143 –20.5 0.28 2003 –13,282 –8.7 2004 –49,783 –32.7 2006 –9,949 –6.5 0.00 1980 1985 1990 1995 2000 2005 2010 Total: –136,503 –89.7 Source: DAPS. 78 Senegal FIGURE D.6. POTATO, 1980–2012 40 Yield (MT/Ha) Linear (0.3 trend) 0.3 trend Linear (Yield (MT/Ha)) Loss Loss Years (in MT) (US$, millions) 1992 –6,529 –0.947 30 2000 –4,662 –0.676 2003 –2,523 –0.366 Yield (MT/Ha) 2004 –1,614 –0.234 20 2005 –1,859 –0.270 2008 –1,942 –0.282 Total: 19,129 –2,774 Source: DAPS. 10 0 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 60.00 Yield (MT/Ha) 0.3 trend Trend Linear (0.3 trend) FIGURE D.7. TOMATO, 1980–2012 Loss Loss Years (in MT) (US$, millions) 50.00 1994 –12,052 –2.4 1996 –4,984 –1.0 40.00 Yield (MT/Ha) 1997 –14,103 –2.8 1998 –17,432 –3.5 30.00 1999 –17,737 –3.6 2000 –41,486 –8.4 2001 –32,456 –6.5 20.00 2004 –27,109 –5.5 2008 –11,304 –2.3 10.00 Total: –178,665 –36.0 Source: DAPS. 0.00 1980 1985 1990 1995 2000 2005 2010 40 Yield (MT/Ha) Trend Linear (Yield (MT/Ha)) FIGURE D.8. ONION, 1980–2012 Loss Loss Years (in MT) (US$, millions) 1991 –5,130 –2.8 30 1995 –4,039 –2.2 1999 –13,475 –7.4 Yield (MT/Ha) 2000 –25,111 –13.8 2001 –13,575 –7.5 20 2002 –30,095 –16.5 2003 –11,917 –6.5 2010 –28,524 –15.7 Total: –131,867 –72.5 10 Source: DAPS. 0 1980 1983 1986 1989 1992 1995 1998 2001 2004 2007 2010 Agricultural Sector Risk Assessment 79 FIGURE D.9. GREEN BEAN, 15 1989–2012 Loss Loss Years (in MT) (US$, millions) 11 1994 −1,387 −0.71 Yield (MT/Ha) 2000 −1,592 −0.82 2001 −939 −0.48 8 2002 −1,286 −0.66 2007 −3,846 −2.0 2008 −2,602 −1.3 4 2009 −4,522 −2.3 Total: −16,173 −8.33 Source: DAPS. 0 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 FIGURE D.10. MANGO, 12.50 Yield (MT/Ha) Trend 0.3 trend 1989–2012 Loss Loss 10.00 Years (in MT) (US$, millions) 1980 −1,990 −0.79 1981 −2,316 −0.92 7.50 Yield (MT/Ha) 1983 −1,967 −0.78 1984 −2,292 −0.91 5.00 1991 −11,768 −4.68 1992 −9,120 −3.62 1999 −7,297 −2.90 2.50 2006 −12,160 −4.83 2007 −4,819 −1.91 2008 −14,947 −5.94 0.00 2009 −21,056 −8.37 1980 1983 1986 1989 1992 1995 1998 2001 2004 2007 2010 2010 −8,661 −3.44 2011 −10,041 −4.0 2012 −11,922 −4.7 Total: −120,355 −47.84 Source: DAPS. 80 Senegal APPENDIX E CROP PEST ANALYSIS TABLE E.1. RICE: PREHARVEST Species (common name) Losses Devastating pest Locusta migratoria (migratory locust) Total Oedaleus senegalensis (Senegalese grasshopper) Mythimnauni loreyi (maize caterpillar) Mythimnauni puncta (rice armyworm) Major pest Agrotis segetum (turnip moth) Dimorphopterus Major pest—role in virus and disease Orseolia oryzivora (African rice gall midge) Up to 15% transfer Cofana spectra (white leafhopper) Leptoglossus gonagra (squash bug) Trichispase ricea (rice, hispid) Cob borers and stem borers Chilo diffusilineus Chilo zacconius Corcyra cephalonica (rice meal moth) Earias insulana (Egyptian stem borer) Eldana saccharina (African sugarcane borer) Heteronychus licas (black sugarcane beetle) Sesamia calamistis (African pink stem borer) Spodoptera exempta (black armyworm) Spodoptera exigua (beet armyworm) Spodoptera littoralis (cotton leafworm) Minor pests Atherigona orientalis (pepper fruit fly) Gryllotalpa africana (african mole cricket) Nezara viridula (green stink bug) Pachnoda interrupta (chafer beetle) Parapoynx stagnalis (rice case worm) Agricultural Sector Risk Assessment 81 TABLE E.2. RICE: POSTHARVEST Species (common name) Losses Major postharvest pests Rhyzopertha dominica (lesser grain borer) Up to 15% Sitophilus zeamais (greater grain weevil) Sitotroga cerealella (grain moth) Stegobium paniceum (drugstore beetle) Tribolium castaneum (red flour beetle) Minor postharvest pests Cadra cautella (driedcurrantmoth) Up to 5% Liposcelis