A FEASIBILITY STUDY ASSESSING THE POTENTIAL FOR LARGE-SCALE AGRICULTURAL CROP AND LIVESTOCK INSURANCE IN PUNJAB PROVINCE, PAKISTAN JUNE 2018 A FEASIBILITY STUDY ASSESSING THE POTENTIAL FOR LARGE-SCALE AGRICULTURAL CROP AND LIVESTOCK INSURANCE IN PUNJAB PROVINCE, PAKISTAN June 2018 CONTENTS Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi Acronyms and Abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii Executive Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv Chapter 1 Introduction and Objectives of the Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.1. The Importance of Agriculture and Agricultural Growth in Punjab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.2. Climatic Risks to Crop and Livestock Production in Punjab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 1.3. Agricultural Insurance for Crops and Livestock in Pakistan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 1.4. The SMART Punjab Program to Transform Agriculture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 1.5. Government of Punjab Request to the World Bank Group for Technical Assistance . . . . . . . . . . . . . . . . . . . . . . . . . 4 1.6. Scope and Objectives of This Feasibility Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 1.7. Organization of This Report . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 Chapter 2 Key Features of Agriculture in Punjab and the Agricultural Impacts of Climatic and Natural Disasters . . . . . . 7 2.1. A Densely Populated Province Where Small Farms Predominate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 2.2. Crop and Livestock Production in Punjab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 2.3. Access to Agricultural Credit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 2.4. Exposure of Agriculture to Climatic and Natural Disasters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 Chapter 3 Agricultural Insurance Provision and Natural Disaster Relief Programs in Punjab . . . . . . . . . . . . . . . . . . . . . . . . 23 3.1. Crop Insurance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .23 3.2. Livestock Insurance Scheme for Borrowers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 3.3. Innovations in Crop and Livestock Insurance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 3.4. National and Provincial Disaster Management Programs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 Chapter 4 Agricultural Crop and Livestock Insurance Opportunities for Punjab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 4.1. Crop Insurance Types, Opportunities, and Data Needs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 4.2. Building on the Crop Loan Insurance Scheme . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 4.3. International Experience with Crop Area Yield Index Insurance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 4.4. Area-Yield Index Insurance for Semicommercial/Progressive Farmers in Punjab . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 4.5. Crop AYII Insurance for Small Subsistence Farmers in Punjab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 4.6. Crop Insurance for Cash Crops (Fruits and Vegetables) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 4.7. Livestock Insurance Opportunities for Punjab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 Chapter 5 Legal, Institutional, and Operational Considerations for a Punjab Agricultural Insurance Program . . . . . . . . . 65 5.1. Legal and Regulatory Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 5.2. Institutional and Organizational Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 5.3. Operational Considerations for Commercial Crop Insurance for Semicommercial/Progressive Farmers . . . . . . 66 5.4. Key Roles of Public-Private Partnership Players . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69 Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan iii Chapter 6  A Five-Year Plan and Budget for Building and Scaling Up a Large-Scale Crop Insurance Program in Punjab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 6.1. A Five-Year Build-up Plan, Portfolio Projections, and Financial Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 6.2. Costs to Government of Crop Insurance Premium Financing and Other Program Operating Costs . . . . . . . . . . 77 6.3. Crop Insurance Sensitivity Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 Chapter 7  Launch of Punjab Agricultural Crop Insurance Program in Kharif Season 2018: Planning Considerations and Implementation Challenges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 7.1. Steps and Timetable for Launching a Crop Insurance Pilot in Kharif 2018 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 7.2. Key Lessons and Challenges Experienced in Rolling Out Crop AYII in Kharif 2018 . . . . . . . . . . . . . . . . . . . . . . . . . . 89 Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93 Annex 1 Area, Production, and Yields of Major Crops in Punjab, Pakistan, 2006/07–2015/16 . . . . . . . . . . . . . . . . . . . . . . . . 95 Annex 2 Pakistan CRED-EM-DAT Database for Natural Disasters, 1900–2017 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .105 Annex 3 Compensation Paid to Flood Victims in Punjab, 2010–15 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 Annex 4 International Experience with Crop Area-Yield Index Insurance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109 Annex 5 Results of Preliminary AYII Coverage and Rating Analysis for Lodhran District, Punjab . . . . . . . . . . . . . . . . . . . . . 115 Annex 6 The CADENA Program in Mexico: State-Level Catastrophe Insurance as a Safety Net for Smallholders . . . . . 129 Annex 7 Examples of Agricultural Insurance Pool Programs in Spain and Turkey . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139 Annex 8 Possible Options for Coinsurance Pools in Punjab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145 Annex 9  Details Regarding Number of Insured Farmers, Insured Area, Sum Insured and Premium, under Alternative Scenarios . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149 TABLES Table ES1: Crop Insurance Products and Suitability for Punjab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xix Table ES2: Proposed Phased Introduction of New Crop Insurance Products and Programs in Punjab . . . . . . . . . . . . . . . . . xxi Table ES3:  Punjab Five-Year Crop Insurance Portfolio Showing Estimated Number Insured Farmers, Insured Crop Area, Sum Insured and Premium Income (US$) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxiv Table ES4:  Indicative Costs of GoPunjab Premium Subsidy Support and Other Financial Support to Crop Insurance Programs 2018/19 to 2022/23 (US$) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxv Table 2.1: Punjab: Number and Area of Farms by Size of Farm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 Table 2.2: Pakistan: Agricultural Growth Percentages (base = 2005–06) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 Table 2.3: Punjab: Livestock and Poultry Numbers, 1960–2006 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 Table 2.4: Pakistan: Supply of Agricultural Credit by Lending Institutions (PKR billion) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 Table 2.5:  Agricultural Loan Rescheduling: Number of Branches Rescheduling by Province and Value of Rescheduled Loans (2007–16) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 Table 2.6: Punjab: Losses and Damages Caused by Floods, 2010–13 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 Table 2.7: Punjab: Estimated Value of Flood Damage to Agriculture 2010–13 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 Table 3.1: SBP Task Force: Crop Loan Insurance Framework . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 iv A Feasibility Study Table 3.2:  CLIS: Number of Insured Borrowers, Premium, and Average Premium per Loanee per Season and Year, 2008/09–2015/16 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 Table 3.3: CLIS: Increased Claims Costs for 400 Percent or 500 Percent Loss Ratio Cap . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 Table 3.4: Framework of Livestock Insurance Scheme for Borrowers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 Table 3.5: Livestock Insurance Scheme for Borrowers: Underwriting Results, 2014–16 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 Table 3.6: Bangladesh: Results of Five Formal and Informal Livestock (cattle) Insurance Programs . . . . . . . . . . . . . . . . . . 31 Table 3.7: Pakistan: Seven Options for a Disaster Risk Financing Strategy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 Table 3.8:  Punjab: Payments (PKR) to Districts to Compensate Flood Victims for Loss of Livelihoods, Housing, and Crops, 2010–15 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 Table 4.1: Types of Crop Insurance Product Available and Potential Suitability for Punjab . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 Table 4.2: Preconditions, Advantages, and Disadvantages of Area-Yield Index Insurance . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 Table 4.3:  India: Insured Farmers, Average Premium Rates, and Average Loss Ratios by State and Season for the Area-Yield Index Insurance Program of the Modified National Agricultural Insurance Scheme, 2001–14 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 Table 4.4:  Lodhran District, Punjab: Average Cultivated Area of Main Crops in Kehror Placa, Dunyapur, and Lodhran Tehsils . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 Table 4.5: Punjab: Comparison of Methods for Calculating Average or Expected Yields for Sample Tehsils . . . . . . . . . . . 52 Table 4.6: Dunyapur Tehsil, Punjab: Levels of Insured Yield Coverage for Maize (kg/acre) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53 Table 4.7: Crop Loan Insurance Scheme, Pakistan: Limits on Sum Insured per Acre (maximum loan size) . . . . . . . . . . . . . 54 Table 4.8:  Dunyapur Tehsil, Punjab: Historical Burning Cost Rating Analysis for Ground-up AYII Cover for Actual Maize Yields from 2007–08 to 2016–17 (kg/acre) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 Table 4.9:  Dunyapur Tehsil, Punjab: Historical Burning Cost Rating Analysis for Ground-up AYII Cover for Detrended Maize Yields from 2007–08 to 2016–17 (kg/acre) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 Table 4.10:  Kehror Pacca Tehsil, Punjab: Comparison of HBA Results for (1) Top-up AYII Coverage and (2) Ground-up AYII Coverage for Cotton . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 Table 4.11:  Types of Traditional Indemnity and Index-based Livestock Insurance Products and Their Suitability for Punjab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64 Table 6.1: Uptake and Pricing Scenarios Analyzed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74 Table 6.2:  Portfolio Projections, FY2018/19 to FY2022/23: Number of Insured Farmers, Insured Area, and Sum Insured for Scenario 1 (average commercial premium rates 5.0 percent in Kharif season and 3.5 percent in Rabi season) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75 Table 6.3:  Uptake Projections for Area Yield Index Insurance for Semicommercial/Progressive Farmers and Subsistence Farmers (AYII programs 1 and 2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76 Table 6.4:  Punjab: Costs of Government Support to Crop Insurance Premium Subsidies and Program Implementation Costs, Scenario 1 (high uptake rates and average premium rates of 5.0 percent in Kharif season and 3.5 percent in Rabi season) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .78 Table 6.5:  Scenario 2: Five-Year Crop Insurance Portfolio Projections for High Uptake and Higher Average Premium Costs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81 Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan v Table 6.6:  Scenario 3: Five-Year Crop Insurance Portfolio Projections for Medium Uptake and Low Average Premium Costs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82 Table 6.7:  Scenario 4: Five-Year Crop Insurance Portfolio Projections for Low Uptake and Higher Premium Costs . . . . . 83 Table 7.1:  Area-Yield Index Insurance Program 1 (for semicommercial/progressive farmers linked to crop credit): Implementation Work Plan and Timetable Leading up to Launch in Kharif Season 2018 . . . . . . . . . . . . . . . . . . . 92 ANNEX TABLES Table A1.1: Punjab: Area, Production, and Yields of Major Cops, 2006/07–2015/16 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96 Table A1.2: Punjab: Average Wheat Yields (kg/acre) by District, 2006/07–2015/16 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 Table A1.3: Punjab: Average Cotton Yields (kg/acre) by District, 2006/07–2015/16 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98 Table A1.4: Punjab: Average Rice Yields (kg/acre) by District, 2006/07–2015/16 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99 Table A1.5: Punjab: Average Maize Yields (kg/acre) by District, 2006/07–2015/16 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100 Table A1.6: Punjab: Average Sugarcane Yields (kg/acre) by District, 2006/07–2015/16 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101 Table A1.7:  Punjab: Sown Area, Unirrigated Area, and Irrigated Area by Division and District, 2013–14 . . . . . . . . . . . . . . 102 Table A1.8:  Punjab: Correlation between Percentage of 2013–14 Sown Area Unirrigated per District and Coefficient of Variation in Wheat Yields, 2006–07 to 2015–16 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103 Table A2.1: Pakistan: Top 10 Disasters by Total Number of Deaths . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 Table A2.2: Pakistan: Damage Record by Type of Event, 1900–2017 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106 Table A2.3: Pakistan: Analysis of Damage Record by Decade . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106 Table A3.1:  Punjab: Compensation Paid (PKR) to Flood Victims for Loss of Livelihood, Housing, and Crops, 2010–15, by District . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 Table A4.1: India: Public Sector Crop Index Insurance Coverage, 2012–13 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111 Table A4.2:  India: Summary of Coverage and Performance of Public Sector Subsidized Crop Index Insurance for Small Farmers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112 Table A5.1: Wheat Cultivated Area and Average Yields, 2007–08 to 2016–17 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115 Table A5.2: Wheat AYII Rating Analysis Based on Actual Average and Detrended Yields . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117 Table A5.3: Cotton Cultivated Area and Average Yields, 2007–08 to 2016–17 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118 Table A5.4: Cotton AYII Rating Analysis Based on Actual Average and Detrended Yields . . . . . . . . . . . . . . . . . . . . . . . . . . . 