Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management 2 Measures, and Financing Mechanisms Through a Regional Lens © 2021 International Bank for Reconstruction and Development / The World Bank Some rights reserved This work is a product of the staff of The World Bank with external contributions. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does guarantee the accuracy, completeness, or currency of the data included in this work and do not assume responsibility for any errors, omissions, or discrepancies in the information, or liability with respect to the use of or failure to use the information, methods, processes, or conclusions set forth. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Nothing herein shall constitute or be construed or considered to be a limitation upon or waiver of the privileges and immunities of The World Bank, all of which are specifically reserved. 3 Rights and Permissions This work is made available under the Creative Commons Attribution-NonCommercial-ShareAlike 3.0 IGO licence (CC BY-NC-SA 3.0 IGO; https:// creativecommons.org/licenses/by-nc-sa/3.0/igo/legalcode). Under the Creative Commons Attribution license, you are free to copy, distribute, transmit, and adapt this work, for non-commercial purposes, under the following conditions: Attribution — Please cite the work as follows: World Bank. 2022. Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens. Washington, D.C.: The World Bank. Translations — If you create a translation of this work, please add the following disclaimer along with the attribution: This translation was not created by The World Bank and should not be considered an official World Bank translation. The World Bank is not liable for any content or error in this translation. Adaptations — If you create an adaptation of this work, please add the following disclaimer along with the attribution: This is an adaptation of an original work by The World Bank. Views and opinions expressed in the adaptation are the sole responsibility of the author or authors of the adaptation and are not endorsed by the World Bank. Third-party content — The World Bank does not necessarily own each component of the content contained within the work. The World Bank therefore does not warrant that the use of any third-party-owned individual component or part contained in the work will not infringe on the rights of those third parties. The risk of claims resulting from such infringement rests solely with you. If you wish to re-use a component of the work, it is your responsibility to determine whether permission is needed for that re-use and to obtain permission from the copyright owner. Examples of components can include, but are not limited to, tables, figures, or images. Any queries on rights and licenses, including subsidiary rights, should be addressed to World Bank Publications, The World Bank Group, 1818 H Street NW, Washington, DC 20433. Cover design: Fernanda Rubiano Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens Table of Contents Acknowledgments 11 Abbreviations  12 Executive Summary 16 1. INTRODUCTION TO FOOD SECURITY, AGRICULTURE, AND RISKS IN WEST AFRICA 23 2. THE ARM FRAMEWORK AT THE REGIONAL LEVEL  28 2.1 EXPANDING THE FRAMEWORK: REGIONAL INTEGRATION OF RISKS 30 2.2 A REGIONAL AGRIFOOD SYSTEMS RISK MANAGEMENT APPROACH 33 2.3 REGIONAL RISK MANAGEMENT INSTRUMENTS: WHAT NEEDS TO BE IN PLACE?  35 3. FOOD, TRADE, AND KEY PRODUCTION RISKS  37 3.1 FOOD CROPS AND LIVELIHOODS 38 3.2 MARKETS AND INTRAREGIONAL TRADE  42 3.3 KEY PRODUCTION RISKS  44 3.4 OTHER RISKS AFFECTING AGRICULTURAL PRODUCTION  51 3.4.1 Locusts 51 3.4.2 Conflict 52 4. FOOD CROPS, RANGELANDS, AND PRODUCTION AND LOSS ASSESSMENTS  55 4.1 IMPACTS OF PRODUCTION SHOCKS IN THE REGION 60 4 4.2 FOOD CROP PRODUCTION, RISK IMPACT, AND LOSSES IN THE REGION 62 4.2.1 Yam production and Crop Risk Assessment  64 4.2.2 Cassava production and Crop Risk Assessment  67 4.2.3 Rice production and Crop Risk Assessment  70 4.2.4 Maize production and Crop Risk Assessment  73 4.2.5 Sorghum production and Crop Risk Assessment  76 4.2.6 Millet production in West Africa  80 4.2.7 Cowpea production in West Africa 83 4.2.8 Plantain production in West Africa 85 4.3 CORRELATIONS BETWEEN YIELD AND WEATHER-RELATED RISKS 88 4.4 RANGELAND PRODUCTION RISK ASSESSMENT  91 5. HOTSPOTS FOR RISKS TO WEST AFRICA’S FOOD CROP PRODUCTION  94 5.1 HOTSPOT #1 CROP RISK ASSESSMENT 98 5.2 HOTSPOT #2 CROP RISK ASSESSMENT 100 5.3 HOTSPOT #2 CROP RISK ASSESSMENT 103 5.4 HOTSPOT #4 CROP RISK ASSESSMENT 105 6. FOOD PRICE RISKS AND MARKET INTEGRATION IN WEST AFRICA  109 6.1 FOOD PRICE RISKS IN WEST AFRICA 110 6.1.1 The impacts of weather-related and other risks on crop prices 111 6.1.2 Impact of social unrest on food prices  112 6.2 MARKET INTEGRATION IN WEST AFRICA 119 Table of Contents 6.2.1 Evolution of foreign trade and intraregional trade in West Africa 119 6.2.2 Food markets are integrated between countries  121 7. IMPACTS OF FOOD CROP PRODUCTION RISKS ON FOOD SECURITY  128 7.1 The Costs of Required Food Security-Related Humanitarian Support 132 8. WORRYING GAPS IN AGRICULTURE RISK FINANCING: INSIGHTS INTO FINANCING ARRANGEMENTS IN SIX COUNTRIES IN WEST AFRICA  134 8.1 RATIONALE FOR DRF IN THE AGRICULTURE AND FOOD SPACE 135 8.2 EXISTING NATIONAL AGRICULTURE RISK FINANCING INITIATIVES 140 8.2.1 National disaster funds 141 8.2.2 Strategic Food Reserves 143 8.2.3 Scalable safety nets 145 8.2.4 Agricultural insurance schemes 155 8.3 EXISTING REGIONAL AGRICULTURE RISK FINANCING INITIATIVES 148 8.3.1 Regional cooperation on food reserves 148 8.3.2 African Risk Capacity  151 8.3.3 Other regional efforts on DRF 152 8.3.4 Indicative funding gap analysis 153 9. REGIONAL RECOMMENDATIONS TO STRENGTHEN RESILIENCE AGAINST 5 AGRICULTURAL RISKS TO FOOD SECURITY  158 9.1 REGIONAL RISK MANAGEMENT IN WEST AFRICA 160 9.2 REGIONAL RISK FINANCING 163 9.3 STRENGTHENING RISK MANAGEMENT AT THE NATIONAL LEVEL TO ENABLE REGIONAL RISK SOLUTIONS 166 9.4 AREAS FOR FUTURE ASSESSMENT  169 REFERENCES 170 APPENDICES 176 APPENDIX A 177 APPENDIX B 201 APPENDIX C 217 BOXES BOX 3-1 MAPS 39 BOX 3-2 The Think Hazard Methodology 46 BOX 4-1 Methodology for Food Crop Loss Assessment 58 BOX 4-2 Analyzing Future Food Crop Yield Declines Due to Climate Change Impacts 63 BOX 4-3 Analysis of Correlation Between Selected Indexes and Crop Yields: Methodology and Detailed Results 89 BOX 5-1 Tools Used to Assess Crop Losses for this Report 108 BOX 7-1 Model Methodology 129 Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens Table of Contents FIGURES FIGURE ES 1 Cropwise Distribution of the Exposures for Main Food Crops in West Africa, Per Hectare and Per Value 19 FIGURE 2-1 The World Bank ARM Framework 29 FIGURE 2-2 Three Types of Risk 30 FIGURE 2-3 Agriculture and Food Systems Risk Indicators Assessed in this Report 32 33 FIGURE 2-4 Agricultural Risk Layering33 FIGURE 2-5 FSRP Regional Risk Management Layering Approach 36 FIGURE 3-1 Crop-Wise Distribution of the Exposures for Main Food Crops in West Africa, Per Value 40 FIGURE 3-2 Crop-Wise Distribution of the Exposures for Main Food Crops in West Africa, Per Hectare 40 FIGURE 3-3 Annual Agriculture Value-Added Growth is Highly Volatile in the Region (% growth) 45 FIGURE 3-4 Countries’ Relative Exposure to Water Scarcity and River Floods 50 FIGURE 3-5 Frequency of Locust Invasions 52 FIGURE 3-6 Conflict Hotspots Data Points or Events, by Type 53 6 FIGURE 4-1 Selection of Crops and Countries for the Analysis 57 FIGURE 4-2 Exposure by Crop Across the 15 Countries in the Region (value US$, millions, 1983–2018)63 FIGURE 4-3 West Africa: Historic Evolution of Yam Production 65 FIGURE 4-4 West Africa: Historic Evolution of Yam Crop Yields by Country 65 FIGURE 4-5 West Africa, Contribution of Each Country to the Expected AAL for Yam Crops 66 FIGURE 4-6 West Africa: Historic Evolution of Cassava Production 68 FIGURE 4-7 West Africa: Historic Evolution of Cassava Crop Yields 68 FIGURE 4-8 West Africa, Contribution of Each Country to the Expected AAL for Cassava Crops 69 FIGURE 4-9 Historic Evolution of Rice Production 71 FIGURE 4-10 Historic Evolution of Rice Yields 72 FIGURE 4-11 West Africa, Contribution of Each Country to the Expected AAL for Rice Crops 73 FIGURE 4-12 West Africa: Historic Evolution of Maize Production 74 FIGURE 4-13 West Africa: Historic Evolution of Maize Yields 75 FIGURE 4-14 West Africa, Contribution of Each Country to the Expected AAL for Maize Crops 76 FIGURE 4-15 West Africa: Historic Evolution of Sorghum Production 78 FIGURE 4-15 West Africa: Historic Evolution of Sorghum Production 78 FIGURE 4-16 West Africa: Historic Evolution of Sorghum Crop Yields 78 FIGURE 4-17 West Africa, Contribution of Each Country to the Expected AAL for Sorghum Crops 79 FIGURE 4-18 West Africa: Historic Evolution of Millet Production 81 FIGURE 4-19 West Africa: Historic Evolution of Millet Yields 81 Table of Contents FIGURE 4-20 West Africa, Contribution of Each Country to the Expected AAL for Millet Crops 82 FIGURE 4-21 West Africa: Historic Evolution of Cowpea Production 83 FIGURE 4-22 West Africa: Historic Evolution of Cowpea Yields 84 FIGURE 4-23 West Africa, Contribution of Each Country to the Expected AAL for Cowpea Crops 85 FIGURE 4-24 West Africa: Historic Evolution of Plantain Production 86 FIGURE 4-25 West Africa: Historic Evolution of Plantain Yields 86 FIGURE 4-26 West Africa, Contribution of Each Country to the Expected AAL for Plantain Crops 87 FIGURE 5-1 Selection of Crops and Countries for the Analysis 95 FIGURE 6-1 Number of Conflict Events (1997 to 2021) 112 FIGURE 6-2 Cassava Retail Price in US$ per kilogram in Benin, Côte d’Ivoire, Liberia and Sierra Leone from Jan 2005 to Sept 2020 113 FIGURE 6-3 Yam Retail Price in US$ Per Kilogram in Benin, Côte d’Ivoire, Ghana and Nigeria from July 2011 to July 2020 114 FIGURE 6-4 Maize Retail Price in US$ per kilogram (kg) in Chad, Côte d’Ivoire, Guinea-Bissau, Mali, Senegal, Togo from Jan 2000 to Sept 2020 115 FIGURE 6-5 Millet Retail Price in US$ Per Kilogram in Burkina Faso, Chad, Mali, Niger and Senegal from Jan 1990 to Aug 2020 to Sept 2020 116 FIGURE 6-6 Rice Retail Price in US$ Per Kg in Chad, Côte d’Ivoire (CIV), Ghana, 7 Guinea, Guinea-Bissau (GB), Mali, Senegal and Sierra Leone (SL), from Jan 2004 to Sept 2020 117 FIGURE 6-7 Sorghum Retail Price in US$ Per Kilogram in Chad, Mali, Nigeria and Togo from Jan 2000 to Sept 2020 118 FIGURE 6-8 Average Retail Prices for Cereals (Maize, Millet, Rice, and Sorghum) Across the Countries, US$ Per Kilogram, January 2000 to September 2020 122 FIGURE 7-1 a–d. Incremental Increase in Number of People Undernourished at Different Production Shock Return Periods (Mali, Chad, Burkina Faso, and Niger) 132 FIGURE 7-2 Comparison Between Status Quo Level of Humanitarian Appeals in Sahel Countries and the Modeled Increase in Humanitarian Appeals in Case of Disasters 133 FIGURE 8-1 Humanitarian Aid Inflows in US$, Millions and as a Percent of GDP (Burkina Faso, Chad, Mali, Niger) 135 FIGURE 8-2 Humanitarian Aid Inflows in US$, Millions and as a Percent of GDP (Sierra Leone, Togo) 136 FIGURE 8-3 Risk Layering Strategy for Government Ex-Post Risk Response 139 FIGURE 8-4 ECOWAS Food Storage Strategy 154 FIGURE 9-1 Layered Risk Management at Country and Regional Level for West Africa 160 FIGURE 9-2 Summary of National and Regional Options to Strengthen Financial Resilience for Food Security by Type of Shocks 163 FIGURE 9-3 Schematic of Potential Regional Risk Transfer Solution for Adaptive Social Protection of Sahel Countries 165 Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens Table of Contents FIGURE 9-4 Schematic of Potential Risk Finance Backstop to ECOWAS Regional Food Security Reserve 166 MAPS MAP ES 1 Sahel’s Six Livelihood Types are Transnational2 17 MAP 3-1 West Africa’s Main LVZ are Transnational8 39 MAP 3-2 Intraregional Trade Flows, 2017 43 MAP 3-3 Water Scarcity and Drought 45 MAP 3-4 Extreme Heat 47 MAP 3-5 River Floods (Admin 2 level) 48 MAP 3-6 River Floods with Added Layer Displaying 2016 Population Estimates 48 MAP 3-7 Density Map of Conflict Hotspot Data 53 MAP 4-1 Yam is Mainly Produced in the Humid Areas in West Africa 64 MAP 4-2 Harvested Areas of Cassava in the Region 67 MAP 4-3 Rice Production Across West Africa 70 MAP 4-4 Rice Yields, kg/ha, West Africa 71 MAP 4-5 Maize Production in West Africa 74 MAP 4-6 Sorghum Production Across West Africa 77 8 MAP 4-7 Sorghum Productivity Across West Africa 77 MAP 4-8 Millet Production in West Africa 80 MAP 4-9 Distribution of Rangeland Areas in West Africa 91 MAP 4-10 Expected Peak Season Accumulated NDVI 93 MAP 4-11 Exposures93 MAP 5-1 Administrative Level 1 Crop Risk Assessment Portfolio, Exposures 96 MAP 5-2 Admin 1 Level Crop Risk Assessment Portfolio, Expected AAL 96 MAP 5-3 Selected Food Insecurity Hotspots for Crop Risk Assessment 98 MAP 6-1 West Africa Market Integration, Cassava 123 MAP 6-2 West Africa Market Integration, Yam 123 MAP 6-3 West Africa Market Integration, Maize 124 MAP 6-4 West Africa Market Integration, Millet 125 MAP 6-5 West Africa Market Integration, Rice 125 MAP 6-6 West Africa Market Integration, Sorghum 126 TABLES TABLE 1-1 Natural Disasters Reported in 17 West African and Sahelian Countries, 1980–2020 24 TABLE 3-1 National Proportion of Admin 2 Level, by Risk Level (Per TH Methodology) 49 TABLE 3-2 Main Weather-Related Risks Per Crop in the Region 50 TABLE 4-1 Country Contributions to Aggregated Expected Food Crop Losses and Most Vulnerable Crops 60 TABLE 4-2 West Africa, Expected LaR Values for Yam Crops 66 Table of Contents TABLE 4-3 West Africa, Cassava Crops, Expected LaR Values for Different Return Periods 69 TABLE 4-4 West Africa, Rice Crops, Expected LaR Values for Different Return Periods 72 TABLE 4-5 West Africa, Maize Crops, Expected LaR Values for Different Return Periods 75 TABLE 4-6 West Africa, Sorghum, Expected LaR Values for Different Return Periods 79 TABLE 4-7 West Africa, Millet, Expected LaR Values for Different Return Periods 82 TABLE 4-8 West Africa, Cowpeas, Expected LaR Values for Different Return Periods 84 TABLE 4-9 West Africa, Plantains, Expected LaR Values for Different Return Periods 87 TABLE 4-10 Correlations of crop yields and number of people food insecure food, expressed in R2 90 TABLE 5-1 Expected Crop Area and Exposures for Selected Crops in Hotspot #1 99 TABLE 5-2 Hotspot #1, Expected LaR Values 99 TABLE 5-3 Expected Crop Area and Exposures for Selected Crops in Hotspot #1 100 TABLE 5-4 Expected Crop Area and Exposures for Selected Crops in Hotspot #2 108 TABLE 5-5 Hotspot #2, Expected LaR Values 101 TABLE 5-6 Expected Percentage AAL and LaR for Each Crop and Admin 1 Level Selected in the Portfolio in Hotspot #2 102 TABLE 5-7 Expected Crop Area and Exposures for Selected Crops in Hotspot #3  103 TABLE 5-8 Hotspot #3, Expected LaR Values 104 9 TABLE 5-9 Expected Percentage AAL and LaR for Each Crop and Admin 1 Level Selected in the Portfolio in Hotspot #2 104 TABLE 5-10 Expected Crop Area and Exposures for Selected Crops in Hotspot #4 106 TABLE 5-11 Hotspot #4, Expected LaR Values 106 TABLE 5-12 Expected AAL and LaR for Each Crop and Admin 1 Level Selected in the Portfolio in Hotspot #4 107 TABLE 6-1 Differences Between International and National Sources of Food Price Risks for Small, Highly Import-Dependent Countries. 110 TABLE 6-2 Africa Regional Integration Index 2015 120 TABLE 6-3 Share of Intraregional and Extraregional Food Trade 120 TABLE 7-1 Modelled loss in agricultural output from main crops and Depth of undernourishment130 TABLE 7-2 Total estimated number of people undernourished by type of event 131 TABLE 8-1 Mapping of Exemplary Agriculture Policy Measures to Risk Management Framework138 TABLE 8-2 Use of National DRF Instruments in the Six Focus Countries 141 TABLE 8-3 National Disaster Funds in Focus Countries 143 TABLE 8-4 Eight-Year Plan for Building the Regional Reserve and Increasing National Public Stocks (2014–21, in Megatons) 144 TABLE 8-5 Eight-Year Plan for Building the Regional Reserve and Increasing National Public Stocks (2014–21, in Megatons) 149 Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens Table of Contents TABLE 8-6 Financing Structure for the Establishment, Maintenance, and Governance of the Regional Reserve (in US$, millions) 150 TABLE 8-7 ARC Sovereign Drought Insurance Cover Over Time in Burkina Faso, Chad, Mali, Niger, Sierra Leone, and Togo 151 TABLE 8-8 US$ Amounts Kept at National Disaster Funds in the Focus Countries 154 TABLE 8-9 National and Regional Grain Storage Levels 155 TABLE 8-10 Retail Price by Type of Grain Held in the Storage Reserve (US$ Per Ton) 155 TABLE 8-11 National and Regional Grain Storage Levels 156 TABLE 8-12 Funding Gap for a Lower Severity Disaster in the Agriculture Sector 156 TABLE 8-13 Funding Gap for a Lower Severity Disaster in the Agriculture Sector 157 10 Acknowledgments This report was prepared by a World Bank team Economist), Ernest Ruzindaza (Senior Agriculture led by Åsa Giertz (Senior Agriculture Economist) Economist), Makoto Suwa (Senior Disaster Risk and consisting of Tenin Fatima Dicko (Financial Management Specialist), Cecile Lorillou (Disaster Sector Specialist), Cristina Stephan (Disaster Risk Risk Management Specialist), Sebastian Heinz Finance Specialist), Felix Lung (Social Protection (Agriculture Analyst), and Kaja Waldmann Specialist), Ramiro Iturrioz (Senior Agricultural Risk (International Development Analyst) for their Management Specialist), Clemence Tatin-Jaleran support in coordinating with stakeholders and (Actuary, Agriculture and Inclusive Insurance), their useful inputs on the alignment with the Mitik Ayalew Zegeye (Economist), and Aicha West Africa Food System Resilience Program; Lucie Sanou (Economist). This report forms part of and peer reviewers Diego Arias (Lead Agriculture the Food Systems Resilience Facility led by Katie Economist), Antoine Bavandi (Senior Financial Kennedy Freeman (Senior Agriculture Economist). Sector Specialist), and Oscar Anil Ishizawa Escudero (Senior Disaster Risk Management Specialist) The team gratefully acknowledges the with inputs from Thibaut Humbert (Disaster Risk 11 collaboration of the Economic Community of Modeling Specialist); and Platform for Agricultural West African States (ECOWAS) Department Risk Management (PARM), with inputs from of Agriculture on this work, including Sekou Ilaria Tedesco (Senior Technical Advisor, PARM/ Sangare, ECOWAS Commissioner for Agriculture, International Fund for Agricultural Development Environment and Water Resources; and Alain [IFAD]), Carlos Arce (Senior Technical Advisor, Sy Traore, ECOWAS Director of Agriculture and PARM), Jean Claude Bidogeza (Country Technical Rural Development. The team is grateful for the Specialist, PARM), and Emily Coleman (Agricultural contributions to this work from the AGRHYMET Insurance Technical Lead, Insurance for rural Regional Center, a specialized institution of the resilience and economic development), whose Permanent Interstate Committee for Drought valuable comments strengthened the quality of Control in the Sahel (CILSS), including Dr. this report. Critical guidance and support in the Souleymane Ouedraogo, Executive Director of the work for this report were provided by Deborah L. AGHRYMET Regional Center of the AGHRYMET/ Wetzel (Director, World Bank) and Chakib Jenane CILSS, Dr. Abdou Ali, Head of Information and (Practice Manager, World Bank). Research of CILSS/AGHRYMET. The report was copyedited by Helen Overmeyer The team would also like to thank World Bank from Dina Towbin and Associates LLC. The graphic colleagues Tobias Baedeker (Senior Agriculture design was carried out by Fernanda Rubiano. Economist), Amadou Ba (Senior Agriculture Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens Abbreviations AAL Annual Average Loss ACLED Armed Conflict Location and Event Data Project AIC Akaike Information Criterion ADF Augmented Dickey Fuller ADRiFi Africa Disaster Risk Financing AfCFTA African Continental Free Trade Area AfDB African Development Bank AFSLD African Food Security Leadership Dialogue AGRHYMET Regional Training and Application Center in Agrometeorology and Operational Hydrology AMU Arab Maghreb Union APTERR ASEAN Plus Three Emergency Rice Reserve ARC African Risk Capacity ARM Agricultural Risk Management ARV Africa RiskView ASRA Agrifood Sector Risk Assessment 12 ASY Aggregate Synthetic Index AU African Union AUC African Union Commission CAADP Comprehensive Africa Agriculture Development Program CAT DDO Catastrophe Deferred Drawdown Option CH Cadre Harmonisé CHIRPS Centre for Human Rights and Policy Studies CILSS Comité inter-états de lutte contre la sécheresse dans le Sahel (Permanent Interstate Committee for Drought Control in the Sahel) COMESA Common Market for Eastern and Southern Africa CORAF Conseil ouest et centre africain pour la recherche et le dével- oppement agricoles CSA Cimate-Smart Agriculture CSAIP Climate-Smart Agriculture Investment Plans CSI Cumulated Synthetic Index DID Développement international Desjardins DNPGCA Dispositif National de Prévention et Gestion des Crises Alimen- taires DOLS Dynamic Ordinary Least Square DRA Disaster Risk Assessment DRF Disaster Risk Financing DRR Disaster Risk Reduction EAC East African Community ECCAS Economic Community of Central African States ECOWAP Economic Community of West African Agricultural Policy ECOWAS Economic Community of West African States EM-DAT Emergency Events Dababase eMODIS EROS Moderate Resolution Imaging Spectroradiometer EO Earth Observation EROS Earth Resources Observation and Science ETLS ECOWAS Trade Liberalization Scheme EU European Union EWS Early Warning System FAO Food and Agriculture Organization FAOSTAT Food and Agriculture Organization Corporate Statistical Data- base FARM Agriculture and Rural Financing in Mali FCD Fonds Commun des Donateurs 13 FCI/CDRF Finance, Competitiveness, and Innovation/Crisis and Disaster Risk Finance FECONTRAF Food Emergency Contingent Trade Financing Facility FEWS NET Famine Early Warning Systems Network FSCCP Food Security Under Climate Change Program FSRF Food System Resilience Facility FSRP Food System Resilience Program GDP gross domestic product GFDRR Global Facility for Disaster Reduction and Recovery GGW Great Green Wall GHG Greenhouse Gas GIIF Global Index Insurance Facility GIZ Gesellschaft für Internationale Zusammenarbeit (Society for International Cooperation) GRiF Global Risk Financing Facility GVP Gross Value of Production HQIC Hannan Quinn Information Criterion IFAD International Fund for Agricultural Development IFPRI International Food Policy Research Institute ILRI International Livestock Research Institute IMI Internal Market Information System Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens IPC Integrated Food Security Phase Classification LaR Loss at Risk LGA Local Government Area LVZ Livelihood Zones MIS Market Information Systems MoU Memorandum of Understanding NDVI Normalized Difference Vegetation Index NGO Nongovernmental Organization OECD/SWAC Organisation for Economic Co-operation and Development/ Sahel and West Africa Club PARM Platform for Agricultural Risk Management PDF Probability Distribution Function PML Probable Maximum Loss R&D Research and Development REC Regional Economic Community RESOGEST Réseau des sociétés et offices chargés de la gestion des stocks 14 de sécurité alimentaire au Sahel et en Afrique de l’Ouest (Network of Companies and Offices Responsible for Inventory Management of Food Security in the Sahel and West Africa) RFE2 Rainfall Estimates 2 RPCA Food Crisis Prevention Network RRSA Regional Food Security Reserve SAARC South Asian Association for Regional Cooperation SADC Southern African Development Community SASPP Sahel Adaptive Social Protection Program SBIC Schwarz Bayesian Information Criterion SFR Strategic Food Reserve SI Stocks d’Intervention SNS Stocks National de Sécurité SPAM Spatial Production Allocation Model SPS Sanitary and phytosanitary SRSN Shock Responsive Safety Net TH ThinkHazard! TSU Technical Support Unit UNCTAD United Nations Conference on Trade and Development UNDP United Nations Development Programme UNECA United Nations Economic Commission for Africa UNOWAS United Nations Office for West Africa and the Sahel USAID United States Agency for International Development USGS United States Geological Survey VAR Vector autoregression VECM Vector Error Correction Model WAEMU West African Economic and Monetary Union WBG World Bank Group WFP World Food Programme WFP-VAM World Food Programme-Vulnerability and Mapping WMO World Meteorological Organization WRSI Water Requirement Satisfaction Index 15 Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens Executive Summary Addressing Risks through Regional Risk Synopsis Analysis This report provides an overview of agricultural A risk in West Africa1 by analyzing a variety of griculture is an increasingly risky risks and their impacts to food production and business in much of the world, including food security. The objective of the analysis is the West African region. The World Bank to identify the potential for risk management has developed an Agricultural Risk Management collaboration at regional levels by assessing (ARM) framework that assesses risks in systemic (1) the region’s most important crops for food production, markets, and enabling environments production, the production and price risks they to understand their total sectoral impacts and face, and the production losses experienced, to prioritize them. Prioritizing risks improves which impacts food security; (2) key regional targeting of risk management measures so that hotspots that concentrate high risk exposure scarce resources can be allocated where they and high losses; (3) costs of humanitarian have the most impact. It also helps identify how assistance to respond to these risks and the to align other agriculture, environment, and social current financial gaps in the ex-post financing of protection policies to manage existing risks. humanitarian interventions; and (4) existing risk These risks are usually identified and managed financing instruments in six select countries and at national levels, and the three key types are proposed measures for regional agricultural risk production risks, market risks, and enabling 16 management. environment risks. This Regional Risk Architecture and Financing This report focuses on how West African Mechanisms deliverable is one of the flagship countries can benefit from collaboration in initiatives that has been prioritized under the managing agrifood system risks and on the World Bank’s regional work. It extends World resulting need to adapt a regional lens to Bank work on agricultural risk management the ARM framework. Since both crop-specific by including the regional implications of risks growing areas and the risks they face often span and capturing impacts on food security for national borders, there are substantial advantages West African countries. This report is a joint effort that can be gained by stronger collaboration. There between the World Bank’s Agriculture Global is a need to build layered approaches to manage Practice and the Crisis and Disaster Risk Finance risk that combine risk-mitigating, risk-transfer, and (CDRF) Team in the Finance, Competitiveness, risk-coping instruments. These risk management and Innovation (FCI) Global Practice and is approaches are needed within countries, with closely coordinated with the Social Protection regional approaches building on national efforts. Global Practice. Further, the report is a result of This report provides a foundational analysis to a partnership between the Global Risk Financing begin identifying needed actions for West African Facility (GRiF), the Dutch Ministry of Foreign Affairs, countries and at regional levels. and the World Bank. This report was prepared as an output of the World Bank West Africa Food System Resilience Facility (FSRF) and its initial 1 For this report, West Africa includes the combined membership Regional Blueprint report that was developed to of ECOWAS and CILSS: Benin, Burkina Faso, Cabo Verde, Chad, Côte guide further analytical work. d’Ivoire, The Gambia, Ghana, Guinea, Guinea-Bissau, Liberia, Mali, Mauritania, Niger, Nigeria, Senegal, Sierra Leone, and Togo. Addressing Risks through Regional Risk Analysis D espite some progress, West Africa year timespan (1980–2020), droughts affected the and the Sahel are seeing chronic most people (over 99 million), while flooding has food insecurity, principally driven inflicted the highest economic damage (US$1.6 by complex interactions of climate change, billion). Insect infestations such as locust swarms population growth, an eroding natural resource are increasing. Climate modeling suggests that base, and civil conflict, with COVID-19 placing these hazards will intensify as climate change further strains on food and livelihood systems. impacts increase. Climate impacts worsen the Other shocks to the agrifood system are region’s civil insecurity by increasing food losses increasingly prevalent, and estimates suggest and food insecurity and increasing price spikes for that 19.6 million people were experiencing a food food, which has triggered unrest. crisis across 16 countries in the Sahel and West Africa region in March through May 2021, with an West African countries are largely agriculture- additional 27.1 million people at risk during the based economies and agriculture is the largest 2021 lean season (Food Crisis Prevention Network source of livelihood and income for most of (RPCA) April 2021). the population. The region’s agriculture sector has grown consistently and currently accounts Weather-related risks, especially droughts and for 30.5 percent of West Africa’s economy, while flooding, and agricultural pests and diseases also providing income and livelihoods to 70–80 have large impacts on West Africa’s food percent of the population. There are two main production, livelihoods, and economy. Across ecological divisions—the Sahelian countries of 17 17 West African and Sahelian countries and a forty- Burkina Faso, Chad, Mali, and Niger, which have MAP ES 1 Sahel’s Six Livelihood Types are Transnational2 Note: FEWSNet LVZ are not available for The Gambia, Guinea-Bissau, Ivory Coast, Ghana and Benin. Nigeria. Guinea, Sierra Leone, Liberia are still to be grouped and mapped. 1 https://fews.net/livelihoods Source: Original data based on Famine Early Warning Systems Network (FEWS NET) classification 2 Note: LVZ are not available for Benin, Côte d’Ivoire, The Gambia, Ghana, Guinea, Guinea-Bissau, Liberia, Nigeria, and Sierra Leone. FEWS NET LVZ for these countries are still to be grouped and mapped. Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens an arid and semiarid climate, and the rest of to climate change, and even a small swarm of the region, comprised of more humid coastal locusts covering 1 square kilometer can destroy countries. Agropastoralism is important regionally, the equivalent of what 35,000 people consume in with between 17 and 25 million agropastoralists a day. Understanding risks, impacts, and relative across West Africa and up to 17 percent of the exposure across countries can indicate the need population in Sahelian countries (United Nations and incentives for risk collaboration. Office for West Africa and the Sahel 2018; Food and Agriculture Organization [FAO] 2018b). Conflicts can strike rural areas and We identify six transnational livelihood zones transportation corridors, impacting both food (LVZ) that cross national boundaries in map 1.1: production and food security, and can be both pastoral, agropastoral, cereals production, mixed directly and indirectly triggered by a variety production, irrigated production, and other. of climate impacts. Security risks can affect production, transport, processing, marketing, West African countries import around 20 international trade, or distribution of crops with percent of total food consumed mostly from rural conflict waves causing food security ripples outside the region; intraregional trade in in urban areas as prices increase due to scarcity. ECOWAS is likely around 12 percent over Nigeria, Côte d’Ivoire, Mali, and Burkina Faso are a decade. The value of intraregional food the most impacted by conflict, while Mali, Burkina export and import was estimated at US$2.2 Faso, Benin, and the Lake Chad area have multiple billion and US$2.3 billion, respectively, in 2019 border areas that have experienced significant (United Nations Conference on Trade and insecurity in the past few years. Development [UNCTAD] 2020). West Africa’s low intraregional trade is mainly due to inadequate internal transport infrastructure (roads and rail 18 networks) and to road harassment, leading to higher transport and transaction costs and weak competitiveness of regional products. The West African region is among the world’s most vulnerable to climate change, and impacts are already affecting food production. Erratic rainfall has made both drought and flooding more frequent events with severe impacts on food and water security, leading to high volatility in annual agriculture value-added growth. Extreme heat, water scarcity, and drought happen more often in the Sahelian countries, while flooding events are more common in coastal countries. Flood events are more frequent, more expensive, and more localized than droughts, which can span huge areas. The Gambia, Mali, and Togo are high-risk countries for both droughts and floods, while Côte d’Ivoire, Liberia, Chad, and Ghana are low risk. Locust invasions, which are also linked Food Crop Production and Losses N igeria produces from around 35 to 65 and fonio. Figure ES-1 shows the large difference percent of the region’s total production, in crop importance in terms of overall area making it by far the region’s biggest planted (where the top three are sorghum with crop producer. Overall crop production varies maize and millet tied for second) and economic across LVZ, but the main crops consumed and value (where yams are most important, followed vital for the region’s food security are cassava, yam, by cassava and, distantly, rice). rice, maize, millet, sorghum, plantain, cowpea, FIGURE ES 1 Cropwise Distribution of the Exposures for Main Food Crops in West Africa, Per Hectare and Per Value Cow peas Sweet potatoes 9% 0% Millet Plantains 17% Yams Fonio 1% 40% 1% Cassava 27% Yams 11% Sorghum 20% VALUE HECTARE Cassava 12% Rice 13% 19 Fonio 0% Rice Cow peas 12% 1% Plantains Maize Maize 3% Millet 7% 17% 3% Sorghum 6% Source: Original figure based on Food and Agriculture Organization Corporate Statistical Database (FAOSTAT) The team analyzed food crop production and millet, plantains, cowpeas, and fonio together and losses for the most important food crops account for about 7 percent. Each of the countries for each country. For each country and each and eight of the most important food crops are crop, they identified the Average Annual Loss assessed in chapter 4. In a separate analysis, the (AAL) and the Loss at Risk (LaR), a percentile of team attempted to identify correlations between loss distributions for different time scales. The production and climatic (drought) indicators region’s 15 countries lose an average of US$4.7 that would support early estimations of yield billion through production shortfalls of their most reductions and decision-making for risk responses. important food crops; Nigeria loses US$3.1 billion The findings were not sufficiently robust given alone. Countries vary greatly in the loss value and the data used, but the approach could prove the time frame for expected losses. Across the 15 important in building rapid responses if better countries, yam, cassava, and rice have the largest quality local data becomes available. aggregate AAL in terms of value (40, 27, and 13 percent, respectively), while maize and sorghum each only contribute 7 and 6 percent to losses, Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens Food Security Hotspots A hotspot approach was used to look main production areas for the high value yam at the food supply risk in critical West and cassava crops. Overall, the impacts for three African regions. Six fragility and food hotspots would be around 1–13 percent for 1 in insecure hotspots were identified through the 10-year risks and up to 20–25 percent losses for 1 analysis, and a detailed food crop risk assessment in 100-year risks. This hotspot approach allows for was done for four hotspots. One hotspot cross-border solutions to managing agriculture alone accounts for 61.2 percent of the total risk and food security risks—that is, regional risk exposures because this hotspot includes the management. From Risks to Reality: Both Shocks and Bene- fits Travel Across Regions W eather risk mitigation and political amount on average to more than US$700 million stability in West Africa are necessary per year and climb to more than US$1 billion every to ensure food price stability and five years. Climate-linked shocks and other events food system resilience and to further improve will intensify, and current humanitarian funding is 20 food security based on a market integration not covering chronic food insecurity. Additional analysis. Price volatility is a major issue in the funding will be needed to support chronic food region. There are significant effects on food shortages and even low probability (5 percent) prices from climate shocks and social unrest. shocks that are likely to happen, moving millions Yet positive links were found for different crops of people from food insecure to undernourished. across countries; for example, reducing millet prices in Senegal could lower prices and benefit households in Burkina Faso, Mali, and Niger. Risks to food crop production are becoming more frequent, turning into food crop losses and threatening the region’s food security. Food crop production losses in Burkina Faso, Chad, Mali, Niger, Sierra Leone, and Togo could Mind the Gaps: Supporting Agricultural Risk Financing I ncreasing populations, increasing risks, creation of an ECOWAS Regional Food Security and increased hazard events signal that Reserve (RRSA). mechanisms for reducing, transferring, and risk coping are urgently needed. Countries Most countries have significant funding gaps need solid plans, approaches, institutions, and and would face challenges in responding to funds ready to respond to food production shocks even small-scale disasters. To identify funding and, more importantly, working to prevent them. gaps, it is essential to know likely government Yet the six countries facing the highest risks lack shortfalls, so analysis of both risk retention and funding to cover humanitarian and food security risk transfer mechanisms were made, considering needs and are less equipped to balance out both agricultural yield shocks and humanitarian broader economic losses. There is little possibility aid for food insecurity. Large funding gaps are to use agricultural risk financing to anticipate and likely for all countries. The combined humanitarian mitigate shocks; virtually all funding is delivered and economic funding gaps are largest for ex-post crisis. Key tools such as national disaster Niger, which faces a gap of US$768 million for funds, strategic food reserves, scalable safety 1 in 5 years food production shocks and US$1.1 nets, and agricultural insurance schemes exist billion for 1 in 10 years food production shocks. but there is extremely limited use of these. While Mali follows, with respective figures of US$341 regional efforts need to link to country efforts, million and US$491 million. All countries have 21 there are regional initiatives emerging, including high exposure to disasters and limited national or the strong use of sovereign insurance coverage regional financial instruments to respond. from the African Risk Capacity (ARC) and the Recommendations for Resilience: Reducing Risks to Food Security T he findings of this report show that disasters. Regional financing solutions could there are opportunities to share risk use risk diversification benefits and operational management instruments between economies of scale, lowering overall cost, and countries in West Africa. Countries in the could support sharing disaster response expertise region face similar agricultural production across countries, especially when countries face and humanitarian shocks and risk transmitting disasters across wide areas. between countries, which provides a strong basis to improve regional risk management and Nevertheless, the cost of risk impacts, financing instruments. Further, most countries especially when converted to humanitarian have significant funding gaps and would face costs, are currently so high that a combination challenges in responding to even small-scale of risk management measures (mitigation- Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens transfer-coping) at national and regional levels will have to be implemented at all layers of risks. The region should prioritize those risks that have the biggest impacts on the region, whether from a country perspective or at crop level. The regional food reserve should be aligned with the risk-based needs. Further, for several countries, risks transmit through prices into countries through trade, and addressing risks in certain countries (for example, Senegal) would therefore benefit consumers in more countries in the region. The scope of risk impacts on food crop production will also require a combination of risk financing instruments at national and regional levels. Effective ARM in the region requires a comprehensive risk financing approach at both national and regional levels. Regional solutions include regional risk pooling for adaptive social safety nets and creating a risk financing backstop for the regional food reserve. 22 Going forward, it will be important for all countries to strengthen their ability to manage risks and effetely integrate risk management into general agriculture and food security policy objectives. National level Agrifood Sector Risk Assessments (ASRAs) and Disaster Risk Assessments (DRAs) should therefore be conducted for those countries where such assessments are not yet available. Other recommendations for national-scale actions include strengthening investments in agricultural productivity, social protection, and Disaster Risk Reduction (DRR) measures; and the capacity of national response finance systems to respond to lower severity shocks. A fter years of progress, Africa is experiencing a decline in food security, and West Africa and the Sahel region are seeing chronic food insecurity exacerbated by spikes in transient food insecurity. The current decline in food insecurity is principally driven by complex interactions of climate change, population growth, an eroding natural resource base, and the incidence of conflict. With persistent chronic food insecurity, shocks to the agrifood system are increasingly prevalent, giving 1 rise to transient food insecurity. In the March– INTRODUCTION TO May 2021 period, some 19.6 million people were estimated to be experiencing a food crisis across 16 countries in the Sahel and West Africa regions FOOD SECURITY, (Food Crisis Prevention Network [RPCA] 2021). An additional 27.1 million people are expected to be AGRICULTURE, at risk during the 2021 lean season (RPCA 2021). 23 While COVID-19 has put additional strains AND RISKS IN WEST on the region’s food security, the increase in transient food insecurity results from AFRICA prevailing and deepening vulnerabilities in the food system. Naturally occurring events such as weather-related risks and agricultural pests and diseases have large impacts on West Africa’s food production, livelihoods, and economy. Especially in the dry regions, recurring weather- related risks cause food crises, resulting in loss of lives and livelihoods. Between 1981 and 2020, all West African and Sahelian countries recorded 757 natural disasters, with 52 percent due to extreme weather and hydrometeorological events. These events accounted for 98 percent of affected populations (around 104.17 million people) and 98 percent of damages (US$1.74 billion) (see table 1.1). Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens TABLE 1.1 Natural Disasters Reported in 17 West African and Sahelian Countries, 1980–2020 Count of Total Damages ('000 Total Affected Popula- Disaster Type Event US$) tion Drought 68 71754 99,127,840 Earthquake 1 0 21,436 Epidemic 310 0 1,280,866 Extreme temperature 3 47000 1,000,000 Flood 306 1620637 29,513,568 Insect infestation 37 0 500,000 Landslide 8 30000 24,256 Storm 32 5257 181,890 Volcanic activity 2 0 8,806 Wildfire 6 0 8,793 24 Total 773 1,774,648 131,667,455 Hydrometeorological 409 1,774,648 129,823,298 disasters Hydrometeorological 53% 98% 99% (% of total) Insect infestation (% of 4.8% N/A 0.4% total) Source: Original table based on data from the Emergency Database (EM-DAT), Université Catholique de Louvain (UCL) 1980-2020 Climate change already affects the region, led to increased pests and disease outbreaks, with more frequent extreme weather events accounting for 5 percent of total natural disasters that result in other food production risks in the region in the last 70 years. In 2003–05, the such as pest and disease outbreaks. The region faced a major desert locust invasion that World Meteorological Organization’s (WMO) destroyed over 12 million hectares, affecting most recent Global Annual to Decadal Climate approximately 8 million people (Emergency Update indicates that from 2020–24, the Sahel database). region is likely to become wetter than the recent past (WMO 2020), which is expected to cause Other risks such as conflicts and price volatility flooding and increase displacement of vulnerable also affect food security in the region. A sharp populations. The increased frequency of extreme uptake in violence that started in 2012 has weather and hydrometeorological events has turned into diverse forms of conflict across the subregion, encompassing violent extremism, more resilient African food system through joint armed rebellion, and banditry. In the past, price action, the African Union Commission, African shocks have affected the region. During the 2008 Development Bank (AfDB), IFAD, FAO, and the international food price crisis, maize and rice prices World Bank co-convened the African Food in Côte d’Ivoire, Guinea, Mali, and Niger reached Security Leadership Dialogue (AFSLD) in Kigali their highest levels in the previous 10 years. West in August 2019. An outcome of this event is the African urban consumers and net cereal buyers in continent-wide Food Security under Climate rural areas were the hardest hit by the food crisis, Change Program (FSCCP) for Africa, under given the region’s high dependence on imports preparation by the World Bank, regional partners, from the international market and the impact of and other actors. West Africa has been selected price increases passed on to domestic consumers as the first region for the FSCCP, and measures and producers. will be implemented under the proposed World Bank financed West Africa Food System Resilience The COVID-19 pandemic has exposed existing Program (FSRP). FSRP is led by the mandated weaknesses of the region’s food system and regional bodies ECOWAS and CILSS.3 pushed millions into food insecurity. While confirmed COVID-19 cases in West Africa are The Regional Risk Architecture and Financing low compared to other regions and food supply Mechanisms deliverable is one of the flagship disruptions have remained limited so far, the initiatives that has been prioritized under the COVID-19-induced health and economic crises World Bank’s regional work. Deepening the reduced access to food, for example, through understanding of regional risks and potential rising unemployment and reduced purchasing risk instruments for further collaboration was power, particularly for urban poor (RPCA 2020). also championed and supported by partners 25 It has also disrupted cross-border migratory during the regional consultative conference movements, blocking nearly 57,000 livestock “Under the Palaver Tree: Unpacking Food Systems farmers with some 1.5 million cattle in January Resilience in West Africa” that was hosted by 2021 (RPCA 2021). The Sub-Saharan Africa ECOWAS, CILSS, and the Conseil ouest et centre economy is projected to decline by 3.3 percent africain pour la recherche et le développement in 2020, confirming the region’s first economic agricoles (West and Central African Council for recession in 25 years. The region’s fragile countries Agricultural Research and Development) (CORAF) are expected to experience a stronger decline in July 2020. The virtual conference brought 400 in growth as COVID-19 worsens the drivers of participants together from West African countries, fragility. The number of people living in poverty regional bodies, and development partners and is expected to increase significantly (up to 40 representatives from the private sector, academia, million additional people in Sub-Saharan Africa, and civil society. with negative effects on food security in the short and long-term [World Bank in Africa, 2020]). The objective of this report is to increase knowledge of systemic risks to food The urgency in making Africa’s agrifood production across the region and their systems more resilient is being recognized impacts on food security and to assess the and countries and development partners are potential for regional risk management mobilizing resources. To outline a vision for a collaboration. The report looks at risks through 3 The program objective is to strengthen regional food system risk management, improve sustainability of the productive base in targeted areas, and develop regional agricultural markets. The main beneficiaries of the program are ECOWAS, CILSS, and member countries across phases of a multiphase approach. Phase 1 and Phase 2 countries are Burkina Faso, Chad, Mali, Niger, Sierra Leone, and Togo. Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens a regional lens to see how countries in West work on ARM but will extend the existing Africa and the Sahel can collaborate around risks, framework to include regional implications largely focusing on production risks and certain of these risks and to capture impacts on food price risks. It also assesses the losses for major security. ARM and Disaster Risk Financing (DRF) food crops and impacts on food security and have long been a priority for the World Bank estimates the costs of humanitarian assistance to and this report builds on previous work and the respond to these risks and the current financial wealth of experience that the World Bank has gap for countries. The analysis for this report was gained on this topic. The World Bank has over conducted in support of the preparation of the the past decades developed a series of ASRAs West Africa FSRP lending operation financed in Africa, Asia, Central Asia, and Latin America, by the World Bank and other partners that also together with frameworks and methodological determined which countries were selected for guidelines for ARM. Taking a regional approach certain analyses.4 to agricultural risks and impacts on food security is a more recent endeavor, for example, with the This report is a joint effort between the World 2019 Southern African Development Community Bank’s Agriculture Global Practice and the (SADC) Agriculture Risk Financing Report. CDRF team in the FCI Global Practice and is Previous work has mainly focused on agricultural closely coordinated with the Social Protection losses without assessing impacts on food access. Global Practice. Further, it is a result of a Thus, this report aims to contribute to the partnership between the GRiF, the Dutch Ministry discourse on regional approaches to assessing of Foreign Affairs, and the World Bank. This agricultural production risks at the regional level, report was prepared as an output of the World the impacts of these risks on food security across 26 Bank West Africa FSRF and its initial Regional countries, and regional approaches for related risk Blueprint report that was developed to guide management. further analytical work. Other, related outputs under the FSRF are “Hotspots, Fragility, and Further, the World Bank’s CDRF team has Integrated Approaches” on the linkages between addressed financial resilience for natural agriculture production risks and conflicts, “Digital disaster shocks and crises in more than 60 Climate Information and Ag Advisory Delivery countries. In West Africa particularly, FCI/CDRF Mechanisms,” which will identify viable delivery is supporting countries with (1) risk analytics models for climate information and agriculture tools to quantify fiscal impact and financial cost advisory to reach farmers at scale, and “Trade: of shocks; (2) integrating financial planning for Toward More Data and a Scorecard Methodology” shocks in fiscal risk management framework and that will create a broadly owned methodology financing strategies; (3) defining policy actions to for a country scorecard on intraregional trade of strengthen financial protection against disasters, food commodities. Sections on West Africa’s food shocks and crises by developing DRF strategies system and regional trade in chapter 2 in this built on a cost-effective use of funding sources report draw heavily from the Blueprint report. through a risk-layered approach; (4) structuring risk financing instruments (reserve funds, This report adds to the extensive World Bank Catastrophe Deferred Drawdown Option, risk 4 FSRP will be implemented as a multiphased program approach where the full set of countries is expected to include Benin, Burkina Faso, Chad, Côte d’Ivoire, The Gambia, Ghana, Guinea, Guinea-Bissau, Liberia, Mali, Niger, Nigeria, Senegal, Sierra Leone, and Togo. The first phase of the FSRP will include Burkina Faso, Chad, Mali, Niger, Sierra Leone, and Togo. transfer products such as insurance); (5) scalable Finally, chapter 9 provides concluding remarks safety nets; and (6) capacity building. This report and proposes a few next steps. In addition to also builds on other work in West Africa: Earth the main volume, two background reports are Observation (EO) for the CDRF program in Africa also published: (1) Detailed Agricultural Loss that specifically addresses requests for support Estimates at National and Subnational Levels and on designing new, comprehensive, and robust (2) Agriculture Risk Financing in Western Africa: EO-based drought indices for early triggers for Analysis and Options for Enhancing Food Security early action; Global Index Insurance Facility (GIIF) in Burkina Faso, Chad, Mali, Niger, Sierra Leone, and climate and disaster risk insurance programs and Togo. providing technical assistance and delivering agriculture risk transfer solutions in countries; Shock Responsive Safety Net (SRSN) as part of the Sahel Adaptive Social Protection Program (SASPP) and other social protection projects; and analytical works such Sahel risk financing solutions for the livestock sector. The report is structured as follows: Chapter 2 describes the World Bank’s ARM framework and extends it to the regional level. Chapter 3 provides a brief overview of the region’s food crops and LVZ, markets and intraregional trade, and key agricultural production risks. Chapter 4 is 27 an overview of production for the most important food crops, the combined costs of production risks, and production for rangelands in the region. Chapter 5 presents an analysis of potential volatility in different LVZ that span country borders and identifies particularly vulnerable hotspots or areas of high concern from a food security perspective. Chapter 6 analyzes food price risk and market integration in West Africa through some of the main crops considered in chapter 4. Chapter 7 then provides an overview of production losses from risks impact food insecurity in six countries and shows existing gaps in the ex-post financing of humanitarian interventions. Chapter 8 presents a set of proposed measures for regional ARM and explores to what extent existing in-country financing resources to respond to food security crises suffice to respond to the estimated financing needs for different production shocks. Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens Key points in this chapter: • The World Bank’s ARM framework assesses risks in systemic production, markets, and enabling environments to understand their 2 total sectoral impacts and to prioritize them. Prioritizing risks improves targeting of risk THE ARM management measures so scarce resources can be allocated where they have the most FRAMEWORK AT impact and so other agriculture, environment, and social protection policies can be aligned with existing risks. THE REGIONAL • Managing risks through a layered approach means that the impacts of low-severity, high- LEVEL 28 frequency risks should be mitigated or reduced to the extent possible, while medium-severity risks can be managed by actors in the sector through risk transfers offered to the private sector, with high-severity, infrequent risks This chapter describes the World Bank’s ARM managed through established risk response framework and the layered approach to ARM mechanisms such as social safety nets or and proposes expanding the framework humanitarian assistance, financed by pre to support regional agricultural and food arranged contingency funds, budget support, systems risk management. sovereign risk transfer products, and so on. All risk management strategies, whether private or public, include a combination of risk management instruments for the three layers. • For regional ARM, countries must have an incentive to pool and jointly manage risks. For certain regional risk instruments to have an affordable price for a regional pool, risks need to be diversifiable between countries to a certain extent. T he World Bank’s ARM framework from constraints and trends in that they are not was developed as a comprehensive a constant in the regular environment in which approach to managing systemic risks actors operate. Systemic or covariate risks differ to the agriculture sector (figure 2.1). Risks are from nonsystemic or idiosyncratic risks in that uncertain events that lead to losses5 and differ they simultaneously affect a group or sector. FIGURE 2-1 The World Bank ARM Framework Instruments  Investments Technical Assistance Policy Stakeholders Producers Commecial Sector Risks Strategies Public Sector Production Mitigate Market Transfer 29 Enabling Evironment Cope Source: World Bank ARMT The World Bank’s ARM framework the exchange rate to fluctuate. To effectively distinguishes between three types of risks: manage a risk, it is therefore important to production risks, market risks, and enabling identify its root cause. environment risks (figure 2.2). These risks can be either manmade or natural and exogenous Production risks are often those most visible (caused by external factors) or endogenous risks in the agriculture sector, including (caused by domestic factors). Risks are also often weather-related events (such as droughts, interlinked; for example, production risks often floods, and cyclones), outbreaks of agricultural cause price fluctuations due to production pests and diseases, and damage caused by volatility, or erratic policy changes may cause animals, windstorms, or fire. But lack of access 5 Some literature, notably the World Bank’s World Development Report 2014: Risk and Opportunity—Managing Risk for Development (2013), also assigns potential positive returns from risks. In the ASRA framework, the term “risk” comprises only unpredictable events that, unmanaged, lead to losses. Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens FIGURE 2-2 Three Types of Risk Production risks Market risks Enabling environment risks $ - Weather events - Price volatility - Con ict/instability - Pests and diaseases - Exchangerate volatility - Erratic policy changes - Bush res - Interest rate volatility - Erratic trade restrictions - Counterparty default Source: World Bank ARMT to labor due to pandemics or other lockdowns consumers are worried by high prices (raising can also be considered a production risk. their expenditure). Another market risk is Production risks are mostly associated with exchange rate volatility, which can affect the yield reductions but can also affect product price of outputs and inputs. quality. 30 Enabling environment (macro-level) risks Market risks affect the price and availability also affect the agriculture sector. Macro-level of outputs and inputs. Commodity markets risks cover unexpected changes in the broader can have a high degree of volatility caused by economic environment that agriculture changing local and global supply and demand. operates in and can include changes in At times, markets disappear overnight due to government or business regulations, fiscal rapid changes in domestic or external policies, and monetary policy settings, external trade which happened in many countries during the restrictions, political instability, corruption, COVID-19 lockdown. Producers are concerned regional conflict, and domestic unrest. about low prices (reducing their income); 2.1 EXPANDING THE FRAMEWORK: REGIONAL INTEGRATION OF RISKS T his report focuses on how West risks management measures will still happen African countries can benefit from at the national level. Yet for certain measures, collaboration in managing agrifood collaboration between countries brings system risks and the resulting need to adapt advantages beyond the sum of the individual a regional lens to the ARM framework. Even country contributions, such as diversification when countries collaborate around risks, most of risk across a larger portfolio, enhanced negotiation power in an eventual placement and food markets. Understanding risk of a risk transfer product in the market, transmission is especially important for economies of scale in research, and access to getting countries to collaborate around more information through shared information coping instruments. A prerequisite for systems. Countries will generally give up well-functioning ex-post risk response fewer tools than those countries that share mechanisms (such as transfers and safety a regional currency, but there are still certain nets) is that the impacts of risks are costs that come with collaboration. Countries mitigated to the extent possible ex-ante. need to have important and equal incentives to participate. It is therefore important to assess • The degree of symmetry of risks whether countries in the region face similar risk between countries is important to challenges, risk transmission between countries, understand. While there may be cases and are affected by symmetric or asymmetric where multinational instruments are shocks. The underlying assumptions for these optimal in times of symmetric risk events criteria are as follows: (for example, regional EWS), symmetric risk impacts are likely to put constraints • Facing similar risk challenges in terms on ex-post risk financing mechanisms. of scope and peril could incentivize countries to collaborate around risk Other factors could impede regional management. Risks see no borders, and collaboration among West Africa countries. effective risk management increasingly These include the absence of an integration requires multicountry partnerships culture and a development culture generally around Early Warning Systems (EWS), in the countries of the region; lack of trust 31 livestock diseases surveillance and between member states; the difference in risks vaccination systems, sovereign risk and their intensity; differences in ideology, insurance, and so on. Incentives for language, and approach; political instability; countries to collaborate around and and the sovereignty of states. These factors invest in common risk management make collaboration between countries difficult instruments that are mutually beneficial (USAID 2020). Taking as an example the fact increase if they face the same risk perils that states are sovereign and each has its (especially when the instrument targets own approach, they are not bound to respect a specific peril) and a similar frequency the decisions made at the regional level. This and impact of the risks. has been the case with Nigeria, which has closed its borders and reduced imports of • Risk transmission across countries certain foodstuffs, even though ECOWAS has is important to understand to advocated the free movement of goods and better manage risks and develop services in the zone to mitigate certain risks to joint policy approaches. Examples food security and the market. Even if Nigeria of risk transmission are the ways that had its reasons for doing so, this decision has production risks in one country can had negative repercussions on the food trade affect food prices in another or how of neighboring countries such as Benin (Golub domestic policy decisions or local et al. 2019). conflicts can affect international trade Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens FIGURE 2-3 Agriculture and Food Systems Risk Indicators Assessed in this Report Regional agricultural production risks Food crop Regional production price losses at integration regional level Regional Risks Management Regional risks Identi cation transmission of regional food on prices production risk hotspots Cost of 32 ex-post humanitarian aid Source: Courtesy by Wageningen University & Research 2020 Figure 2.3 provides an overview of the as risk impacts on food production and impacts agriculture and food systems risk indicators that on prices from shocks; and (2) may have scope have been assessed for the regional agrifood for regional collaboration. Omitted from this systems risk assessment in this report. assessment is, for example, the assessment of risks It should be noted that the assessment of the of a nutritiously adequate food basket for which food system in this report is not comprehensive information across the region was not possible to but is an attempt to assess factors that (1) have collect within the scope of this report. direct impacts on food security in terms of availability and access of food commodities, such 2.2 A REGIONAL AGRIFOOD SYSTEMS RISK MANAGEMENT APPROACH R egional-level agrifood system risks not substitute for, country-level agrifood management should benefit the system risk management. The discussion participating countries beyond the below outlines the layered approach to ARM at returns that each country can achieve from the country level that is necessary to effectively investing in risk management individually. manage agricultural risks. The role of the In many areas, regional collaboration provides regional risk instrument will be to collaborate benefits for countries that exceed those of the around the solutions identified in the specific sum of efforts of individual countries. Examples ARM strategy. Such instruments can include, are trade and research collaboration, currency for example, risk transferring instruments at unions, and agricultural support schemes. The farm level, or for sovereign risk financing, they level of success depends on mutual interest can include cross-border agricultural diseases and adherence to shared principles and management systems, weather information objectives, and this also applies to regional risk systems, research programs for risk resilient management initiatives. While many initiatives agricultural inputs, regional trade agreements have been attempted, many have also failed. to mitigate supply shocks and price volatility, This section will disentangle what components and so on. need to be in place for regional collaboration 33 around risks to add value and be sustainable. A layered approach can be used to manage risks. A well-designed risk management A regional agrifood system risks strategy—whether for a public or a private actor management approach would build on, but FIGURE 2-4 Agricultural Risk Layering Source: Giertz et. al. 2018 Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens in the sector—usually involves a combination context of risk layering, stakeholders may of risk-mitigating, risk-transfer, and risk- decide to retain or transfer risks depending on coping instruments. While risk management their financial capacity and appetite for risk. instruments provide solutions to manage risks’ Insurance is best suited to infrequent but severe impacts either before (ex-ante) or after a risk events, where fast payment of a claim can be event (ex-post), all instruments that are part of made (the risk transfer layer). Insurers rely on the strategy should be set up before risks occur. reinsurance companies to boost their effective The most appropriate type of risk management capital by entering reinsurance contracts. instrument to address a specific risk or set of risks can be determined with the help of a risk- Risk-coping mechanisms enable those layering approach (figure 2.4). The frequency, suffering losses caused by a risk event impacts, and priority of risks assessed through to manage the impact of these losses. the ASRA will guide where in the risk-layering Coping mechanisms can be both public and figure the relevant risks fall and how to best private, depending on the risk management manage them. instruments available in a country and the resources available to the actors in the sector. Risk mitigation comprises the first layer. For Especially in the case of smaller risk impacts, risks that are frequent but with relatively low most private sector actors cope by themselves impacts, the preferred approach would be to by using savings, selling assets like livestock eliminate or reduce the frequency of negative or equipment, borrowing from relatives or events or to reduce the severity of the losses neighbors, finding income elsewhere, or resulting from these events. Examples of risk- receiving remittances from family members 34 mitigating instruments are irrigation, water- abroad. For larger risk events, governments draining infrastructure, crop diversification, use often provide support to those involved of fast-maturing varieties, extension services, through food aid, social safety nets and cash and livestock vaccination. It is important to transfers, replanting material for farmers, or know the value of the losses to determine livestock compensation for mass slaughter of whether the benefit of the investment exceeds animals during livestock disease outbreaks. its cost.6 Ideally, risk management should mitigate Risk-transfer instruments transfer the the impacts of the negative event to risks to a willing third party at a cost. the extent possible—an approach that Typical examples in the agriculture sector are also often pays off in normal years. Any insurance contracts or capital market solutions risk management strategy, whether private for managing production risk and hedging or public, should attempt to minimize the instruments for transferring price risk. In the impacts of risks because even with ex-post risk 6 While much of the cost for risk mitigation is borne by the private sector, governments have an important role in providing an enabling en- vironment and the right institutional infrastructure for farmers and other agriculture sector actors to operate within. More direct public sector support includes providing public goods for risk management, including research and development (R&D) of risk-mitigating varieties, breeds, and practices; livestock vaccination and disease management programs, livestock registers, and traceability systems; EWS; large public irrigation infrastructure; agricultural extension services, agribusiness development support, and education programs; and environmental services such as soil management and afforestation. Where market failures exist, the public sector may also support investments of private goods—for example, by availing credit or subsidizing private, risk-mitigating investment. management mechanisms in place such as and to environmental management. In the long transfer and coping, once a risk has impacted term, these investments can enable farmers the sector, investments in production, to become more adaptive to future impacts processing, or other parts of the supply chain of climate change. Similarly, many measures are lost. In many cases, investments in risk- intended to mitigate market risks are related mitigating measures have a range of positive to improving transparency and increasing the impacts beyond merely mitigating against availability of information related to prices, the risk of the negative impact of an event. production, quality, and trade, which may have At production level, these investments may a positive impact on the business climate apart promote general improvements to productivity from playing a role in the mitigation of risk. 2.3 REGIONAL RISK MANAGEMENT INSTRUMENTS: WHAT NEEDS TO BE iN PLACE? D epending on the risks in the region, management instrument intends to manage. there are multiple instruments that Any instrument that is intended to respond can be suitable for collaboration to food insecurity needs to mitigate impacts across countries; nevertheless, a layered on both availability and access to the extent approach still needs to be applied. possible at national levels. An example of the 35 Certain risk management instruments can regional risk management layering approach advantageously be established at the regional proposed for FSRP is provided in figure 2.5. level rather than in individual countries. Perhaps Low-impact risks are managed at local levels the most frequently considered are risk pooling through risk impact mitigation or risk reduction, instruments such as sovereign risk insurance. with complementary risk mitigation measures However, regional research exchange, mutually at the national and regional levels. Medium- recognized sanitary and phytosanitary (SPS) impact risks are managed with an added layer and food safety standards, and regional of coping mechanisms at the national level; information and EWS are other examples of high-impact, rarer risks are further prepared important regional infrastructure for ARM. for with regional financial instruments or food reserves. For regional risk management instruments to work properly, institutional governance and financing and appropriate domestic policies in participating countries need to be in place. The exact components that should be in place depend on what the regional risk Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens FIGURE 2-5 FSRP Regional Risk Management Layering Approach Risk Management strategy * Sovereign risk transfer * Public and private insurance scheme for farmers and others Transfer in the sector * International insurance and reinsurance solutions * Price hedging instruments * Community and National * Regional food reserve (third level food reseves ( rst and line of defense) second line of defense) * Information systems to trigger Coping * Scalable social safety nets response and payouts * Response SOP using regional * Regional risk collaboration information Systems * CSA, improved productive * Early warming systems measures * Other information systems Mitigation * Extension services * Regional pest & disease warning systems * Landscape management * Irrigation * Regional R&D * Post-harvest * R&D in resilient approaches * Trade facilitation Infrastructure & storage * Pest and disease management * Market integration IPC Phase 1 IPC Phase 2 “stressed” IPC Phase 3 “crisis” Severity of shocks Source: World Bank 36 3 FOOD, TRADE, crops such as cassava, yam, maize, and rice in the region. • Food crop production generally has AND KEY increased over the past decades regionally but yields for most crops are relatively low and PRODUCTION have remained largely constant. Production increases have been driven by expanding land use. RISKS • Many important livelihood activities, such as pastoral livestock systems, stretch across country borders. • About 20 percent of the total food consumed This chapter provides insights into West in West Africa comes from imports mainly from Africa’s food crop production in terms of value, outside the region. hectares planted, contribution to the diet in • Intraregional trade in West Africa is relatively the region, different agroecological patterns low—US$2.2 billion in export and US$2.3 billion in the region, and existing LVZ. It also provides in imports in 2019—mainly due to inadequate a brief overview of intraregional food trade internal transport infrastructure (roads and rail and related challenges. Finally, this chapter networks) and to road harassment, leading to discusses some of the main production risks higher transport and transaction costs and in the region, namely droughts, floods, locust, weak competitiveness of regional products. and conflict. • The West African region is among the world’s most vulnerable to climate change, and impacts are already affecting food production, 37 Key points in this leading to high volatility in annual agriculture value-added growth. chapter: • Extreme heat, water scarcity, and drought happen more often in the Sahelian countries, while flooding events are more common • West Africa’s agriculture sector currently in coastal countries. Flood events are more accounts for 30.5 percent of its economy; it frequent, more expensive, and more localized is the largest source of income and provides than droughts, which can span huge areas. The livelihoods for 70–80 percent of the population. Gambia, Mali, and Togo are high-risk countries • Agroclimatic zones stretch across the region for both droughts and floods, while Côte and largely determine production patterns, d’Ivoire, Liberia, Chad, and Ghana are low-risk rather than country borders. Sorghum and countries. millet are mainly produced in the arid areas, • Other important food systems risks in West while cassava and yam are produced in the Africa are locust and conflict. Locust invasions, coastal areas. also linked to climate change, or only a small • Climate patterns are largely defined by the swarm of locusts covering 1 km2 can destroy West African Monsoon, so major crop failures the equivalent of what 35,000 people consume usually happen across entire climatic zones, in a day. Conflict-related risks can affect impacting all crops produced in that climatic production, transport, processing marketing, zone. or distribution of crops with rural conflict • The main crops consumed, which are vital waves causing food security ripples in urban for food security, are cassava, yam, rice, maize, areas as prices increase due to scarcity. Burkina millet, sorghum, plantain, cowpeas, and fonio. Faso, Côte d’Ivoire, Mali, and Nigeria are the Nigeria produces from 35 to 65 percent of most impacted by conflict. Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens 3.1 FOOD CROPS AND LIVELIHOODS W est African countries are largely Livelihood means and production include agriculture-based economies and millet, rice, other cereals, tubers, cash crops, agriculture is the largest source animal husbandry, other rural activities, of livelihood and income for most of the mixed-income generating activities, and other population. The region’s agriculture sector economic activities for the growing urban has been consistently growing and currently and periurban populations. While production accounts for 30.5 percent of West Africa’s grown and consumed varies by LVZ, these economy. It is also the largest source of zones can be grouped to reflect common income and livelihoods for 70–80 percent of features in West Africa. the population, mainly women, who are on the frontline (Toulmin and Guèye 2016). Despite There are six common livelihood profiles in the sector’s seemingly impressive growth in West Africa that cut across country borders, recent decades, agricultural production has showing that risks and coping mechanisms not reached its full potential for meeting the are likely shared across the region. The FEWS growing population’s food needs. West African NET database has documented and defined countries have been importing around 20 LVZ7 for select West African countries based percent of total food consumed to meet food on household profiles, income-generating demands, and several continue to receive activities, food sources, and hazards. The six food aid (FAO 2017). LVZ in West Africa are seen in map 3.1: pastoral, agropastoral, cereals production, mixed 38 West African countries can be divided into production, irrigated production, and other. two agroecological tiers: Sahelian and The six grouped LVZ clearly overlap national coastal countries. The Sahelian countries borders. of Burkina Faso, Chad, Mali, and Niger have an arid and semiarid climate. The rest are Staple crops comprise the bulk of West coastal countries, and most have a humid African diets. The main crops consumed, that or subhumid climate. This difference is also are vital for food security, are cassava, yam, rice, reflected in the respective main food crops; maize, millet, sorghum, plantain, cowpeas, and while millet is the predominant food crop in fonio. The impacts of production shocks on the Sahel, more water-intensive crops, such as these crops can have devastating consequences rice and cassava, play a more important role in for the local population. West Africa imports coastal countries. about 20 percent of its total food, however. Total crop area with main food crops in West LVZ in West Africa are numerous and varied. Africa amounts to 75.4 million hectares and the Populations in West Africa have largely adapted gross value of production (GVP) for main crops their livelihoods to the various agroecological is US$104.6 billion. Yams and cassava have the zones, from the coast to the desert latitudes. greatest total value, accounting for 40 percent 7 FEWS NET LVZ, https://fews.net/livelihoods. MAP 3-1 West Africa’s Main LVZ are Transnational8 Note: FEWSNet LVZ are not available for The Gambia, Guinea-Bissau, Ivory Coast, Ghana and Benin. Nigeria. Guinea, Sierra Leone, Liberia are still to be grouped and mapped. 1 https://fews.net/livelihoods Source: Original data based on FEWS NET classification 39 BOX 3-1 MAPS Throughout this chapter and chapter 4, maps are used to visualize the regional aspects of food crop production and risks. Many of them have been produced specifically for this report based on the International Food Policy Research Institute Spatial Production Allocation Model (IFPRI SPAM) database. IFPRI analysis is based on GIS and other work to estimate production zones at a ‘pixel level’ by devel- oping SPAM. For this model, production areas are extrapolated from coarser administrative production figures and remote sensing data. Based on these modeling and output estimations from IFPRI SPAM, production areas for some of West Africa’s main staple crops can be mapped. Two variables have been mapped to reflect the drivers of areas’ importance to food safety: harvested area (hectares) and production (kilograms per hectare). The crops mapped for this study are maize, sorghum, millet, yam, plantains, and cassava (more crops are available from SPAM). Overall, the dark pixels identify areas with greater acreage and higher productivity. Estimates are for all types of technologies used. While in-field or in-country verifications are undertaken, the SPAM output provides a model and maps that illustrate these assessments. Source: World Bank 8 Note: LVZ in West Africa. FEWS NET LVZ are not available for Benin, Côte d’Ivoire, The Gambia, Ghana, Guinea-Bissau, and Nigeria. Guinea, Liberia, and Sierra Leone LVZ are still to be grouped and mapped. Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens and 27 percent respectively, of the total value indicating their importance as subsistence of the region’s main food crops, followed by crops, although their importance in terms of rice (12.8 percent). Grain crops such as maize, value is relatively low (figure 3.1 and figure 3.2). sorghum, and millet occupy large land areas, FIGURE 3-1 Crop-Wise Distribution of the Exposures for Main Food Crops in West Africa, Per Value Yams 40% Cassava 27% VALUE Rice 13% Fonio 0% Cow peas 1% Plantains Maize 3% Millet 7% 3% Sorghum 40 6% Source: Original data based on FAOSTAT FIGURE 3-2 Crop-Wise Distribution of the Exposures for Main Food Crops in West Africa, Per Hectare Cow peas Sweet potatoes 9% 0% Millet Plantains 17% Fonio 1% 1% Yams 11% Sorghum 20% HECTARE Cassava 12% Rice 12% Maize 17% Source: Original data based on FAOSTAT The ECOWAS region has immense potential, The agropastoralist food system is creating a strong production base for a range characterized by the households’ reliance of crops and encouraging complementarity on livestock keeping with varying degrees between major production areas. More of mobility and transhumance. In total, West than anywhere else, West Africa is home to a Africa is home to between 17 and 25 million diversity of ecosystems, from the humid coastal agropastoralists with strong demographic zone to the Sahel’s dry and arid northern areas, growth rates. Seventeen percent of the Sahelian and from the desert to the central Sudanian countries’ population can be considered part and semihumid zones. These ecosystems are of the agropastoralist community (FAO 2018). favorable to producing a large variety of crops Animal products account for 12–19 percent of and to livestock farming and have an abundance the GDP in the Sahel countries and less than of natural resources —including land suitable 6 percent in coastal countries. Intraregional for farming and surface and subterranean trade of livestock is estimated at nearly US$400 water resources. The harvested areas for all the million per year for the 2013–15 period, which is major food crops span large areas with similar six times higher than for cereals (Tondel 2019). climates and across borders. Production of each crop is concentrated in specific areas, at Some countries have only one or two major both the national and regional levels, and each production zones or crops (for example, climatic zone has specialized production in Niger), while others have more diversified both acreage cultivated and productivity. The production across their geography (for Sahelian zone produces mostly cereals—millet, example, Nigeria) and in crop variety (for sorghum, and rice under irrigation. The coastal example, Senegal). Climate patterns are latitudes and inland produce mostly tubers, largely defined by the West African Monsoon, 41 rice, and maize. so major crop failures usually happen across entire climatic zones, impacting all crops The grains and legume-based food system produced in that climatic zone. Less productive is characterized by both highly market- areas (light color on productivity maps) occupy oriented and subsistence-based farming large production areas, meaning that the local households and is the typical agricultural population uses lots of land and may thus rely production system in the Sudano-Sahelian on the displayed crop (which could be critical zone. An estimated 32 million people are at for the local population). the heart of this grains- and legumes-based food system and represent about 15 percent of West Africa’s total population, excluding Nigeria. Trends over the last two decades show an increased production capacity of grains and pulses responding to the rising demand, although closing yield gaps remains a challenge. Since 2012, the security situation in Sahelian countries has worsened in parts of Burkina Faso, Mali, and Niger, with spillover effects in other areas. For instance, farmers in central Mali cannot reliably cultivate their fields because of recurrent attacks and massacres. Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens 3.2 MARKETS AND INTRAREGIONAL TRADE S everal policy frameworks promoting ECOWAS is now one of the regional intraregional trade and regional economic communities in Africa with the integration have been implemented highest intraregional trade shares for both in West Africa and more widely in the total trade and agricultural trade (Bouët et continent. The 2014 Heads of State Malabo al. 2019). Unfortunately, ECOWAS intraregional Declaration on Accelerated Agricultural Growth trade is largely underreported because trade and Transformation for Shared Prosperity and data are fragmented and of uncertain quality. Improved Livelihoods committed to triple intra- In 2008, the AfDB estimated the value of African trade in food commodities and services intraregional trade at US$8.6 billion. In 2010, the by 2025. The declaration’s targets are part of the volume of intraregional trade of all commodities Comprehensive Africa Agriculture Development was estimated at 16 percent of the total value Program (CAADP) and are monitored by the of commercial trade of the region. Yet intra- African Union (AU) biennial review mechanism. ECOWAS imports decreased from 13.2 percent The Economic Community of West African in 2000 to 10.4 percent in 2009, with an average Agricultural Policy (ECOWAP) operationalizes of 12 percent over the decade. This suggests CAADP within the ECOWAS region. Alongside that member countries import more from these agriculture development frameworks, the rest of the world than from neighboring market integration is promoted through countries. For example, in 2017, 35 percent of both regional and continental level trade the rice consumed in ECOWAS was imported agreements. In 1979, the ECOWAS Trade from Thailand and other Asian countries. 42 Liberalization Scheme (ETLS) was established and in 2019 the African Continental Free Trade West Africa’s low intraregional trade is Area (AfCFTA) has entered into force. The ETLS mainly due to inadequate internal transport aims to remove barriers to intraregional trade infrastructure (roads and rail networks) and to guarantee free movement of basic and to road harassment, leading to higher food staples within ECOWAS. However, ETLS transport and transaction costs and weak implementation is incomplete and many competitiveness of regional products. Terms obstacles to trade persist. Considering AfCFTA, of trade are biased, favoring food imports from the agreement aims to reduce tariffs between outside ECOWAS to the coastal cities because member countries and covers policy areas of low shipping and transaction costs. Sourcing such as trade facilitation and services, as well as food from hinterlands imposes high domestic regulatory measures such as sanitary standards transport and transaction costs. Between and technical barriers to trade (Maliszewska 1996–2000 and 2006–10, imports of most basic and Ruta 2020). By giving greater attention food commodities from outside the ECOWAS and political oversight to trade, AfCFTA offers region grew at an accelerating rate, including an opportunity to improve trade facilitation rice and wheat, palm oil, dairy products, poultry in West Africa at borders and along corridors and other meats, tomato (paste and peeled), between countries. Trade, including the carrots and turnips, potatoes, green onions, agricultural sector, will grow substantially as the and various processed vegetables. Among volume of total exports is expected to increase fruits, accelerating net imports include apples by 2035. Among countries, the largest increase (16 percent), grapes (14 percent), oranges (14 in the value of exports to regional partners is percent), dates (23 percent), and all kinds of expected to be in Nigeria and Côte d’Ivoire. fruit juices (Hollinger et al. 2015). The top five intraregional food importing Ghana, and Niger (Elbehri et al. 2013). Map 3.2 countries are Nigeria, Burkina Faso, Ghana, below presents the intraregional trade flows of Mali, and Niger, which absorb 77 percent cereals, root crops, livestock, fish, fruit, pepper, of intraregional food imports. The largest and colas. This map shows that the main cereals livestock exporters are Burkina Faso, Mali, exporters are Burkina Faso, Mali, and Nigeria. and Niger, while Nigeria is the main exporter Nigeria exports its cereals mainly to Niger. of cereals and cassava. The largest livestock ECOWAS and CILSS aim to expand the number importers are Nigeria, Ghana, Côte d’Ivoire, of monitored corridors to improve knowledge and Senegal; the largest maize importers of trade levels and promote better circulation are Niger, Senegal, and Burkina Faso; and the of goods and services. largest sorghum and millet importers are Benin, MAP 3-2 Intraregional Trade Flows, 2017 43 Source: CILSS 2017 The value of intraregional food export and rice, millet, and sorghum; (2) a central area import was estimated at US$2.2 billion comprising Côte d’Ivoire, Ghana, Togo, Mali, and US$2.3 billion, respectively, in 2019 and Burkina Faso, trading mainly in maize; (UNCTAD 2020). Soule et al. (2010) identified (3) an eastern area comprising Nigeria and five principal market areas: (1) a western area its neighbors Benin, Niger, and Chad, which centered on Senegal, trading mainly in local accounts for 60 percent of total intraregional Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens flows of millet, sorghum, maize, cowpeas, and and importing countries have a market share reexported rice (from Benin to Nigeria); (4) the of around 70 percent of the total intraregional Ibadan-Lagos-Accra conurbation, comprising trade. The most imported and exported food agglomerations in Nigeria, Benin, Togo, and products are tobacco, vegetables, edible food Ghana with flows of maize and reexports of rice; preparations, and live animals. The top five and (5) the Sahelian belt spanning Mauritania, intraregional food exporters are Côte d’Ivoire, Mali, Burkina Faso, Niger, and Nigeria (millet Senegal, Niger, Nigeria, and Ghana, with a and sorghum). The top five food exporting market share of 76 percent. 3.3 KEY PRODUCTION RISKS T he West African region is among the that many African countries have committed to world’s most vulnerable to climate under CAADP.9 However, due to the numerous change and impacts are already initiatives aimed at promoting climate resilient affecting food production. Since the 1990s, techniques to improve national production rains have become more erratic in terms of and the development of value chains in many quantity, timing, and geographical distribution, West African countries such as Senegal, a less making droughts and poor harvests more significant variation in the growth rate has frequent. These changes are having an been observed over the last decade (Blein, R. enormous impact on the region’s farmers et al 2008). and pastoralists given their dependence on 44 rainfall for their livelihoods. Weather events For water scarcity and drought, Burkina Faso, accentuate their vulnerability and affect their northern Chad, The Gambia, eastern Mali, Niger, resilience capabilities since they lack sufficient and northern Nigeria and are the riskiest areas time to recover from one crisis before the next in the region. TH West Africa subset data can crisis hits. Natural disasters are increasing in be used to map, at the Admin 2 level, exposure number and frequency and affecting most for water scarcity and extreme heat, which West African countries. Droughts and floods are major determinants of annual production have become recurrent and more acute over outcomes (map 3.3 and map 3.4). In addition recent years, with severe impacts on food and to the above-mentioned countries, northern water security. This has led to high volatility Ghana, Benin, and areas in Nigeria and Senegal in annual agriculture value-added growth also have water scarcity and drought risks in (figure 3.3), which shows the most volatile high production zones. Some countries that countries. These large and frequent volatilities overall appear to be spared still have small high- are especially problematic for countries where exposure pockets, such as Côte d’Ivoire, Ghana, the agriculture sector is a large contributor to and Benin, which may have to be considered. the overall economy. When agriculture value- Pastoral areas at higher latitudes may face added growth dips into negative territory as further water scarcity. The return periods set by frequently as it does for the countries in the TH for the Low and Medium classification are, figure, this has important consequences for respectively, 5 and 20 years for extreme heat overall growth. It is also difficult for countries and 5 and 50 years for water scarcity. to achieve the objective of 6 percent growth 9 Volatility in agriculture value added reflects all types of risks to the sector, including production, marketing, and enabling environment risks (see chapter 2). FIGURE 3-3 Annual Agriculture Value-Added Growth is Highly Volatile in the Region (% growth) Source: WDI Database Note: Because the data in this figure is at aggregate level, it may mask some of the volatilities in the sector that occur at subnational level. Chapter 5 attempts to address this by assessing risk hotspots in the region based on subnational data. 45 MAP 3-3 Water Scarcity and Drought HIG MED LOW VLO Source: GFDRR TH Data Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens BOX 3-2 The Think Hazard Methodology One of the only hazard information source readily available for the whole region is the World Bank’s DRR resource ThinkHazard! (TH). The TH mod- el includes perils such as river and urban floods, extreme heat, water scarcity, and cyclones. The models are forward looking and should thus be suited to support policy and project planning. The platform has access to several hazard mod- els and a methodology to select the best fit for a specific peril and location. The TH model outputs are at Admin 2 level (a cercle in Mali or a Local Government Area (LGA) in Nigeria). The detailed TH methodology and data processing is laid out by GFDRR and can be summarized as follows: (1) intensity thresholds for exposure are predefined for each peril; (2) the output of hazard frequency models at Admin 2 level are compared to these thresholds; (3) each Admin 2 level is classified as low, medium, or high risk based on the compar- ison of frequency hazard; and (4) exposure levels are then aggregated at Admin 1 level (regions or 46 states). The TH methodology has limitations since global models may not capture local particularities adequate- ly. The TH outlook differs from an agriculture risk assessment. In particular, the drought and flood return periods and threshold are less meaningful for agriculture risk assessment than for DRR. Also, water scar- city is not the same as water requirements for agriculture, and the focus on infrastructure and project impacts does not match what is needed for agriculture and food production. The TH process excludes the preexisting population’s vulnerability and time frames that are relevant for agriculture. As such, the TH outputs provide a coarse estimation, with the objective of highlighting the relative exposure of each subpart of the West Africa region. Hazard specific classification methods: water scarcity hazard levels This hazard is ranked using a water stress index, which reflects water availability per person per year. It is a measure of water stress based on hydrologic drought and water use. Water scarcity is the only TH haz- ard that uses an “inverse damage intensity threshold.” All other hazards are ranked based on exceeding the intensity value. Water scarcity becomes more severe as water availability decreases, so lower water availability per capita per year results in higher risk. The hazard is classified as high when water availability is <500 m3/capita/year at the 5-year return period. The hazard is classified as medium when water avail- ability is <1000 m3/capita/year at the 50-year return period. Finally, the hazard is classified as low when water availability is <1700 m3/capita/year at the 100-year return period. Source: GFDRR, ThinkHazard!, https ://thinkhazard.org/en/. WATE and • • MAP 3-4 Extreme Heat • EXTREME HEAT • Th are for ret cla Source: GFDRR TH Data 47 Coastal countries and communities may be Overall, West Africa is more exposed to more exposed to flooding as precipitation high-impact flood events than to high- events produce more rain and can be extreme; impact drought events, but low-level however, Mali, southern Niger, southern droughts significantly affect the region. Chad, Senegal, and northern Nigeria are also When assessing the proportion of a country at exposed, according to the DRR platform data. a specific risk level (low, medium, high, or very TH West Africa subset data can be used to map high), some countries are at greater risk than the exposure for river floods and urban floods others, as seen below in table 3.1. More countries as seen in map 3.5 below. Rainfed production are exposed to floods than to drought (12 versus of maize, rice, groundnut, and tubers may be 8 countries). Although flood events tend to be the most exposed to river floods and excess more localized than droughts, heavy rains and rainfall. In addition, irrigated areas fall under floods do occasionally simultaneously affect high-risk areas, which could impact rice and multiple countries in the region.10 Nevertheless, other irrigated crops, particularly in Chad, The from the point of view of the severity and scale, Gambia, Mali, Niger, and Senegal. Finally, the droughts are more pervasive than floods. This is West African population is denser and more particularly true in the Sahel; except for Ghana, numerous in coastal areas; river and urban all coastal countries have limited exposure to floods can impact economic activities in these drought. For countries with only a portion of areas, in turn impacting the ability to earn the territory exposed, analysis and comparison money and purchase food (map 3.5). with production and productivity levels would be required to assess vulnerability further. 10 UN Office for the Coordination of Humanitarian Affairs, “West and Central Africa: Flooding Situation As of 6 November 2020,” https://reliefweb. int/report/niger/west-and-central-africa-flooding-situation-6-november-2020 Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens MAP 3-5 River Floods (Admin 2 level) HIG MED LOW VLO Source: GFDRR TH Data 48 MAP 3-6 River Floods with Added Layer Displaying 2016 Population Estimates HIG MED LOW VLO Source: GFDRR TH Data TABLE 3-1 National Proportion of Admin 2 Level, by Risk Level (Per TH Methodology) FLOODS WATER SCARCITY - DROUGHT Level of Very Level of Very Low Medium High Low Medium High exposure: high exposure: high Benin 14,3% 32,5% 15,6% 33,8% Benin 70,1% 11,7% 10,4% 7,8% Burkina Burkina 0,0% 28,9% 44,4% 26,7% 0,0% 2,2% 0,0% 97,8% Faso Faso Côte Côte 0,0% 0,0% 54,5% 45,5% 84,8% 6,1% 6,1% 3,0% d'Ivoire d'Ivoire Cape Cape 0,0% 0,0% 0,0% 0,0% 0,0% 0,0% 0,0% 0,0% Verde Verde Chad 11,8% 23,3% 15,6% 47,8% Chad 7,2% 64,3% 7,2% 21,3% Ghana 11,6% 30,1% 24,1% 32,9% Ghana 0,9% 38,9% 53,2% 6,9% The Gam- The Gam- 18,9% 21,6% 0,0% 59,5% 0,0% 0,0% 0,0% 100,0% bia bia Guinea 10,5% 26,3% 34,2% 26,3% Guinea 89,5% 0,0% 10,5% 0,0% Guin- Guin- 23,1% 7,7% 10,3% 59,0% 94,9% 0,0% 5,1% 0,0% ea-Bissau ea-Bissau Liberia 14,0% 14,0% 19,9% 49,3% Liberia 100,0% 0,0% 0,0% 0,0% Mali 2,0% 14,0% 8,0% 76,0% Mali 4,0% 2,0% 36,0% 58,0% 49 Mauritania 11,4% 40,9% 25,0% 22,7% Mauritania 0,0% 2,3% 79,5% 18,2% Niger 5,3% 21,1% 23,7% 50,0% Niger 0,0% 0,0% 5,3% 94,7% Nigeria 19,9% 25,5% 13,5% 38,5% Nigeria 38,1% 11,5% 23,7% 26,7% Senegal 4,7% 30,2% 27,9% 37,2% Senegal 0,0% 0,0% 81,4% 18,6% Sierra Leone 14,3% 14,3% 0,0% 71,4% Sierra Leone 100,0% 0,0% 0,0% 0,0% Togo 16,7% 30,0% 23,3% 26,7% Togo 36,7% 56,7% 0,0% 6,7% West Africa 14,5% 24,1% 17,5% 40,9% West Africa 32,2% 21,5% 21,7% 23,5% Note: some admin2 are missing data thus total does not add up to 100% More than 50% of Admin2 at Level 4/Very High More than 90% of Admin2 at Level 4(Very High) More than 50% at level 3 (High) or 4/(Very High) More than 90%(+) at Level 3 (High) or 4 (Very High) More than 60%(+) at Level 3 (High) or 4 (Very High) Source: Original table based on ThinkHazard For exposure to both droughts and floods, need and incentives for risk collaboration. The Gambia, Mali, and Togo are high risk While the first assessment of drought and flood countries, while Chad, Côte d’Ivoire, Ghana, exposure shows that there are large differences and Liberia are on the low-risk end of the in the level of combined exposure between spectrum (figure 3.4). Understanding relative countries, most countries in the region are exposure between countries can indicate the highly exposed to either droughts or floods. Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens FIGURE 3-4 Countries’ Relative Exposure to Water Scarcity and River Floods Relative Level of Risks (5) Illustration of Relative exposure to Water Scarcity and River Floods THINK HAZARD FLOODS (based on table from previous slide) EXPOSURE Sierra Mali Gambia, Leone Guinea-Bissau The Benin Togo DROUGHT Liberia EXPOSURE Ghana Senegal Burkina Côte Faso d’Ivoire Chad Niger Nigeria Mauritania Source: Original table based on ThinkHazard Regarding crop level, the main weather-related risks are summarized in table 3.2: 50 TABLE 3-2 Main Weather-Related Risks Per Crop in the Region Main Risk is indicated as follows: Drought/heat XS Moinsture IDEAL CONDI- CROP CLIMATE RISKS DETAILS TIONS RF: 500–700 mm Temp: 21–30 Dry spells, waterlogging, ex- 10+ day dry spell, XS water throughout cycle, tem- Maize Cycle: 65–120+ treme temperatures perature >35 at flowering (which is the critical stage) days RF: 400–700 mm Dry spells and high tempera- Waterlogging / immersion destroys. XS rain at Millet Temp: 23–35 Cy- tures (especially seedlings and flowering. Requires evenly distributed RF. Also used cle 60–180 days at flowering), XS water as fodder. Pest and diseases. High RF: 800–1200 mm Drought-resistant crop. 450–650 mm RF areas are temperatures (>40), drought Sorghum Temp: 20–35 Cy- adequate. Second phase (reproduction) is the most (especially terminal) or water cle: 90–120 days critical as it accounts for 70 percent of yield. logging for > 15 days RF: 1400– Hot and humid climate grass. Requires bright 1900 mm Drought; High temperatures days (300 h over the last 45 days). Annual demand Rice Temp: 25–32 (>35); wind and storms. growth: 6 percent; 20 million farmers in West and Cycle: 100–180 Central Africa days IDEAL CONDI- CROP CLIMATE RISKS DETAILS TIONS RF: 500–1500 mm Hardy perennial tuber that can survive dry seasons Cassava Temp: 20–30 No tolerance to waterlogging. of up to eight months. RF can be as low as 500 and Cycle: perennial as high as 5000 mm. Susceptible to pest and diseas- RF: 400–1000 mm Drought-tolerant plant adapted for harsh condi- es, and frost. Mostly drought (dep. on variety) tions. Intercropping with maize, sorghum, millet, or Cowpea sensitive during flowering Temp: 24–36 Cy- cassava. Food and animal feed crop. RF can be as (3–5 dry days max; otherwise, cle: 50–120 days low as 300 mm. impact > 7–15 dry days). Source: Review of 15 crops cultivated in the Sahel, USAID and others 3.4 OTHER RISKS AFFECTING AGRICULTURAL PRODUCTION 3.4.1 Locusts T he frequency of locust invasions in change effects such as warming oceans and 51 West Africa had been under control more frequent tropical storms worsen, locust during the second half of the twentieth invasions could be more common. In addition century, but there has been a resurgence, to climate change, shifts in land use and political as seen in figure 3.5. The impact of a locust instability in countries where monitoring and swarm is extensive: a small swarm covering prevention are required also explain the current 1 km2 can destroy the equivalent of the daily and potential future outbreaks. consumption of 35,000 people in a day. Since the 1980s, there have been 37 recorded events of insect infestation (including locusts) in West Africa (Emergency Events Database [EM-DAT]). The highest number of insect infestations in recent years was observed in 2004, with a total of nine events in a single year.11 Swarm developments are linked to climatic events. The current outbreak is traced back to 2018 when some heavier than usual rains and warmer temperatures occurred. As climate 11 According to the EM-DAT database, the number of insect infestation events in West Africa by year starting from 1980 were as follows: 1985, 2 events; 1986, 8 events; 1987, 9 events; 1988, 8 events; 2004, 9 events; 2009, 1 event. Information on the scale of the impacts of these events has not been found by this study. Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens stible to locust ring the 2nd half ed. rm, i.e. one he daily FIGURE 3-5 Frequency of Locust Invasions current han usal rains frequent ust could be olitical instability uired also Source: Meynard C.N., 2020.Source: Global “On the relative Change role of climateBiology, change andVolume: 26, Issue: management 7, Pages: in the current 3753-3755, desert First locust outbreak published: in East Africa” . Global Change Biology 29 April 2020, DOI: 26,(10.1111/gcb.15137) no. 7 (2020): 3753–55, https://doi.org/10.1111/gcb.15137. 52 3.4.2 Conflict M any parts of West Africa are • 40 percent of violent events take place increasingly suffering from various within 100 kilometers of borders types of insecurity, violence, and • One-half of the conflicts are low protracted conflict from insurgencies, illegal intensity, lingering, and spatially trafficking of drugs and arms, or conflict clustered; the other 50 percent are over access to natural resources. The period spatially clustered and high intensity between 2014 and 2019 (see figure 3.6 below) • Failing political institutions are the has been the most violent on record. From 2011 primary reason for conflict to 2019, violent events in the region soared from • Ethnicity and religion are frequently 581 to 3,617 incidents. Over the same period, instrumentalized. the number of associated fatalities rose from 3,361 to 11,911 (Organisation for Economic Co- Map 3.7 below shows past hotspots for unrest operation and Development/Sahel and West and conflicts. Nigeria, Côte d’Ivoire, Mali, and Africa Club [OECD/SWAC] 2020). According Burkina Faso are the most impacted by conflict, to an analysis conducted by the OECD/SWAC while Mali, Burkina Faso, Benin, and the Lake (2020), conflict and violence in West Africa is Chad area are conflict areas with multiple defined by the following characteristics: border areas that have experienced significant insecurity in the past few years. • Violence is increasingly targeting civilians FIGURE 3-6 Conflict Hotspots Data Points or Events, by Type Source: Armed Conflict Location & Event Data Project (ACLED) MAP 3-7 Density Map of Conflict Hotspot Data 53 Source: ACLED Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens By affecting rural areas and transportation that is likely to worsen with climate change. corridors, conflict impacts both food Increasing demographic pressures on land production and food security in various and deteriorating livestock-to-land ratios ways. Security risks can undermine farmers’ have led to reduced fallow periods, increased ability to produce, impacting otzher livelihoods soil degradation, and losses in forests and and related incomes. As a result of conflict, vegetative cover. This has resulted in reduced crops are often lost as farmers are forced off land productivity and increased competition their fields and the provision of crucial services for resources. Together with high population is interrupted. Insecurity-caused border growth and stagnant productivity, public closures and interrupted supply chains can support for agricultural lifestyles has led severely restrict trade and market access, which farmers to increasingly encroach on traditional may result in rising food prices and thus rising pastureland. In a context where government levels of food insecurity. institutions are frequently hindered by capacity and legitimacy constraints, these trends have Food security outcomes have worsened significantly contributed to the observed rise in significantly in conflict-affected areas violent clashes between farmers and herders. of Burkina Faso, Mali, Niger, Chad, and Accelerating climate change is widely expected northeastern Nigeria. Conflict has been to further increase conflict risks in the mid- and among the most important drivers in the rise long term. Forthcoming work under the FSRF12 in food insecurity observed between 2015 and will explore climate-conflict linkages in greater 2020 (OECD 2020b). Major hotspots in West detail and develop spatially disaggregated Africa for food insecurity include the Lake Chad climate-conflict vulnerability profiles based 54 Basin, which consists of subnational areas in on adaptive capacity parameters. In addition, Cameroon, Chad, Niger, and northern Nigeria; the study will provide an overview of practices and the Central Sahel (Liptako Gourma region) that may lessen conflict risks in the context of overlapping between Burkina Faso, Mali, and development interventions. Niger. In these regions, the presence of armed groups led to large-scale displacement of people, destruction or closure of basic social services, and disruption of productive activities, markets, and trade flows. Frequently cutting across borders, hotspots are concentrated in remote, drought-prone areas. Except for the region around Lake Chad, these areas are characterized by low population density, a lack of government-provided basic services, and weak states. Escalating competition between livelihood groups over natural resource access is fueling the region’s conflict risks—a trend 12 Publication expected in early 2022. Key points in this chapter: 4 • Yam and cassava are the most important crops by total value, accounting for 40 percent FOOD CROPS, and 27 percent, respectively, of the region’s main food crops, followed by rice (12.8 percent). Grain crops such as maize, sorghum, RANGELANDS, and millet occupy extensive production areas in West Africa, showing their importance for subsistence, even though their overall value is AND relatively low. While West Africa is a big cash crop producer, food crops are not necessarily the most economically important. PRODUCTION AND • The region’s 15 countries lose an average of US$4.7 billion annually (or US$1.6 billion if Nigeria is excluded) through production shortfalls of their most important food crops. LOSS ASSESSMENTS • Food crop losses are high already for medium- level production risks: for medium-severity shocks (occurring once every 5 years), losses 55 exceed US$2.1 billion across the 14 countries assessed, and Nigeria loses an additional This chapter provides a detailed description US$5.5 billion; this figure increases to US$2.5 of the geographic scope of production and billion plus US$8.6 billion for 1 in 10-year level the productivity of the most important food risks and to US$2.9 billion plus US$12.2 billion crops—yam, cassava, rice, maize, cowpeas, for 1 in 25-year level risks. millet, sorghum, and plantain—and shows • There are large differences between countries how production zones span country borders in the value of production losses. Nigeria in the region. It also assesses the scope of contributes by far the most to the region’s AAL, agricultural production losses for the most with 66 percent of total value of losses. The rest important food crops in each of the 15 West of the countries that substantially contribute African countries and rangelands in the Sahel to regional losses range between a little over region. To establish indicators for policy and 6 percent for Ghana and Niger to 3.8 percent risk response, this chapter correlates different for Mali. On the other end of the spectrum are indexes with food crop production losses and Guinea-Bissau, Liberia, and Senegal, accounting food insecurity. for less than 1 percent of the regions total AAL. • For the medium level of risks (1 in 5-year), countries such as Chad, The Gambia, Guinea- Bissau, Mali, Niger, Nigeria, Sierra Leone, and Togo see losses around more than 10 percent, while Benin, Côte d’Ivoire, and Ghana see losses around or below 5 percent of their food crop production. Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens • For more extreme events (1 in 25-year • No formal correlation was found for level events), The Gambia, Guinea-Bissau, food insecurity and drought, likely due to Mali, Niger, Nigeria, and Sierra Leone inadequate data quality to reveal policy continue to see very high impacts on triggers—for example, for a regional their food crop production—over 20 food reserve or other safety nets. percent, or 32 percent for Niger—while • Rangeland production losses were the impacts for extreme events are assessed at the regional level (Admin slightly lower for Chad and Togo. Benin, 1) in the Sahel region in West Africa to Côte d’Ivoire, and Ghana experience understand the potential impacts on lower impacts on their own food crop the livestock sector from fodder crop production, even for these 1 in 25-year production risks. The selected countries level events, with losses around 8 percent lose an estimated 3.39 percent of the of total food crop production. rangeland production, on average, every • Across the 15 countries, yam, cassava, year, with 14.4 percent loss for 1 in 100 and rice have the largest aggregate AAL year-events and 18.2 percent for 1 in 250- in terms of value (40, 27, and 13 percent year events. The estimates were based respectively), while maize and sorghum on the Normalized Difference Vegetation each only contribute 7 and 6 percent Index (NDVI). As the region’s livestock to losses and added together, millet, production systems are quite fragile, plantains, cowpeas, and fonio about 7 even slight variations in NDVI may lead to percent. significant stress situations, jeopardizing • The expected AAL for each crop is production or even causing the loss of calculated based on the deviation of stock. 56 Monte Carlo-generated actual GVP from expected GVP, but the tool fails to capture all risk impacts: First, because the food crop risk assessment tool has been designed based on country-level average crop yields, the volatility in crop yields is underestimated. Second, due to the scale of the aggregation of the information, the tool is not picking up eventual regional crop losses. Third, as the crop yields used for the analysis are calculated on a harvested area basis, crop losses due to loss of area are not factored in the analysis. • Analyzing the correlations for each of the phenology stages of the analyzed crops shows that it is not possible to identify a pattern for which crop phenology stage is better to measure the evolution of the index to infer crop yield volatility. This is presumably due to inadequate data quality, but more analysis is needed before an index-based trigger or payout can be established. T o assess the scope of losses, a crop risk a key measure to infer the size of a loss that assessment was performed for 15 West could be reached with a given probability, African countries: Benin, Burkina Faso, which is inversely related to the severity of a Chad, Côte d’Ivoire, The Gambia, Ghana, shock. The crop risk assessment is performed Guinea, Guinea-Bissau, Liberia, Mali, Niger, for the five main food crops in each country Nigeria, Senegal, Sierra Leone, and Togo. based on their importance in terms of The risk assessment analysis for main food area harvested, since this indicates their crops in West Africa assessed two dimensions importance as local food crops. A summary of production risk. The first dimension is the of the methodology is presented in box 4.1, expected AAL, which is an indicator of the with a more detailed description provided in loss that can be expected in any given year for appendix A. The selection of crops for each each crop. The second dimension is the LaR, country is presented in figure 4.1. FIGURE 4-1 Selection of Crops and Countries for the Analysis Cow peas Maize Cassava Rice Maize Maize Millet Maize Cassava Burkina Faso Millet Sierra Leone Liberia Chad Mali Millet Rice Rice Plantains Rice Rice Sorghum Cow peas Maize Sorghum Sorghum Yams Sweet potatoes Yams Sweet potatoes Cow peas Cassava Cassava Maize Yams Maize Maize Sorghum 57 Côte d'Ivoire Cassava Cassava Nigeria Benin Niger Togo Millet Rice Maize Yams Rice Rice Sorghum Yams Sorghum Plantains Sorghum Yams Rice Rice Maize Millet Millet Maize Rice Rice Guinea Bissau Rice Maize Cassava Fonio Sorghum Gambia, The Senegal Guinea Ghana Maize Sorghum Yams Maize Plantains Sorghum Rice Plantains Millet Millet Cassava Cow peas Rice Cassava Maize Source: Original figure based on Area Harvested from FAOSTAT Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens BOX 4-1 Methodology for Food Crop Loss Assessment The methodology used in this analysis is based on a stochastic model based on the probability that the actual yields of main crops fall short of their expected yields. Monte Carlo simulation was used to generate simulated samples of 10,000 hypothetical years of d-trended yields for the crops in the portfolio. The food crop risk assessment for West African countries is performed based on official records from FAOSTAT. The analysis is based on country-level FAOSTAT historic records of crop harvested area, production, and yield for each of the selected crops in each of the selected countries for the period of 1983 to 2018. The model is based on stochastic scenarios of the GVP of the crops and locations that com- Fig B4-1 page:51 prise the portfolio. The stochastic GVP is calculated as the product between the Monte Carlo-generated deviation from the expected crop area (EHA), the Monte Carlo-generated deviation from yield (YPD), the corresponding expected yield for 2018 and the corresponding crop’s average price at the harvest month. Figure B4-1 Description of the Methodology Used for the Assessment 58 EXPECTED YIELD: MAIZE EXPECTED MAIZE YIELD: KG/HA 467 KG/HA 467 EXPECTED AREA: MAIZE EXPECTED MAIZE AREA: HA 1,449,675 HA 1,449,675 EXPECTED MAIZE EXPECTED MAIZE PRICE: PRICE: 390US$/TN 390US$/TN 5 5 Source: World Bank The main outputs of the crop risk assessment model are the expected AAL and the LaR. The ex- pected AAL for each crop is calculated based on the deviation of Monte Carlo-generated actual GVP from expected GVP. If the Monte Carlo-generated actual GVP falls short of the expected GVP, then there is a loss proportional to the size of shortfall. For this model, the expected AAL for a given unit is determined by the average of the Monte Carlo GVP shortfalls in respect to the expected GVP. The LaR or Probable Maximum Loss (PML) is a key measure to infer the potential losses in the portfolio. The LaR is a percentile of the loss distribution, calculated in function of the probability of occurrence of a catastrophic event. For example, the LaR for an exceedance probability “p” of 1 percent (or return period of 1 in 100 years) is the value of the loss distribution that accumulates 99 percent of probability, that is, the 99th percentile. For this model, the LaR for a given unit is determined by the percentile of the Monte Carlo GVP shortfalls in respect to the expected GVP associated with a given probability related to a return period. The param- eters used in this report are the expected AAL and the PML for different return periods, which captures extreme events by precisely assessing the expected loss for different probabilities of occurrence. These two parameters are widely used by risk management practitioners and the insurance industry to assess risks at the portfolio level. The expected AAL provides a notional about the expectancy of losses, while the PML provides an indication of the potential dispersion of the risk measured in term of expectancies of portfolio losses for different probabilities of occurrence. Monte Carlo simulation considers a wide range of possible outcomes from frequent events occurring every two years (1 in 2) to extreme events with a return period of 1 in 100 years. The Monte Carlo simulation differs from the ASRA Loss Methodology, which looks at trends based on past events and excludes events with impacts below a certain threshold. Appendix A provides more details on the methodology used to model food crops shocks. The results of the risk assessment must be interpreted considering the limitations of the tool. The food crop risk assessment tool for West Africa has been designed based on country-level average crop yields, which means that the analysis will have some limitations. The first limitation is that the AAL and PML es- timations are based on the only data that is available. The second limitation is that the country-level crop yield is resulting from an average, thus the volatility in crop yields is underestimated. The third limitation is that, due to the scale of the aggregation of the information, the tool is not picking up eventual regional crop losses. The fourth limitation is that, as the crop yield used for the analysis is calculated on harvested area basis, crop losses due to loss of area are not factored in the analysis. Source: World Bank 59 4.1 IMPACTS OF PRODUCTION SHOCKS IN THE REGION F rom an economic loss perspective, production risks: for shocks of medium severity the 15 countries lose on average (occurring once every 5 years), losses exceed around US$4.7 billion per year US$2.1 billion across the 14 countries assessed, through production shortfalls of their most and Nigeria loses an additional US$5.6 billion; important food crops (table 4.1). This figure this figure increases to US$2.5 + US$8.6 billion is dramatically reduced to US$1.6 billion if for 1 in 10-year level risks and to US$2.9 + Nigeria is excluded from the sample. Food crop US$12.2 billion for 1 in 25-year level risks. losses in the region are high for medium level Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens TABLE 4-1 Country Contributions to Aggregated Expected Food Crop Losses and Most Vulnerable Crops Expect- Expected Exposure Contribution Recurrence PML PML ed AAL AAL(% Main Country GVP (US$, to regional period (% (US$, (US$, GVP) risk crop millions) AAL (%) (years) GVP) millions) millions) levels 5 3.4 117 10 5.3 179 Yam, Benin 3,412.08 67 1.95 1.41 Cassava, 25 7.8 267 Maize 50 9.6 329 5 7.4 136 10 10.4 192 Sorghum, Burkina 1,845.25 77 4.20 1.64 Maize, Faso 25 14.1 259 Millet 50 16.1 297 5 9.7 111 10 13.4 153 Sorghum, Chad 1,137.69 62 5.45 1.32 Millet, 25 17.8 203 Rice 50 20.8 237 5 5.6 413 10 7 518 Yam, Côte 7,379.17 273 3.70 5.79 Cassava, d’Ivoire 60 25 8.7 645 Rice 50 10.1 744 5 12.7 16 10 17.8 23 Millet, Gambia, 126.93 9 7.04 0.19 Maize, The 25 23.1 29 Rice 50 26.7 34 5 2.9 540 10 4.9 922 Cassava, Ghana 18,837.26 313 1.66 6.64 Yam, 25 7.1 1,340 Plantains 50 8.8 1,653 5 7.7 226 10 11.1 328 Rice, Guinea 2,946.97 131 4.45 2.78 Fonio, 25 15.9 467 Maize 50 19.4 572 5 10.4 22 10 15.2 32 Rice, Guinea- 212.52 11 5.23 0.24 Plantains, Bissau 25 20.6 44 Sorghum 50 24.3 52 Expect- Expected Exposure Contribution Recurrence PML PML ed AAL AAL(% Main Country GVP (US$, to regional period (% (US$, (US$, GVP) risk crop millions) AAL (%) (years) GVP) millions) millions) levels 5 8.9 38 10 12.2 52 Rice, Liberia 428 20 4.74 0.43 Cassava, 25 15.9 68 Plantains 50 18.4 79 5 10.9 336 10 16.2 502 Maize, Mali 3,095.80 179 5.78 3.8 Millet, 25 21.6 669 Rice 50 24.9 770 5 16.6 525 10 24.2 765 Cowpeas, Niger 3,163.13 284 8.98 6.04 Millet, 25 32 1,012 Sorghum 50 37.1 1,175 5 9.4 5,549 10 14.5 8,562 Yam, Nigeria 59,210.32 3,124 5.28 66.35 Cassava, 25 20.5 12,163 Rice 61 50 24.6 14,593 5 8.9 75 10 12 101 Rice, Senegal 841.67 44 5.23 0.93 Millet, 25 15.2 128 Maize 50 17.3 146 5 11.8 106 Rice, 10 17.5 158 Sierra Sweet 902.08 54 5.99 1.15 Leone potatoes, 25 23.4 211 Cassava 50 27.2 246 5 9.4 105 10 13.7 153 Cassava, Togo 1,119.51 61 5.41 1.29 Maize, 25 18.9 211 Rice 50 22.3 250 Note: Risk levels: Low (green), Medium (yellow), High (red) Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens Expect- Expected Exposure Contribution Recurrence PML PML ed AAL AAL(% Main Country GVP (US$, to regional period (% (US$, (US$, GVP) risk crop millions) AAL (%) (years) GVP) millions) millions) levels 5 6.9 7,236 Aggregate 10 9.9 10,341 Portfolio 104,658.34 4,709 4.50 100 Results 25 13.6 14,204 50 15.9 16,664 5 4.7 2,125 Aggregate Portfolio 10 5.4 2,476 Results 45,448.02 1,584 3.49 43.4 (excluding 25 6.3 2,875 Nigeria) 50 6.9 3,156 Source: Original figure based on loss assessment Nigeria and Ghana contribute the most to production. Countries such as Chad, The the overall loss of the aggregated portfolio, Gambia, Guinea-Bissau, Mali, Niger, Nigeria, followed by Niger, Côte d’Ivoire, and Mali. Sierra Leone, and Togo see losses of around Proportional to its contribution to total crop or above over 10 percent at the medium-level area and to the overall economic output, risks (1 in 5-year), while Benin, Côte d’Ivoire, and Nigeria is the country contributing the most Ghana see losses around or below 5 percent of to the AAL, with 66 percent of the region’s their food crop production for medium-level 62 total losses, with the rest of the countries that events. For more extreme events (1 in 25-year substantially contribute to the regional losses level events), The Gambia, Guinea-Bissau, Mali, ranging between a little over 6 percent for Niger, Nigeria, and Sierra Leone see very high Ghana and Niger and 3.8 percent for Mali. On impacts on their food crop production—over the other end of the spectrum are Guinea- 20 percent, or 32 percent for Niger—while the Bissau, Liberia, and Senegal, accounting for less impacts of high-level events is slightly lower for than 1 percent of the region’s total AAL. Chad and Togo. Benin, Côte d’Ivoire, and Ghana experience lower impacts on their own food For the region’s individual countries, crop production, even for these 1 in 25-year- vulnerability to agricultural risks varies level events, with losses around 8 percent of in the share of losses in total food crop total food crop production. 4.2 FOOD CROP PRODUCTION, RISK IMPACT, AND LOSSES IN THE REGION A cross the portfolio of the 15 countries, percent), maize (7 percent), and sorghum (6 yam, cassava, and rice are the largest percent). Millet, plantains, cowpeas, and fonio contributors to aggregate AAL in terms jointly contribute about 7 percent to aggregate of value. Yam contributes 40 percent of the losses (figure 4.2). AAL, followed by cassava (27 percent), rice (13 FIGURE 4-2 Exposure by Crop Across the 15 Countries in the Region (value US$, millions, 1983–2018) Millet 3.179 Plantains 2.651 Cassava 28.073 Cow peas, 1.408 Fonio, 477 Sweet potatoes 159 Sorghum 6.280 Yam 42.088 Maize 6927 Rice, paddy 13.417 Source: Authors 63 BOX 4-2 Analyzing Future Food Crop Yield Declines Due to Climate Change Impacts A series of Climate-Smart Agriculture Investment Plans (CSAIPs) have been developed by the World Bank and partners for select countries to identify specific interventions and investments to increase agriculture productivity in an environmentally and socially sustainable way. The studies have conducted climate modeling scenarios to understand how climate change affects key food crops in the countries. In Mali, the average scenario-based modeling showed the percent difference in yields by 2050 for food crops will decline by the following amounts: yams, 1.9 percent; cassava, 2.6 percent; cowpeas, 3.9 per- cent; sorghum, 4 percent; rice, 4.5 percent; millet, 4.7 percent; wheat, 8.3 percent; maize, 15.8 percent. In Ghana, percentage difference in crop yields over a no-climate change reference scenario for 2050 un- der high GHG concentration scenario showed decline in yields for maize (-20.73 percent), millet (-6.04 percent), rice (-1.71 percent), sorghum (-6.19 percent), cassava (-0.99 percent), and yams (-0.93 percent). Applying a similar scenario-based modeling in Burkina Faso, the percent difference in yields by 2050 are as follows: maize (-21.69 percent); millet (-5.03 percent); rice (-1.99 percent); sorghum (-9.32 percent); cowpeas (-0.57 percent); cassava (-3.92 percent); yams (0.64 percent). While the loss assessment here in this report is based on the frequency and impacts of risk events in the past (the ASRA methodology), these studies offer valuable inputs for managing future climate risks in the region and provide comple- mentary information for investments in risk management going forward. Nevertheless, to fully under- stand the impact of climate change on the tail of the distribution, additional assessment is needed. Source: World Bank series of country CSAIPs Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens 4.2.1 Yam production and crop risk assessment Y am is a staple crop used for both global yam production, with Nigeria producing household consumption and income 71 percent of this amount (that is, some 53 generation in the equatorial and percent of the world production), and Côte humid areas of West Africa (map 4.1). The d’Ivoire, Ghana, and Benin as the region’s other region produces more than 75 percent of main yam producing countries. MAP 4-1 Yam is Mainly Produced in the Humid Areas in West Africa Yam Harvested Areas Cassava Harvested Areas 64 Source: World Bank map based on IFPRI SPAM data Yam production in the West Africa region Yam is a medium-risk crop compared with shows a steady uptrend, but yields have other crops in West Africa. The expected AAL declined in some places such as Nigeria cost as a percentage of total exposure is 5.1 and Togo due to pest and disease attack. percent for yam crops or US$2.1 billion per year. However, this decline in yield has been offset The LaR analysis indicates that yam crops may by the increase in the planted area in these face a loss equivalent to 27.5 percent of their countries. Currently, there are almost 8 million regional exposure (or US$11.6 billion) once in hectares planted with yams in West Africa, 50 years and a loss of 36.2 percent of its regional amounting to US$42 billion in total exposures. exposure (or US$15.2 billion) in a 250-years Figure 4.3 and figure 4.4 show the evolution of return period. Table 4.2 shows the expected LaR yam production and yields in West Africa for values for yam crops in West Africa. the period 1983–2018. FIGURE 4-3 West Africa: Historic Evolution of Yam Production Source: Original figure based on FAO database FIGURE 4-4 West Africa: Historic Evolution of Yam Crop Yields by Country 65 Source: Original figure based on FAO database Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens TABLE 4-2 West Africa, Expected LaR Values for Yam Crops Recurrence (years) 10 25 50 100 150 200 250 500 LaR (% exposure) 15.7% 22.7% 27.5% 32.1% 34.2% 36.2% 37.4% 40.7% LaR (US$, millions) 6,618 9,554 11,588 13,522 14,379 15,226 15,762 17,121 Source: Original table based on estimation results The volatility of yam production in West Africa with AAL values under 2 percent. Yam crop is greatly explained by the volatility of yam crop losses can be as high as 46 percent of the production in Nigeria. Nigeria accounts for 67 exposures in Nigeria for events with recurrence percent of yam exposures in the selected West period of 1-in-100 years. Figure 4.5 presents the African portfolio, followed by Ghana, which contribution of each country in the portfolio accounts for 20 percent. Nigeria (6.37 percent) to the expected AAL for yams in West Africa. followed by Côte d’Ivoire (5.39 percent) are Losses in yams for each country at different the countries that show the highest expected levels of risk are provided in the background AAL. Conversely, Benin, Ghana, and Chad show report West Africa Food Crop Risk Assessment. exceptionally low expected AAL for yam crops FIGURE 4-5 West Africa, Contribution of Each Country to the Expected AAL for Yam Crops Liberia 4% 66 Ghana 7%A Chad 1% Nigeria 31% Côte d'Ivoire 27% Benin 10% Togo 20% Source: Original figure based on estimation results 4.2.2 Cassava production and Crop Risk Assessment C assava is a versatile, staple crop and for 72 percent of the area and 65 percent of West Africa is one of the main cassava- the regional production of cassava. Nigeria producing regions in the world. is followed by Ghana, which accounts for 11 Nigeria, Côte d’lvoire, and Ghana are the main percent of the regional area and 22.6 percent countries producing cassava. Nigeria accounts of the total cassava production (map 4.2). MAP 4-2 Harvested Areas of Cassava in the Region Cassava Harvested Areas 67 Source: World Bank map based on IFPRI SPAM data Currently, there are more than 9.5 million Cassava is a relatively low-risk crop compared hectares planted with cassava in the selected with the production of other crops in West countries in West Africa. Regional total Africa. The expected AAL for cassava crops in exposures for this crop are estimated at US$28 West Africa accounts for 3.41 percent of its total billion, accounting for 26.8 percent of the total exposures or US$958.3 million per year. The exposures in the selected portfolio. Figure 4.6 expected LaR for this crop is 23.6 percent of and figure 4.7 show the evolution of cassava the exposures (US$6.62 billion) for a recurrence production and cassava crop yields over the period of 100 years and 29 percent of the period from 1983 through 2018. exposures (US$8.13 billion) for a recurrence period of 250 years. Table 4.3 presents the Cassava crops in West Africa are mostly expected LaR values for cassava crops in West affected by droughts, pests, and diseases. Africa for different return periods. Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens FIGURE 4-6 West Africa: Historic Evolution of Cassava Production Source: Original figure based on FAO database FIGURE 4-7 West Africa: Historic Evolution of Cassava Crop Yields 68 Source: Original figure based on FAO database TABLE 4-3 West Africa, Cassava Crops, Expected LaR Values for Different Return Periods Recurrence (years) 10 25 50 100 150 200 250 500 LaR (% exposure) 9.8% 15.6% 19.6% 23.6% 26.0% 27.8% 29.0% 33.1% LaR (US$, millions) 2,750 4,379 5,491 6,620 7,295 7,812 8,130 9,290 Source: Original table based on estimation results The volatility of cassava crop production in Nigeria, which governs the cassava results in West Africa is mainly driven by the cassava the portfolio, show an expected AAL of 4.27 crop yields in Nigeria. Nigeria accounts for 62 percent. While the losses for the whole cassava percent of cassava exposures in the selected portfolio are expected to be equivalent to 23.6 portfolio. Nigeria is followed by Ghana, which percent of the exposures for a return period accounts for 25.5 percent of cassava exposures of 1 in 100 years, it should be noted that the in the portfolio. Sierra Leone and Togo are the expected losses for the same recurrence period countries that show the highest expected AAL could be as high as 90 percent in Sierra Leone in the portfolio with values of 10.73 percent or 61 percent in Togo. Figure 4.8 presents the and 8.13 percent, respectively. Benin, The contribution of each country in the portfolio to Gambia, Ghana, and Guinea-Bissau are the the expected AAL and LaR in the cassava crop countries that show the lowest expected AAL portfolio in West Africa. for cassava crops with values under 2 percent. 69 FIGURE 4-8 West Africa, Contribution of Each Country to the Expected AAL for Cassava Crops Côte d'Ivoire 8% Gambia, The 1% Togo 23% Benin 5% Guinea 5% Ghana 5% Nigeria 12% Liberia 10% Sierra Leone 31% Source: Original figure based on estimation results Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens 4.2.3 Rice production and Crop Risk Assessment R ice is a staple food in many West Africa Except for a few countries that have attained countries and forms a major part of self-sufficiency in rice production, rice demand the diet in many others. During the exceeds production, and large quantities of past three decades, rice demand has increased rice are imported to meet demand at a huge consistently, and its growing importance is cost in hard currency. Up to half of regional rice evident in the strategic food security planning consumption needs are met through imports policies of many countries. The bulk of rice in the ECOWAS countries. production in West Africa comes from small- and medium-scale family farms, and over Rice crops cover an area of 9.1 million three million (small-scale) family farms or 18– hectares in West Africa. Rice production 24 million people13 are involved in rice and amounts to US$13.4 billion, which makes irrigated horticulture production. Rice is a staple this crop the third most important crop in food for almost all West African households, importance in the region. Most harvested areas including poor urban and rural households, and seem to be in Côte d’Ivoire, Guinea, Nigeria, and is the most consumed cereal after sorghum- Sierra Leone and along the rivers in The Gambia, millet, accounting for more than a third of all Niger, and Senegal. Nigeria is by far the main grain consumption. In 2017, West African rice producer of rice in the region, accounting for consumption was 15.86 million metric tons. It is 36 percent of the total planted area, followed projected to grow to 22 million metric tons by by Mali at 17 percent and Côte d’Ivoire at 11 70 2025 based on the trends in the last five years. percent. Map 4.3 and map 4.4 show the rice This is close to a 50 percent increase between producing areas and the differences in rice 2017 and 2025, with per capita consumption productivity. Most of the rice (40–45 percent) equally expected to rise from 43 kilograms in is produced in the rainfed lowland areas where 2017 to 49 kilograms in 2025 (ECOWAS 2019c). the potential rice yield is double that of highland MAP 4-3 Rice Production Across West Africa Rice Harvested Areas Source: Original map based on IFPRI SPAM data 13 areas, with irrigation further increasing yields producing areas are not necessarily those with by 20–80 percent. Yet the areas with the most the highest yields. MAP 4-4 Rice Yields, kg/ha, West Africa Rice Productivity Source: Original map based on IFPRI SPAM data Rice production shows an uptrend pattern, countries (such as Benin, Mali, Niger, and with several West African governments Senegal) have had a sharp increase in rice yields 71 devising specific policies to promote in recent years. Figure 4.9 and figure 4.10 show rice production, resulting in a significant the evolution of rice production and yields, increase in area sown. Yield increases are respectively, over the period 1983–2018. not evident at regional levels, although some FIGURE 4-9 Historic Evolution of Rice Production Source: Original figure based on FAO database Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens FIGURE 4-10 Historic Evolution of Rice Yields Source: Original figure based on FAO database Rice crops present a relatively low-level of loss equivalent to 16.3 percent of the exposure 72 risk compared with other crops in the region. (or US$2.19 billion) for a recurrence of 1 in 100 The expected AAL cost for rice crops in West years or a loss of 18.42 percent of the exposure Africa is 4.05 percent of the total exposures or (US$2.47 billion) once in 250 years. Table 4.4 US$543 million per year. The LaR for rice crops presents the expected LaR values for rice crops indicates that this crop may face an aggregate in West Africa for different return periods. TABLE 4-4 West Africa, Rice Crops, Expected LaR Values for Different Return Periods Recurrence (years) 10 25 50 100 150 200 250 500 LaR (% exposure) 9.4% 12.8% 14.6% 16.3% 17.3% 17.8% 18.4% 19.7% LaR (US$, millions) 1,258 1,711 1,965 2,188 2,327 2,390 2,471 2,650 Source: Original table based on estimation results The volatility of rice production in West Africa expected AAL with values as of 11.99 percent is greatly explained by the performance of this and 10.14 percent, respectively. Countries crop in Nigeria. Nigeria accounts for 50 percent situated in the equatorial humid region, with of rice exposures in the selected portfolio, AALs between 2 percent and 3 percent of the followed by Guinea, which accounts for 11.8 exposures, show the lowest expected AAL in percent of the exposures in the portfolio, and the portfolio. Nigeria, which is the main driver Côte d’Ivoire, which accounts for 10.5 percent of the rice crop performance in the selected of rice exposures in the portfolio. Chad and The portfolio, has an expected AAL of 3.92 percent Gambia are the countries that show the highest of the exposure. While the losses for the whole rice portfolio in West Africa are expected to be 4.11 presents the contribution of each country equivalent to 16 percent for a return period in the portfolio to the expected AAL and LaR of 1 in 100 years, it should be noted that the in the rice crop portfolio in West Africa. More expected losses for the same recurrence details on the rice sector and its volatilities can period could be 70 percent of the exposures in be found in Background Report 1. countries such as Chad and The Gambia. Figure FIGURE 4-11 West Africa, Contribution of Each Country to the Expected AAL for Rice Crops Gambia, The Nigeria 15% 5% Guinea 5% Mali 4% Chad 13% Ghana 3% Senegal 3% Burkina Faso 9% Côte d'Ivoire 3% 73 Benin 3% Togo Guinea-Bissau 8% 7% Sierra Leone Niger 7% 7% Source: Original figure based on estimation results 4.2.4 Maize production and Crop Risk Assessment M aize is a staple crop that plays a about one-fifth of the calories and protein central role in fulfilling population consumed by households. For countries such food requirements in West Africa. as Benin, Ghana, Mali, and Nigeria, maize is a Maize crops are currently grown on over 12.5 major source of food and cash for smallholder million hectares in West Africa, mostly rain-fed farmers. Figure 4.13 shows the evolution of areas (map 4.5). West African maize production maize production and crop yields in West Africa has grown significantly since the 1980s (figure from 1983 through 2018, with Burkina Faso 4.12) and is increasingly important for the having the highest yields consistently. region’s food security. Maize accounts for Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens MAP 4-5 Maize Production in West Africa Maize Harvested Areas Source: IFPRI SPAM data Other constraints to West African maize availability of improved seeds and access production include diseases, high cost of land to inputs. Inconsistent policies across the preparation, inadequate storage facilities and region have crippled investments to increase access to extension services, and inadequate productivity. 74 FIGURE 4-12 West Africa: Historic Evolution of Maize Production Source: Original figure based on FAO database FIGURE 4-13 West Africa: Historic Evolution of Maize Yields Source: Original figure based on FAO database Drought is the main peril affecting West African 100 years and 19.49 percent of the exposures maize crops. The expected AAL for maize in (US$1.35 billion) for a recurrence period of 250 West Africa accounts for 4.53 percent of the total years. Table 4.5 shows the expected LaR values 75 exposures (US$3.13 million). The expected LaR for maize crops in West Africa for different for this crop is 14.43 percent of the exposures return periods. (US$1.2 billion) for a recurrence period of TABLE 4-5 West Africa, Maize Crops, Expected LaR Values for Different Return Periods Recurrence (years) 10 25 50 100 150 200 250 500 LaR (% exposure) 10.1% 13.4% 15.6% 17.4% 18.3% 18.7% 19.5% 22.0% LaR (US$, millions) 701 926 1,078 1,207 1,265 1,295 1,350 1,524 Source: Original table based on estimation results Nigeria drives the results for maize crops in as high as 11.56 percent and 11.42 percent, the selected portfolio in West Africa. Nigeria respectively. Maize expected AAL in countries accounts for 38 percent of maize exposures situated in the equatorial and humid regions in the selected portfolio, followed by Ghana, present the lowest values for the portfolio. which accounts for 15 percent of the exposures Nigeria, which is the main driver of the maize in the portfolio, and Mali, which accounts for crop performance in the selected portfolio, 10.5 percent of maize exposures in the portfolio. has an expected AAL of 4.45 percent. Most of Niger and Sierra Leone are the countries that the countries in the region are expected to show the highest expected AAL with values experience maize crop losses of 30–40 percent Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens of the exposures once in 100 years. However, 4.14 presents the contribution of each country in countries such as Niger or Sierra Leone, the in the portfolio to the expected AAL and LaR in loss expectancy rises to values of 80 percent the maize crop portfolio in West Africa. for recurrence periods of 1 in 100 years. Figure FIGURE 4-14 West Africa, Contribution of Each Country to the Expected AAL for Maize Crops Benin 2% Guinea-Bissau 3% Niger Côte d'Ivoire 14% 4% Ghana 5% Guinea 5% Nigeria 5% Sierra Leone 14% Togo 6% Gambia, The Chad 76 10% 9% Mali 8% Senegal Burkina Faso 9% 8% Source: Original figure based on estimation results 4.2.5 Sorghum production and Crop Risk Assessment Sorghum is crucially important to food Burkina Faso, southern Mali, southeastern Niger, security in West Africa as it is uniquely northern Nigeria, and Togo (map 4.6). Nigeria, drought resistant among cereals and can Niger, and Burkina Faso account for 48 percent, withstand periods of high temperature. 15 percent, and 13 percent of West African Sorghum is extensively produced throughout sorghum production, respectively. The highest the region with 15 million hectares currently productivity areas appear to be in Burkina Faso, under production, with a total value estimated northern Ghana, Guinea, Nigeria, and Senegal at US$6.2 billion. Sorghum is especially suited (map 4.7) Despite a large, harvested area, for the semiarid areas in the Sahel region such as Nigeria sees low yields. Average yields for the Figure 4.15 and figure 4.16 show the evolution region have declined for this sorghum overall of sorghum production and sorghum yields despite its importance for food security and its over the period 1983–2018. drought resilience. MAP 4-6 Sorghum Production Across West Africa Sorghum Harvested Areas Source: IFPRI SPAM data MAP 4-7 Sorghum Productivity Across West Africa 77 Sorghum Harvested Areas Source: World Bank map based on IFPRI SPAM data Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens FIGURE 4-15 West Africa: Historic Evolution of Sorghum Production Source: Original figure based on FAO database FIGURE 4-16 West Africa: Historic Evolution of Sorghum Crop Yields 78 Source: Original figure based on estimation results The main peril affecting sorghum crops in West years and 31 percent of the exposures for a Africa is drought. The expected AAL for this crop recurrence period of 250 years. Table 4.6 shows accounts for 5.26 percent of the exposures. The the expected LaR values for sorghum crops for expected LaR for this crop is 26.44 percent of different return periods. the exposures for a recurrence period of 100 TABLE 4-6 West Africa, Sorghum, Expected LaR Values for Different Return Periods Recurrence (years) 10 25 50 100 150 200 250 500 LaR (% exposure) 13.4% 19.4% 23.2% 26.4% 28.6% 29.4% 31.0% 34.8% LaR (US$, millions) 842 1,216 1,457 1,660 1,794 1,847 1,949 2,186 Source: Original table based on estimation results The performance of sorghum crops in the Nigeria, which is the main driver of the sorghum selected portfolio of crops and countries for risk crop performance in the selected portfolio, assessment is governed by the performance the expected AAL for sorghum is 4.73 percent. of sorghum in Nigeria. Considering sorghum Most of the region’s countries are expected exposures in the selected portfolio, Nigeria to experience sorghum crop losses of 20–40 accounts for 65 percent, followed by Niger, percent of the exposures once in 100 years. which accounts for 9.4 percent, then Burkina However, in countries showing high risk, such Faso and Mali. The Gambia, Niger, and Sierra as Niger, sorghum crop losses are expected Leone show the highest expected AAL with to be 80 percent of the exposures for 1 in 100 values above 8 percent. For the remaining years. Figure 4.17 presents the contribution of countries in the region, sorghum losses are each country in the portfolio to the expected expected to be from 3 percent to 6 percent. In AAL and LaR in the sorghum crop portfolio. FIGURE 4-17 West Africa, Contribution of Each Country to the Expected AAL for Sorghum Crops 79 Niger Guinea-Bissau 16% 5% Benin 5% Chad 5% Sierra Leone 14% Burkina Faso 6% Gambia, The 13% Nigeria 8% Togo 8% Mali Senegal 10% 10% Source: Original figure based on estimation results Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens 4.2.6 Millet production in West Africa M illet is an important, staple food for still significant shortfalls between production most people in the Sahel and West and demand, which is increasing. The increase Africa and is planted on almost in millet production is mostly explained by the 12.9 million hectares. It is important for the increase in area sown, as yields have remained food security of most countries and accounts constant over the period. Figure 4.18 shows for more than 30 percent of the region’s total that millet production has increased steadily cereal production, with Niger alone accounting in West Africa from 1983 through 2018. Yields for 54 percent of the total regional area planted differ significantly between countries, but they with millet, followed by Mali with 16.7 percent have remained fairly constant over the past of the total areas sown (map 4.8). Despite the four decades. recent increase in millet production, there are MAP 4-8 Millet Production in West Africa Millet Harvested Areas 80 Source: IFPRI SPAM data The increase in millet production is mostly significantly between countries, they have explained by the increase in sown area as remained fairly constant over the past four yields remained constant along the period. decades (figure 4.19). Figure 4.18 shows that millet production has increased steadily in West Africa over the period from 1983 through 2018. But while yields differ FIGURE 4-18 West Africa: Historic Evolution of Millet Production Source: Original figure based on FAO database FIGURE 4-19 West Africa: Historic Evolution of Millet Yields 81 Source: Original figure based on FAO database Millet is high risk relative to other crops (equivalent to US$821 million) and the LaR for analyzed in the selected portfolio. The expected a recurrence period of 1 in 250 years is 29.53 AAL for millet accounts for 6.44 percent of the percent (equivalent to US$939 million). Table exposures. The LaR for millet for a recurrence 4.7 presents the expected LaR values for millet period of 1 in 100 years is 25.84 percent crops for different return periods. Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens TABLE 4-7 West Africa, Millet, Expected LaR Values for Different Return Periods Recurrence (years) 10 25 50 100 150 200 250 500 LaR (% exposure) 15.2% 19.9% 22.7% 25.8% 27.5% 28.6% 29.5% 31.6% LaR (US$, millions) 482 634 722 821 875 908 939 1,004 Source: Original table based on estimation results The volatility of West African millet production Niger, as the main driver of the millet crop is greatly explained by millet’s performance in performance, presents an expected AAL of 6.16 Niger, which accounts for 45 percent of millet percent. While the losses for the whole millet exposures in the portfolio. Niger is followed portfolio in West Africa are expected to be 26 by Mali, which accounts for 24 percent of percent of the exposures for a return period the exposures in the portfolio, and Burkina of 1 in 100 years, it should be noted that the Faso, which accounts for 11 percent of millet expected AAL for the same recurrence period exposures. Mali and Chad have the highest could be 40 percent to 50 percent in each of expected AAL, with values of 7.68 percent the portfolio countries. Figure 4.20 presents the and 7.41 percent, respectively. Burkina Faso, contribution of each country in the portfolio The Gambia, and Guinea, with AAL between to the expected AAL and LaR in the millet crop 4 percent and 5 percent, are the countries portfolio in West Africa. showing the lowest expected AAL (figure 4.20). 82 FIGURE 4-20 West Africa, Contribution of Each Country to the Expected AAL for Millet Crops Chad Burkina Faso 18% 10% Gambia, The 12% Mali 18% Guinea 12% Niger 15% Guinea-Bissau 15% Source: Original figure based on estimation results 4.2.7 Cowpea production in West Africa C owpea is grown as both a food and accounting for 80 percent of the region’s animal feed crop in West Africa’s total area planted. Niger is followed by semiarid tropics. Cowpea’s high protein Burkina Faso, which accounts for 17.7 percent content, its adaptability to different types of of the total area sown with cowpeas. Burkina soil and intercropping systems, its drought Faso has higher cowpea yield than Niger, but resistance, and its ability to improve soil fertility Niger’s yields have been steadily increasing and prevent erosion makes it an important over time. Cowpea production has been economic crop in many areas of the Sahel consistently increasing from 1983 through region. The sale of the stems and leaves as 2018. The increase in cowpea production is animal feed during the dry season also provides explained by both planting cowpeas over a vital income for farmers. increasing areas and higher yields in Niger. Figure 4.21 and figure 4.22 show the historic The area planted with cowpea amounts evolution of cowpea production and cowpea to almost 7 million hectares, with Niger yields in West Africa from 1983 through 2018. FIGURE 4-21 West Africa: Historic Evolution of Cowpea Production 83 Source: Original figure based on FAO database Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens FIGURE 4-22 West Africa: Historic Evolution of Cowpea Yields Source: Original figure based on FAO database Relative to other crops analyzed in the selected US$738 million) and the LaR for a recurrence 84 portfolio, cowpea is a high-risk crop. The period of 1 in 250 years could be as high as expected AAL for cowpeas accounts for 9.57 57.9 percent (equivalent to US$815 million). percent of the exposures. The LaR for this Table 4.8 presents the expected LaR values for crop for a recurrence period of 1 in 100 years cowpea crops for different return periods. could be as high as 52.4 percent (equivalent to TABLE 4-8 West Africa, Cowpeas, Expected LaR Values for Different Return Periods Recurrence (years) 10 25 50 100 150 200 250 500 LaR (% exposure) 30.8% 41.3% 47.4% 52.4% 55.1% 56.5% 57.9% 61.1% LaR (US$, millions) 434 581 668 738 776 795 815 860 Source: Original table based on estimation results The volatility of cowpea production in West Africa are expected to be 70 percent to 80 Africa is explained by the performance of this percent of the exposures for a return period crop in Niger. Niger accounts for 74 percent of 1 in 100 years. Figure 4.23 presents the of cowpea exposures in the selected portfolio. contribution of each country in the portfolio to Niger shows the highest expected AAL with the expected AAL and LaR for the cowpea crop values as high as 12.38 percent. The expected portfolio. AAL for the whole cowpea portfolio in West FIGURE 4-23 West Africa, Contribution of Each Country to the Expected AAL for Cowpea Crops Burkina Faso 10% Niger 15% Senegal 10% Source: Original figure based on estimation results 85 4.2.8 Plantain production in West Africa W est Africa is one of the major plantain- d’Ivoire accounts for 50 percent of the region’s producing regions of the world, area planted with plantains, followed by Ghana accounting for approximately 32 with 40 percent. While Ghana occupies second percent of worldwide production. Plantains place in planted area with plantains, Ghana is are an important staple crop in the region with first in total plantain production, accounting a high nutritional content, variety of preparation for 68 percent of the region’s total plantain methods, and a production cycle that is less output. Ghana’s plantain yields are almost three labor intensive than many other crops. Plantains times higher than the region’s other plantain- are a key component in sustainable agricultural producing countries. Plantain production has systems in high rainfall zones. Major constraints grown consistently from 1983 through 2018, to plantain production include pests and and this production increase is explained by disease, short shelf life, and damage during a larger sown area and by Ghana’s increased transportation. plantain yields and production. Figure 4.24 and figure 4.25 show the historic evolution of The area planted with plantains in West plantain production and crop yields in West Africa is almost 911,000 hectares. Côte Africa over the period from 1983 through 2018. Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens FIGURE 4-24 West Africa: Historic Evolution of Plantain Production Source: Original figure based on FAO database FIGURE 4-25 West Africa: Historic Evolution of Plantain Yields 86 Source: Original figure based on FAO database Relative to other crops analyzed in the selected million) and the LaR for a recurrence period of portfolio, plantains are low risk. The expected 1 in 250 years could be as high as 8.16 percent AAL for plantains accounts for 1.19 percent (equivalent to US$216 million). Table 4.9 of the exposures. The LaR for plantains for a presents the expected LaR values for plantains recurrence period of 1 in 100 years could be for different return periods. as high as 6.96 percent (equivalent to US$185 TABLE 4-9 West Africa, Plantains, Expected LaR Values for Different Return Periods Recurrence (years) 10 25 50 100 150 200 250 500 LaR (% exposure) 3.6% 5.1% 6.0% 7.0% 7.5% 7.8% 8.2% 8.8% LaR (US$, millions) 95 135 160 185 199 208 216 234 Source: Original table based on estimation results The volatility of plantain production in West for the whole plantain portfolio in West Africa Africa is explained by the performance of this is expected to be 8 percent to 9 percent of the crop in Ghana, which accounts for 66.8 percent exposures for a return period of 1 in 100 years of plantains exposures in the portfolio. In in the main production countries. Figure 4.26 general, the expected AAL for plantain crops in presents the contribution of each country in the portfolio are between 0.5 percent and 1.5 the portfolio to the expected AAL and LaR in percent, which is extremely low compared with the plantain portfolio in West Africa. other crops in the portfolio. The expected AAL FIGURE 4-26 West Africa, Contribution of Each Country to the Expected AAL for Plantain Crops Liberia 13% 87 Guinea-Bissau 17% Côte d'Ivoire 39% Ghana 31% Source: Original figure based on estimation results Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens 4.3 CORRELATIONS BETWEEN YIELD AND WEATHER- RELATED RISKS T o understand if and how weather- that it is not possible to identify a phase of the related events affect yields, this report crop on which the selected index provides attempted to identify correlations the best estimation across all the districts. For between production and climatic (drought) instance, for correlations between eMODIS indicators; however, this type of correlation NDVI and millet crops in Senegal, the best was limited in the region. If such correlations correlations were found in two districts during could be established, relevant indexes could be the vegetative phase, in four districts during used as potential proxies for early estimations the reproductive stage, and in nine districts of yield volatility and, consequently, for during the crop’s maturity phase. decision-making around risk response measures. However, the results showed that no From a policy point of view, the fact that meaningful correlation could be established existing data don’t show any meaningful between the indexes and the data analyzed.14 correlations between indexes and crop The methodology and detailed results are production means that more analysis based described in box 4.3. on better data must be conducted before mechanisms such as a weather-based The analysis of correlations for each of the index trigger or transfer can be established 88 phenology stages of the analyzed crops without a potentially high-cost basis. The showed that it is not possible to identify information on crop yield that is available is not which crop phenology stage is better for standardized and thus not of sufficient quality measuring the evolution of the index to to develop remote sensing weather-based infer crop yield volatility. The hypothesis for triggers, given existing, huge uncertainties this analysis was that it would be eventually about the correlations between selected possible to identify one phase of the crop drought indexes and crop yields. Therefore, on which the index provides a meaningful the basis risk15 could potentially be high. It can correlation with crop yields, to be used as be noted that a proportion of districts showed an early trigger of eventual coverage—for meaningful correlation, which indicates that example, using eMODIS NDVI during the there is a potential for using indexes as triggers. flowering stage of maize crops can be a good Also, other initiatives (for example, ARC/ estimator of the likely harvestable crop yield. Africa RiskView [ARV]) do find some degree of The outcome of the analysis showed, however, correlation between weather indicators and 14 For meaningful correlation, for all the indexes, except for Cumulated Synthetic Index (CSI), the threshold is as follows: R2>0.33 = meaningful; R2 in between -0.33 and 0.33 were considered independent; R2< - 0.33 were considered meaningless. In the case of CCI, for which the inverse correlations between yields and the CCO was sought, the threshold used was R2>0.33 = meaningless; R2 in between -0.33 and 0.33 were consid- ered independent; and R2< - 0.33 was considered meaningful. Note that a meaningful correlation does not mean that the index should be used as a proxy for estimating yields. Meaningful correlation means that, given the quality of the crop yield information we are using for the analysis, finding an R2 greater than 0.33 (or lower than -0.33 in the case of CCI) will merit an effort to procure more accurate information to perform a more accurate analysis (which wasn’t possible with the quality of available information). 15 Basic risk arises when indices are imperfectly correlated to actual losses. This can result in risk transfer products triggering payouts that are higher than the actual losses or not triggering payouts or triggering lower payouts than actual losses. BOX 4-3 Analysis of Correlation Between Selected Indexes and Crop Yields: Methodology and Detailed Results Objective: To assess the correlations between various indicators of selected indexes chosen as potential proxies for early estimations of yield volatility of key crops in West Africa. This assessment focuses on the correlation between the selected indicators at different stages of the crop cycle rather than assessing consecutive drought years (drought cycles). The correlation analysis be- tween selected indexes and crop yields was performed over crop yield series aggregated at the first level administrative division (Admin 1 level) that were procured from official publications from the Ministries of Agriculture and National Statistical Services at the national level. The team acknowledges that the reli- ability of the crop yield information collected for the analysis may have quality issues. Caution should be exercised when interpreting the results as the reliability of the analysis produced from this information might affect the conclusions. Selected Indexes: • eMODIS AQUA NDVI reported on a decadal basis for 2002–20 by the U.S. Geological Survey (USGS-EROS). • CHIRPS Seasonal Rainfall Accumulation, reported on five-year intervals for 2001–20 by FEWS NET. • Rainfall Estimates 2 data (RFE2), Seasonal Rainfall Accumulation, reported on a decadal basis 89 for 2001–20 by FEWS NET. • Water Requirement Satisfaction Index (WRSI) for the selected crops reported on a decadal basis for 2001–20 by FEWS NET. Selected Risk Models: • Evolution of the selected indexes for the selected crops’ vegetative, reproductive, and maturity stages. • Cumulated Synthetic Index (CSI), considering the number of decadal periods along the crop season on which the actual value of the selected index falls short of its expected value. • Aggregate Synthetic Index (ASY), considering the weighted average of the selected indexes for each decadal along the crop season by using the share of the demand of water of the crop in each decadal over the total demand of water along the season as weighting factor. Results of the analysis: • Only 11 percent of the selected Admin 1 level units showed meaningful correlations (R2>0.5) with yields. • Across all the analyzed indexes, the strongest correlations were found for millet crops in the Sahel, while the weakest correlations were found for sorghum crops. • CHIRPS and RFE2 provided meaningful correlation (R2>0.5) in 30 percent of the analyzed units when it is measured during the reproductive stages of maize and millet crops. • CHIRPS-ASY presented meaningful correlations (R2>0.5) in 23 percent of the units, while RFE2-ASY presented meaningful correlations in 27 percent of the Admin 1 level units in the portfolio. • It was not possible to identify a pattern to determine on which crop phenology stage the evolution of the selected indexes provides the best correlation with crop yields. Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens crop yields. Thus, there is a need to explore found. The analysis used food insecurity data further, for example, by procuring crop yield as measured by the Cadre Harmonisé (CH). A information directly from the local sources to strong negative correlation would be expected refine the analysis. between yield and the number of people who are food insecure and require food aid; instead, A negative correlation between production about half of the R2 coefficients resulting from and food insecurity data could not be the analysis show positive correlations (table established; instead, the reverse was 4.10). TABLE 4-10 Correlations of crop yields and number of people food insecure food, expressed in R2 Current Food Security Assistance phase 3 to 5 Country Admin 1 Crop Sep-Dec T0 Jan-May T+1 Jun-Aug T+1 Niger Diffa Millet -0.23 -0.17 s/d Niger Dosso Millet 0.1 0.28 s/d Niger Maradi Millet -0.42 -0.63 s/d Niger Niamey Millet -0.94 0.51 s/d Niger Tahoua Millet 0.42 -0.93 s/d Niger Tillaberri Millet s/d s/d s/d Niger Zinder Millet 0.36 -0.74 s/d 90 Mali Bamako Millet s/d s/d s/d Mali Kayes Millet -0.33 -0.04 s/d Mali Koulikoro Millet -0.83 -0.44 s/d Mali Mopti Millet 0.01 -0.3 s/d Mali Ségou Millet s/d s/d s/d Mali Sikasso Millet 0.59 0.63 s/d Burkina Faso Sud-Ouest Sorghum 0.37 -0.7 s/d Burkina Faso Boucle-du-Mouhoun Sorghum s/d s/d s/d Burkina Faso Centre Sorghum 0.44 0.04 s/d Burkina Faso Centre-Est Sorghum -0.82 -0.59 s/d Burkina Faso Nord Sorghum -0.33 -0.13 s/d Burkina Faso Centre-Sud Sorghum -0.82 -0.36 s/d Burkina Faso Centre-Nord Sorghum -0.62 -0.74 s/d Burkina Faso Centre-Ouest Sorghum -0.57 -0.78 s/d No or missing data Burkina Faso Plateau-Central Sorghum s/d s/d s/d R2 makes sense R2 relatively makes Burkina Faso Est Sorghum -0.79 -0.86 s/d sense Burkina Faso Hauts-Bassins Sorghum 0.04 -0.11 s/d R2 does not make Togo Plateaux Maize -0.12 s/d s/d sense R2 does not make Togo Maritime Maize 0.41 0.22 s/d sense at all While it seems likely that these frequent, the region, no formal correlations could be large, food-crop production losses are at established. The lack of a robust correlation least partially due to recurrent drought and likely points to the diversity of food insecurity that they are an important contributor to drivers, the existence of chronic food insecurity, the high occurrence of food insecurity in and data inaccuracies. 4.4 RANGELAND PRODUCTION RISK ASSESSMENT D rought and locust attacks can lead people in the Sahelian zone. The main source of to widespread livestock death and fodder for livestock production in the region is severely deplete livestock assets for from rangelands, with the main rangeland areas affected households. To understand food being located in Tahua, Tillaberry, and Zinder systems vulnerabilities, it is therefore also Administrative Regions in Niger and Mopti and important to assess risks to rangelands. the Timbuktu Administrative Region in Mali. 2 Fig4.27 Livestock provide a source of livelihood Many pastoralists in the region are becoming for millions of pastoralists and contribute more vulnerable to natural disasters such as significantly to national economies in Sahel locust invasions and severe droughts. Studies countries. Indeed, the sector contributes to show that rebuilding a herd that has suffered 10–15 percent of the GDP in Sahelian countries a 30 percent loss takes an average of 10 years. and a source of income for 69 percent of rural 91 MAP 4-9 Distribution of Rangeland Areas in West Africa Source: Original areas Main rangeland in figure based on West estimations Africa from areLand Collection 6 MODIS Tahua, Tillaberry, Cover - MCD12Q1 and -University of Maryland Zinder Administrativ (UMD) Regions in Niger and Mopti and Timbuktu Administrative Regions in Mali. Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens As rangeland is the main source of fodder The results of the analysis performed for livestock sector in the region, a pasture indicate that selected countries lose an rangeland risk assessment based on the estimated average 3.39 percent of rangeland evolution of eMODIS AQUA NDVI over the production (measured through the land classified as rangeland was performed.16 accumulated NDVI during the peak season) NDVI is measured using satellite imagery and every year. To understand the potential provides a good indicator of vegetation growth impacts on the livestock sector from fodder conditions on the ground. NDVI imagery can crop production risks, rangeland production be used to distinguish between different losses were assessed at Admin 1 level in the types of land use cover, for example between Sahel region in West Africa. The analysis shows natural pasture and grazing lands and areas of that expected losses can be as high as 14.4 sparse vegetation or bare soil or water; and to percent of the exposures once in 100 years and measure the condition of the vegetative cover 18.2 percent of the portfolio exposures once in and distinguish between healthy and dead 250 years. Slight variations in the accumulated or dry vegetation. NDVI is also very closely NDVI during the peak season, as the livestock correlated with climatic variables such as production systems in the region are quite precipitation and potential evapotranspiration. fragile, may lead to significant stress situations The close correlation between NDVI, plant that can jeopardize the production or even photosynthesis, plant vigor, and amount of cause the loss of stock. rainfall points to the fact that NDVI is potentially a particularly good proxy to use to measure the impact of progressive drought on pasture 92 quality and productivity. The underline assumption of this analysis is that the critical herd management practices that drive livestock production are synchronized with the expected availability of fodder to feed the herd. In this regard, deviations from the expected supply of fodder during the critical periods such as breeding and calving create significant disturbances in herd management and, consequently, in livestock production. Thus, the rangeland risk assessment is performed based on the evolution of the rangeland production, measured through the NDVI, during peak seasons. Peak seasons are defined as the decadal periods on which the expected eMODIS NDVI falls in the first tercile in terms of NDVI values. 16 The detailed methodology is presented in appendix B and in the West Africa crop and rangeland risk assessment background report to be published. 3 Fig 4.28 MAP 4-10 Expected Peak Season Accumulated NDVI 3 Fig 4.28 In terms of rangeland productivity measured through accumulated NDVI during the peak Source: Original figure based on FEWS NET eMODIS NDVI season, subhumid and semiarid areas in the south of the region are up to three time more productive that the MAP arid and 4-11 semiarid areas in the north. Exposures 93 In terms of rangeland productivity measured through accumulated NDVI during the peak Source: Original figure based on FEWS NET eMODIS NDVI season, subhumid and Note: The World Bank semiarid areas and the International Livestock in Institute Research south the (ILRI) of are currently the region conducting are of work on the feasibility drought risk finance and insurance solutions for the livestock sectors in Burkina Faso, Mali, Niger, and Senegal. up to three time more index-based productive that the arid and semiarid areas in the north. Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens 94 5 HOTSPOTS FOR RISKS TO WEST AFRICA’S FOOD CROP PRODUCTION Key points in this chapter: • The report identifies six hotspots that are particularly vulnerable to food insecurity as a results of production risks. • Food crop losses were assessed at the district Admin 1 level for four of the hotspots. • For three of the hotspots, the impacts on food crop production would be 1–13 percent This chapter looks at food crop losses with for 1 in 10-year risks and up to 20–25 percent subnational level data to provide a more losses for 1 in 100-year risks. nuanced picture of production risk impacts • Countries need to improve production data across the region. Further, the chapter zooms collection going forward: only eight of the in on four vulnerable hotspots in West countries covered in this report had Admin 1 Africa, assesses the impacts of production level production data and the remaining seven risks in these specific areas, and details the countries could therefore not be included in distribution of losses between districts and the assessment at subnational level. countries. 16 The Admin 1 level food crop risk assessment for selected West African countries was based on official FAOSTAT records and complemented with agriculture production records for local Ministries of Agriculture and Departments of Statistics in the selected countries. The analysis is based on Admin 1 level historic records of crop harvested area, production, and yield for each selected crop in each of the selected countries from 2003 through 2018. T his chapter aims to describe the The overall crop risk assessment potential volatility of the food crop performed at Admin 1 level comprises portfolio in those regions of high seven West African countries that were concern from a food security perspective. included based on data availability: Six fragility hotspots were identified through Burkina Faso, Ghana, Liberia, Mali, Niger, the mapping exercise in chapter 4. The crop Senegal, and Togo. The crop risk assessment risk assessment performed for the selected is performed over the eight main food food security hotspots provides relevant crops selected based on their importance information about the food supply risk in in planted area in each country and also on critical West African regions. The hotspot the availability of information for the analysis. approach allows for cross-border solutions to The crops selected in each of the countries in managing agriculture and food security risks, the analysis is presented in figure 5.1. The full thereby providing opportunities for regional analysis for the countries and crops below at collaboration. It also enables assessing the Admin 1 level can be found in the background risk at the zonal level and gaining a better report West Africa Crop Risk Assessment. understanding of the tradeoffs in designing a tailored risk financing strategy for each zone of food security concern. FIGURE 5-1 Selection of Crops and Countries for the Analysis Maize Maize 95 Burkina Faso Millet Millet Rice Liberia Mali Rice Maize Rice Cassava Sorghum Sorghum Millet Fonio Rice Ghana Sorghum Yams Cassava Maize Millet Cassava Maize Senegal Millet Maize Niger Togo Plantains Rice Rice Sorghum Sorghum Sorghum Rice Yams Source: Original figure based on FAOSTAT A detailed analysis of exposures in the evidence for the existence of at least two risk Admin 1 level portfolio of selected crops accumulation hotspots in the region. The first and countries in West Africa presents hotspot comprises Northern, Brong Ahafo, evidence of high-risk accumulations in Ashanti, Eastern, Central, and Volta Regions certain areas in the region. Map 5.1 below, in Ghana. This hotspot accounts for 61.2 showing the distribution of the exposures at percent of the total exposures in the selected the Admin 1 level in the portfolio, provides portfolio. The high-risk accumulation in this Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens hotspot is because these regions are the exposures in the selected portfolio. The main main production areas for the high-value driver of risk accumulation in the second yam and cassava crops. The second hotspot hotspot is rice, which accounts for more than comprises the Segou and Sikasso Regions in 62 percent of the hotspot’s exposures. Mali and accounts for 13.3 percent of the total MAP 5-1 Administrative Level 1 Crop Risk Assessment Portfolio, Exposures 96 Source: Original map based on estimation results Crops in Admin 1 level units in the Sahelian A detailed analysis of the expected AAL at regions are riskier than the crops cultivated the Admin 1 level portfolio of selected crops toward the humid and equatorial regions. and countries in West Africa confirms that Fonio, sorghum, and millet, which are the highest expected AAL are in drought- cultivated in the subhumid and semiarid areas, prone areas toward the Sahel region. The show high expected AAL in the portfolio. Segou Region in Mali followed by Kayes (also Conversely, the crops cultivated in the humid in Mali) and Tambacounda, Louga, and Thies Guinean region present lower expected AAL. Regions in Senegal are the Admin 1 level- For instance, plantains and yams present AAL units that show the highest expected AAL. of 1 percent of the exposures. Countries in Map 5.2 presents the expected AAL for the semiarid areas toward the northernmost part Admin 1 level units considered in the crop of the region show the highest expected AAL. risk assessment portfolio. Mali and Senegal show the highest expected AAL in the selected portfolio, while countries in humid tropical areas, such Ghana and Togo, show the lowest level of risk. MAP 5-2 Admin 1 Level Crop Risk Assessment Portfolio, Expected AAL Source: Original map based on estimation results Map 5.3 presents the selected hotspots for are factored in on the crop yield series used, food crop risk assessment in West Africa. As and the actual historic yields over the 30-year 97 part of the analysis, food crop production risk period was d-trended to remove the effects of was assessed for four selected hotspots in technology and management practices. The western Mali (Hotspot #1), south–southwest procedure for D-trending yields is described Burkina Faso–northwest Ghana (Hotspot #2), in the methodology section of this report eastern Burkina Faso–western Niger–north (see appendix A). Benin (Hotspot #3), and central–southwest Ghana (Hotspot #4). For those hotspots comprising areas in Benin, Côte d’Ivoire, and Nigeria, it was not possible to perform the food crop risk assessment due to an inability to obtain a full series of crop production statistics to perform the analysis. A detailed food crop risk assessment for each of the selected food insecurity hotspots for which crop production information is available is presented in the following subsections. This analysis leverages 30 years of Admin 1 level-yield data. Major shifts during that time frame such as technology application and change in climate conditions A Blueprint for Strengthening Food System Resilience in West Africa: Regional Priority Intervention Areas MAP 5-3 Selected Food Insecurity Hotspots for Crop Risk Assessment Source: Original map based on estimation results 98 5.1 HOTSPOT #1 CROP RISK ASSESSMENT H otspot #1 comprises the Kayes and determines whether a year is good or bad Koulikoro Regions in Mali. The zone for food security. FEWS NET reports also cite has a moderate population density and pests, diseases, floods, and bushfires as other is primarily characterized by rainfed agriculture important hazards affecting crop production and livestock rearing. The zone consists in the hotspot. primarily of plains, hills, and woodland. Annual precipitation ranges from 600–800 millimeters Hotspot #1 accounts for 6.55 percent of the in the south to 400–500 millimeters in the north area and 3.0 percent of the exposures of the and allows for cultivating sorghum, millet, and selected Admin 1 level food crop portfolio maize, which are the main food crops. Sorghum for West Africa. The selected food crops for is more dominant toward the south and millet risk assessment in Hotspot #1 are rice, maize, toward the north. fonio, sorghum, and millet. Maize and sorghum (accounting for 35 percent and 27 percent of According to USAID report in 2010 18, drought exposures, respectively) are the main crops in is the main hazard affecting crop production the portfolio. Table 5.1 presents the crop area in this hotspot. Drought is one of the major and the exposures of each crop considered in production risks in the Sahel and largely the analysis. 18 Livelihood Zoning and Profiling Report: Mali, USAID January 2010. TABLE 5-1 Expected Crop Area and Exposures for Selected Crops in Hotspot #1 Expected Crop Area Exposure Crop (hectares) % (US$, millions) 7.0% Fonio 8,748 0.5% 5 0.5% Maize 459,562 26.4% 330 35.0% Mill 400,629 23.0% 131 13.9% Rice 111,616 6.4% 220 23.3% Sorghum 761,711 43.7% 257 27.2% Total 1,742,266 100.0% 943 100.0% Source: Original table based on estimation results This analysis finds that the expected AAL of million) once in 100 years or losses as high the selected food crops for Hotspot #1 is 6.13 as 22.6 percent of the exposure (or US$213 percent or US$57.8 million per year. The LaR million) once in 250 years. Table 5.2 shows the analysis indicates that the whole area selected expected LaR values estimated for the portfolio under the hotspot may expect losses as high in Hotspot #1. as 19.7 percent of the exposure (or US$185 TABLE 5-2 Hotspot #1, Expected LaR Values Recurrence (years) 10 25 50 100 150 200 250 500 LaR (% exposure) 13.00% 16.06% 17.95% 19.65% 20.88% 21.70% 22.59% 24.06% 99 LaR (US$, millions) 122.5 151.4 169.3 185.3 196.9 204.6 213.0 226.8 Source: Original table based on estimation results The volatility of the selected food crop portfolio the Koulikoro Region presents an AAL of 4.8 in Hotspot #1 is explained by the volatility of percent of total exposures. Table 5.3 presents maize and sorghum crops, which account for the contribution of each crop to the expected 35 percent and 27 percent, respectively, of AAL and LaR for each crop and Admin 1 level in exposures in Hotspot #1. Maize crops present Hotspot #1. an AAL of 4.29 percent, while sorghum crops present an AAL of 8.09 percent. The expected losses for these crops for recurrence periods for 1 in 100 years are 26.73 percent of the exposures for maize and 32.66 percent of the exposures for sorghum. The Kayes Region presents an AAL of 10.04 percent of the total exposures while Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens TABLE 5-3 Expected Crop Area and Exposures for Selected Crops in Hotspot #1 LaR (in percent for recurrence period years) Unit Exposure % AAL % 10 25 50 100 150 200 250 500 Mali-Kayes-Fonio 0.5 8.6 22.6 24.8 25.6 26.0 26.1 26.2 26.3 26.4 Mali-Koulikoro-Fonio 0.1 10.1 31.3 38.4 42.0 44.4 45.5 46.2 46.6 47.7 Mali-Kayes-Maize 7.7 13.1 39.3 50.1 56.1 60.8 63.0 64.4 65.5 68.0 Mali-Koulikoro-Maize 27.3 1.8 6.6 10.9 14.1 17.2 19.0 20.2 21.3 24.3 Mali-Kayes-Mill 1.7 5.4 17.4 22.4 25.0 27.0 28.0 28.5 28.9 29.9 Mali-Koulikoro-Mill 12.3 4.0 13.9 20.6 25.0 29.1 31.3 32.8 33.9 37.2 Mali-Kayes-Rice 3.4 4.9 17.0 25.4 31.0 36.2 39.0 41.0 42.4 46.7 Mali-Koulikoro-Rice 19.9 8.2 28.0 40.2 48.1 55.0 58.8 61.3 63.0 68.3 Mali-Kayes-Sorghum 12.3 10.3 28.9 34.6 37.4 39.4 40.2 40.7 41.1 41.9 Mali-Koulikoro-Sor- 15.0 6.3 19.4 23.7 25.8 27.2 27.8 28.1 28.4 29.0 ghum Total 100.0 6.1 13.0 16.1 18.0 19.6 20.9 21.7 22.6 24.1 Source: Original table based on estimation results Note: All table numbers are expressed as percentages. 5.2 HOTSPOT #2 CROP RISK ASSESSMENT H 100 otspot #2 comprises the Cascades, FEWS NET also reports wild animals consuming Haut-Basis, Boucle du Mouhoun, crops, pest attacks, and excess of rains during Centre-Ouest, and Sud-Ouest Regions August as other important hazards affecting in Burkina Faso, the Upper West Region crop production in the hotspot. in Ghana, and the Zanzan and Savannes Regions in Côte d’Ivoire. This is a zone of Hotspot #2 accounts for 8.4 percent of the area rainfed agriculture and livestock rearing that and 3.4 percent of the exposures of the selected receives between 900–1000 millimeters of Admin 1 level food crop portfolio for West Africa. rainfall yearly. The region is relatively sparsely The selected food crops for risk assessment in populated, and people mainly consume cereals Hotspot #2 are rice, maize, sorghum, and millet. (sorghum, millet, and maize), followed by rice Maize and millet (accounting for 35 percent and and yams. 33 percent of exposures, respectively) are the main crops in the portfolio. Table 5.4 presents Drought, associated with delayed onset of the the crop area and the exposures of each crop rain season and lack of rain during the crops’ considered in the analysis. flowering season, is the main hazard affecting crop production in Hotspot #2 (USAID, 2010])19. 19 Livelihood Zoning and Profiling Report: Burkina Faso, USAID January 2010. TABLE 5-4 Expected Crop Area and Exposures for Selected Crops in Hotspot #2 Expected Crop Area Exposure Crop (hectares) % (US$, millions) 7.0% Maize 741,786 33.3% 377 35.1% Millet 518,833 23.3% 359 33.3% Rice 90,728 4.1% 76 7.1% Sorghum 876,667 39.3% 263 24.5% Total 2,228,013 100.0% 1,076 100.0% Total 1,742,266 100.0% 943 100.0% Source: Original table based on estimation results According to the risk assessment, the expected millions) once in 100 years or sustain losses as AAL of the selected food crops for Hotspot #2 is high as 18.1 percent of the exposure (or US$195 4.27 percent or US$46 million per year. The LaR million) once in 250 years. Table 5.5 shows the analysis indicates that the whole area selected expected LaR values estimated for the portfolio under the hotspot may expect to lose as high in Hotspot #2. as 15.9 percent of the exposure (or US$171 TABLE 5-5 Hotspot #2, Expected LaR Values Recurrence (years) 10 25 50 100 150 200 250 500 101 LaR (% exposure) 9.43% 12.07% 14.04% 15.86% 16.69% 17.60% 18.12% 19.77% LaR (US$, millions) 101.5 129.9 151.1 170.6 179.6 189.4 195.0 212.8 Source: Original table based on estimation results The volatility of the selected food crop portfolio an AAL of 5.41 percent of the total exposures, in Hotspot #2 is explained by the volatility of while the Hauts Basins Region in Burkina Faso maize and millet crops, which account for 35 presents an AAL of 4.96 percent of the total percent and 33 percent of the exposures in exposures. Considering a recurrence period of Hotspot #2, respectively. Maize crops present 100 years, the expected losses in these regions an AAL of 4.79 percent, while millet crops are estimated as 38.2 percent for the Upper present a AAL of 3.27 percent. The expected West region in Ghana and 27.5 percent in Haut losses for these crops for recurrence periods Basins Region in Burkina Faso. Table 5.6 presents of 1 in 100 years are 28.2 percent of the the contribution of each crop to the expected exposures for maize crops and 19.4 percent AAL and LaR for each crop and Admin 1 level in of the exposures for millet crops. The highest Hotspot #2. AALs were observed for the Upper West Region in Ghana and Hauts Basins Region in Burkina Faso. The Upper West Region in Ghana presents Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens TABLE 5-6 Expected Percentage AAL and LaR for Each Crop and Admin 1 Level Selected in the Portfolio in Hotspot #2 Exposure LaR (in percent for recurrence period years) Unit AAL % % 10 25 50 100 150 200 250 500 BF-Boucle du Mouhoun- 5.6 6.5 22.1 30.5 35.7 39.9 42.2 43.6 44.6 47.6 Maize BF-Cascades-Maize 4.4 2.9 9.9 14.0 16.6 18.8 19.9 20.7 21.2 22.8 BF-Centre-Ouest-Maize 0.6 3.5 11.7 16.3 19.2 21.8 23.2 24.1 24.8 27.0 BF-Hauts-Bassins-Maize 3.2 5.1 15.8 21.2 24.6 27.5 29.0 30.0 30.7 32.7 BF-Sud-Ouest-Maize 0.2 5.2 17.2 23.6 27.7 31.3 33.3 34.7 35.7 38.6 BF-Boucle du 3.7 3.0 9.9 13.5 15.6 17.5 18.5 19.2 19.7 21.2 Mouhoun-Millet BF-Cascades-Millet 1.5 7.6 25.9 36.9 43.8 49.9 53.1 55.3 56.8 61.3 BF-Centre-Ouest-Millet 2.5 4.2 13.9 19.2 22.6 25.6 27.2 28.3 29.2 31.6 BF-Hauts-Bassins-Millet 14.7 5.0 15.5 21.0 24.5 27.4 29.0 30.0 30.8 32.9 BF-Sud-Ouest-Millet 0.3 6.2 19.3 23.8 26.0 27.5 28.2 28.5 28.8 29.5 BF-Boucle du Mouhoun- 0.9 6.7 21.1 26.3 28.9 30.8 31.6 32.2 32.5 33.4 Rice BF-Cascades-Rice 0.1 5.7 19.1 26.5 31.0 34.7 36.7 37.9 38.9 41.4 BF-Centre-Ouest-Rice 3.1 7.9 26.6 37.1 43.5 49.1 51.9 53.8 55.1 59.0 BF-Hauts-Bassins-Rice 27.1 8.0 26.1 35.5 41.4 46.8 49.6 51.5 52.9 57.1 BF-Sud-Ouest-Rice 0.1 6.3 19.0 25.4 29.4 32.8 34.5 35.7 36.6 38.9 BF-Boucle du 0.4 4.0 13.0 17.5 20.3 22.7 24.0 24.8 25.5 27.3 Mouhoun-Sorghum 102 BF-Cascades-Sorghum 1.2 8.2 26.5 36.1 42.1 47.5 50.4 52.3 53.8 58.1 BF-Centre-Ouest-Sor- 1.9 3.4 12.1 19.1 24.2 29.0 31.8 33.8 35.3 39.6 ghum BF-Hauts-Bassins-Sor- 2.4 2.5 8.5 11.8 14.0 16.0 17.0 17.8 18.3 19.8 ghum BF-Sud-Ouest-Sorghum 1.2 3.5 10.8 14.3 16.4 18.2 19.1 19.7 20.1 21.3 Ghana-Upper West-Maize 2.3 4.0 14.0 21.1 25.9 30.4 32.8 34.5 35.9 39.5 Ghana-Upper West-Millet 1.4 3.5 11.7 15.7 18.3 20.5 21.7 22.4 23.0 24.7 Ghana-Upper West-Rice 4.1 15.1 48.5 62.3 69.8 75.6 78.3 80.0 81.1 84.3 Ghana-Upper West-Rice 4.1 15.1 48.5 62.3 69.8 75.6 78.3 80.0 81.1 84.3 Ghana-Upper West-Sor- 1.3 5.7 19.7 31.3 39.9 48.3 53.3 56.6 59.5 67.9 ghum Total 100.0 4.3 9.4 12.1 14.0 15.9 16.7 17.6 18.1 19.8 Source: Original table based on estimation results Note: All table numbers are expressed as percentages 5.3 HOTSPOT #2 CROP RISK ASSESSMENT H otspot #3 comprises the Tilabéri, flowering season, is the main hazard affecting Niamey, and Dosso Departments in crop production in Hotspot #3 according to Niger; the Est and Sahel Regions in FEWS NET (USAID FEWS NET 2010; USAID FEWS Burkina Faso; and the Atakora and Alibori NET 2011)21 22. The FEWS NET report also cites Departments in Benin . This is a zone of pest attacks, diseases, and floods as other rainfed agriculture that receives between important hazards affecting crop production in 400–600 millimeters of yearly rainfall. The the hotspot. zone’s population is of medium density. Fertility and crop yields vary according to Hotspot #3 accounts for 20 percent of the soil types and rainfall, with rainfall ranging area and 3.6 percent of the exposures of the from 400 millimeters per year in the north to selected Admin 1 level food crop portfolio more than 600 millimeters in the south. Millet for West Africa. The selected food crops for predominates where rainfall is lower and soils risk assessment in Hotspot #3 are rice, maize, are sandier, while substantial amounts of sorghum, and millet. Millet and sorghum sorghum are grown elsewhere. Livestock— (accounting for 55 percent and 45 percent of cattle, goats, and sheep—are an important exposures, respectively) are the main crops in source of cash income (particularly for wealthier the portfolio for Hotspot #3. Table 5.7 presents households) and the principal form of savings the crop area and the exposures of each crop or investment. considered in the analysis. 103 Drought, associated with dry spells during the crop season and lack of rain during the crops’ TABLE 5-7 Expected Crop Area and Exposures for Selected Crops in Hotspot #3 Expected Crop Area Exposure Crop (hectares) % (US$, millions) 7.0% Maize 86,895 33.3% 29 2.6% Millet 3,041,596 23.3% 627 55.3% Rice 22,667 4.1% 19 1.7% Sorghum 2,149,298 39.3% 459 40.5% Total 5,300,456 100.0% 1,134 100.0% Total 1,742,266 100.0% 943 100.0% Source: Original table based on estimation results 20 Departments in Benin has not been assessed as it was not possible to obtain a complete data set of crop production statistics for the period 2003–18. 21 Livelihood Zoning “Plus” Activity in Niger, USAID FEWS NET, August 2011. 22 Livelihood Zoning and Profiling Report: Burkina Faso, USAID FEWS NET, January 2010. Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens The analysis shows that the expected AAL of million) once in 100 years or losses as high the selected food crops for Hotspot #3 is 5.30 as 27.7 percent of the exposure (or US$314 percent or US$60 million per year. The LaR million) once in 250 years. Table 5.8 shows the analysis indicates that the whole area selected expected LaR values estimated for the portfolio under the hotspot may expect losses as high in Hotspot #3. as 24.4 percent of the exposure (or US$276 TABLE 5-8 Hotspot #3, Expected LaR Values Recurrence (years) 10 25 50 100 150 200 250 500 LaR (% exposure) 12.56% 17.56% 21.10% 24.37% 25.77% 27.07% 27.72% 29.67% LaR (US$, millions) 142.4 199.2 239.3 276.4 292.3 307.0 314.4 336.5 Source: Original table based on estimation results The volatility of the selected food crop portfolio the highest AALs in the region. The Sahel in Hotspot #3 is explained by the volatility of Region in Burkina Faso presents an AAL of 7.70 millet and sorghum crops, which account for percent of the total exposures while the Tilabéri 55.3 percent and 40.5 percent of exposures in Department in Niger presents an AAL of 5.64 Hotspot #3, respectively. Millet crops present percent of the total exposures. Considering an AAL of 4.6 percent, while sorghum crops a 100 year recurrence period, the expected 104 present an AAL of 6.0 percent. The expected losses in these regions are estimated to be losses for these crops for recurrence periods for 34.8 percent for Burkina Faso’s Sahel Region 1 in 100 years are 27.5 percent of the exposures and 33.5 percent for the Tilabéri Department for millet and 42.3 percent of the exposures for in Niger. Table 5.9 presents the contribution sorghum. The Sahel Region in Burkina Faso and of each crop to the expected AAL and LaR for the Tilabéri Department in Niger presented each crop and Admin 1 level in Hotspot #3. TABLE 5-9 Expected Percentage AAL and LaR for Each Crop and Admin 1 Level Selected in the Portfolio in Hotspot #2 Exposure LaR (in percent for recurrence period years) Unit AAL % % 10 25 50 100 150 200 250 500 BF-Boucle du 5.6 6.5 22.1 30.5 35.7 39.9 42.2 43.6 44.6 47.6 Mouhoun-Maize BF-Cascades-Maize 4.4 2.9 9.9 14.0 16.6 18.8 19.9 20.7 21.2 22.8 BF-Centre-Ouest- 0.6 3.5 11.7 16.3 19.2 21.8 23.2 24.1 24.8 27.0 Maize BF-Hauts-Bassins- 3.2 5.1 15.8 21.2 24.6 27.5 29.0 30.0 30.7 32.7 Maize BF-Sud-Ouest-Maize 0.2 5.2 17.2 23.6 27.7 31.3 33.3 34.7 35.7 38.6 BF-Boucle du 3.7 3.0 9.9 13.5 15.6 17.5 18.5 19.2 19.7 21.2 Mouhoun-Millet BF-Cascades-Millet 1.5 7.6 25.9 36.9 43.8 49.9 53.1 55.3 56.8 61.3 BF-Centre-Ouest- 2.5 4.2 13.9 19.2 22.6 25.6 27.2 28.3 29.2 31.6 Millet LaR (in percent for recurrence period years) Exposure Unit AAL % % 10 25 50 100 150 200 250 500 Niger-Dosso-Millet 24.4 2.6 8.9 12.8 15.3 17.5 18.7 19.5 20.1 21.8 Niger-Niamey-Millet 0.5 5.8 20.0 31.8 40.6 49.1 54.2 57.7 60.3 68.7 Niger-Tillabéri-Millet 22.4 5.6 18.4 25.3 29.7 33.5 35.6 37.0 38.1 41.2 Niger-Dosso-Maize 0.1 7.0 23.5 31.9 37.0 41.1 43.2 44.5 45.4 48.0 Niger-Dosso-Sorghum 0.4 11.4 38.0 51.5 59.6 66.3 69.7 71.7 73.3 77.6 Niger-Niamey-Sor- 18.5 6.4 22.3 35.3 45.0 54.5 60.1 63.8 66.8 76.3 ghum Niger-Tillabéri-Sor- 14.0 5.7 19.1 25.8 29.7 32.8 34.4 35.4 36.1 38.0 ghum Burkina Faso-Est-Maize 2.4 5.9 17.7 21.0 22.5 23.4 23.8 24.1 24.2 24.6 Burkina Faso-Sa- 0.1 12.9 40.9 51.8 57.5 61.8 63.7 64.9 65.8 67.9 hel-Maize Burkina Faso-Est-Millet 2.2 8.1 26.4 36.0 42.0 47.3 50.2 52.2 53.6 58.0 Burkina Faso-Sa- 6.0 8.4 25.9 31.7 34.5 36.5 37.3 37.8 38.2 39.0 hel-Millet Burkina Faso-Est-Rice 1.7 7.6 24.3 32.5 37.6 42.0 44.3 45.8 47.1 50.5 Burkina Faso-Est-Sor- 5.6 5.3 17.1 22.0 24.7 26.6 27.6 28.2 28.5 29.6 ghum Burkina Faso-Sa- 2.1 5.4 16.6 22.1 25.6 28.5 30.1 31.1 31.8 33.9 hel-Sorghum Total 100.0 5.1 12.46 17.34 20.73 24.33 25.56 27.04 27.60 30.23 Source: Original table based on estimation results Note: All table numbers are expressed as percentages 105 5.4 HOTSPOT #4 CROP RISK ASSESSMENT H otspot #4 comprises the Western, are identified as the main perils affecting crop Central, Ashanti, Eastern, and Brong production. Ahafo Regions in Ghana. The regions included in Hotspot #4 constitute Ghana’s Hotspot #4 accounts for only 5.8 percent of the breadbasket; apart from its extensive forest area but 32.8 percent of the total exposures of reserve and cocoa plantations, Ghana has a large the selected Admin 1 level food crop portfolio acreage in food crops. These regions enjoy a for West Africa. The selected food crops for heavy to moderate amount of rainfall. Generally, risk assessment in Hotspot #4 are rice, maize, there is a gradual decrease in rainfall from the cassava, yams, and plantains. Cassava and yams south, where the average annual rainfall is well (accounting for 47.4 percent and 40.2 percent over 1,650 millimeters, to the northern parts of of exposures, respectively) are the main crops in the region, where the average annual rainfall the portfolio for Hotspot #4. Table 5.10 presents is 1,100 millimeters. Crop production accounts the crop area and the exposures of each crop for 70 percent of regional agricultural output in considered in the analysis. Hotspot #423. Droughts, floods, and bush fires 23 Ministry of Food and Agriculture, Republic of Ghana. Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens TABLE 5-10 Expected Crop Area and Exposures for Selected Crops in Hotspot #4 Expected Crop Area Exposure Crop (hectares) % (US$, millions) 7.0% Cassava 531,263 34.6% 4,859 47.4% Maize 484,574 31.6% 263 2.6% Plantains 273,529 17.8% 942 9.2% Rice 51,011 3.3% 71 0.7% Yams 193,259 12.6% 4,117 40.2% Total 1,533,637 100.0% 10,252 100.0% Source: Original table based on estimation results The results of the risk assessment for Hotspot expect losses as high as 4.25 percent of the #4 present low risk compared with the region’s exposure (or US$572 millions) once in 100 years other hotspots, and the expected AAL of the or losses as high as 4.92 percent of the exposure selected food crops is 0.71 percent or US$95.5 (or US$661 million) once in 250 years. Table 5.11 million per year. The LaR analysis indicates that shows the expected LaR values estimated for the whole area selected under the hotspot may the portfolio in Hotspot #4. TABLE 5-11 Hotspot #4, Expected LaR Values 106 Recurrence (years) 10 25 50 100 150 200 250 500 LaR (% exposure) 2.12% 3.02% 3.67% 4.25% 4.55% 4.79% 4.92% 5.29% LaR (US$, millions) 285.4 406.7 493.4 571.6 611.6 643.9 660.9 711.6 Source: Original table based on estimation results The volatility of the selected food crop portfolio exposures. Considering a recurrence period of in Hotspot #4 is explained by the volatility of 100 years, the expected losses in these regions cassava and yams crops, accounting for 47.4 are estimated to be 20.1 percent in the Western percent and 40.2 percent of exposures in Region and 13.3 percent in the Central Region. Hotspot #4, respectively. Cassava crops present The observed AALs for rice and maize, although an AAL of 0.77 percent, while yams present a lower that the AALs for these crops in other AAL of 0.82 percent. The expected losses for hotspots, are well above the AALs observed for these crops for recurrence periods of 1 in 100 tuber crops (cassava and yams) and fruit crops years are 5.46 percent of the exposures for (plantains). According to the analysis, the AAL cassava and 5.58 percent of yam exposures. The for rice is 4.59 percent and for maize is 3.51 Western Region presents an AAL of 2.69 percent percent. Table 5.12 presents the contribution of the total exposures while the Central Region of each crop to the expected AAL and LaR for presents an AAL of 1.99 percent of the total each crop and Admin 1 level in Hotspot #4. TABLE 5-12 Expected AAL and LaR for Each Crop and Admin 1 Level Selected in the Portfolio in Hotspot #4 Exposure LaR (in percent for recurrence period years) Unit AAL % % 10 25 50 100 150 200 250 500 Ghana-Ashanti-Cas- 9.8 0.7 2.7 4.1 4.9 5.7 6.1 6.4 6.6 7.2 sava Ghana-Brong 14.0 0.5 1.8 2.8 3.4 4.0 4.3 4.5 4.7 5.2 Ahafo-Cassava Ghana-Central-Cas- 7.8 1.9 6.6 9.4 11.2 12.7 13.6 14.2 14.6 15.9 sava Ghana-Eastern-Cas- 17.6 0.0 0.0 0.0 0.0 0.1 0.2 0.3 0.3 0.4 sava Ghana-Greater 0.3 0.8 2.8 4.2 5.1 5.9 6.3 6.6 6.8 7.5 Accra-Cassava Ghana-Ashan- 0.5 6.1 16.6 18.4 19.1 19.4 19.6 19.6 19.7 19.8 ti-Maize Ghana-Brong Ahafo- 0.9 0.8 3.0 4.5 5.4 6.3 6.7 7.0 7.3 7.9 Maize Ghana-Cen- 0.4 3.7 13.1 19.2 23.2 26.7 28.6 29.9 30.9 33.7 tral-Maize Ghana-East- 1.0 0.5 2.1 3.2 3.9 4.5 4.9 5.1 5.3 5.8 ern-Maize Ghana-Greater 0.0 0.1 0.4 0.9 1.3 1.6 1.7 1.9 1.9 2.2 Accra-Maize Ghana-West- 0.1 7.8 25.9 34.6 39.6 43.7 45.6 46.9 47.8 50.1 107 ern-Maize Ghana-Ashanti-Plan- 2.8 0.1 0.5 1.0 1.4 1.7 1.9 2.0 2.1 2.3 tains Ghana-Brong 2.2 1.3 4.7 6.8 8.1 9.3 9.9 10.4 10.7 11.6 Ahafo-Plantains Ghana-Central-Plan- 0.4 1.6 5.9 8.6 10.3 11.7 12.5 13.0 13.3 14.3 tains Ghana-Eastern-Plan- 1.8 0.1 0.3 0.7 1.0 1.3 1.5 1.6 1.6 1.9 tains Ghana-West- 1.5 1.5 5.3 8.1 9.9 11.5 12.4 13.0 13.4 14.7 ern-Plantains Ghana-Ashanti-Rice 0.2 7.9 26.3 35.4 40.7 45.0 47.2 48.5 49.6 52.3 Ghana-Brong Ahafo- 0.1 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Rice Ghana-Central-Rice 0.0 7.9 25.6 34.9 40.9 46.2 49.0 50.9 52.4 56.5 Ghana-Eastern-Rice 0.2 2.6 8.7 12.2 14.4 16.4 17.5 18.2 18.8 20.4 Ghana-Greater 0.2 4.3 14.2 19.7 23.1 26.2 27.9 29.0 29.9 32.4 Accra-Rice Ghana-Western-Rice 0.2 3.8 12.7 17.6 20.7 23.5 25.0 26.0 26.8 29.0 Ghana-Ashanti-Yams 5.3 0.8 2.9 4.3 5.3 6.1 6.5 6.8 7.1 7.7 Ghana-Brong Ahafo- 23.3 0.6 2.0 2.8 3.4 3.8 4.1 4.3 4.4 4.8 Yams Ghana-Central-Yams 0.2 1.8 6.4 9.0 10.7 12.3 13.1 13.6 14.0 15.3 Ghana-Eastern-Yams 8.3 1.1 3.9 5.6 6.7 7.8 8.3 8.7 9.0 9.8 Ghana-West- 1.0 3.0 10.5 15.0 17.8 20.2 21.4 22.2 22.8 24.5 ern-Yams Total 100.0 0.71 2.12 3.02 3.67 4.25 4.55 4.79 4.92 5.2 Source: Original table based on estimation results Note: All table numbers are expressed as percentages A Blueprint for Strengthening Food System Resilience in West Africa: Regional Priority Intervention Areas BOX 5-1 Tools Used to Assess Crop Losses for this Report The Administrative Level 0 and 1 Stochastic Crop Risk Assessment Tool and the Administrative Level 0 and 1 HBCA Crop Risk Assessment Tool, delivered jointly to the study report, are especially useful tools to assist users and policymakers in assessing food crop risk in West Africa and to be used as a basis for designing risk financing strategies. The risk assessment analysis for main food crops in West Africa assesses the risk in two dimensions. The first dimension is the expected AAL. The AAL is an indicator of the loss that can be expected in each year for each crop. The second dimension is the LaR. The LaR is a key measure used to infer the size of a loss that could be reached with a given probability or return pe- riod. These tools allow users to perform actuarial analysis to estimate key risk portfolio parameters such as the expected AAL and the PML, either on an experience-based approach or based on a stochastic approach. These tools are also flexible enough to enable users to perform the analysis with different breakdown levels (by crop, by country, by region) and to combine different units. Source: World Bank [1] Livelihood Zoning and Profiling Report: Mali, USAID January 2010 [2] Livelihood Zoning and Profiling Report: Burkina Faso, USAID January 2010 108 [3] Departments in Benin has not been assessed as it was not possible to obtain a complete data set of crop production statistics for the period 2003–18. [4] Livelihood Zoning “Plus” Activity in Niger, USAID FEWS NET August 2011 [5] Livelihood Zoning and Profiling Report: Burkina Faso, USAID FEWS NET January 2010 [6] Ministry of Food and Agriculture, Republic of Ghana. Key points in this 6 chapter: FOOD PRICE RISKS AND MARKET • Weather risk mitigation and political stability in West Africa are necessary to ensure food INTEGRATION IN price stability and food system resilience and to further improve food security. Analysis in this chapter shows that there are significant WEST AFRICA effects on food prices from climate shocks and social unrest. • The level of impacts on prices can vary depending on the type of crops and countries; 109 This chapter assesses whether weather- for example, prices for millet, sorghum, and related risks and conflicts affect the region’s yam are more susceptible to impacts from food prices and explains why risk events other risks. Similarly, prices in Chad, Mali, and in one country may impact food prices in Niger are sensitive to outside shocks. another country. It also analyzes the level of • Market integration analysis reveals potential food market integration between countries benefits on price levels from trade within West in West Africa and how they are affected by Africa depending on the type of crops. different outside events. • Improved risk management that reduces the price of millet in Senegal could benefit households in Burkina Faso, Mali, and Niger by reducing the price level in those countries. • Rice market analysis also reveals that Senegal has a significant effect on prices in the region. • Improved risk management in Mali that reduces sorghum prices could also decrease sorghum prices in Togo and Chad. Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens 6.1 FOOD PRICE RISKS IN WEST AFRICA P rice volatility acts as a signal for buyers West African countries, it is not uncommon for and sellers in a market, impacting both people to spend up to half of their money on consumers and producers. Food price food, so even small price changes can have a volatility is a directionless measure of the extent significant impact. of food price variability and is mostly captured by the standard deviation of logarithmic Several direct and indirect causes can price changes on a monthly average (Gilbert, explain the volatility of food prices in C. L., & Morgan, C. W. 2010). High and rapid most net food-importing countries, such price fluctuations can destabilize economies, as those in West Africa. Table 6.1 shows the particularly in West Africa, where producers international and national sources of food and consumers often have limited access to price risk for small, highly importing countries. credit and insurance, two key protections The main national risk factors appear to be against volatility. For small-scale producers, domestic supply and regional cross-border price volatility makes investments uncertain trade, while the international risk factors are and risky (Benzie 2015). Returns are difficult changes in international supply and demand to predict when prices change dramatically, and market dynamics. Focusing on national so farmers often try to remain flexible to risks, domestic supply in West Africa may be switch from lower-priced crops to higher- negatively affected by factors such as climate priced crops—for example, when markets change and social unrest. Thus, the analysis 110 change. This may benefit them in the short focuses on the impact of certain climate shocks term but discourages investments that could and insecurity events on food price levels. In improve their livelihoods in the long term. For addition, regional cross-border trade will be consumers, the volatility of staple food prices assessed by analyzing market integration. can effectively reduce household income. In TABLE 6-1 Differences Between International and National Sources of Food Price Risks for Small, Highly Import- Dependent Countries. International National • Changes in international supply • Domestic supply (harvest, and Principal risk • Changes in international demand processing capacity) drivers • Market dynamics (perception, buyer • Regional cross-border trade behavior, exporter and importer policy responses) National Global to national Scope and direc- • International flows directly influence • Domestic flows have little or no influence on international prices in highly tion of impacts domestic price (the degree of price import-dependent countries transmission varies between markets) Risk type Likelihood: LOW Likelihood: MEDIUM Magnitude: HIGH Magnitude: LOW to MEDIUM International National Moderate • Able to directly influence agriculture policy (driver of productivity) and policies Importing coun- None to facilitate food processing sector (which try government’s • Importers are at the mercy of convert domestic agriculture production international drivers and responses from influence on or into food commodities) but not to directly other buyers (for example, other national control of risk control prices at the farm gate rice boards) • Unable to control climate variables, market price of inputs, and so on Source: Stockholm Environment Institute 2015 6.1.1 The impacts of weather-related and other risks on crop prices F ood security is a crucial issue in West Climate change increases interannual Africa, and climate change will increase variability, negatively affecting production food insecurity. Higher temperatures of major food crops such as wheat, rice, and and changes in rainfall induced by global maize. In individual countries, climate change warming are threatening rainfed agriculture in directly impacts domestic agriculture, leading 111 this region. The Sudano-Sahelian zone is highly to price spikes in local markets and occasionally vulnerable to climate hazards. It is characterized to shortages of domestically-produced food. by a monsoon season stretching between Countries with agricultural systems that are May and September, which concentrates more exposed to these direct impacts and that most of the annual precipitation (between are otherwise less able to adapt will be most 200 and above 1,200 millimeters per year). vulnerable. Combined with other underlying The region’s rural population depends on this drivers, such as rapid population growth and rainfall season for subsistence rainfed farming urbanization, climate change will also add to and pastoralism (Ickowicz et al. 2012). Maize, the volatility of future agricultural commodity millet, and sorghum play a very important markets and worsen the effect of price shocks role in West Africa, accounting for 60 percent on import-dependent countries. to 98 percent of total crop production for the 2008–17 period for Burkina Faso, Mali, Niger, Nigeria, and Senegal (FAO database). Climate change and its impacts on agriculture could reduce crop yield and crop production, with high uncertainty for the twenty-first century (Kummu et al. 2021). When weather conditions reduce agricultural production over large areas, the resulting widespread supply reductions cause significant increases in local food prices (Brown et al. 2015). Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens 6.1.2 Impact of social unrest on food prices C onflict exacerbates the conditions Civil wars in Côte d’Ivoire or in Sierra Leone, leading to malnutrition, such as high for instance, contributed to many deaths and food prices, inadequate household migration that reduced the level of human food security, and poor diet (FAO 2002). Figure labor in many sectors such as agriculture. In 6.1 shows conflict events in West Africa, mainly Côte d’Ivoire, the percentage of children under in Nigeria, Sierra Leone, Côte d’Ivoire, Mali, and five suffering from stunting increased from 30 Burkina Faso. Conflicts reduced average daily percent in 2004 to 32.9 percent in 2006, a period calories by 7 percent, with a reduction in overall corresponding to the civil war that divided the food supply and an increase in food prices. country into two halves (Saumik 2015). FIGURE 6-1 Number of Conflict Events (1997 to 2021) 20000 18000 16000 14000 12000 112 10000 8000 6000 4000 2000 0 Benin Burkina Chad Côte Gambia, Ghana Guinea Guinea- Liberia Mali Mauritania Niger Nigeria Senegal Sierra Togo Faso d'Ivoire The Bissau Leone Battles Explosions Protests Riots Strategic development Violence agains civils Source: Original figure based on ACLED database Some conflicts can result from high 200 deaths due to worsening living conditions. commodities prices, and Guinea, Mauritania, Demonstrations occurred sporadically and Senegal had food riots during the throughout 2007, and in mid-February 2008, 2007/08 food crisis (UN New Centre 2008). rioting broke out in Ratoma caused by high In Guinea, in January and February 2007, food prices (Harsh 2008). In Mauritania in antigovernment demonstrations caused nearly November 2007, bread riots lasted for 10 days. Hungry people looted food stores and with over a thousand people demonstrating attacked police stations in Nouakchott. There in Dakar over high living costs (Bush 2010). were at least nine other towns where protests Three Burkina Faso cities—Bobo-Dioulasso, raged, and at least one demonstrator was killed Ouahigouya, and Banfora—had food riots in during the protests (Bush 2010). In late 2007 in February 2008 due to high food prices and the Senegal, unions and civil society groups were government collecting taxes from small-scale active in the demonstrations. On the March merchants (Harsh 2008). Food price volatility 30, 2008, in Dakar, two consumers’ associations thus tends to be a politically sensitive issue called for a march with the slogans “we are for both urban and rural communities. Most hungry” and “life has become difficult” against governments in developing countries see rising prices. In April 2008, more protests were domestic food price stability as a priority for pushed by some of the country’s trade unions food security and also to maintain social and (Harsh 2008), caused by curbs to informal street political stability. hawking and leading to large street protests, CASSAVA FIGURE 6-2 Cassava Retail Price in US$ per kilogram in Benin, Côte d’Ivoire, Liberia and Sierra Leone from Jan 2005 to Sept 2020 Cassava price 113 .6 Retail price (USD/kg) .2 0 .4 2005m1 2010m1 2015m1 2020m1 Benin CIV Liberia Sierra Leone Source: Original figure based on data from FAOSTAT and WFP-VAM Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens According to stationarity tests and availability on price levels. In Côte d’Ivoire, for example, of data, the vector autoregression (VAR)24 a 1 percent increase in the level of conflict model was performed with Benin, Côte d’Ivoire, leads to a 0.8 percent increase in the price of Liberia, and Sierra Leone cassava prices. The cassava, and a 1 percent increase in the level figure describes the trends in cassava prices of protest contributes to a 0.4 percent cassava from January 2005 to September 2020 for the price increase. Rainfall and rainfall shocks have four countries considered for our analysis. On a negative effect on prices, while temperature average, there is a high volatility in food prices and heat shocks lead to higher prices, mainly mainly for Côte d’Ivoire (CIV), considered to in Liberia and Sierra Leone, with no significant be one of the biggest cassava producers and effect in Benin. This finding could be because consumers in the region (see table B-4 and rainfall helps improve cassava production table B-5 in appendix B). The result of the and market availability, but an increase in VAR model (table 6 in the appendix B) allows temperature depresses cassava production and one to conclude that for cassava, insecurity availability. variables considered, such as riots, conflicts and protests, have positive and significant effects Yam FIGURE 6-3 Yam Retail Price in US$ Per Kilogram in Benin, Côte d’Ivoire, Ghana and Nigeria from July 2011 to July 2020 Yam price 2 114 1.5 Retail price (USD/kg) .5 0 1 2011m7 2014m7 2017m7 2020m7 Benin CIV Ghana Nigeria Source: Original figure based on data from FAOSTAT and WFP-VAM 24 The methodology for estimating the impact of events (policy or shocks) on the levels and variability of food prices in the short run relies on VAR models. VAR models are useful for estimating effects of policies or shocks on economic variables, especially where there is uncertainty about the correct structural model or where data is limited and some variables are unavailable (Sims 1980; Jayne et al. 2008; Fackler 1988). Prices for food commodities are often nonstationary (nonconstant mean and infinite variance), which makes difficult the estimation of relationships between prices. Therefore, the Augmented Dickey Fuller (ADF) unit root test was performed to check the stationarity of the data, and to see if they were integrated of order 1. The ADF test is applied using a lag length that minimizes the information criterion (Schwarz Bayesian Information Criterion [SBIC], Hannan Quinn Information Criterion [HQIC], Akaike Information Criterion [AIC] and Final Prediction Error Criterion [FPE]). The data considered in this section met the conditions for applying an VAR model. The yam price analysis was estimated with also called strategic development 25 in the prices from Benin, Côte d’Ivoire, Ghana, and ACLED database) have, on average, positive Nigeria. The attached figure shows the trends and significant effects on the price levels of in yam prices from July 2011 to July 2020 for yams. For example, a 1 percent increase in the four countries considered for our analysis. disturbance and property destruction arrests, On average, there was high volatility in food looting or destruction of property, disrupted prices and high prices mainly for Côte d’Ivoire weapons use, and so on) in Nigeria or conflicts (CIV) and Nigeria, considered as two of the in Benin, respectively, contributes to a 0.8 largest yam producers in the region (see table percent and 5 percent increase in the price B-4 and table B-5 in appendix B). The results level of yams (all else being constant). We did of the VAR model (table B-7 in appendix B) not find a significant effect for the climate and allow us to conclude that the security variables climatic shock variables. considered (conflicts, riots, and other disorders, maize FIGURE 6-4 Maize Retail Price in US$ per kilogram (kg) in Chad, Côte d’Ivoire, Guinea-Bissau, Mali, Senegal, Togo from Jan 2000 to Sept 2020 Maize 3 115 Retail price (USD/KG) 1 0 2 2000m1 2005m1 2010m1 2015m1 2020m1 Chad CIV GB Mali Senegal Togo Source: Original figure based on data from FAOSTAT and WFP-VAM Figure 6.4 shows a large variation in maize prices test, we considered the price of maize in Chad, in most countries, with a huge price variation Côte d’Ivoire, Guinea-Bissau, Mali, Senegal, and mainly in Guinea-Bissau. Due to the high level of Togo for this analysis. The results (table B-10 in missing values and the results of the stationarity the appendix B) conclude that an increase in 25 ACLED includes some activity that can broadly be described as nonviolent but differs in its role within contexts of disorder. These events, named Strategic Developments, include incidences of looting, peace talks, high-profile arrests, nonviolent transfers of territory, recruitment into nonstate groups, and so on, and accounts for a small proportion of the total dataset. These common events suggest the context of disorder. Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens the level of conflict and rainfall in Mali increases destruction and reduces local maize supply. the price level of maize of 0.2 percent and 0.02 Increases in the level of maize in Togo and Chad percent, respectively. Conflicts by their nature lead to rising maize prices in Mali due to trade can destroy crops or limit the ability of farmers relationships—also because Togo is one of the to work in the fields. The effect of rainfall could highest maize producers in the region. be due to heavy rainfall that contributes to crop millet FIGURE 6-5 Millet Retail Price in US$ Per Kilogram in Burkina Faso, Chad, Mali, Niger and Senegal from Jan 1990 to Aug 2020 to Sept 2020 Millet price .8 Retail price (USD/kg) .2 .4 .6 116 0 1990m1 2000m1 2010m1 2020m1 Burkina Faso Chad Mali Niger Senegal Source: Original figure based on data from FAOSTAT and WFP-VAM Figure 6.5 displays a strong variation in millet temperatures, violence against the civilian prices with a similar trend for the countries population in Chad (for example, terrorism), and considered (Burkina Faso, Chad, Mali, Niger, violence in Burkina Faso contribute to higher and Senegal). These Sahelian countries are price levels in Chad because they primarily the most important for millet production and affect trade and reduce the availability of local they are also part of a specific corridor in West maize. Africa. They were also considered for the VAR analysis because they meet the nonstationarity criteria. According to the results presented in table B-8 (in the appendix B) the price of millet in Niger has a positive and significant effect on the price level in Chad and Burkina Faso. High rice FIGURE 6-6 Rice Retail Price in US$ Per Kg in Chad, Côte d’Ivoire (CIV), Ghana, Guinea, Guinea-Bissau (GB), Mali, Senegal and Sierra Leone (SL), from Jan 2004 to Sept 2020 Rice price 2.5 Retail price (USD/kg) 2 1.5 1 .5 0 2002m1 2004m1 2006m1 2008m1 2010m1 2012m1 2014m1 2016m1 2018m1 2020m1 Chad CIV Guinea GB Mali Ghana Senegal SL 117 Source: Original figure based on data from FAOSTAT and WFP-VAM Figure 6.6 shows a large variation in the price leads to significant rice price increases in in the of rice in most countries analyzed. There was price level in Senegal leads to significant rice a large increase in the price of rice in Chad price increases in in the price level in Senegal in 2020 that could be attributed to COVID-19 leads to significant rice price increases in in the impacts. The result of the VAR model shows price level in Senegal leads to significant rice that in the short run, there is a positive and price increases in Guinea-Bissau and Mali. significant effect of rice price in Côte d’Ivoire on Mali and Guinea. For instance, a 1 percent increase in the level of rice price in Côte d’Ivoire leads to price increases of, respectively, 17 percent and 9 percent in Guinea and Mali. Due to trade relationships, in the short run, a 1 percent increase in the price level in Senegal Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens sorghum FIGURE 6-7 Sorghum Retail Price in US$ Per Kilogram in Chad, Mali, Nigeria and Togo from Jan 2000 to Sept 2020 Sorghum price 2.5 Retail price (USD/kg) .5 1 0 1.5 2 1990m1 2000m1 2010m1 2020m1 Benin Chad Gambia Ghana Guinea-Bissau Mali Niger Nigeria Senegal Togo 118 Source: Original figure based on data from FAOSTAT and WFP-VAM There is a large variation in the price of sorghum a positive and significant effect on sorghum in most West African countries, with a huge prices in Mali. Due to the trade relationship price mainly in Guinea-Bissau (figure 6.7). Due to between Mali and Chad, we also see from the the high level of missing values and the results results that an increase in the price of sorghum of the stationarity test, we considered the price in Chad favors an increase in the price in Mali. of sorghum in Chad, Mali, Nigeria and Togo for this analysis. The results displayed in table B-7 (appendix B) conclude that an increase in the level of conflict and temperature in Mali has 6.2 MARKET INTEGRATION IN WEST AFRICA M arkets and trade are essential so united by the relations of unrestricted trade for the proper functioning and that prices take the same level everywhere development of economic systems. with ease and rapidity” (Cournot 1971). In other Economic historians have therefore shown words, the equilibrium level of prices must be a keen interest in markets and trade and in equal (law of one price) and prices must easily market integration. An integrated market can and quickly return to this level after any shock. be defined as “a whole territory whose parts are 6.2.1 Evolution of foreign trade and intraregional trade in West Africa T he Africa Regional Integration Index member state in an REC). Free movement of constructed by the United Nations people is constituted by ratification (or not) of Economic Commission for Africa the REC protocol on free movement of persons; (UNECA) considers several dimensions for proportion of REC member countries whose each regional economic community (REC) nationals do not require aW visa for entry; and 119 in a comprehensive manner and reveals proportion of REC member countries whose that ECOWAS appears to be one of the most nationals are issued a visa on arrival. Finally, important RECs in Africa. The first dimension, financial and macroeconomic integration trade integration, includes several indicators includes regional convertibility of national such as level of customs duties on imports, currencies and inflation rate differential (based share of intraregional goods exports (percent on the harmonized consumer price index). GDP), share of intraregional goods imports (percent GDP), and share of total intraregional Although the ECOWAS region is considered goods trade. The second, regional infrastructure, to be one of the continent's subregions, the considers the infrastructure development level of intraregional food trade is lowbased index (transport; electricity; information and on the table below and previous reports by communications technology; water and RESAKSS or UNCTAD. Indeed, the table below sanitation); proportion of intraregional flights; highlights West Africa’s heavy dependence total regional electricity trade (net) per capita; on food imports from outside the region, as and average cost of roaming. Productive few commodities are traded internally. In this integration considers the share of intraregional section, market integration was analyzed in intermediate goods exports (percent of the region through a few specific food crops total intraregional exports goods); share of (maize, millet, rice, sorghum, yams, and cassava) intraregional intermediate goods imports and the factors affecting it. The importance of (percent of total intraregional imports goods); agriculture is reflected in the growing share and merchandise trade complementarity index of agricultural exports in West Africa’s foreign (total absolute value of the difference between trade. Agricultural exports (mainly cash crops) share of imports and share of exports of a for all countries combined account for nearly Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens US$6 billion, or 16.3 percent of the region’s total atypical due to oil, regional agricultural exports exports of goods and services. Yet the region exceed imports, representing 30 percent and imports US$5.4 billion worth of food products. 10 percent, respectively. Excluding Nigeria, whose export structure is TABLE 6-2 Africa Regional Integration Index 2015 Index COMESA ECCAS SADC AMU ECOWAS ECA Average Trade integration 0.57 0.53 0.51 0.63 0.44 0.78 0.58 Regional infrastructure 0.44 0.45 0.5 0.49 0.43 0.5 0.47 Productive integration 0.45 0.29 0.35 0.48 0.27 0.55 0.40 Free movement of people 0.27 0.4 0.53 0.49 0.8 0.73 0.54 Financial & macro integration 0.34 0.6 0.4 0.2 0.61 0.16 0.39 Average 0.41 0.45 0.46 0.46 0.51 0.54 Source: Original figure based on data from FAOSTAT and WFP-VAM TABLE 6-3 Share of Intraregional and Extraregional Food Trade Share of intraregional and extraregional food trade for the biggest crop consumed (average quantity in tons from 2016 to 2018) 120 share imported share imported Country Crop Production Import (intraregional) (extraregional) Maize 1,446,908 961 0% 100% Benin Rice 339,820 9,144 0% 100% Sorghum 1,653,192 92 100% 0% Burkina Faso Millet 974,128 244 100% 0% Cassava 4,971,889 0 0% Côte d’Ivoire Yam 7,098,524 0 0% Maize 4,837 35,599 0% 100% Cabo Verde Rice 0 84 0% 100% Yam 7,718,488 0% Ghana Maize 2,013,158 64,809 12% 88% Rice 53,542 3 0% 100% Gambia, The Millet 100,955 0% 100% Rice 2,224,632 2,028 48% 52% Guinea Cassava 1,368,098 0% Rice 180,930 4 0% 100% Guinea-Bissau Maize 7,261 59 0% 100% Rice 280,223 2 0% 100% Liberia Cassava 604,789 0% Rice 2,909,779 546 92% 8% Mali Millet 1,817,813 1 0% 100% Rice 235,167 83 14% 86% Mauritania Millet 3,261 0% 100% Sorghum 1,951,196 7,392 60% 40% Niger Millet 3,844,150 4,549 60% 40% Rice 6,993,694 3,369 0% 100% Nigeria Cassava 59,400,000 0% Groundnuts 826,673 27 100% 0% Senegal Rice 702,470 350 0% 100% Rice 897,392 181 8% 92% Sierra Leone Cassava 22,423 0% Maize 856,071 819 1% 99% Togo Rice 141,038 27 0% 100% Sorghum 974,966 25,000 0% 100% Chad Millet 714,156 0 0% 121 Source: Original table based on FAO database 6.2.2 Food markets are integrated between countries T here are many more net buyers than of hungry people (Wodon et al. 2008) while net sellers, so integration into world boosting the income of only a few relatively markets has generally improved food well-off farmers. security in West Africa. Integration into world markets benefits net buyers, as it lowers peak Integrated markets could enhance food prices while decreasing the profits of net security because spatial market integration sellers. Yet this makes local food prices both allows food to move from surplus to deficit dependent on and vulnerable to shocks from regions, thus improving food availability the world markets. For example, the increase and access across space. Most households rely in global corn (maize) prices in 2008 had a on food markets; therefore, integrating those strong influence on the price of grain in the markets is a major determinant of their ability rural, informal commodity markets in West to meet their food needs. Following these Africa, despite their relative isolation. The positive effects, market integration could also period of high global prices had the immediate help mitigate the impacts of climate change consequence of sharply increasing the number on food and nutritional security. The extent Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens to which markets are integrated defines the are higher than the price difference between boundaries of a food price stabilization policy. markets, arbitrage possibilities are eliminated, trade does not occur, and market integration Yet market integration could be affected by is compromised. Finally, insecurity negatively the management of domestic food prices affects trade and then market integration. The policies, trade transaction costs, poor results below indicate, on average, a low level transport infrastructure, climate change, of market integration in West Africa, mainly due and security issues. Transaction costs often to the causes mentioned above. However, for inhibit exchange and, therefore, hinder market countries with some level of integration, social integration. A deficient transport infrastructure unrest, such as conflict in one country, can (road, rail, and air) is associated with relatively affect another, leading to long run variation in higher transport costs. Poor transport food prices. Figure 6.8 shows how retail prices infrastructures represent a higher trade barrier of cereals (maize, millet, rice, and sorghum) vary than import tariffs or other trade restrictions and across 12 countries in the region, highlighting stifles market integration. When transport costs the constraints of food trade. FIGURE 6-8 Average Retail Prices for Cereals (Maize, Millet, Rice, and Sorghum) Across the Countries, US$ Per Kilogram, January 2000 to September 2020 2.5000 122 2.0000 1.5000 1.0000 0.5000 0.0000 Benin Burkina Chad Côte Ghana Guinea Guinea- Mali Niger Nigeria Senegal Sierra Faso d'Ivoire Bissau Leone Maize Rice Sorghum Millet Source: Original figure based on data from FAOSTAT, WFP-VAM, FEWS NET, and FAO-GIEWS For cassava, there is a long run relationship and that in the long run, a shock that increases between Benin, Côte d’Ivoire, Liberia, and the price of cassava in Sierra Leone and Côte Sierra Leone. This is illustrated in map 6.1 d’Ivoire will lead to a decrease in the price in (in which the darker color indicates strong Benin and Liberia, respectively. However, there integration). Specifically, there is a significant is no long run effect of other markets on Sierra negative long run relationship between Sierra Leone and Côte d’Ivoire, and in the long run, Leone and Benin, Sierra Leone and Liberia, and there is a significant negative effect of social Côte d’Ivoire and Liberia. This could lead to the unrest variables and weather shocks on the conclusion that these markets are integrated level of integration between these four markets. MAP 6-1 West Africa Market Integration, Cassava Level of integration no long-run relationship Level of integration Long-run relationship no long-run relationship Long-run relationship Source: Original map based on estimation results Market integration analysis: Cassava For yam, market integration show Cassava results analysis: Market integration Nigeria. This could lead to the conclusion that a long run relationship between Benin, these markets are integrated and that in the Côte d’Ivoire, Ghana, and Nigeria (map 6.2). long run, a shock that increases the price level There is a negative and significant long run of yam in Nigeria, Ghana, and Benin will increase 123 relationship between Côte d’Ivoire and Nigeria the price in Ghana, Côte d’Ivoire, and Nigeria, and between Côte d’Ivoire and Benin, while a respectively. There is also a significant effect of positive and significant effect exists between social unrest variables and climate shocks on Benin and Nigeria and between Ghana and the level of integration between these markets. MAP 6-2 West Africa Market Integration, Yam Level of integration No Long-run relationship Long-run relationship Level of integration No Long-run relationship Long-run relationship Source: Original map based on estimation results Market integration analysis: Yam Market integration analysis: Yam Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens For maize, the market integration analysis between Côte d’Ivoire and Guinea-Bissau, reveals a long run relationship between Senegal, and Togo. In the long run, there is a Chad, Côte d’Ivoire, Mali, Guinea-Bissau, significant effect of social unrest variables and Senegal, and Togo (map 6.3). For example, climatic shocks on the level of integration there is a positive and significant long run among these six markets. For instance, due to relationship in the maize prices from Togo to the trade relationship, a protest that happens Chad, Côte d’Ivoire, Guinea-Bissau and Mali, in Chad leads to a positive and significant effect while a negative and significant effect exists in Togo. MAP 6-3 West Africa Market Integration, Maize 124 Level of integration No long-run relationship Long-run relationship Level of integration No long-run relationship Source: Original map based on estimation results Long-run relationship Market integration analysis: Maize For millet, the analysis displays a long run Senegal, and Mali. Social unrest variables and relationship between Burkina Faso, Chad, analysis: Market integration Maizeshocks show a significant, long run climate Mali, Niger, and Senegal (map 6.4). There is effect in the level of integration between these a negative and significant long run relationship five markets. between Senegal and Chad, Niger and Burkina Faso, and Chad and Niger, while there is a For rice, the analysis reveals a long run positive and significant effect between Chad relationship between Chad, Côte d’Ivoire, and Senegal, Chad and Burkina Faso, Mali and Ghana, Guinea, Guinea-Bissau, Mali, Chad, Senegal and Burkina Faso, Senegal and Senegal, and Sierra Leone (map 6.5). For Mali, and Senegal and Niger. The results could instance, table B-16 (appendix B) shows a lead one to conclude that these markets are positive and significant long run relationship integrated, and in the long run, a shock that will between Senegal and Guinea, Mali and Guinea- increase the level of millet price in Mali, Burkina Bissau, and Côte d’Ivoire and Sierra Leone, Faso , Niger, Chad, and Senegal will respectively while there is a negative and significant effect raise the price in Chad, Niger, Burkina Faso, between Ghana, Chad, and Côte d’Ivoire. These MAP 6-4 West Africa Market Integration, Millet Level of integration No long-run relationship Long-run relationship Level of integration No long-run relationship Source: Original map based on estimation results Market integration analysis: Millet Long-run relationship results could lead to conclusions that these that, in the long run, there are significant and Market integration analysis: Maize markets are integrated and in the long run, a positive effects of social unrest variables, mainly shock that will increase rice prices in Senegal conflict, on the level of integration in these 125 will raise the price in its neighbors Guinea, eight markets. Guinea-Bissau, and Mali. We could also observe MAP 6-5 West Africa Market Integration, Rice Level of integration No long-run relationship Long-run relationship Level of integration Source: Market Original map based integration Rice results on estimation analysis: No long-run relationship Long-run relationship Market integration analysis: Maize Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens For sorghum, the market integration could lead one to conclude that these markets analysis indicates a long run relationship are integrated and in the long run. For instance, between Chad, Mali, Nigeria, and Togo a shock that increases sorghum prices in Mali (map 6.6). Specifically, there is a positive and will raise the price in Togo. In the long run, there significant long run relationship between Togo are significant effect of social unrest variables and Chad, Mali and Togo, and Chad and Mali, and climate shocks in the level of integration while there is a negative and significant effect between these four markets. For instance, between Nigeria and Togo. Given the high due to the trade relationship, civil violence production of sorghum in Nigeria compared that happens in Mali leads to a positive and with Togo, the long run impact of Nigeria in significant effect in Togo. Togo is higher than in the reverse. These results MAP 6-6 West Africa Market Integration, Sorghum 126 Level of integration No long-run relationship Long-run relationship Level of integration Source: Original map based on estimation results No long-run relationship Market integration analysis: Sorghum Long-run relationship Food markets are integrated between Regional market information systems countries, but the combination of countries need Market integration analysis: to be upgraded and harmonized to Maize varies depending on crops. As this section reduce market price volatility and facilitate shows, there is no static set of countries for cross-border trade (World Bank and FAO which markets are integrated; rather, this 2021). MIS provide publicly available data on depends on the food crop. Yet a few countries prices and traded volumes to market actors to (Chad, Côte d’Ivoire, Mali, and Senegal) show enhance decision-making. However, limited higher levels of integration than the other capacity to monitor market information and countries assessed. Similarly, price variations for low uptake by users continue to hold back millet, yam, and sorghum result in higher price system effectiveness at the regional level. Going transmission across the region than for rice, forward, the challenge is how to harmonize maize, and cassava. existing systems across countries (World Bank and FAO 2021) (for example, information on stock levels, inputs, and so on) through an integrated MIS and beef up monitoring capacity at the regional level. One entry point for intervention could be dispatching innovators from the regional start-up landscape to regional bodies to make existing MIS more attuned to user needs. In addition, a regional agricultural- input market can be organized to reduce differentiation of input prices across countries, which is currently the case, and to cut down the price of most basic inputs for farmers. Building on the experience of other regional MIS—for example, the European Union Internal Market Information System (IMI)—the MIS developed in West Africa could help authorities to fulfill their cross-border administrative cooperation obligations in multiple Single Market policy areas. Furthermore, the system needs to be flexible and accommodate any national centralized or fully decentralized administrative structure (Lottini 2014). 127 Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens Key points in this chapter: 7 • Food crop production losses in Burkina Faso, Chad, Mali, Niger, Sierra Leone, and Togo could amount on average to more than US$700 IMPACTS OF million per year and climb to more than US$1 billion with 20 percent probability (once every FOOD CROP five years). These large production shocks could also lead to recurrent food security crises in the six countries. PRODUCTION • The modeled food security impact of food crop production losses is particularly severe RISKS ON FOOD in the Sahel: in Burkina Faso, Chad, and Niger, large numbers of people are thrown into food insecurity even by relatively frequent, low- 128 SECURITY severity food production shocks. • In Chad and Niger, where the baseline or chronic food insecurity is particularly elevated, there is a 10 percent probability that food This chapter analyzes how the estimated production shocks could lead to the majority of production losses from risks impact food the national population being undernourished. insecurity in Burkina Faso, Chad, Mali, • Four Sahel countries—Burkina Faso, Chad, Niger, Sierra Leone, and Togo and the cost Mali, and Niger—request an average US$1 of associated humanitarian assistance at billion of international food security-related different levels of events severity. The country humanitarian support every year, while only sample group in this chapter was selected US$360 million gets funded on average, based on the group of countries that will barely covering chronic food insecurity needs. participate in the initial phase of the West Additional aggregate funding needs for severe Africa Regional FSRP (see chapter 1). production shocks in the four countries are large; for example, for production shocks occurring with a 5 percent probability (1 in 20 years), additional funding needs could rise to US$800 million.26 26 See appendix C for the limitations of this model, which was developed by the World Bank for this report. A subsequent analysis for future iteration will be to compare the modeled number of undernourished people with IPC (Integrated food security Phase Classification) or past historical reports of estimated people in food insecurity condition. T o understand the impacts that food crop works, which country and crop is most exposed losses have on food security around allows targeted risk management policies the region, any loss model would need and investments for the highest impacts and to be expanded beyond assessing the value discussions can be initiated to design more of food crop losses so that it has a direct sophisticated disaster risk finance instruments. link to food insecurity indicators and is not Complementary analysis is needed to quantify limited to an economic and geographic view the protection needed by the most severely of losses. The simulations made in chapter food insecure populations, according to the 4 are a proposed way to assess the scale of model in this chapter. The methodology production losses resulting from yield variations developed for this purpose is described in box across the region. Knowing, through modeling 7.1 and in more detail in appendix C. BOX 7-1 Model Methodology To estimate the impacts of agricultural production losses on food insecurity, a complementary food insecurity model was developed for this report. This model builds on chronic food insecurity 129 numbers, namely, the prevalence of undernourished people as a percentage of the total population. The following data sets are used: • Average number of calories of a normal diet per country (Humanitarian Data Exchange) • Basket of the main three food crops consumed by country (FAO, FAOSTAT) • Baseline prevalence of undernourishment per country (Humanitarian Data Exchange) • Baseline daily calorie deficit for the undernourished per country (Our World in Data, Oxford).. Using the data on the average basket of food crops and calories per person per day in each coun- try, the model recomputes a daily intake of calories derived from the main crops for a well-nour- ished and an undernourished person. The results of the food crop risk assessment are used to calcu- late the likely drop in production and the consequences in terms of caloric deficit for each country. The caloric deficit is then translated into a recomposed food basket for each country. Based on the caloric deficit per person for each simulated event, the model calculates the impact as either people who are newly food insecure (the so-called horizontal scale-up of the food insecure) or the depth of the food deficit for the already existing food insecure (the so-called vertical scale-up). The advantage of such a model is that targeted measures can be put in place depending on how the food insecurity situation evolves and these measures can then be linked to suitable scale-up of emergency response. Another advantage of using a food insecurity model is that the estimated number of people undernourished for different production shocks can be translated into the expected humanitarian cost. Source: World Bank series of country CSAIPs Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens Depth of undernourishment crops. The basis was the modeled drop in yields (or loss of production; see table 7.1) The most The countries experiencing the highest severe economic losses from production (as a modeled crop losses and the highest expected share of total expected output from the five levels of food insecurity from agricultural main crops) are in Niger and Sierra Leone. Of production losses are not always the same. the six countries assessed, Mali and Niger are For each simulation of loss in production, the the countries with the highest impact of yield corresponding daily intake of calories was shocks on the daily diet. recomputed based on the basket of main food TABLE 7-1 Modelled loss in agricultural output from main crops and Depth of undernourishment Modeled Loss in production Mali Chad Burkina Niger Togo Sierra Leone Baseline (before shock) 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 1-2 years 3.06% 2.32% 2.78% 5.43% 2.96% 1.78% 1-5 years 10.29% 9.70% 8.88% 16.28% 7.99% 11.97% 1-10 years 14.92% 14.96% 12.73% 23.68% 12.41% 17.82% 1-15 years 17.36% 17.83% 15.01% 27.66% 14.99% 20.61% 130 1-20 years 19.00% 19.47% 16.36% 29.80% 16.83% 22.62% 1-50 years 23.08% 24.64% 20.70% 36.76% 22.27% 27.92% 1-100 years 26.14% 28.57% 24.32% 41.21% 26.85% 31.37% AAL 5.44% 5.13% 4.76% 8.83% 4.85% 6.04% Depth of undernourishment (variation in daily intake vs a standard diet) Mali Chad Burkina Niger Togo Sierra Leone Baseline 15.75% 31.42% 32.05% 12.40% 14.95% 26.45% 1-2 years 20.52% 52.04% 38.66% 27.80% 31.77% 43.42% 1-5 years 28.93% 56.28% 44.02% 38.28% 37.48% 50.23% 1-10 years 35.51% 59.96% 48.20% 45.79% 42.38% 55.80% 1-15 years 42.02% 63.70% 52.49% 52.49% 47.56% 61.09% 1-20 years 47.92% 67.05% 56.47% 57.71% 52.51% 65.70% 1-50 years 57.63% 72.63% 63.53% 64.67% 61.43% 72.83% 1-100 years 61.66% 74.99% 66.76% 66.87% 65.60% 75.54% AAL 18.87% 51.36% 37.57% 25.37% 30.39% 42.73% Source: World Bank Number of people undernourished Extreme food production risk events could the model, Chad, Niger, and Togo could see more result in very high incidents of food insecurity than 90 percent of their population in a state of in Chad, Niger, and Togo. For each simulation undernourishment for a very severe shock to of production loss, it was possible alternatively production (1 in 100 years). Mali is displaying to recompute the number of people who will the lowest prevalence of undernourishment fall into the undernourished category (that is, (20 percent), even for extreme events. Mali having a lower average daily intake of calories is also the country with the lowest baseline than a normal diet) (see table 7.2). According to prevalence of undernourishment. TABLE 7-2 Total estimated number of people undernourished by type of event Number of people undernourished (millions) Modeled scenario Mali Chad Burkina Niger Togo Sierra Leone Baseline 0.92 5.77 3.58 3.64 1.55 1.91 1-2 years 1.19 9.76 4.44 8.17 3.42 3.22 1-5 years 1.40 10.05 4.66 9.46 3.68 3.41 1-10 years 1.68 10.48 5.04 11.23 4.02 3.70 1-15 years 2.07 11.14 5.46 13.42 4.51 4.10 1-20 years 2.45 11.79 5.93 15.39 5.03 4.47 131 1-50 years 2.79 12.38 6.38 16.95 5.56 4.79 1-100 years 3.10 12.96 6.77 18.10 6.01 5.06 AAL 1.10 9.43 4.20 7.38 3.18 3.08 Source: World Bank The baseline percent of population frequently. In Niger, the lower severity shocks experiencing undernourishment differs are contributing to a significant step increase between the countries as does the in the number of undernourished people, percent increase of such undernourished reaching almost the entire population for very populations at different levels of shocks severe shocks. Burkina Faso has a high baseline (figure 7.1, a–d). In Chad, the baseline or of chronically food insecure people, and relative chronic undernourishment is very elevated, so to it, shocks add a small fraction of newly food less severe shocks (one- to two-year event) can insecure people. already push a significant number of people into undernourishment and subsequently into food insecurity. In Mali, increasingly severe shocks will add a proportional number of newly undernourished people, as opposed to Chad, where a potential big increase in undernourishment could be observed more Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens FIGURE 7.1 a–d. Incremental Increase in Number of People Undernourished at Different Production Shock Return Periods (Mali, Chad, Burkina Faso, and Niger) MALI CHAD Increase Discrease Total Increase Discrease Total 4,000,000 16,000,000 234,000 3,500,000 370,000 14,000,000 436,800 436,800 582,400 306,000 582,400 3,000,000 12,000,000 655,200 342,000 655,200 436,800 378,000 10,000,000 3,989,440 291,200 2,500,000 387,000 8,000,000 2,000,000 279,000 6,000,000 5,765,760 1,500,000 216,000 270,000 918,000 4,000,000 1,000,000 2,000,000 500,000 - Baseline 1-2 1-3 1-5 1-10 1-15 1-20 1-25 1-50 1-100 Baseline 1-2 1-3 1-5 1-10 1-15 1-20 1-25 1-50 1-100 Years Years BURKINA FASO NIGER Increase Discrease Total Increase Discrease Total 8,000,000 25,000,000 354,350 373,000 7,000,000 391,650 447,600 20,000,000 832,000 728,000 6,000,000 466,250 1,144,000 128,950 1,560,000 373,000 1,976,000 5,000,000 857,900 223,800 15,000,000 2,184,000 4,000,000 3,580,800 1,768,000 3,000,000 10,000,000 1,289,600 4,534,400 2,000,000 5,000,000 3,640,000 1,000,000 - - Baseline 1-2 1-3 1-5 1-10 1-15 1-20 1-25 1-50 1-100 Baseline 1-2 1-3 1-5 1-10 1-15 1-20 1-25 1-50 1-100 Years Years 132 Source: World Bank 7.1 The Costs of Required Food Security- Related Humanitarian Support T he four Sahel countries request food humanitarian response cost figure per person security-related humanitarian support of US$40.27 The modeled need for humanitarian of around $US1 billion every year, of assistance could then be as high as US$2.5 which approximately $US360 million gets billion for extreme scenarios and a 1 in 5 years funded (OCHA FTS 2020). Calculating the cost scenario would require a total funding of more of chronic food insecurity uses the modeled than US$1 billion (figure 7.2). number of undernourished people, assuming they will require food relief assistance. These figures are multiplied by an estimated 27 US$40 for humanitarian response cost per person is an approximation used by the ARC for different countries in Africa and was developed through consultations with the WFP. . FIGURE 7.2 Comparison Between Status Quo Level of Humanitarian Appeals in Sahel Countries and the Modeled Increase in Humanitarian Appeals in Case of Disasters Source: World Bank 133 Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens 8 WORRYING GAPS Key points in this IN AGRICULTURE RISK FINANCING: INSIGHTS INTO FINANCING chapter: • Agricultural risk finance is one of several key components in a comprehensive ARM strategy. • None of the six countries is equipped to respond even to moderate food production shocks occurring once every five years, much less to more severe shocks occurring once every 10 years with their current response funding ARRANGEMENTS IN arrangements. Governments lack funding to cover humanitarian and food security needs SIX COUNTRIES IN and are much less equipped to balance out broader economic losses. • The combined humanitarian and economic 134 WEST AFRICA funding gap is largest for Niger, which faces a gap of US$768 million for 1 in 5 years food production shocks and US$1.1 billion for 1 in 10 years food production shocks. Mali follows, with respective figures of US$341 million and This chapter introduces agricultural risk US $491 million. financing as part of a comprehensive ARM • All six countries lack an overall coordinated approach and explores to what extent approach—for example, a national DRF existing in-country financing resources to strategy—ensuring that adequate financial respond to food security crises suffice to resources exist to respond to food security respond to the estimated financing needs crises. for different production shocks. The analysis • Despite the strong exposure to shocks, there is conducted for the six countries of Burkina is low use of risk financing arrangements at Faso, Chad, Mali, Niger, Sierra Leone, and Togo. national levels, with governments relying on The main existing agriculture risk financing ex-post relief, particularly donor-provided mechanisms and resources are reviewed, humanitarian aid. both at the national level and at the regional • Also, at the regional level, there are very few level. Then, the identified amounts for each risk-financing instruments available. However, country are contrasted with the expected there are some encouraging regional initiatives humanitarian needs and economic impacts emerging, including the strong take-up of ARC for food production shocks of different and the establishment of an ECOWAS regional magnitudes. reserve. 8.1 RATIONALE FOR DRF IN THE AGRICULTURE AND FOOD SECURITY SPACE B urkina Faso, Chad, Mali, Niger, Sierra Ebola outbreak that affected Sierra Leone and Leone, and Togo regularly rely on ad drew in massive amounts of aid funding— hoc international humanitarian aid US$712 million or 7.4 percent of Sierra Leonean to respond to shocks. Data from the OCHA GDP in 2014 alone. The large amount of external shows that during the period 2012–19, the four support to the Sahel countries illustrates Sahel countries received almost US$1 billion their overall low degree of financial resilience per year on average or 1.9 percent of their GDP or preparedness to deal with humanitarian in international humanitarian assistance. For shocks. Instead, humanitarian response finance the two coastal countries, the ratio of external tends to be arranged ex-post—through donor humanitarian assistance to GDP was much support or, to a lesser extent, through budget smaller during this period, except for the 2014 reallocations. FIGURE 8-1 Humanitarian Aid Inflows in US$, Millions and as a Percent of GDP (Burkina Faso, Chad, Mali, Niger) 135 Source: UN OCHA FTS Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens FIGURE 8-2 Humanitarian Aid Inflows in US$, Millions and as a Percent of GDP (Sierra Leone, Togo) Source: UN OCHA FTS Note: For better legibility, Ebola-related 2014 humanitarian inflows are cut off in this diagram; they amounted to US$714 million or 7.4% of Sierra Leonean GDP This ex-post approach to funding proactive response mechanisms could reduce 136 humanitarian expenditure tends to be humanitarian cost and losses by US$3. Besides inefficient; prearranging response finance, the lack of speed, the key challenge with ex- that is, disaster risk finance instruments, post arranged finance is that it is unreliable— instead can yield large benefits in terms of affected governments, and ultimately affected speed, cost, and reliability. Funding that is populations, will always be uncertain of the only gathered after a shock has occurred tends amount they can receive from donors or of the to arrive with great delay, thus enabling the government budget available for reallocation. humanitarian crisis to unfold even further. For Thus, planning for recurring humanitarian droughts in Africa, the average delay between expenditure makes sense. DRF is “the system of a failed harvest and the arrival of emergency budgetary and financial mechanisms to credibly assistance is estimated to amount to seven to pay for a specific risk or risks, arranged before nine months (Clarke and Hill 2013). Conversely, shocks occur” (Centre for Disaster Protection studies from East Africa and elsewhere 2019). DRF, for example, includes instruments show that the value for money of response such as dedicated disaster funds, budget lines, finance increases greatly when responding or insurance contracts that provide finance earlier: for example, Cabot Venton (2018) for specific shock-related expenditure. Such showed that each US$1 invested in early and prearranged funding can help provide finance 28 The World Bank’s Crisis and Disaster Risk Finance team has captured the most important concepts of disaster risk financing in the “Five DRF Fundamentals”: (1) relevance of data and risk analytics; (2) risk layering approach, that is, the recognition that specific financing instruments are suited to specific risks; (3) timeliness of funding; (4) rules guiding disbursement; and (5) the benefits of risk diversification. An overview is given in an e-learning course, https://olc.worldbank.org/content/fundamentals-disaster-risk-finance-0. much faster, more cost effectively and more plans, and building the capacity of reliably, and thus ultimately help to save lives. disaster responders. Preparedness measures are taken to ensure a country’s As per the methodological frameworks in capacity to respond effectively when a chapter 2, there are a host of different ARM risk is realized. instruments and approaches, and risk financing 4. Risk financing, including the use is just one element that needs to be embedded of ex-ante and ex-post financing in a comprehensive ARM framework. Separate mechanisms for shock response. These elements of an effective ARM process include are the financial sources used to pay the following:29 for surge costs associated with a risk occurring (for example, humanitarian 1. Risk identification, such as developing relief following a drought). It is agriculture risk assessments and natural important to plan for and establish hazard risk profiles to understand the these mechanisms ahead of the risk. extent to which certain parts of the 5. Resilient recovery (and sector are exposed to which risks. reconstruction), including effective ex- Risk identification also includes the ante design of institutional structures assessment of contingent liabilities, that to ensure that disaster recovery and is, quantifying fiscal exposure linked reconstruction are timely and of high to the occurrence of certain risks. It is quality. critical to enable appropriate pricing and application of DRF instruments that All these elements are critical for an are often the first step to establishing a effective ARM process. No single element 137 more effective financing framework. alone—including risk financing instruments— 2. Risk mitigation, such as irrigation can address the risks that the agriculture sector systems, improving storage is exposed to. They are also interdependent, as, infrastructure, land-use planning, and for example, the needed amount and timing of conservation agriculture. Climate- response funding (4) depend on the planned Smart Agriculture (CSA) is an approach contingency measures (3). Also, the more that combines various risk reduction risks are reduced (2), the more affordable risk tools—conservation agriculture, soil financing solutions such as insurance (4) will and water conservation, and improved be. Thus, a comprehensive approach is needed. livestock management practices— Table 8.1 gives an overview of how various with emission-reducing elements instruments map the different elements of (PARM 2018). Generally, agricultural risk ARM. mitigation measures seek to bolster farm-level resilience against risks that the sector is exposed to. 3. Preparedness building, such as the use of EWS, developing contingency 29 These are the five Pillars of Action of the GFDRR. Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens TABLE 8-1 Mapping of Exemplary Agriculture Policy Measures to Risk Management Framework Risk identifica- Risk reduc- Preparedness Risk fi- Resilient Exemplary ARM measures tion tion building nancing recovery Develop agricultural risk profile X Assess contingent liabilities for X X agriculture sector Improve crop storage infrastruc- X ture Land-use planning X CSA X Warehouse receipt system X National or regional disaster EWS X Develop disaster contingency X plans Capacity building of shock re- X sponse stakeholders National disaster fund X Strategic food reserve (to counter X food insecurity) 138 Public borrowing for shock X response Contingent line of credit for shock X response Risk transfer instruments, includ- X ing insurance Price hedging (commodity ex- X changes) Strengthen financial mechanisms X X for recovery Improve public capacities for X recovery planning Strengthen central policy frame- X works for recovery Source: World Bank There is no single financing instrument that • Medium-risk layer. In cases where can address all agriculture risks; multiple governments must respond to medium- instruments should be combined to ensure scale, less frequent events, such responses availability of resources for governments could be financed through contingent when they are needed most. Just as with ARM facilities. These facilities are for many in general, financial planning for government countries provided by international responses to agricultural risks should follow a financial institutions. Contingent credit layered approach. This approach (represented arrangements that are preagreed on schematically in figure 8.3 and explained below) allow governments to access liquidity prioritizes the most cost-effective solution for quickly after a disaster. different layers of risk and ensures that the most expensive instruments are used only for • High-risk layer. The financial risk for the extreme events that rarely occur. extreme events that occur infrequently could be transferred to the international • Low-risk layer. In cases where frequent markets using instruments such as but low-impact events cannot be insurance or commodity hedges. Such mitigated in full, such events could be risk transfer solutions tend to be relatively financed primarily through risk retention expensive but can unlock large amounts mechanisms in the form of a disaster of funding when they are needed most. fund, a dedicated budget line, or a contingency budget. If needed, some minor in-year budget reallocations could also be used. 139 FIGURE 8-3 Risk Layering Strategy for Government Ex-Post Risk Response FRECUENCY / SPECIFIC INSTRUMENTS PRIMARY RECEIPIENT SEVERITY OF RISK GENERAL INSTRUMENTS OF FUNDS National Insurance Sovereign Risk Transfer Commodity Hedge National Funds Risk transfer Schemes (Diasasters) (Price Volatility) Vulnerable Populations (Weather and Disease) Contingent Finance Contingent lines of credit (WB, Donors, AfDB) National Budget Support Reserves National National National Vulnerable Populations Diasaster Fund Safety Net Fund Food Reserves TOOLS THAT SUPPORT THE DRF STRATEGY National Early Warning National Disaster Risk Finance National Disaster Risk Finance National Vulnerability Systems Diagnosis Strategy Assessments Source: World Bank None of the six target countries has a national can set policy priorities aimed at strategically DRF strategy in place, but they provide meeting post-disaster financing needs. A sound a useful framework for a government to DRF strategy is based on an in-depth analysis of plan the financing of shock-related costs. risks, contingent liabilities, and the remaining Through a national strategy, governments risk management framework. A DRF strategy Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens builds on this analysis to clarify financing for agriculture sector and addresses all costs the identified risks and the best instruments associated with certain risks (for example, to use. The best instruments will be those that reconstruction costs linked to potential floods). provide reliable, sufficient, rapid, and cost- However, in countries where the agriculture effective disaster funding and that also account sector plays an important role and is exposed for broader governance considerations. This to significant risks—as in the six focus means that instruments should match in- countries—risk financing instruments targeted country structures and capacity—a frequent at agriculture should feature prominently. So challenge with implementing technical risk far, none of the six countries has a DRF strategy financing approaches such as insurance. A in place. comprehensive DRF strategy focuses on the 8.2 EXISTING NATIONAL AGRICULTURE RISK FINANCING INITIATIVES T he overall use of disaster risk finance is purchasing ARC replica insurance to mirror its instruments in the six focus countries in-country finances for responses led by NGOs. is low but there is significant interest Four of the six countries have experimented from ECOWAS, international donors, and with agricultural insurance, but none of the 140 national governments to improve the programs have reached scale. It is striking that response strategy to disasters in West Sierra Leone, despite its significant exposure to Africa. This report reviewed the DRF landscape climatic shocks, has the least experience with in each of the six focus countries (table 8.2). using prearranged financial disaster response Three of the six countries operate a dedicated mechanisms. The country DRF profiles are national disaster fund that provides emergency described in detail in the background report relief funding to respond to food emergencies, Agriculture Risk Financing in West Africa. but these are often funded by donors and not consistently funded. Five of the six countries operate strategic food reserves with the explicit purpose of providing food relief for emergencies; however, target sizes tend to be too small to cope with the need and even the target stockage is often not met. Five of the target countries have purchased sovereign insurance coverage from the ARC, demonstrating the strong interest in ARC across Western Africa. However, the size of the insurance coverage is often relatively small, and there has been inconsistent participation by countries. For Burkina Faso and Mali, the WFP TABLE 8-2 Use of National DRF Instruments in the Six Focus Countries Risk Retention** Risk Transfer** Sover- eign Risk National ARC Country National Transfer National Local physi- Scalable replica financial (ARC) Agriculture Disaster SFRs cal SFR Safety 2019–20 SFR (tar- – cover insurance Fund (tons) (target Nets , US$, get size) 2019–20, size)* millions US$, millions Burkina 25,000t (X – signed X 15,000t 75,000t (X – pilot) 7 (X – pilot) Faso equivalent MoU) Chad 15,000t 35,000t - (X – pilot) 1.2 - - 25,000t Mali X 15,000t 60,500t (X – pilot) 15 12.6 (X – pilot) equivalent 60,000t Niger X 15,000t 80,000t (X – pilot) 5.3 - (X - pilot) equivalent Sierra - - - - (X – pilot) - - - Leone Togo - - 25,000t - (X – pilot) 4 - - Sources: Various Note: SFR is Strategic Food Reserves; ARC is African Risk Capacity *The national physical SFR is aggregating all national food storage capacities in each country (including, for example, SNS and SI) ** Risk retention is a strategy to cover from national funds the cost of recovery after low- and medium-severity disasters, while risk transfer is the strategy to cover through insurance-like mechanisms the costs of very severe but infrequent events. The funds are in theory readily available to be disbursed and they incur 141 relatively low transaction costs (yet they have an opportunity cost), but they are limited, while risk transfer instruments can channel significant amounts of financial support, but they have higher set-up times and transaction costs. This is the reason for layering them depending on the type of event and the required funds for response. 8.2.1 National disaster funds D isaster funds are most suited to budget allocated for these funds given the provide finance for higher frequency, limited fiscal space available in countries and lower severity shocks and are an the opportunity cost related to them, they allow instrument used by governments around to respond to small- to medium-sized events the world to provide shock response and increase the predictability of resources. funding. Disaster funds are off-budget It is best practice for such funds to have clear financial pools holding finance to be disbursed rules regarding access (for example, preagreed for responding to selected shocks. Such a risk triggers); the eligibility of funded activities; and finance instrument is appropriate to cope disbursement processes. The key advantage with frequent and low severity shocks that of a disaster fund vis-à-vis a simple line in the will not draw attention from the international budget to be used for disaster response is that community, and mobilization of donations funding can accumulate over multiple years, or loans from the donor community are thus allowing the funding of costs of larger usually slow to materialize and the amount disasters to be spread over a longer period. By unpredictable. Despite the limited size of the being established separately from the budget, Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens disaster funds may also facilitate the inflow prepare for emergencies (Coulibaly 2014; of external donor funding. The disaster funds République du Mali 2015). serve as a channel to aggregate funding and to distribute more efficiently, effectively, and • Niger operates a very large central transparently rather than having scattered (varying annual budget of almost interventions across donors. US$15 million) relief fund that is a key funding instrument for food security In the six focus countries, only three interventions. This Fonds Commun des governments have set up such disaster Donateurs (FCD) is financed exclusively reserve funds. Although all six countries through donor contributions. The FCD experience significant shocks to production and receives financial contributions every food security on a regular basis, only Burkina year and serves as the central instrument Faso, Mali, and Niger have set up dedicated for donors to channel and coordinate disaster funds. They differ significantly in size their humanitarian food relief funding. and setup: Some contributions seem to be made in advance, in anticipation of potential • Burkina Faso has set up the moderately- future crises, while others are short-term sized (annual budget of approximately support contributions tied to specific US$750,000) Fonds National de Solidarité ongoing crises (Achirou 2017).30 with a broad mandate to contribute to the care of individuals, disadvantaged Despite their significant exposure to disasters, groups, or groups in difficulty and victims no specific disaster funds could be identified 142 of natural disasters and humanitarian in the remaining three countries as part of this crises. The fund is financed through study (table 8.3). the national budget and can be used for a variety of purposes, including relief and rehabilitation programs following disasters and social assistance and general socioeconomic support programs (Zongo 2019). • In Mali, only a small portion (approximately US$84,000) of the National Agricultural Support Fund is set aside for disaster response. The fund is financed through the national budget. Sectoral government structures are expected to rely on their own operating funds to 30 For the purposes of this study, it could not be determined what percentage of FCD funds are made in anticipation of abstract future crises (and are thus to be categorized as “ex-ante prearranged response funding”) and what percentage is made in response to concrete ongoing or upcoming crises (and are thus to be categorized as “ad-hoc response funding”). For the purposes of the indicative funding gap analysis conducted in Section 5.4, all FCD funds have thus been given the benefit of the doubt and were assumed to be provided as ex-ante funding. TABLE 8-3 National Disaster Funds in Focus Countries Country Disaster fund Annual size in US$ Responsible entity Sources Fonds National de Approx. 750,000 (FCFA 400 Ministry for Social Action Zongo 2019; Alpha Burkina Faso Solidarité million) and National Solidarity and Pemou 2019 Approx. 84,000 (1.5% of Coulibaly 2014; National Agricultural Ministry of Rural Devel- Mali budget of FCFA 3 billion in République du Mali Support Fund opment 2015) 2015 Approx. 14.8 million Fonds Commun des Niger (average annual donor DNPGCA Alimentaires Achirou 2017 Donateurs contributions 2008–17) Sources: As indicated 8.2.2 Strategic Food Reserves I nternational experience holds that SFRs themselves. Local stocks or cereal banks are can be an effective tool to counter food collective community stocks and not publicly insecurity. They store food (in physical or owned. Mali is the exception where each of the financial and virtual form) to be available 700 municipalities in the country has its own 143 swiftly for emergency response when needed. cereal bank. A rough estimate is that in each of Similar to disaster funds, SFRs are thus reserve the Sahel countries, around 1,000 local cereal instruments that are most cost effective when banks operate, each holding around 15 tons of used to respond to higher frequency, lower grain (Galtier 2019). They are financed by a mix severity disasters. They can be an effective of national support, financing from national instrument to counter food insecurity for such and international NGOs, and community types of shocks when managed properly. contributions. With this financing in place, they There is an important difference between SFRs tend to purchase grain just after the harvest and so-called buffer stocks. These also store when prices are relatively low and hold it to grain but use it not for distribution to food- be sold or distributed at favorable conditions insecure populations during times of need to households in need during the lean season but to buy and sell grains to influence market (SOS Faim 2016). prices. International consensus has largely held national buffer stocks to be an ineffective tool At the national level, SFRs have long been as their size tends to be too small to impact an important instrument to counter food grain prices in a significant way. Most advanced insecurity in Sahel countries. Burkina Faso, economies have abandoned them (World Bank Mali, and Niger operate two national SFRs 2012; FAO et al. 2011). each: (1) The Stocks National de Sécurité (SNS) were created at the end of the 1980s and In the Sahel countries, SFRs exist both at the are managed by the respective government local and at the national level; at the local together with donors. For these, food stocks level, stocks are managed by communities can only be disbursed after an EWS alert and Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens with the agreement of both the government target of 50,000 tons was met over the period and the donors’ group (double signature). The 2005–15 (Alpha and Pemou 2019). In Mali, SNS are composed of a physical and a financial on average 82 percent of the SNS stocking reserve; and (2) the Stocks d’Intervention (SI) (or target of 35,000 tons was met over the period in Niger, the Réserve Alimentaire Stratégique) 2006–15 (FAO 2017). For Niger, a study found that were created after the food crisis in that the SNS was stocked on average at only 2005 and are managed by the respective 41 percent of the target amount during the government alone (Galtier 2019). Chad and period 2002–10 (Club du Sahel et de l’Afrique Togo also operate SFRs. de l’Ouest 2010). And the SFR in Togo tends to hold around 10,000–15,000 tons, that is, around The examined national SFRs often fall short 50 percent of the stocking target of 25,000 tons of their stocking targets. In Burkina Faso, on (WFP 2018). A snapshot of the latest figures average only 47 percent of the SNS stocking available is provided in table 8.4. TABLE 8-4 Strategic Food Reserve Storage in Focus Countries Level of public stocks in tons - actual and (target), Public as per different dates Total – actual Country storage Source and (target) capacities SNS* SNS – financial SI** Burkina 38,147 (50,000), 3,600 (25,000), Alpha and 98,100 41,747 (100,000) n.a. (25,000) Faso Dec 2015 Jan 2016 Pemou 2019 144 Chad n.a. 1,860, (35,000) 1,860 (35,000), n.a. n.a. FEWS NET 2019 2019 35,000 (35,000), 10,000 Mali 136,700 45,000 (85,500) n.a. (25,000) FAO 2017 2015 (25,500), 2015 32,000 (50,000), 30,000 (n.a.), Achirou 2017; Niger 154,700 62,000 (140,000) n.a. (60,000) March 2011 March 2011 Galtier 2019 12,500 (25,000), Togo 92,500 12,500 (25,000) n.a. n.a. WFP 2018 2018 Source: Alpha and Pemou 2019; FEWS NET 2019; FAO 2017; Achirou 2017; Galtier 2019; WFP 2018 *SNS is Stocks National de Sécurité **SI is Stocks d’Intervention 8.2.3 Scalable safety nets S calable social safety nets are emerging shocks. In the four countries, the respective as an important approach to responding social safety net programs have delivered cash to disasters in Africa. Scalable safety nets payments to shock-affected households. In are part of the shock-responsive or adaptive Niger, the government is working on a pilot social protection agenda that seeks to help program using satellite imagery for a rules- households build resilience against shocks— based rapid trigger of shock-response scale- that is, their “ability to prepare for, cope with, ups of the safety net (World Bank 2019). In and adapt to shocks in a manner that protects Sierra Leone, the government has repeatedly their well-being; ensuring that they do not fall scaled up existing safety nets to deliver cash into poverty or become trapped in poverty transfers to households affected by Ebola and due to impacts” (Bowen et al. 2020). Scalable floods (Sandford et al. 2020). Likewise in Togo, safety nets are existing continuous safety net the government has used an existing cash programs (for example, cash transfer programs) transfer program to deliver payments to shock- that can scale up rapidly in the event of a affected households, most recently in response shock such as a severe drought. They can scale to COVID-19 (Simonite 2020). up either vertically, increasing the amount of cash transfer payments to shock-affected So far, no prearranged DRF instruments beneficiaries, or horizontally, adding additional have been used to finance scalable safety shock-affected households to the beneficiary nets in the focus countries, but respective roster. Some safety nets can also tweak the efforts are emerging. The shock-related cash 145 delivery of their routine services in response to transfer payments in the six focus countries shocks—for example, by waiving conditionality. have so far largely not been financed through Others are also used for piggybacking, that is, established prearranged mechanisms but they provide the infrastructure to enable the through ad hoc payments. However, in the delivery of additional shock-related services Sahel countries, the World Bank has set aside (O’Brien et al. 2018). Different parts of the resources for scalable cash transfers through world have experimented widely with scalable the Sahel Adaptive Social Protection Program safety nets and in Africa, they have been and works with governments to institutionalize implemented—for example, in Ethiopia, Kenya, this further (for example World Bank 2018). Also and Uganda (Ninno, Coll-Black, and Fallavier the Government of Sierra Leone is thinking 2016). about pre-arranging finance for shock response (Sandford et al. 2020). Different programs are experimenting with scalable safety nets in all six focus countries. Supported by the World Bank through the Sahel Adaptive Social Protection Program since 2014, the four governments of the Sahelian focus countries are establishing social protection programs to support the most vulnerable and build their resilience against Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens 8.2.4 Agricultural insurance schemes S upporting agricultural insurance based on rainfall data (for example, weather schemes can be an important way for index insurance), area yield estimates (for governments to ensure that shock- area yield index insurance), or satellite-based affected agricultural producers receive evaluations of pasture availability (for livestock rapid relief after shocks. In many countries index insurance) (Lung 2020). By foregoing across the world, governments support farmer-level loss adjustment, agricultural national agricultural insurance schemes index insurance holds the promise of greater through various support measures such as affordability than traditional indemnity subsidy payments, required data collection, insurance and reduces the problems of moral or by making insurance compulsory. There are hazard and adverse selection. Its key challenge, different policy objectives governments pursue however, is the exposure to basis risk—that with this, including addressing typical market is, the risk of a “difference between an index failures of agricultural insurance markets, and the shock that the index is supposed to building the resilience of agricultural producers be a proxy for” (Centre for Disaster Protection to production shocks, and protecting their 2020). Over the last two decades, a multitude fiscal balance against recurrent humanitarian of index-based agricultural insurance programs expenditure shocks. Today, it is considered have been implemented across the African best practice for national agricultural insurance continent, with major programs operating, for programs to be run via public-private example, in Ethiopia, Kenya, Malawi, Senegal, 146 partnerships between the government and and Zambia today (GIZ 2016). Nevertheless, private sector insurers, as each have their most of these have failed to reach scale. comparative advantages for different required elements for such a program (Mahul and Two of the six focus countries—Burkina Faso Stutley 2010). and Mali—have a fair amount of experience with agricultural insurance. Niger has limited Agricultural index insurance programs agricultural insurance experience and the other have received much interest and support three countries have no experience. In all six in many countries across Africa over countries, agricultural insurance has not yet the last two decades. There are different reached the national scale. types of agricultural insurance programs. For traditional indemnity-based insurance • In Burkina Faso, PlaNet Guarantee programs, policyholders are indemnified and Oxfam have operated two insurance following producer specific loss-adjustments pilots, a credit-linked area yield index procedures. These tend to be costly, however, insurance program for cotton and an driving up the insurance premium and making evapotranspiration index insurance it unaffordable for poorer subsistence farmers program for maize. However, both and prone to challenges such as moral hazard initiatives have been unable to reach and adverse selection. Instead, index insurance scale, together selling less than 10,000 solutions use quantified indices that use policies in their 2014 peak year. From proxy data to estimate agricultural production 2011–13, an interministerial technical losses. Such indices can, for example, be committee (CTI) assessed opportunities for agricultural insurance. In 2014, the CTI • In Chad, Sierra Leone, and Togo, no ultimately recommended that a public- existing experience with agricultural private partnership be set up, which has insurance products could be identified. not been realized to date (Stoppa and In Togo, a 2019 study mandated by the Dick 2018). Ministry of Agriculture analyzed the feasibility of introducing index-based • In Mali, PlaNet Guarantee, together insurance and suggested a roadmap of with other development and actions to operationalize an agricultural insurance partners, is supporting an insurance program for maize, rice, and evapotranspiration index insurance cotton farmers (République du Togo and program providing cover to more than AfDB 2019). 40,000 maize and sesame farmers (GIIF 2017; USAID 2018). Furthermore, another • Other initiatives in the region are insurance program supported by government program development Développement international Desjardins WFP R431 in Burkina Faso; GCF-financed (DID) with the Agriculture and Rural program in Sahel (forthcoming); and Financing in Mali project (FARM) provides IFAD, WFP, AfDB, ARC and The Africa cover to approximately 1,500 rice, maize, Integrated and onions farmers (DID 2016; WFP, UN Women, and USAID 2017). • Climate Risk Management Programme. Building the resilience of smallholder • In Niger, there have been a few farmers to climate change impacts seven agricultural insurance initiatives. One Sahelian countries of the Great Green 147 is the preliminary development of a Wall (GGW).32 weather index by the IRI for Climate and Society at Columbia University as part of the Program Africain d’Adaptation of the UN Development Program (UNDP) that, however, has not been translated into an insurance program (IRI 2015). Besides, a pasture-drought insurance product for the pastoral areas has been developed and tested by IBISA (Inclusive Blockchain Insurance Using Space Assets) in 2019 (Luxembourg Space Agency 2020). 31 World Food Programme, R4 Rural Resilience Initiative, March 2021, https://docs.wfp.org/api/documents/WFP-0000128805/download/?_ ga=2.6130995.1248932254.1624433472-864948064.1624433472. 32 Green Climate Fund, The Africa Integrated Climate Risk Management Programme: Building the resilience of smallholder farmers to climate change impacts in 7 Sahelian Countries of the Great Green Wall (GGW), April 2021, https://www.greenclimate.fund/sites/default/files/document/ funding-proposal-fp162.pdf RRSA interventions to date have been 5,000 megatons to Nigeria in January 2019; 2,750 megatons to Ghana in December 2018; 4,303 megatons to Burkina Faso in August 2018; 6,528 megatons to Niger in August 2018; and 1,130 megatons to Nigeria in August 2017 (ECOWAS 2019). Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens 8.3 EXISTING REGIONAL AGRICULTURE RISK FINANCING INITIATIVES 8.3.1 Regional cooperation on food reserves I nternational best practice holds that reserve stocks to support each another regional food reserves can be a useful tool in times of need; (2) capacity building to respond to food security crises. Regional of national food reserve agencies as food reserves that hold grain to be provided capacity is stronger in some countries, for food relief to multiple countries can offer particularly Sahelian ones; and (3) benefits across two dimensions through (1) risk- information exchange through regular pooling benefits when food security shocks are statutory meetings and establishing an imperfectly correlated across the participating information system of institutional stocks. countries (Koester 1986); and (2) economies of RESOGEST comprises the member states scale as operational costs are shared. Regional of ECOWAS and those of CILSS, that is, reserves have thus been advocated for as a tool Chad and Mauritania (ECOWAS 2012). for countering food insecurity (for example, FAO et al. 2011). Some regions in the world have thus • Following a public debate triggered by 148 implemented such regional reserves, including the 2008/09 food crisis and supported by the ASEAN Plus Three Emergency Rice Reserve the G20, ECOWAS created, in February (APTERR) (Briones 2011) and the South Asian 2013 at the ECOWAS Heads of State Association for Regional Cooperation (SAARC) Conference in Yamoussoukro, the RRSA. (Kumar and George 2019). For years, there have Resourced by and in support of ECOWAS also been discussions to set up regional food member states, the RRSA is to act as a reserves, for example for the SADC (Koester regional mechanism for member states 1986) and on a global level (FAO et al. 2011). that need emergency food during food crises. Its target size is 411,554 Triggered by food crises in the 2000- megatons, with 140,000 megatons 2010, West African countries committed held in a physical reserve and 271,554 to implementing two new regional food megatons held in a financial reserve. In reserve cooperation mechanisms. line with international best practice, it does not aim to influence grain prices • In response to the 2005 food crisis, but to support countries affected by West African countries created a network severe shocks for relief measures. The of their national food reserve agencies, RRSA has, to date, been used five times, RESOGEST, that pursues three objectives: although small amounts were disbursed. (1) committing participating countries In October 2020, the physical reserve to set aside 5 percent of national food held 32,000 megatons and was expected had been committed to the financial reserve to increase to about 42,000 megatons by (ECOWAS 2020); (2) Alongside the RRSA growth, March 2021 (ECOWAS 2020; 2012). ECOWAS countries were supposed to gradually also increase their national-level food reserve ECOWAS countries follow a three-tier stocks (also shown in table 8.5), reaching a strategy to counter food insecurity using total of 600,774 megatons by 2020; by 2021 food reserves. ECOWAS has adopted a (year 8), existing food stocks were supposed to regional Food Storage Strategy that categorizes increase by a factor of three to four. However, the use of different types of food reserves into no significant growth of national food reserves three lines of defense: food from local cereal is evident anywhere in ECOWAS member banks is to be used first; national food reserves states so far (Galtier 2019); (3) The costs for second; and regional food reserves third. Thus, establishing the RRSA were supposed to be a principle of subsidiarity is implied in the divided across different actors (table 8.6). The strategy (Galtier 2019). EU, through an ongoing project, provided €38 million in support of the RRSA, with €20 million Progress on establishing the RRSA used for grain purchases. However, ECOWAS has been delayed. The slow progress of member states and the ECOWAS Commission implementing ECOWAS RRSA can be seen have at best committed minimal funds. Initially, across four dimensions: (1) The goal for the there had been plans to introduce an ECOWAS RRSA was to build its physical and financial wide zero-hunger tax to also fund the RRSA, stock progressively from 2013 and reach full but it has not been implemented (Galtier 2019); capacity by year eight, or 2021. The intended (4) Finally, the RRSA operational structures progress can be seen in table 8.5. By 2020, the setup has also been delayed. The reserve 149 physical and financial reserves were supposed Management Committee was officially set up to have reached 100,000 megatons and the and a Manual of Procedures elaborated only in equivalent of 193,967 megatons, respectively. 2018. Management Committee members were By the end of 2020, the physical reserve only appointed and the Manual of Procedures was amounted to 32,000 megatons and no funds validated in 2019 (ECOWAS 2017; 2019). TABLE 8-5 Eight-Year Plan for Building the Regional Reserve and Increasing National Public Stocks (2014–21, in Megatons) Year 0 Year 1 Year 2 Year 3 Year 4 Year 5 Year 6 Year 7 Year 8 RRSA 0 176,380 176,380 176,380 176,380 293,967 293,967 293,967 411,554 Physical 0 60,000 60,000 60,000 60,000 100,000 100,000 100,000 140,000 Reserve Financial 0 116,380 116,380 116,380 116,380 193,967 193,967 193,967 271,554 Reserve National 227,000 360,464 360,464 360,464 360,464 600,774 600,774 600,774 841,083 Stocks Source: ECOWAS 2012 Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens TABLE 8-6 Financing Structure for the Establishment, Maintenance, and Governance of the Regional Reserve (in US$, millions) Year Year 2 Year 3 Year 4 Year 5 Year 6 Year 7 Year 8 1 National government contributions 15 0 0 0 12 0 0 15 (grains) ECOWAS/West Africa Economic and Monetary Union (WAEMU) contribu- 20 15 15 15 15 15 15 15 tions Technical and financial partner contri- 12 12 12 12 12 12 12 12 butions Annual balance (resources–usage) -47.8 20.8 22.6 20.8 -26.9 18.3 21.2 -29.3 Total costs 94.8 6.2 4.4 6.2 65.9 8.7 5.7 71.3 Source: ECOWAS 2012 The regional reserve also may be subject to National available crisis coping capacities must multiple technical design challenges. While be below 66 percent of the needs estimated no full review of the governance structure to respond to the crisis; A crisis plan needs of the RRSA was conducted for this study, to be in place; and The requesting country 150 there are four potential areas of concern: (1) must demonstrate a clear commitment to Resourcing questions seem to be insufficiently refund the received cereal amount to the addressed to achieve adequate and sustainable regional reserve after use. There is reported stocking of the reserve (ECOWAS 2020); (2) The overlap between using country stocks and target total size of the reserve was estimated by those from the regional reserve (ECOWAS considering country-level food needs during 2020). Quantifiable criteria enabling country the worst food crisis year over a 12-year time access could improve RRSA performance and horizon and applying a ratio of food needs to sustainability; and (4) Disbursement rules could be met regionally vs. nationally (ECOWAS 2012). benefit from further review. Specifically, they While taking a historical view seems to be a could outline a framework for integrating in- suitable approach, the chosen methodology kind relief with cash-based operations that may benefit from further refinement using are often considered to be a more effective actuarial principles; (3) Rules for country response. In addition, grain disbursements access to grains from the regional reserve may are primarily conducted as loans, requiring benefit from further review—currently, grains receiving countries to reimburse the grain they are distributed to countries per their request are provided. As a regional reserve, this may not following criteria—Affected areas must be always be the most efficient way of sharing risk. subject to Phase 3 classification in CH analysis; 8.3.2 African Risk Capacity T he ARC is a specialized agency of (MoU) with ARC. They also need to present and the AU offering sovereign drought have approved a contingency plan outlining insurance to member states to support how a potential payout would be spent (ARC emergency relief efforts. Active since 2014, n.d.). ARC operates as a mutual insurance captive for AU member state governments. The idea ARC insurance has been met with a great deal is that by offering insurance for drought- of enthusiasm in West Africa and has also related emergency relief expenditure, ARC can been purchased by five of the focus country improve governments’ financial management governments. ARC insurance contracts are of drought shocks. By providing payouts early, signed on an annual basis. All ECOWAS country pooling drought risk across the continent and governments have signed the ARC MoU but operating on a not-for-profit basis, ARC aims to not all of them have purchased ARC insurance. offer significant cost savings to participating Of the six focus countries, all except for Sierra governments compared with available Leone have purchased ARC drought insurance. commercial insurance options. Before being Only Niger has thus far received an ARC payout able to participate, AU country governments (table 8.7). need to sign a Memorandum of Understanding TABLE 8-7 ARC Sovereign Drought Insurance Cover Over Time in Burkina Faso, Chad, Mali, Niger, Sierra Leone, and Togo 151 Agricultural season Maximum Payout (US$, Country cover 2019–20 millions) 2014–15 2015–16 2016–17 2017–18 2018–19 2019–20 (US$, millions) Burkina - - x x x - - - Faso Chad - - - - - x 1.2 - Mali - - x x - x 15.0 - Niger x x x - - x 5.3 3.5 (2015) Sierra - - - - - - - - Leone Togo - - - - - x 4.5 - Source: PwC 2020, excludes ARC Replica sums insured The current ARC insurance product is based product is a satellite rainfall-deficit (drought) on a rainfall-based crop yield model, but ARC index insurance product for all of Africa termed is developing and testing new approaches, ARV, based on the FAO’s WRSI. ARV monitors including in the focus countries. ARC’s core the rainfall during the growing season and the Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens impacts on drought on vulnerable populations national institutional strengthening, policy and the response costs based on drought development, risk profiling, and contingency severity. The drought policy is sold to the planning for disasters. Five ADRiFi country government in each participating country projects have so far been approved, including and payouts are in turn settled by ARC, which a €4.8 million grant to the government of Niger is then responsible for implementing the (AfDB 2020). Of the focus countries, Burkina preagreed response plan. In recent years, ARC Faso, Chad, and Mali have likewise expressed has worked on expanding its product portfolio their interest in participating in the program and has, for example, developed an alternative (AfDB 2018). drought index insurance product targeting rangeland and thus pastoral rather than Besides government-purchased drought farming populations. ARC is testing the product insurance, ARC also offers replica coverage in selected pilot countries, including in Niger that is being purchased by the WFP in two (ARC 2020a). ARC has also developed a flood of the focus countries. In the focus countries index insurance pilot product that is currently during the 2019/20 season, the WFP purchased being tested in four countries, including in replica cover of US$7 million in Burkina Faso and Togo (ARC 2020b). US$12.6 million in Mali (PwC 2020). ARC replica coverage is an insurance product offered by The AfDB has launched a new program co- ARC that is subject to the same requirements financing ARC premium payment, including as the ARC main drought cover purchased by in the focus countries. In 2018, the AfDB governments. Replica cover can be purchased launched the Africa Disaster Risk Financing by nongovernmental development partners to 152 Program (ADRiFi). Through this program, the receive a potential insurance payout at the same AfDB cofinances ARC premium payments of time as the government, thus contributing to a recipient governments. As accompanying better coordination of response efforts. measures, the program also finances 8.3.3 Other regional efforts on DRF T he World Bank and the ILRI are Senegal). Focusing on pastoral populations, the exploring the feasibility of drought risk project also aims to identify potential concrete finance solutions in Sahel countries, application and implementation modalities as including in three of the focus countries. part of wider drought risk management and The aim of the project is to assess the feasibility pastoral development initiatives. The feasibility of implementing financial protection index- assessment considers technical (product based solutions against drought in four design), socioeconomic (potential demand and Sahelian countries (Burkina Faso, Niger, Mali, value), and operational (supply chain) factors. 8.3.4 Indicative funding gap analysis T he relative lack of disaster risk finance impact of agricultural yield shocks at instruments in most countries analyzed different return periods (1–5 years and creates significant funding gaps in 1–10 years) and the humanitarian cost the response mechanisms to even small- of food insecurity at different return scale disasters. We will present a simplified periods (1–5 years and 1–10 years). way to assess the potential shortfall of funds governments might face after a disaster. The • The funding gap is then expressed objective is to identify the scale of the gap as the as the difference between the total basis for dialogue with governments on options available amounts from financial for strengthening financial preparedness for instruments and the total funding disasters. needs due to disasters. We note that donor assistance is not accounted for Estimating funding gaps precisely requires since this is not a certain source of comprehensive data on public contingent financing. We also note that none of the liabilities for all types of hazards experienced countries in scope for the analysis have by a country, the hazards’ return periods contingent lines of credit. at varying intensities, and all existing ex- ante DRF mechanisms (both formal and • In terms of national disaster funds, informal), together with assumptions on the funds listed in table 8.8 are the scale of fund flow from ex-post DRF considered. These correspond to the 153 instruments. While estimating funding gaps observed values in national disaster precisely requires extensive data that are funds described in section 5.2.1. unavailable, calculations were made based on the desk work analysis of existing national • In terms of SFRs, three types of and regional DRF instruments combined SFRs are considered: local SFRs with available information on probable PML (cereal banks at village or town level), and food insecurity costs from the models national SFRs, and the ECOWAS RRSA. presented in chapter 7. All three forms of SFRs are considered as they form the structure of the ECOWAS • The financial instruments considered Food Storage strategy as described in the analysis are risk retention in section 5.2.2 (figure 8.4). Local SFRs mechanisms (national disaster funds, were assumed to amount to 15,000 local SFRs, national SFRs, regional SFR) megatons in each of the Sahel countries. and risk transfer mechanisms (ARC National (physical and financial) SFRs sovereign insurance and ARC replica were assumed to hold 70 percent of insurance). their target storage capacity. The current physical stock of the regional food • The costs to be covered by the reserve was also considered, even if the instruments are both the economic current level of 32,000 megatons is far Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens TABLE 8-8 US$ Amounts Kept at National Disaster Funds in the Focus Countries Country National Disaster Fund Burkina Faso US$750,000 Chad None Mali US$84,000 Niger US$14,800,000 Sierra Leone None Togo None Source: Zongo 2019; Achirou 2017; Coulibaly 2014 below the targeted level for its eighth percent of the physical storage capacity. year of functioning. Countries have The country-specific quotas have been assigned ECOWAS utilization quotas of applied to the regional reserve’s current the regional reserve and Burkina Faso, storage level. Table 8.9 sums up the Mali, and Niger are the main country available total food storage via different beneficiaries with a total quota of 89.6 SFRs for each country. 154 FIGURE 8-4 ECOWAS Food Storage Strategy FOO RESERVE PRIORITY USAGE Regional SGR-s FOOD INSECURITY (ECOWAS food reserve) SEVERITY OF National SGR-s (SNS/SIE) Local SGR-s (cereal banks) Source: World Bank TABLE 8-9 National and Regional Grain Storage Levels SFR assumed storage level (tons) Country SNS physical SNS financial SI physical (70% of ECOWAS Local SFRs Total (70% of target) (70% of target) target) RRSA Burkina 15,000 35,000 17,500 17,500 5,568 90,568 Faso Chad 15,000 24,500 39,500 Mali 15,000 24,500 17,500 17,850 6,624 81,474 Niger 15,000 35,000 42,000 21,000 16,160 129,160 Sierra 0 Leone Togo 17,500 17,500 Source: Various To include SFRs in the analysis, the food multiplied by the retail prices for each food stocks need to be expressed in monetary crop in each country that were also used for amounts. For each food stock in each country, the crop risk assessment (table 8.10). The US$ the composition of food crops was analyzed, amounts obtained are shown in table 8.11. data allowing34 These compositions were then 155 TABLE 8-10 Retail Price by Type of Grain Held in the Storage Reserve (US$ Per Ton) Sorghum Maize Millet Rice Burkina Faso 290 270 340 530 Chad 310 410 360 Mali 340 340 390 270 Niger 330 380 400 830 Togo 260 Source: WFP-VAM; FAO-GIEWS; FEWS NET Large funding gaps are expected for all With the information on national disaster funds countries, with gaps exceeding US$1 billion and food reserves in terms of US$, we can now for Niger for higher severity events based on calculate the approximate funding gap for the modeled results and findings from the each country for two types of disasters: a lower review of countries’ financing frameworks. severity one (1 in 5 years) and a moderately 34 Food stock compositions found for this analysis are described in the respective country reports in appendix C. Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens TABLE 8-11 Monetary Value of Storage Reserves SFR value of storage level (US$, millions) Country ECOWAS food Local SFRs SNS total SI Total reserve Burkina Faso 4,500,000 15,750,000 4,987,500 1,670,400 26,907,900 Chad 5,400,000 8,820,000 — — 14,220,000 Mali 5,475,000 15,330,000 2,362,560 2,362,560 25,530,120 Niger 5,475,000 28,490,000 5,979,200 5,979,200 45,923,400 Sierra Leone — — — — — Togo — 4,550,000 — — 4,550,000 Source: Various severe one (1 in 10 years). For the 1 in 5 years through ARC and ARC replica, where applicable. event, it was assumed that no resources from The results are shown in table 8.12 and table ARC were available. For the 1 in 10 years event, 8.13. it was assumed that resources were available TABLE 8-12 Funding Gap for a Lower Severity Disaster in the Agriculture Sector 156 1 in 5 years Country Modeled Value of ARC sum in- Modeled National Di- food security Gap (US$, food storage sured (ARC and economic saster Fund humanitarian millions) reserves ARC replica) loss cost Burkina Faso 0.75 26.91 — 136 58 (167) Chad — 14.22 — 111 189 (285) Mali 0.08 25.53 — 336 31 (341) Niger 14.80 45.92 — 525 304 (768) Sierra Leone — — — 106 72 (178) Togo — 4.55 — 105 99 (199) Source: World Bank TABLE 8-13 Funding Gap for a Lower Severity Disaster in the Agriculture Sector 1 in 10 years Country Modeled Value of ARC sum in- Modeled National Di- food security Gap (US$, food storage sured (ARC and economic saster Fund humanitarian millions) reserves ARC replica) loss cost Burkina Faso 0.75 26.91 7.00 192 75 (233) Chad — 14.22 1.20 153 215 (352) Mali 0.08 25.53 27.60 502 46 (495) Niger 14.80 45.92 5.30 765 391 (1,090) Sierra Leone — — — 158 88 (246) Togo — 4.55 4.50 153 118 (262) Source: World Bank The analysis shows that all countries are large gap between the available funding and extremely exposed to disasters of any kind, the expected humanitarian cost, both for less given their limited national and regional severe and for more severe events. In terms of financial instruments available for response. economic cost, none of the countries can buffer Mali is the only country where the value of the economic impact of either 1 in 5 years or 1 food storage reserves is close to covering the in 10 years shocks with the available funding 157 modeled humanitarian need of food insecurity through the analyzed funding instruments. after a disaster. In all other countries, there is a Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens 9 REGIONAL REC- OMMENDATIONS TO STRENGTHEN RESILIENCE AGAINST AGRICULTURAL RISKS national and regional levels will have to be implemented at all layers of risks. This will also require a combination of risk financing instruments at national and regional levels. • The region should prioritize those risks that have the biggest impacts on the region, whether from a country perspective or at crop level. The regional food reserve should be aligned with the risk-based needs. • For several countries, exogenous risks TO FOOD SECURITY transmit through prices and addressing the root causes for the risks in certain countries (for example, Senegal) would benefit consumers in other countries in the region. This chapter compiles the main takeaways As such, there is an incentive for countries in from the analyses in this report and provides the region to collaborate on risk mitigation recommendations on the next steps for in particularly affected areas. strengthening aspects of agrifood systems • Effective ARM in the region requires a risk management at the regional level in comprehensive risk financing approach, 158 response to the challenges identified in West both at country and regional levels. Africa. Regional solutions include regional risk pooling for adaptive social safety nets and creating a risk financing backstop for the Key points in this regional food reserve. • For the analysis, obtaining standardized data proved to be a challenge and has chapter: consequences for the possibility of using index-based triggers—for example, for insurance instruments or scalable social • Production risks are prevalent and the impacts safety nets. Better information systems are on West Africa’s food systems are significant. therefore needed. • The findings of this report show that there • Going forward, it will be important for are opportunities for sharing risk management all countries to strengthen their ability to instruments between countries in West Africa. manage risks and effectively integrate risk The type of risks varies across the regions, but management into general agriculture and most countries see important impacts already food security policy objectives. Country- for low-level risks events. level ASRAs and DRAs should therefore be • Currently, the cost of risk impacts, especially conducted for those countries where such when converted to humanitarian costs, are so assessments are not yet available. high that a combination of risk management measures (mitigation-transfer-coping) at A s this report has shown, food crop differ across the regions—for example, some losses in West Africa are expected countries are more affected by droughts and to be high. According to the risk others by floods—but most countries see assessment performed for the portfolio of similar impacts already at even low-level risks food crops selected for West Africa, the region events. Except for Benin, Côte d’Ivoire, and should expect to suffer food crop losses Ghana, which see risk impacts of below 5 equivalent to 4.50 percent of the portfolio percent, and Niger, which sees 17 percent for exposures or US$4.7 billion annually, or 1.6 1 in 5-year risk events (increasing to 24 and 32 percent not including Nigeria. In addition to percent for 1 in 10 years and one in 25 years the expected losses, the region can expect risk events, respectively) all countries see risk to have huge losses on relatively short impacts of a relatively similar scope, ranging time recurrence periods. The LaR analysis from 7.4 percent for 1 in 5 years risk events performed as part of this study indicates food in the case of Burkina Faso to 12.7 percent in crops portfolio in West Africa may face losses the case of The Gambia. The analysis in this as high as 9.9 percent of total exposures (or report also showed multiple hotspots that US$10.3 billion) once in ten years, 15.9 percent cross country borders, where mitigation and of its total regional exposures (or US$16.7 response mechanisms can be managed jointly. billion) for once in 50 years risk events. This indicates an opportunity to manage certain risks jointly at the regional level. Currently, the cost of risk impacts, Similarly, for production risks, as much of risk especially when converted to humanitarian management will have to be mitigation, there costs, are so high that a combination of risk are multiple areas where the region would management measures at the national and benefit from common instruments around 159 regional levels will have to be implemented research, information systems, and support to at all layers of risks. In addition to the transborder risk hotspots. Further, food markets extensive crop losses described above, for the are unevenly integrated, which in some parts six countries where a financing assessment of the region causes unnecessarily high food was conducted (chapter 8), the financing gap prices. For other countries, risks transmit amounted to US$1.8 billion for humanitarian through prices into countries that trade and needs as a result of the food production addressing risks in certain countries would risks during 1 in 5 year risk events, and close therefore benefit consumers in more countries to US$2.6 billion for 1 in 10 years events. 35 in the region. As such, there is an incentive To be able to effectively respond to risks in for countries in the region to collaborate on the coming decades, especially given the risk mitigation in certain particularly affected expected increase in negative impacts as a areas or in countries where the impacts of result of climate change, it is necessary for the risks transmit into other countries in the countries in the region to manage these risks form of price increases. Reflecting the figures through a layered approach. presented in chapter 2, figure 9.1 lays out the principles for layered risk management. While The findings of this report show that the top layer constitutes a regional layer for there are opportunities to share risk high-level risks—for example, risk financing management instruments between and risk pooling like the RRSA all layers of risk countries in West Africa. The types of risks management comprise regional components. 34 Note that these are the combined financing gaps for the six countries for such level of events, but the events may not strike simultaneously for the six countries over the 5- or 10-year period—that is, the financial needs for the country group can be spread out in the respective period. Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens FIGURE 9-1 Layered Risk Management at Country and Regional Level for West Africa Risk Management strategy Sovereing risk transfers, regional risk pooling Government coping mechanisms; micro- and meso-level transfers Risk mitigation Risk Mitigation Low / IPC 1 Low / IPC 2 Low / IPC 3 Level of Risk Source: World Bank 160 9.1 REGIONAL RISK MANAGEMENT IN WEST AFRICA Y am, cassava, and rice crops, at the crops should also be targeted with ex-ante regional level, show the highest losses interventions in combination with ex-post in terms of value, so risk management measures such as the RRSA or sovereign risk measures should seek to decrease the transfer instruments. impacts of risks for these crops. These three crops account for 69.8 percent of the total At the regional level, certain countries and exposures and explain 77.5 percent of the areas emerge as those contributing the expected AAL for the crops assessed in this most to the overall risks to regional food report. Any type of ex-post risk management production, but resources should also mechanism is likely to be costly and for countries seek to decrease risk in targeted hotspots. that have a smaller share of these crops in their Overall, the analysis showed that for the region, portfolio, joint risk sharing instruments may be the biggest food crop production losses can less attractive. Nevertheless, crops for which be found in Nigeria, which accounts for 66.4 value is lower but on which a large share of the percent of the value of total food crop losses population in some countries depends on for in the region, followed by Ghana (6.6 percent), food security (for example, millet and sorghum) Niger (6.0 percent), and Côte d’Ivoire (5.8 may also come at high costs for humanitarian percent). If countries are to collaborate around response. As for yam, cassava, and rice, these regional risk response instruments, better managing risks in these countries would be depending on the type of crops and countries. a priority as they contribute to large shares For example, prices for millet, sorghum, and of the region’s total losses. At the same time, yam are more susceptible to impacts from the analysis identified certain transboundary other risks. As prices in Chad, Mali, and Niger hotspots that are particularly exposed to risks are sensitive to outside shocks, these countries (for example, south-southwest Burkina Faso– should have an incentive to participate in risk northwest Ghana and eastern Burkina Faso– mitigating activities also outside of their own western Niger–north Benin) and even though countries. Similarly, multiple countries in the some of them are part of low-risk countries, region would have an interest in mitigating these areas should also be prioritized under risk impacts on rice and millet in Senegal, as more targeted risk management measures. this is transmitted to prices in other countries (including Burkina Faso, Mali, and Niger), as ECOWAS should seek to promote areas would Togo and Chad in reducing risk impacts for collaboration around agricultural risk on sorghum in Mali. As the analysis may also mitigating measures. This report mainly suffer from lack of reliable data at both Admin focuses on regional solutions to ARM but while risk 1 level and Admin 0 level, the West Africa transfers and risk response are more traditional MIS network needs to be strengthened. This areas for regional collaboration, there are areas will allow access to data from all countries to of risk mitigation where regional collaboration provide up-to-date and accurate information. already exists and should be strengthened. Examples of these are weather information Align the food reserve with the risk-based systems and pest and disease information needs and integrate it in the layered risk systems, digital farmer-to-farmer extension management approach. This study has 161 systems, regional agricultural research centers provided detailed analysis of the scope of and seeds recognition systems, and so on. As risks impacts at various levels of risks and how discussed earlier, the regional food security this translates into humanitarian aid for six impacts of certain country-level risks can also countries in the region. The enormous gap motivate collaborative solutions and regional for both the reserve and other risk financing support for country-level investments, such as should be strengthened. While section 9.2 in from the disaster risk fund. These collaborative this chapter provides more detailed advice on approaches should also target cross-border how to strengthen the financial stock of the hotspot areas—for example, around drought reserve, the findings in this report can provide risks to pastoral systems, agriculture pest guidance on where assistance will be needed and disease systems, and similar areas where and with what types of food crops. Importantly, asymmetrical management of risks between the Admin 1 level assessment in hotspot areas countries diminishes the positive effects can give detailed guidance on where to place impacts of disease management in individual community-based reserves—that is, the RRSA’s countries. first line of defense. Analysis in this report shows that weather- Developing a standardized, joint system related risks and social unrest affect prices for agricultural production data. For the in the region beyond the domestic markets. analysis, obtaining standardized data proved Thus, countries should have an interest in jointly to be a challenge and lower-level data (Admin seeking to mitigate the impacts of weather- 1 level) was only available for a limited set of related risks. The level of impacts on prices varies countries (see chapter 5). In the analysis, this Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens meant that it was not possible to establish Given the high levels of risk in the region robust correlations between weather- and the mutual benefits for countries in related risk events and yield losses or food collaborating around managing these risks, insecurity. This, in turn has consequences for regional risk management capacity should measuring the impacts on agriculture and be strengthened—for example, by creating risk management policy and, importantly, the a regional technical support unit (TSU) possibility of using index-based triggers, for on agrifood-sector risk management and example for insurance instruments or scalable disaster risk finance. Preparing and structuring social safety nets. Better information systems ARM and DRF solutions frequently requires very are therefore needed at the national level, but specific technical skills that may not always they should be harmonized at the regional level be easily accessible by country authorities, through common standards for application in including, for example on risk assessment, risk shared risk management systems, such as the modeling, risk layering, actuarial analytics, and RRSA and regional transfer instruments. the structuring of insurance markets. A regional TSU could be set up to provide countries with Explore the possibility of setting up support to understand countries’ risk exposure a pasture information system and and vulnerability better and to design and the collaboration around pasture implement appropriate financing solutions. management. The assessment of risk impacts Potential tasks of a regional TSU could include on pastures showed AAL of 3.4 percent across (1) supporting countries in conducting ASRAs the region, increasing to 14.4 percent loss and developing standardized, probabilistic once in 100 years. This aggregate loss estimate risk profiles; (2) building capacity on technical 162 is likely masking much higher variations issues related to ARM and risk financing; (3) at local levels, and as systems are fragile, sharing best practice guidance and facilitating small variations in vegetation can have large south-south knowledge exchange36 on ARM impacts. Many pastoralist systems in the region and risk financing strategies and policy; and (4) cross borders and there may be incentives to pool national and subnational data to develop collaborate around regional pasture EWS and a monitoring platform for weather events, harmonize pasture management to mitigate price variations, disease outbreaks, food stocks, risk impacts on pastures. Satellite information and populations at risk, which could enable enable EWS for pastoralists and can guide more effective EWS across the region. The TSU herders away from areas experiencing high could be housed at the ECOWAS Commission losses in a season. A precondition for such or at a technical partner such as AGRHYMET. systems, however, is that all parts manage their Linkages could be exploited, for example, with pastures and thus joint agreements on pasture ECOWAS activities on disaster risk mitigation management, herd size and composition, and and reduction or with climate and data services research exchange, together with a system for provided by AGRHYMET. enforcement, should accompany any regional pasture information system. 36 South-South Knowledge Exchange connects policy makers and development practitioners within and across countries to learn from each others' experiences and to identify workable development solutions and policies. 9.2 REGIONAL RISK FINANCING F or the countries in the region that that the response to shocks that affect the are exposed to similar production region are better coordinated. It should be and humanitarian shocks, there is a noted that while the figure 9.2 presents strong rationale for a regional approach recommendations on potential agricultural to financing these countries. This was risk financing solutions for the focus countries, confirmed by the six- country humanitarian any such solution must be embedded in a needs assessment conducted in chapter 7. broader ARM framework. As described in Regional financing solutions could make use chapter 8, risk financing solutions can only of risk diversification benefits and operational be part of a comprehensive approach to risk economies of scale and thus lower overall management and require the existence of cost. They could also facilitate the sharing of other elements. Where relevant, this has been DRF expertise across countries and ensure pointed out below. FIGURE 9-2 Summary of National and Regional Options to Strengthen Financial Resilience for Food Security by Type of Shocks Risk nance backstop for Regional food reserve High Severity Regional SP scale up nancing mechanism National Regional Shocks risk technical 163 Establish sub- regional country aggregation for assessment support maximun positive externalities and synergies unit on National disaster risk DRF nance Low severity shocks National risk retention mechanisms and strategies disater risk reduction Chronic food insecurity National investments in agricultural productivity and SP system Source: World Bank For any regional risk response solution, countries seem to be mostly connected consider the Sahelian and the coastal to droughts, in the coastal countries they countries separately from each other. Sahelian are mostly due to floods. In addition, both and coastal countries’ risk profiles are very country groups display very different food different: both groups are situated largely security vulnerability profiles. There are larger in different agroecological zones, with the chronically food-insecure populations in the Sahelian countries experiencing a similar mix Sahel than in the coastal countries and a of food crops as well as mostly an arid and greater share of the population at risk of falling semiarid climate, whereas coastal countries are into becoming food insecure. These important largely humid and subhumid. Whereas climatic differences should be considered when agricultural production shocks in Sahelian designing any regional risk financing solution, Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens as the insurable risks will not be similar, nor • Lower operational cost through will the sums insured that are needed (ARC has shared infrastructure. Rather than in place coverage for some of these countries, operating a similar financing structure but given the funding gaps estimated, these in each Sahel country, a regional insured amounts are insufficient). Adapting solution offers the opportunity to share to local needs, diminishing basis risk by a infrastructure and thus make use of thorough risk assessment, will make risk operational economies of scale. For transfer affordable, while increasing coverage. example, program institutional and technical design could be standardized • Consider establishing a regional across countries; there could be a single risk pooling solution for adaptive program coordination body and data social safety nets. The World Bank Sahel management systems could be shared. Adaptive Social Protection Program (SASPP) supports the governments • Risk diversification benefits. A of the four Sahel focus countries in regional initiative could pool climatic the development of shock-responsive (for example, drought) risk across a safety nets. The shock-responsive greater geographical area and diversity elements are still being developed than in any one single country. While the and institutionalized by the respective underlying risk exposure for individual governments. Once operational, countries would remain the same, a they will require flexible financing four-country regional pool should offer arrangements to provide rapid resources the potential for geographical risk 164 in case of a shock for horizontal or diversification—namely, the fact that vertical expansion. Such financing even during a severe event, not all four could be pooled and prearranged at the countries would be equally affected in regional level—for example, through a the same year. For a potential regional regional fund or a common risk transfer risk transfer solution—for example, mechanism. Expected benefits vis-à- with ARC—this would likely offer cost- vis each country pursuing their own effective gains compared to countries financing solution include the following: insuring their individual risk. • The sharing of expertise. Structuring A schematic of the potential structure is a prearranged financing solution is provided in figure 9.3. a complex technical area requiring understanding of risk modeling, risk layering, actuarial analytics, and potentially the structuring of insurance markets. For many countries it can be difficult to develop this knowledge locally. Establishing a regional program that centralizes and shares expertise on the respective technical areas can be an effective investment from a country perspective. FIGURE 9-3 Schematic of Potential Regional Risk Transfer Solution for Adaptive Social Protection of Sahel Countries Reinsurance Burkina Faso Regional risk SP project nance for social protection and Chad food security SP project Mali SP project Niger SP project Source: World Bank 165 Explore a risk finance backstop to the RRSA. Benefits of the approach include (1) creating The RRSA has been conceptualized by ECOWAS dependable financing for severe shocks as the third line of defense behind local and and protecting RRSA capital; (2) ensuring national food stocks. While its establishment rapid RRSA replenishment at the occurrence has been delayed, there continues to be a of a shock; (3) using potential risk pooling strong rationale for a regional reserve, as benefits—for example, by transferring parts witnessed by regional reserves established of the risk to the insurance markets; and (4) in other parts of the world. A risk financing helping operationalize the RRSA as additional backstop to the RRSA could help operationalize financing might be used as an incentive for it by ensuring that it receives the required ECOWAS and countries in the region to build resources to intervene at the occurrence of a greater ownership of the mechanism. When shock. The backstop could be in the form of implementing such a financing solution, a prearranged funding window or a formal countries should also consider using the insurance arrangement (for example, with opportunity to review the technical design ARC). The solution could provide finance to of the facility and potentially revise country the RRSA upon the occurrence of objective access, procurement, disbursement, and data, such as indicating the occurrence of a financing rules in light of the difficulties that drought in a participating ECOWAS country. the RRSA has experienced in recent years. Triggers would have to be carefully aligned Figure 9.4 presents a schematic of the potential with the country access rules to the reserve. structure of the RRSA risk finance backstop. Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens FIGURE 9-4 Schematic of Potential Risk Finance Backstop to ECOWAS Regional Food Security Reserve Prearranged nance / reinsurance ECOWAS Regional Reserve ECOWAS Country 1 ECOWAS Country 2 ECOWAS Country 3 166 ECOWAS Country 4 ECOWAS Country 5 Source: World Bank 9.3 STRENGTHENING RISK MANAGEMENT AT THE NATIONAL LEVEL TO ENABLE REGIONAL RISK SOLUTIONS A precondition for regional risk necessary risk management measures being management solutions, and in place at the national level are likely to be especially response instruments overly costly and transfer too much of the such as regional food reserves and regional risks to other countries in the region. If some risk financing instruments, is that countries countries manage risks better at the national effectively manage their lower layers of level and mitigate their impacts, while others risks. As discussed in chapter 2, regional do not, this may also constitute a disincentive risk management instruments without the for the former group to participate in risk pooling instruments. Further, the existing for how to cover expected shock-related financing gap that was shown for the six costs. Such a strategy determines the most focus countries is too high to share with other suitable sources of financing and instruments countries and is likely too costly to transfer (retention and transfer) to deal with the against a premium. Effectively managing different types and severities of disasters. It lower-level risks at country and local levels also helps identify gaps and opportunities for is therefore a precondition for several of the improving the current funding framework in regional recommendations. It was not within order to increase the efficiency, timeliness, and the scope of this report to conduct detailed coherence of different financial mechanisms country-level agriculture risk assessments; and the transparency of disaster response nevertheless, this report recommends three interventions. important next streps for the countries in the region to bring ARM to the next level. Invest in agricultural productivity and social protection and in agriculture risk Identify, quantify risk, and prioritize risks mitigating measures. at country and local levels to strengthen ARM and bolster national financial • Invest in risk mitigation and risk planning for shocks.37The food crop risk reduction for lower-level risks. assessment analysis at the country level that All focus countries show significant was conducted for this report provides a exposure and vulnerability to frequent, general overview of the risk faced by food low-level agricultural production risks. crops in the region; nevertheless, further For such risks, investments in mitigating analysis at the subnational level and into the measures to prevent impacts of the 167 scope of different risk is a needed design of a risk, rather than in preparedness and risk management and risk financing strategy response, tend to be most cost-effective for food crop production for individual and appropriate. Such measures will countries. While there are a few exceptions include investing in national seeds in the region (for example, Ghana, Niger, systems and in regional seeds testing and Senegal), few countries have conducted recognitions, promoting climate-smart ASRAs or DRF diagnostic studies. This is practices, expanding the coverage key to prioritize risks and develop effective and use of irrigation and drainage risk management strategies. As discussed, technology, enabling extension services the layered approach requires a nuanced and access to risk management tools approach to risk management depending on such as pest and disease management, the nature and scope of risks, and especially, and strengthening agricultural MIS. The risk mitigation and risk reduction requires exact measures for each country will detailed insights into the nature and root depend on their respective prioritized causes of risks in particular locations. risks, target crops, localities, and so Further, such analyses can guide national on, which the ASRAs are useful in DRF strategies with individual country plans 36 In this context, countries may draw further use of the food crop risk assessment Excel-based tool developed for this report and described in Box 4.1. The tool enables the user to assess the risk for different scenarios combining different crops and countries and is designed to work both at the level of crop and with different aggregation levels in the portfolio. It can be applied to data at different administrative levels. Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens identifying. The scope of the country- Strengthen national response finance wide production risks, priority in terms systems for lower-severity shocks. The of crops, and for some countries, the analysis has shown that even less severe, most affected localities, can be found in more frequent production shocks can have the background report West Africa Food devastating consequences for food security Crop Risk Assessment. in the focus countries. At the same time, the existing national financial instruments are • Invest in agricultural productivity insufficiently resourced to accommodate and social protection to address such shocks. Investing in means to address chronic food insecurity and to create less severe, more frequent shocks should be space for risk response instruments. prioritized over less frequent, more severe Many of the countries in West Africa are shocks in order to ensure that the assets of marked by relatively high levels of food vulnerable populations are not depleted every insecurity. This is particularly the case few years. Lower-level shocks also tend to be for the four Sahel countries, where even overlooked by the international humanitarian during the best of the last eight years, community, while for the more extreme ones, more than 15 percent of the population more humanitarian support is often available. were considered to be at least in CH For such lower-level events, national risk Phase 2 (stressed), indicating a very high retention instruments such as disaster funds level of chronic food insecurity in these and local and national food reserves may be countries. But also in coastal countries, better suited to provide financing than regional food insecurity remains an important ones that are typically more appropriate to problem. Risk financing instruments are provide backup finance for the more extreme 168 not suited to address issues of chronic disasters. Policymakers should also consider food insecurity but instead deal with the establishing suitable delivery channels for financial impacts of shocks. However, such established response finance systems. among the six countries included in the For example, scalable safety nets are one key risk financing assessment (chapter 8), mechanism to disburse funds rapidly and cost the study found that the Sahel countries effectively and their adoption or expansion use their existing relief financing should be considered in all focus countries. resources largely to address chronic food insecurity rather than shocks. Thus, in order to create the space for risk finance instruments to operate, the first step is to address expenditure related to chronic food insecurity. Investments in agricultural productivity and crowding in the private sector as well as in social protection can be useful approaches to address this. To a somewhat lesser extent, this also applies to the two coastal countries where chronic food insecurity is less pervasive but low agricultural productivity and poverty are still challenges to be addressed. 9.4 AREAS FOR FUTURE ASSESSMENT B roadening the concept of food security technical expertise and capital from financial risks. For future work, it would be valuable systems in the resilience of food systems may to include a more comprehensive food be an area that countries in the region are basket in this analysis, as a nutritious diet is interested in pursuing in the future. the basis for food security, and to assess risks along the supply chains in line with the ASRA methodology. While the report tries to extend beyond production of risks to also include impacts on food security and prices from production shocks, it was beyond the scope of the report to elaborate the concept further. Nevertheless, expanding the assessment under future work, especially for country specific ASRAs, should be considered. Climate change and tail risk. The impact of climate change is likely to further increase the impacts of extreme risk events. It was beyond the scope of this report to conduct such analysis, especially given the regional focus of 169 this report, and that agricultural production will be differently exposed depending on the location in the region. Nevertheless, understanding to what extent crops such as cassava and maize will be more exposed to extreme drought events in the future, which will affect production, will be important for future agricultural sector investments in the region. The existing literature is patchy on these aspects, and it is therefore recommended that this type of assessment is included under the proposed country-level ASRAs. Assessing the potential contribution from the financial sector. The financial sector could build and support resilience in the agriculture sector amid an increase in damaging risk events. For example, the Afrexim Bank will soon launch Food Emergency Contingent Trade Financing Facility (FECONTRAF). Further exploring the possibility of leveraging the Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens Achirou, Y. A. M. 2017. Financement des Risques de Catastrophes: Une Etude sur le Mecanisme de Financement des Risques de Catastrophes au Niger. Washington, D.C. : World Bank Group. https://imagebank2.worldbank.org/ search/32939736 ACLED (Armed Conflict Location & Event Data Project). https:// acleddata.com/data-export-tool/ AfDB (African Development Bank). 2018. “African Development Bank Rolls out Program to Boost Climate Risk Financing and Insurance for African Countries.” African Development Bank: Building Today, a Better Africa Tomorrow. African Development Bank Group. https://www.afdb.org/en/ news-and-events/african-development-bank-rolls- out-programme-to-boost-climate-risk-financing-and- insurance-for-african-countries-18618 AfDB. 2020. “Niger: Don de plus de Quatre Millions d’euros Au Program de Financement de La Gestion Des Risques de Catastrophe En Afrique (ADRiFi).” African Development Bank: Building Today, a Better Africa Tomorrow. African Development Bank Group. September 24, 2020. https:// www.afdb.org/pt/news-and-events/press-releases/ niger-don-de-plus-de-quatre-millions-deuros-au- programme-de-financement-de-la-gestion-des-risques- de-catastrophe-en-afrique-adrifi-37992 170 References Allen, Thomas, Philipp Heinrigs, and Inhoi Heo. 2018. “Agriculture, Food and Jobs in West Africa.” April. https:// doi.org/10.1787/dc152bc0-en. Alpha, A., and B. Pemou. 2019. Articuler Stocks Publics et Stocks de Proximité Pour Améliorer La Sécurité Alimentaire Au Burkina Faso. Food Reserves, Working Paper #3. Development Alternatives Incorporated (DAI). ARC. n.d. “How the African Risk Capacity Works: African Risk Capacity.” Accessed January 11, 2021. https://www. africanriskcapacity.org/about/how-arc-works/. ARC (African Risk Capacity). 2020a. “ARC Natural Hazard Risk Modeling.” Webinar presentation at sensitization workshop for IGAD member states. ARC. 2020b. “Updates: African Risk Capacity; The African Risk Capacity (ARC) Group Positions Itself as a Leader in Disaster Risk Management in Africa.” https://reliefweb.int/ report/world/african-risk-capacity-arc-group-positions- itself-leader-disaster-risk-management-africa Benzie, M., 2015. “National adaptation plans and the indirect impacts of climate change”. Stockholme Environmental Institute. Technical report. Available at: https://www. weadapt.org/ sites/weadapt.org/files/legacy-new/ initiatives/ files/49/5475ef5f947cesei-pb-2014- Centre for Disaster Protection. 2020. Centre For Disaster indirectclimate-impacts-naps.pdf [Accessed July 22, Protection: Glossary of Terms. Centre for Disaster 2015]. Protection. https://www.disasterprotection.org/glossary. Blein, R., Soulé, B. G., Dupaigre, B. F., & Yérima, B. (2008). Les Cervigni, Raffaelo, and Michael Morris. 2016. “Confronting potentialités agricoles de l’Afrique de l’Ouest (CEDEAO). Drought in Africa’s Drylands: Opportunities for Enhancing Paris: Fondation pour l’agriculture et la ruralité dans le Resilience.” Washington, DC: World Bank Group. https:// monde (FARM), 116. https://fondation-farm.org/IMG/pdf/ openknowledge.worldbank.org/handle/10986/23576. etudepotentialites_rapport.pdf Clarke, D., and R. V. Hill. 2013. “Cost-Benefit Analysis of the African Bouët, A., & Odjo, S. P. (2019). Africa agriculture trade monitor Risk Capacity Facility.” SSRN Scholarly Paper, ID 2343159. 2019. International Food Policy Research Institute. https:// Rochester, NY: Social Science Research Network. https:// www.ifpri.org/publication/africa-agriculture-trade- doi.org/10.2139/ssrn.2343159. monitor-2019 Club du Sahel et de l’Afrique de l’Ouest. 2010. Étude Régionale Bowen, T., C. del Ninno, C. Andrews, S. Coll-Black, U. Gentilini, K. sur Les Stocks d’Urgence en Afrique de l’Ouest et au Johnson, Y. Kawasoe, A. Kreyziu, B. Maher, and A. Williams. Sahel. https://www.inter-reseaux.org/wp-content/ 2020. Adaptive Social Protection: Building Resilience uploads/Etude_Regionale_sur_les_Stocks_d_Urgence_ to Shocks. World Bank Group. https://openknowledge. en.pdf. worldbank.org/handle/10986/33785. Collins, Jennifer. 2011. “Temperature Variability over Africa.” Briones, Roehlano. 2011. Regional Cooperation for Food Security: Journal of Climate 24 (July): 3649–66. https://doi. The Case of Emergency Rice Reserves in the ASEAN Plus org/10.1175/2011JCLI3753.1. Three. Asian Development Bank. https://www.adb.org/ publications/regional-cooperation-food-security-case- Coulibaly, M. 2014. «Fonds national d’appui à l’agriculture: Le emergency-rice-reserves-asean-plus-three. Comité de Pilotage Précède Aux Ultimes Réglages.» Niarela.Net. http://niarela.net/economie/fonds-national- 171 Brown, M. E., & Kshirsagar, V. (2015). “Weather and international dappui-a-lagriculture-le-comite-de-pilotage-prcede-aux- price shocks on food prices in the developing world”. ultimes-reglages. Cournot. 1971. Global Environmental Change, 35, 31-40. https://doi. org/10.1016/j.gloenvcha.2015.08.003 Diallo, Ismaila, Mouhamadou Sylla, Moctar Camara, and Amadou Gaye. 2012. “Interannual Variability of Rainfall over Brown, Molly, Beat Hintermann, and Nathaniel Higgins. the Sahel Based on Multiple Regional Climate Models 2009. “Markets, Climate Change, and Food Security Simulations.” Theoretical and Applied Climatology 113, in West Africa.” Environmental Science & Technology 351–362 (October). https://doi.org/10.1007/s00704-012- 43 (November): 8016–20. https://doi.org/10.1021/ 0791-y. es901162d. DID (Développement international Desjardins). 2016. “DID to Bush, R. (2010). Food riots: Poverty, power and protest 1. Structure Rural Finance in Mali.” https://www.did.qc.ca/ Journal of Agrarian Change, 10(1), 119-129. https://doi. en/news/did-structure-rural-finance-mali-138/. org/10.1111/j.1471-0366.2009.00253.x Dosio, Alessandro, Richard G. Jones, Christopher Jack, Cabot Venton, C. 2018. Economics of Resilience to Drought. Christopher Lennard, Grigory Nikulin, and Bruce USAID. https://www.usaid.gov/sites/default/files/ Hewitson. 2019. “What Can We Know about Future documents/1867/Summary_Economics_of_Resilience_ Precipitation in Africa? Robustness, Significance and Final_Jan_4_2018_BRANDED.pdf. Added Value of Projections from a Large Ensemble of Regional Climate Models.” Climate Dynamics 53 (9): Centre for Disaster Protection. 2019. Stopping Disasters 5833–58. https://doi.org/10.1007/s00382-019-04900-3. Devastating Lives: Centre for Disaster Protection’s Strategy 2019–2024. Centre for Disaster Protection. Doso Jnr, Stephen. 2014. “Land Degradation and Agriculture United Kingdom https://static1.squarespace.com/ in the Sahel of Africa: Causes, Impacts and static/5c9d3c35ab1a62515124d7e9/t/5de425dab93c Recommendations.” Journal of Agricultural Science c92d39c0f1f6/1575233014799/Centre_For_Disaster_ and Applications 3 (September): 67–73. https://doi. Protection%E2%80%99s_Strategy_2019-2024.pdf. org/10.14511/jasa.2014.030303. Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens ECOWAS (Economic Community of West African States). 2012. FEWS NET (Famine Early Warning Systems Network). 2019. “Regional Food Security Reserve.” https://www.inter- Tchad: Perspectives Sur La Sécurité Alimentaire; Février reseaux.org/wp-content/uploads/Faisabilite_Reserve_ à Septembre 2019. https://fews.net/sites/default/files/ Regionale_EN.pdf. documents/reports/TD_OL_2019_Final.pdf. ECOWAS. 2017. “Echos des Stocks Alimentaires : Bulletin Galtier, F. 2019. “Can the ECOWAS Regional Reserve Project d’information trimestriel de la Réserve régionale de Improve the Management of Food Crises in West sécurité alimentaire de la Cedeao,” 8. http://reserve.araa. Africa?” Food Reserves Working Paper #4. Development org/images/Bulletin/Bulletin-N1---decembre-2017.pdf. Alternatives Incorporated (DAI). https://europa.eu/ capacity4dev/file/90201/download?token=d_lFmoSn. ECOWAS. 2019. “Echos Des Stocks Alimentaires : Bulletin d’information trimestriel de La Réserve régionale Gautier, D., D. Denis, and B. Locatelli. 2016. “Impacts of Drought de sécurité alimentaire de la Cedeao.” https:// and Responses of Rural Populations in West Africa: A www.ecowas.int/wp-content/uploads/2019/03/ Systematic Review.” Center for International Forestry Bulletindinformation-N5-6.pdf. Research (CIFOR). 2016. https://doi.org/10.1002/wcc.411. ECOWAS. 2020. “Briefing Note on the Regional Storage System Giertz, Asa; Braimoh, Ademola; Mwanakasale, Alex; Chapoto, and Its Prospects in the Context of the PRSP Formulation.” Antony; Rubaiza, Rhoda; Chisanga, Brian; Mubanga, ECOWAS, Division of the RRSA/RAAF. Ngao; Samboko, Paul; Obuya, Grace. 2018. “Increasing Agricultural Resilience through Better Risk Management Elbehri, A., Kaminski, J. & Samake, M. (2013). An assessment in Zambia”. World Bank, Washington, DC. © World of sorghum and millet in Mali and implications for Bank. https://openknowledge.worldbank.org/ competitive and inclusive value chains. Rebuilding West handle/10986/29779 License: CC BY 3.0 IGO.” Africa’s Food Potential: Policies and Market Incentives for Smallholder-Inclusive Food Value Chains. FAO/IFAD, GIIF (Global Index Insurance Facility). 2017. “Global Index Rome, 481-500. Insurance Facility Result Stories: Mali.” https://www. 172 indexinsuranceforum.org/publication/result-stories-mali. EM-DAT (Emergency Database). Université Catholique de Louvain. [please provide date – cited table 1.1] Gilbert, C. L., & Morgan, C. W. (2010). Food price volatility. Philosophical Transactions of the Royal Society B: Fackler, P. L. (1988). Vector autoregressive techniques for Biological Sciences, 365(1554), 3023-3034. structural analysis (No. 1908-2017-1838). GIZ (Gesellschaft für Internationale Zusammenarbeit). 2016. FAO. 2002. Food, Security, Justice and Peace. https://www.fao. “Innovations and Emerging Trends in Agricultural org/worldfoodsummit/msd/y6808e.htm Insurance.” https://www.giz.de/en/downloads/giz-2016- en-innovations_and_emerging_trends-agricultural_ FAO. 2014. “Burkina Faso: Country Fact Sheet on Food and insurance.pdf. Agriculture Policy Trends.” FAPDA. http://www.fao.org/3/ i3760e/i3760e.pdf. Golub, S. S., Mbaye, A. A., Diop, C. A., & Igué, J. O. (2019). Benin’s informal trading. https://edi.opml.co.uk/wpcms/wp- FAO. 2017a. Améliorer l’efficacité et l’efficience de La Stratégie de content/uploads/2019/08/08-ch8-Benin-Nigeria_new.pdf Stockage Public Au Mali. Partie 2: Diagnostic. http://www. fao.org/in-action/mafap/resources/detail/en/c/877008/. Harsh, E. 2008. Price protests expose state faults. Africa Renewal https://www.un.org/africarenewal/magazine/july-2008/ FAO. 2017b. “The Future of Food and Agriculture: Trends and price-protests-expose-state-faults Challenges.” https://www.fao.org/3/i6583e/i6583e.pdf. Hollinger, Frank, John M. Staatz, FAO (Food and Agriculture FAO (Food and Agriculture Organization). 2018. “The State of Organization of the United Nations), AfDB (African Agricultural Commodity Markets 2018: Agricultural Trade, Development Bank), and ECOWAS (Economic Community Climate Change and Food Security”. Rome. of West African States). 2015. “Agricultural Growth in West Africa: Market and Policy Drivers.” Rome, Italy: FAO and the FAO, IFAD, IMF, OECD, UNCTAD, WFP, WBG, WTO, IFPRI, and UN AfDB. http://www.fao.org/3/i4337e/i4337e00.htm. High Level Task Force on Global Food and Nutrition. 2011. “Price Volatility in Food and Agricultural Markets.” https:// IRI (International Research Institute). 2015. Programme Africain doi.org/10.1596/27379. d’Adaptation et de Sécurité Alimentaire (PAA), PNUD Niger. Rapport Final. UNDP and IRI, Earth Institute, Global change biology, 26(7), 3753-3755., https://doi. Columbia University. org/10.1111/gcb.15137. Jalloh, Abdulai, Gerald Nelson, Timothy Thomas, and Robert Ngcamu, B.S.; Chari, F. Drought Influences on Food Insecurity in Zougmoré. 2013. West African Agriculture and Climate Africa: A Systematic Literature Review. Int. J. Environ. Res. Change: A Comprehensive Analysis. Washington, DC: Public Health 2020, 17, 5897. https://doi.org/10.3390/ International Food Policy Research Institute. https://doi. ijerph17165897. org/10.2499/9780896292048. Ninno, C. del, S. Coll-Black, and P. Fallavier. 2016. “Social Jayne, T. S., Myers, R. J., & Nyoro, J. (2008). The effects of NCPB Protection Programs for Africa’s Drylands.” https:// marketing policies on maize market prices in Kenya. openknowledge.worldbank.org/handle/10986/24817. Agricultural Economics, 38(3), 313-325. O’Brien, C., Z. Scott, G. Smith, V. Barca, A. Kardan, R. Holmes, C. Koester, Ulrich. 1986. “Regional Cooperation to Improve Food Watson, and J. Congrave. 2018. “Shock-Responsive Social Security in Southern and Eastern African Countries.” Protection Systems Research: Synthesis Report.” Oxford Research Reports 53, IFPRI. https://econpapers.repec.org/ Policy Management. https://www.opml.co.uk/files/ paper/fprresrep/53.htm. Publications/a0408-shock-responsive-social-protection- systems/srsp-synthesis-report.pdf?noredirect=1. Kumar, Nagesh, and Joseph George. 2019. “Regional Cooperation for Sustainable Food Security in South Asia.” Routledge OCHA (UN Office for the Coordination of Humanitarian Affairs) India. https://doi.org/10.4324/9780429346507. Financial Tracking Service (FTS). 2020. https://fts.unocha. org/. Kummu, M., Heino, M., Taka, M., Varis, O., & Viviroli, D. (2021). Climate change risks pushing one-third of global food OCHA. 2020. “West and Central Africa: Flooding Situation As of 6 production outside the safe climatic space. One Earth, November 2020.” https://reliefweb.int/report/niger/west- 4(5), 720-729. and-central-africa-flooding-situation-6-november-2020. Lottini, M. 2014. “An Instrument of Intensified Informal Mutual OECD/SWAC (Organisation for Economic Co-operation and 173 Assistance: The Internal Market Information System (IMI) Development/Sahel and West Africa Club). 2020.“The and the Protection of Personal Data.” European Public Law Geography of Conflict in North and West Africa.” West 20 (1) Volume 20, Issue 1 (2014) pp. 107 – 125. African Studies. Paris: OECD Publishing. https://doi. org/10.1787/02181039-en. Lung, F. 2020. “Being Timely: Creating Good Triggers and Plans in Disaster Risk Financing.” Centre for Disaster Protection. PARM (Platform for Agricultural Risk Management). 2018. https://www.disasterprotection.org/s/Centre_DRF_ “Agricultural Risk Management Tools: Module 3 of Paper3_9Oct.pdf. e-Learning Course ‘Agricultural Risk Assessment and Management for Food Security in Developing Countries’.” Luxembourg Space Agency. 2020. “Transforming Agriculture International Fund for Agricultural Development. https:// Insurance through Technology.” http://space-agency. p4arm.org/app/uploads/2019/01/PARM_Agricultural-risk- public.lu/en/news-media/news/2020/ibisa_and_intech. mangement-for-food-security_Oct2018.pdf. html. Projet d’Appui au Pastoralisme au Sahel (PRAPS). 2017. “L’élevage Mahul, O., and C. J. Stutley. 2010. Government Support to pastoral au Sahel et en Afrique de l’Ouest.” https://www. Agricultural Insurance: Challenges and Options for inter-reseaux.org/wp-content/uploads/int-17-broch- Developing Countries. Washington, DC: World Bank. pastoralismefr-bd-corr.pdf https://doi.org/10.1596/978-0-8213-8217-2. PwC. 2020. “African Risk Capacity Insurance Company Limited Maliszewska, M., & Ruta, M. (2020). The African Continental Free Notes to the Audited Financial Statements, December 31, Trade Area: Economic and Distributional Effects. World 2019.” https://www.africanriskcapacity.org/wp-content/ Bank Group. https://www.gtap.agecon.purdue.edu/ uploads/2020/05/ARCLtd_2019_Audited_Financial_ resources/download/9747.pdf StatementsEN.pdf. Meynard, C. N., Lecoq, M., Chapuis, M. P., & Piou, C. (2020). On République du Mali. 2015. « Rapport d’évaluation Des Capacités the relative role of climate change and management Nationales Pour La Réduction Des Risques, La Préparation in the current desert locust outbreak in East Africa. et La Réponse Aux Urgences Au Mali.” https://www.cadri. Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens net/system/files/2021-06/Mali-Evaluation-des-capacites- SOS Faim. 2016. « Assurer la viabilité et promouvoir le nationales-RRC-Draft-2015.pdf développement des systèmes de stockage alimentaire de proximité en Afrique de l’Ouest. » https://www.sosfaim. République du Togo and AfDB. 2019. « Étude de Faisabilité Pour be/assurer-la-viabilite-et-promouvoir-le-developpement- La Mise En Place d’une Assurance Agricole Au Profit Des des-systemes-de-stockage-alimentaire-de-proximite-en- Producteurs Agricoles Au Togo. » https://www.afdb.org/ afrique-de-louest/. pt/news-keywords/togo?page=2 Soulé, Bio Goura and Sanni Gansari. 2010. La dynamique des RPCA (Food Crisis Prevention Network). 2020. Policy Brief « échanges régionaux des céréales en Afrique de l’Ouest. COVID-19 risks overshadowing and compounding the Bamako: Syngenta Foundation severe food and nutrition crisis in the Sahel and West Africa. Several tens of millions of people are threatened. Stock, J. H., AND M. W. Watson (1993): "A Simple Estimator The region’s stability could be at risk. » https://www.food- of Cointegrating Vectors in Higher Order Integrated security.net/wp-content/uploads/2020/04/NAD2020_ Systems," Econometrica, 61, 783-820. EN.pdf Stockholm Environment Institute 2015 “Reducing vulnerability to food price shocks in a changing climate” RPCA. 2021. Policy Brief «The Sahel and West Africa are facing a major food and nutrition crisis for the second consecutive Stoppa, A., and W. Dick. 2018. Agricultural Insurance in Burkina year and 27.1 million people could be at risk during the Faso: Challenges and Perspectives. [please provide city]: 2021 lean season.» https://www.food-security.net/wp- Oxfam. content/uploads/2021/05/NAD2021_EN.pdf Sultan B and Gaetani M (2016) Agriculture in West Africa in the Twenty-First Century: Climate Change and Impacts Saikkonen, P. (1991): "Asymptotically Efficient Estimation of Scenarios, and Potential for Adaptation. Front. Plant Sci. Cointegration Regressions," Econo- metric Theoiy, 7, 1-21 7:1262. doi: 10.3389/fpls.2016.01262. 174 SWAC/OECD (Sahel and West Africa Club/Organisation for Sandford, J., S. Rajput, S. Coll-Back, and A. Kargbo. 2020. “Safety Economic Co-operation and Development). 2020. “Food Nets, Health Crises, and Natural Disasters: Lessons from and Nutrition Crisis 2020, Analyses and Responses, Maps Sierra Leone.” Washington, DC: World Bank Group. https:// and Facts, No. 3, November 2020.” Paris: OECD. openknowledge.worldbank.org/handle/10986/34952. Sylla, Mouhamadou Bamba, Jeremy S. Pal, Aissatou Faye, Kangbeni Dimobe, and Harald Kunstmann. 2018. “Climate Saumik, P. (2015). Conflict, food security and crop diversification Change to Severely Impact West African Basin Scale strategies: evidence from cote d’Ivoire. Institute of Irrigation in 2 °C and 1.5 °C Global Warming Scenarios.” developing economies. Scientific Reports 8 (1): 14395. https://doi.org/10.1038/ s41598-018-32736-0. Seth, Anji, Sara A. Rauscher, Michela Biasutti, Alessandra Tondel, F. 2019. Dynamiques régionales des filières d’élevage Giannini, Suzana J. Camargo, and Maisa Rojas. 2013. en Afrique de l’Ouest. Etude de cas centrée sur la côte “CMIP5 Projected Changes in the Annual Cycle of d’ivoire dans le bassin commercial central. Precipitation in Monsoon Regions.” Journal of Climate 26 (19): 7328–51. https://doi.org/10.1175/JCLI-D-12-00726.1. Toulmin, C., & Guèye, B. (2016). Transformations in regional agriculture and family farming. In West African Worlds (pp. 167-206). Routledge. Simonite, T. 2020. “A Clever Strategy to Distribute Covid Aid— With Satellite Data.” Wired. https://www.wired.com/story/ UN (United Nations) New Centre. 2008. “The Global food crises” clever-strategy-distribute-covid-aid-satellite-data/. https://www.un.org/esa/socdev/rwss/docs/2011/ chapter4.pdf Sims, C. A. (1980), « Macroeconomics and Reality », UNCTAD (United Nations Conference on Trade and Econometrica, 48(1) : 1-48. Development). 2020. “UNCTAD Annual Report.” https:// search.library.wisc.edu/catalog/9911035603502121. USAID (United States Agency for International Development). World Bank. 2013. “World Development Report 2014: Risk 2010a. “Livelihood Zoning and Profiling Report: Burkina and Opportunity; Managing Risk for Development.” Faso.” Washington, DC. https://openknowledge.worldbank.org/ USAID. 2010b. “Livelihood Zoning and Profiling Report: Mali.” handle/10986/16092 License: CC BY 3.0 IGO. USAID. 2018. “Constraints to Accessing Finance and Insurance in World Bank. 2018. “Niger: Second Niger Adaptive Safety Mali’s Livestock Sector.” https://www.agrilinks.org/sites/ Net Project.” World Bank Group. https://documents. default/files/resources/final_call_order_4_ag_finance_ worldbank.org/en/publication/documents-reports/ and_insurance_report.pdf. documentdetail/777931546830037773/Niger-Second- Niger-Adaptive-Safety-Net-Project. USAID. 2020. “West Africa regional development cooperation strategy 2015-2020”. https://www.usaid.gov/sites/ World Bank. 2019. “Sahel Adaptive Social Protection Program: default/files/documents/1860/RDCS_West_Africa_ Annual Report 2019.” Sahel Adaptive Social Protection December_2020_External_508_1_2.pdf Program. World Bank Group. http://documents1. worldbank.org/curated/en/680361585895594749/pdf/ USAID-FEWS NET. 2010. “Livelihood Zoning and Profiling Report: Sahel-Adaptive-Social-Protection-Program-Annual- Burkina Faso.” Report-2019.pdf. USAID-FEWS NET. 2011. “Livelihood Zoning ‘Plus’ Activity in World Bank in Africa. 2020. Monitoring Global Poverty. Niger.” https://openknowledge.worldbank.org/bitstream/ handle/10986/34496/9781464816024_Ch1.pdf Wageningen Food Systems Paper 2021 “Transforming Food Systems Towards nutritious, inclusive, sustainable and World Bank and FAO of the United Nations. 2021. “A Blueprint efficient outcomes” https://www.wur.nl/en/show/ for Strengthening Food System Resilience in West Supporting-paper-1-.htm Africa: Regional Priority Intervention Areas.” Washington, DC: World Bank and UN FAO. © World Bank and WDI (World Development Indicators) Database. UN FAO. https://openknowledge.worldbank.org/ handle/10986/35618 License: CC BY-NC-SA 3.0 IGO. 175 WFP (World Food Programme). 2018. « Révue Stratégique Faim Zéro Au Togo. » https://docs.wfp.org/api/documents/ Zongo, Y. 2019. Fonds national de solidarité: Les contributions WFP-0000103375/download/. sont les bienvenues. Lefaso.Net. https://lefaso.net/spip. php?article90474. WFP, UN Women, and USAID. 2017. “Gender, Access and Use of Credit, Capital, and Insurance Services in Mali.” https:// docs.wfp.org/api/documents/WFP-0000100869/ download/. WMO (World Meteorological Organization). 2020. https:// public.wmo.int/en/media/press-release/new-climate- predictions-assess-global-temperatures-coming-five- years Wodon, Q. T., Tsimpo, C., Backiny-Yetna, P., Joseph, G., Adoho, F., & Coulombe, H. (2008). Potential impact of higher food prices on poverty: summary estimates for a dozen west and central African countries. World Bank Policy Research Working Paper, (4745). https://openknowledge. worldbank.org/bitstream/handle/10986/6931/WPS4745. pdf?sequence=1 World Bank. 2012. “Using Public Food Grain Stocks to Enhance Food Security.” World Bank Group. https://documents. worldbank.org/en/publication/documents-reports/ documentdetail/412711468336603745/Using-public- food-grain-stocks-to-enhance-food-security. Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens 176 APPENDICES Appendix A. Detailed Methodology Description (Crop and Rangeland Risk Assessment)1 A.1. Crop Risk Assessment A.1.1. Data analysis The crop risk assessment for the selected fifteen West African countries is performed based on official records from FAOSTAT. The analysis is based on country-level FAOSTAT historic records of crop harvested Area, Production, and Yield for each of the selected crops in each of the selected countries for the period from 1983 up to and including the year 2018. The crop risk assessment performed is based on the volatility of the GVP. The selection of the GVP as the main underlying variable is because this figure allows to find a common denominator and to include different crops in the portfolio. In this regard, the portfolio volatility is measured in monetary amounts and the contribution of the different crop to the portfolio volatility is weighted by the price of each crop. For purposes of the analysis the GVP is valued at retail prices. Treatment of the information Crop harvested Area, Production, and Yield records were treated before to be used in the analysis. This subsection describes the methodologies followed to clean and fill Yield records, to perform Yield d- rending and, finally, to adjust the historical Yield to the current expected Yields. - Outliers 177 Crop Yield series are treated for outliers. The reason for detecting outliers is to clean abnormal records that may introduce bias in the assessment. The criterion used to identify the outliers was to set an upper bound and a lower bound for the records and to remove all the records that felt out of these boundaries. For the purposes of the analysis, the upper bound is set at average value for the variable plus 3.5 times its standard deviation and the lower bound is set at the average value for the variable less 3.5 times its standard deviation. - Data Cleaning The analysis is focused on major food crops. The crops that show a Crop Area smaller than 2,000 hectares were not considered in the analysis. The deletion of those Units smaller than 2,000 hectares is to avoid errors in the model. Crop yield statistics for minor crops do not have the same quality as the crop yield statistics for major crops. The resources of the Crop Statistical Services deployed to minor crops are not the same as those deployed for major crops. Crop sampling methodologies applied to minor crops are not as rigorous as those applied for major crops. 1 This appendix refers to chapter 4. 1 Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens Crop Area Data Analysis - Crop Harvested Area data Crop Harvested Area records, formally HAt;k;j, are included in the 40-years series for each crop “k” and country “j” comprising the portfolios that were analyzed in the report. - Expected Crop Area The modeled Expected Crop Area is calculated as per the average of the crop area for the latest three years available in the crop yield series received for the analysis; the period considered for the calculation for the expected crop area is from 2016 up to and including 2018. Where, HA2016;k;j , HA2017;k;j, and HA2018;k;j the actual crop area for the years 2016, 2017, and 2018. Table A-1 presents the Expected Crop Area per crop and country included in the crop risk assessment. It should be noted that sorghum, followed by maize, millet, and rice are the main crops in terms of area in the portfolio. Sorghum crop is predominant in the transitional tropical and tropical climates of the Sudano. Millet is the predominant crop in the countries situated in the semiarid tropical climates in the Sahel, while maize is predominant in the transitional tropical and transitional equatorial climates. Sorghum, millet, and maize account for 53 percent of the total area sown with the selected crops in West Africa. Nigeria, accounting for 37 percent of the total area planted with the selected crops in West Africa, is the country to contribute most to the agriculture output. Nigeria is followed by Niger, which 178 accounts for 22 percent of the area, and Mali and Burkina Faso, with a share of 7 percent each. 2 Table A-1: Expected Crop Area Per Crop and Country (in Thousands of Hectares) Source: Original table based on estimation results Yield data 179 The crop risk assessment for main food crops in West Africa is based on 35 years of actual annual country average crop yield series from 1983 up to and including 2018 calculated from FAOSTAT. The crop yield database used for the yield assessment includes records for each main crop in the fifteen selected countries in West Africa. The information was provided in terms of kilograms per hectare for each year “t” and crop “k” and country “j.” The yields are calculated as follows: There are several considerations to make regarding the crop yield information used for the crop yield risk assessment for the main food crops in West Africa. The first consideration is that the yields provided by FAOSTAT are calculated over harvested area. The fact that the crop yield was calculated over harvested area might lead to situations on which total losses are not factoring while computing yields. The second consideration is that the crop yield information used for the risk assessment is resulting from an aggregate average by crop and country. The fact that the crop yield information used for the risk assessment is resulting from an aggregate average by crop and country could lead to underestimations of crop yield volatilities hence underestimation of risks. The third consideration resulting from working with country-level average crop yields is that regional cross border crop yield shortfalls situations, as the crop yield are aggregated a country level, would not be factored in the 3 Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens analysis. Table A-2 presents the Expected Yields for each crop “k” and country “j” considered in the portfolio. Table A-2: Expected Yields for Selected Crops and Countries in West Africa (Kilograms per Hectare) Source: Original table based on estimation results 180 - Expected Yield The modeled Expected Yield has been calculated for each crop “k” and country “j” and for each year “t” as the result of a simple average of the three central years out of the previous five years of yield data. Formally, the Expected Yield for year “t” and crop “k” and country “j” is calculated as follows: where the average yield reported in year “Yt−i;k;j” for crop “k” and country “j” and lt−i;k;j;Not MaxMin equals to 1 if is neither the maximum nor the minimum yield of the previous five years and equals to 0 if is the maximum or the minimum of the previous five years. Although there could be a long run trend, the objective in setting the Expected Yield as a simple average is to avoid an overestimation of yields due to the application of the long-term crop yield trend. - Trend Analysis and D-trended Yields Crop Yield Series are d-trended to remove the effects of technology and management practices. A simplified method is adopted for determining the central tendency of the average annual yields for each crop and country. Crop yield historical trends in each country has been calculated as the moving average of the three central years out of the previous five years (that is, excluding the maximum and minimum). 