bostrychophila (book louse) Liposcelis entomophila (grain psocid) TABLE E.3. SORGHUM: PREHARVEST Species (common name) Losses Devastating pest Locusta migratoria (migratory locust) Total Oedaleus senegalensis (Senegalese grasshopper) Schistocerca gregaria (desert locust) Quarantine pest Thaumatotibia leucotreta (false codling moth) Major pest Atherigona soccata (shootfly) Agrotis ipsilon (black cutworm) Mythimnauni loreyi (maize caterpillar) Stenodiplosis sorghicola (sorghum midge) Diabolocatantops axillaris (devil grasshopper) Rhinyptia infuscata Trichoplusia ni (cabbage looper) Major pest—role in virus and disease Aphis spiraecola (Spirea aphid) transfer Melanaphis sacchari (yellow sugarcane aphid) Rhopalosiphum maidis (green corn aphid) Major pest—cob borers and stem borers Eldana saccharina (African sugarcane borer) Helicoverpa armigera (cotton bollworm) Sesamia calamistis (African pink stem borer) Spodoptera exempta (black armyworm) Spodoptera littoralis (cotton leafworm) Heteronychus licas (black sugarcane beetle) Minor pests Anoplocnemis curvipes (giant coreid bug) Less than 2% Atherigona orientalis (pepper fruit fly) Dimorphopterus spp. Nezara viridula (green stink bug) Pachnoda interrupta (chafer beetle) 82 Senegal TABLE E.4. SORGHUM: POSTHARVEST Species (common name) Losses Major postharvest pests Araecerus fasciculatus (cocoa weevil) Up to 30% Cadra cautella (driedcurrant moth) Rhyzopertha dominica (lesser grain borer) Sitophilus zeamais (greater grain weevil) Sitotroga cerealella (grain moth) Tribolium castaneum (red flour beetle) Quarantine pest Trogoderma granarium (khapra beetle) Total—needs to be destroyed Minor postharvest pests Alphitobius laevigatus (black fungus beetle) Corcyra cephalonica (rice meal moth) Liposcelis bostrychophila (book louse) Liposcelis entomophila (grain psocid) TABLE E.5. MILLET: PREHARVEST Species (common name) Losses Devastating pest Locusta migratoria (migratory locust) Total Major pest Pachnoda interrupta (chafer beetle) Up to 50% Cob borers and stem borers Spodoptera exempta (black armyworm) Minor pest Mythimna unipuncta (rice armyworm) TABLE E.6. MILLET: POSTHARVEST Species (common name) Losses Quarantine pest Trogoderma granarium (khapra beetle) Total—needs to be destroyed Major postharvest pests Cadra cautella (driedcurrant moth) Up to 30% Tribolium castaneum (red flour beetle) Minor postharvest pests Corcyra cephalonica (rice meal moth) Agricultural Sector Risk Assessment 83 TABLE E.7. COWPEA: PREHARVEST Species (common name) Losses Devastating pest Locusta migratoria (migratory locust) Up to total Oedaleus senegalensis (Senegalese grasshopper) Quarantine pest Amsacta moorei (tiger moth) Major pest Anoplocnemis curvipes (giant coreid bug) Aphis craccivora (groundnut aphid) Clavigralla tomentosicollis (African pod bug) Maruca vitrata (lima bean pod borer) Megalurothrips sjostedti (bean flower thrips) Major pest—role in virus and disease Aphis gossypii (cotton aphid) transfer Aphis spiraecola (Spirea aphid) Ferrisia virgata (striped mealybug) Frankliniella schultzei (cotton thrips) Nezara viridula (green stink bug) Minor pests Agrius convolvuli (sweet potato moth) Less than 2% Agrotis ipsilon (black cutworm) Aspidiotus destructor (coconut scale) Diaphania indica (cucumber moth) Helicoverpa armigera (cotton bollworm) Lampides boeticus (pea blue butterfly) Liriomyza trifolii (American serpentine leafminer) Ootheca mutabilis (leaf beetle, brown) Ophiomyia phaseoli (bean fly) Oxycarenus hyalinipennis (cotton seed bug) Spodoptera exigua (beet armyworm) Spodoptera littoralis (cotton leafworm) Trichoplusia ni (cabbage looper) TABLE E.8. COWPEA: POSTHARVEST Species (common