119 Table A5.5: Rice Cultivated Area and Average Yields, 2007–08 to 2016–17 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120 Table A5.6: Rice AYII Rating Analysis Based on Actual Average and Detrended Yields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121 Table A5.7: Maize Cultivated Area and Average Yields, 2007–08 to 2016–17 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122 Table A5.8: Maize AYII Rating Analysis Based on Actual Average and Detrended Yields . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125 Table A5.9: Sugarcane Cultivated Area and Average Yields, 2007–08 to 2016–17 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126 Table A5.10: Sugarcane AYII Rating Analysis Based on Actual Average and Detrended Yields . . . . . . . . . . . . . . . . . . . . . . 127 Table A6.1: Mexico: CADENA Crop and Livestock Insurance Products and Programs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131 vi A Feasibility Study Table A6.2: Eligibility Criteria for CADENA Programs (Direct Support and SAC) in 2012 . . . . . . . . . . . . . . . . . . . . . . . . . . . .132 Table A6.3: CADENA Consolidated Agricultural Insurance Results 2003–2011 (MXN ’000) . . . . . . . . . . . . . . . . . . . . . . . . . 134 Table A6.4: Cost of Direct Compensation Payments (MXN) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134 Table A9.1:  Punjab Crop Insurance Program: Number of Insured Farmers, Insured Area, Sum Insured and Premium: Scenario 1, High Level of Uptake Percent Target Average Premium Rates: Kharif 5.0 Percent and Rabi 3.5 Percent (US$) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150 Table A9.2:  Punjab Crop Insurance Program: Number of Insured Farmers, Insured Area, Sum Insured, Premium and Premium Subsidies: Scenario 2, High Levels of Uptake and Higher Target Average Premium Rates: Kharif 7.5 Percent and Rabi 5.0 Percent (US$) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151 Table A9.3:  Punjab Crop Insurance Program: Number of Insured Farmers, Insured Area, Sum Insured, Premium and Premium Subsidies: Scenario 3, Medium Levels of Uptake and Lower Average Premium Rates: Kharif 5.0 Percent and Rabi 3.5 Percent (US$) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .152 Table A9.4:  Punjab Crop Insurance Program: Number of Insured Farmers, Insured Area, Sum Insured, Premium and Premium Subsidies: Scenario 4, Medium Levels of Uptake and Higher Average Premium Rates: Kharif 7.5 Percent and Rabi 5.0 Percent (US$) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153 FIGURES Figure 2.1: Punjab: District Map . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 Figure 2.2:  Punjab: Area, Production, and Yields for Major Crops, 2006/07–2015/16 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 Figure 2.3:  Wheat in Punjab: District-Level Average Yield and Standard Deviation (kg/acre), and Coefficient of Variation from 2005–06 to 2015–16 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 Figure 2.4: Punjab: Sown Area, Irrigated Area, and Percentage Sown Area Irrigated, 2013–14 . . . . . . . . . . . . . . . . . . . . . . . 12 Figure 2.5: Pakistan: Flood Map . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 Figure 2.6: Pakistan: Geographic Distribution of People Affected by Natural Disasters, 1973–2012 . . . . . . . . . . . . . . . . . . . 17 Figure 2.7: Pakistan: Record of Damage by Type of Natural and Climatic Event, 1900–2017 . . . . . . . . . . . . . . . . . . . . . . . . 18 Figure 2.8: Pakistan: Analysis of Natural and Climatic Disaster Damage Records by Decade, 1990–2017 . . . . . . . . . . . . 19 Figure 2.9: Punjab: Districts Flooded in 2010 and Number of People Affected . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 Figure 3.1: CLIS: Number of Insured Borrowers and Premium Income, 2008–09 to 2015–16 . . . . . . . . . . . . . . . . . . . . . . . . 27 Figure 3.2: Number and Value of CLIS Claims in Punjab, 2008–15 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 Figure 3.3:  Punjab: Distribution of Crop Compensation Payments (PKR) by District, 2010–15 . . . . . . . . . . . . . . . . . . . . . . . 35 Figure 4.1:  Suitability of Crop and Livestock Insurance Products for Different Types of Farmer . . . . . . . . . . . . . . . . . . . . . . 39 Figure 4.2:  Option for Linking CLIS and GoPunjab AYII Top-up Coverage for Semicommercial/Progressive Farmers . . . . 45 Figure 4.3:  Full-Value or “Ground-up” AYII Coverage for Semicommercial/Progressive Farmers Not Insured under CLIS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 Figure 4.4:  Example of an AYII Contract Providing Ground-up Coverage for Maize Grown in Village X . . . . . . . . . . . . . . . . . 49 Figure 4.5:  Dunyapur Tehsil, Punjab: Actual Historical Maize Yields and Detrended Maize Yields (kg/acre) . . . . . . . . . . . . 55 Figure 4.6: Kehror Pacca, Dunyapur, and Lodhran Tehsils, Punjab: 10-Year Average Cotton Yields (kg/acre) . . . . . . . . . . . 57 Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan vii Figure 4.7:  Kehror Pakka Tehsil, Punjab: Example of HBA for Top-up AYII Cover with an 80 Percent Trigger Yield and 50 Percent Exit Yield for Cotton, Based on Actual 12-Year Historical Yields (kg/acre) . . . . . . . . . . . . . . . . . 58 Figure 6.1: Punjab: Proposed Phasing of New Crop Insurance Programs, FY2018/19 to FY2022/23 . . . . . . . . . . . . . . . . . . 74 Figure 6.2: Number of Insured Farmers by Year and Program Type, 2018/19–2022/23 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76 Figure 6.3: Estimated Costs of Crop Insurance Premium Subsidies FY2017/18 to FY2021/22 (US$) . . . . . . . . . . . . . . . . . . 78 Figure 6.4:  Estimated Costs of Government Support to Crop Insurance Program Start-up and Operating Costs, FY2017/18–FY2021/22 (US$) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 Figure 6.5:  Total Costs of GoPunjab Support to Crop Insurance (premium subsidies and subsidies on operating costs), FY2017/18–FY2021/22 (US$) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 ANNEX FIGURES Figure A3.1:  Punjab: Compensation Paid (PKR) to Flood Victims for Loss of Livelihood, Housing, and Crops, 2010–15, by District . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108 Figure A5.1: Wheat Cultivated Area and Average Yields, 2007–08 to 2016–17 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116 Figure A5.2: Cotton Cultivated Area and Average Yields, 2007–08 to 2016–17 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118 Figure A5.3: Rice Cultivated Area and Average Yields, 2007–08 to 2016–17 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120 Figure A5.4: Maize Cultivated Area and Average Yields, 2007–08 to 2016–17 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123 Figure A5.5: Maize Yield and Shortfall Based on Actual 10-Year Yields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123 Figure A5.6: Maize Yield and Shortfall Based on Detrended 10-Year Yields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124 Figure A5.7: Sugarcane Cultivated Area and Average Yields, 2007–08 to 2016–17 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126 Figure A6.1:  Mexico: Evolution of the CADENA Program, 2003–11 (insured crop area in hectares and number of insured livestock) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130 Figure A6.2: CADENA Cost of Premium Subsidies to State and Federal Governments (MXN ’000) . . . . . . . . . . . . . . . . . . 