EE T OH CC O IM NN C IACL AAN SD SE S IS AT TC NOCRE WORK AGRICULTURAL SECTOR RISK ASSESSMENT IN NIGER: Moving from Crisis Response to Long-Term Risk Management JANUARY 2013 REPORT NUMBER: 74322-NE T E C H N I C A L A S S I S TA N C E AGRICULTURAL SECTOR RISK ASSESSMENT IN NIGER: Moving from Crisis Response to Long-Term Risk Management A gr icultur e and Envir onm ental Ser vices (A ES ) D epartm ent and A gr icultur e, Rur al D evelopm ent, and Irr igation (A FTAI ) Unit in the A fr ica Region Report No. 74322-NE © 2013 International Bank for Reconstruction and Development/International Development Association The World Bank 1818 H Street NW Washington DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org This volume is a product of the staff of the International Bank for Reconstruction and Development/The World Bank. The �ndings, interpretations, and conclusions expressed in this paper do not necessarily reflect the views of the Executive Directors of The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this work is subject to copyright. Because The World Bank encourages dissemination of its knowledge, this work may be reproduced, in whole or in part, for noncommercial purposes as long as full attribution to this work is given. Any queries on rights and licenses, including subsidiary rights, should be addressed to the Of�ce of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2422; e-mail: pubrights@worldbank.org. Cover photos: Vikas Choudhary and El Hadj Adama Toure C O NT E NT S III Table of CONTENTS List of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii List of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi Acronyms and Abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii Executive Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv Chapter 1: Niger: Introduction and Context . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.1 Context: Living with Risk . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.2 Traditional Capacity to Manage Risk . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 1.3 Agricultural Sector Risk Review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Chapter 2: Agricultural System in Niger . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 2.1 Agro-Climatic Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 2.2 Land and Water Resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 2.3 Crop Production System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 2.4 Production Trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 2.5 Cereal Crops: Food Supply and Demand . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 2.6 Cash Crops . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 2.7 Livestock . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 2.7.1 Transhumant-Nomadic Patoral Zone . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 2.7.2 Agro-Pastoral Zone . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 2.7.3 Livestock in the Agricultural Zone. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 Chapter 3: Agriculture Sector Risks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 3.1 Rainfall Trends 1980–2009 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 3.2 Production Risks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 3.2.1 Drought . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 3.2.2 Locusts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 3.2.3 Livestock Diseases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 3.2.4 Floods. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 3.2.5 Crop Pests and Diseases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 3.2.6 Bush�res and Windstorms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 3.3 Market Risks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 3.3.1 Trends in Average Annual Cereal and Cowpea Prices . . . . . . . . . . . . . . . . . . . . . . . . . 20 3.3.2 Livestock-Cereal Price Ratios . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 3.3.3 Seasonal Price Movements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 3.3.4 Exchange Rate Risk . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 TEC H N I C A L AS S I STANCE IV C ONTENTS 3.4 Conflicts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 3.5 Insecurity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 3.6 Macroeconomic Shocks in Nigeria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 3.7 Political Instability and Regulatory Risks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 3.8 Interlinkages between Risks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 3.8.1 Independent Risks Occurring Simultaneously . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 3.8.2 Dependent Risks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 3.8.3 Regional Shocks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 3.8.4 Management of One Risk Creating Another Risk . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 3.8.5 Differential Impact of Risk . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 Chapter 4: Adverse Impact of Agricultural Risks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 4.1 Conceptual and Methodological Basis for Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 4.1.1 Loss Thresholds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 4.1.2 The Indicative Value of Losses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 4.1.3 Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 4.2 Crop Production Risks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 4.2.1 Major Cereal Crops . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 4.3 Crop Price Risk. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 4.4 Locust Attacks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 4.5 Floods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 4.6 Expected Losses and Risk Priorities for Crop Production. . . . . . . . . . . . . . . . . . . . . . . . . . . 31 4.7 Losses to the Livestock Sector Due to Major Droughts . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 4.8 Impact of Agricultural Risks on National GDP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 Chapter 5: Risk Prioritization and Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 5.1 Risk Prioritization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 5.2 Agricultural Risk Management Measures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 5.2.1 Drought Tolerant and Improved Seed Varieties . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 5.2.2 Soil and Water Conservation and Natural Resource Management . . . . . . . . . . . . . . . . . . 39 5.2.3 Desert Locust Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 5.2.4 Irrigation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 5.2.5 Shortening Emergency Response Time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 5.2.6 Contingent Financing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 5.2.7 Strategic Destocking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 5.2.8 Vaccination and Veterinary Services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 5.2.9 Community Level Food and Fodder Banks. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 5.3 Prioritization of Risk Management Measures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 5.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 Annex 1: Disaster Risk Reduction and Crisis Management: Lessons Learned and Positive Changes since 2005, and Remaining Challenges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47 Annex 2: Rainfall Analysis and Drought Classi�cation . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 AGR ICULTURA L SEC TOR RISK ASS ES S MENT IN NIGER C O NT E NT S V Annex 3: Livelihood Zones and Principle Risks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55 Annex 4: Grain Reserves—OPVN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 Annex 5: AGRHYMET’s Analysis of Climate Determinants on Crop Yield . . . . . . . . . . . . . . . 59 A5.1 Relationship between Grain Yield and the Climatic Factors . . . . . . . . . . . . . . . . . . . . . . . . 60 A5.1.1 Case of Millet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 A5.1.2 Case of Sorghum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61 A5.2 Prinipal Component Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 A5.3 Trends of Rainfall and on Rainy Season Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 A5.3.1 Context . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 A5.3.2 Trend Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64 TEC H N I C A L AS S I STANCE LI ST O F F I G U R E S V II List of FIGURES Figure E.1: Annual GDP Growth (%) and GDP per Capita (%) (1984–2010) . . . . . . . . . . . . . . . . . . xv Figure E.2: Major Shocks to Crop and Livestock Production . . . . . . . . . . . . . . . . . . . . . . . . . . xvi Figure E.3: Expected Average Losses for Adverse Crop Production Events . . . . . . . . . . . . . . . . . xviii Figure E.4: Risk Prioritization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xviii Figure 2.1: Agro-Climatic Zones . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 Figure 2.2: Crop Area and Rural Population Trends. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 Figure 2.3: Rainy Season Crop Area Trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 Figure 2.4: Millet Production . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 Figure 2.5: Sorghum Production. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 Figure 2.6: Cowpea Production . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 Figure 2.7: Cereal Food Production vs. Consumption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 Figure 3.1: Average Annual Rainfall . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 Figure 3.2: Major Shocks to Crop and Livestock Production . . . . . . . . . . . . . . . . . . . . . . . . . . 16 Figure 3.3: Real Cereals and Cowpea Prices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 Figure 3.4: Livestock-Millet Price Ratios (Meat) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 Figure 3.5: Monthly Millet Prices, Maradi . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 Figure 3.6: Exchange Rate: FCFA/Naira (Parallel Market) . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 Figure 4.1: Expected Average Loss for Adverse Crop Production Events . . . . . . . . . . . . . . . . . . . 32 Figure 4.2: Chronology of Adverse Crop Production Events. . . . . . . . . . . . . . . . . . . . . . . . . . 32 Figure 4.3: Annual GDP Growth and GDP per Capita . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 Figure 5.1: Risk Prioritization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 Figure 5.2: Risk Layering Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 Figure A5.1: Agro-Climatic Zones in Niger . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 Figure A5.2: Principal Component Analysis of Dakoro (1980–2010) . . . . . . . . . . . . . . . . . . . . . . . 64 Figure A5.3: Trend of Total Annual Rainfall at Gaya (Soudano-Sahelian Zone) . . . . . . . . . . . . . . . . . 66 Figure A5.4: Trend of Number of Rain Days per Year at Gaya (Soudano-Sahelian Zone) . . . . . . . . . . 66 TEC H N I C A L AS S I STANCE VIII LIST O F FIGUR ES Figure A5.5: Trend of Total Annual Rainfall at Filingue (Sahelian Zone) . . . . . . . . . . . . . . . . . . . 67 Figure A5.6: Trend of Duration of Rainy Season at Filingue (Sahelian Zone). . . . . . . . . . . . . . . . . 67 Figure A5.7: Trend of the Longest Dry Spell at Filingue (Sahelian Zone) . . . . . . . . . . . . . . . . . . . 68 Figure A5.8: Yearly Development of Minimum Temperatures Observed at Birni Koni . . . . . . . . . . . . 68 AGR ICULTURA L SEC TOR RISK ASS ES S MENT IN NIGER LI ST O F TA B L E S IX List of TABLES Table E.1: Decision Filters and Intervention Classi�cation . . . . . . . . . . . . . . . . . . . . . . . . . . . . xix Table 1.1: Reported Household Impacts of Shocks Ranked as Most Important . . . . . . . . . . . . . . . . . 1 Table 2.1: Agro-Climatic Zones . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 Table 2.2: Percent Change in Food Crop Production . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 Table 2.3: Coef�cients of Variation: Production, Area, and Yields . . . . . . . . . . . . . . . . . . . . . . . . 11 Table 2.4: Characteristics of Major Cash Crops . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 Table 3.1: Summary of Drought Events . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 Table 3.2: Summary of Locust Infestations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 Table 3.3: Incidence of Livestock Disease 1995–2010 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 Table 3.4: Recorded Losses Due to Selected Major Floods . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 Table 3.5: Crop Price Trends and Characteristics: 1980–2011 . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 Table 4.1: Frequency, Severity, and Cost of Adverse Events for Crop Production . . . . . . . . . . . . . . . 28 Table 4.2: Frequency, Severity, and Cost of Adverse Events for Millet Production . . . . . . . . . . . . . . . 29 Table 4.3: Frequency, Severity, and Cost of Adverse Events for Sorghum Production . . . . . . . . . . . . . 30 Table 4.4: Frequency, Severity, and Cost of Adverse Price Events for Crops . . . . . . . . . . . . . . . . . . 30 Table 4.5: Frequency, Severity, and Estimated Costs of Locust Attacks . . . . . . . . . . . . . . . . . . . . . 30 Table 4.6: Frequency, Severity, and Estimated Costs to Agriculture of Floods . . . . . . . . . . . . . . . . . 31 Table 4.7: Losses to the Livestock Sector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 Table 5.1: Indicative Agricultural Risk Management Measures . . . . . . . . . . . . . . . . . . . . . . . . . 37 Table 5.2: Relative Bene�ts of Risk Management Measures . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 Table 5.3: Decision Filters and Intervention Classi�cation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 Table A2.1: Frequency of Adverse Rainfall Events by Region . . . . . . . . . . . . . . . . . . . . . . . . . . 49 Table A2.2: Comparison of 2004 and 2009 Drought by Department . . . . . . . . . . . . . . . . . . . . . . 51 Table A2.3: Niger: Regional Share of Major Crop Production (2010) . . . . . . . . . . . . . . . . . . . . . . 51 Table A2.4: Frequency of Adverse Rainfall Events by Department . . . . . . . . . . . . . . . . . . . . . . . 52 Table A2.5: Frequency of Adverse Rainfall Events (Drought) at Departmental Level . . . . . . . . . . . . . 54 TEC H N I C A L AS S I STANCE X LIST OF TA B LES Table A5.1: Weather Stations Used for Different Agro-climatic Zones . . . . . . . . . . . . . . . . . . . . . 59 Table A5.2: Correlation between Millet Grain Yield and Rainy Season Variables . . . . . . . . . . . . . . . . 60 Table A5.3: Correlation between Sorghum Grain Yield and Rainy Season Variables . . . . . . . . . . . . . . 61 Table A5.4: Percentage of Variation Explained by the Principal Axes (F1 and F2) . . . . . . . . . . . . . . . 62 Table A5.5: Signi�cant Trend of the Variables (1980–2010) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 AGR ICULTURA L SEC TOR RISK ASS ES S MENT IN NIGER AC K N OW L E D G ME NT S XI ACKNOWLEDGMENTS Niger agricultural sector risk assessment was conducted jointly by Agriculture and Environmental Services (AES) Department and Agriculture, Rural Development, and Irrigation (AFTAI) Unit in the Africa Region. This report was prepared by a team led by El Hadj Adama Toure (AFTA1, Task Team Leader) and Vikas Choudhary (AES, Co-task Team Leader), and consisting of Garry Christenson (Consultant, Agricultural Economist, AES), and Henri Josserand (Consultant, Crop and Livestock Specialist, AES). Edgar Uribe (Consultant, Weather Risk Management, AES) and AGRHYMET conducted weather analysis for the report, while Mohamadou Issaka Magha (Consultant, AFTA1) provided local coordination and data collection support. The team is grateful to the Government of Niger, in particular to Amadou Allahoury Diallo (High Commissioner of the 3N Initiative), Illa Djimra, (Secretary General), and M. Dagna, (Deputy Secretary General of the Ministry of Agriculture), Diamoitou Boukary, (Secretary of the Ministry of Livestock), and their staff for their helpful collaboration and contributions to the work during and beyond the �eld mission. The team would like to thank Ousmane Diagana (Country Director, Niger), Nestor Kof� (Country Manager, Niger), Martien Van Nieukoop (AFTA1), and Marc Sadler (AES) for their valuable guidance and support. Christoph Pusch (Global Facility for Disaster Risk Reduction [GFDRR]), Erick C.M. Fernandes (Latin America and the Caribbean Region Agricultural and Rural Development Unit [LCSAR]), and Loraine Ronchi (Investment Climate for Industry Unit [CICIN]) were peer reviewers of the assessment. Valuable comments were received from David Rohrbach (AFTA3), Steven N. Schonberger (Agriculture, Rural Development, and Irrigation Department Practice 2 [AFTA2]), and Robert Utz (Poverty Reduction and Economic Management Practice 4 of the Africa Region [AFTP4]). Kathryn Bach provided valuable editorial input. Administrative support was provided by Hadidia Djimba (World Bank Niamey Of�ce) and Germaine M. Ethy (AFTA1). Guidance and suggestions were provided by Amadou Alassane (Senior Agricultural Services Specialist, AFTA1), Carlos Arce (Senior Economist, AES), and Janet Owens (Senior Economist, AFTP4) throughout the assessment process and the team would like to thank them for their support. The work was �nanced by the World Bank and a Multidonor Trust Fund supported by the Netherlands Ministry of Foreign Affairs and the Swiss Secretariat for Economic Cooperation (SECO). TEC H N I C A L AS S I STANCE AC R O N Y MS A N D A BBRE VIAT IONS X III ACRONYMS AND ABBREVIATIONS January 1 – December 31 Currency Equivalents Currency unit: Franc CFA (FCFA) US$1= 495 FCFA (Exchange rate 2010) Weights and Measures Metric system AES Agriculture and Environmental Services CIC Centre d’Information et de Communication AFTA1 Agriculture, Rural Development, and Irrigation Department Practice 1 CICIN Investment Climate for Industry Unit AFTA2 Agriculture, Rural Development, and CILSS Permanent Inter-State Committee for Irrigation Department Practice 2 Drought Control in the Sahel AFTA3 Agriculture, Rural Development, and CMC Commission mixte Etat-Donateurs Irrigation Department Practice 3 CNEDD Counseil National de l’Environnement AFTA4 Poverty Reduction and Economic pour un Développement Durable Management Practice 4 of the Africa CofoCom Commissions foncières communales Region CofoB Commissions foncières de base AGRHYMET Agriculture, Hydrology, Meteorology CofoDep Commissions foncières départementales APCAn Assistance in the Management of the National Prevention System of the CPI consumer price index National Prevention System for the CRC Comité Restreint de Concertation Nutrition Crisis of Niger CV coef�cient of variation ARD Agriculture and Rural Development Department CVCA Climate Vulnerability and Capacity Assessment BCEAO Banque Centrale des États de l’Afrique de l’Ouest (Central Bank of West African DNPGCA Dispositif National de Prévention et de States) Gestion des Crises Alimentaires au Niger CAADP Comprehensive African Agricultural ECB Emergency Capacity Building Development Program ECOWAS Economic Community of West African CAP Community Action Program States CCA Cellule Crises Alimentaires ENBC Enquete Nationale sur le Budget et la Consommation des Ménages CC/SAP Cellule de Coordination du Système d’Alerte Précoce ENVAM Enquete sur la Vulnérabilité Alimentaire des Ménages CGE computable general equilibrium EWS early warning system CIS Centre for International Cooperation FAO Food and Agriculture Organization CLCPRO Commission for Controlling the Desert Locust in the Western Region FAO EMPRESS FAO Emergency Prevention System TEC H N I C A L AS S I STANCE XI V AC R ONY MS A ND A BB R EV IATIONS FAOSTAT FAO Statistics Division PCA Principal Component Analysis FCFA West African CFA Franc PANA Programme d’Action National pour l’Adaptation FCMPS Food Crisis Management and Prevention System PRODEX Agro-sylvo-pastoral Export and Markets Development Project FEWS NET Famine Early Warning System Network RGAC Recensement General de L ’Agriculture et FMNR Farmer Managed Natural Regeneration du Cheptel, Government of Niger GAO gross agricultural output SAFEX South African Commodity Exchange GDP gross domestic product SCAP-RU Système Communautaire d’Alerte Précoce GFDRR Global Facility for Disaster Risk Reduction et de Réponse aux Urgences GFRP Global Food Crisis Response Program SDR Stratégie pour le Développement Rural (Country’s Rural Development Strategy) GIEWS Global Information and Early Warning System SIMA Système d’Information sur le Marché Agricole GON Government of Niger SNS Stock National de Sécurité HC3N Haut Commissariat à l’Initiative 3 N (les nigériens nourissent les nigériens) SSA Sub-Saharan Africa INS Institut National de la Statistique (Niger) TFP total factor productivity LCSAR Latin America and the Caribbean Region TLU tropical livestock units Agricultural and Rural Development Unit. UEMOA L’Union Economique et Monétaire Ouest NGAC Niger General Agricultural Census Africaine NGO nongovernmental organization UNDP United Nations Development Programme NRM natural resource management UNEP United Nations Environment Programme ODA of�cial development assistance UNICEF United Nation Children’s Fund OECD Organization for Economic Co-operation UN/OCHA Of�ce for the Coordination of Humanitarian and Development Affairs of the United Nations Secretariat OIE The World Organisation for Animal Health USAID United States Agency for International Development OPVN Of�ce des Produits Vivriers du Niger VAM Vulnerability Analysis and Mapping OSV Observatory for Monitoring Vulnerability WFP World Food Programme PAC Community Action Plan Vice President: Makhtar Diop Sector Director: Jamal Saghir Sector Manager: Nestor Kof� Sector Manager: Martien Van Nieukoop Task Manager(s): El Hadj Adama Toure and Vikas Choudhary AGR ICULTURA L SEC TOR RISK ASS ES S MENT IN NIGER EX E C U T I V E S U MMARY XV EXECUTIVE SUMMARY Niger, owing to its climatic, institutional, livelihood, economic, and environmental context, is one of the most vulnerable coun- tries of the world. Poverty is pervasive in Niger and it ranks low on almost all the human development indicators. Agriculture is the most important sector of Niger’s economy and accounts for over 40 percent of national gross domestic product (GDP) and is the principle source of livelihood for over 80 percent of the country’s population. The performance of the agricultural sector, however, due to its high exposure to risks, is very volatile. Niger has experienced multiple shocks, largely induced by agricultural risks over the past 30 years, which impose high welfare cost in terms of food availability, food affordability, and malnutrition. It also adversely affects household incomes, performance of the agricultural sector, the government’s �scal balance, and the growth rate of Niger’s economy (see �gure E.1 below). Recognizing the need to explicitly and comprehensively address agricultural risks, the Government of Niger, through 3N high commissioner, requested the World Bank to conduct an agricultural sector risk assessment of Niger. This risk assess- ment enriches the existing knowledge base of the agricultural sector in Niger and provides the following contributions: (1) systematically analyzes a whole range of agricultural risks and its impact over a longer time period (1980–-2012); (2) helps situate drought in the context of other agricultural risks; (3) prioritizes the most important agricultural risks for the country FIGURE E.1: Annual GDP Growth (%) and GDP per Capita (%) (1984–2010) GDP growth (annual %) GDP per capita growth (annual %) 10 5 0 percent 2000 2002 2003 2004 2005 2006 2007 2008 2009 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2001 2010 –5 drought drought political drought drought –10 1987 1990 instability +locust political drought 2009 1999* 2004 instability 2000 –15 1992* –20 drought 1984 year Sources: World Development Indicators Database, 2012; and Author’s calculations. *1992 political instability (transitional government November 1991–April 1993)1 and 1999 political instability (sssassination of President Ibrahim Bare’ Mainassare).2 1 President Ali Saibou’s regime acquiesced to demands for elections, and a transitional government was installed in November 1991 to manage the affairs of state until the institutions of the Third Republic were put in place in April 1993. The economy deteriorated over the course of the transition, leading to a fall in GDP in 1992. 2 In April 1999, President Baré Mainassare was assassinated in a coup led by Major Daouda Malam Wanke, who established a transitional National Reconciliation Council to oversee the drafting of a constitution for a Fifth Republic. The Nigerien electorate approved the new constitution in July 1999 and held legislative and presidential elections in October and November 1999. The council transitioned to civilian rule in December 1999; however, the period of transitional government, similar to 1992, had severe consequences for the economy. TEC H N I C A L AS S I STANCE XV I EX EC UTIV E S UM M A RY based on objective criterion; (4) provides a framework of mitigation-transfer-coping to manage priority risks; and (5) offers a �ltering mechanism to select high return interventions for agricultural risk management. PRODUCTION RISKS Drought, locusts, livestock diseases, crop pests and diseases, floods, windstorms and bush�res are the main sources of pro- duction risk. Farmers also complain of the risks to crop production from livestock herds, although the incidence and severity of these events is dif�cult to determine. The impact of some of these events on crop production for the period 1980–2011 is indicated in �gure E.2 below, using indices of agricultural production. Drought is the principle risk in Niger and the country has experienced seven droughts between 1980–2010, with adverse impact on national agricultural production. Over the past 12 years, Niger has witnessed four years (2001, 2005, 2010, and 2012) of severe food insecurity that resulted in appeal for international humanitarian assistance and food relief. Drought is also the principle trigger for spikes in food prices and conflicts over pasture and water; it is highly correlated with some crop pests and diseases, and it aggravates mortality and morbidity due to livestock diseases. Locust outbreak is another high frequency-high severity risk in Niger. Almost one-third of losses during the 2004–05 crises were due to locust, with adverse impact on both the crop and livestock sectors. Considering the signi�cance of livestock for Niger’s economy, livestock diseases, especially pasteurellose, anthrax, peste des petits ruminants, and Newcastle disease (for poultry), are another principle risk for the country. Flood incidences are increasing in Niger; however, they do not pose a serious risk to the broader agricultural sector, due to localized nature of flooding. Furthermore, most of the flood years are usually associated with bumper harvest because of higher than average rainfall at an aggregate level for the country. Crop pests and diseases, like striga and fungal diseases, are a perennial problem among most crops; however, overall losses from a majority of these pests and diseases, barring the exception of grasshoppers, grain eating bird, and millet borer, are frequent but relatively low. FIGURE E.2: Major Shocks to Crop and Livestock Production 200 cereals crops 180 livestock 160 production index (2004–2006 = 100) 140 120 100 80 drought floods 60 drought locusts 40 drought drought floods 20 drought pests locusts locusts locusts drought 0 locusts 19 0 19 1 19 2 19 3 84 19 5 19 6 19 7 88 19 9 19 0 19 1 92 19 3 19 4 19 5 19 6 19 7 19 8 20 9 20 0 01 20 2 20 3 20 4 20 5 20 6 20 7 20 8 09 10 8 8 8 8 8 8 8 8 9 9 9 9 9 9 9 9 9 0 0 0 0 0 0 0 0 19 19 19 19 20 20 years Source: Author’s calculations. AGR ICULTURA L SEC TOR RISK ASS ES S MENT IN NIGER EX E C U T I V E S U MMARY X V II Bush �re is often a problem for pastoral areas and windstorms damage young plants at the beginning of copping season. Nonetheless, the overall impact of these two risks on the agricultural sector is negligible. MARKET RISKS Food price volatility is a big concern for consumers and major spikes in nominal prices occurred in 1998, 2001, 2002, 2005, 2009, and 2010. These spikes are also observed for real prices, although they were much less pronounced in 2009 and 2010. There is a strong association between seasonal price movements and the incidence of drought and other adverse events. The hardship endured during these periods of adversity seem to con�rm the growing consensus that reduced access to food (high prices), along with reduced food availability, may be the most critical impact of drought and locust attacks in Niger. High seasonal price spikes appear to be more closely and systematically associated with these events than inter-annual changes in production or prices. Due to Niger’s heavy reliance on trade with Nigeria, adverse movements in the West African CFA Franc (FCFA)/naira exchange rate could also be a potential source of risk. However, market data shows that the FCFA/naira exchange rate is relatively stable, with low adjusted coef�cients of variation (.06-–0.08) for monthly exchange rates for the period January 2003 to October 2011. ENABLING ENVIRONMENT RISKS Political instability is a major risk in Niger, which has had a tumultuous political history with four coups d’état since indepen- dence (1974, 1996, 1999, 2010). Niger has witnessed two transitional governments that were associated with rapid deteriora- tion of the economy (�gure E.1) and a sharp decline in GDP growth rates. The impact on the agricultural sector, however, was much less pronounced and more indirect and might include: (1) reduced access to particular regions, which means that rural markets are more restricted, food prices rise, and aid can’t get through; (2) reduced public and private