name) Losses Quarantine pest Trogoderma granarium (khapra beetle) Total—needs to be destroyed Major postharvest pests Callosobruchus maculatus (cowpea weevil) Up to 30% Minor postharvest pests Cadra cautella (driedcurrant moth) Corcyra cephalonica (rice meal moth) Sitophilus zeamais (greater grain weevil) 84 Senegal APPENDIX F RATIONALE FOR RISK ASSESSMENT METHODOLOGY The rationale behind the risk prioritization exercise is based upon the nature of risk in agriculture, and it is necessary to define risk to understand its nature. First it is key to recognize that risk is an abstract concept associated with an activity, and cannot exist of itself. Only if an activity is undertaken does the possibility of its outcomes being different to those hoped. Secondly, risk involves a detrimental outcome normally but not always associated with reduced returns on investment. Farmers the world over are aware that there is “a risk” inherent in crop production and that in doing so they “risk” the assets that they invest in the process. Thirdly, the risk inherent in an activity is mag- nified as the ratio between the potential for downside loss and the capacity to absorb that loss (Weeks 1970) Commercial farmers who risk $1,000 per ha but have the assets to sustain a loss of that size take on less risk than a smallholder for whom the loss of $1,000 might result in complete bankruptcy and loss of livelihood. Since the word risk is widely used colloquially, It is also important to define what, for the purposes of this prioritization exercise at least, risk is not. It is not a specific type of event—such as drought. Although we talk of such events as “risks,” they are not risks of themselves. Although the potential for a drought to occur might contribute to the risk inherent in crop production, the fact that a drought might contribute much more to rain-fed crop production that it would to irrigated vegetable production highlights the fact that risk is more a property of an activity than of the event itself. Neither is risk the probability that an event might occur. Although the work risk is often colloquially used in place of chance or probability, the words are not interchangeable. A greater prob- ability that a detrimental event might occur can increase the risk inherent in an activity that is affected by that event, but does not itself constitute that risk. Risk as described above is an abstract and subjective concept, but it is a critical aspect of the stochastic processes that typify agricultural production systems. Although deter- ministic value chain analyses can help to determine the optimal path for rural develop- ment, the reality is often at odds with deterministic optima, frequently as a result of the impacts of risk upon rural investment decisions. Unless the perceived risks faced by Agricultural Sector Risk Assessment 85 smallholders are understood, it is difficult to design strate- being perhaps the most important), or they may be covar- gies to meet the objectives of rural development policies. iate, either through being dependent upon a common causal factor or through one affecting the other (the rela- A key aspect of risk that confounds analyses is that there tion between price and production can often result in is no standardized procedure for its quantification and covariate risks being faced by different groups of consum- measurement. Although it is possible to measure the ex ers). Each of these aspects of risk must be understood so post impacts of events that contribute to risk in terms of that the analysis of data and prioritization of risks can be the loss of yield or income resulting from those events, it is