133 Figure A6.3:  CADENA: Comparison of Insurance Coverage Purchased (TSI) versus Total Cost of Government Financial Support (premium subsidies and direct payments) (MXN) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135 Figure A6.4:  Hypothetical Analysis of Alternative Use of Crop Insurance Premiums to Make Direct Support Payments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136 Figure A7.1: Agroseguro Spain: Institutional Framework . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140 Figure A7.2: Agroseguro Premiums and Share Paid by Farmers and by the State . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141 Figure A7.3: Turkey: Institutional Framework of the Tarsim Agricultural Insurance Pool . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143 Figure A7.4: Tarsim Agricultural Insurance Pool: Risk Transfer Mechanisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144 Figure A7.5: Tarsim: Growth in Number of Policies Sold, TSI and Premium Income, 2007–2012 . . . . . . . . . . . . . . . . . . . . . 144 BOXES Box 1.1: Agriculture’s Lackluster Performance in Punjab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 Box 5.1: Benefits and Limitations of Coinsurance Pool Arrangements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 viii A Feasibility Study ANNEX BOXES Box A4.1: Key Features of India’s Area-based National Agricultural Insurance Scheme for Crops . . . . . . . . . . . . . . . . . . . 110 Box A4.2: Main Features of India’s Modified NAIS (mNAIS) Scheme for Rabi 2010–11 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111 Box A4.3: Brazil’s Maize AYII Program in Rio Grande do Sul State . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113 Box A7.1: Objectives of the Tarsim Pool . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143 Box A8.1: Benefits and Limitations of Coinsurance Pool Arrangements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147 Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan ix ACKNOWLEDGMENTS This report was prepared at the request of the Government of Punjab and authored by Charles Stutley (Agricultural Insurance Specialist Consultant), Vijayasekar Kalava- konda (Senior Financial Sector Specialist GFM02), and Johannes (Hans) Jansen (Senior Agriculture Economist, GFA06). The authors would like to thank the Government of Punjab, especially the Depart- ment of Agriculture, for support in facilitating access to key stakeholders, including the Crop Reporting Services (CRS), Department of Livestock and Dairy Develop- ment (DL&DD), Punjab Irrigation Department (PID), Meteorological Department, National and Provincial Disaster Management Authorities, Securities and Exchange Commission of Pakistan, authorities on crop insurance of the Federal Ministry of Finance, leading commercial banks—including Zarai Taraqiati Bank Limited (ZTBL, formerly the Agriculture Development Bank of Pakistan)—and insurance companies, officials of the State Bank of Pakistan (SBP), National Rural Support Program (NRSP), Pakistan Poverty Alleviation Fund (PPAF), and the Pakistan Microfinance Investment Company. Funding support from the Global Index Insurance Facility (GIIF) and the Global Facility for Disaster Reduction and Recovery (GFDRR) is gratefully acknowl- edged, as is final editing of the report by Kelly Cassaday. Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan xi ACRONYMS AND ABBREVIATIONS AAL Annual average loss MPCI Multiple Peril Crop Insurance AYII Area-yield index insurance MT Metric ton CADENA Componente Atención a Desastres Naturales NAIS National Agricultural Insurance Scheme (Natural Disaster Response Component of NDMA National Disaster Management Authority the Ministry of Agriculture, Livestock, Rural NDVI Normalized Difference Vegetation Index Development, Fisheries, and Food in Mexico) NPCI Named Peril Crop Insurance CCE Crop cutting experiments NRSP National Rural Support Program CLIS Crop Loan Insurance Scheme PDMA Provincial Disaster Management Authority, CTL Constructive Total Loss Government of Punjab CRS Crop Reporting Services PDRMA Punjab Disaster Risk Management Agency CV Coefficient of variation PforR Program for Results DoA Department of Agriculture, Government of PKR Pakistan rupee Punjab PMD Punjab Meteorological Department GIC General Insurance Corporation of India PMFBY Pradan Mantri Fasal Bima Yogana (Prime GDP Gross Domestic Product Minister’s Crop Insurance Scheme, India) GFDRR Global Fund for Disaster Reduction and PPAF Pakistan Poverty Alleviation Fund Recovery PPP Public-Private Partnership GoP Government of Pakistan SBP State Bank of Pakistan GoPunjab Government of Punjab SMS Short message system GRP Group Risk Plan TSI Total Sum Insured HBA Historical burning cost rating analysis TSU Technical Support Unit IFAD International Fund for Agricultural Development UAI Unit area of insurance LISB Livestock Insurance Scheme for Borrowers US$ United States dollar LTA Long-term average WFP World Food Programme MFI Microfinance Institution WII Weather Index Insurance mNAIS modified National Agricultural Insurance ZTBL Zarai Taraqiati Bank Limited Scheme (India) Weights and Measures 1 ton = 1,000 kg 1 maund = 40 kg in Pakistan Currency Exchange Rate Used in This Report 1 US Dollar (US$) = 100 Pakistan rupees (PKR) Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan xiii EXECUTIVE SUMMARY This report presents the findings and recommendations of a World Bank agricultural insurance feasibility study commissioned by the Government of Punjab (GoPunjab) for the design and implementation of large-scale crop and livestock insurance programs for Punjab’s 5.2  million mainly small-scale farmers and livestock producers.1 The crop and livestock insurance programs will be developed and rolled out under the World Bank financed US$300 million project for “Strengthening Markets for Agriculture and Rural Transformation in Punjab” Program-for-Results (SMART Punjab PforR) from 2018 to 2022. The report consists of seven chapters, starting with the background and objectives of the feasibility study. Chapter 2 includes a review of crop and live- stock production in Punjab Province, an assessment of the main natural and climatic risk exposures faced by the farmers, and the value of losses caused by climatic disasters. Chapter 3 presents a review of the existing crop and livestock products and schemes which are being implemented in Pakistan along with their issues and challenges along with an analysis of the flood disaster compensation payments made to farmers by the Provincial Disaster Management Authority (PDMA) Punjab. Chapter 4 presents options and proposals for new crop and livestock insurance products and programs for GoPunjab to consider, based on international best practice. This is followed in Chap- ter 5 by a review of the legal, institutional, and operational considerations and options for Punjab. Chapter 6 presents an outline of a five-year crop and livestock buildup plan and budget along with estimates of the potential costs of GoPunjab financial support to this program for premium subsidies and subsidies on insurance startups and operating costs. Finally, Chapter 7 deals with the next steps in designing and implementing agri- cultural insurance in Punjab, starting with the launch in Kharif (summer) season 2018 of a crop area-yield index insurance (AYII) program for small-scale s ­ emicommercial/ progressive farmers. IMPORTANCE OF AGRICULTURE IN PUNJAB Agriculture is a key economic sector in Punjab. The agricultural crop, live- stock, fisheries, and forestry sectors in Punjab account for 26 percent of Punjab’s GDP and 40 percent of employment. In 2013–14 Punjab produced 19.7 million tons of wheat, equivalent to 76 percent of total wheat production in Pakistan. It is also a major producer of maize, cotton, sugarcane, rapeseed, and mustard. The province is also a very important producer of horticultural crops (fruits and vegetables). Livestock pro- duction for both milk and meat is very important in Punjab, and all farming families own some livestock. 