investment in response to higher levels of uncertainty; (3) the diversion of public expenditure to military purposes to the detriment of other public services; and (4) loss of donor support. Political instability may have a bigger impact on the agriculture sector when it coincides with other shocks like drought, such as in 1995–97 . Political instability induces changes in government priorities and contributes to volatility in agricultural sector funding. Rising pressure on common property resources, or on resources used within the bounds of customary law, have led to frequent, but localized, conflicts affecting herders and farmers. Insecurity has always been an issue for herders practicing transhumance, but of late, the situation has deteriorated, especially in the border areas of Mali, Libya, and Nigeria (Chad is an endemic problem). Despite their signi�cance for herders in some locations, from a macroperspective, the impact of insecurity and conflict on the broader agricultural sector is relatively small. Macroeconomic shocks in Nigeria, as the largest trading partner of Niger, can have potentially serious repercussions on the agricultural sector; however, so far, past impacts of such shocks have been moderate and short lived. The assessment team analyzed downside deviation from the trend and correlated them with adverse events to calculate frequency and indicative losses from major production risks for crops. Figure E.3 highlights the result of that analysis. A combination of qualitative and quantitative measures, based on the assessment team’s evaluation, was used to prioritize major risks for the entire agricultural sector, both livestock and crops (�gure E.4). This analysis highlights six priority risks, namely 1) drought (crop), 2) drought (livestock), 3) locust outbreaks, 4) consumer price risk, 5) livestock diseases and 6) politi- cal instability. To address the priority risks, the assessment deployed a holistic agricultural risk management framework, comprising of mitigation (action taken to reduce the likelihood of events, exposure, and/or potential losses), transfer (risk transfer to a will- ing party, at a fee or premium), and coping (activities geared to help cope with losses) solutions to identify a list of potential interventions. Risk transfer solutions (insurance and hedging), owing to Niger’s speci�c context, have limited applicability and will be quite challenging to implement. Coping solutions (social safety net programs) are required and quite important in Niger; TEC H N I C A L AS S I STANCE XV I I I EX EC UTIV E S UM M A RY FIGURE E.3: Expected Average Losses for Adverse Crop Production Events 90 80 70 indicative loss $US m 60 drought 50 40 30 20 locusts price 10 floods 0 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 frequency Source: Author’s calculations. FIGURE E.4: Risk Prioritization crop pest and diseases (0.4) high Livestock pest and drought- diseases (0.35) crop (0.35) conflict frequency of occurrence flood (0.3) consumer price risk (0.3) insecurity political instability/ NIRA-CFA exchange regulatory risk locust (0.2) rate (0.2) in Niger (0.2) macro economic drought- shocks in Nigeria livestock (0.1) producer price risk (0.05) low negligible catastrophic severity of impact Source: Author’s calculations. AGR ICULTURA L SEC TOR RISK ASS ES S MENT IN NIGER EX E C U T I V E S U MMARY X IX TABLE E.1: Decision Filters and Intervention Classi�cation POTENTIAL ADVERSE IMPACT ON RELATIVE EASE OF IMPACT ON POVERTY SCALABILITY COST IMPLEMENTATION RETURN TIME ENVIRONMENT ALLEVIATION Drought tolerant/improved seed High Medium Medium Short Low High varieties (M) Soil and water conservation (M) High Medium Medium Medium Low High Irrigation (M) Low High Low Short–medium Moderate High Early detection and destruction of High Medium High Short Moderate Low locusts (M) Community-level food and fodder High Medium Medium Short Low High banks (M, C) Vaccination programs (M) High Medium Medium Medium Low High Contingent �nancing (C) High Low High Short Low Low Shortening emergency response Medium Low Medium Short Low Low time (C) Strategic de-stocking (C) Low Medium Low Medium Low Low Insurance (T) Low Low Medium Medium Low Low Source: Authors. Note: M is Mitigation, C is coping, and T is Transfer. however, they do not address fundamental risk issues in the agricultural sector and have limited applicability as a long-term solution. Risk mitigation is perhaps the most required, but much ignored, with the highest returns in addressing short- and long-term issues in Niger’s agricultural sector. It is important to highlight that most of these potential interventions are comple- mentary in nature and most of them are required to effectively address agricultural risks in Niger. Nonetheless, considering the resource-constrained environment of Niger, decision �lters (see table E.1) were used to help prioritize the interventions. Using these �lters, the following types of interventions were recommended with the potential to generate sizable risk man- agement bene�ts:  Drought-tolerant, high-yielding crop varieties. Despite its importance, less than 6 percent of farming households in Niger have access to drought-tolerant cereal varieties. Early warning about the impending weather season coupled with ready availability of drought-tolerant varieties could help mitigate the risk of crop failure. This will necessitate development of a “sustainable seed system, � consisting of seed research, seed multiplication, and seed delivery on a large scale.  Soil and water conservation; natural resource management (NRM) techniques. Effective soil and water conservation techniques in Niger have successfully contributed to (1) conserving rain water, (2) increasing its in�ltration, and (3) enhancing plant growth, which improves the resilience of crops during water stress and serves as a useful drought mitigation intervention. Further expansion and consolidation of water conservation and NRM interventions will contrib- ute to greater integration of the agriculture-livestock sector, yield improvements, and drought risk management.  Irrigation. Expansion of irrigation facilities can serve as an important drought risk management tool, considering that uneven rainfall distribution is one of the principle causes of crop failure in Niger. Despite the desert conditions, there is considerable potential to increase the area under irrigation in Niger, which could contribute to improved nutrition, access to diversi�ed food, and improved household income, thereby reducing food affordability issues and improving household food security.  Continuous support to early detection and eradication of desert locusts. Ex-ante preventive action reliant on monitor- ing of seasonal reproductive areas, localization, and destruction of �rst locust populations has been effective so far by successfully controlling a potential locust outbreak in 2009, largely through use of biopesticides. Therefore, this approach should be supported and strengthened. TEC H N I C A L AS S I STANCE XX EX EC UTIV E S UM M A RY  Supporting community-level interventions for food and fodder banks. To ensure availability of food and fodder at the local level, support for community food and fodder banks, on an as-needed basis, should be provided in areas known to be suffering, or are likely to, suffer from food shortages or food price spikes. Targeting could be based on indicators of food access and food availability, using price data. The aim should be to intervene earlier in the seasonal cycle, well before prices reach their seasonal peak. Besides ensuring food and fodder availability for vulnerable populations, such interventions will also help stabilize food and fodder prices for wider populations and can help respond to local level market failures without creating major distortions.  Livestock vaccination. Vaccination is perhaps one of the most signi�cant measures to reduce the risk of livestock dis- eases. With limited resources, the government of Niger could focus on preventive vaccination campaigns against the biggest threats and respond to some of the worst outbreaks. Niger is a case of living perpetually with risk, thus more emphasis on long-term structural solutions, rather than short-term quick �xes, is required to improve the resilience of the agricultural sector. Designing and implementing a comprehensive agricultural risk management strategy will require sustained and substantial �nancial investments, shifting the focus from short-term crisis response to long-term risk management, streamlining disparate donor investments and isolated interven- tions toward the core problem, supporting decentralized community- and farm-level decision making, integrating agricultural risk management into the existing development frameworks, prioritizing agricultural risks into government and donor strate- gies, and focusing on implementation. AGR ICULTURA L SEC TOR RISK ASS ES S MENT IN NIGER C H A P T E R 1 — N I G E R: INT RODUCT ION AND CONT E XT 1 Chapter 1: NIGER: INTRODUCTION AND CONTEXT 1.1 CONTEXT: LIVING WITH RISK in the economy. Most of the chronic poor are crop farmers of The climatic, institutional, livelihood, economic, and environ- which almost 8 in 10 live in households where the principal mental contexts of Niger make it one of the most vulner- activity is crop farming. Livestock, on the other hand, fares able countries of the world. Over the years, Niger has faced slightly better with only 2 percent of the chronic poor en- multiple incidences of droughts, famine, locust invasion, gaged primarily in livestock rearing. political instability, violent conflicts, floods, and cholera and Though the majority of households engage in some farming, meningitis outbreaks. Niger is rated as the most susceptible almost all households are net purchasers of food. Over 60 country to risks and is the second-most vulnerable next only percent of households rely in part on their own production to to Afghanistan, among a total of 173 countries analyzed in meet their consumption needs. Yet, food accounts for over 60 the World Risk Report (2011). percent of household expenditures, since most households Poverty is pervasive in Niger and almost 65 percent of rural do not produce enough to meet their consumption needs. households are poor compared to 41 percent of urban house- Food insecurity is a big concern for Niger. In 2006, more than holds. Considering that 84 percent of Niger’s population is 50 percent of Niger’s population was estimated to be chroni- concentrated in rural areas, poverty dominates the lives of cally food insecure, with 22 percent suffering from extreme rural households. Niger, with a per capita income of $358 chronic food insecurity (per capita caloric consumption of (2010), fares low on almost all the development indicators less than 1800 calories per person per day). In addition, much and ranks among the lowest (186) in the United Nations of population frequently suffers from seasonal and transitory Development Programme (UNDP) human development rank. food insecurity. There have been marginal drops in poverty rates since 2005; however, the burgeoning population of 16 million, with the Niger is highly prone to “shocks, � which exacerbate a high highest population growth rate in Africa (3.5 percent), have level of chronic poverty and food insecurity. Furthermore, eroded any gains made in poverty alleviation. these shocks serve as poverty traps and aggravate the con- ditions for transient poor populations and create transitory More than 80 percent of working-age adults are employed in food insecurity. The World Bank’s Poverty Assessment (2011) agriculture, yet this sector has the lowest level of productivity highlighted poor harvest as the biggest shock (table 1.1) for TABLE 1.1: Reported Household Impacts of Shocks Ranked as Most Important SELF-REPORTED IMPACTS (PERCENT OF HOUSEHOLDS THAT REPORTED THE SHOCK/PROBLEM AS MOST IMPORTANT) NUMBER OF HOUSEHOLDS LOSS OF ASSET LOSS PLUS RANKING SHOCK AS MOST LOSS OF INCOME LOSS DURABLE OR INCOME LOSS IMPORTANT DURING PAST HOUSEHOLD AND REDUCED PRODUCTIVE OR REDUCED NO REPORTED SHOCK/PROBLEM 12 MONTHS INCOME CONSUMPTION ASSETS CONSUMPTION CONSEQUENCES Poor harvest 1,303 0 4 11 28 56 Lack of money 945 2 6 11 20 61 Food price increase 587 2 2 5 9 82 Lack of water 188 0 2 3 6 89 Serious illness or 143 9 8 15 24 45 disease Source: World Bank, May 2011. TEC H N I C A L AS S I STANCE 2 CH A PTER 1 — NIGER: INTROD UC TION A ND C ONTEX T rural households, which led to reduced consumption, loss of or explicitly addressed certain aspects of agricultural risks. productive and durable assets, and loss of income for about Following the 2005 food crisis, the government enhanced its 45 percent of affected households. The World Bank (2011) Food Crisis Management and Prevention System (FCMPS) further estimated that the average per capita consumption in (see annex 1) to incorporate controlling risks, improving food 2010, due to the 2009 drought, was about one-quarter less security, and ensuring sustainable management of natural re- (24 percent lower) than the average in 2007 . Also, households sources as objectives of the government’s Rural Development living in areas where rainfall was at least 100 millimeters less Strategy (RDS). Currently, there are four dominant institutional than the 20-year mean had a per capita consumption of about systems; however, they generally focus more on emergency 7 to 13 percent less than the reference households not ex- response rather than on risk management. This current insti- posed to shock. tutional environment suffers from a number of well-known limitations, including the following: Not only does the frequent occurrence of these shocks impose high welfare costs in terms of food availability, food  There is a high reliance on external actors and donors, affordability, and malnutrition challenges for individual house- which complicates the task of enforcing strategic focus holds, it also adversely affects household incomes, the per- and consistency over time and cross-sectoral coordina- formance of the agricultural sector, the government’s �scal tion, especially since the short-term priorities of some balance, and the growth of Niger’s economy. of these actors have fluctuated widely over time.  There are a number of good strategy documents; however, the weak implementation of these strategies 1.2 TRADITIONAL CAPACITY TO MANAGE RISK is further complicated by political instability. Niger is a high-risk environment and traditional livelihood  The combination of political and institutional changes strategies have been designed to cope with a harsh, dry, un- has led to the creation of new strategies or frequent certain, and high-stress environment. The traditional symbi- changes in key framework documents, introducing otic relationship between livestock and farming community, uncertainty about previous commitments and render- including, farming practices using the limited water available; ing coordination dif�cult among all parties. growing crop types and varieties that can withstand water  There are established coordination mechanisms stress for long dry spells; seasonal and long-term migration; and some of their technical components have been and relying on communal networks for coping, have all been showing de�nite improvements, but in times of crisis, an integral part of livelihood strategies to survive in Niger’s it remains dif�cult and time consuming to reach a high-risk environment. consensus regarding assessment and response. These traditional livelihood strategies are becoming weaker  The integration of crisis prevention into various na- and less effective in the face of the changing context. Erratic tional development strategies remains partial, and the rainfall patterns, increasing temperature, movement of iso- coexistence of disparate systems, more focused on hyets, encroachment of crop cultivation on cattle corridors, crisis response than on risk management, reinforces declining soil fertility, reduction in land holding size, increas- the conceptual divide between emergencies and risk ing household size, and high pressure on land are some of management. the factors making it increasingly dif�cult to manage risks using traditional livelihood strategies. Public support to agricultural risk management requires ad- equate information and analysis systems as well as good To support risk management strategies, there are a large num- operational capacity (for dissemination of techniques, infor- ber of institutions, projects, and almost 10 different line minis- mation, and support to risk management measures taken by tries of the government of Niger that are playing various roles. individuals in the sector). This is constrained by the weak ca- A number of government and donor3 initiatives have implicitly pacity of national institutions responsible for collecting basic data and delivering information, inputs, or technical services to producers (resources available to the typical technical 3 The Bank provided with emergency instruments (two Global Food Crisis Response Program [GFRP] supported projects since ministry appear today much more limited than they were in 2008, and a Safety Nets operation approved in 2011) and through the 1980s). Furthermore, funding in the agricultural sector making more and more flexible ongoing operations, namely the is quite volatile with signi�cant interannual volatility, which Community Action Program (CAP) and the Agro-sylvo-pastoral Export and Markets Development Project (PRODEX). further limits investments for agricultural risk management. AGR ICULTURA L SEC TOR RISK ASS ES S MENT IN NIGER C H A P T E R 1 — N I G E R: INT RODUCT ION AND CONT E XT 3 Despite weak institutional infrastructure and resource limita- The Agricultural Risk Management Team of the Agriculture and tions, the government of Niger has been able to effectively Rural Development Department (ARD) and the Agriculture, manage various crises in the past. It has taken a number of Rural Development, and Irrigation Department in the Africa measures to improve the effectiveness and ef�ciency of the Region (AFTA1) conducted this Niger Agriculture Sector Risk existing crisis management systems. Some interventions by Review in March 2012 with the objective to: the government and donors have focused on long-term risk management, for example investment in irrigation, distribu-  Analyze the principal agricultural risks in Niger tion of drought resistant varieties, and improving information (production risks, market risks, and enabling systems. However, acute resource scarcity, chronic poverty environment risks) of households, and �nancial limitations of the government  Analyze the impact of agricultural risks have prevented a notable scaling up of those measures.  Prioritize the most signi�cant agricultural risks for Niger The situation of frequent crises, regular food availability and  Provide an illustrative list of risk management affordability problems, and low agricultural productivity pose measures (mitigation-transfer-coping) a serious development challenge for the government: how to  Prioritize risks management measures using decision feed its burgeoning population while contributing to income �lters growth in a sustainable manner. Addressing this question will require a better understanding and management of agricul- This report relies on several primary and secondary sources tural risks, which are often the underlying factors and triggers of information. It takes a holistic approach to agricultural risk behind crises. management and relies on long time-series historical data to arrive at an empirical and objective assessment of agricul- tural risks and their impacts on Niger. To gather primary data, 1.3 AGRICULTURE SECTOR RISK REVIEW a 2-week mission was conducted from March 5–15, 2012, The current government, shortly after it took power in April during which the study team travelled to the Tahoua, Dosso, 2011 after a 1-year military transition, announced a new strat- Maradi, Tillabery, and Niamey regions and interacted with egy aimed at dealing more effectively with food production, all the major agricultural sector stakeholders, including rel- 3N (les Nigériens Nourissent les Nigériens), and appointed a evant government organizations and technical and �nancial High Commissioner, who reports directly to the President’s partners, farmers, traders, and so forth. The mission team or- Of�ce. The speci�c objective of the 3N Commission is to ganized two wrap-up stakeholder consultations at the World strengthen national food production capacity and improve Bank of�ce and at the Ministry of Planning on March 15 to resilience in the face of crisis and natural catastrophe. share preliminary results and solicit feedback. Recognizing the need to explicitly and comprehensively ad- dress agricultural risks, the government of Niger, through the This report takes a holistic approach to agricultural risk and 3N High Commissioner, requested the World Bank to con- contributes to the debate by showing how best to prioritize duct an agricultural sector risk review. resources and use existing institutions more effectively. The assessment is designed to help decision makers understand Agricultural risk management is an old issue for Niger and it the risk exposure of the agriculture sector and develop ap- has been analyzed, although using different terminology with propriate risk management strategies that will strengthen focus on certain aspects, in detail by multiple stakeholders. the implementation of the government’s new 3N strategies, The World Bank’s Poverty Assessment (2011), Niger Food as well as other new and ongoing agriculture development Security and Safety Nets (2009), the Food and Agriculture programs. Organization’s (FAO) Review of Risk Management Tools and Policies in Niger’s Rural Sector (2008), and UNDP’s Analyse The rest of the document is organized as follows: Chapter Integree des Facteurs de Risque Au Niger (2007) have also 2 describes the agricultural system in Niger; Chapter 3 ana- covered some aspects of agricultural risks. This agriculture lyzes various agricultural risks; and, Chapter 4 estimates the sector risk review by the World Bank incorporates the exist- magnitude of losses. Chapter 5 prioritizes the risks and risk ing work and takes a step further in performing a systemic management measures based on a simple �ltering mecha- and holistic assessment of agricultural risks, their impacts, nism. The document closes with a brief conclusion and and a discussion of various potential measures for improved suggested next steps for improving agricultural risk manage- agricultural risk management in Niger. ment in Niger. TEC H N I C A L AS S I STANCE C H A P T E R 2 — AG R ICULT URAL SYST E M IN NIGE R 5 Chapter 2: AGRICULTURAL SYSTEM IN NIGER An overview of the agriculture sector is presented in this parameter for each zone, as there is limited variation in chapter, as the context for analysis and discussion of agri- soil type and topography at an aggregate level. Eighty per- culture sector risk. Sector characteristics most pertinent to cent of soils are sandy and there are few mountain areas. risk are thus given particular attention. The analysis of the A more detailed classi�cation is developed within each frequency and severity of risk is best done over an extended agro-climatic zone, based on local variations in soil type, time period; the review is based on trends for the period topography, and vegetation. Production conditions improve 1980-2010, wherever suitable data are available. from north to south, with most agricultural production in the four southern agro-climatic zones (table 2.1). The main characteristics of the �ve agro-climatic zones are listed in this table. 2.1 AGRO-CLIMATIC CONDITIONS The main agro-climatic zones of Niger are depicted in Available evidence shows that the isohyets that delimit these �gure 2.1 below. Rainfall isohyets are the main de�ning agro-climatic zones shifted southward by 100–200 kilometers FIGURE 2.1: Agro-Climatic Zones LEGENDE Zonage agro-écologique du Niger Limites des Régions Les zones climatiques + Villes principales Routes principales Routes bitumées Routes permanentes + + + + Zone Saharienne Zone Saharo-sahélienne Zone Sahélienne Zone Sahélo-soundanienne Zone Soundanienne + + + + + + + + + + N + W E + + 200 0 200 400 kilometers + S Source: Comité Interministérial de Pilotage de la Stratégie de Developpement Rural; Secrétariat Exécutif. September 2004. Le Zonage Agro-écologique du Niger. TEC H N I C A L AS S I STANCE 6 CH A PTER 2 — AGRICULTURA L SY STEM IN NIGER TABLE 2.1: Agro-Climatic Zones ZONE CLIMATE PRODUCTION PERCENT OF TERRITORY Sahara Arid: << 200 mm rainfall Livestock (nomadic and transhumant). Irrigated crop production in Oases 74% Saharo-Sahel Rainfall: 200–300 mm Traditional zone for pastoral livestock production. Encroaching crop production. 10% Sahel Rainfall: 300–400 mm Mixed livestock and millet-legume based cereal production. Some off-season cash 8% cropping in fossil river valleys. Sahelo-Soudain Rainfall: 400–600 mm Mixed livestock and cereal production (millet and sorghum). Higher potential for 7% off-season cash crops in fossil river valleys. Soudain Rainfall: >> 600 mm Mixed livestock and cereal production (millet and sorghum). Widespread practice 1% of irrigated, off-season cash crop production. Source: Comité Interministérial de Pilotage de la Stratégie de Developpement Rural; Secrétariat Exécutif. Septembre 2004. Le Zonage Agro-écologique du Niger. from 1950–90, following a decline in rainfall,4 with a pro- 85,000–100,000 hectares of agriculture land is irrigated, out longed period of below average rainfall from 1970–90. Recent of 270,000 hectares of potentially irrigable land.7 Although analysis of long-term rainfall trends shows that this trend has there is some scope to expand large-scale irrigation along now reversed, with average rainfall increasing again since the Niger’s rivers (from 15,000 to 30,000 hectares), there is 1990s.5 This suggests that the rainfall isohyets of 350–400 greater potential for irrigation from underground water millimeters, which delimit the zones where crop production resources. is viable, are shifting north again. As a result, the area suitable for crop production may have increased since the 1990s. Despite continued expansion of the area cultivated (�gure 2.2), per capita land use is declining. Farms are small The marked climatic constraints to agricultural production in (average 4.1 hectares) and getting smaller because Niger’s Niger are evident from this taxonomy. Less than 5% of agri- high population growth rate (nearly 3.5 percent per annum) cultural land is in the higher potential soudanian zone. Most exceeds the rate of area expansion. The consequent pres- agricultural land is in the sahel and sahelo-soudanian zones sure on agricultural land resources has risen in the last de- where rainfall is low and erratic. Crop production is even risk- cades and is now very high. ier in the pastoralist, saharo-sahelian zones, although both pastoralists and farmers are now trying to establish mixed Crop production thus takes place in a context of low and vari- crop and livestock farming systems there. A combination of able rainfall, and high and increasing pressure on cultivable higher rainfall (as noted above), reduced access to grazing, land. The potential for irrigation is limited, and its use is rela- and increased land pressure in traditional crop producing tively low. Production is further constrained by the predomi- areas may partly account for this transition from pastoralism nance of traditional management systems, with limited use to mixed farming in the saharo-sahel agro-climatic zone. of improved seeds, fertilizer, and mechanization. High land pressure also limits the scope for fallowing land as a means to preserve and restore soil fertility, with only 5 percent of 2.2 LAND AND WATER RESOURCES land in fallow in 2005 (Recensement General de L ’Agriculture et du Cheptel, Government of Niger [RGAC], op cit) . The land suited to crop production is limited, both in absolute terms and relative to population size. Approximately 6.5 mil- lion hectares are cultivated in the rainy season, with a further 2.3 CROP PRODUCTION SYSTEMS 73,000 hectares used for intensive horticulture production (“culture marâichère�) in the dry season.6 Approximately Faced with these constraints, most farmers opt for crop production systems based on the intercropping of millet, sor- ghum, and cowpea. All have low yield potential, but are well 4 ECOWAS, CILSS, OECD. 2006. The Ecologically Vulnerable Zones suited to low rainfall conditions. Intercropping further miti- of the Sahelian Countries. gates production risk by combining crops with differing sea- 5 UNEP. 2011. Livelihood Security: Climate Change, Migration and sonal growth patterns. In a year of early drought, the millet Conflict in the Sahel. 6 Ministere du Developpement Agricole/Ministere des Ressourc- es Animales. Juin 2007. Recensement General de l’Agriculture et du Cheptel 2005-2007. 