undertaken in an informed and effective manner. more difficult to estimate income foregone by producers who limit their level of investment due to the risks that A key aspect of risk prioritization is the perspective from they perceive to be associated with the production of a which the priorities should be determined. From a national specific crop. Although such ex ante impacts of risk might perspective, the highest priority risks may be those that be quantified as potential losses, but their attribution and result in the greatest impact on national and rural eco- measurement is extremely complex. nomic development, but from the perspective of a small- holder, risk is determined with regard to the livelihood of This is even more the case in the livestock sector. The the household, and especially its food security. It is the fear majority of ruminants in Africa are kept under extensive of food insecurity that drives the majority of investment management in ancient systems that were not set up to decisions faced by smallholders so that, when faced with achieve maximum profit or production but to enable their limited resources for investment in crop production, the adherents to be able to live and make use of specific smallholder will allocate those resources in such a way resources in areas where there may not be other viable that the risk to household food security is minimized. It is means of making a living or earning an income. These important therefore to qualify the priorities developed extensive livestock production systems are strongly driven from national data and to note and explain the differences by tradition, heritage and cultural issues. Normal fiscal that may occur when these priorities are assessed at the indicators may not capture the key characteristics of the household level. system where significant cultural and noncommercial drivers may be as, or even more, important than market The methodology thus focuses on three aspects of agri- drivers. cultural risk: (1) the frequency and impacts of specific events upon production at the national and individual To understand and prioritize the risks in agricultural and household level, (2) the measures employed to mitigate livestock sectors, this assessment has adopted a semi- those impacts, and (3) vulnerability, that is, the inverse of quantitative approach that does not focus on risk per se. the capacity of rural households to absorb the negative Rather it considers the impacts of the events that contrib- impacts of such events. The frequency and impacts of ute to the risk inherent in undertaking a given agricultural specific events (such as drought, disease epidemics of enterprise or livelihood. Such impacts are moderated by locust swarms) are assessed from available data and from mitigating measures and their final effect upon a house- interviews with key stakeholders. Mitigating measures are hold in terms of perceived risk will depend upon the assessed primarily through interviews with stakeholders, capacity of each household to absorb the moderated while the capacity of households to absorb negative impacts. All three aspects of risk are assessed to generate impacts has been assessed from both the literature and an eventual listing of key risks and their relative impor- from discussions with smallholders themselves. Such an tance to production. approach focuses mainly upon the ex post impacts of events that contribute to risk, while the ex ante impacts of Risks in agriculture can be both idiosyncratic, linked quite risk itself upon investment decisions are largely ignored. It specifically to individual commodities (especially prices is recognized