1The Feasibility Study Report was presented to the GoPunjab in July 2017. Based on the report’s findings and recommendations, GoPunjab decided to launch a pilot crop insurance program in Kharif 2018 and scale up the program over the next five years. At GoPunjab’s request the Report was updated in 2018. The key changes to the earlier draft include Chapter 7, which has been updated to highlight key issues and challenges that were encountered in the launch of the pilot crop insurance program in April 2018 for the Kharif 2018 season. Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan xv There are a total of 5.25 million farms in Pun- reducing the risk exposure of default in the event of jab (63.5 percent of the total 8.26 million farms major climate-induced crop failure, financial institutions in Pakistan) and the majority of these farms are are more likely to extend production credit to small farm- small.2 In Punjab, the 2010 census reported an average ers to enable them to invest in improved seed and fer- farm size of 5.6 acres: the small farm size is illustrated by tilizer technology and to increase their production and the fact that 91 percent of all farms are under 12.5 acres incomes. (5 hectares) and 41  percent of farms are smaller than 2.5 acres (1 hectare). The very small size of farms poses major challenges for the design and operation of suitable crop and livestock insurance products and programs. CLIMATIC RISK EXPOSURE IN AGRICULTURE Agricultural growth is, however, lagging in Pun- Pakistan (including Punjab), whose risks are jab and declined from 3.3 percent annual growth further exacerbated by a rapidly growing popu- over the last decade to below 3 percent during lation, growing water scarcity, and uncontrolled 2011–2015, and zero growth in the financial year, 2016. urbanization is highly vulnerable to climate Growth turned positive again in financial year 2017. Both change. The country is ranked among the top ten most crop and livestock productivity are low compared to lev- climate vulnerable countries in the world in the Global els achieved by semicommercial/­ progressive farmers and Climate Risk Index and has seen a considerable increase other Asian countries, and yield growth is largely flat for in frequency and intensity in extreme weather events and major crops such as wheat and rice.3 Reasons forwarded natural disasters, causing huge losses. A recent study by for low agricultural growth include limited adoption of the World Bank established that the melting of the Hindu modern technologies, poor service delivery, inefficient Kush-Karakoram-­ Himalayan glaciers could affect water irrigation water delivery and pricing, poorly function- flows into the Indus River system with implications for ing agricultural markets, and overreliance on provincial agricultural production.4 Concerted efforts at adaptation government subsidies on crop inputs and output (World to conserve water and build resilience in the agriculture Bank 2017). sector are therefore required. Access to production credit appears to be a major Monsoon flooding is a major risk exposure to crop constraint to increasing farm investments thus and livestock production in Punjab, and can result impacting productivity and yield gains in Pun- in major loss of crop production and death of live- jab. According to the State Bank of Pakistan (SBP), cur- stock. Other risks include drought in rain-fed areas of the rently there are about 1 million small farmers with less province, localized windstorm and hail damage, and pests than 25 acres (or 12 percent of total farmers in Pakistan) and diseases of crops and livestock. In the case of livestock who borrow credit from the banks and who are insured severe flooding results in death of animals by drowning, but on a compulsory basis under the Crop Loan Insurance also starvation due to lack of fodder and grazing and dis- Scheme (CLIS). It is estimated that about 70 percent ease outbreak. There is a significant risk of earthquakes, (700,000) of these farmers are located in Punjab. How- but this mainly causes loss of life and damage to urban pri- ever, the majority (>85 percent) of farmers in Punjab do vate and public infrastructure and property rather than to not have access to seasonal crop credit. Several commer- agriculture. cial banks reported that they were reluctant to lend to farmers because of the high historical default rates. In the 2010 floods, Pakistan incurred total direct and indirect flood damage estimated at Paki- Agricultural insurance could potentially play an stan rupee (PKR) 854.8 billion (US$10.1  billion) important role in leveraging access to produc- of which 50 percent was incurred by agriculture, tion credit by small farmers in Punjab. Where livestock, and fisheries. The floods affected more crop insurance is linked to credit provision, thereby than 20 million people (over one-tenth of Pakistan’s population) with over 1,980 reported deaths and nearly 2The latest agricultural census (2010) is available at http://www.pbs.gov.pk/ 2,946  injured; about 1.6 million homes were destroyed content/agricultural-census-2010-pakistan-report. 3Maize is one exception and the cultivated area, production, and yields have and thousands of acres of crops and agricultural lands increased significantly in recent years on account of the strong demand for maize by the animal and poultry feed industries and adoption of high- yielding varieties (HYVs), especially hybrids. 4Yu et al. (2013). xvi A Feasibility Study were damaged. The total costs of reconstruction were these aims government has allocated a US$3.90  bil- estimated at between PKR 577.9 billion (US$6.8 billion) lion capital budget for agriculture, livestock, and irriga- and PKR 757.8 billion (US$8.9 billion) while the costs of tion over five years (FY 2017 to FY 2021), out of which reconstruction of the agricultural sector were estimated US$1.6 billion is allocated to agriculture and livestock at between PKR 21.9 billion (US$257 million) and PKR (excluding the Kissan package). 