7 FAO. 2007. Aquastat. AGR ICULTURA L SEC TOR RISK ASS ES S MENT IN NIGER C H A P T E R 2 — AG R ICULT URAL SYST E M IN NIGE R 7 FIGURE 2.2: Crop Area and Rural Population Trends 18,000,000 1.80 total crops (ha) rural population 16,000,000 1.60 total area/capita (ha/person) 14,000,000 1.40 12,000,000 1.20 10,000,000 1.00 8,000,000 0.80 6,000,000 0.60 4,000,000 0.40 2,000,000 0.20 19 0 19 1 19 2 19 3 19 4 19 5 19 6 19 7 19 8 89 19 0 19 1 19 2 19 3 19 4 95 19 6 19 7 98 20 9 20 0 01 20 2 20 3 04 20 5 20 6 20 7 08 20 9 10 11 8 8 8 8 8 8 8 8 8 9 9 9 9 9 9 9 9 0 0 0 0 0 0 0 19 19 19 19 20 20 20 20 Source: INS; Author’s calculations. or sorghum will suppress cowpea, for instance, while the during a severe drought as production falls for all three crops— density of cowpeas can be increased in the event that plant- as occurred in 2010 (�gures 2.4, 2.5, and 2.6). ing rains come late and cereals establish poorly. This form of diversi�cation also helps stabilize income and spreads the Dry-season production of high-value vegetable crops (culture use of scarce soil moisture and labor during planting and marâichère) provides a further means to raise and diversify early weeding. Given the right mix of crops and appropriate farm incomes. It provides sizeable income and nutritional planting densities, there is also ample evidence that inter- bene�ts to households engaged in it; however, from a mac- cropping of this nature will increase yields, suppress weeds, ro-perspective, its impact is modest due to the limited area and improve soil nitrogen levels.8 cultivated—equivalent to 1 percent of the area cultivated dur- ing the rainy season. Approximately 73,000 hectares were The advantages of this strategy are offset by the low yields cultivated in 2005 (RGAC, op cit), of which 19,000 hectares in of these three crops, which dominate production. Together, flood recession areas (“culture de décrue�) and 54,000 hect- they account for 85 to 90 percent of the total area planted, ares in irrigated lowland areas. Onions grown for export are limiting the scope for other, higher-yielding crops and for the main dry season crop (30,200 hectares in 2005), followed further diversi�cation (�gure 2.3). Although numerous other by gombo (7 ,300 hectares), eggplant (6,000 hectares), and crops are grown (groundnuts, rice, maize, fonio, sorrel, cabbage (3,500 hectares). voandzou, gombo, sesame, and souchet), the area planted to each is relatively small. Groundnuts account for 4 percent Lack of irrigation is the main constraint to increase dry- of the total area cropped and most of the others account for season crop production. Of the total area irrigated (approxi- less than 1 percent of total cropped area. Note that while mately 100,000 hectares), full-control irrigation schemes this intercropping mix provides useful protection in the event account for only 15,000 hectares. Farm productivity is low of lesser climatic shocks, it does not protect against more on these schemes. Producers are thus unable to �nance the severe shocks. In fact, covariate risk may be accentuated full costs of maintaining pumps, canals, dams, and so forth and the infrastructure is falling into disrepair. It should be noted, however, that the rice-growing areas bene�tted from 8 Lessons Learned from Long-Term Soil Fertility Management Experiments in Africa. Edited by Andre Bationo, Boaz Waswa, the organization of cooperatives, setting of management Job Kihara, Ivan Adolwa, Bernard Vanlauwe, Koala Saidou. 2012. fees, and maintenance of hydraulic systems and equipment. TEC H N I C A L AS S I STANCE 8 CH A PTER 2 — AGRICULTURA L SY STEM IN NIGER FIGURE 2.3: Rainy Season Crop Area Trends 8,000 7,000 millet 6,000 hectares (in thousands) 5,000 cowpea 4,000 3,000 sorghum 2,000 g'nuts oseille 1,000 sesame voandzou 0 other cereals 2000 gombo 2001 2002 souchet 2003 2004 2005 2006 2007 year 2008 2009 2010 2011 Source: INS; Author’s calculations. FIGURE 2.4: Millet Production 8,000,000 production (mt) 600 area (ha) 7,000,000 yield (kg/ha) 500 6,000,000 400 5,000,000 mt and ha kg/ha 4,000,000 300 3,000,000 200 2,000,000 100 1,000,000 0 0 80 82 84 86 88 90 92 94 6 8 0 2 04 06 08 10 9 9 0 0 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 20 year Source: Author’s calculations. AGR ICULTURA L SEC TOR RISK ASS ES S MENT IN NIGER C H A P T E R 2 — AG R ICULT URAL SYST E M IN NIGE R 9 FIGURE 2.5: Sorghum Production 3,500,000 production (mt) 500 area (ha) yield (kg/ha) 450 3,000,000 400 2,500,000 350 300 2,000,000 mt and ha kg/ha 250 1,500,000 200 1,000,000 150 100 500,000 50 0 0 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 00 01 02 03 04 05 06 07 08 09 10 11 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20 20 20 20 20 year Source: Author’s calculations. FIGURE 2.6: Cowpea Production 6,000,000 400 production (mt) area (ha) yield (kg/ha) 350 5,000,000 300 4,000,000 250 mt and ha kg/ha 3,000,000 200 150 2,000,000 100 1,000,000 50 0 0 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 00 01 02 03 04 05 06 07 08 09 10 11 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20 20 20 20 20 year Source: Author’s calculations. TEC H N I C A L AS S I STANCE 10 CH A PTER 2 — AGRICULTURA L SY STEM IN NIGER As a result, cultivation intensity, as well as yield, have im- TABLE 2.2: Percent Change in Food Crop Production proved signi�cantly. A further 60,000–70,000 hectares is PRODUCTION AREA YIELD partially irrigated from groundwater reserves, usually by 1980–2011 small-scale, informal farmer groups. Although production Millet 102.6% 129.5% –11.7% levels are similar to those observed on full-control irriga- Sorghum 117.7% 292.4% –44.5% tion schemes, these partial water control schemes appear Cowpea 483.9% 320.3% 38.9% more viable. Capital costs are much lower and the physi- cal infrastructure (wells, pumps, polythene pipes) is much Groundnut 75.0% 264.4% –52.0% simpler and less expensive to maintain. Flood recession Maize –36.0% –45.4% 17.2% cropping accounts for a further 18,000–20,000 hectares. All Rice –59.2% –56.7% –5.7% of these irrigation systems experience water shortages dur- 2000–2011 ing droughts. These systems are quite diverse in terms of Millet 64.5% 33.1% 23.6% cropping, water control system, technology, management, Sorghum 107.7% 38.2% 50.4% market opportunity, social organization, source of �nancing, Cowpea 497.2% 20.8% 394.5% and land tenure security, which makes it dif�cult to make an authoritative statement about the relative viability of these Groundnut 249.6% 91.7% 82.3% systems. Recent projects supported by the World Bank have Maize 68.4% 36.1% 23.8% demonstrated high pro�tability of individual, small-scale irri- Rice 5.2% 112.2% –50.4% gation projects. This often requires handing over of manage- Source: INS; Author’s calculations. ment responsibilities to producers’ collectives and adequate supervision for maintenance and management of irrigation projects. scope to raise crop yields, given the low current rates of fer- tilizer use and the low use of improved seed varieties, but These deep-seated limitations to both rainy and dry season strong supply and demand side constraints limit the use of crop production, and their attendant vulnerability to drought, these inputs. Total fertilizer use is estimated at 20,000 tons highlight the importance of livestock to rural livelihoods. Niger per year9 for 6.5 million hectares, an average of only 3 ki- is fortunate in having both favorable conditions for livestock lograms per hectare. Even for more intensive, off-season production and a huge, readily accessible export market in crop production, only 25 percent of producers reported neighboring Nigeria. Mixed crop and livestock production fertilizer use in the 2005 Census of Agriculture (RGAC, op systems are thus a critical strategy for raising and diversify- cit). Similarly, improved seed varieties accounted for only 6 ing farm income. percent of the total seed used for millet, sorghum, cowpeas, groundnuts, and sesame10 in 2010. This combination of low input use, low and declining soil fertility, and dif�cult climatic 2.4 PRODUCTION TRENDS conditions means that nationwide yield gains are likely to be Production has increased steadily for all of the major crops slow, and that production increases may be slower and more (�gures 2.4, 2.5, and 2.6). Millet and sorghum production variable in the future. doubled from 1980–2011 and cowpea production increased Production variability, as measured by coef�cients of varia- by almost �ve times (table 2.2). Area expansion largely tion, con�rms the higher drought tolerance of millet and sor- drove this production increase. Yields fell initially and then ghum, particularly the former (table 2.3). The higher produc- rose after 2000, consistent with underlying rainfall trends. tion variability of cowpea production is attributed to its higher The pronounced increase in cowpea production, which was yield variability. Production variability increases yet again for driven by both area and yield increases, reflects its increasing groundnut, rice, and maize, which is consistent with their pro�tability for export. Maize and rice production fell overall limited importance to overall crop production. from 1980–2011, although maize has recovered somewhat since 2000. The small contribution of maize and rice to total crop production has declined and they now account for less than 1 percent of total cereal output. 9 Reseau National des Chambres d’Agriculture du Niger. Note d’Information/Intrants No. 13. Decembre 2010. As the scope for further area expansion is limited, future 10 Ministère de l’Agriculture et de L’Elevage. Annuaire sur la Dis- production growth will depend on yield increases. There is ponibilité en Semences Améliorées 2010–2011. Niger. AGR ICULTURA L SEC TOR RISK ASS ES S MENT IN NIGER C H A P T E R 2 — AG R ICULT URAL SYST E M IN NIGE R 11 TABLE 2.3: Coef�cients of Variation: Production, Area, below for joint production of millet, sorghum, maize, and rice and Yields (fonio was excluded due to lack of data and its small contri- PRODUCTION AREA YIELD bution to production—approximately 5,000 mt in 2011). The Millet 0.19a 0.06a 0.18 analysis is based on production and consumption only, and does not reflect imports, food aid, or changes in stocks. Sorghum 0.32a 0.13a 0.33 a a Cowpea 0.53 0.15 0.46 The results of the analysis show the following major trends: a a Groundnut 0.65 0.43 0.43  Driven by the expansion of millet and sorghum area, Rice 0.68a 0.57a 0.39 growth in total cereal production for food has kept Maize 0.84 0.73 0.78 pace with population growth for the period 1980–2011. Source: Authors calculations. This trend was also shown for other Sahelian coun- a Adjusted for trend using the Cuddy Delle-Valle Index. tries in a recent study by Organization for Economic Co-operation and Development (OECD: Atlas on 2.5 CEREAL CROPS: FOOD SUPPLY AND DEMAND Regional Integration in West Africa, 2009). The comparison of domestic cereal production for food with  Cereal food production de�cits occurred frequently domestic demand for cereals is based on the following as- during the lower rainfall years of 1984–2000, including sumptions: (1) non-food use of cereals for seed, livestock a number of years when cereal food production fell feed, and exports estimated at 15 percent of total cereal below the minimum consumption requirement. production; (2) average domestic consumption for food of  Cereal food production de�cits have been less 225 kilograms per capita per year, based on the most recent frequent and less severe since 2000, in response to Global Information and Early Warning System (GIEWS) analy- higher rainfall, with no cases of aggregate production sis by FAO; and (3) a minimum consumption requirement of falling below the minimum consumption requirement. 165 kilograms per capita per year, equivalent to the minimum  The period since 2000 has also been characterized by dietary energy requirement of 1680 calories per capita per frequent surpluses, including record surpluses in 2008 day (FAO). The resultant trends are reported in �gure 2.7 and 2010, which has facilitated the rebuilding of stocks. FIGURE 2.7: Cereal Food Production vs. Consumption 5,000,000 cereal food supply 4,500,000 average consumption minimum consumption 4,000,000 3,500,000 3,000,000 tons 2,500,000 2,000,000 1,500,000 1,000,000 500,000 – 19 0 19 1 19 2 19 3 84 19 5 19 6 19 7 19 8 89 19 0 19 1 19 2 19 3 19 4 19 5 19 6 19 7 19 8 20 9 20 0 20 1 20 2 20 3 04 20 5 20 6 07 20 8 20 9 10 11 8 8 8 8 8 8 8 8 9 9 9 9 9 9 9 9 9 9 0 0 0 0 0 0 0 0 19 19 19 20 20 20 year Sources: INS, GIEWS, FAO, Author’s calculations. TEC H N I C A L AS S I STANCE 12 CH A PTER 2 — AGRICULTURA L SY STEM IN NIGER 2.6 CASH CROPS Of the other rainy season cash crops, souchet is grown as a Cash cropping is based on rainy season production of cow- mono-crop, while sesame is grown on small garden plots— pea, sesame, and souchet and irrigated, off-season produc- often by women. Data for these two crops are limited to tion of vegetable crops. Onions and cowpea are the main the period 2000–11. Production of both crops has increased export crops with estimated revenues of $88 million and $68 signi�cantly since 2000, although the increase observed for million respectively in 2006 (World Bank, 2010). Souchet ex- souchet is from a very small base. Both crops exhibit marked ports amount to a further $10 million and sesame $3 million. interannual variability in production, area, and yield (table 2.4). All of these crops are sold in regional markets of Nigeria, Niger has established a strong export market for onions, Benin, and Cote d’Ivoire. particularly in the large urban markets of Nigeria, Ghana, and Niger is the second largest producer of cowpeas in West Cote d’Ivoire. Onions have thus become an important source Africa, after Nigeria. Cowpeas are valued both as an export of off-season revenue for small-scale farmers, helping them crop (50 to 75 percent of production is sold to Nigeria), and smooth and increase farm incomes. Most production is on as a source of animal feed for domestic use (cowpea hay). small plots, irrigated with simple pump irrigation systems. Domestic bean consumption is relatively low (6 kilograms per Available data suggest that the area planted has increased capita). Production has increased considerably since 2000, signi�cantly in response to the introduction of onions to ar- largely in response to increased yield (table 2.4). Intercropping eas around Agadez and the continued growth of small-scale with millet and/or sorghum is the main form of production, irrigation along the fossil-river valleys in south-central Niger. reducing production risk but also lowering yield. Interannual Production has not changed much however, due to declining variation in cowpea production is quite high, nevertheless yields. Marketing issues may represent the main constraint (table 2.4), despite its suitability for low rainfall conditions and to increased production. Growers are highly vulnerable to light, lower fertility soils. Farmers seeking to expand cowpea periodic market gluts due to their limited bargaining power, production will eventually face a choice of whether to switch limited knowledge of trends and conditions in major export to growing it as a single crop, with higher yields but much markets, and lack of storage and processing facilities. higher attendant risks, or to continue intercropping. 2.7 LIVESTOCK TABLE 2.4: Characteristics of Major Cash Crops Sahel countries have a long tradition of pastoralism, rely- PERCENT COEFFICIENT ing on their comparative advantage in extensive livestock AVERAGE CHANGE OF VARIATION management systems to supply the larger coastal markets Niebe (2000–11) of Nigeria, Cote d’Ivoire, and Ghana. Niger, with 10.5 million Area 4,167,512 20.8% 0.15a tropical livestock units (TLUs),11 has the largest herd popula- Production 852,538 497.2% 0.53a tion in the Sahel region. The contribution of livestock to gross agriculture domestic product is 40 percent but it used to be Yield 193 394.5% 0.46 much higher in the past. But the proportion of the national Sesame (2000–11) budget invested into this subsector has fallen.12 Area (ha) 89,786 94.1% 0.40a Production (ton) 33,940 283.9% 0.63a Over the past 50 years, Niger’s national herd is estimated to Yield (kg/ha) 380 97.8% 0.49 have grown at an average annual rate of 2.47 percent if one considers the total number of head, or 2.33 percent p.a. in Souchet (2000–11) TLU terms13; the small difference being due to the fact that Area (ha) 6,489 448.0% 0.51a herd composition overall has not changed drastically (mod- Production (ton) 18,628 1921.8% 0.54a Yield (kg/ha) 3,011 268.9% 0.46 11 The tropical livestock unit concept brings together metabolism Onions (2000–10) and feed requirements to various animal species; one TLU is Area (ha) 14,005 84.7% 0.11a generally equivalent to 1 camel, 0.7 cattle, and 0.1 small rumi- nant. Production (ton) 345,286 9.5% 0.10 12 In the early postindependence period, the share of government Yield (ton/ha) 26.2 −40.7% 0.09 a budget going to livestock was about 4 percent compared to about 0.5 percent today. Sources: INS, FAOSTAT, Author’s calculations. a Adjusted for trend using the Cuddy Delle-Valle Index. 13 Livestock statistics in Sahel countries are notoriously unreliable and must be taken with caution. AGR ICULTURA L SEC TOR RISK ASS ES S MENT IN NIGER C H A P T E R 2 — AG R ICULT URAL SYST E M IN NIGE R 13 est proportional reduction in cattle, small increase in camels, whose annual quantity, quality, and distribution over space is some shift from goats to sheep within small ruminants). highly variable and uncertain. Bigger changes have involved the redistribution of livestock over space, within production systems, and ownership. Repeated weather and other shocks have affected herd sur- vival and productivity as well as access by pastoral house- The livestock system in Niger could be classi�ed into the pas- holds to food, on the basis of the livestock, and grain, in toral zone, where transhumance and nomadism dominate; terms of trade. As a result, all herding families have tried to the agro-pastoral zone; and the agricultural zone. In each develop complementary income-generating activities, such zone, animal husbandry systems may be further classi�ed as trade with Algeria or Libya in the north and seasonal or by species composition, type of management, and degree of long migration. Other emerging trends in this zone include a association with agriculture, to the extent that these factors redistribution of livestock ownership within the transhumant- are relevant to risk exposure and management. nomadic pastoral zone,15 a reduction in family herd sizes,16 and a move toward the south by many herding families.17 Two researchers14 analyzing changes in Niger’s livestock sec- tor over the past 30 years have proposed a typology of live- stock ownership and management including seven distinct 2.7.2 Agro-Pastoral Zone groups based on two main criteria: (1) relative vulnerability This falls within an ecological area with an annual rainfall to climatic risk and (2) extent to which they are connected of 300–400 millimeters, spanning the country from east to to, and experienced in, dealing with livestock markets. This west, and all regions except for Dosso and the Niamey UCN, typology reflects the widening gap between the livestock and accounts for two-thirds of the cattle herd in Niger. The owners and trading groups who have managed to develop main feature here has been the expansion of agriculture, both relatively effective and resilient risk management strategies by people escaping the very high population densities (and (civil servants and other wage earners, well-connected live- shrinking farm sizes) of the south, and by herding families stock owners and traders, and entrepreneurs) on the one becoming increasingly involved in agriculture, after having hand, and groups who have remained highly vulnerable to lost much livestock during recent crises. In the agro-pastoral weather and other risks, on the other. The latter groups in- zone, about 60 percent of heads of households reported live- clude pastoralists having suffered major losses and without stock herding as their main activity, with 26 percent stating signi�cant market connection, those in the process of herd that they practiced a combination of animal husbandry and reconstitution, and agro-pastoralists having partly reduced farming.18 their dependence on livestock and grain markets through The combination of low and variable rainfall (over 20 per- limited agricultural production: cent interannual variability) puts crops, pastures, and water  Entrepreneurs operating in relatively secure systems  Traders and investors seeking �nancing 15 Following a crisis, the herding families that had implemented the most successful management strategies have the means to  Herders and small traders with market connections acquire some or all of the remaining livestock belonging to less  Wage earners with livestock assets successful and poorer herding groups. In some cases, the latter merely become herdsmen for large owners until they accumu-  Herders’ families with agricultural production late the means to start a small herd of their own again.  Households rebuilding herds 16 According to recent study by Salla, Atte, Oumarou, Gouverne- ment du Niger, 2011, the average family herd size has reduced  Poor herding households, without market connections to 15.5 TLUs compared to an estimated 23.4 TLUs required to sustain the average herding household. 2.7.1 Transhumant-Nomadic Pastoral Zone 17 This move (likely permanent) by herding families to agro-pastoral This largely falls within the Agadez Region, and covers or agricultural areas mostly took place during the mid-1980s and the 2004–05 and 2009–10 periods. After moving southward in the northern parts of most other regions (Tillaberi, Tahoua, an attempt to save surviving animals, many families decided Maradi, Zinder, and Diffa). The composition of herds reflects not to return to their traditional northern grazing areas. Some environmental conditions and the combination of camels and lost everything and ended up as urban or peri-urban refugees; most settled in agro-pastoral areas and turned to agriculture and small ruminants in this region accounted for 85 percent of all other activities to complement their limited income from remain- TLUs in 2009. Transhumance and nomadism represent the ing herds. Here too, a transfer of ownership took place, from most effective strategies to use forage and water resources pastoral to salaried or agricultural families, which contributed in a small way to the signi�cant increase in livestock observed in agricultural areas. 14 A. Marty, B. Bonnet, 2006. 18 Gouvernement du Niger, 2011. TEC H N I C A L AS S I STANCE 14 CH A PTER 2 — AGRICULTURA L SY STEM IN NIGER resources at high risk. Transhumance remains a major fea- include : (1) shrinking farm size and marginalization of many ture of animal husbandry in the agro-pastoral zone, including smallholders,20 (2) improved resource conservation and use long-term growing out in parts of southeastern Mali, north- due to large-scale Farmer Managed Natural Regeneration ern Benin, or Nigeria. Agriculture is very uncertain, especially (FMNR) of woodlands,21 and (3) a considerable increase in for poorer herding families now settled in this zone, with the livestock population (mostly cattle and small ruminants). majority (77 percent) being able to grow enough food for a period ranging from 1 to 3 months of their annual grain re- One of the most striking features of the recent agricultural quirements.19 In addition, the combination of high livestock census of Niger is that about 60 percent of livestock in density, extensive transhumance, and expansion of farmed the country are now considered to be sedentary. Another areas have raised tensions and conflicts between groups us- feature of livestock in the agricultural area, aside from the ing open-access resources (among herders, between herd- coexistence of well and poorly integrated systems, is that ers and farmers, between agro-pastoralists, etc.). ownership is reportedly skewed: some 90 percent of cattle and 75 percent of small ruminants are reportedly owned by the richer half of all farming households.22 Relatively better- 2.7.3 Livestock in the Agricultural Zone off farmers have invested in livestock, during years of good This southern belt, wider in the west than in the east, production or when livestock prices have been particularly spans the entire country and receives on average 400–600 low, as a way to diversify their investment and savings port- millimeters of rainfall per year (up to 800 millimeters in folios (women have followed the same strategy, although on southern Dosso areas). Some emerging trends in this zone a smaller scale, and with small rather than large ruminants). 20 High population growth rate, limited arable land, and consolida- tion by resource-endowed households to grow high-value export crops (souchet, tobacco, etc.) have put tremendous pressure on available land resources, resulting in fragmentation of plots, re- duced land availability, and marginalization of small holders. 21 There is remarkable and spontaneous uptake by farmers of FMNR to develop agro-forestry systems in high population density areas where most land is cultivated and where natural vegetation had almost disappeared. The scale of this on-farm re- greening is at least 5 million hectares, making it the largest envi- ronmental transformation in the Sahel, and possibly, in Africa. 19 Gouvernement du Niger, 2011. 22 FEWSNet, 2011. AGR ICULTURA L SEC TOR RISK ASS ES S MENT IN NIGER C H A P T E R 3 — AG R ICULT URAL SE CTOR RISKS 15 Chapter 3: AGRICULTURAL SECTOR RISKS The main sources of risk are reviewed in this chapter: produc- period of analysis (�gure 3.1). Detailed analysis conducted by tion risk, market risk, and a general set of risks associated AGRHYMET for the 1980–2010 time period (see annex 5) in- with the enabling environment for agriculture. The incidence dicates that in the sahelian and Sahelo Saharan zones, more and implications of multiple or successive shocks are also than half of the rainfall stations analyses recorded a signi�- considered. An analysis of rainfall characteristics, at both na- cant increase in rainfall. More than 33 percent of stations had tional and regional levels, deepens the context for analysis. a signi�cant increase in the number of rain days, a signi�cant decrease in duration of the longest dry spell and cessation dates of rainfall. No such trend was found in the soudano sahelian zone. AGRHYMET concludes that, apart from the 3.1 RAINFALL TRENDS 1980–2009 severe droughts of 2004 and 2011, Niger has experience Trends in average annual rainfall indicate years of lower rainfall conditions much better over the past two decades rainfall and con�rm that rainfall has increased during the compared to the dry decades of the 1970s and 1980s. FIGURE 3.1: Average Annual Rainfall 600 500 400 mm 300 200 100 0 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 00 01 02 03 04 05 06 07 08 09 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20 20 20 years Source : INS. Annuaire Statistique des Cinquantes Ans d’Independence du Niger. Edition Speciale. Novembre 2010. TEC H N I C A L AS S I STANCE 16 CH A PTER 3 — AGRICULTURA L SEC TOR R IS K S FIGURE 3.2: Major Shocks to Crop and Livestock Production 200 cereals crops 180 livestock 160 production index (2004–2006 = 100) 140 120 100 80 drought floods 60 drought locusts 40 drought floods drought 20 drought pests locusts locusts locusts drought locusts 0 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 00 01 02 03 04 05 06 07 08 09 10 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20 20 20 20 year Sources: FAOSTAT; Author’s calculations. A more nuanced indicator of the range and extent of adverse 3.2.1 Drought rainfall events are obtained from analysis of standardized Drought is the most important agricultural risk with high anomalies of rainfall (see annex 2) for the same period by probability and severity affecting both crop and livestock region (table A2.5). This analysis con�rms the range and production. An agricultural drought occurs when a soil mois- extent of the catastrophic droughts in 1984 and 1987 . Less ture de�cit signi�cantly reduces crop yields and output. severe adverse rainfall events are observed in 1990, 1993, It can occur in response to low overall annual rainfall or to 1997 , 2000, 2004, and 2009. abnormalities in the timing and distribution of annual rainfall. Inadequate rainfall at key periods during the crop production cycle (seeding, flowering, and grain �lling) affects crop yields, 3.2 PRODUCTION RISKS even when overall rainfall is comparable to long-term norms. Drought, locusts, livestock diseases, crop pests and diseas- During these periods, a soil moisture de�cit for a period as es, floods, windstorms, and bush�res are the main sources short as 10 days can have a major impact on crop yields. of production risk. Farmers also complain of the risks to crop AGRHYMET’s principal component analysis reveals that long production from livestock herds, although the incidence and dry spells (number of consecutive days without rainfall) and severity of these events is dif�cult to determine. The impact late onset of rains are the two biggest factors responsible for of these events on crop production for the period 1980–2011 yield losses and crop failure in Niger. is indicated in �gure 3.2 below, using indices of agricultural Drought is typically de�ned relative to some long-term aver- production.23 The index of livestock production is reported for age balance between precipitation and evapotranspiration, the sake of completion, as the data on which it is based are which is considered normal for a particular location at a not considered realistic. particular time of year. Yield loss is also assessed relative to some normal long-term average yield for a particular crop in a particular location. 23 Shows relative level of aggregate volume of production for each year relative to a base period of 2004–06. Calculated as the sum Analysis of rainfall data (annex 2) indicated that between of price-weighted quantities of main commodities using average international prices. 1980–2009, there were nine years, namely, 1984, 1987 , AGR ICULTURA L SEC TOR RISK ASS ES S MENT IN NIGER C H A P T E R 3 — AG R ICULT URAL SE CTOR RISKS 17 TABLE 3.1: Summary of Drought Events NUMBER OF NUMBER OF FODDER BALANCE CEREAL BALANCE DEPARTMENTS SEVERE DEPARTMENTS SHEET (SURPLUS OR SHEET (SURPLUS OR SEVERE FOOD YEAR DROUGHT CATASTROPHIC DROUGHT DEFICIT %) DEFICIT %) CRISIS 1984 12 8 NA NA (1985) 1987 11 13 NA NA 1990 8 12 NA NA 1993 5 12 NA NA 1995 6 4 NA NA 1997 8 9 NA NA (1998) 2000 8 9 −12% −28% 2004 7 7 −25% −21% (2005) 2009 10 2 −70% −12% (2010) 2011* −41% −18% (2012) Source: Author’s calculations; World Food Programme (WFP). Situation Alimentaire Et Nutritionnelle Au Niger. February 2012. * While 2011 was a drought year, relevant meteorological information was not available when this analysis was conducted. 