that in terms of foregone production the ex risks), systemic, where large populations and different ante impacts of risk upon agricultural GDP are poten- crops are affected in the same way (inadequate rainfall tially as great if not greater than ex post impacts of risk 86 Senegal events (Elbers et al. 2007). Nevertheless, the measurement decisions. For the most part, it was considered that small- of perceived risk and its impact upon investment is a com- holders are homogeneous in their response to risk. This is plicated task beyond the resources available in this pre- recognized to be an approximation since the large pro- liminary assessment. Instead this assessment focuses upon ducers and traders tend to be less risk averse than their available data on the basis that ex ante impacts can be smaller counterparts, but it is not anticipated that the expected to be roughly proportional to ex post losses, so approximation would affect the final results of the that after taking into account the qualitative input from prioritization. interviews and focus groups, the priority components of risk can be readily identified and responses recommended. The qualitative results of interviews and focus groups was combined with the quantitative data to develop an overall The first step of the assessment is to determine the timing assessment of the key risks facing the main subsectors of and extent of reductions in yield below a predetermined agriculture (crop production, horticulture and livestock threshold, in this case, a third of a standard deviation production) to allow for the final prioritization and recom- below the trend line for national yield over time. Losses mendation of response measures. that exceeded this threshold are judged to have had a sig- nificant impact upon national production and the indica- The timescale of the risk is important. In the preparation tive value of such losses is calculated using either domestic of this assessment, the greatest significance has been or international prices according to the nature of the ascribed to those elements of risk that have discernible enterprise. Data for such analyses has been sourced from impact within the period of influence of government pol- national statistics or when these have not been readily icy, estimated in this case to be 10 years. Within such a available through FAOSTAT. For each year that a signifi- time frame of manageable interest, long-term trends such cant reduction in yield occurred, key stakeholders were as climate change contribute little to the ex ante risks per- canvassed as to the primary causes for the reduction in ceived by stakeholders, whereas ex post impacts remain that year. This information was triangulated from differ- below the statistical thresholds used to identify risk events. ent sources to obtain the most accurate assessment of The changing climate may indeed contribute to agricul- which events had led to particular crop losses. tural risk, but its impacts are not identified by the empiri- cal methodology used to develop this risk prioritization. National data can provide a useful indication of the main events that resulted in substantial loss of production and The response model of this methodology is graphically reduced agricultural GDP, but such data tends to obscure depicted in figure 6.5. It relies upon three different levels events of a more localized impact that may nevertheless of response according to the frequency and severity of contribute significantly to the risk inherent in a particular impact of risk events. Where risk events may be common, enterprise. These events were captured by extensive can- but their