89.1 billion (US$1,049 billion). Sind and Punjab Prov- inces were the most heavily flood damaged provinces in The “Strengthening Markets for Agricul- 2010: in Sind damages were valued at PKR 371.3 bil- ture and Rural Transformation in Punjab” lion (US$4.4 billion) or 43.6 percent of the total value of Program-for-Results (SMART Punjab PforR) is ­ damage (ADB and World Bank 2010). a US$300 million loan from 2018 to 2022 by the World Bank to the Government of Pakistan (GoP) In Punjab the 2010 flood damage was estimated to assist the GoPunjab’s programs to trans- at PKR 219.3 billion (US$2.6 billion) or 25.7 per- form agriculture for the province’s 5.25 million cent of total damage, and the costs of reconstruction mainly smallholder farmers. The SMART Pun- were estimated at between PKR 93.5 billion (US$1.1 bil- jab PforR will assist the GoPunjab in three results areas: lion) and (PKR 117.6 billion (US$2.1 billion) (ADB and (1) increased on-farm productivity and value of crops World Bank 2010). and livestock to reduce unit production costs through improving agricultural research and extension systems Currently in Pakistan, the government oper- and targeting subsidies to smallholders; (2) increased ates ex-post natural disaster compensation pro- value addition and competitiveness of crops and live- grams for the affected population, but budget stock through modernizing the wheat marketing system, constraints mean that only a fraction of the lost stimulating high-value agriculture, deregulating crop and value of agricultural crop production is com- livestock markets with increased private sector participa- pensated. Natural disaster management is coordinated tion, improving livestock health and breeding, regulating at the national level by the National Disaster Manage- food safety and inspecting and testing food quality, and ment Authority (NDMA) and implemented by the Pro- developing agribusiness including post-harvest manage- vincial Disaster Management Authorities (PDMAs). ment and value-addition; and (3) enhanced resilience In the case of agriculture, post disaster relief is very of smallholder farmers to climate change and natural restricted. For example, PDMA Punjab reports 3.3 mil- disasters through improvements in the financial sustain- lion acres of crops damaged due to the severe floods of ability of surface irrigation systems through better water 2010 to 2013. The World Bank team estimates these charge assessment and collection, regulation of ground- crop losses at about US$1.6 billion; however, over this water use, and improved water service delivery, as well as period, compensation payments to farmers in Punjab improved climate resilience through crop insurance and amounted to only US$67 million or just over 4 percent introduction of CSA technologies. of the total flood losses. Agricultural insurance could play an important role in complementing the existing This agricultural insurance feasibility study ex-post disaster compensation schemes. has been conducted under Results Area 3 of the SMART Punjab PforR project which seeks to intro- duce new innovative crop and livestock insurance prod- ucts and programs to meet the needs of all sectors of GOPUNJAB GOALS Punjab’s farming community. TO TRANSFORM AGRICULTURE AND SMART AGRICULTURAL INSURANCE PUNJAB PROGRAM GoPunjab aims to stimulate growth in the agri- PROVISION cultural sector by facilitating increases in crop and Private commercial agricultural crop and live- livestock productivity, enhancing resilience, increasing stock insurance is poorly developed in Pakistan, competitiveness in agriculture marketing and trade by and few farmers in Punjab are insured against providing a conducive climate for private investment, and loss of their crops and livestock. The largest improving supply chains and value addition. To achieve crop insurance program in Pakistan is the Crop Loan Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan xvii Insurance Scheme (CLIS), which was introduced in 2008 international markets and their potential suit- by the State Bank of Pakistan (SBP) in conjunction with ability for introduction in Punjab. This includes a group of about 12 private local insurers. The CLIS is traditional indemnity-based crop insurance products a “catastrophe loss of yield coverage,” which is triggered and new innovative index insurance products, details of when a disaster is declared by the provincial and/or dis- which are summarized in Table ES1. trict authorities and when crop losses exceed 50 percent of normal expected production and yields. CLIS consists One size does not fit all, or in other words, crop of a Constructive Total Loss (CTL) Policy such that when insurance products must be tailored to the risk crop losses exceed the 50 percent area yield threshold, the transfer needs of different types of farmer in sum insured (outstanding seasonal crop production credit Punjab including: medium and large commercial farm- loan value) is settled in full to the lending institution. ers, small semicommercial/progressive farmers, and sub- CLIS is mandatory for all farmers accessing seasonal sistence farmers. loans from the commercial banks. Small loanee farmers with up to 25 acres of land qualify for 100 percent CLIS premium subsidies up to a maximum premium rate of CROP INSURANCE PRODUCTS FOR 2.0 percent paid by SBP. Large loanee farmers with >25  acres do not qualify for CLIS premium subsidies. MEDIUM AND LARGE FARMERS The main drawback of the CLIS is that although it pro- IN PUNJAB tects the lending institutions against farmers defaulting For commercial farmers in Punjab who account on their loan repayments in times of catastrophe losses, for just 3 percent of all farming households and the insurance program does not directly benefit farmers who have more than 25 acres of cropping, indi- themselves. vidual grower Multiple Peril Crop Insurance (MPCI) or named-peril crop insurance may be In Punjab about 700,000 farmers (loanees) are a suitable product for the largest of these farm- automatically insured under CLIS, representing ers with more than 100  acres (40 hectares) of about 13 percent of all 5.2 million farmers in insured crops. Insurers can offer MPCI coverage to this province and approximately 25 percent of the 2.9 large farmers because the premium generated by each million smallholders (farmers having 2.5–25 acres). The risk is adequate to cover the costs of pre-acceptance risk CLIS does not, however, provide adequate protection inspections, mid-season monitoring inspections, and for these farmers against loss of their production costs end of season crop yield assessment (Table ES1). invested in growing their crops. There is currently no crop insurance coverage for high value vegetable crops (e.g., potatoes) and tree fruit crops (e.g., citrus, mangoes). CROP INSURANCE PRODUCTS Moreover, the insurance sector does not offer any pro- FOR SMALL SEMICOMMERCIAL/ tection for small and marginal farmers with less than 2.5 PROGRESSIVE FARMERS IN PUNJAB acres, which amounts to 2.2 million farmers or 42 per- Individual farmer MPCI is not a suitable product cent of all farmers in Punjab. for small semicommercial/progressive farmers who typically own between 2.5 acres5 and 25 acres Since 2011 SBP has also designed a livestock and who account for 55 percent of all farmers investment mortality coverage which attracts in Punjab. These small-scale semicommercial farmers premium subsidies. Finally, there have been a few produce crops both for family consumption and for sale. private sector crop index insurance pilot programs and They are increasingly accessing seasonal crop loans to livestock insurance initiatives. invest in improved hybrid seed and fertilizer technology, and they face a financial exposure in the event of crop loss. For these farmers, an index-based insurance prod- uct such as weather index insurance (WII) or area-yield CROP INSURANCE OPTIONS FOR PUNJAB 5It should be recognized that there are also “semicommercial” farmers with Chapter 4 of this report presents a detailed less than 2.5 acres who are receiving crop production credit either through the Microfinance Institutions (MFIs) or GoPunjab’s E-Kissan credit scheme. overview of the different types of individual These farmers would also be targeted by commercial crop insurance products farmer crop insurance products available in and services. xviii A Feasibility Study TABLE ES1: CROP INSURANCE PRODUCTS AND SUITABILITY FOR PUNJAB Type of agricultural Basis of insurance insurance product and indemnity Availability Suitability for Punjab a. Indemnity-based crop insurance Named Peril Crop Insurance Percent damage 1.  