1990, 1993, 1995, 1997 , 2000, 2004, and 2009, wherein 10 or Locusts eat all vegetative matter in their path, starting more departments suffered severe or catastrophic drought with grazing lands and then moving to agricultural crops. (table 3.1). However, not all of them had a major impact on But whereas the impact on crop production is direct and agricultural sector. Only seven of these rainfall de�cit years, immediate, the impact on livestock production is indirect namely 1984, 1997 , 1993, 1997 , 2000, 2004, and 2009, were and delayed due to the loss of grazing. The level of dam- strongly associated with crop failure (�gure 3.2). Furthermore, age is highly localized and proportional to the size of the these droughts did not always lead to a severe food crisis24 swarms, but is devastating during major attacks. While the and major food shortages resulted only in 1984–85, 1997–98, incidence of locust attacks is adequately recorded (table 2004–05, and 2009–10. Although the 2011 drought was not 3.2), it is more dif�cult to quantify losses to crops and analyzed systematically, it led to large losses in Niger and livestock production, especially when it occurs in combi- resulted in severe food crisis in 2012. nation with drought. FAO Emergency Prevention System/ (FAO EMPRES) estimated that the crop losses incurred during 2004–005 were due two-thirds to drought and one- 3.2.2 Locusts third to locusts; but this applies only to areas most heavily Locusts (especially desert locusts, Schistocerca gregaria) are affected. a constant threat in the Sahel. They are dif�cult to detect and eradicate as their breeding grounds are located in remote 3.2.3 Livestock Diseases areas, and seasonal desert winds facilitate rapid, widespread dispersal of locust swarms. For the period of analysis, the Of�cial statistics for the period 1995–2010 indicate that, two worst infestations occurred in 1987–88 and 2004–05. pasteurellose, anthrax, and peste des petits ruminants are Most reported infestations start in the northern desert re- the most widely reported diseases among ruminants, and gions of Niger where crop and livestock production are low Newcastle disease is the most common disease risk for (see table 3.2 below), but they often move southward to poultry (table 3.3). Although the statistics on reported cases cultivated areas. per year and deaths per year are improbably low (less than 0.1 percent of total animal numbers), the reported frequency (about 0.35) of disease outbreaks is probably indicative of actual trends. Of the diseases reported above, vaccines exist for all except pasteurellose (although it is dif�cult to vacci- 24 Aker, Jenny. “Rainfall Shocks, Markets and Food Crises: Evi- nate for foot and mouth as different strains require different dence from the Sahel.� Center for Global Development. Working Paper no. 157. December 2008. vaccines). Hence, for most major livestock diseases, risk can TEC H N I C A L AS S I STANCE 18 CH A PTER 3 — AGRICULTURA L SEC TOR R IS K S TABLE 3.2: Summary of Locust Infestations YEAR CLASSIFICATION AND AREAS OF INFESTATION AREA SPRAYED BY THE GOVERNMENT 1978 (Oct–Nov) Minor infestation in Akarbai 6,000–7,000 ha 1980 (Sept–Dec) Widespread infestation in western Air and Tamesna 100,000–120,000 ha 1986 (Oct–Dec) Widespread infestation in northern Tamesna and Air 60,000–70,000 ha 1987 (Nov–Dec) Moderate infestation in Tamesna and Air 25,000–30,000 ha 1988 (Jun–Dec) Locust plague catastrophe 700,000–800,000 ha 1989 (Jan–May) Localized spread in Tamesna 4,500–5,000 ha 1995 (Jan–June) Minor infestation in various areas in Tamesna region 9,000–10,000 ha 2000 (Feb) Localized infestation in the area of Wadi Ta�det 700 ha 2003 (Nov) Localized infestation in Tamesna between Agadez and In Abangharit 300–400 ha 2004 (Jan–Dec) Locust upsurge (large-scale infestation) in Tamesna and Air 300,000–400,000 ha 2005 (June) Localized infestation 263 ha 2006 (Nov–Dec) Minor infestations in Abangharit, Tassara, and central Tamesna, primarily west of 4,000–5,000 ha In Abangharit in the Azaouak Valley Source: Compiled from FAO’s Desert Locust Situation Update and Desert Locust Information System. TABLE 3.3: Incidence of Livestock Disease 1995–2010 For livestock conditions such as internal and external AVERAGE parasites, mastitis, and foot ailments, the risk of infection NUMBER AVERAGE AVERAGE NUMBER increases when animal husbandry and feeding are poor— OF YEARS NUMBER OF NUMBER OF especially when grazing and forage are limited. Livestock OBSERVED OUTBREAKS/ OF CASES/ DEATHS/ 1995–2010 YEAR YEAR YEAR mortality is thus particularly high during periods of drought. The incidence of these ailments is also likely to increase as Sheep and 7 92 710 195 goat pox livestock systems intensify, especially in peri-urban areas.25 Pasteurellose 4 157 586 123 Adequate feeding, good animal husbandry, and ready access to veterinary services are the best way to reduce these risks, Anthrax 4 28 194 147 whether there is a drought or not. Unfortunately, farmer Peste des 6 39 155 118 petits knowledge of animal husbandry is limited, as is the national ruminants veterinary service’s ability to deliver animal health care. Foot and 5 58 151 23 mouth 3.2.4 Floods Blackleg 5 28 56 26 While it has generally been favorable to crop production, the Contagious 6 6 48 28 bovine pleuro- increase in average annual rainfall since the mid-1980s has pneumonia resulted in an increased incidence of flooding. Most flooding Newcastle 5 5 518 444 occurs in the rainy season months of July through September, disease when the main crops are under cultivation and livestock are Source: The World Organisation for Animal Health (OIE) Database. con�ned to households—particularly small ruminants. It is usually the result of heavy rainfall during a short time period, which results in flash flooding and localized flooding along the banks of the Niger River, so there is little time to prepare. be dramatically reduced with easily administered, low-cost Damage and losses tend to be localized, but are devastating vaccinations. A recent study on the impact of the 2009–10 where they occur. Houses and buildings are damaged or de- crisis on the livestock subsector showed that vaccination stroyed, �elds are inundated, and high numbers of livestock rates in surveyed areas had increased from about 30 percent in the mid-2000s to over 40 percent by the end of the de- cade. Much more can be done. 25 Ly, Fall, and Okike, 2010. AGR ICULTURA L SEC TOR RISK ASS ES S MENT IN NIGER C H A P T E R 3 — AG R ICULT URAL SE CTOR RISKS 19 TABLE 3.4: Recorded Losses Due to Selected Major Floods DISPLACED HOUSES DAMAGED YEAR REGIONS AFFECTED PEOPLE OR DESTROYED CROP LOSSES LIVESTOCK LOSSES 1988 Southern, eastern and western parts of 307,000 N/A 7,500 ha N/A Niger and Niger River 1994 Southern and Western Niger 20,000 N/A N/A N/A 1998 Niamey, Mokko 41,590 230 9,196 ha 1254 large ruminants, 6,544 small ruminants 1999 Northern Niger, Agadez region 21,679 1,996 2,736 ha 26 large ruminants, 215 small ruminants 2006 Maradi, Dosso, Tillaberi, and Agadez 7,965 387 437 ha and 7431 small ruminants, 896 cattle, 86 �elds 208 donkeys, 59 camels 2009 Northern Niger, Agadez region 99,900 6,261 424 ha 23,585 animals (mostly small ruminants) 2010 Niger River and Southern Niger 252,900 10,708 6,872 �elds 115,114 animals (mostly small ruminants) Sources: Dartmouth Flood Observatory Dartmouth University; Cellule de Coordination du Systeme d’Alert Precoce. are washed away. The number of displaced people can also although millet is naturally tolerant to striga. Farmers invari- be high, especially where these floods affect urban areas. ably complain about the losses caused by plant pest and diseases, but also acknowledge that they seldom use the Serious floods have been recorded in 9 of the last 30 years chemical treatments available for prevention and treatment, in Niger, a frequency of 0.30 (table 3.4). It is more dif�cult to claiming that they are too expensive. Overall losses from get accurate information on the agricultural losses caused by majority of these pest and diseases, barring the exception flooding. Detailed records were not kept until quite recently of grasshoppers, grain-eating birds, and millet borer, are fre- and information is not recorded systematically where it is quent but relatively low. available. But available data suggests that the main direct �- nancial cost of these floods is the result of damaged houses and buildings as well as livestock losses. Aggregate crop 3.2.6 Bush�res and Windstorms losses are of a lesser magnitude, as farm size is small. The Fire is used to clear land for cultivation to improve grazing indirect costs of flooding are likely to be much higher than and facilitate hunting. Burn-offs also help control pests and direct costs, however, as people lose their livelihoods when diseases such as grasshoppers, locusts, ticks, anthrax, and they are displaced, in addition to the direct losses of crops livestock parasites. Burning poses a risk when �res get out of and livestock. control, however. This risk is localized and highest in pastoral areas where rainfall is low and vegetation is dry,26 and it leads to loss of pastures. Nonetheless, frequency of bush�res and 3.2.5 Crop Pests and Diseases the direct adverse impact on Niger’s agricultural sector is low. The main crop pests and diseases are as follows: Windstorms pose a risk for agriculture when they occur at  Insects: grasshoppers, leaf-hoppers, millet stem borer the beginning of the cultivation season, damaging young and head borer plants and forcing farmers to replant their crops (Programme  Diseases: downy mildew, smut, ergot d’Action National pour l’Adaptation [PANA], op cit). Direct  Weeds: striga damage to crop or livestock is minimal, although sandstorms  Pests: grain birds, rodents and dust storms cause soil loss and remove organic matter and the nutrient-rich lightest particles, thereby reducing agri- There is no information on the level of damage and losses cultural productivity. Nonetheless, the overall impact of sand- that these pests and diseases cause annually in Niger. storms and dust storms on the agricultural sector is limited. Grasshoppers, leafhoppers, and grain-eating birds are known to cause high losses on an intermittent basis. There are an- nual government programs to spray for grasshoppers, but no programs to control bird damage. Striga and fungal dis- 26 Programme d’Action National pour l’Adaptation. Identi�cation et eases are perennial among most cereal and cowpea crops, Evaluation des Phenomenes Extremes. June 2005. TEC H N I C A L AS S I STANCE 20 CH A PTER 3 — AGRICULTURA L SEC TOR R IS K S FIGURE 3.3: Real Cereal and Cowpea Prices 400 millet sorghum cowpea 350 imp rice local rice 300 250 FCFA/kg 200 150 100 50 0 19 0 19 1 19 2 19 3 19 4 19 5 19 6 19 7 19 8 19 9 19 0 19 1 19 2 19 3 19 4 19 5 96 19 7 98 20 9 20 0 01 20 2 20 3 20 4 20 5 20 6 07 20 8 09 20 0 11 8 8 8 8 8 8 8 8 8 8 9 9 9 9 9 9 9 9 0 0 0 0 0 0 0 1 19 19 19 20 20 20 year Source: SIMA. 1996 = 100. 3.3 MARKET RISKS that the global food price crisis of 2007–08 had limited im- pact on prices for staple foods in Niger. Long-term trends in agricultural commodity prices are re- viewed initially to assess interannual price volatility, followed Local and imported rice prices exhibit differing trends (local by an analysis of seasonal price movements. The analysis rice prices are only available from 1990 onward). Imported is based on consumer prices for millet, sorghum, local rice, rice prices follow world markets, falling steadily until 2008 imported rice, and cowpea as reported in local markets and when they rose by 13 percent in response to the global food producer prices for livestock. Ratios of millet to livestock price crisis. In contrast, local rice prices increased steadily prices are also analyzed to assess the volatility of changes from 1990 until 2008–09, and then decreased. Overall, the in purchasing power. Real prices are used as the basis for gap between imported and local rice prices has narrowed as analysis, deflated by the consumer price index. Crop price rice consumption has increased. data are drawn from Système d’Information sur le Marché Agricole (SIMA) and Global Information and Early Warning Interannual price variability for all of the crops analyzed above System (GIEWS) and livestock price data from FAOSTAT. is moderate, with coef�cients of variation ranging from 0.31 for imported rice to 0.21 for local rice (table 3.5). Niger’s free and active trade with Nigeria reduces the price fluctuations 3.3.1 Trends in Average Annual Cereal and Cowpea that would occur otherwise in response to fluctuations in Prices domestic cereal and cowpea production. Millet and sor- Trends in real average annual cereal and cowpea prices are ghum prices track each other very closely in terms of price illustrated in �gure 3.3. Real millet, sorghum, and cowpea level and price changes,27 con�rming that they are very close prices have all fallen signi�cantly relative to the early 1980s (table 3.5). Most of this decrease occurred during the 1980s, 27 The exception is an increase in millet prices in 1996, due to the followed by an increase in prices since the 1990s. The de- combined impact of lower than average production and an 18 crease in real millet and sorghum prices since 2005 indicates percent appreciation in the naira relative to the FCFA. AGR ICULTURA L SEC TOR RISK ASS ES S MENT IN NIGER C H A P T E R 3 — AG R ICULT URAL SE CTOR RISKS 21 TABLE 3.5: Crop Price Trends and Characteristics: 1980–2011 MILLET SORGHUM COWPEA LOCAL RICE IMPORTED RICE Percent change 1980–90a −42.3% −37.6% −29.5% N/A 0.9% b Percent change 1990–2011 25.0% 20.4% 6.6% 12.7% 3.7% Percent change 1980–2011 −27.8% −24.9% −24.8% N/A 4.7% Coef�cient of variation 0.30 0.25 0.28c 0.21c 0.31 Sources: SIMA; Author’s calculations. a Average for 1980–83 versus 1989–91. b Average for 1989–91 versus 2009–11. c Adjusted for trend using the Cuddy Delle-Valle Index. substitutes. This ready access to close substitutes also helps power occurred in 1996, a year not considered as especially to reduce price variability of staple foods. There is no indica- severe. Moreover, the ratio does not reflect the problems tion that the level of interannual price volatility has increased extant in 1993 as a result of drought and locusts. These two or decreased over time. cases are pertinent in that they suggest �rst that price shocks can occur in the absence of drought, as in 1996, and second The price spikes that are observed do not always coincide that drought is not always associated with price shocks, as with major drought or other adverse events. Of the seven in 1993. major drought shocks observed since 1980 (�gure 3.2), sharp increases in real millet and sorghum prices are evident in 1997–98, 2000–01, and 2004–05. Cowpea prices rose sharp- 3.3.3 Seasonal Price Movements ly in 1984–85, 1995–96, 2000–01, 2004–05, and 2009–10. Closer analysis of seasonal price movements provides fur- This characteristic is further evidence of the role of trade with ther insight into the nature of price variability. Analysis is Nigeria and other countries such as Benin in assuring food based on monthly millet prices for Maradi—a large regional supply and reducing price variability. market center close to the Nigerian border and a major lo- cus of poverty.28 Analysis is based on monthly millet prices Recent empirical work has shown that cereal markets in for the period January 1990 to December 2011. Real prices, Niger are highly integrated (Akers, 2008 op cit), and that deflated by the monthly consumer price index (CPI), are also prices in different regional markets move in close unison. used for January 2000 onward (monthly CPI data are not Hence, regional price trends are assumed to be similar to available before January 2000 due to a change in the basis the trends observed for aggregate national prices. In fact, for CPI calculation). These seasonal trends are presented in Akers �nds that markets become even more integrated dur- �gure 3.5 below. ing drought years. The following trends and characteristics are apparent: 3.3.2 Livestock-Cereal Price Ratios 1. Major spikes in nominal prices occurred in 1998, Livestock-millet price ratios provide an indication of changes 2001, 2002, 2005, 2009, and 2010. These spikes are in the purchasing power of pastoral households in response also observed for real prices, although they are much to adverse events. As livestock are frequently sold to buy less pronounced in 2009 and 2010. food in periods of hardship, any change in the price of one 2. The price spike in 2005 was the most severe. Price or both can have major implications for household welfare. volatility has been much less pronounced since then. The marked interannual changes in these ratios in �gure 3.4 3. Prices typically start to rise after the end of the below con�rm this premise, although the changes are driven harvest in January and peak in July and August. This largely by changes in millet prices due to the limited variabil- prolonged period (6–7 months) of increasing prices ity of the livestock price series used. is followed by a relatively short period (3–4 months) Despite the data limitations, this ratio appears to be a more sensitive indicator of hardship than price movements alone. 28 With a poverty rate of 78 percent in 2010 and the highest re- Reduced purchasing power is evident in 1996, 1998, 2000– gional population, Maradi accounts for approximately 25 percent 01, and 2004–05. Note that the sharpest fall in purchasing of Niger’s poor (World Bank, 2011 op cit). TEC H N I C A L AS S I STANCE 22 FCFA/kg ratio – 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 18.0 0 50 100 150 200 250 300 Jan ’90 19 91 Jul ’90 Source: GIEWSS. Jan ’91 19 Jul ’91 92 Jan ’92 real 19 Jul ’92 93 nominal Jan ’93 goats/millet cattle/millet sheep/millet Jul ’93 19 94 Jan ’94 Jul ’94 19 Jan ’95 95 Jul ’95 19 Jan ’96 96 Jul ’96 Jan ’97 19 97 Jul ’97 FIGURE 3.5: Monthly Millet Prices, Maradi Jan ’98 19 Sources: FAOSTAT (livestock prices), SIM Agricole (millet prices). Jul ’98 98 FIGURE 3.4: Livestock-Millet Price Ratios (Meat) Jan ’99 Jul ’99 19 99 Jan ’00 Jul ’00 20 Jan ’01 00 year Jul ’01 20 Jan ’02 01 month and year Jul ’02 Jan ’03 20 02 Jul ’03 Jan ’04 20 Jul ’04 03 Jan ’05 20 Jul ’05 04 Jan ’06 Jul ’06 20 05 Jan ’07 Jul ’07 20 Jan ’08 06 Jul ’08 20 Jan ’09 07 Jul ’09 Jan ’10 20 08 Jul ’10 Jan ’11 20 Jul ’11 09 AGR ICULTURA L SEC TOR RISK ASS ES S MENT IN NIGER CH A PTER 3 — AGRICULTURA L SEC TOR R IS K S C H A P T E R 3 — AG R ICULT URAL SE CTOR RISKS 23 of rapidly falling prices, from July and August to scope to improve the effectiveness of targeted, subsidized October. cereal sales by starting interventions earlier in response to 4. Closer scrutiny of the major price spikes shows that above average price increases, and by targeting this support the pattern of seasonal price movements differs from on low-income households in areas served poorly by local other years in two ways: (i) the postharvest increase and regional markets. in prices starts earlier, lasts longer, and rises faster than in other years; and (ii) that the subsequent fall in 3.3.4 Exchange Rate Risk prices is more rapid and slightly shorter than in other years. This increased asymmetry suggests that there Niger’s heavy reliance on trade with Nigeria, for both agricul- is a signi�cant window of time for intervention during tural exports (livestock, cowpeas) and imports (coarse grains), the period of price increases, and that effective inter- means that movements in the FCFA/naira exchange rate are vention can quickly reduce prices. a potential source of market risk. In fact, parallel market data 5. Other than in 2005, the observed price spikes oc- from both sides of the border show that the FCFA/naira ex- cur in a sequence of two to three years (1996–98, change rate is relatively stable (�gure 3.6), with low adjusted 2001–02, 2008–10). The cumulative effect of these coef�cients of variation (0.06–0.08) for monthly exchange price spikes on low-income households may thus be rates for the period January 2003 to October 2011. Moderate as or more important than the year-speci�c impacts. (5 to 15 percent) changes occur occasionally due to border closures, as in 2003 (presidential elections in Nigeria), 2005, The strong association between seasonal price movements, and 2008 (food shortages in Niger); but these are short lived. the incidence of drought and other adverse events, and the Moreover, the Nigerian restrictions on foreign currency trade hardship endured during these periods of adversity seem to do not appear to restrict the parallel foreign currency market. con�rm the growing consensus according to which reduced This is attributed to the relatively open border between the access to food, along with reduced food availability, may two countries, active trade in both directions, and the prefer- be the most critical impact of drought and locust attacks. ence of Nigerian people to hold the FCFA as a reserve cur- High seasonal price spikes appear to be more closely and rency rather than naira. Most other trade is with other FCFA systematically associated with these events than interannual countries in the UEMOA (Union Economique et Monétaire changes in production or prices. This suggests that there is Ouest Africaine) region, which removes exchange rate risk. FIGURE 3.6: Exchange Rate: FCFA/Naira (Parallel Market) 7.0 Malanville (Benin) Jibia (Nigeria) Illéla (Niger) 6.0 5.0 4.0 rate 3.0 2.0 1.0 0 2003M1 2003M3 2003M5 2003M7 2003M9 2003M11 2004M1 2004M3 2004M5 2004M7 2004M9 2004M11 2005M1 2005M3 2005M5 2005M7 2005M9 2005M11 2006M1 2006M3 2006M5 2006M7 2006M9 2006M11 2007M1 2007M3 2007M5 2007M7 2007M9 2007M11 2008M1 2008M3 2008M5 2008M7 2008M9 2008M11 2009M1 2009M3 2009M5 2009M7 2009M9 2009M11 2010M1 2010M3 2010M5 2010M7 2010M9 2010M11 2011M1 2011M3 2011M5 2011M7 2011M9 2011M11 Source: SIMA. TEC H N I C A L AS S I STANCE 24 CH A PTER 3 — AGRICULTURA L SEC TOR R IS K S The gradual appreciation observed in �gure 3.6 is due to situation has seriously deteriorated, especially in the border exogenous factors, namely the appreciation of the euro29 areas of Mali, Libya, and Nigeria (Chad is an endemic prob- and the higher level of inflation in Nigeria. This appreciation lem). The combination of rebellion, banditry, and even ter- enhances Niger’s capacity to import essential food commodi- rorism exposes herding families and their assets to greater ties and energy, but reduces the competitiveness of agricul- danger. Insecurity in these areas also sharply curtails the tural exports. The reduction in competitiveness is beginning extent to which government services and national and to be felt in coastal markets such as Abidjan where products international nongovernmental organizations (NGOs) can from UEMOA are facing competition from Chinese products. intervene, or merely collect basic information. Furthermore, insecurity: 3.4 CONFLICTS  Induces population movements (this is one of the main risks covered in Niger’s contingency plan). These Rising pressures on common property resources, or on re- place additional heavy burdens on local populations, sources used within the bounds of customary law, have led government resources, and the natural resource base to frequent conflicts over access to pasture and water within (e.g., the recent influx of refugees from Mali). livestock herding groups, and between farmers and herders,  Invariably fosters vicious cycles. For instance, the dis- often with signi�cant loss of life. The �rst signi�cant attempt ruption of law and order and inflows of armed groups to resolve this issue took place in the mid-1980s, with the and weapons from conflict-affected neighboring coun- advent of the Code Rural process, whose objectives were tries usually heighten existing domestic tensions or to (1) provide for better management of natural resources conflicts. It can also increase opportunities for illegal and (2) reduce conflicts. This was given a major boost by the activities. implementation of the decentralization law of 2004, with the creation of various local commissions30 at the village, com-  Finally, insecurity discourages foreign direct invest- mune, and departmental levels. ment and places a prohibitive risk on domestic ones. In particular, the development of critical sectors on Literature review and in-�eld consultations reveal that con- which future growth expectations are based, such flicts among pastoralists have sharpened in marginal areas as oil and uranium, cannot take place in an insecure where local pastoral associations were unable to play a environment. conflict prevention or resolution role. By the same token, conflicts between farmers and herders have mostly taken The largest adverse impact is on trading community and place in areas where local communities had failed to improve herders engaged in transhumance, but the overall direct natural resource management, and where local (traditional short-term impact on the broader agricultural sector being and decentralized) conflict resolution mechanisms failed. relatively small. Conversely, successful intermediation by heads of associa- tions or traditional leaders31 and signi�cant improvements in the resource base appreciably reduced conflicts between 3.6 MACROECONOMIC SHOCKS IN NIGERIA farming and herding communities.32 Nigeria is the major trading partner of Niger and because of its size and generally stronger effective demand, it has been the major market for Niger’s agricultural exports, in particular 3.5 INSECURITY livestock and cowpeas. It has also been, in the past, an im- Insecurity has always been an issue for people moving valu- portant source of fertilizer (mostly through informal channels) able assets through large, unprotected spaces. Of late, the and fuel. Overall trade with Nigeria is critical for Niger to stabi- lize prices and supplies; however, at times, shocks in Nigeria, for example, political uncertainly, violent conflicts, droughts, 29 The FCFA is �xed to the euro at a rate of 1 euro = 655.957 FCFA. and changes in macroeconomic policies, have a destabiliz- 30 Commissions foncières départementales (CofoDep); Commis- ing impact on Niger’s agricultural sector, in terms of volatility sions foncières communales (CofoCom); Commissions fonci- in demand and supply of agricultural products. Unexpected ères de base (CofoB), and so on. changes in trade agreements or policy implementation, such 31 Zakara and Abarchi, 2007. as import restrictions on basic commodities (e.g., rice) or 32 Comité Inter-Etate pour la Lutte contre la Sécheresse au Sahel (French: Permanent Inter-State Committee for Drought Control in export restrictions (e.g., 2004–05), can reduce Niger’s abil- the Sahel) and Centre for International Cooperation (CIS), 2009. ity to manage supply shocks caused due to production risks; AGR ICULTURA L SEC TOR RISK ASS ES S MENT IN NIGER C H A P T E R 3 — AG R ICULT URAL SE CTOR RISKS 25 however, in the past, their impacts have been moderate and drought and locusts are taken to be two independent events, short lived. although dry climatic conditions favor both; but in 2004, Niger suffered big losses due to simultaneous occurrence of both these risks. Simultaneous occurrence of multiple risks 3.7 POLITICAL INSTABILITY AND REGULATORY has led to most devastating crises in Niger and needs to be RISKS monitored closely. Niger has had a tumultuous political history. The �rst two republics spanned a 30-year political era of single-party rule 3.8.2 Dependent Risks and military juntas (1961–1991). Various types of govern- Drought is the principal trigger that leads to sudden spikes ments were in power over the next 20 years, until the recent in commodity prices; however, price volatility can also be advent of the seventh republic through free democratic elec- caused by sudden supply de�cits due to other production tions (March 2011). There have been four coups d’état since shocks, exchange rate risks, or import and export restrictions independence (1974, 1996, 1999, and 2010) and two major by neighboring countries. Drought also creates fodder scar- periods of Tuareg/Toubou rebellion (1990–95 and 2007–08). city, thereby contributing to and intensifying conflict risk be- Niger has witnessed two transitional governments, each last- tween herders and farmers, as well as making animals more ing over a year. This political instability contributed to frequent susceptible to diseases. Grain-eating birds and grasshoppers changes in policy and institutional environment. are usually reliant on pastures to meet their food require- This instability, especially the period of transitional govern- ments; however, fodder scarcity during drought forces them ments, led to rapid deterioration of the economy. In 1992, it to move to cultivated areas, causing severe damage to ma- resulted in reduction of GDP growth rate by 7 percent, while ture crops. These interdependencies between different risks in 1999 the decline was 1 percent. The impact on the agricul- need careful attention to ensure that the combined problem, tural sector, however, was much less pronounced and more rather its separate symptoms, is being addressed. indirect and might include: (1) reduced access to particular regions, which means that rural markets are more restricted, 3.8.3 Regional Shocks food prices rise, and aid can’t get through (This is more criti- Due consideration needs to be given to regional systemic cal in food de�cit areas, but its aggregate impact depends risks (e.g., regional droughts such as 1973 and 1984) and on how important the region is for food production and as a regional locust invasions (as in 1988), which can severely trade corridor.); (2) reduced public and private investment in impair people’s ability to cope with risks in Niger. Regional response to higher levels of uncertainty; (3) the diversion of events can severely curtail the ability to respond to a crisis public expenditure to military purposes to the detriment of by markets, governments, and other actors’ (e.g., 2004), and other public services; and (4) loss of donor support. Political can cause grave suffering to people across the region. instability may have a bigger impact on the agriculture sec- tor when it coincides with other shocks like drought (e.g., 3.8.4 Management of One Risk Creating Another Risk 1995–97). Political instability induces changes in government Action by individual actors in the supply chain, of by the priorities and contributes to volatility in agricultural sector governments of neighboring countries, to manage their own funding. risks could severely curtail the ability of other stakeholders to manage risk. In the past, action by individual countries, for example banning of grain exports by Mali and Burkina Faso 3.8 INTERLINKAGES BETWEEN RISKS in 2004 and border closures by Nigeria (2005) reduced the While the individual risks described above are signi�cant flow of grains and led to further increase in prices, restrict- threats to the agriculture sector, it is important to understand ing Niger’s ability to cope with the drought shock or a sharp interlinkages between these risks. There are some speci�c increase in food prices. dimensions of agricultural risks that need special attention. 