direct impacts can be effectively reduced, then vassing of stakeholders who were asked to identify the key mitigation measures are appropriate. Where events are elements of the risk that they faced when growing a crop less frequent and those impacts generally exceed capacity or undertaking a specific livestock enterprise. Neither for mitigation, then it is more appropriate to undertake does national data cover the extent or effectiveness of the transfer of risk impacts to another party (for example, mitigation measures that may already exist to reduce the through insurance). Finally, for the more exceptional risk impact of these events (and hence to reduce risk), but events that have a substantial impact, coping strategies these too were captured through stakeholder interviews. may be the most effective response. The application of Finally, national data cannot reveal the effect of vulnera- this model is described in more detail in the final section bility to real or potential loss upon producers’ investment of this report. Agricultural Sector Risk Assessment 87 APPENDIX G ANALYSIS OF WEATHER RISK EVENTS TABLE G.1. FREQUENCY OF LOW RAINFALL EVENTS BY REGION, 1981–2010 North North Central South Central South East South West Zone Saint Year Louis Louga Matam Dakar Thies Diourbel Fatick Kaolack Kaffrine Tambacounda Kedougou Kolda Sedhiou Ziguinchor 1980 –0.03 –0.28 –1.40 –0.09 –0.80 –1.02 –1.32 –1.39 –0.39 –1.16 –0.51 –1.94 –1.31 –1.40 Ext Dry 1981 0.98 –0.37 –0.08 –0.39 –0.10 –0.46 –0.33 0.10 0.24 0.63 0.63 –0.17 0.60 0.60 Normal 1982 –0.88 –0.04 –0.86 –0.57 –0.29 –0.86 0.06 –0.51 –0.89 –1.17 –0.97 –0.16 –0.74 –0.79 Normal 1983 –1.81 –0.62 –1.68 –1.57 –1.61 –1.78 –2.13 –1.22 –1.05 –1.57 –1.01 –1.46 –1.53 Ext Dry 1984 –2.09 –0.75 –1.54 –1.11 –0.84 –0.33 –0.21 –0.75 –0.50 –1.34 0.19 –0.14 0.14 –0.23 Dry 1985 –0.63 –0.62 –0.28 0.83 –0.43 –0.91 0.34 –0.67 –0.24 –0.36 0.38 –0.97 –0.80 –0.16 Normal 1986 –0.97 –0.59 –0.34 –0.01 –0.95 –1.15 –0.70 0.30 –0.04 –0.07 0.02 0.64 –0.07 –1.06 Normal 1987 0.25 1.06 0.64 0.38 0.03 –0.68 0.03 0.20 –0.01 –0.75 –0.28 –0.48 0.46 –0.64 Normal 1988 0.65 0.99 0.35 0.49 0.91 0.74 0.28 0.98 0.21 1.09 –0.87 –0.36 0.04 0.26 Normal 1989 0.44 1.84 –0.19 1.47 1.23 1.57 0.72 0.60 –0.18 0.33 –0.62 0.41 0.82 1.05 Excess 1990 –1.19 –0.49 –0.24 –0.95 –0.58 –0.09 –1.53 –1.12 –1.33 –0.58 –1.69 –1.14 –0.43 –0.10 Dry 1991 –0.89 –0.17 –1.28 –0.82 –0.70 –0.97 –0.84 –1.50 –2.04 –1.09 –0.19 –1.46 –1.18 0.94 Dry 1992 –1.16 –1.07 –1.87 –1.45 –0.88 –0.73 –0.78 –0.39 –0.34 –0.72 –1.08 1.14 –0.08 –1.06 Dry 1993 0.11 –0.04 –0.57 –0.46 –0.95 –0.68 –0.17 –0.08 –0.01 –1.23 –0.25 0.33 0.72 –0.15 Normal 1994 –0.41 –0.92 0.73 –0.99 –0.46 0.57 –0.12 0.08 0.44 1.55 0.50 0.89 0.76 –0.39 Normal 1995 0.37 –0.31 –0.51 0.48 0.04 1.20 0.87 –0.62 –0.70 –0.32 –0.58 –0.34 –3.21 –0.07 Normal 1996 –1.44 0.56 –0.10 0.46 –0.09 –1.21 –0.78 –1.24 –0.53 –0.28 –0.19 –1.61 0.21 –0.14 Dry 1997 –0.30 –1.29 –0.33 –1.34 –0.86 –0.47 –1.03 –0.68 –1.11 –0.17 2.01 0.61 –0.19 0.04 Dry 1998 0.16 0.12 –0.85 –0.34 –0.77 –1.02 –0.58 –1.11 –0.50 0.02 0.37 0.96 –0.50 –0.34 Normal 1999 0.64 0.40 1.26 0.69 0.88 1.00 0.62 1.95 1.38 1.31 1.24 1.99 0.76 2.70 Ex Excess 2000 0.19 0.74 2.68 0.54 1.83 0.50 1.27 1.07 0.58 0.26 1.26 –0.21 –0.60 0.25 Excess 2001 1.05 –0.39 0.26 –0.79 0.81 1.10 –0.10 0.26 –0.56 –0.83 0.17 –1.20 –0.18 0.26 Normal 2002 –1.10 –1.54 0.49 –0.71 –1.58 –0.93 –0.99 –0.59 –0.78 –1.50 –0.40 –1.32 –1.47 –1.89 Ext Dry 2003 1.21 –0.65 1.62 0.18 –0.73 –0.73 –0.06 0.57 0.23 2.46 2.89 1.04 1.38 –0.64 Excess 2004 –1.31 0.72 0.80 –1.15 –1.19 –0.59 –0.35 0.06 1.73 0.83 0.93 0.71 0.59 –0.84 Normal 2005 0.67 0.54 0.36 1.93 0.77 1.57 0.59 0.69 1.67 0.54 0.11 