Widespread Possible (e.g., hail, frost, wind) (NPCI) Multiple Peril Crop 2.  Yield loss Widespread Only for large growers >40 ha cereals Insurance (MPCI) 3.  Crop Revenue Insurance Yield loss and price loss Very restricted (USA) Not available b) Index-based crop insurance Crop Weather Index 4.  Weather index payout Widespread Limited weather station density Insurance (WII) based on scale (30 Punjab). Not best suited to microlevel Ground Weather Stations insurance for small cereal farmers <2 ha. Possible applications for horticulture and fruit crops Crop Weather Index 5.  Weather index payout Fairly widespread Satellite data freely available. Not best Insurance (WII) based on scale suited to micro-level insurance for small Synthetic Satellite Rainfall farmers <2 ha 6.  Crop Area Yield Index Area yield loss Fairly widespread Potential for small farmers (cereals, Insurance (AYII) cotton, sugarcane) using Department of Agriculture/Crop Cutting Services (CCE) yield data 7.  Specialist indexes, e.g., flood Flood index payout Very restricted Major research required to launch cover index Bangladesh scale index insurance (AYII) which do not require costly pre-­ of an AYII product is that it does not indemnify crop inspections or individual field-by-field loss assessment, yield losses at the individual farmer or field level. Rather, may offer solutions to their risk transfer needs. an AYII product makes indemnity payments to farmers according to yield loss or shortfall against an average area In Punjab, the density of weather stations is very yield (the index) in a defined geographical area (e.g., a low and insufficient to support a p ­ rovincial-level district or Tehsil or Union Council or Village). crop WII program for major crops grown by the majority of the province’s 5.2  million farmers. The key advantages of the Area-Yield Index The Punjab Meteorological Agency has a network of Insurance approach are that moral hazard and only 30 official synoptic weather stations or less than one anti-selection are minimized and the costs of weather station per district. This density is inadequate administering such a policy are significantly to develop a large-scale WII program for the 5.2 million reduced, making this product suitable to offer to farmers in Punjab. Alternatives exist to develop satellite small-scale farmers. Under an AYII policy yield losses indexes, including synthetic rainfall, evapotranspiration, are settled against the area average yield index as opposed or Normalized Difference Vegetation Index (NDVI) to settling losses on individual farmers’ fields. This means in certain crops, and under the SMART Punjab PforR that individual farmers cannot influence the yield out- project, opportunities to develop satellite index insurance come, for example by purchasing coverage only for fields may be explored further. in low lying areas which are subject to flooding and water logging (anti-selection) or by applying sub-­ optimal levels Starting in 2018, there appears to be considerable of husbandry and pest and disease and weed controls scope to design and implement AYII for semi- (moral hazard) in the expectation of then claiming the commercial/progressive farmers in Punjab who yield loss on their crop insurance policy. The costs of grow Rabi (winter) wheat and Kharif (summer) operating AYII are much lower than for a MPCI policy, rice, maize, cotton and sugarcane. The key feature especially because individual farmer pre-inspections and Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan xix in-field crop loss assessment are not required, and this These farmers account for 2.2 million farms or 42 percent offers the potential to market this product at lower pre- of all farms in Punjab. AYII could be used as an ex-ante mium costs to small and medium sized farmers. crop insurance coverage to trigger objective payouts to the large number of subsistence farmers in Punjab and The main disadvantage of an AYII policy is could operate as a complementary disaster risk financing “basis risk,” i.e., the difference in the actual yield out- and insurance coverage to the government’s existing nat- come achieved by individual farmers on their own fields ural disaster compensation program, which is operated and the average area yield. For example, an individual through the Punjab Disaster Risk Management Agency. farmer may incur severe crop production and yield losses due to localized perils, e.g., hail or flooding by a nearby river, but because these localized losses do not impact on CROP INSURANCE PRODUCTS the county or departmental average yield, the grower FOR HORTICULTURE CROPS does not receive any indemnity. In Punjab, there may be scope for developing Named Peril Crop Insurance (NPCI) to protect To operate an AYII coverage, it is necessary to against specific perils such as frost, hail, and have (1) accurate historical yield data at the local excess rain in high-value horticultural crops such area levels on which basis to construct a yield as potato, or frost, hail, and wind damage in tree index, and (2) an objective and accurate method crops including mango and citrus. Potentially, hail of establishing the actual average yield in the and wind damage insurance could also be offered for insured growing season to determine if a payout cereals if there is a significantly high exposure in these is due or not. In Punjab, the Crop Reporting Services crops, especially at the time of grain maturity and har- (CRS) of the Department of Agriculture (DoA) has for vest. In addition, there may be scope for developing WII many years been involved in implementing seasonal crop coverage for these crops in areas which are supported by yield surveys based on a random selection of farmers ground weather stations (Chapter 4). and fields which are then subjected to randomly placed crop cutting experiments (CCEs) to estimate crop yields A phased program is proposed for the planning for major crops, including wheat, rice, maize, cotton, and and design and implementation of the three sugarcane. It is proposed to base the AYII program on large-scale crop insurance programs starting in the CRS data from CCEs, recognizing that the density of Kharif 2018 with the AYII Program for semicom- those CCEs will have to increase significantly over time. mercial farmers. This would be followed in 2019 by the launch of the AYII program for small subsistence Chapter 4 describes the key contract design fea- farmers and the NPCI program for tree fruit and vegeta- tures of an AYII policy and explains the approach to ble farmers (Table ES2). setting insured yield coverage levels and methodology for rating such a coverage to derive technical and commercial premium rates. Some preliminary rating analysis is pre- sented for the five