3.8.5 Differential Impact of Risk 3.8.1 Independent Risks Occurring Simultaneously Finally, the location and timing of risk has different impacts. Flood and drought are usually considered as two independent The 2009 drought, owing to its prevalence in the rangeland risks; however, there are instances when both these events and forage area, was a catastrophic event for the livestock affect the country in the same year (e.g., 2009). Similarly, population, while the 2011 drought was more severe for crop TEC H N I C A L AS S I STANCE 26 CH A PTER 3 — AGRICULTURA L SEC TOR R IS K S production. This more granular distinction of risk will help in succession of droughts during the 1980s and the successive identifying the relevant affected population and selecting ap- years of higher and higher seasonal price peaks for cereals propriate measures to risk management. are other examples. The higher rainfall since the late 1990s has reduced the incidence of such successive shocks and Brief discussion of the impact of a serial succession of replaced it with a combination of droughts and floods, but shocks might also be useful here. The combination of political the possibility is still there. instability and drought in the mid-1990s for example, and the AGR ICULTURA L SEC TOR RISK ASS ES S MENT IN NIGER C H A P T E R 4 — A DV E RSE IMPACT OF AGRICULT URAL RISK S 27 Chapter 4: ADVERSE IMPACT OF AGRICULTURAL RISKS The conceptual and methodological basis used for analysis and catastrophic losses at more than 0.66 standard deviation is outlined �rst, then applied to production and market risks. from trend. These thresholds captured the various levels of The various sources of risk are then prioritized on the basis of known adverse events during the period of analysis. expected loss estimates. 4.1.2 The Indicative Value of Losses Available data on actual losses due to adverse events are not 4.1 CONCEPTUAL AND METHODOLOGICAL BASIS always accurate or consistent enough to facilitate compari- FOR ANALYSIS son and ranking of the costs of these events. Analysis was For the purposes of this study, risk is de�ned as an expo- thus based on estimates of the indicative value of losses, sure to a signi�cant �nancial loss or other adverse outcome which provide a more effective basis for comparison. While whose occurrence and severity is unpredictable, but for these estimates draw on actual data as much as possible, which some probability of occurrence can be estimated on they do not represent actual losses. Indicative loss values the basis of historical experience. Thus, risk implies exposure are also compared to agricultural GDP in the relevant year in to substantive losses, over and above the normal costs of do- order to provide a relative measure of the loss. ing business. In agriculture, farmers incur small losses each year due to unexpected events such as suboptimal climatic Indicative losses were calculated as follows: conditions at different times in the production cycle and/or For production risks, the total value of gross agricultural modest departures from expected output or input prices. output (GAO) lost for each event was �rst calculated in FCFA Risk refers to the more severe and unpredictable adverse as the difference between the actual and trend values of the events that occur beyond these smaller events. relevant crop or crops, using real producer prices (2010 = 100). The proportion of this total loss value in excess of the 4.1.1 Loss Thresholds threshold for trend production losses was deemed to repre- As agricultural production is inherently variable, the �rst step sent the loss attributable to the adverse event. The resultant for analysis is to de�ne loss thresholds, which distinguish value was also converted into $US at 2010 exchange rates. adverse events from smaller, interannual variations in output. Note that this measure reflects the combined impact of in- This is achieved by �rst estimating a time trend of expected terannual changes in both production and price. Comparable production in any given year, based on actual production, and deviations from trend can thus translate into quite different treating the downside difference between actual and expect- levels of indicative loss, depending on the extent to which ed production as a measure of loss. Loss thresholds are then production falls and prices change. set for these downside deviations from trend, to distinguish between losses due to adverse events and those that reflect Production risks were analyzed only for crops as the available the normal costs of doing business in an uncertain environ- and livestock data were considered inadequate. Preliminary ment. Two thresholds are used to represent differing levels analysis based on available livestock data showed that se- of severity: severe losses and catastrophic losses. These vere livestock losses were observed in 1995 and 1998 only, below-threshold deviations from trend allow quantitative and that the level of loss in these two years was relatively analysis of the frequency, severity, and cost of loss over a low. However, actual livestock losses are known to be more given time period. severe and more frequent. As in many countries, reported annual livestock numbers and production in Niger are based For purposes of analysis, the threshold for severe losses on a series of coef�cients, which remain fairly constant ir- was set at more than 0.33 standard deviation from trend, respective of actual production conditions. Hence, livestock TEC H N I C A L AS S I STANCE 28 C H A PTER 4 — A DV ER S E IMPACT OF AGR ICULTUR A L R IS K S TABLE 4.1: Frequency, Severity, and Cost of Adverse Events for Crop Production PERCENT DEVIATION OF YEAR PRODUCTION FROM TRENDa INDICATIVE LOSS VALUEbc (2010) CONTEXT FCFA (m) $US (m) PERCENT Ag GDP 1995 −23.2% −41,382 −122.1 −24.1% Drought, localized locust attacks, political uncertainty 1996 −13.6% −12,504 −35.0 −6.2% Political uncertainty 1997 −23.1% −49,892 −135.8 −23.7% Drought 2000 −9.1% −4,159 −10.8 −1.4% Drought 2004 −17.2% −50,920 −125.3 −11.6% Drought, locusts 2005 −7.0% −1,827 −4.2 −0.3% Low rainfall 2009 −10.4% −27,244 −55.6 −3.1% Drought, floods FREQUENCY OF ADVERSE EVENTS Severe 4/20 Catastrophic 3/20 Sources: FAOSTAT; Author’s calculations. a Estimated as a linear regression by ordinary least squares. b Calculated as the value of actual minus trend production, less the threshold for normal losses from trend. c In 2010, values based on real FCFA prices (2010 = 100), and $US/FCFA exchange rates for 2010. production losses due to drought or disease are not ad- events was based on qualitative information on the incidence equately captured. and magnitude of known events drawn from of�cial reports, media accounts, and interviews with stakeholders. For price risks, the trend level of production for the relevant crop was used as the point of reference. The total loss due to a price fall was then calculated in (current) FCFA as the 4.2 CROP PRODUCTION RISKS difference between GAO at trend prices minus GAO at actual Measured in terms of gross agricultural value,33 crop produc- prices, and the remainder of the calculation was derived as tion was signi�cantly reduced seven times by adverse events for production risks. The use of trend production (rather than from 1991–2010, an overall frequency of 0.35 (table 4.1). actual production) as the basis for analysis allowed the loss Three of these events were catastrophic, with a frequency of due to adverse price events to be calculated independently 0.15. Most of these adverse events resulted in a 10 to 20 per- of losses due to an adverse fall in production. cent fall from underlying production trends. Indicative losses were high for the catastrophic events, as would be expected, 4.1.3 Data whether measured in �nancial terms or as a percentage of Analysis of this nature requires a consistent set of data on agricultural GDP. both production and prices, for an extended time period, with a reasonable level of disaggregation. Of the various sources of Drought was the main cause of these shocks, sometimes, in data available, FAOSTAT’s data series on the value of gross ag- combination with other events. It was also the main cause of ricultural production (1991–2009) was the most suitable. This two of the three catastrophic events (1997 and 2004). A se- series facilitates analysis of risk over a 19-year period, which vere, continuous period of hardship occurred from 1995–97 was increased to 20 years on the basis that 2010 was a record in response to the combination of drought and political uncer- year for agricultural production (with no consequent produc- tainty. The cumulative impact of multiple, successive shocks tion shocks). National level data on the main crops are available such as these is devastating, and beyond the resources of from 1980 onward, but there are no parallel data on producer most low-income countries to manage. The lesser shocks in prices, and production data are patchy for most cash crops. 1996, 2000, and 2009 reflect lower level droughts and the political uncertainty during 1995–96. Note also that the lower The derivation of adverse events and indicative losses, based on the methodology outlined above, is purely quantitative. 33 Aggregate value of volume of production for each crop times pro- The attribution of these downside deviations to actual adverse ducer price. AGR ICULTURA L SEC TOR RISK ASS ES S MENT IN NIGER C H A P T E R 4 — A DV E RSE IMPACT OF AGRICULT URAL RISK S 29 TABLE 4.2: Frequency, Severity, and Cost of Adverse Events for Millet Production PERCENT DEVIATION OF YEAR PRODUCTION FROM TRENDa INDICATIVE LOSS VALUEbc (2010) CONTEXT FCFA (m) $US (m) PERCENT Ag GDP 1995 −19.7% −17,897 −53 −10.4% Drought, localized locust attacks, political uncertainty 1996 −26.9% −30,985 −87 −15.3% Political uncertainty 1997 −24.5% −28,872 −79 −13.7% Drought 2000 −13.4% −12,754 −33 −4.2% Drought 2004 −20.0% −31,184 −77 −7.1% Drought, locusts 2005 −6.0% −831 −2 −0.1% Low rainfall 2009 −12.4% −18,911 −39 −2.2% Drought, floods FREQUENCY OF ADVERSE EVENTS Severe 3/20 Catastrophic 4/20 Sources: FAOSTAT; Author’s calculations. a Estimated as a linear regression by ordinary least squares. b Calculated as the value of actual minus trend production, less the threshold for normal losses from trend. c In 2010, values based on real FCFA prices (2010 = 100), and $US/FCFA exchange rates for 2010. level drought in 1993 did not result in a fall in production be- rather than actual production as the basis for analysis (sec- low the threshold for adverse events. tion 4.1.2). This methodology was applied to producer prices for millet, sorghum, groundnut, fonio, sesame, and onions for 4.2.1 Major Cereal Crops the period 1991–2009 (there were no producer price data for Millet and sorghum are the staple foods in Niger, accounting cowpeas). Results of this analysis elicited only one adverse for 45 to 50 percent and 10 to 12 percent of the value of gross price event—for millet in 2007 (table 4.4). The indicative loss agricultural output, respectively (FAOSTAT). The characteris- of this price shock was relatively low. tics of shocks to millet production follow the same general Ostensibly surprising, this low incidence of adverse price pattern as for overall crop production (table 4.2), consistent shocks is actually consistent with the low interannual variabil- with its high contribution to total crop production. ity of the producer price data during the period of analysis. Sorghum production exhibits a higher frequency of cata- Coef�cients of variation (adjusted for trend) ranged from 0.13 strophic shocks (0.25), consistent with its lower tolerance to for millet down to 0.04 for sesame seed (for nominal prices). drought (table 4.3). The overall incidence of shocks is also It is unclear whether the low level of variability for producer higher (0.40). But the indicative costs of adverse events are prices corresponds to actual trends during the period of anal- much lower, due to sorghum’s lower contribution to the ag- ysis or reflects weaknesses in the data. For instance, some- gregate value of crop production. what higher levels of variability were observed for consumer prices for these commodities in the analysis in chapter 2. The slightly higher frequency of adverse events for individual But, overall, these results are consistent with the high level crops relative to that for all crops combined also indicates of activity in local and regional markets in Niger, the grow- two important characteristics of agricultural risk in Niger: ing evidence of strong market integration both internally and �rst, that the diversi�cation of crops grown does reduce risk, within the region, and the relatively free and active trade with and second, that this potential advantage remains limited, neighboring Nigeria, Benin, Cote d’Ivoire, and Ghana. Strong, since millet still dominates production. active markets invariably help minimize price variability. The limited apparent impact of interannual price variability on 4.3 CROP PRICE RISK agricultural sector risk does not mean that price variability is The methodology used to examine adverse price events not an issue. The analysis examines downside deviations and captures the impact of price change alone, by using trend the risk of these deviations to producers. Upside deviations, TEC H N I C A L AS S I STANCE 30 C H A PTER 4 — A DV ER S E IMPACT OF AGR ICULTUR A L R IS K S TABLE 4.3: Frequency, Severity, and Cost of Adverse Events for Sorghum Production PERCENT DEVIATION OF YEAR PRODUCTION FROM TRENDa INDICATIVE LOSS VALUEbc (2010) CONTEXT FCFA (m) $US (m) PERCENT Ag GDP 1995 −37.7% −9,151 −18 −3.7% Drought, localized locust attacks, political uncertainty 1996 −18.6% −1,452 −3 −0.5% Political uncertainty 1997 −36.2% −10,524 −21 −3.7% Drought 1998 −18.7% −2,746 −6 −0.6% Unidenti�ed 1999 −22.0% −5,263 −11 −1.3% Unidenti�ed 2000 −40.4% −16,226 −33 −4.2% Drought 2004 −19.9% −7,165 −14 −1.3% Drought, locusts 2009 −21.5% −9,887 −20 −1.1% Drought, floods FREQUENCY OF ADVERSE EVENTS Severe 3/20 Catastrophic 5/20 Sources: FAOSTAT; Author’s calculations. a Estimated as a linear regression by ordinary least squares. b Calculated as the value of actual minus trend production, less the threshold for normal losses from trend. c In 2010, values based on real FCFA prices (2010 = 100), and $US/FCFA exchange rates for 2010. TABLE 4.4: Frequency, Severity, and Cost of Adverse Price Events for Crops PERCENT DEVIATION OF NOMINAL YEAR PRICE FROM TRENDa INDICATIVE LOSS VALUEbc (2010) FCFA (m) $US (m) PERCENT Ag GDP 2007 −9.6% −3,051 −6.2 −0.4% FREQUENCY OF ADVERSE EVENTS Severe 1/19 Catastrophic 0/19 Sources: FAOSTAT; Author’s calculations. a Estimated as a linear regression by ordinary least squares. b Calculated as the value of actual minus trend production, less the threshold for normal losses from trend. c In 2010, values based on real FCFA prices (2010 = 100), and $US/FCFA exchange rates for 2010. and particularly sharp seasonal price spikes are a major risk for low-income consumers, as discussed in chapter 3. TABLE 4.5: Frequency, Severity, and Estimated Costs of Locust Attacks 4.4 LOCUST ATTACKS AREA SPRAYED ESTIMATED CROP INDICATIVE Estimates of the indicative crop losses from locust attacks (ha) AREA LOST (ha) LOSSa ($US m) were based on the assumption that 50 percent of the area 1980 110,000 55,000 4.3 sprayed was grazing land and 50 precent was crop land. 1986 65,000 32,500 2.4 Production losses were calculated in 2010 prices based on 1987 27,500 13,750 0.8 the average millet yield for the relevant year multiplied by the 1988 750,000 375,000 32.8 millet producer price for 2010. 1995 9,500 4,750 0.3 Signi�cant locust attacks (de�ned as those resulting in spray- 2004 350,000 175,000 11.1 ing of more than 5,000 hectares) occurred six times from Sources: FAO Locust Updates; Author’s calculations. a $US/FCFA exchange rate for 2010. 1980–2010 (table 4.5). Two of these attacks (in 1988 and AGR ICULTURA L SEC TOR RISK ASS ES S MENT IN NIGER C H A P T E R 4 — A DV E RSE IMPACT OF AGRICULT URAL RISK S 31 TABLE 4.6: Frequency, Severity, and Estimated Costs of Floods to Agriculture YEAR CROPS LIVESTOCK AREA (ha) INDICATIVE LOSS ($US m) LIVESTOCK LOST INDICATIVE LOSS ($US m) 1988 7,500 0.66 No data No data 1994 No data No data No data No data 1998 9,196 0.72 1,254 large ruminants, 6,544 small ruminants 0.65 1999 2,736 0.20 26 large ruminants, 215 small ruminants 0.02 2005 446 0.04 7,431 small ruminants, 896 cattle, 208 donkeys, 59 camels 0.70 2007 2,210 0.17 215 small ruminants 0.01 2008 55 0.01 20 small ruminants No data 2009 424 0.03 23,585 (all animals) 2.08 2010 687 0.06 115,114 (all animals) 10.17 Average 0.24 Average 1.95 Sources: Dartford Flood Observatory; Cellule de Coordination du Système d’Alert Précoce; Author’s calculations. 2004) had catastrophic results. This represents a frequency The available data show that indicative losses for livestock of 0.20 for all attacks and 0.07 for catastrophic attacks. The are 10 times higher than for crops, suggesting that affected average indicative loss for all locust outbreaks was $8.6 mil- households can lose a signi�cant part of their asset base in lion, an amount raised signi�cantly by the two catastrophic addition to their immediate source of food. As for locusts, attacks in 1988 and 2004. For the four less severe outbreaks, the calculation of indicative losses was based on price and the average loss was $1.95 million. (millet) yield data for 2010. While the crop production losses are high, the ex-post cost of controlling the outbreak is also high. In 2004, approximately 4.6 EXPECTED LOSSES AND RISK PRIORITIES $11 million34 was spent in Niger to control the outbreak. But FOR CROP PRODUCTION the cost for the region was much higher, with an estimated Results of the preceding analysis can be summarized in $400 million spent to control the outbreak for all of the Sahel terms of the expected loss associated with the main risks countries in 2004–05, including some 13 million liters of to crop production (�gure 4.1). The frequency of each risk is (mostly organo-phospate) pesticides.35 According to FAO/ based on its occurrence during the past 20 to 30 years, and EMPRES and the Commission for Controlling the Desert the associated loss is estimated as the average of the indica- Locust in the Western Region (CLCPRO) estimates, this tive costs for each type of risk during the period of analysis. amount could have paid for the equivalent cost of 170 years of prevention activities. Preventive measures are much cheaper. The pre-eminence of drought as the major source of risk is Niger currently spends approximately $400,000 annually on clear. It surpasses all other sources of risk in terms of both fre- regular locust monitoring, detection, and early control. quency and cost. Locusts emerge as the next most important source of risk (even without an accounting of losses to the live- stock subsector), followed by producer price risk and floods. 4.5 FLOODS Two further conclusions emerge from this analysis. The �rst Floods are a medium frequency (0.3) event, with low direct is the high frequency of adverse shocks to agriculture, in one losses in terms of crops destroyed and livestock lost (table 4.6). form or another (�gure 4.2). Adverse events of varying de- grees have affected crop production in 11 of the past 20 years. 34 In 20 countries, 130,000 square kilometers of area were treated Four of these shocks have been catastrophic. In a country as at the cost of $400 million, averaging $3,076 per square kilome- ter. Total area treated in Niger was approximately 3,750 square poor and risk prone as Niger, it is not enough to just identify meters (375,000 hectares), with an average cost of $3,076 per the most important risk or risks to address, and to strengthen square kilometer, and therefore the cost of locust control in Ni- ger comes out to approximately $11.5 million. the capacity to manage such risks. This capacity must also be 35 Mohamed Lemine Ould Ahmedou. 2012; Plan national de ges- built to manage a constant stream of adverse events, of dif- tion du risque acridien. Government of Niger. 2011. fering types and differing magnitudes, often in combination. TEC H N I C A L AS S I STANCE 32 C H A PTER 4 — A DV ER S E IMPACT OF AGR ICULTUR A L R IS K S FIGURE 4.1: Expected Average Loss for Adverse Crop Production Events 90 80 70 60 indicative loss $US m drought 50 40 30 20 locusts price 10 floods 0 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 frequency Source: Author’s calculations. FIGURE 4.2: Chronology of Adverse Crop Production Events 140 drought locust 120 flood price 100 indicative loss $US m 80 60 40 20 0 91 92 93 94 95 96 97 98 99 00 01 02 03 04 05 06 07 08 09 10 19 19 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20 20 20 20 year Source: Author’s calculations. AGR ICULTURA L SEC TOR RISK ASS ES S MENT IN NIGER C H A P T E R 4 — A DV E RSE IMPACT OF AGRICULT URAL RISK S 33 TABLE 4.7: Losses to the Livestock Sector YEAR EVENT COMBINED IMPACTS OF WEATHER RISKS AND LIVESTOCK DISEASES 1973–74 Generalized drought in the Sahel countries It is estimated that Niger lost 45% of cattle, 27% of sheep, and 15% of goat as a result of drought of 1973.a 1984–85 Generalized drought in the Sahel countries. It is estimated that Niger lost 40% of cattle, 35% of sheep, and 33% of goat as a result of drought of 1984.a 2004–05 Combination of drought and desert locust Losses to the livestock sector were estimated at 20% for cattle and 13% for small ruminants in sample invasion with a high impact on grazing areas parts of the agro-pastoral zone.b 2009–10 Combination of drought and loss of pasture In the sample area of a study on the impact of this crisis (14 departments of 7 regions in the pastoral and with subsequent heavy rains and flooding agro-pastoral zones), livestock mortality rates were estimated at 25.5% for cattle, 38.6% for sheep, 31.3% for goats, and 2.6% for camels.c. a Analysis of livestock statistics data from Institut National de Statistique. Annuaire Statistique des Cinquantes Ans d’Independence du Niger. Edition Speciale. Novembre 2010. b Charasse and Gouteyron, 2005. c The study breaks down the livestock mortalities due to the various risks associated with a drought and other weather events in 2009: lack of forage due to drought (38%), diseases (35%), heavy rains and floods (23%), and lack of water (4%), although the actual percentages were clearly different in the northern part of the pastoral zone and the southern fringes of the agro-pastoral one. The second conclusion is the implication of political instability 4.7 LOSSES TO THE LIVESTOCK SECTOR DUE TO for agricultural production and the management of agricultural MAJOR DROUGHTS risk. The period from 1995–96 was a period of deep, continu- Considering data limitations, it is dif�cult to quantify the ous shocks to agricultural production, even though produc- losses to the livestock sector. However, based on second- tion conditions were not unduly harsh. The country and its ary data, table 4.7 attempts to illustrate potential losses to population were thus poorly prepared for the drought in 1997 . the livestock sector from major droughts. From the herders’ Three consecutive years of extreme hardship thus occurred. point of view, the worst years are those when droughts lead FIGURE 4.3: Annual GDP Growth and GDP per Capita GDP growth (annual %) GDP per capita growth (annual %) 10 5 0 2000 2002 2003 2004 2005 2006 2007 2008 2009 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2001 2010 percent –5 drought drought political drought drought –10 1987 1990 instability +locust 2009 political 1999* drought 2004 instability 2000 –15 1992* –20 drought 1984 year Sources: World Development Indicators Database, 2012; and Author’s calculations. *1992 political instability (transitional government November 1991–April 1993)36 and 1999 political instability (assassination of President Ibrahim Bare’ Mainassare).37 36 President Ali Saibou’s regime acquiesced to demands for elections, and a transitional government was installed in November 1991 to manage the affairs of state until the institutions of the Third Republic were put in place in April 1993. The economy deteriorated over the course of the transition, leading to a fall in GDP in 1992. 37 In April 1999, President Baré Mainassare was assassinated in a coup led by Major Daouda Malam Wanke, who established a transitional National Reconciliation Council to oversee the drafting of a constitution for a Fifth Republic. The Nigerien electorate approved the new constitution in July 1999 and held legislative and presidential elections in October and November 1999. The council transitioned to civilian rule in December 1999; however, the period of transitional government, similar to 1992, had severe consequences for the economy. TEC H N I C A L AS S I STANCE 34 C H A PTER 4 — A DV ER S E IMPACT OF AGR ICULTUR A L R IS K S to poor animal conditions, low livestock prices, and high revenue, and requiring substantial �nancial resources for grain and other basic food prices. This was the case in 2004, crisis management. Figure 4.3 demonstrates the volatility 2009, and to a lesser extent, 2011. Loss frequency is 0.1 (four of national GDP growth rate and per capita growth rate in events in 40 years). Niger over a period of 26 years (1984–2010). GDP growth rate was negative, or zero, in 8 out of 26 years analyzed. There is strong correlation between drop in GDP growth rates and 4.8 IMPACT OF AGRICULTURAL RISKS ON occurrence of risk events. Two of these drops could be largely NATIONAL GDP attributed to political instability, characterized by long dura- Besides impacting agricultural production, consequences of tion of transitional government, crippling the decision making agricultural risks are far wider, affecting foreign exchange earn- while six of these drops can be partly explained by droughts ings, GDP growth rate, per capita income, loss of government in those years. AGR ICULTURA L SEC TOR RISK ASS ES S MENT IN NIGER C H A P T E R 5 — R I S K P RIORIT IZ AT ION AND MANAGEMENT 35 Chapter 5: RISK PRIORITIZATION AND MANAGEMENT 5.1 RISK PRIORITIZATION crops. This analysis highlights six priority risks: (1) drought To better utilize scarce resources, it is important to under- (crops), (2) drought (livestock), (3) locust outbreaks, (4) stand which risks, or subset of risks, are causing maximum consumer price risk, (5) livestock diseases, and (6) political losses, and at a much greater frequency. Figure 4.1 in the instability. previous chapter highlights the priority risks, using quantita- The relative signi�cance of these risks to different livelihood tive measures, for the crop subsector. Due to the paucity practices vary and annex 3 summarizes the principle risks for of data, some of the risks could not be quanti�ed; however, major livelihood groups. From a macro perspective, among �gure 5.1 combines qualitative and quantitative measures, the top �ve risks identi�ed, the capacity to manage risk is based on the assessment team’s evaluation, to prioritize ma- currently high for locust outbreaks, due to recent actions jor risk for the entire agricultural sector, both livestock and FIGURE 5.1: Risk Prioritization crop pest and diseases (0.4) high livestock pest and Drought- diseases (0.35) crop (0.35) conflict flood (0.3) consumer price frequency of occurrence risk (0.3) insecurity political instability/ NIRA-CFA exchange regulatory risk locust (0.2) rate (0.2) in Niger (0.2) macro economic drought- shocks in Nigeria livestock (0.1) producer price risk (0.05) low negligible catastrophic severity of impact Source: Author’s calculations. Note: Figure in parenthesis denotes the frequency of adverse events based on tabulation of past events. In many cases (insecurity, conflict, macroeconomic shocks), the team was unable to tabulate the historical occurrence and their prioritization is based on subjective assessment. Sandstorms and bush�res were of minor signi�cance for the agriculture sector and therefore was not including in this prioritization exercise. TEC H N I C A L AS S I STANCE 36 C H A PTER 5 — R ISK PRIOR ITIZATION A ND MA NAGEM ENT FIGURE 5.2: Risk Layering Approach layer 3 very low frequency, very high losses layer 2 risk mitigation low frequency, + risk transfer medium losses + risk coping probability layer 1 risk mitigation + risk transfer high frequency, low losses risk mitigation severity Source: World Bank; Agricultural Sector Risk Assessment Brochure (2012) of the Agricultural Risk Management Team of the World Bank taken by the government of Niger and other stakeholders. form of compensation (cash or in-kind), social protec- Capacity to manage drought, political instability, crop pest tion programs, and livelihood recovery programs (e.g., and diseases, livestock pests and diseases, and consumer government assistance to farmers, debt restructuring, price volatility is relatively low and therefore, this section contingent risk �nancing). concentrates on management of these risks. A risk layering approach (�gure 5.2), based on probability of There is no silver bullet to manage any given risk. Effective occurrence and potential losses, is used to select an appro- risk management typically requires a combination of mea- priate risk management strategy. Risk mitigation cuts across sures, some designed to remove underlying constraints and all the three layers and is the dominant approach across all others designed to directly address the risk. Resource avail- frequency and severity levels. Risk transfer is more appro- ability will often determine what is possible, and integrated priate for low frequency and moderate or high losses while risk management programs are often more effective than coping mechanisms trigger in for catastrophic losses, which stand-alone programs. Risk management measures could be are usually less frequent events. classi�ed into the following three categories: Table 5.1 highlights some of the interventions that could  Risk mitigation (ex ante). Actions designed to reduce be undertaken to manage selected risks in Niger, classi�ed the likelihood of risk or to reduce the severity of by management strategy. The following list is by no means losses (e.g., water harvesting and irrigation infrastruc- exhaustive, but is meant to illustrate the type of invest- ture, crop diversi�cation, extension). ments that, based on the analysis, have a strong potential to improve agricultural risk management in Niger. Although  Risk transfer (ex ante). Actions that will transfer the agricultural risk management measures are discussed risk to a willing third party. These mechanisms usually sequentially, many of these interventions, if implemented will trigger compensation in the case of a risk-gen- jointly, can have positive effects on each other and address erated loss (e.g., purchasing insurance, reinsurance, multiple risks. �nancial hedging tools).  