1.10 0.98 –0.05 Excess 2006 0.64 –1.16 –1.35 0.21 –0.09 –0.35 0.23 0.38 –0.10 –0.90 –0.39 –0.59 1.09 0.61 Normal 2007 0.03 –0.43 0.60 –0.87 –0.22 0.56 –1.39 –0.50 –0.78 –0.16 –1.78 0.16 –0.19 –1.12 Normal 2008 –0.28 0.66 0.35 0.86 1.09 1.87 1.63 0.58 0.62 1.12 –0.68 0.89 0.77 2.05 Excess 2009 1.03 2.15 0.57 1.19 1.72 1.24 1.96 0.77 0.06 1.21 –0.31 0.76 0.36 0.66 Ex Excess 2010 2.92 2.65 1.65 1.61 2.16 0.95 1.61 2.51 2.93 1.33 1.05 2.00 1.27 1.08 Ex Excess 2011 –0.16 0.49 –0.32 –0.68 0.21 –0.41 –0.38 –0.06 0.02 –0.53 0.98 –0.81 –0.31 –0.23 Normal 2012 1.00 0.16 1.31 1.84 1.13 1.55 2.06 1.42 1.70 1.13 –0.38 0.52 1.83 1.54 Ex Excess 2013 0.51 –0.17 0.04 1.27 1.28 0.79 1.16 0.82 0.43 0.39 0.03 –0.24 –0.05 0.79 Normal Source: ANACIM 2014. Agricultural Sector Risk Assessment 89 TABLE G.2. FREQUENCY OF HIGH RAINFALL EVENTS BY REGION, 1981–2010 North North Central South Central South East South West Zone Saint Year Louis Louga Matam Dakar Thies Diourbel Fatick Kaolack Kaffrine Tambacounda Kedougou Kolda Sedhiou Ziguinchor 1980 –0.03 –0.28 –1.40 –0.09 –0.80 –1.02 –1.32 –1.39 –0.39 –1.16 –0.51 –1.94 –1.31 –1.40 Ext Dry 1981 0.98 –0.37 –0.08 –0.39 –0.10 –0.46 –0.33 0.10 0.24 0.63 0.63 –0.17 0.60 0.60 Normal 1982 –0.88 –0.04 –0.86 –0.57 –0.29 –0.86 0.06 –0.51 –0.89 –1.17 –0.97 –0.16 –0.74 –0.79 Normal 1983 –1.81 –0.62 –1.68 –1.57 –1.61 –1.78 –2.13 –1.22 –1.05 –1.57 –1.01 –1.46 –1.53 Ext Dry 1984 –2.09 –0.75 –1.54 –1.11 –0.84 –0.33 –0.21 –0.75 –0.50 –1.34 0.19 –0.14 0.14 –0.23 Dry 1985 –0.63 –0.62 –0.28 0.83 –0.43 –0.91 0.34 –0.67 –0.24 –0.36 0.38 –0.97 –0.80 –0.16 Normal 1986 –0.97 –0.59 –0.34 –0.01 –0.95 –1.15 –0.70 0.30 –0.04 –0.07 0.02 0.64 –0.07 –1.06 Normal 1987 0.25 1.06 0.64 0.38 0.03 –0.68 0.03 0.20 –0.01 –0.75 –0.28 –0.48 0.46 –0.64 Normal 1988 0.65 0.99 0.35 0.49 0.91 0.74 0.28 0.98 0.21 1.09 –0.87 –0.36 0.04 0.26 Normal 1989 0.44 1.84 –0.19 1.47 1.23 1.57 0.72 0.60 –0.18 0.33 –0.62 0.41 0.82 1.05 Excess 1990 –1.19 –0.49 –0.24 –0.95 –0.58 –0.09 –1.53 –1.12 –1.33 –0.58 –1.69 –1.14 –0.43 –0.10 Dry 1991 –0.89 –0.17 –1.28 –0.82 –0.70 –0.97 –0.84 –1.50 –2.04 –1.09 –0.19 –1.46 –1.18 0.94 Dry 1992 –1.16 –1.07 –1.87 –1.45 –0.88 –0.73 –0.78 –0.39 –0.34 –0.72 –1.08 1.14 –0.08 –1.06 Dry 1993 0.11 –0.04 –0.57 –0.46 –0.95 –0.68 –0.17 –0.08 –0.01 –1.23 –0.25 0.33 0.72 –0.15 Normal 1994 –0.41 –0.92 0.73 –0.99 –0.46 0.57 –0.12 0.08 0.44 1.55 0.50 0.89 0.76 –0.39 Normal 1995 0.37 –0.31 –0.51 0.48 0.04 1.20 0.87 –0.62 –0.70 –0.32 –0.58 –0.34 –3.21 –0.07 Normal 1996 –1.44 0.56 –0.10 0.46 –0.09 –1.21 –0.78 –1.24 –0.53 –0.28 –0.19 –1.61 0.21 –0.14 Dry 1997 –0.30 –1.29 –0.33 –1.34 –0.86 –0.47 –1.03 –0.68 –1.11 –0.17 2.01 0.61 –0.19 0.04 Dry 1998 0.16 0.12 –0.85 –0.34 –0.77 –1.02 –0.58 –1.11 –0.50 0.02 0.37 0.96 –0.50 –0.34 Normal 1999 0.64 0.40 1.26 0.69 0.88 1.00 0.62 1.95 1.38 1.31 1.24 1.99 0.76 2.70 Ex Excess 2000 0.19 0.74 2.68 0.54 1.83 0.50 1.27 1.07 0.58 0.26 1.26 –0.21 –0.60 0.25 Excess 2001 1.05 –0.39 0.26 –0.79 0.81 1.10 –0.10 0.26 –0.56 –0.83 0.17 –1.20 –0.18 0.26 Normal 2002 –1.10 –1.54 –0.49 –0.71 –1.58 –0.93 –0.99 –0.59 –0.78 –1.50 –0.40 –1.32 –1.47 –1.89 Ext Dry 2003 1.21 –0.65 1.62 0.18 –0.73 –0.73 –0.06 0.57 0.23 2.46 2.89 1.04 1.38 –0.64 Excess 2004 –1.31 0.72 0.80 –1.15 –1.19 –0.59 –0.35 0.06 1.73 0.83 0.93 0.71 0.59 –0.84 Normal 2005 0.67 0.54 0.36 1.93 0.77 1.57 0.59 0.69 1.67 0.54 0.11 1.10 0.98 –0.05 Excess 2006 0.64 –1.16 –1.35 0.21 –0.09 –0.35 0.23 0.38 –0.10 –0.90 –0.39 –0.59 1.09 0.61 Normal 2007 0.03 –0.43 0.60 –0.87 –0.22 0.56 –1.39 –0.50 –0.78 –0.16 –1.78 0.16 –0.19 –1.12 Normal 2008 –0.28 0.66 0.35 0.86 1.09 1.87 1.63 0.58 0.62 1.12 –0.68 0.89 0.77 2.05 Excess 2009 1.03 2.15 0.57 1.19 1.72 1.24 1.96 0.77 0.06 1.21 –0.31 0.76 0.36 0.66 Ex Excess 2010 2.92 2.65 1.65 1.61 2.16 0.95 1.61 2.51 2.93 1.33 1.05 2.00 1.27 1.08 Ex Excess 2011 –0.16 0.49 –0.32 –0.68 0.21 –0.41 –0.38 –0.06 0.02 –0.53 0.98 –0.81 –0.31 –0.23 Normal 2012 1.00 0.16 1.31 1.84 1.13 1.55 2.06 1.42 1.70 1.13 –0.38 0.52 1.83 1.54 Ex Excess 2013 0.51 -0.17 0.04 1.27 1.28 0.79 1.16 0.82 0.43 0.39 0.03 –0.24 –0.05 0.79 Normal Source: ANACIM 2014. 