4 Therefore, the trended yield is just the modeled Expected Yield that has been defined in equation (3)10. Figure A-1 presents the evolution of the average crop yield for sorghum crops in Burkina Faso, along with the trended or Expected Yield. It can be seen the low yield values in 1990, 1994, 1997, 2009, 2011, 2016, and 2017 as compared to their crop yield expectations. Figure A-1: Burkina Faso, Sorghum: Historic Crop Yields and Expected Yields Source: Original figure based on estimation results The second phase of the d-trending analysis consists in the estimation of the yield variability. Aiming for 181 this purpose, the Percentage Deviations (YPD) between the Actual yields with respect to the corresponding Expected Yield for each crop and country are calculated. Formally, the Percentage Deviations for the year “t,” crop “k,” and country “j” ( YPDt;k;j ) can be defined as follows: Production The assessment of the production for the main food crops in the selected countries in Western Africa is the result of the multiplication for crop Area and. More formally the d-trended Production in year “t” for the crop “k” and country “j” Pt;k;j, is expressed as follows: Table A-3 presents the Expected Production for each crop and county in the portfolio. 5 Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens Table A-3: Expected Production for Main Crops in Selected Countries in West Africa (in thousand toms) Source: Original table based on estimation results - Expected Production 182 The Expected Production for each crop in each country and year is the result of the multiplication of the Expected Crop Area for crop “k” and country “j” (Equation 1) and the Expected Crop Yield for the crop “k” in the country “j” and year “t” (Equation 3). Formally, the equation can be expressed as follows: Prices used for Valuation The valuation of the crops for the calculation of the GVP is based on the Average Retail Price for the period 2016–20 for the selected main food crops in Western African countries. The reason for using the retail price instead of the market price is that the aim of the report is to assess food availability for food security purposes. If the ultimate objective of the analysis is for Food Security Purposes, the financial loss should be measured at the replacement cost of the food, hence products should be valued at the retail or the wholesale price. Retail price information comes from different sources that include the WFP-VAM, FEWS NET, FAO-GIEWS, and Bureau of Statistics and Ministry of Agriculture of the respective countries. Table A-4 presents the Average Price in US$ per kilogram for the period 2016–20 for the selected crops and countries in the portfolio. 6 Table A-4: Average Price for the Period 2016–20 for the Main Food Crops in the Selected West African Countries (US$ Per Kilogram) Source: Original table based on estimation results - Exposures – GVP The exposures used in the risk assessment model are based on the GVP, EGVPk;i. The GVP results from 183 the product of the Expected Production for the crop “k” in the country “j,” EPk;j, times the Average Price for the crop “k” in the country ‘j,’ APavg;k;j. The selection of the GVP as the main underlying variable is because the GVP allows finding a common denominator to include different crops in the portfolio. The portfolio volatility is measured in monetary amounts and the contribution of the different crop is weighted by its GVP. Formally, the Expected GVP in year “t” for the crop “k” and country “j,” EGVPk;j can be defined as follows: Table A-5 exhibits the Exposures, or GVP- GVPk;j, for each crop that is considered in the risk assessment. It should be noted that the portfolio is driven by the results of yam and cassava crops, which are the crops that show the highest value, accounting for 67 percent of the GVP of the portfolio. The portfolio is also driven by the volatility of the GVP in Nigeria, which accounts for 56 percent of the GVP in the portfolio. 7 Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens Table A-5: Exposures or GVP by Selected Crop and Country for Main Food Crops in West Africa (US$, millions) Source: Original table based on estimation results 184 A.1.2. Average loss cost and LaR calculation Average loss Crop losses were calculated based on the deviation of Actual GVP from Expected GVP. If the Actual GVP falls short of the Expected GVP, then there is a loss proportional to the size of shortfall. Formally, the percentage Average loss for crop “k” and country “j” is calculated as follows: Whereas the Actual GVP for the crop “k” in the country “j” and is the Expected GVP for the crop “k” in the country “j.” The LaR or PML is a key measure used to infer the potential losses in the portfolio and to, eventually, establish the appropriate level of risk retention in the portfolio. LaR estimates the size of a loss that could be exceeded with a given probability; that is, is the minimum loss for a given Exceedance Probability. The LaR is a percentile of the Loss Distribution, calculated according to the probability of occurrence of a catastrophic event. For example, the LaR for an Exceedance Probability “p” of 1% (or return period of 1 in 100 years [2]), is the value of the Loss Distribution that accumulates 99 percent of probability—the 8 99th percentile. That is, there is a 1 percent probability of observing loss higher than the estimated LaR. The LaR for a 2 percent Exceedance Probability (or for a 1 in 50 years event) is the 98th percentile; and so on. Formally, the LaR for the Expected GVP “α” associated to an Exceedance Probability “p,” and the loss cost, then the LaR is implicitly defined in the following formula: The LaR can be calculated for each crop and country or at an aggregated level for the portfolio. Furthermore, the LaR can be calculated by crop in isolation or for a complete portfolio that includes several crops. For a given crop, the estimation of LaR is based on the Loss Distribution for that crop, which depends exclusively on the probability distribution of Lαk;j, as defined in the equation above. If several crops are considered simultaneously in a portfolio, the Loss Distribution should consider the covariance among losses for all the crops. Furthermore, when several crops are combined in one portfolio, the correlation across crops should be also included in the analysis. The risk assessment study performed for main food crops in West Africa includes the calculation of (1) LaRs for each crop and country and (2) Portfolio LaR (that is, by pooling all crops and countries comprised in the portfolio). Figure A-2: Loss Exceedance Probability Curve for Main Selected Food Crops in West Africa 185 Source: Original figure based on estimation results A.1.3. Simulation model The best practices suggest performing risk assessment and rating based on risk modeling executed through Monte Carlo simulation to extend the size of the sample of yield data. This improves the confidence levels of the average loss costs and the LaR calculations. This section describes the simulation model that has been developed for the risk analysis performed for main food crops in West Africa. 9 Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens Inputs The probabilistic model that is specifically devised to assess the risk of the main food crops in West Africa works on the base of three inputs: Expected Crop Area (EHA); Historical Percentage Deviations from Expected Yield (YPD); and crop Prices. The EHA is the result of the average crop area for each selected crop and country for the period 2016–18. The YPD are the stochastic variables of the probabilistic model. Crop Price is a deterministic variable in the model. EHA The modeled EHA for each combination of crop and country comprising the portfolio in West Africa region is calculated as per the average of the crop area for the latest three years available in the crop yield series received for the analysis; the period considered for the calculation for the expected crop area is from 2016 up to and including 2018. Where, HA2016;k; , HA2018;k;j. and is the actual crop area for the years 2016, 2017, and 2018. YPD The Percentage Deviation from Expected Yield for crop “k” and county “j” is defined in equation below as: 186 The historical values of YPDt;k;j for sorghum in Burkina Faso. The same procedure is applied for getting the historical sample of Percentage Deviations from Expected Yield for each crop and country in West Africa region. The YPD jointly with the Percentage Deviation from Expected Yield are key inputs for the risk assessment methodology based on Monte Carlo simulation. Correlation matrices of YPDs per crop and country The historical data of YPD for each crop are correlated to reflect the systemic risk and the portfolio diversification effect. If a perfect correlation between deviations from the expected yields (EYt;k;j) is assumed (that is, all the yields for each crop and country vary on the same proportion and direction), the losses would be overestimated because a low crop yield in one combination of crop and country (high negative percentage deviation) would imply low yields in all the units conformed by the different crops and countries in the portfolio. On the other hand, if zero correlation is assumed, the cost of losses would be underestimated as the variation of crop yield losses across the units conformed by the combination of crops and countries that conform the portfolio would be assumed to have no relationship and to vary independently each from the other. Neither of these extreme scenarios would be a reasonable assumption as there is some level of correlation in yields across the different crops and countries that conform the portfolio, but this correlation is not perfect. - Crop Prices The valuation of the crops for the GVP is based on the Average Price for the month of harvest for the period 2016–18 for the main selected food crops in West Africa. 10 Distribution Fitting of YPDs The stochastic variable of the model, YPD for each crop and country in the selected portfolio, is fit to a to a parametric probability distribution function. The probability distribution fitting process is conducted on an individual basis for each combination of crop and country in the portfolio: a continuous distribution function is fitted to the historical d YPDt;k;j for each crop and country in the portfolio. The fit PDF is tested against three criteria: (1) Chi-squared, (2) Anderson-Darling, and (3) Kolmogorov-Smirnov. An analysis of some critical values of the fitted probability distributions is done to select the best fit for rating purposes, particularly as follows: • 1st percentiles were compared to minimum historical values. • Probability of get a value lower than the minimum historical. • Probability of obtain a value lower than -100% (minus hundred percent, that is, a negative yield). • Theoretical skewness coefficients were compared to sample skewness. • Theoretical means were compared to sample averages. • Theoretical Expected Loss Costs were compared to AAL of the Historical Burning Cost Analysis for different coverage levels. • Theoretical Losses at Risk were compared to Worst Annual Loss scenario of the Historical Burning Cost analysis for different Yield Threshold levels. Figure A-3 shows the comparison of parametric-based metrics with the corresponding values calculated through Historical Burning Cost Analysis for sorghum in Burkina Faso. The probability distribution function that best fits with the Percentage Deviations for sorghum in Burkina Faso is a Weibull 187 distribution centered in 15.84 with α = 1.35 and β=-1.30. Notice that differences between Technical Premium Rates in Panel (c) are mainly explained by the difference in LaR (calculated with an Exceedance Probability of 1 percent, that is, a 1 in 100 years event) vs. WAL (based on a sample of 35 years) exhibited in Panel (b). With this process, the volatility of crop yield percentage deviations for each combination of crop and country within the portfolio is represented by probability distribution function fitted that belongs to a different distribution family (for example, Logistic, Normal, Weibull, and so on) and the parameters estimated for each combination of crop and country could also be different. 11 Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens Figure A-3: Panel (a) Burkina Faso, Sorghum: Comparison of Parametric-Based Rating Metrics with the AAL from HBCA Analysis Source: Original figure based on estimation results Figure A-4: Panel (b) Burkina Faso, Sorghum: Comparison of Parametric-Based Rating Metrics with the WAL from HBCA Analysis 188 Source: Original figure based on estimation results D-Trended Yield Simulation Once the PDFs are fit to the historical YPDt;k;j data are calculated and the correlations among the crops and countries in the portfolio are assessed, YPD simulation process is performed. Specifically, Monte Carlo simulation is used to generate simulated samples of 10,000 hypothetical years of d-trended Yields for the crops included in the portfolio. Fitted distribution functions for Historical Percentage Deviations from Expected Yield (YPDt;k;j) in conjunction with the corresponding correlation matrix are used to simulate 10,000 data of Percentage Deviation from Expected Yield (YPD) for each crop and country. The simulated Historical Percentage Deviations from Expected Yield are then applied to the Expected Yield (EY) to obtain a simulated sample of 10,000 d-trended Yields. Once the values of YPDs are simulated, the calculation presented in equation below is used to simulate samples of (d-trended) crop Area as follows: 12 is the Percentage Deviation from the Expected Yield for a given crop t “k” country “j” and simulation “i” (i =1 to 10,000) based on Monte Carlo simulation. The simulated samples of d-trended Yields for each crop are then used then to perform the crop risk assessment. D-Trended GVP The D-Trended GVP ( DGVPk;i;MCS ) is the result of the multiplication between the D-Trended Production for each crop “k” and country “j” and the corresponding expected Price. Formally: The D-Trended Production ( DPk,j ) is, in turn, the result of the multiplication between the Expected Crop Area and the D-Trended Yield for crop “k” and country “j.” Formally: The Expected Crop Price (APk;j) and the Expected Crop Area for the crop “k” and country “j” is considered as a deterministic variable of the model and remains fixed with the 10,000 simulations. The D-Trended GVP ( DGVPk;j;i ) can be formally expressed as follows: As a result of the application of this Equation, a sample of 10,000 D-Trended GVP ( DGVPk;i;MCS) simulated is obtained. This sample is used as underlying to assess the risk for the crop “k” and country “j.” 189 The D-Trended GVP ( DGVPk;I;j ) for a Portfolio of crops and province results from the addition of the D- trended GVPs for each crop “k” and country “j” in any particular year particular simulation. The formal Equation for D-Trended GVP for the portfolio ( DGVPportfolio;MCS ) is: A.1.4. Risk analysis based on simulation The 10,000 simulated d-Trended GVP for each crop and country in the portfolio ( DGVPk;I;j ) generated through Monte Carlo simulation in the model are then used to carry out the risk assessment for the main food crops in West Africa. An Excel-based rating tool has been designed to give maximum transparency and flexibility to the user in calculating Average Losses and LaR for each main food crop and country in West Africa. Simulation-based Loss Cost The simulation-based Loss Cost (defined as percentage of Expected Value of Production) in simulation “i” for crop “k” and country “j” for the Expected Yield “α” is calculated using the Simulated D-trended GVP defined as follows: 13 Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens The Monetary Loss Cost (in US$) in each simulation is calculated by multiplying the percentage simulated Loss Cost of equation (13) times the Expected GVP of crop “k” and country “j.” The Total Loss Cost (in monetary amount) in each simulation is the sum of the Monetary Loss Cost across all the crops comprised in the Portfolio under analysis. Simulation-based average Loss Cost The Average Loss Cost for the crop “k” and country “j” in the risk assessment is calculated in this study using the following generic equation: Where, E(Lαjk;j) is the Expected Value of losses for the crop “k” and country “j” in respect to the Expected Yield which is used to calculate the Expected GVP. Considering the equation below, which is based on Monte Carlo simulation, is the simple average of the Simulated Percentage Loss Costs. Formally, the Average Loss Cost for the crop “k” and country “j” based on Monte Carlo simulation crop “k is calculated as follows: 190 Simulation-Based LaR Per Province and Per Crop The LaR was defined in section 3.2 and the calculation based on Monte Carlo simulation for the crop “j” and country “k” and Exceedance Probability “p,” LaRαj;k;pMCS is defined by the following equation: The aggregated LaR for the whole portfolio is also a key variable to analyze the global exposure of the portfolio of main food crops in West Africa. This measure is obtained from the Total Loss Distribution, which results from the simple aggregation of the losses for each crop and country but considering the effect of the covariance among crop and countries comprising the portfolio and the relative weight of each combination of crop and country over the total exposures in the portfolio. The Total Loss Distribution per Crop in the Portfolio is obtained as a by-product of the simulation process; the Simulated Total Loss in simulation “i” is the weighted average of the simulated loss for the crop “k” and county “j” where weights are given by the Exposure of the crop “k” and country “j” over the whole portfolio. Formally: 14 The aggregated LaR for the portfolio of main food crops in West Africa based on Monte Carlo simulation and an Exceedance Probability “p,” LaRαp;MCS is implicitly defined in the following equation: It is worth noticing that when the risks are pooled, there is a diversification effect that results in an aggregated LaR lower than the sum of the LaR per crop. A.2. Rangeland Risks Assessment: Detailed Methodology Description The rangeland risk assessment performed for selected West African countries is based on eMODIS NDVI raster maps composed exclusively by pasture rangelands pixels. Aimed at using relevant NDVI information for rangeland risk assessment, the eMODIS raster maps were masked to remove those pixels that correspond to (1) those areas that are not classified as grassland and (2) those areas where livestock production is not the predominant agriculture activity or if livestock production is not based on rangelands production as the main source of forage (for example, industrial dairy). The eMODIS smoothed 250m NDVI raster maps were masked using the pixels classified as rangelands according to global crop and rangeland mask, Joint Research Centre.2 For the purposes of the selection of the rangeland pixels, only those pixels that present 75 percent or more of the area classified as 191 rangeland were selected. The following sequence presented in map A-1, map A-2, and map A-3 shows the eMODIS smoothed 250m NDVI data-masking process with Grassland Land Use as per the Joint Research Centre. 2 The Joint Research Centre (JRC), also known as the Batti Research Centre, is the European Commission’s science and knowledge service that employs scientists to carry out research in order to provide independent scientific advice and support to EU policy. The JRC is a Directorate-General of the European Commission under the responsibility of Tibor Navracsics, Commissioner for Education, Culture, Youth & Sport. 15 Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens Map A-1: eMODIS 250m Temporally Smoothed NDVI Map A-2: Grassland Cover (global crop and rangeland mask, Joint Research Centre >75 percent) Map A-3: Masked Grasslands eMODIS NDVI 192 Source: Original maps based on eMODIS-NDVI results The pasture rangeland masked eMODIS NDVI was also screened against main agriculture production systems throughout West Africa to select only those rangeland pixels that belong to production systems on which livestock production is predominant. Aimed at discriminating those pixels belonging to areas where livestock production is not the main activity in the predominant agricultural production systems, grassland Land Use masked eMODIS raster were screened against the LVZ developed by FEWS NET. Only those pixels classified as “grasslands” that belonged to areas associated with Pastoral and Agropastoral Zones were selected. Regional (Admin 1 level) Administrative Division shapefiles were overlaid over the masked eMODIS pixels to assimilate the masked eMODIS NDVI to the Regional boundaries. Regional Administrative Boundaries were used as the boundaries to define the Risk Unit Areas to perform the rangeland risk assessment. The Regional Administrative Boundaries used for the rangeland risk assessment is based on Admin 1 level shapefiles that were downloaded from Global Administrative Areas (GADM) (https://gadm.org/maps.html). The result of the eMODIS NDVI raster map masking exercise for those pixels classified as “grasslands” that belonged to Pastoral and Agropastoral Zones is presented in map A-4, map A-5, and map A-6. 16 Map A-4: Masked Grasslands eMODIS Grouped into Admin 1 Level Boundaries Map A-5: Pastoral and Agropastoral LVZ per FEWS NET Classification Map A-6: Masked Grasslands eMODIS Grouped for Pastoral and Agropastoral Zones 193 Source: Original maps based on eMODIS-NDVI results Considering the reliability of the eMODIS NDVI that will be used as underlying for rangeland risk assessment, the masked raster maps were tested against additional criteria for a Region to be selected in the portfolio. The criteria employed was that the region must have more than 10,300 hectares of area to qualify to be included in the selected portfolio. The objective of this last criteria was to select only those wards on which livestock production over rangelands is significant. The last step of processing eMODIS information for NDVI-IBLI Product design Purposes is to average the eMODIS NDVI masked raster for each dekadal period and selected Region. The objective of averaging the eMODIS NDVI masked raster is to obtain one single eMODIS NDVI masked value per ward for each dekadal period from July 2002 up to June 2020 to be used as underlying for the coverage. Figure A-5 shows the evolution of the grassland masked eMODIS value for Louga Region in Senegal for the period July 2002–June 2020. 17 Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens Figure A-5: Evolution of the actual eMODIS NDVI in Louga Region, Senegal Source: Original figure based on estimation results A.2.1. Rangeland shortfall event definitions The rangeland risk assessment for the selected five West African countries was performed under the assumption that the critical herd management practices that drive livestock production are synchronized with the expected availability of fodder to feed the herd. In this regard, deviations from the expected supply of fodder create significant disturbances in the herd management and, consequently, on livestock production. For instance, heifers and breeding cows have high nutritional 194 requirements during the breeding seasons (as the cow must be in good health to get pregnant at the time they are calving). As a result, breeding seasons are synchronized with the peak supply of fodder. Rangeland production systems in West Africa are fragile. In this regard, slight deviations on the expected rangeland production during the peak season may cause severe stress over the rangeland production systems. As the performance of the herd is mostly driven by the availability of fodder during the critical periods such as breeding and calving, the rangeland risk assessment is performed based on the evolution of the rangeland production, measured through the NDVI, during peak seasons. Peak seasons are defined as the decadal periods on which the expected eMODIS NDVI falls in the first tercile in terms of NDVI values. The expected accumulated NDVI along the four-month peak season accounts for approximately 50 percent of the expected accumulated NDVI along the year. The Expected eMODIS NDVI during the peak season, EYk, has been calculated for each Region “k” as the result of a simple average of the eighteen years of eMODIS NDVI peak Season data. Formally, the Expected eMODIS NDVI during the peak season for the region “k” is calculated as follows: 18 Figure A-6: Louga Region, Senegal: Expected Dekadal eMODIS NDVI and the Expected Accumulated Peak Season NDVI Source: Original figure based on estimation results Rangeland shortfalls are measured through the deviations between the actual and the expected peak season accumulated NDVI for each Risk Unit. In the example presented in figure A-7 for the year 2002 in Louga Region in Senegal, the actual Accumulated Peak Season eMODIS NDVI (third dekadal of July–2nd dekadal of November) indicates a value of 3.52, which is 24.6 percent short of the expected Accumulated Peak Season eMODIS NDVI of 4.67. Figure A-7: Louga Region, Senegal: 2002 Actual Versus Expected Accumulated Peak Season eMODIS NDIV (Jul-3/Nov-2) 195 Source: Original figure based on estimation results Figure A-8 presents the historic evolution of the Actual Accumulated NDVI during the Peak Season in Louga Region in Senegal compared with the expected Actual Accumulated NDVI for the Peak Season. In this regard, it can be observed that the Actual Accumulated NDVI for the Peak Season fell short of the Expected values in 10 years out of the 18 years considered for the analysis—the years 2002 and 2014, the worst years in the series—with shortfalls of 24.5 percent and 21.4 percent, respectively. 19 Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens Figure A-8: Louga Region, Senegal: Evolution of Actual Accumulated NDVI During the Peak Season Compared with the Expected Value Source: Original figure based on estimation results Processing of the actual accumulated NDVI during the Peak Season Actual accumulated NDVI during the Peak Season Series are d-trended to remove the long-term central tendency. The analysis performed for eMODIS NDVI measured over rangelands in West Africa did not show a significant long-term central tendency. Based on this finding, no d-trending analysis was 196 performed over the actual eMODIS NDVI during the peak season in the selected West African countries. The second phase of the analysis consisted in the estimation of the yield variability. Aiming for this purpose, the Percentage Deviations (YPD) between the actual accumulated NDVI during the Peak Season with respect to the corresponding Expected value for each Region were calculated. Formally, the Percentage Deviations for the year “t” and region “k” ( YPDt;k ) can be defined as follows: Figure A-9: Louga Region, Senegal: Historic Actual Accumulated NDVI Deviations from Expected NDVI Along the Peak Season Source: Original table based on estimation results 20 Later, the Percentage Deviations (YPD) are applied to the last Expected peak season accumulated NDVI for each Risk Unit to obtain an adjusted and d-trended annual average yield series for each crop in the portfolio. More formally, the d-trended actual peak season accumulated NDVI in year “t” and region “k” in the portfolio is calculated by applying the Percentage Deviations to the Expected peak season accumulated NDVI for the region “k,” EYk , as is represented in the equation (3): Randomness in equation (3) arises exclusively from the uncertainty in the Percentage Deviation of equation (2). Therefore, for the purpose of the risk assessment, the historical sample of Percentage Deviations from Expected peak season accumulated NDVI has been taken as the main input to be used in the Simulation Model. Exposures - Expected Rangeland Area The Expected Rangeland Area is estimated based on the number of pixels classified as rangeland with more than 75 percent of its area belonging to rangeland as per the Joint Research Institute classification. Joint Research Institute rangeland classification is performed over pixels of 1 square kilometer each or 100 hectares. The rangeland area for each risk unit “k” considered in the analysis is EHAk , and results for adding up all the pixels classified as rangeland in each Risk Unit “k” multiplied times 100 to convert it into hectares as it is presented in the following equation: Table A-6 presents the Expected Rangeland Area per country included in the rangeland risk assessment. 197 It should be noted that Niger and Mali, with 32.6 million hectares and 21.8 million hectares, respectively, account for almost 65 percent of the rangeland area in the portfolio of selected countries for risk assessment in West Africa. Table A-6: Expected Rangeland Area per Country Area Country (hectares) (%) 32,611,70 Niger 0 39% 21,795,50 Mali 0 26% 14,989,10 Chad 0 18% Burkina Faso 9,182,800 11% Senegal 5,732,400 7% 84,311,50 Total 0 100% Source: Original table based on estimation results 21 Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens Production The assessment of the production for rangelands in the selected countries in West Africa is the result of the multiplication of the rangeland area and the actual peak season accumulated NDVI. It should be noted that the production is estimated through the accumulated eMODIS NDVI along the peak seasons, which is a convention unit that is directly associated with rangeland production. More formally, the d- trended rangeland production in year “t” and region “k,” Pt;k, is expressed as follows: The Expected Production for each region “k” and year is the result of the multiplication of the Expected Rangeland Area for the region “k” and the Expected peak season accumulated NDVI. Formally, the equation can be expressed as follows: Table A-7: Expected Rangeland Production per Country Area Country (NDVI Units) (%) Niger 90,357,748 31% Mali 75,073,599 26% Chad 58,596,020 20% Burkina 36,646,047 13% Faso 198 Senegal 27,407,823 10% Total 288,081,237 100% Source: Original table based on estimation results A.2.2. Average loss cost and LaR calculation A.2.3. Simulation model The best practices suggest performing risk assessment and rating based on risk modeling executed through Monte Carlo simulation to extend the size of the sample of rangeland eMODIS NDVI data. This improves the confidence levels of the average loss costs and the LaR calculations. This section describes the simulation model that has been developed for the risk analysis performed for main rangelands in West Africa. Inputs The probabilistic model that is specifically devised to assess the risk of the main food crops in West Africa works on the base of three inputs: EHA and Historical Percentage Deviations from Expected Accumulated eMODIS NDVI along the Peak Season (YPD). The Expected crop Area (EHA) is the result of the average crop area for each selected crop and country for the period 2016–18. The Historical Percentage Deviations from Expected Yield (YPD) are the stochastic variables of the probabilistic model. Crop Price is a deterministic variable in the model. 22 - EHA The Expected Rangeland Area for each region “k” in the selected portfolio was estimated based on the number of pixels classified as rangeland with more than 75 percent of its area belonging to rangeland as per the Joint Research Institute classification. The Rangeland area for each risk unit “k” considered in the analysis is and results for adding up all the pixels classified as rangeland in each risk unit “k” multiplied times 100 to convert it into hectares as it is presented in the equation 4 above. - Historical Percentage Deviations from Expected Accumulated NDVI along the Peak Season (YPD) The Percentage Deviation of the Actual accumulated NDVI Deviations from Expected NDVI along the Peak Season (YPD) in year “t” and region “k” was defined as follows: The historical values for rangelands in Louga region in Senegal are shown in graph 5. The same procedure is applied for getting the historical sample of Percentage Deviations from accumulated NDVI Deviations from Expected NDVI along the Peak Season for Region in West Africa. The expected accumulated Peak Season NDVI, jointly with the Percentage Deviation of the Actual accumulated NDVI Deviations from Expected NDVI along the Peak Season (YPD), are key inputs for the risk assessment methodology based on Monte Carlo simulation. - Correlation matrices of YPDs per crop and country The historical data of Percentage Deviation of the Actual accumulated NDVI Deviations from Expected NDVI along the Peak Season (YPD) in each region are correlated to reflect the systemic risk and the portfolio diversification effect. If a perfect correlation between Percentage Deviation of the Actual 199 accumulated NDVI Deviations from Expected NDVI along the Peak Season is assumed (that is, all the actual NDVI peak season accumulated values for each region vary on the same proportion and direction), the losses would be overestimated. On the other hand, if zero correlation is assumed, the loss cost for the portfolio would be underestimated as the variation of eMODIS NDVI shortfalls across the regions conforming the portfolio would be assumed to have no relationship and to vary independently each from the other. Neither of these extreme scenarios would be a reasonable assumption as there is some level of correlation in NDVI values across the different regions that conform the portfolio. A.2.4. Risk analysis based on simulation The 10,000 simulated d-Trended Rangeland Production for each region “k” in the portfolio (DPk;i ) generated through Monte Carlo simulation in the model are then used to carry out the risk assessment for rangeland in West Africa. An Excel-based rating tool has been designed to give maximum transparency and flexibility to the user in calculating Average Losses and LaR of Rangeland Production at the regional level in West Africa. Simulation-based Loss Cost The Simulation-based Loss Cost (defined as percentage of Expected Rangeland Production) in simulation “i” and region “k” for the Expected accumulated eMODIS NDVI during the peak season “α” is calculated using the Simulated d-trended Production: 23 Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens Simulation-based Rangeland Average Loss Cost The Average Rangeland Loss Cost for the region “k” in the risk assessment is calculated in this study using the following generic equation: Where E(Lαk) is the Expected Value of rangeland losses for the region “k” in respect to the Expected value of accumulated eMODIS NDVI during the peak season. Considering Equation (12), which is based on Monte Carlo simulation, is the simple average of the Simulated Percentage Loss Costs generated through equation (11). Formally, the Average Rangeland Loss Cost for the region “k” based on Monte Carlo simulation is calculated as follows: Simulation-based Rangeland LaR per region The LaR was defined in section 3.2, and the calculation based on Monte Carlo simulation for the region “k” and Exceedance Probability “p,” LaRαk;p;MCS , is defined by the following equation: The aggregated LaR for the whole portfolio is also a key variable to analyze the global exposure of the 200 rangeland portfolio in West Africa. This measure is obtained from the Total Loss Distribution, which results from the simple aggregation of the losses for the region but considering the effect of the covariance among regions comprising the portfolio and their relative weight on the total exposures in the portfolio. The Total Rangeland Loss Distribution per region in the Portfolio is obtained as a by- product of the simulation process; the Simulated Rangeland Total Loss in simulation “i” is the weighted average of the simulated loss for region “k,” where weights are given by the Exposure of each region “k” over the whole portfolio. Formally: The aggregated LaR for the rangeland portfolio in West Africa based on Monte Carlo simulation and an Exceedance Probability “p,” LaRαp;MCS is implicitly defined in the following equation: It is worth noticing that when the risks are pooled, there is a diversification effect that results in an aggregated LaR lower than the sum of the LaR per crop. 24 Appendix B. Food Price Risks and Market Integration Methodology3 B.1. Food Price Risks B.1.1. Methodology The methodology for estimating the impact of events (policy or shocks) on the levels and variability of food prices relies on VAR models. VAR models have proven useful for estimating effects of policies or shocks on economic variables, especially where there is uncertainty about the correct structural model or where data is limited and some variables are unavailable (Sims 1980; Jayne et al. 2008; Fackler 1988). The VAR model extends the idea of univariate autoregression to time series regressions, where the lagged values of all series appear as regressors. Put differently, a VAR model regresses a vector of time series variables on lagged vectors of these variables. The VAR model has proven to be especially useful both for describing the dynamic behavior of time series and for forecasting. It often provides superior forecasts to those from univariate time series models and elaborate theory-based simultaneous equations models. The case is being considered where there is a climate and or security shocks St to influence the price of a food commodity Yt. The relationship between climate, security shocks, and outcomes is represented as follows: Where the βi, Ay, Di, and Gi,Ap, are matrices of unknown parameters, “k” is the maximum number of lags 201 allowed in any equation, and µyt and µpt are vectors of mutually uncorrelated (orthogonal) innovations representing random shocks to underlying demand, supply, and policy processes that are generating observable data Yt and St. B.1.2. Data This analysis has been performing to see the impact of some shocks that happen in the region on food prices. Hence, price data are used for the biggest food crops identified in West Africa: maize, millet, rice, sorghum, cassava, and yam. This price information mostly comes from the WFP-VAM database and over the period January 1998 to September 2020. As predictor variables, some conflicts variables are used from the ACLED database, such as protests, conflicts, riots, and violence against civil and strategic developments (arrests, looting or property destruction). These variables allow an assessment of the impact of security risks on food prices and further on market integration. Analyzing food prices also requires paying attention to some climate invents, mainly climate variability, which can affect food production and then food prices through the availability dimension. Hence, monthly data are included on rainfall and temperature that come from the Olivier Santoni/cerdi database as an indicator of climate variability. As monthly data for some climate shocks’ variables such 3 This appendix refers to chapter 6. 25 Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens as drought and flood were not available, rainfall and temperature deviation have been computed through the Z-score. A Z-score is a numerical measurement that describes a value’s relationship to the mean of a group of values and is measured in terms of standard deviations from the mean. Hence, with rainfall and temperature monthly data (January 1992 to September 2020), climate shocks are computed as the difference between the value for each month and the total average value for each month over the period divided by the standard deviation. Therefore, it allows to obtain the gap between the current value and the average value for each month that rainfall and temperature shocks are considered. B.1.3. Results Table B-1: VAR Model, Cassava 202 Source: Original table based on estimation results 26 Table B-2: VAR Model, Sorghum 203 Source: Original table based on estimation results 27 Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens Table B-3: VAR Model, Yam 204 Source: Original table based on estimation results 28 Table B-4: VAR Model, Millet 205 Source: Original table based on estimation results 29 Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens Table B-5: VAR Model, Maize 206 Source: Original table based on estimation results 30 Table B-6: VAR Model, Rice 207 31 Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens B.2. Market Integration B.2.1. Methodology Market integration has been analyzed using the cointegration regression based on long-run covariance with dynamic ordinary least square (DOLS) estimators (Saikkonen 1992; Stock and Watson 1993). Many research papers used the Vector Error Correction Model (VECM) to assess the long-run relationship between prices. However, among all the estimators reported by Stock and Watson (1993), the DOLS has the smallest bias in their Monte Carlo results, excluding Johansen’s Maximum Likelihood Estimation. Hence, the long-run covariance between prices with DOLS estimators has been used because it plays a major role in much of times series inference and cointegration regression. To implement this method, we follow the same criteria necessary to apply a VECM and VAR model such as stationarity condition and if monthly food prices are Integrated of order 1, I(1). When this happens, the prices are said to be co-integrated, which allows estimating the long-run relationship between the prices. When prices are non-stationary and based on our previous VAR model, the following equation can be written: The DOLS estimators are obtained by adding the lead and lag of ΔSt to soak up the long-run correlation between error. In testing for cointegration, it is important to note that the prices of food commodities of interest are 208 likely to be affected by supply-side factors such as seasonality of production, weather, and so on. The long-run relationship of prices due to arbitrage across markets would be estimated more precisely by removing that part of the price co-movement driven by seasonality of production. The co-integrating vector gives the long run relationship between country prices for each commodity or market. 32 B.2.2. Results Table B-7: Market Integration, Cassava 209 Source: Original table based on estimation results 33 Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens Table B-8: Market Integration, Sorghum 210 Source: Original table based on estimation results 34 Table B-9: Market Integration, Yam 211 Source: Original table based on estimation results Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens Table B-10: Market Integration, Maize 212 Source: Original table based on estimation results 36 Table B-11: Market Integration, Millet 213 Source: Original table based on estimation results 37 Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens Table B-12: Market Integration, Rice 214 Source: Original table based on estimation results 38 Table B-13: Biggest Crops in Terms of Production, Consumption, and Area Harvested Share of Biggest crop area Area Biggest Biggest crop produced Yield harvested Country harvested crop produced (value in (hg/ha) in the (ha) consumed tons) total ag land (%) Benin Cassava 79,349,688 115,022 212,190 8.36 Maize Burkina Sorghum 43,525,056 9,233 1,510,210 15.6 Sorghum Faso Cabo Sugar cane 694,794 186,968 1,182 1.69 Maize Verde Chad Sorghum 18,157,424 7,089 802,658 1.65 Sorghum Côte Yams 150,527,568 82,474 641,406 3.47 Cassava d’Ivoire Gambia, Groundnuts, 3,061,498 9,799 101,599 18.11 Rice The with shell Ghana Cassava 319,169,632 134,581 724,733 5.48 Yam Guinea Rice, paddy 41,141,096 15,776 912,763 6.42 Rice 215 Guinea- Rice, paddy 4,113,840 16,800 79,993 5.41 Rice Bissau Liberia Cassava 13,506,523 69,548 62,654 2.42 Rice Mali Rice, paddy 36,465,540 23,677 448,195 1.27 Rice Mauritania Rice, paddy 3,387,942 43,171 23,980 0.06 Rice Niger Millet 79,129,720 4,352 5,759,614 16.25 Sorghum Nigeria Cassava 1,193,362,432 103,655 3,847,899 6.16 Rice Senegal Sugar cane 26,286,032 1,130,811 7,505 0.08 Peanuts Sierra Cassava 38,170,156 65,131 155,791 5.12 Rice Leone Togo Cassava 22,910,140 57,555 141,529 4.33 Maize Source: Original table based on FAO database 39 Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens Table B-14: First and Second Biggest Crops Produced, Consumed, and Harvested Biggest crops produced Biggest crops in terms of Biggest crops (value in tons) area harvested consumed Country First Second First Second First Second Benin Cassava Yams Maize Seed cotton Maize Rice Burkina Sorghum Millet Sorghum Millet Sorghum Millet Faso Cabo Sugar cane Tomatoes Pulses nes Maize Maize Rice Verde Groundnuts, Chad Sorghum Millet Sorghum Sorghum Millet with shell Côte Cocoa, Coffee, Yams Cassava Cassava Yam d’Ivoire beans green Gambia, Groundnuts, Groundnuts, Millet Millet Rice Millet The with shell with shell Cocoa, Ghana Cassava Yams Maize Yam Maize beans 216 Guinea Rice, paddy Cassava Rice, paddy Maize Rice Cassava Cashew Cashew Guinea- Rice, paddy nuts, with nuts, with Rice, paddy Rice Maize Bissau shell shell Rubber, Liberia Cassava Sugar cane Rice, paddy Rice Cassava natural Mali Rice, paddy Millet Millet Sorghum Rice Millet Mauritania Rice, paddy Sorghum Sorghum Pulses nes Rice Millet Cow peas, Cow peas, Niger Millet Millet Sorghum Millet dry dry Nigeria Cassava Yams Sorghum Maize Rice Cassava Groundnuts, Groundnuts, Senegal Sugar cane Millet Peanuts Rice with shell with shell Sierra Cassava Rice, paddy Rice, paddy Cassava Rice Cassava Leone Togo Cassava Yams Maize Sorghum Maize Rice Source: Original table based on FAO database 40 Appendix C: Methodology and Limitations Food Insecurity Model4 C.1. Methodology As demonstrated by the stochastic model used to model yield levels, there is no straightforward explanation of changes in yields coming purely from exogenous shocks. Economic impact, derived from the prices of food crops on the market, is not linked to food insecurity, given the fact that in all countries in scope for the analysis, there is some baseline chronic food insecurity that exacerbates the impact even of smaller events. To link agricultural production to food insecurity, a complementary food insecurity model is proposed. Such a model builds on chronic food insecurity numbers, namely the prevalence of undernourished people as a percentage of the total population. The following data sets are used: • Average number of calories of a normal diet per country (Humanitarian Data Exchange) • Basket of the three main food crops consumed by country (FAO, FAOSTAT) • Baseline prevalence of undernourishment per country (Humanitarian Data Exchange) • Baseline daily calories deficit for the undernourished per country (Our World in Data, Oxford). Using the data on the average basket of food crops and calories per person per day in each country, the model recomputes a daily intake of calories derived from the main crops for a well-nourished and an undernourished person. The results of the food crop risk assessment are used to calculate the likely drop in production and the consequences in terms of calories deficit for each country. The calories 217 deficit is then translated into a recomposed food basket for each country. Based on the calories deficit per person for each simulated event, the model calculates the impact in terms either of new people who are food insecure (the so-called horizontal scale-up of the food insecure) or of the depth of the food deficit for the already existing food insecure (the so-called vertical scale-up). The advantage of using such a model is that targeted measures can be put in place depending on how the food insecurity situation evolves, and these measures can then be linked to suitable scale-up of emergency response. Another advantage of using a food insecurity model is that the estimated number of people undernourished for different production shocks can be translated into expected humanitarian cost. C.2. Limitations of the Food Insecurity Model To calculate the vertical or the horizontal scale-up of food insecurity, the model makes significant simplifying assumptions. Key simplifications include the following: • The most recent available figures of percent of population undernourished were taken as baseline data. No additional analyses were conducted to verify whether this represented the 4 This appendix refers to chapter 7. 41 Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens number of chronically undernourished in any given year—for example, other shocks could have inflated temporarily the share of the undernourished population. • The model uses the average depth of undernourishment for undernourished people as per the baseline data. After shock events, the average depth may vary, however, as undernourishment of those undernourished before the shock deepens and previously well-nourished people become newly undernourished. These dynamics are reflected in the model separately (that is, the model either keeps the number of people undernourished equal to the baseline and allocates the drop in food supply to the same group, thus deepening the gap in consumed calories [see table 4.2], or it keeps constant the baseline depth of undernourishment, in which case the drop in food supply is allocated to people who were not initially food insecure, thus increasing the number of people undernourished [see table 4.3]). • Potential balancing effects of trade, food substation, or activities other than agriculture are not considered. The model assumes that all production lost translates by 100 percent into a reduction in food consumption. Thus, the results on food insecurity are likely not to be as extreme in reality as calculated by the model. This also means that the model would be expected to provide more accurate results for the four Sahel countries, where employment in agriculture exceeds 70 percent or 80 percent of employment in all countries and most farmers are subsistence farmers. Meanwhile, in Sierra Leone and Togo, where employment in agriculture amounts to 68 percent and 63 percent, respectively, results would be expected to be less accurate. In these countries, the model would be expected to overestimate food insecurity impacts due to production shocks. • The model does not establish a direct link between the number of undernourished people and levels of food insecurity, per the IPC classification system.5 For the purpose of the model, undernourished people are those who consume a daily diet with a nutritional value inferior to 218 the recommended daily intake of calories for their gender and age group; for the purpose of the calculation of the humanitarian need associated to food insecurity, the same value of food aid is allocated to each person undernourished. In reality, a severe food insecurity situation (IPC3+) will require more assistance than lower levels of food insecurity (IPC1, IPC2). 5 IPC overview and classification system, http://www.ipcinfo.org/ipcinfo-website/ipc-overview-and-classification- system/en/. 42 219 Regional Risks to Agriculture in West Africa: Agricultural Risk Impacts, Management Measures, and Financing Mechanisms Through a Regional Lens