Risk coping (ex post). Actions that will help the affect- The following section provides a brief description of nine ed population cope with the loss. They usually take the major interventions. AGR ICULTURA L SEC TOR RISK ASS ES S MENT IN NIGER C H A P T E R 5 — R I S K P RIORIT IZ AT ION AND MANAGEMENT 37 TABLE 5.1: Indicative Agricultural Risk Management Measures MITIGATION TRANSFERa COPING Improving availability of existing drought-resistant seed Farm level crop Use of weather index for triggering early varieties to the farmers insurance warning and response Adoption of soil and water conservation/NRM techniques Macro (government) Contingent �nancing and other �nancial level crop insurance instruments for �nancing coping strategies Drought (Crop) Improved farming techniques (e.g., intercropping, conservation Decentralized disaster contingent fund for tillage) rapid response to local emergencies Investment in small-scale irrigation (dry season farming) to Cash for work and food for work program improve nutrition and food diversity to support soil and water conservation Community level food banks Interventions to improving livestock feed and fodder Livestock insurance Facilitate early destocking of livestock availability (pasture improvement and forage production) Community level fodder and forage banks Livestock food and forage delivery Drought (Livestock) Modifying transhumance pattern More strategic destocking Health interventions (Improving access to deworming medicines and salt licks) Improved and sustained support for early detection and Social protection programs Locusts destruction of locusts Deeping of democratization process Social protection programs Political instability Poverty alleviation and economic growth Community level price stabilization Improving ef�ciency of emergency grain reserve Improving the ef�ciency of private storage (lower losses, etc.) Improved use of existing market information for earlier Consumer price risk response to food price spikes Lowering barriers of trade Market information Increasing production Supporting vaccination services Quarantine measures Livestock diseases and pests Improving veterinary services Source: Authors. a There are signi�cant challenges for potential risk transfer products in Niger (see box 5.1). 5.2 AGRICULTURAL RISK MANAGEMENT and high-yielding varieties could help mitigate the risk of MEASURES crop failure. During the 1980s, there was an emphasis on developing short-cycle varieties to mitigate the effect of late 5.2.1 Drought-Tolerant and Improved Seed Varieties seasonal onset or early cessation of rains. A number of suc- Widespread availability of drought-tolerant seed varieties cessful varieties, such as improved HKP and short-cycle (zat- and short-maturing varieties for cereals will help in ensuring ib, ICMD) millets, the 90-SN7 , the drought-tolerant IRAT 204 crop production during drought years. Compared to longer and other sorghums, the IT90, and TN88-63, were developed maturing varieties, these short-maturing varieties will have and disseminated by a local research OR outreach institution. higher yields in a drought year, but lower yields in a normal However, no national system was put in place to ensure sus- year, and they might be able to resist diseases a little longer tainable delivery of seed, at affordable or subsidized prices, during drought years. There is often a trade-off between to the farmers. As a result, despite all the work being done higher yield (in normal years) and drought tolerance (in by the government, partners, and private sector, less than six drought years), and early warning about impending weather percent of farming households have access to these drought seasons coupled with ready availability of drought-tolerant tolerant varieties for cereals. TEC H N I C A L AS S I STANCE 38 C H A PTER 5 — R ISK PRIOR ITIZATION A ND MA NAGEM ENT BOX 5.1: Potential for Risk Transfer Products in Niger Risk transfer, wherein a third party accepts a stakeholder’s situation, formal sophisticated instruments such risks at a price (premium), is often considered an effective as insurance might be dif�cult to comprehend risk management strategy to manage residual risks and is and adopt by the majority of the population. best suited for low-frequency and moderate- or high-severity Furthermore, the institutional infrastructure is weak events. Agriculture insurance and hedging in commodity ex- and there are limited distribution channels for change are two risk transfer instruments. large-scale delivery of agricultural insurance. Agricultural Insurance: There has been no past experi- These factors make it challenging to pilot and scale up ag- ence with agricultural insurance in Niger. There is reason- riculture insurance programs either at the farm level or at ably good quality of crop yield data (at departmental and the aggregate macro level. There might be some opportu- regional levels), presence of time-series weather informa- nities of insurance for irrigated commercial farmers, but for tion (daily and monthly rainfall data at more than 40 weath- the bulk of the rain-fed farmers, risk mitigation solutions er stations spread across the country), and a high correla- are better suited to meet their needs. tion between yield losses and weather variables (rainfall). Commodity Hedging: Commodity price volatility is a big These three factors make it technically feasible to design concern for the government of Niger and there is some in- an agricultural insurance contract; however, there are a terest to explore the feasibility of using commercial hedg- large number of other factors that makes agricultural insur- ing products, namely options, to lock in minimum prices ance a challenging proposition in Nigerian context. for food. Theoretically, it’s an interesting proposition; how- 1. High frequency. Insurance is feasible for low- ever, it faces several practical challenges. frequency events, since high frequency means 1. Lack of commodity Exchange to hedge risk. high payouts jeopardizing �nancial sustainability of Currently, there is no operational West African insurance companies. Drought, on the other hand, commodity exchange where the government could is a high-frequency risk at the departmental and hedge its exposure to price spikes of millet and national levels in Niger (annex 2). sorghum. 2. High premium rates. The combination of high fre- 2. Food Security. Food Security of Niger is reliant on quency and high severity of drought might result accessibility and affordability of millet and sorghum in prohibitively high insurance premium rates. In and these two commodities, unlike maize, has other developing countries, premium ratae for very little global trade or production. South African agricultural insurance, in general, are high, rang- Commodity Exchange (SAFEX) has a sorghum con- ing from anywhere between 5 to 20 percent of tract, but it has limited liquidity. Furthermore, there the crop value being insured. In a high-frequency might be little correlation between sorghum price case like Niger, premium rates are likely to be on a in West Africa and sorghum contract in SAFEX higher side. (called basis risk), which might make it dif�cult to 3. Affordability. The majority of the farmers in Niger use hedging products. are engaged in production of cereals primarily for 3. Landing cost. In the absence of any regional com- household consumption (millet and sorghum), and modity exchange, any potential hedging might their little cash income largely comes from the sale have to done in South Africa’s SAFEX; however, of surplus production of cowpeas, ground nuts, actual delivery cost from South Africa due to high and sesame. In a cash-strapped economy, where transportation cost, will make any option contract a majority of the farmers are not commercial farm- prohibitively expensive. ers, affordability of premium payments is a big challenge. 4. Option premium. The option premium for widely traded commodities, like coffee and maize, runs 4. Integration with �nancial sector. With less than in the range of 4 to 8 percent of the price of the 1 percent of the population having access to commodity being locked in. For a narrowly traded commercial banking, the bulk of the farming commodity like millet, the options premium might households have no prior experience or access to go up further, making it unaffordable. commercial saving and borrowing. Under such a AGR ICULTURA L SEC TOR RISK ASS ES S MENT IN NIGER C H A P T E R 5 — R I S K P RIORIT IZ AT ION AND MANAGEMENT 39 BOX 5.1: Potential for Risk Transfer Products in Niger (Continued ) These factors might make using options to hedge against price crisis times; and transparency in the regional stock availability spikes in Niger a challenging proposition. Perhaps, improving might be better instruments to ensure food availability and the existing system of emergency grain reserve; improving manage price volatility in the context of Niger. trade flows between neighboring countries, especially during Source: Authors. Addressing this issue from a risk management perspective 5.2.3 Desert Locust Control will require a national or regional level approach to develop After the 2004 desert locust invasion, the Government of a system that ensures the sustainable delivery of drought- Niger in 2007 decided, with FAO and World Bank support, to resistant and high-yielding varieties. This might entail further establish a national center for grasshopper and locust con- support and expansion of seed multiplication by seed pro- trol, Centre National de Lutte Anti-Acridienne and approved ducers cooperatives, expansion of seed multiplication by pro- a law earmarking an annual amount in the national budget ducers’ organizations and private commercial �rms, further to �nance locust outbreak prevention activities ($400,000 in support to agricultural input shops (boutiques d’intrants), and 2011). This funding for preventive action by a national budget continuation of social protection activities whereby NGOs line removes most of the problematic dependency on donor provide improved seeds to poor smallholder households. funding, which is usually activated only after the crisis. The Furthermore, the possibility of research and development approach in Niger is now reliant on monitoring of seasonal into newer varieties that are drought tolerant as well as high reproductive areas and localization and destruction of �rst lo- yielding (during normal year) could be explored. cust populations, before they move to the gregarious stage, when groups of larvae and swarms of juveniles are formed. 5.2.2 Soil and Water Conservation and Natural Resource This approach successfully controlled a potential outbreak in Management 2009, largely with the use of biopesticides.38 This approach has vastly improved Niger’s ability to manage locust outbreak Effective soil and water conservation techniques (zaï, demi- risk. However, there are challenges of having to operate in lunes, tied ridges, stone or vegetation lines, etc.) in Niger insecure parts of the country (Agadez) requiring military es- have successfully contributed to (1) concentrated rain water, corts, and regional information exchange and effective coor- (2) increased in�ltration, and (3) enhanced plant growth, dination with neighboring countries still needs improvement. which improves the resilience of crops during water stress and could serve as useful drought mitigation intervention. Improved natural resources management techniques such 5.2.4 Irrigation as the FMNR of parklands, besides contributing to manag- Irrigation has the potential to generate a sizeable gain in ing drought risk, also help to lower wind speed, reduce ero- household welfare, boost agricultural growth, improve food sion from water runoff, and contribute to mitigating the risks security, and promote overall economic growth in Niger. Over of flood and wind damage. The FMNR approach described the past few years, small-scale irrigation has expanded for above (soil and water conservation/NRM techniques) is best off-season cropping, largely to meet the growing demand for suited to promote effective agriculture-livestock integration, onion, other horticulture crops, and export crops. However, as the regenerated woodland park provides ample tree fodder irrigation and off-season cropping cover a very small area for various species, promotes additional recharge of ground- relative to total land cultivated. Irrigation is a useful instru- water resources, and so forth. An integrated agriculture- ment of drought risk mitigation; however, it has several limi- livestock sector will help in reducing livestocks’ exposure to tations. Past experiences have demonstrated that droughts drought, bush�re, insecurity, and conflict, and might allow for in Niger indicate reduced water availability for existing irriga- better timing of livestock sale, acting as a commercial buffer tion systems, leading to drops in production even in irrigated for strategic destocking by pastoral groups in bad years and land during the severe drought years. Furthermore, in Niger, restocking in good ones. Expanding the scale and scope of cereals, barring the exception of rice, are not grown under existing soil and water conservation/NRM intervention could provide high return on investment in the long run due to the 38 Based on Metharizium, a fungal disease affecting juvenile lo- multiple bene�ts they generate. custs and grasshoppers. TEC H N I C A L AS S I STANCE 40 C H A PTER 5 — R ISK PRIOR ITIZATION A ND MA NAGEM ENT irrigated conditions, and any irrigation increase is unlikely to 5.2.6 Contingent Financing lead to increase in cereal production. Securing adequate �nancing to support emergency response and crisis management is a big challenge for the government Nonetheless, there is a strong case for investment in irriga- of Niger. Delay in funding comes at a huge cost in poverty tion. While it might not be able to address severe systemic and humanitarian terms. Contingent line of credit is a World droughts, in the case of localized drought or poor rainfall dis- Bank �nancing instrument for middle-income countries that tribution, it could help ensure food availability in food de�cit provides rapid disbursement of funds during crisis. While areas. These is considerable potential to increase the area this fund is currently not available for low-income countries, under irrigation in Niger and more widespread irrigation could the possibility of a contingent grant facility for rapid dis- also contribute to improved nutrition, by access to diversi�ed bursement of funds might be a useful option for a country food, as well as improve household income, thereby reduc- like Niger. Considering that the bulk of the crisis in Niger is ing food affordability issues and improving household food caused by drought, an objective weather index or combina- security. tion of multiple indexes could trigger and activate release of emergency response funds to the government. Several 5.2.5 Shortening Emergency Response Time donors could pool resources into such a fund, which could Rapid response prevents a crisis from escalating to a higher be country or region speci�c, and could be managed by a degree and helps reduce asset losses for vulnerable house- multi-donor trust fund. holds. In Niger, however, despite the widespread knowledge of eminent crisis, the response is much delayed39 with 5.2.7 Strategic Destocking consequences on poverty, food insecurity, and malnutrition. The primarily marketable animals in a cattle herd are adult Usually, the performance of the agricultural season is largely steers and older cows; however, during crisis times, such as determined by the rainfall performance, known as early as in 2009–10, many herders had to resort to crisis destocking August; however, the response mechanism usually kicks in (i.e., selling all types of animals at rock-bottom prices), and by February or March of the following year and actual relief thousands of head had to be slaughtered to reduce pressure operations commence by May or June. By then, it’s usually on grazing resources and to provide herders with at least too late and conditions have deteriorated. The government of some income. Strategic offtake based on better information Niger, in collaboration with other development partners, were can allow pastoral groups to reduce herds to a more manage- able to act early in response to the drought of 2011, with able and less risky size when drought or desert locusts have planning response as early as October/November 2011 with reduced edible biomass, and when animals are more prone ground relief activities as early as January/February 2012. to disease. Access to seasonal forecasts, information on the There is a need for institutionalization and decentralization state of grazing and water resources, market conditions (sup- of early response, considering its scope in reducing losses ply, prices), exchange rates, and even the price of forage or and improving coping ability of rural households. Objective feed in areas where animals may have to be held before or af- and early assessment of impending crisis, using surrogate ter sale, are critical for herders while making decisions about measures like weather index in collaboration with quick strategic destocking. Since strategic destocking also implies vulnerability assessments, in combination with contingent access to markets, an important supporting measure is the funding and changes in government and partners’ operational effective implementation of cross-border veterinary and live- policies could help in shortening emergency response time. stock trade agreements between Niger, Nigeria, Mali, and Furthermore, more emphasis on indicators of reduced food Benin considering that animal mobility might be restricted by access and less emphasis on indicators of reduced food government actions during drought years. Last but not least, availability could also help shorten the response time. Price pastoral associations and their partners must provide herding movements reflect changes in supply and demand quite families with alternative and secure forms of savings. quickly and could act as a proxy for food access. Information of price movements could be collated from SIMA and could 5.2.8 Vaccination and Veterinary Services be used to identify thresholds where the rate and level of price change justi�es an immediate response. Vaccination is perhaps one of the most signi�cant measures to reduce the risk of livestock diseases. With limited re- sources, the government of Niger focuses on preventive vac- 39 Because the focus is on collecting information on food availability cination campaigns against the biggest threats and responds (production), this type of information takes longer to gather. to some of the worst outbreaks. However, given the size of AGR ICULTURA L SEC TOR RISK ASS ES S MENT IN NIGER C H A P T E R 5 — R I S K P RIORIT IZ AT ION AND MANAGEMENT 41 current losses in good and bad times, the returns on invest- respond to local level market failures without creating major ing in improved access to vaccination and veterinary services distortions. Despite the similarity in principles of managing and its effectiveness as a risk management measure are very community food and fodder banks, unique and separate dy- high. Just as we are making the case for an extension of namics of fodder availability for cattle and food availability for the ‘boutique d’intrants’ approach in agriculture, we argue for humans should be taken in due consideration while design- greater access to vaccination and veterinary services at the ing interventions to support them. commune or groupement pastoral levels. 5.3 PRIORITIZATION OF RISK MANAGEMENT 5.2.9 Community Level Food and Fodder Banks MEASURES The government of Niger undertakes a number of measures, It is important to highlight that almost all of the measures de- such as emergency grain reserves (see annex 4), delivery of scribed above are complementary in nature and will contrib- food at subsidized prices by OPVN, collection and dissemina- ute to improved risk management in the short, medium, and tion of food prices by SIMA, and so forth, that play a crucial long term. However, in a resource-constrained environment role in ensuring availability of food during emergencies and like Niger, decision makers are compelled to �nd the quick- stabilization of food prices for consumers. The Government est, cheapest, and most effective measures among a myriad of Niger has invested a lot on food and fodder banks over of possibilities. Ideally, a detailed, objective, and exhaustive the past few years, including under the World Bank funded cost-bene�t analysis will help in selecting the most appro- Community Action Plan (PAC1 and 2) and the Second priate intervention options. But conducting a cost-bene�t Emergency Food Security project under the GRFP (PUSA2/ analysis of so many different options in itself could be costly GFRP), however, more could be done to expand such pro- and time consuming. grams on an as-needed basis. Using price data as a proxy of supply shortages, the aim should be to intervene early in Using decision �lters to evaluate and prioritize among a list of the seasonal cycle, well before prices reach their seasonal potential interventions could help in making rational resource peak. Besides ensuring food and fodder availability for vulner- allocation decisions in lieu of a detailed cost-bene�t analysis. able populations, such interventions will also help stabilize The �lters described below (table 5.2 and 5.3) are indicative food and fodder prices for the wider population and can help and imperfect; nonetheless, they present a �rst step in the TABLE 5.2: Relative Bene�ts of Risk Management Measures YIELD/ ADDRESS REDUCES THE REDUCES THE REDUCES THE COMPENSATES PRODUCTIVITY MULTIPLE HAZARD EXPOSURE LOSSES AFTER THE LOSS IMPROVEMENTS RISKS Drought-tolerant seed No Yes Yes No Yes (in a drought, not No varieties (M) otherwise) Soil and water conservation/NRM (M) No Yes Yes No Yes Yes Irrigation (M) No Yes Yes No Yes No Early detection and No Yes Yes No Yes No destruction of locusts (M) Community level food and fodder No Yes Yes No No No banks stabilization (M and C) Vaccination programs (M) No Yes Yes No Yes No Insurance (T) No No No Yes No Yes Shortening emergency response No Yes Yes No No Yes time (C) Strategic destocking (C) No Yes Yes No Yes No Contingent �nancing (C) No Yes No Yes No Yes Source: Authors. Note: M = Mitigation, T = Transfer and C= coping. TEC H N I C A L AS S I STANCE 42 C H A PTER 5 — R ISK PRIOR ITIZATION A ND MA NAGEM ENT TABLE 5.3: Decision Filters and Intervention Classi�cation ADVERSE POTENTIAL IMPACT RELATIVE EASE OF IMPACT ON ON POVERTY SCALABILITY COST IMPLEMENTATION RETURN TIME ENVIRONMENT ALLEVIATION Drought-tolerant/improved High Medium Medium Short Low High seed varieties (M) Soil and water conserva- High Medium Medium Medium Low High tion (M) Irrigation (M) Low High Low Short–medium Moderate High Early detection and de- High Medium High Short Moderate Low struction of locusts (M) Community level food and High Medium Medium Short Low High fodder banks (M, C) Vaccination programs (M) High Medium Medium Medium Low High Contingent �nancing (C ) High Low High Short Low Low Shortening emergency Medium Low Mediium Short Low Low response time (C) Strategic destocking (C) Low Medium Low Medium Low Low Insurance (T) Low Low Medium Medium Low Low Source: Authors. Note: M is Mitigation, C is coping, and T is transfer. right direction. The government of Niger and partners could project is usually much higher than cost involved in choose other criteria as �lters, but it is important to en- setting up a system for seed distribution. sure clarity, consistency, and objectivity while using them  Scalability. While some interventions, due to pre- to evaluate decision options. The following criteria were requisites, might be able to bene�t a small group of used by the World Bank team, with the interventions ap- stakeholders, other interventions have a much greater propriately rated low, medium, and high, or short, medium, propensity to reach scales and bene�t a much larger and long. There are a number of complex analytic screening group of stakeholders. tools to assess all of these decision �lters and this study  Ease of implementation. Technical complexity of does not claim methodological rigor while assessing these interventions and the capacity of local stakeholders to �lters. Instead, the study team applied these �lters as a implement are �lters that could be used to prioritize sort of rapid assessment to obtain �rst order of approxi- decisions. Simpler interventions might �nd greater mation, based on their assessment of the situation on the acceptability and will be easier to implement. ground.  Return time. Some interventions have a much longer  Relative bene�ts. This �lter attempted to assess the gestation period, while others could yield quick potential bene�t of a particular intervention option, in results. While risk management will require short-, comparison to others. The most bene�cial interven- medium- and long-term perspectives, quick wins are tions are the ones that reduce the likelihood of oc- often a high priority for decision makers. currence of risk or reduce the losses, if the risk were  Adverse impact on environment. Some of the risk to occur. In addition, if the intervention could lead to management interventions, especially large-scale additional bene�ts, for example, yield improvements, spraying of chemicals for locust destruction, could ef�ciency improvements, and cost reduction, they have long-term catastrophic consequences for the were rated high. environment. Hence, it is important to scrutinize the  Relative cost. Without detailed assessments, it is potential adverse impact of a given intervention on the dif�cult to estimate the cost of some interventions. environment. However, based on the experience of the assess-  Potential impact on poverty alleviation. While some ment team, the relative cost of interventions could be interventions would directly contribute to improved assessed. The cost involved in a large-scale irrigation income and poverty alleviation, others might indirectly AGR ICULTURA L SEC TOR RISK ASS ES S MENT IN NIGER C H A P T E R 5 — R I S K P RIORIT IZ AT ION AND MANAGEMENT 43 contribute toward the goal. Using this �lter helps in projects take into account, directly or indirectly, agri- identifying risk management interventions that might cultural risks. These actions will be strengthened by yield large poverty alleviation dividends. development alliances with government interventions and other technical and �nancial partners, and aligned Based on prioritization of risk and intervention measures, the with the speci�c objectives of the 3N initiative. following six interventions might yield greatest risk manage-  Shift of focus from short-term crisis response to long- ment bene�ts: term risk management. Funding priorities reflect a bias 1. Drought-tolerant crop varieties toward crisis management and emergency response that results in responding to symptoms rather than 2. Soil and water conservation/NRM interventions resolving underlying problems. 