90 Senegal APPENDIX H CROP PRODUCTION AND YIELDS FIGURE H.1. MILLET PRODUCTION, 2003–12 1.10 Production Area Yield 1,200,000 and area planted (Ha) 1,000,000 Production (MT) 0.83 Yield (MT/Ha) 800,000 0.55 600,000 400,000 0.28 200,000 0.00 0 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 Source: DAPS. FIGURE H.2. SORGHUM PRODUCTION, 2003–12 1.10 Production Area Yield 300,000 250,000 and area planted (Ha) 0.83 Production (MT) Yield (MT/Ha) 200,000 0.55 150,000 100,000 0.28 50,000 0.00 0 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 Source: DAPS. Agricultural Sector Risk Assessment 91 FIGURE H.3. COWPEA PRODUCTION, 2003–12 0.55 Production Area Yield 300,000 and area planted (Ha) 250,000 Production (MT) Yield (MT/Ha) 200,000 0.28 150,000 100,000 50,000 0.00 0 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 Source: DAPS. FIGURE H.4. RICE PRODUCTION, 2003–12 and area planted (Ha) 5.00 Production Area Yield 450,000 Production (MT) 400,000 Yield (MT/Ha) 4.00 350,000 300,000 3.00 250,000 2.00 200,000 150,000 1.00 100,000 50,000 0.00 0 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 Source: DAPS. FIGURE H.5. MAIZE PRODUCTION, 2003–12 1.38 1,625,000 1.10 1,300,000 and area planted (Ha) Production (MT) Yield (MT/Ha) 0.83 975,000 0.55 650,000 0.28 325,000 0.00 0 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 Source: DAPS. FIGURE H.6. GROUNDNUT PRODUCTION, 2003–12 1.375 1,625,000 and area planted (Ha) Production (MT) 1.1 1,300,000 Yield (MT/Ha) 0.825 975,000 0.55 650,000 0.275 325,000 0.000 0 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 Source: DAPS. 92 Senegal FIGURE H.7. COTTON PRODUCTION, 2003–12 1.50 60,000 and area planted (Ha) 1.20 45,000 Production (MT) Yield (MT/Ha) 0.90 30,000 0.60 15,000 0.30 0.00 0 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 Source: DAPS. FIGURE H.8. ONION PRODUCTION, 2002–11 40 200,000 and area planted (Ha) Production (MT) Yield (MT/Ha) 30 150,000 20 100,000 10 50,000 0 0 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 Source: DAPS. FIGURE H.9. TOMATO PRODUCTION, 2002–11 60 225,000 and area planted (Ha) 180,000 Production (MT) Yield (MT/Ha) 45 135,000 30 90,000 15 45,000 0 0 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 Source: DAPS. FIGURE H.10. POTATO PRODUCTION, 2003–12 40 22,500 and area planted (Ha) Production (MT) Area planted (Ha) Yield (MT/Ha) Production (MT) 18,000 Yield (MT/Ha) 30 13,500 20 9,000 10 4,500 0 0 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 Source: DAPS. Agricultural Sector Risk Assessment 93 FIGURE H.11. MANGO PRODUCTION, 2002–11 12.5 150,000 and area planted (Ha) Production (MT) 10.0 120,000 Yield (MT/Ha) 7.5 90,000 5.0 60,000 2.5 30,000 0.0 0 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 Source: DAPS. FIGURE H.12. GREEN BEAN PRODUCTION, 2002–11 15.0 14,000 and area planted (Ha) Production (MT) 11.3 10,500 Yield (MT/Ha) 7.5 7,000 3.8 3,500 0.0 0 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 Source: DAPS. 94 Senegal A G R I C U LT U R E G L O B A L P R A C T I C E T E C H N I C A L A S S I S TA N C E P A P E R W O R L D B A N K G R O U P R E P O R T N U M B E R 96296-SN 1818 H Street, NW Washington, D.C. 20433 USA Telephone: 202-473-1000 Internet: www.worldbank.org/agriculture