3. Expansion of small-scale irrigation  Streamlining disparate donor investments and isolated 4. Support to community level food and fodder banks interventions toward the core problem. Niger is highly 5. Continuous support to early detection and destruc- reliant on of�cial development assistance (ODA) to tion of locusts fund its development programs. While �nancial sup- 6. Livestock vaccination programs. port from donors is necessary in a resource-scarce environment, it often results in diluting government attention toward multiple shifting priorities. While all 5.4 CONCLUSION other problems are important and critical, the govern- Agricultural risks and the implications for Niger are perhaps ment of Niger needs to focus on the core problem. well known, and much analyzed, discussed, and debated. Frequent occurrence of drought is the root problem In some ways, this document reinforces much of what is and unless that is resolved, issues of poverty alle- already known and acknowledges the informal risk manage- viation, economic growth, malnutrition, health, and ment already taking place in Niger, the steps being taken to environmental protection cannot be addressed. improve the situation, and the limitations—�nancial, human,  Instead of a new framework, integrate with the and institutional—that Niger faces in overcoming the chal- existing development framework. Niger already has a lenge of agricultural risk. number of strategies (e.g. Code Rural, climate change, 3N) which address risk management, implicitly or This documents contributes and enriches the existing knowl- explicitly. Rather than creating a new framework, risk edge base of the agricultural sector in Niger and provides the management interventions need to be integrated into following contributions: (1) systematically analyzing a whole the existing development framework to avoid duplica- range of agricultural risks and its impact over a longer time tion and create synergy. period (1980–2012), (2) helping situate drought in the context  Decentralized (farm and community levels) decision of other agricultural risks, (3) prioritizing the most important making. Despite the push for decentralization, the agricultural risk for the country based on objective criteria, (4) current systems are geared for top-down decision providing a framework of mitigate-transfer-cope to manage making. More can be done to empower local govern- risk prioritization, and (5) offering a �ltering mechanism to ment authorities and to give them the resources and select best possible interventions for agricultural risk man- flexibility to act locally in response to adverse events. agement. Implementing these measures would require: Improved access to �nancial and human resources  Sustained and substantial �nancial investments. A to support community level institutions such as majority of the risk management interventions that cereal and fodder banks, and enabling household and are capital intensive require substantial and sustained community level decision making for agricultural risk �nancial investments. In Niger where agriculture sec- management, will reduce the response time losses tor funding has been low and highly volatile, securing and prevent the spread of risks. resources for risk management will be a challenging  Prioritizing agricultural risks into government and donor proposition. Strong interdepartmental and intermin- strategies. There is an urgent need to incorporate isterial coordination, and increased leveraging of agricultural risk management explicitly in the govern- resources from all relevant sector wherever comple- ment of Niger’s strategy documents, policy priorities, mentarities exists, will be required. For example, most work plan, and national budget. Incorporating risk of the actions of the current portfolio of World Bank perspective in Country Assistance Strategies, Poverty TEC H N I C A L AS S I STANCE 44 C H A PTER 5 — R ISK PRIOR ITIZATION A ND MA NAGEM ENT Reduction Strategy Papers, and donor strategy docu- is rather weak. Lack of resources, both �nancial and ments will contribute to highlighting and mainstream- human; weak institutional structure; and, excessive ing agricultural risk management and will help ensure focus on strategy development have contributed to that adequate resources are made available to address lackluster performance on the implementation front. the issue. Effective implementation of selected interventions is  Implementation. While there are a number of good fundamental for improved agricultural risk manage- strategies in place, implementation on the ground ment in Niger. AGR ICULTURA L SEC TOR RISK ASS ES S MENT IN NIGER BIBLIOGRAPHY 45 BIBLIOGRAPHY Aker, J. 2008. “Rainfall Shocks, Markets and Food Crises: Evidence Frederic, M., and Mittal, A. 2006. Sahel: A Prisoner of Starvation? A from the Sahel. � Center for Global Development. Working Paper, Case Study of the 2005 Food Crisis in Niger. Oakland Institute. No. 157 . Geesing, D., Djibo, H., and Fourrager, P. 2001. Niger. FAO. Ali, A., and Lebel, T. 2009. The Sahelian standardized rainfall index re- Gubbels, P. 2011. Escaping the Hunger Cycle—Pathways to Resilience visited. International Journal of Climatology. 29(12): 1705–1714. in the Sahel. Sahel Working Group. Bationo, A. Waswa, B., Kihara, J., Adolwa, I., Vanlauwe, B., and Human Development AFTH2, Country Department AFCF2, Africa Saidou, K, eds. 2012. Lessons Learned from Long-Term Soil � World Region. 2009. “Niger: Food Security and Safety Nets. Fertility Management Experiments in Africa. Bank Report No. 44072-NE. Bockel, L., Thoreux, M., and Sayagh, S. 2008. A Review of Risk IMF. 2011. Niger: Ex-Post Assessment of Longer-Term Program Management Tools and Policies in Niger’s Rural Sector Policy Engagement. Brief. EASYPOL Module 208. FAO. Institut National de Statistique. 2010. Annuaire Statistique des Cabinet du Premier Ministre Secretariat Permandent de l SRP. 2007. Cinquantes Ans d’Independence du Niger. Edition Speciale. Strategie de Developpement accelere et de Reduction de la Pauvrete 2008–2012. IPCC. 2007 . Climate change 2007: The physical science basis. Contribution of Working Group I to the Forth Assessment Report Charasse, M, and Gouteyron, A. 2005. Examen en Commission of the Intergovernmental Panel on Climate Change. Cambridge, Sénatoriale, Rép. Française. UK: Cambridge University Press, 1009. CILSS. 2006. “The Ecologically Vulnerable Zones of the Sahelian IRAM. 2007. Les Politiques de Prévention et Gestion des Crises � In Atlas on Regional Integration in West Africa. Countries. Alimentaires: Enseignements de la crise du Niger de 2005. CILSS. 2008. Pro�l Securite Alimentaire: Niger. Kandji, S., Verchot, L., and Mackensen, J. 2006. Climate Change and CILSS. 2009. The Silent Transformation of Environment and Variability in the Sahel Region: Impacts and Adaptation Strategies Production Systems in the Sahel: Impacts of Public and Private in the Agricultural Sector. ICRAF and UNEP . Investments in Natural Resource Management. Ly, C., Fall, A., and Okike, I. 2010. “West Africa: The Livestock Sector Comité Interministériel de Pilotage de la Stratégie de Developpement in Need of Regional Strategies. � In Livestock in a Changing Rural. Septembre 2004. “Le Zonage Agro-écologique du Niger� . Landscape, vol. 2, eds. P . Gerber, H. Mooney, J. Dijkman, S. Secrétariat Exécutif. Tarawali, and C. de Haan. Island Press: Washington, DC. Cornia, G., Deotti, L, and Sassi, M. 2012. “Food Price Volatility over Marty, A., and Bonnet, B. 2006. Nord-Tahoua: Le Pastoralisme Survit the Last Decade in Niger and Malawi: Extent, Sources and Aux Changements. IRAM. Impact on Child Malnutrition. � UNDP Working Paper 2012-002. Michiels, D., Egg, J., et. al 2007. Ministère des Affaires étrangères et Fafchamps, M., and Gavian, S. “Spatial Integration of Livestock européennes. IRAM. � Journal of African Economies. 5(3):366–405. Markets in Niger. Ministère de l’Agriculture et de L’Elevage. Annuaire sur la Disponibilité FAO. Various Years. Desert Locust Situation Update and Desert en Semences Améliorées 2010–2011. Niger. Locus Information System. Ministere du Developpement Agricole/Ministere des Ressources FAO. 2004. Special Report: FAO/WP Crop and Food Supply Animales. Recensement General de l’Agriculture et du Cheptel Assessment Mission to Niger. 2005-2007 . 2007. FAO. 2007. AQUASTAT. Minister de Developpenent Rural. Strategie de Developpement Rural. FAO, Oxfam, AREN, CICR, et al. 2011. Opération de Déstockage au Mohamed Lemine Ould Ahmedou, FAO/EMPRES Seminar on desert Niger. locust control in West Africa, FAO, 2012. FEWS NET. Various Years. West Africa Food Security Outlook Programme d’Action National pour l’Adaptation. 2005. Identi�cation bulletins. et Evaluation des Phenomenes Extremes. TEC H N I C A L AS S I STANCE 46 B IBLIOGR A PH Y Redelsperger, Jean-Luc, Diedhiou A., Flamant C., Janicot, S. Lafore, Union des Cooperatives des Professionnels de la Filiere Oignon de J.-P., et al. 2006. La Météorologie [ISSN 0026-1181], 2006, Série Galmi. 2007 . Evaluation des Productions Oignon et l’Elaboration des 8, N° 54 ; p. 22-32 DOI:2042/20098 Strategies de Commercialisation Galmi. Rapport Final. BETIFOR. Republic of Niger, CNEDD, UNDP, and GEF. 2006. National Adaptation United Nations Environment Program. 2011. Livelihood Security: Programme of Action. Climate Change, Migration and Conflict in the Sahel. Reseau National des Chambres d’Agriculture du Niger. 2010. Note United Nations University. 2011. World Risk Report. d’Information/Intrants No. 13. � Special USAID. 2011. “Livelihoods Zoning “Plus� Activity in Niger. SADEC, Sarl., Etude pour une meilleure prise en charge des ris- Report by the Famine Early Earning Systems Network. ques majeurs liés aux changements climatiques par le système d’alerte précoce (SAP) du Niger. République du Niger, cabinet du World Bank. 2009. Developpement de l’Irrigation au Niger: Diagnostique Premier Ministre, 2011. et Options Strategique. Revue Sectorielle de l’Irrigation. Agriculture and Rural Development. Africa Region. Salla, A., Atte, I., and Oumarou, H. 2011. Evaluation Rapide de l’Impact de la Crise Pastorale 2009–2010 sur la De- World Bank. 2011. Niger: Investing for Prosperity—A Poverty capitalisation du Cheptel et les Moyens de Subsistence des Assessment. Poverty Reduction and Economic Management. Populations Pastotales et Agro-pastorales du Niger. Government Africa Region. of Niger. World Bank. 2010. Niger: Modernizing Trade During a Mining Boom: Sendzimir, J., Reij, C., and Magnuszewski, P . 2011. “Rebuilding Diagnostic Trade Integration Study for the Integrated Framework Resilience in the Sahel: Re-greening in the Maradi and Zinder Program. Poverty Reduction and Economic Management. Africa Regions of Niger. � Ecology and Society. 16(3):1. Region. Toulmin, C. 1986. Drought and the Farming Sector: Farm Animals World Food Program. 2010. Chocs et Vulnerabilité au Niger: Analyse and Post-Drought Rehabilitation. des Données Secondaires. Rapport Global. Trench, P., Rowley, J., Diarra, M., Sano, F., and Keita, B. 2007. Beyond You, L, Ringler, C., Nelson, G., Wood-Sichra, U., Robertson, R., Wood, Any Drought: Root Causes of Chronic Vulnerability in the Sahel. S., Zhe, G., Zhu, T., and Sun, Y. 20009. Torrents and Trickles: Sahel Working Group. Irrigation Spending Needs in Africa. World Bank. UNDP. 2007. Analyse Integree des Facteurs de Risques au Niger. Zakara, G., and Abarchi, H. 2007 . Assessment of the Impacts of Bureau d’Etudes Niger Horizons. Pastoral Policies in Niger. AREN. AGR ICULTURA L SEC TOR RISK ASS ES S MENT IN NIGER ANN E X 1 — D ISAST E R RISK RE DUCT ION AND CRISIS MA NAGEMENT 47 Annex 1: DISASTER RISK REDUCTION AND CRISIS MANAGEMENT: LESSONS LEARNED AND POSITIVE CHANGES SINCE 2005, AND REMAINING CHALLENGES ISSUES LESSONS LEARNED/PROGRESS SINCE 2005 REMAINING CHALLENGES Contribution of the Early • Greater harmonization of indicators for vulnerability (Early • No clear, common conceptual and institutional framework for disaster Warning and Response Warning System [EWS] of Niger, FEWS NET, WFP/VAM, CILSS, prevention and reduction System to reducing etc.) • EW analysis still addresses only some of the speci�c needs of major vulnerability by prevent- • Community-based EWS and response units (SCAP-RU) enabling actors (e.g., remains much more crisis response oriented than risk ing and managing a communities to take direct action and responsibility; method- management focused). food crisis and other ological guide available • Politics influence EWS analysis and its use. disasters • Identi�cation of vulnerable geographic zones and groups by the • SCAP-RU and OSV are still in a small fraction of departments. Observatories for Monitoring Vulnerability (OSV) at the village • Weak articulation between SCAP-RU and OSV and national EWS: and commune levels problems of aggregation, transmission, and consistency. • EWS data begins to come early enough to allow for analysis • Reliability of statistics and analysis for EWS is still weak. and generate an earlier response and support plan. • Remaining weaknesses in vulnerability analysis: • Continued monitoring of market prices, including Northern i. Livelihoods in pastoral zones and urban and peri-urban areas Nigeria ii. Nutritional dimension • Integrate access to, not just availability of, food. • Need to harmonize targeting criteria. • Inadequate communication of EWS data and support to communities to take local action (i.e., related to cereal banks) • Inadequate food stocks and reserves at national and community levels • Inadequate food security database, reference to a baseline year, for objective targeting Contribution of action • Regreening (FMNR) on a large scale, based on research in the • Basic research in areas related to productivity and resilience in the research and learning region, but spontaneously taken up by producers agriculture sector is not supported by the government. initiatives to an • Various models and techniques tested to buy animals and • New improved varieties, techniques, and methods come up through evidence base of safely process meat during the destocking programs in 2010 donor-funded localized projects, or as a result of the work of interna- innovative models • Use of Climate Vulnerability and Capacity Assessment (CVCA) tional research centers. for humanitarian or method. • There is very little capacity to extend research results to producers. development work • CARE-Niger’s Assistance in the Management of the National • Markets for improved varieties, breeds, fertilizer, equipment, and crop Prevention System for the Nutrition Crisis of Niger (APCAN) protection chemicals remain poorly developed. program piloted and developed an effective model of communi- • Access by producers to information about successful experiences ty-based EWS (SCAP-RU) and the commune-level Observatory remains limited and slow. for Monitoring Vulnerability (OSV) that engaged the government EWS (CC/SAP), and AGRHYMET. • CARE/UNDP/PANA initiated pilot projects to develop climate change adaptation models. • Widely disseminated cash transfer pilot project (2008) with positive impact on food security and malnutrition • Action research on pastoral wells; model accepted by the responsible Ministry. Reducing the vulner- • Greater awareness by pastoralists of the need to reduce risk • Development strategy documents place relatively little emphasis on: ability of pastoralists and safeguard their way of life i. Creation of positive synergies between agriculture and livestock • Development of agro-pastoralism and diversi�cation of herders’ production systems income sources ii. Marketing system for animals and animal products • Greater integration of agriculture and livestock systems in iii. Application of land tenure code in pastoral areas many areas iv. Risk reduction (e.g., production and conservation of fodder and • Preparation of the new Pastoral Code fodder banks). • Community-based early warning and disaster response (SCAP- • Access to basic social services in pastoral areas remains limited. RU) operations in certain pastoral areas. • Livestock sector activities are not well integrated with other sectors • Generally positive experiment with emergency destocking and do not adequately address gender issues. • Establishment of cereal banks in pastoral areas • Inadequate �nancial services are available to pastoralists. • Water catchment and land improvements in pastoral areas • Pastoral associations to become more representative and active in show how pastoral communities can help reduce risk. policy dialogue. TEC H N I C A L AS S I STANCE 48 A NNEX 1 — D ISASTER RISK RED UCTION A ND C R ISIS MANAGEM ENT ISSUES LESSONS LEARNED/PROGRESS SINCE 2005 REMAINING CHALLENGES Interagency • The two main national institutions responsible for aspects of 2004–05 Crisis: collaboration and risk management (CC/SAP and CNEDD) both report directly to • Insuf�cient of�cial communication during the crisis; the responsible coordination the Prime Minister’s Directeur de Cabinet. structure, CIC was not effective. • Important framework documents are now available, in • There was insuf�cient coordination between all actors at all levels particular: (e.g., CCA and CC/SAP; national, regional, departmental levels; various i. National Adaptation Programme of Action (2006) and Multi- actors using cash. risk Contingency Plan (updated annually) • NGOs did not all give information to the CCA for coordination and ii. Risk Management Plan against Grasshoppers and Locusts targeting. (2011). • Parallel systems and confused roles between the government and UN • NGOs have been admitted to participate at meetings of the agencies and tensions between UN agencies Comité Restreint de Concertation (CRC), chaired by the Prime Minister. More Recently: • Establishment of a framework of operational coordination • Rapid political changes and institutional instability reduce access by between NGOs, UN agencies, and the Red Cross all actors to a stable and predictable environment for policy dialogue, • Creation of an Emergency Capacity Building (ECB) network, joint strategic planning, and coordination of activities. which conducted a joint review of the 2004–05 humanitarian • Many tools and methods continue to be used without a strong multi- response sectoral and unifying conceptual framework or strategic direction. • Creation of the Niger Disaster Risk Reduction Consortium • Although both CC/SAP and CNEDD report to the Prime Minister’s • Of�ce for the Coordination of Humanitarian Affairs of the of�ce, disaster risk reduction and risk management are not part of a United Nations Secretariat (UN/OCHA) system of clusters single conceptual and institutional framework. (nutrition, food security, logistics, communications, and humani- tarian action plan) AGR ICULTURA L SEC TOR RISK ASS ES S MENT IN NIGER ANN E X 2 — R A I N FAL L ANALYSIS AND DROUGHT CLASS IFIC ATION 49 Annex 2: RAINFALL ANALYSIS AND DROUGHT CLASSIFICATION A meteorological drought is de�ned as precipitation below calculations are based on monthly precipitation data provided the expected value (generally the mean) during a given year. by the Directorate of Metrology, government of Niger. The A widely used concept to identify and classify droughts is data set contains monthly precipitation data from 40 weather based on the estimation of standardized anomalies because stations for a period of 1980–2009. These weather stations negative anomalies indicate relatively lower values of precipi- represent all 8 provinces and 34 departments of the country. tation; while positive anomalies, relatively higher values. The The time-series in this �le are associated with its correspond- formula to estimate standardized anomalies is the following: ing department and region based on the name. Whenever two or more time-series were available for the same depart- ment or region, the time-series were averaged. The resulting SP(t) – m SA (t) = time-series were tested for stationarity (i.e., constant mean). s Most time-series were nonstationary, so the mean was esti- where SA(t) represents the time-series of standardized mated as function of time. anomalies, SP(t) represents the cumulative precipitation during the rainy season (May to September), m represents Tables A2.1 and A2.4 contain standardized anomalies of pre- its mean, and s represents its standard deviation. The cipitation on regional and department levels, correspondingly. TABLE A2.1: Frequency of Adverse Rainfall Events by Region NUMBER OF ADVERSE RAINFALL YEAR AGADEZ DIFFA DOSSO MARADI NIAMEY TAHOUA TILLABERI ZINDER EVENTS SEVERE CATASTROPHIC TOTAL 1980 0.005 0.558 0.305 0.564 0.142 0.147 0.628 1.152 1981 0.384 −0.352 0.364 −0.208 0.085 0.067 −0.110 −0.567 1982 0.298 0.512 −0.467 0.184 −0.642 −0.091 −0.465 0.185 1 1 1983 −0.428 −0.514 0.542 −0.045 1.090 0.248 0.543 −0.284 1984 −0.858 −0.708 −1.423 −0.864 −0.883 −0.655 −0.821 −0.863 7 1 8 1985 0.274 0.814 0.434 0.299 0.261 0.100 0.209 0.617 1986 0.881 0.000 1.014 0.735 0.081 0.282 0.317 0.408 1987 −0.951 −0.972 −1.408 −1.285 −0.678 −0.585 −0.843 −1.380 2 5 7 1988 0.460 1.256 1.122 0.715 0.224 0.569 1.006 1.644 1989 0.275 −0.216 −0.453 0.455 0.759 0.289 −0.166 −0.235 1990 −1.216 −0.598 −0.897 −0.927 −0.407 −1.077 −0.661 −0.747 3 3 6 1991 1.114 0.376 1.574 0.587 −0.409 1.206 0.649 0.279 1992 0.042 −0.400 −0.270 0.539 0.830 −0.126 0.171 0.097 1993 −0.385 −0.565 −0.967 −1.224 −0.731 −1.435 −1.158 −0.740 2 4 6 1994 0.935 1.925 1.661 0.708 0.633 1.937 1.575 1.070 1995 −0.780 −0.657 −0.997 0.197 −0.101 −0.466 −0.056 0.390 2 1 3 TEC H N I C A L AS S I STANCE 50 A NNEX 2 — RA INFA LL A NA LYS IS A ND D ROUGH T C LAS S IFIC ATION TABLE A2.1: Continued NUMBER OF ADVERSE RAINFALL YEAR AGADEZ DIFFA DOSSO MARADI NIAMEY TAHOUA TILLABERI ZINDER EVENTS SEVERE CATASTROPHIC TOTAL 1996 −1.082 −0.080 0.029 −0.260 −0.310 −0.458 −0.580 −0.807 1 1 2 1997 0.521 −0.935 −0.445 −0.620 −1.424 −0.804 −1.347 −0.662 3 3 6 1998 0.833 0.299 0.503 0.096 2.338 0.671 1.311 0.875 1999 0.264 1.298 0.282 1.214 −0.070 1.149 0.742 0.071 2000 −0.774 −0.730 −0.398 −1.028 −1.022 −0.994 −0.779 −0.357 3 3 6 2001 0.287 −0.291 −0.157 0.603 0.595 0.112 −0.101 0.461 2002 −0.421 −0.405 −0.248 −0.787 −0.462 −0.234 −0.025 −0.843 2 2 2003 0.771 1.043 1.018 1.179 0.288 0.391 0.537 0.993 2004 −0.344 −0.912 −0.953 −0.942 −0.577 −0.592 −0.710 −0.720 2 3 5 2005 −0.427 0.471 −0.034 0.388 1.079 0.190 0.352 0.663 2006 −0.016 −0.460 0.563 −0.118 −0.715 0.260 0.047 −0.447 1 1 2007 0.885 0.920 −0.126 0.308 0.003 0.058 0.141 0.206 2008 −0.148 −0.226 0.163 0.066 0.112 0.491 0.079 0.101 2009 −0.555 −0.495 −0.254 −0.533 −0.029 −0.810 −0.420 −0.344 1 1 Sources: Rainfall data from INS. 2010. Annuaire Statistique des Cinquantes Ans d’Independence du Niger. Edition Speciale; Author’s calculations. a Calculated as time series of seasonal precipitation (SP(t)) minus the mean as a function of time (Mn(t)), divided by the standard deviation of SP; based on monthly rainfall data for the growing season (June–September). The standardized anomalies can be interpreted in terms of and 2004. It is important to highlight that 2009 was a se- the standard deviation. For example, an anomaly equal to vere drought year; however, due to aggregation effect, it is −1, for a given year, indicates that the precipitation was one not reflected in the table. The table below highlights that 10 standard deviation lower than the expected value (m). After departments of the country suffered severe drought and 2 inspection of the probability distribution functions of the experienced catastrophic drought in 2009, as opposed to 7 anomalies, it was decided to de�ne a catastrophic drought severe and 7 catastrophic in 2004. Also, although not part of (red cells in the tables) when the anomaly is lower than −0.9, the analysis, 2011 was also a severe drought year, leading to and a severe drought (orange cells in the tables) when the the food crisis of 2012. To sum up, at the national level, Niger anomaly is between −0.9 and −0.6.40 experienced nine severe droughts in the past 30 years. Analysis of these adverse rainfall events by year and region It is important to note that flagging adverse events solely on shows that most regions experienced severe rainfall drops the basis of average precipitation analysis (table A2.1) could in 2 to 3 years out of 10. Maradi and Zinder, the southern be arbitrary in some ways that it may miss years in which total agricultural belt of the country, experienced adverse rainfall rainfall was near normal, but due to factors such as late onset events more often than the drier northern regions. Dosso of rains, early cessation of rains, long dry spell between two and Maradi, the breadbasket of Niger, experienced more rains, and lack of rain during critical growth phase, it leads to catastrophic events than the other parts of the country, lead- signi�cant crop losses. Table A2.1 should be interpreted with ing to severe food shortages nationwide. The total number these caveats, and the need to corroborate these adverse of adverse rainfall events per year, a simple proxy for the events with data from other sources. presence of nationwide drought, suggests that widespread Furthermore, as indicated by the 2009 case, regional level drought occurred in 1984, 1987 , 1990, 1993, 1997, 2000, aggregation often masks the departmental level variation in precipitation and production. Detailed analysis at the depart- 40 Calculated as time series of seasonal precipitation (SP(t)) minus mental level might be more appropriate for understanding the mean as a function of time (Mn(t)), divided by the standard deviation of SP; based on monthly rainfall data for the growing drought and its impact on the agriculture sector. Furthermore, season (June–September). considering that production of certain agricultural commodities AGR ICULTURA L SEC TOR RISK ASS ES S MENT IN NIGER ANN E X 2 — R A I N FAL L ANALYSIS AND DROUGHT CLASS IFIC ATION 51 TABLE A2.2: Comparison of 2004 and 2009 Drought by is more concentrated in certain departments of the country, Department it is more prudent to analyze those departments and regions DEPARTMENT 2004 2009 that really matter for the agricultural sector’s performance. Out of the eight regions in Niger, �ve regions—Maradi, Arlit −0.241 −0.713 Tahoua, Tillabery, Dosso, and Zinder—are more signi�cant Bouza −0.131 −0.802 for rainfed crop production (table A2.3). C.U. Maradi −0.767 −0.041 Dakoro −1.424 −0.804 Table A2.4 provides a department level rainfall variability anal- Diffa −0.494 −0.649 ysis of 34 departments that corresponds to available weather data. It highlights that certain departments are more prone Dosso −0.359 −0.728 to weather variability. Between 1980–2009, there were nine Fillingué −1.202 −0.599 years, namely, 1984, 1987 , 1990, 1993, 1995, 1997 , 2000, Gaya −1.381 0.066 2004, and 2009, wherein 10 or more departments suffered Gouré −0.732 −0.408 severe or catastrophic drought. This analysis also reveals that Illéla −0.841 −0.985 the frequency as well as severity (in terms of area affected Konni −0.509 −1.100 by drought) is decreasing. Nonetheless, they remain the prin- ciple source of risk for the agriculture sector in Niger. Magaria −0.802 0.293 Matamèye −0.760 −0.554 Maïné Soroa −1.259 −0.215 TABLE A2.3: Niger: Regional Share of Major Crop N’Guigmi −0.414 −0.697 Production (2010) Ouallam −0.383 −0.771 COW GROUND Tahoua Arrondis −1.207 −0.339 MILLET SORGHUM PEA NUT SESAME Agadez 0 0 0 0 0 Tanout −0.168 −0.753 Diffa 2 1 1 1 0 Tchintabaraden −0.629 −0.016 Dosso 20 6 21 13 1 Tchirozérine −0.781 −0.586 Maradi 23 25 26 50 31 Tessaoua −1.032 −0.637 Tahoua 19 29 17 7 6 Tillabéry département −1.254 −0.623 Tillabery 20 10 16 3 12 Severe 7 10 Zinder 16 29 19 26 50 Catasthropic 7 2 Niamey 0 0 0 0 0 Total 14 12 Source: Directorate of Statistics, Ministry of Agriculture, Niger. Source: World Bank Staff estimates using rainfall data from Directorate of Metrology, government of Niger. TEC H N I C A L AS S I STANCE 52 A NNEX 2 — RA INFA LL A NA LYS IS A ND D ROUGH T C LAS S IFIC ATION TABLE A2.4: Frequency of Adverse Rainfall Events by Department TAHOUA REGION MARADI REGION TAHOUA TCHIN- C.U. YEAR BOUZA ILLÉLA KONNI KÉITA MADAOUA ARRONDIS TABARADEN MARADI DAKORO MAYAHI TESSAOUA 1980 0.181 0.734 −0.549 0.417 0.081 −0.206 0.758 0.318 0.926 −0.092 1981 −0.143 −0.244 0.058 0.225 0.619 −0.368 0.040 −0.237 −0.527 0.018 1982 0.246 −0.394 0.752 −0.629 −0.597 0.729 −0.491 0.281 0.322 0.441 1983 −0.045 0.302 0.193 0.238 0.193 −0.085 0.572 −0.037 0.166 −0.375 0.085 1984 −0.466 −0.360 −0.378 −1.525 −0.363 0.083 −1.062 −0.555 −0.769 −0.672 −0.790 1985 0.400 −0.641 0.366 0.618 0.574 −0.398 −0.298 0.009 0.354 0.499 0.127 1986 −0.473 1.104 0.213 −0.602 0.452 0.605 1.017 −0.138 0.543 0.890 1987 −0.022 −0.811 −0.630 −0.002 −0.276 −0.723 −1.324 −0.210 −1.375 −1.203 1988 0.565 0.264 0.992 0.229 −0.036 0.820 0.599 0.743 1.035 0.032 1989 0.161 0.441 −0.035 0.769 0.176 0.276 0.944 −0.586 0.543 0.517 1990 −1.395 −0.987 −1.127 −1.428 −0.561 −0.620 −1.035 −0.908 −0.947 −0.203 1991 1.603 1.349 1.114 1.363 −0.264 0.011 1.280 0.323 0.333 1992 −0.226 −0.009 −0.390 −0.515 0.713 1.081 0.178 0.022 0.453 1993 −1.347 −1.402 −0.668 −1.422 −0.807 −1.143 −0.568 −0.909 −1.282 1994 2.138 1.379 1.233 2.011 1.686 0.394 0.417 1.246 0.276 1995 −0.841 −0.430 −0.410 −0.486 −1.142 0.017 −0.527 0.180 0.839 1996 −0.634 −0.731 0.063 −0.208 −0.267 −0.323 −0.731 −0.158 0.690 −0.757 −0.536 1997 −0.511 −1.161 −0.880 −0.651 0.109 −0.844 −0.315 −0.146 −1.249 −0.457 −0.223 1998 0.468 0.817 0.320 0.963 −0.380 0.795 1.146 0.230 0.509 −0.269 −0.115 1999 1.181 0.926 1.047 0.289 1.434 0.819 0.719 0.414 1.182 1.899 0.517 2000 −1.047 −0.652 −0.292 −0.983 −1.560 −0.809 −0.998 −0.631 −0.692 −1.328 −0.695 2001 0.378 0.029 −0.673 1.014 0.351 0.096 0.322 0.846 −0.247 0.503 0.781 2002 −0.312 −0.416 −0.119 −0.441 0.167 0.069 −0.231 −1.070 −0.429 −0.497 −0.561 2003 0.356 0.840 0.168 0.043 0.135 0.405 0.163 0.879 1.304 0.580 1.049 2004 −0.131 −0.841 −0.509 0.272 −0.274 −1.207 −0.629 −0.767 −1.424 0.196 −1.032 2005 −0.460 0.351 0.305 −0.599 −0.080 0.918 0.559 0.249 0.365 −0.305 0.861 2006 0.396 −0.471 0.892 −0.244 0.683 −0.489 0.211 −0.016 0.048 −0.284 −0.121 2007 0.005 0.444 −0.361 0.929 0.247 −0.171 −0.514 1.012 0.470 0.391 −0.713 2008 0.624 0.805 0.621 −0.159 −0.531 0.733 0.324 −0.947 0.331 −0.121 0.834 2009 −0.802 −0.985 −1.100 −0.468 −0.286 −0.339 −0.016 −0.041 −0.804 −0.223 −0.637 Severe 3 5 2 2 3 2 5 2 3 2 4 Catastrophic 3 3 1 4 2 2 3 5 3 4 3 Total 6 8 3 6 5 4 8 7 6 6 7 AGR ICULTURA L SEC TOR RISK ASS ES S MENT IN NIGER DOSSO REGION TILLABERY REGION ZINDER REGION TILLABÉRY YEAR DOSSO DOUTCHI GAYA LOGA FILLINGUÉ KOLLO OUALLAM SAY DÉPARTEMENT TÉRA GOURÉ MAGARIA MATAMÈYE MIRRIAH TANOUT 1980 −0.130 0.419 0.419 0.163 0.171 0.799 0.507 0.656 1.174 1.229 0.924 0.598 0.983 1981 0.224 −0.729 0.873 0.531 −0.005 −0.142 −0.187 −0.393 −0.184 −0.314 −0.993 −0.521 −0.174 1982 0.312 0.122 −0.877 −0.813 −0.321 −0.643 −0.595 0.243 −0.529 −0.481 0.781 0.605 0.039 1983 0.055 0.740 0.009 0.932 0.592 0.652 1.133 −0.399 −0.158 −0.139 0.057 −0.438 −0.345 −0.655 TEC H N I C A L AS S I STANCE 1984 −1.501 −0.713 −1.109 −0.866 −0.702 −0.763 −1.186 0.258 −0.427 −1.006 −0.329 −0.722 −0.640 −1.245 1985 0.980 −0.176 0.714 −0.369 0.415 0.208 0.017 −0.351 0.607 0.444 0.135 0.648 0.328 1.579 1986 0.561 1.016 0.906 0.500 0.400 0.470 0.225 0.136 −0.290 0.511 0.100 0.643 0.475 −0.001 1987 −1.396 −0.851 −1.384 −0.444 −1.014 −0.508 −1.024 −0.139 −0.844 −0.487 −0.817 −1.227 −1.483 −1.324 −1.212 1988 0.706 0.411 0.880 1.318 0.706 0.945 0.578 0.985 0.740 0.778 0.530 2.135 1.381 1.402 1.660 1989 0.530 −0.753 −0.112 −1.110 −0.335 0.064 −0.435 −0.436 0.516 0.111 0.319 −0.761 −0.062 −0.305 −0.114 1990 −0.670 −0.828 0.045 −1.409 −0.739 −0.360 0.125 −0.315 −0.789 −0.875 −0.596 −0.901 −0.543 −0.157 −0.928 1991 0.300 2.603 −0.060 2.226 0.996 −0.209 0.385 0.253 0.630 0.605 0.611 0.641 −0.385 −0.198 0.145 1992 0.481 −0.682 −0.451 −0.117 0.604 −0.105 0.884 −0.667 −0.200 0.798 −0.324 0.451 0.364 −0.542 0.460 1993 −1.178 −1.239 0.135 −0.815 −1.710 −0.511 −1.048 −0.349 −0.230 −1.648 −0.837 −0.580 0.101 −0.518 1994 1.740 1.421 0.984 0.843 1.661 1.215 0.231 2.042 0.354 0.995 0.349 1.421 1.721 0.517 1995 −0.686 −0.980 −1.085 −0.125 −0.248 0.274 0.400 −0.622 0.163 0.643 0.585 −0.316 −0.306 0.156 ANN E X 2 — R A I N FAL L ANALYSIS AND DROUGHT CLASS IFIC ATION 1996 −0.328 0.091 0.501 −0.295 −0.589 −0.561 −0.212 −0.552 −0.355 −0.244 −0.493 −0.589 −1.211 −0.230 1997 −0.474 0.030 −0.596 −0.198 −0.503 −2.003 −0.889 −1.101 −1.277 −1.276 −0.250 −0.950 −0.280 −0.764 1998 0.631 −0.047 0.661 0.154 0.258 2.415 0.452 1.063 1.753 0.927 0.470 0.912 1.223 0.943 0.052 1999 −0.106 0.596 0.020 0.396 1.022 0.272 0.601 1.138 −0.296 0.029 0.444 −0.717 −0.007 −0.029 1.142 2000 −0.341 −0.124 −0.294 −0.418 −1.061 −0.448 −0.192 −0.766 −0.105 −0.636 −0.285 −0.468 −0.007 −0.383 −0.354 2001 0.305 −0.331 −0.369 −0.021 0.585 −0.793 −0.610 −0.826 0.487 0.291 0.479 0.660 0.061 0.516 0.033 2002 0.157 −0.322 −0.426 −0.093 −0.021 0.502 −0.091 0.370 −0.891 0.063 −0.705 −0.600 −0.961 −0.353 −1.134 2003 −0.429 0.876 2.112 0.138 0.604 −0.333 0.999 0.002 0.898 0.361 0.340 1.208 1.269 0.397 1.057 2004 −0.359 −0.398 −1.381 −0.517 −1.202 0.336 −0.383 −0.011 −1.254 −0.171 −0.732 −0.802 −0.760 −0.344 −0.168 2005 0.604 −0.665 −0.559 0.630 0.272 −0.579 −0.223 0.419 0.767 0.643 1.322 0.237 0.606 0.394 0.007 2006 −0.547 1.199 0.750 0.229 0.458 0.396 0.461 −0.240 −0.181 −0.913 −0.584 0.021 −0.218 −1.085 −0.222 2007 0.522 −0.733 0.328 −0.641 −0.491 0.442 −0.106 0.569 0.312 0.164 0.436 −0.222 −0.243 0.772 0.384 2008 0.566 −0.093 −0.497 0.657 0.734 −0.110 0.316 −0.740 0.270 −0.281 −0.247 −0.155 0.565 0.229 0.355 2009 −0.728 0.285 0.066 −0.419 −0.599 −0.419 −0.771 0.111 −0.623 0.405 −0.408 0.293 −0.554 −0.435 −0.753 Severe 3 8 1 4 2 1 5 5 4 2 3 5 2 1 3 Catastrophic 3 2 4 2 4 1 2 2 2 3 1 2 4 3 4 Total 6 10 5 6 6 2 7 7 6 5 4 7 6 4 7 Source: World Bank Staff estimates using rainfall data from Directorate of Metrology, government of Niger. 53 54 A NNEX 2 — RA INFA LL A NA LYS IS A ND D ROUGH T C LAS S IFIC ATION TABLE A2.5: Frequency of Adverse Rainfall Events (Drought) at Departmental Level DEPARTMENT NAME (NUMBER CATASTROPHIC DROUGHT OF DEPARTMENTS) SEVERE DROUGHT FREQUENCY FREQUENCY TOTAL DROUGHT FREQUENCY Kollo (1) 1 1 2/29 Bilma (1) 0 2 2/29 Konni (1) 2 1 3/29 Diffa, Gouré (2) 3 1 4/29 Tahoua Arrondis (1) 2 2 4/29 Mirriah (1) 1 3 4/29 Madaoua (1) 3 2 5/29 Téra (1) 2 3 5/29 Gaya (1) 1 4 5/29 Loga, Tillabéry (2) 4 2 6/29 Arlit, Bouza, Dakoro, Dosso (4) 3 3 6/29 Keita, Mayahi, Fillingué, Matamèye(4) 2 4 6/29 N’Guigmi (1) 7 0 7/29 Ouallam, Say, Magaria, Niamey (4) 5 2 7/29 Tessaoua (1) 4 3 7/29 Tanout (1) 3 4 7/29 C.U. Maradi (1) 2 5 7/29 Agadez, Tchirozeri, Illela, 5 3 8/29 Tchintabaraden, Maïné Soroa (5) Doutchi (1) 8 2 10/29 Source: World Bank staff estimates using rainfall data from Directorate of Metrology, government of Niger. By shifting from aggregate regional drought indices to departmental level. highlights the departments which have departmental level drought indices to de�ne national suffered more than six droughts years between 1980-2009. level drought, wherein 10 or more departments have suf- Out of 34 departments, 24 departments have suffered six fered droughts, improves the explanatory power of such or more than six drought in the past 29 years. Douthchi events. All these drought years corresponds to a sudden have demonstrated most pronounced variability in rainfall drop in cereal production and yields. There have been few and suffered 10 droughts in 29 years. On the other hand, years (1989, 1992, and 1999) wherein yield and produc- departments of Bilma, Konni and Kollo are most stable in tion declined, but which cannot be explained by droughts. terms of average precipitation and have suffered 2 or 3 Table A2.5 summarizes the frequency of droughts at drought years in 29. AGR ICULTURA L SEC TOR RISK ASS ES S MENT IN NIGER ANN E X 3 — L I V E L I HOOD Z ONE S AND P RINCIPAL RIS K S 55 Annex 3: LIVELIHOOD ZONES AND PRINCIPAL RISKS TOTAL ZONE DESCRIPTION POPULATION PRINCIPLE RISKS 1 Northeast Oases: 17,080 • Periodic civil insecurity leading to market disruption Dates, salt, and trade • Food price spike • Disease of date palm 2 Air Massif irrigated 287,019 • Flash floods damaging gardens (localized but frequent since 2007) gardening • Drought leading to lower water table (1 year in 5) • Food price hikes (periodic since 2005) • Civil insecurity disrupting market connection with rest of country (occasional) • Forced return of migrant workers from Maghreb countries, reducing remittances and putting pressure on local families (occasional) 3 Transhumance and 1,284,551 • Localized rain failure (1 year in 3) nomad pastoralism • Livestock epidemic disease (1 year in 10) • Bush �re (every year in dry season months) 4 Agropastoral belt 2,684,996 • Drought (1 year in 5) • Outbreaks of crop pests (yearly) • Outbreak of livestock diseases, notably blackleg, pasteurellosis, pox (la clavelle), and piroplasmosis (yearly) • Flash floods (yearly) • Price hikes (repeatedly in recent years) 5 Rainfed millet and 7,552,232 • Rainfall irregularities , especially in the grain-flowering stage (August–September) (frequent) sorghum belt • Flooding of �elds (minor, about 1 year in 3) • Crop pest (frequent) 6 Cropping and herd- 1,281,416 • Rain failure ing with high work • Market disruption due to local conflict outmigration • Price hikes (especially in lean season) 7 Southern irrigated 2,249,710 • Unusually bad insect attacks (August–September and February–March) cash crops • Rainfall failure/irregularities (June for rain-fed grain sowing; early September for flowering) • Flooding of irrigated �elds (July–August) • Fall in cash crop prices • Hike in prices (including imported grain affected by appreciation of the Nigerian currency [naira]) 8 Southwestern 284,561 • Heavy crop pest attack (1 year in 3) (July–September) cereals with Fan-palm • Erratic rain (1 year in 3 but relatively minor phenomenon) (late start May, irregularity June–July) products • Flooding (1 year in 5) (July–August) • Food price hikes for external reasons (poor hit in main buying months April–September) 9 Niger River irrigated 892,618 • Rainy reason: flooding (July–August) (1 year in 3) rice • Dry season: water breaching river banks (1 year in 5) (December–January) • Hippopotamus damage (yearly risk) (July–September) • Insect pests on rice (�rst harvest) (May–June) • Bird pests on rice (second harvest) (October–December) 10 Dallols-Seasonal 1,241,122 • Flooding (yearly) (July–August) water course irrigated • Irregularities in rainfall (sowing time May–June; flowering time August–September) crops • Unusually serious attacks of insect pests at the cereals flowering stage (August–September) • Bird and locust attacks (August–September) • Insect attacka on garden produce (February–March) • Food price hikes (periodic in past 7 years) TEC H N I C A L AS S I STANCE 56 A NNEX 3 — LIV ELIH OOD ZONES A ND PRINC IPA L R IS K S TOTAL ZONE DESCRIPTION POPULATION PRINCIPLE RISKS 11 Southeastern Natron 187,664 • Pest attacks on date palms (yearly) (November–December) salt and small basin • Rain failure (1 year in 3) (in terms of precipitation over whole season up to September) irrigated dates • Appreciation of the Nigerian naira affecting market grain prices (occasional) 12 Kamadougou irrigated 214,757 • Reduced river flow (1 year in 5) (March–April) peppers • Pepper disease (November–December) • Rain failure for rain-fed millet (June–September) 13 Lake Chad flood 91,989 • Low flood level of the lake by end of rains (1 year in 5) (August–October) retreat cultivation • Unusual severity of crop pest (1 year in 3) (December) with �shing • Unusual outbreak of animal diseases (August–September; March–May) � Special Report by the Famine Early Earning Systems Network. Source: USAID. 2011. “Livelihoods Zoning ‘Plus’ Activity in Niger. AGR ICULTURA L SEC TOR RISK ASS ES S MENT IN NIGER ANN E X 4 — G R A I N RE SE RVE S—OP VN 57 Annex 4: GRAIN RESERVES—OPVN In food-insecure countries, the management and use of a 110,000 metric tons after the 2004–05 crisis41 (80,000 physical grain reserve is a classic risk management strategy of the plus 30,000 metric ton equivalent �nancial), but the maximum coping type. Since the nation is vulnerable to wide fluctua- physical stock level ever reached was 40,000 metric tons. tions in rain-fed food production, Nigerien authorities have al- ways considered this area of public policy as a strategic one. TABLE A4.1: Jointly Managed OPVN Stocks and Funds The country’s main responsible institution OPVN (Of�ce NATIONAL RESERVE STOCK INTERVENTION FUND des Produits Vivriers du Niger, created in 1970), has closely (MAJOR NATIONAL FOOD (MITIGATING LOCALIZED FOOD CRISES) CRISES) matched the evolution of similar grain marketing boards in SNS (80,000 metric tons) Donor’s Common Fund (managed by the Sahel. Up to the early 1980s, most grain markets in the CMC) subregion were highly controlled. Marketing boards had Food Security Fund (30,000 metric Food Aid Counterpart Fund (bilateral, a strong legal power over domestic purchases, sales, and ton equivalent ) e.g. government of Niger, United States) cross-border trade flows. Of�cial prices were set, usually in Source: Authors. a uniform fashion at the national level, and over the entire marketing year. In spite of this, the private sector actually In addition to these jointly managed stocks and funds, OPVN handled most of the marketed cereals and storage. In the has established a Strategic Reserve �nanced from the na- 1980s, Sahel governments gradually liberalized grain markets tional budget. The maximum amount procured under the and shifted, to varying extents, to some type of integrated Strategic Reserve component was reportedly 60,000 metric food security reserve systems comprising three main com- tons in 2010. ponents: (1) food security information and early warning, (2) consultation and coordination with donors, and (3) safety net OPVN is based on a good model, and is reasonably well man- interventions. aged. However, as part of the DNPGCA, it �nds itself pulled between the complex workings of a system jointly managed OPVN boasts considerable experience and signi�cant assets by the government and key external partners, on the one (over 200 warehouses, a 155,000–metric ton storage capacity, hand, and the pursuit of strategic national interests, on the about 100 permanent staff, a fleet of heavy trucks, etc.). It is other. As a consequence, there have been a number of long- also a key component of the Dispositif National de Prévention standing problems, including: et de Gestion des Crises Alimentaires au Niger (DNPGCA), within which it relies on CC/SAP and SIM information to  Widespread dissatisfaction or disagreements about implement subsidized grain sales from February to June and, the quality of CC/SAP data and analysis on which most if necessary (and remaining stocks allowing), free distributions OPVN interventions are based. For example, the CC/ from July through September. Since most OPVN interven- SAP (and thus, OPVN) has been overly focused on tions take place within the framework of the DNPGCA, the food availability since cereal balance de�cits �gure management of the various OPVN stocks and funds involves prominently in vulnerability criteria. Seasonal crises the CMC (Commission mixte Etat-Donateurs), and its Comité drive most of the humanitarian processes without a Restreint de Concertation. Jointly managed OPVN stock and good understanding of how these relate to vulnerabil- funds include the national reserve stock, comprising a physi- ity and to chronic food insecurity even though there cal Stock National de Sécurité (SNS) and a Fonds de Sécurité Alimentaire. The Intervention Fund includes a Donors’ 41 Grain prices were particularly high in 2004. Some countries Common Fund and a Food Aid Counterpart Fund. The national placed limits on exports; even the WFP was unable to procure reserve stock target level has been raised from 80,000 to signi�cant amounts in the subregion. TEC H N I C A L AS S I STANCE 58 A NNEX 4 — GRA IN RES ERV ES — OPV N are vulnerable areas and groups in normal as well as in  The nationally funded Strategic Reserve is an bad years. For this reason, there is not enough empha- unquestionable sovereign choice, but it may sis on such nonfood and complementary approaches undermine the effectiveness of an already dif�cult as cash transfers and vouchers. joint DNPGCA process, at least as far as emergency  Claims by communes and local partners that some stocks are concerned. OPVN interventions have disrupted local initiatives (e.g., cereal banks, warehouse receipts [or warrantage]). In the �nal analysis, OPVN is essentially an instrument of the  Disagreements about the targeting and type of OPVN DNPGCA. Arguments about the proper size of emergency interventions between interested parties have re- stocks have attracted much attention, while in fact, squarely duced the levels of funding for grain procurement and addressing the Dispositif’s fundamental problems will be delayed the process. As a result, OPVN has had been more likely to resolve differences of opinion and help ensure constrained in its purchases, and these have virtually that the of�ce is used more effectively, that is, mostly when never taken place until 4 to 5 months after harvest, by and where subsidized sales and free distributions are truly which time prices have already increased substantially. the best response to a food insecurity crisis. AGR ICULTURA L SEC TOR RISK ASS ES S MENT IN NIGER ANN E X 5 — AG R H YME T ’S ANALYSIS OF CL IMAT E DETERMINA NTS ON C R OP Y IELD 59 Annex 5: AGRHYMET’S ANALYSIS OF CLIMATE DETERMINANTS ON CROP YIELD The economy of Niger is largely depending on agriculture, SQsPL (longest dry spell from onset to cessation date which is highly influenced by climate variability and change. longue de la saison agricole): the longuest number of The occurrence of repeated drought has continued to affect consecutive day without rainfall Niger resulting in catastrophic famine. Rainfall and rainy For synoptics stations: season characteristics (onset, cessation of rainy season, dry spells, length of growing season, etc.) variability have been Tx_MDFH: annual mean of maximal temperature from reported to have signi�cant effect on food productions. In onset to end of rainy season this context, the World Bank requested AGRHYMET, as a Tm_MDFH: annual mean of minimal temperature from specialized institution on climate and agriculture, to deepen onset to end of rainy season the understanding of the links between climate and agricul- Tmoy_DFH: annual mean of the mean of temperature tural production risk over the past 30 years. from onset to end of rainy season AGRHYMET was requested to investigate the links between The calculations of the agroclimatic parameters were done agro-climatic factors and agricultural production risk. This by the software Instat+ v3.36 from 1980 to 2010. Crop grain task involved the following: (1) calculation of the climatic fac- yield (millet, sorghum) are available from 1980 to 2010. The tor, (2) determination of the links between this factor and variables were generated using the software Instat+ v3.36. the crop (particularly for millet), and (3) usage of the princi- pal components analysis to determine the two main influ- ences on yields for each agro-climatic region for the period 1980–2010. TABLE A5.1: Weather Stations Used for the Different Analysis was conducted on the following variables: Agro-climatic Zones AGRO- Cumul: sum of rainfall from onset (start of rainy season) CLIMATIC ZONE NUMBER WEATHER STATIONS to the end of the rainy season Sahelo-saharian 10 Ayerou, Dakoro, Ourafane, Tanout, début_S (onset date): the �rst occasion with more than zone (150–300 mm) Gouré Diffa, Chétimarie, Goudoumaria, Gueskerou, N’guiguimi 20 millimeters in 1 or 2 or 3 consecutive days after May 1 Sahelian zone 34 Tillabery,a Dolbel , �lingué, Gothéye, (300–600 mm) Kollo, Niamey,a Oullam, Say, Tera, Torodi, �n_S (cessation date or end of rainy season): the �rst Toukounous, Dosso, Birni N’gaouré, day after September 1 that the water balance drops Dogondoutchi, Dogonkiria, Tahoua,a Bouza, Konni,a Illéla, Keita , Madaoua, Tamaské, to zero, soil water capacity of 60 millimeters per Maradi,a Chadakori, Gazaoua, Guidan meter, and 5 millimeters per day of water losses due Roumdji, Madarounfa, Mayahi, Tessaoua, Zinder,a Guidimouni, Magaria, Myrriah, to evaporation and transpiration was considered Mainé Soroaa long_S (length of rainy season): difference between the Sahelo-soudanian 1 Gayaa end and the onset dates zone (> 600 mm) NJP (number of rain days): rain day is a day where the a Synoptic weather stations. Source: AGRHYMET. sum of rainfall is more than 1 millimeter TEC H N I C A L AS S I STANCE 60 ANNE X 5 — AGR H Y MET’S A NA LYS IS OF CLIMATE D ETER MINA NTS ON C R OP Y IELD FIGURE A5.1: Agro-Climatic Zones 24 22 Zone Sahélo Soundanienne 600 20 Zone Sahélienne 300 18 latitude Zone Sahélo Saharienne 150 16 Zone Saharienne 14 0 12 0 2 4 6 8 10 12 14 16 longitude Source: Comité Interministérial de Pilotage de la Stratégie de Developpement Rural; Secrétariat Exécutif. September 2004. Le Zonage Agro-écologique du Niger. A5.1 RELATIONSHIP BETWEEN GRAIN YIELD AND THE CLIMATIC FACTORS A5.1.1 Case of Millet TABLE A5.2: Correlation between Millet Grain Yield and Rainy Season Variables SOUDANO-SAHELIAN ZONE LINK (CORRELATION) BETWEEN AGRO-CLIMATIC FACTOR AND MILLET GRAIN YIELD (1990–2010) Stations Cumul début_S �n_S long_S NJP SQsPL Tx_MDFH Tm_MDFH Tmoy_DFH Gaya 0.38 0.31 According to the statistical test of student, only the coef�cient of correlation more than 0.34 are signi�cant (risk of error of 5 percent). SAHELIAN ZONE LINK (CORRELATION) BETWEEN AGRO-CLIMATIC FACTOR AND MILLET GRAIN YIELD (1990–2010) STATIONS CUMUL DÉBUT_S FIN_S LONG_S NJP SQsPL TX_MDFH TM_MDFH TMOY_DFH Tillabery 0.38 Filingue 0.49 0.49 Kollo 0.51 0.34 0.36 0.40 AGR ICULTURA L SEC TOR RISK ASS ES S MENT IN NIGER ANN E X 5 — AG R H YME T ’S ANALYSIS OF CL IMAT E DETERMINA NTS ON C R OP Y IELD 61 SAHELIAN ZONE LINK (CORRELATION) BETWEEN AGRO-CLIMATIC FACTOR AND MILLET GRAIN YIELD (1990–2010) STATIONS CUMUL DÉBUT_S FIN_S LONG_S NJP SQsPL TX_MDFH TM_MDFH TMOY_DFH Ouallam 0.40 0.38 Say 0.34 −0.42 Dosso 0.44 Birni Gaouré 0.42 0.3 Dogon −0.33 Doutchi Tahoua 0.49 Birni Konni 0.46 0.38 0.44 Illela 0.41 0.45 Keita 0.47 0.46 0.40 Madoua 0.46 0.436 Guidan Roumdji Madarounfa 0.37 0.41 Magaria 0.33 Myrriah 0.30 Maine Sora 0.39 0.42 0.67 SAHELO-SAHARIAN ZONE LINK (CORRELATION) BETWEEN AGRO-CLIMATIC FACTOR AND MILLET GRAIN YIELD (1990–2010) STATIONS CUMUL DÉBUT_S FIN_S LONG_S NJP SQsPL TX_MDFH TM_MDFH TMOY_DFH Nguigmi 0.34 0.36 −0.35 Diffa 0.63 0.36 0.64 −0.31 Dakoro 0.45 0.33 0.44 −0.40 Gouré 0.33 0.33 0.44 Tanout 0.52 0.44 −0.43 Source: AGRHYMET. A5.1.2 Case of Sorghum TABLE A5.3: Correlation between Sorghum Grain Yield and Rainy Season Variables SAHELIAN ZONE LINK (CORRELATION) BETWEEN AGRO-CLIMATIC FACTOR AND SORGHUM GRAIN YIELD (FROM 1990–2010) STATIONS CUMUL DÉBUT_S FIN_S LONG_S NJP SQsPL TX_MDFH TM_MDFH TMOY_DFH Tillabery Filingue Kollo 0.28 0.32 Ouallam Say 0.27 0.26 TEC H N I C A L AS S I STANCE 62 ANNE X 5 — AGR H Y MET’S A NA LYS IS OF CLIMATE D ETER MINA NTS ON C R OP Y IELD SAHELIAN ZONE LINK (CORRELATION) BETWEEN AGRO-CLIMATIC FACTOR AND SORGHUM GRAIN YIELD (FROM 1990–2010) STATIONS CUMUL DÉBUT_S FIN_S LONG_S NJP SQsPL TX_MDFH TM_MDFH TMOY_DFH Dosso Birni Gaouré 0.25 0.36 Dogon 0.29 0.35 0.34 Doutchi Tahoua Birni Konni 0.35 Illela 0.38 Keita Madoua 0.34 −0.29 0.38 Guidan Roumdji Madarounfa 0.49 0.38 0.48 Magaria −0.32 Myrriah Maine Sora SAHELO-SAHARIAN ZONE LINK (CORRELATION) BETWEEN AGRO-CLIMATIC FACTOR AND SORGHUM GRAIN YIELD (FROM 1990–2010) STATIONS CUMUL DÉBUT_S FIN_S LONG_S NJP SQsPL TX_MDFH TM_MDFH TMOY_DFH Nguigmi 0.35 0.35 0.36 –0.36 Diffa 0.64 −0.31 0.39 0.67 –0.31 Dakoro 0.47 0.32 0.42 Gouré 0.33 Tanout 0.48 0.4 –0.35 According to the statistical test of student, only the coef�cient of correlation more than 0.34 are signi�cant (risk of error of 5 percrent). Source: AGRHYMET. In general, when the link between climatic factor and millet and sorghum yields is noted, grain yield is positively linked with the sum of rainfall (cumul), the number of rain days (NJP), and the length of the growing period (Long_S) and negatively cor- related with the longest dry spell (SQsPL). A5.2 PRINCIPAL COMPONENT ANALYSIS TABLE A5.4: Percentage of Variation Explained by the Principal Axes (F1 and F2) SOUDANO-SAHELIAN ZONE PRINCIPAL COMPONENT ANALYSIS TOTAL VARIANCE EXPLAINED BY STATION AXIS 1 (F1—HORIZONTAL AXIS) AXIS 2 (F2—VERTICAL AXIS) AXIS 1 AND 2 Gaya 32.61 25.54 58.6 AGR ICULTURA L SEC TOR RISK ASS ES S MENT IN NIGER ANN E X 5 — AG R H YME T ’S ANALYSIS OF CL IMAT E DETERMINA NTS ON C R OP Y IELD 63 SAHELIAN ZONE PRINCIPAL COMPONENT ANALYSIS TOTAL VARIANCE EXPLAINED BY STATIONS AXIS 1 (F1) AXIS 2 (F2) AXIS 1 AND 2 Tillabery 40.12 23.04 63.1 Filingue 38.8 22.21 61.01 Kollo 36.89 23.38 60.26 Ouallam 42.04 24 66.38 Say 40.42 17 57.42 Dosso 36.65 22.20 58.85 Birni Gaouré 39.7 31 65.3 Dogon Doutchi 34.21 31 65.3 Tahoua 43.3 22.98 67.29 Birni Konni 43.35 27.35 70.7 Illela 46.36 22.7 69.11 Keita 45.75 21.13 66.7 Madoua 36.6 29.5 66.16 Guidan Roumdji 32.7 28.08 60.78 Madarounfa 42.57 35.53 78.1 Magaria 43.88 21.16 65.04 Myrriah 42.57 23.53 66.1 Maine Sora 39.13 23.31 62.44 SAHELO-SAHARIAN ZONE PRINCIPAL COMPONENT ANALYSIS TOTAL VARIANCE EXPLAINED BY STATIONS AXIS 1 (F1) AXIS 2 (F2) AXIS 1 AND 2 Nguigmi 56.01 21 77.01 Diffa 52.23 18.92 71.15 Dakoro 43.95 27.77 71.72 Gouré 37.94 25.65 63.59 Tanout 49.44 19.8 69.24 Source: AGRHYMET. For all the stations, the part of information explained by axis In general, the axis 1 opposed two groups of variables. The 1 (F1) and axis 2 (F2) is between 58 and 77 percent. The high result shows that good grain yield corresponds to a rainy scores are recorded in the Sahelo-Saharian zone. season with good rainfall conditions and more rain days. However, years with low grain yield correspond to a long dry The graph of the PCA opposes the variables that form groups spell and a late-onset date. (group 1 and 2) on axis 1 and axis 2. The variables of the same group are positively correlated. While those of the opposing group are negatively correlated. Figure A5.2 shows that A5.3 TRENDS OF RAINFALL AND RAINY SEASON PCA on axis 1 opposes two groups: Group 1 is comprised VARIABLES of performance millet and sorghum yield (rdtmil, rdtsorgho), the cumulative rainfall (cumul), number of days of rain (NJP), A5.3.1 Context and the end date of the rainy season (�n_S). Group 2 is the West Africa, particularly Niger, is one of the areas in the longest dry sequence (SQsPl). world that has had signi�cant climate anomalies in the past TEC H N I C A L AS S I STANCE 64 ANNE X 5 — AGR H Y MET’S A NA LYS IS OF CLIMATE D ETER MINA NTS ON C R OP Y IELD FIGURE A5.2: Principal Component Analysis of Dakoro (1980–2010) variables (axes F1 et F2 : 71.72%) 1 long_S 0.75 0.5 Cumul 0.25 fin_S NJP .77 %) 0 F2 (27 rdtsorgho – 0.25 SQsPL rdtmil – 0.5 – 0.75 debut_S –1 –1 – 0.75 – 0.5 – 0.25 0 0.25 0.5 0.75 1 F1 (43.95 %) Source: AGRHYMET. century. The change from wet conditions in the 1950s to The main results show (table 4) that from 1980 to 2010, in the much drier conditions in the 1970s and 1980s represents Sahelian and Sahelo-Saharian zones: one of the strongest inter-decadal signals on the planet in the twentieth century (Redelsperger, et al. 2006 ). The drought in  More than half of the stations recorded a signi�cant this area since the late 1970s is the most severe and longest increase in rainfall. at continental scale in the world during that century (IPCC,  More than 33 percent of stations have increased 2007). The reduction is extremely clear in the Sahel with high signi�cantly the number of rain days, with a signi�cant de�cit periods in 1972–73 and 1982–84. Since the mid-1990s, decrease of the duration of the longest dry spell and a return to better rainfall conditions has been noted (Ali and cessation dates of increasingly late. Lebel, 2009 ), mostly in the eastern part of Sahel such as There is no trend in the Soudano-Sahelian zone (Gaya). Niger. In fact, apart from the more or less severe droughts of 2004 and 2011, Niger has experienced rainfall conditions Across Niger, the minimum temperatures have increased by much better over the past two decades compared to the dry about 1°C and maximum of 0.3 to 0.5°C. Figure A5.8 shows decades of the 1970s and 1980s. that the minimum temperatures at Birni Konni increased from 21.3 to 22.3°C over 80 years. AGRHYMET analysis shows that yields of crops such as millet A5.3.2 Trend Analysis and sorghum will fall by more than 10 percent in Niger in the To determine trend on the rainfall and rainy season variables case of higher temperatures +2°C and insigni�cant variations from 1980 to 2010, calculate �rst Kendall’s tau to verify the in precipitation in 2050. However millet and sorghum plants, as trend and then calculate the probability of the existence of heat crops, would not be vulnerable to temperature increases the trend at the 5 percent signi�cance levels and Sen’s slope of about 1°C. These trends noted on temperatures probably estimator using the Mann-Kendall test and XLSTAT software explain the weak link between the current rise in tempera- 10 (trend signi�cant at p < 0.05). tures and grain yields of millet and sorghum from 1980–2010. AGR ICULTURA L SEC TOR RISK ASS ES S MENT IN NIGER ANN E X 5 — AG R H YME T ’S ANALYSIS OF CL IMAT E DETERMINA NTS ON C R OP Y IELD 65 TABLE A5.5: Signi�cant Trend of the Variables (1980–2010) SOUDANO-SAHELIAN ZONE VARIABLES CUMUL DÉBUT_S FIN_S LONG_S NJP SQSPL SEN’S SEN’S SEN’S SEN’S SEN’S SEN’S STATION P-VALUE SLOPE P-VALUE SLOPE P-VALUE SLOPE P-VALUE SLOPE P-VALUE SLOPE P-VALUE SLOPE Gaya No signi�cant trend is noticed. SAHELIAN ZONE VARIABLES CUMUL DÉBUT_S FIN_S LONG_S NJP SQSPL SEN’S SEN’S SEN’S SEN’S SEN’S SEN’S STATIONS P-VALUE SLOPE P-VALUE SLOPE P-VALUE SLOPE P-VALUE SLOPE P-VALUE SLOPE P-VALUE SLOPE Tillabery Filingue 0.04 3.62 0.043 1 0.011 −0.25 Kollo 0.049 0.9 Niamey 0.004 0.59 Ouallam 0.032 4.87 0.019 0.294 Say Tera 0.007 0.26 Birni Gaouré 0.006 6.32 0.004 0.857 Tahoua 0.006 4.34 0.026 0.24 0.0019 −0.25 Illela 0.040 4.89 0.02 −0.25 Keita 0.026 4.478 0.014 0.4 0.004 0.316 Madoua 0.001 8.26 0.018 0.455 0.000 0.389 Madarounfa Zinder Myrriah 0.041 3.82 0.005 0.364 0.048 0.2 0.02 0.2 Mainé Soara 0.01 4.967 0.024 −0.909 0.01 0.333 0.001 1.444 SAHELO-SAHARIAN ZONE VARIABLES CUMUL DÉBUT_S FIN_S LONG_S NJP SQSPL SEN’S SEN’S SEN’S SEN’S SEN’S SEN’S STATIONS P-VALUE SLOPE P-VALUE SLOPE P-VALUE SLOPE P-VALUE SLOPE P-VALUE SLOPE P-VALUE SLOPE Dakora 0.004 6.64 0.001 0.33 0.001 −0.6 Tanout 0.013 −1.1 0.008 1.31 0.03 −0.5 Gouré 0.02 5.161 0.008 0.15 Diffa 0.007 4.51 0.006 0.3 0.007 −0.4 Nguigmi Signi�cant at p <0.05. Source: AGRHYMET. Figures A5.3–A5.7 show some trends of rainfall and rainy season parameters. TEC H N I C A L AS S I STANCE 66 ANNE X 5 — AGR H Y MET’S A NA LYS IS OF CLIMATE D ETER MINA NTS ON C R OP Y IELD FIGURE A5.3: Trend of Total Annual Rainfall at Gaya (Soudano-Sahelian Zone) 1300 1100 900 rainfall (mm/yr) 700 500 300 100 19 0 19 1 19 2 19 3 19 4 19 5 19 6 19 7 88 19 9 19 0 19 1 19 2 93 19 4 19 5 19 6 19 7 19 8 20 9 20 0 20 1 20 2 20 3 20 4 20 5 06 20 7 20 8 20 9 20 0 11 8 8 8 8 8 8 8 8 8 9 9 9 9 9 9 9 9 9 0 0 0 0 0 0 0 0 0 1 19 19 19 20 year p > 0.05, no signi�cant trend. Source: AGRHYMET. FIGURE A5.4: Trend of Number of Rain Days per Year at Gaya (Soudano-Sahelian Zone) 70 65 60 number 55 50 45 40 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 year p > 0.05, no signi�cant trend. Source: AGRHYMET. To manage the late onset, shortening of LGP, long dry spells,  Shifting of sowing date and water stress, a wide range of adaptive actions can be  Modifying crop calendar implemented to reduce the adverse effect of climate variabil-  Use of short crop duration ity. In the short term, the mentioned adaptation strategies  Improved seed can be explored by:  Improved water ef�ciency AGR ICULTURA L SEC TOR RISK ASS ES S MENT IN NIGER ANN E X 5 — AG R H YME T ’S ANALYSIS OF CL IMAT E DETERMINA NTS ON C R OP Y IELD 67 FIGURE A5.5: Trend of Total Annual Rainfall at Filingue (Sahelian Zone) 700 y = 4.704x + 274.6 R2 = 0.169 600 500 rainfall (mm/yr) 400 300 200 100 0 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 year p < 0.05, signi�cant trend. Source: AGRHYMET. FIGURE A5.6: Trend of Duration of Rainy Season at Filingue (Sahelian Zone) 140 y = 0.858x + 67.58 R2 = 0.154 120 duration of rainy season (days) 100 80 60 40 20 0 19 0 19 1 19 2 19 3 19 4 19 5 19 6 19 7 88 19 9 19 0 19 1 19 2 93 19 4 19 5 19 6 19 7 19 8 20 9 20 0 20 1 20 2 20 3 20 4 20 5 06 20 7 20 8 20 9 10 8 8 8 8 8 8 8 8 8 9 9 9 9 9 9 9 9 9 0 0 0 0 0 0 0 0 0 19 19 19 20 year p < 0.05, signi�cant trend. Source: AGRHYMET. TEC H N I C A L AS S I STANCE 68 ANNE X 5 — AGR H Y MET’S A NA LYS IS OF CLIMATE D ETER MINA NTS ON C R OP Y IELD FIGURE A5.7: Trend of the Longest Dry Spell at Filingue (Sahelian Zone) 40 35 y = 0.231x + 19.10 30 R2 = 0.099 dry spell duration (days) 25 20 15 10 5 0 19 0 19 1 19 2 19 3 19 4 85 19 6 19 7 88 19 9 19 0 19 1 19 2 19 3 19 4 19 5 19 6 19 7 19 8 20 9 20 0 20 1 20 2 20 3 20 4 20 5 06 20 7 20 8 20 9 10 8 8 8 8 8 8 8 8 9 9 9 9 9 9 9 9 9 9 0 0 0 0 0 0 0 0 0 19 19 19 20 year p < 0.05, signi�cant trend. Source: AGRHYMET. FIGURE A5.8: Yearly Development of Minimum Temperatures Observed at Birni Koni 23,5 23,0 22.3 22,5 22,0 21.3 temperature (°C) 21,5 21,0 20,5 20,0 19,5 19,0 18,5 61 63 65 67 69 71 73 75 77 79 81 83 85 87 89 91 93 95 97 99 01 03 05 07 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 19 20 20 20 20 year Statistical test of Pettit shows mean change on minimal temperature since the 80 years. Source: AGRHYMET.  Promoting soil and water conservation measures In the long term:  Developing water harvesting techniques (e.g., use of harvesting water for supplemental irrigation during  Develop drought and heat-tolerant varieties of Millet rainy season must be explored) and Sorghum AGR ICULTURA L SEC TOR RISK ASS ES S MENT IN NIGER 1818 H Street, NW Washington, D.C. 20433 USA Telephone: 202-477-1000 Internet: www.worldbank.org/ard