major crops based on tehsil-level crop LIVESTOCK INSURANCE yield data for one district provided by CRS. OPPORTUNITIES FOR PUNJAB CROP INSURANCE PRODUCTS A preliminary assessment of opportunities to FOR SUBSISTENCE FARMERS develop livestock insurance has been conducted WITH LESS THAN 2.5 ACRES IN PUNJAB as part of this feasibility study. Punjab province is a major dairy cattle and milk producer. GoPunjab has AND WHO DO NOT HAVE identified a major potential to increase the productivity ACCESS TO SEASONAL CROP CREDIT of dairy farmers by introducing improved breeds of cat- AYII is also identified as a suitable insurance tle coupled with modern husbandry and animal health product that could be offered by GoPunjab as a practices, and improved milk marketing systems. social protection coverage to protect poor subsis- tence farmers with less than 2.5 acres and who There may be opportunities for GoPunjab to pro- do not have access to formal crop credit and who mote individual animal accident and mortality primarily produce crops for family consumption. xx A Feasibility Study TABLE ES2:  PROPOSED PHASED INTRODUCTION OF NEW CROP INSURANCE PRODUCTS AND PROGRAMS IN PUNJAB Financial year FY 2018–19 FY2019–20 FY2020–21 FY2021–22 FY2022–23 Crop season Kharif Rabi Kharif Rabi Kharif Rabi Kharif Rabi Kharif Kharif Crop insurance programs Area Yield Index 1.  Launch Kharif 2018 Insurance for progressive farmers >2.5 Ac <25 Ac Area Yield Index 2.  Launch Kharif 2019 Insurance for subsistence farmers <2.5 Ac Named Peril Crop 3.  Launch Kharif 2019 Insurance for tree fruit and vegetable farmers Livestock insurance programs Dairy Cattle Insurance 1.  Launch in FY2019–20 (indemnity-based accident and mortality cover) coverage for dairy cattle through the banks, dairy pools are common features of major national or regional cooperatives, or (fresh) milk processors. For large Private Partnership (PPP) agricultural insurance Public-­ commercial dairy herds, insurers may be willing to offer programs, including the Agroseguro Program in Spain, All Risk Mortality coverages (see Table ES2). the Tarsim pool program in Turkey, and various regional coinsurance pools in China. Key features of the Spanish There do not, however, appear to be major oppor- and Turkish agricultural insurance pool programs are pre- tunities at present to develop livestock index- sented in Annex 8. Similarly, several developing countries based insurance (e.g., pasture drought NDVI in Africa including Senegal, Malawi, Ghana, and Kenya coverage). The earliest date that livestock insurance have formed agricultural insurance pools in recent years. could be rolled out would be in 2019 given the major commitments to designing the three crop insurance pro- There are potential advantages of pools which grams in 2018 and 2019. include: (1) cost sharing in the research and develop- ment and start-up stages, (2) cost savings in establishing a single underwriting unit, staffing and equipment, either within the lead coinsurer or as a separate underwriting INSTITUTIONAL entity (namely, a Special Purpose Vehicle), (3) ability for AND OPERATIONAL each company to select a share according to its risk appe- tite, and (4) major cost savings in purchasing reinsurance CONSIDERATIONS FOR protection because of the effects of pooling risk and risk LARGE-SCALE CROP diversification. INSURANCE PROGRAMS In the short term it is unlikely that the partici- IN PUNJAB pating insurers would want to create and incor- porate a new pool insurance company for the In the planning and design of the large-scale specific purpose of insuring crops and livestock crop (and livestock) insurance programs for the in Punjab. Rather, they are more likely to seek a simple Punjab, interested insurers will need to consider coinsurance agreement which would allow each of them whether to (1) underwrite this business separately to take up an agreed share of the risk. In this case as the as they currently do for the CLIS, or (2) to form pool would not be a legal entity, it is likely that one com- a coinsurance pool to collectively underwrite pany would be appointed to lead the pool and to issue and settle claims on the programs. Coinsurance policies on their own paper. The pool insurers would also Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan xxi need to agree on how they would manage the business— either by (1) sharing the workload among themselves for POTENTIAL ROLES FOR the key functions of marketing and promotion, education GOPUNJAB TO SUPPORT and training, underwriting and policy issuance and pre- mium collection and in claims settlement and process- THE LARGE-SCALE CROP ing, or (2) by appointing the lead insurer to conduct these activities on their behalves and to contribute to the lead AND LIVESTOCK insurer’s operating expenses. INSURANCE PROGRAMS International experience from developing coun- In Punjab there appear to be major opportuni- tries clearly demonstrates the importance of ties to bundle the AYII program with the Kissan involving both government and the private sec- seasonal crop credit program, which is being tor in agricultural insurance initiatives for small promoted by GoPunjab through the rural and farmers. When only private sector insurance companies commercial banks. There appears to be a major need provide agricultural insurance without government sup- in Punjab to improve farmers’ access to rural finance if port, they seldom have the resources to design and imple- they are to invest in improved seed and fertilizer technol- ment crop and livestock insurance programs for small ogy and to thereby increase their production and yields farmers. When the government alone offers agricultural and farm incomes. The bundling of crop insurance with insurance, its lack of infrastructure and expertise makes credit and input supplies has been shown in many parts distributing policies, delivering payouts, and paying claims of the world to provide a win-win for farmers, credit pro- difficult. Experience from agriculture insurance schemes viders, and insurers. The farmer gains access to seasonal developed across the world (for example in India, Mongo- crop credit and lending institutions are more willing to lia, Morocco, and Kenya) shows that public-private part- lend to small farmers because their loans are protected by nerships (PPPs) can overcome these challenges by building crop insurance. In addition, the insurer benefits from (1) a on the comparative advantages of the respective sectors. reduction in anti-section, (2) less need for pre-­ inspections (3) reductions in the costs of promoting and marketing Chapter 5 of this report identifies a series of the agricultural insurance program, and (4) an insurance areas where GoPunjab support to the crop insur- uptake and spread of risk and premium volume that is ance operations would be critical to the success- generally much higher than under a purely voluntary ful implementation of these programs, including: program. 1) Data strengthening for crop insurance, including most importantly designing and imple- The option of linking commercial crop AYII menting a farmer electronic registration and insurance for semicommercial/progressive database system and in providing the insurers farmers with the CLIS program should also be with crop time series yield data at the tehsil level explored in the design phase of this program. for the major crops. In this instance the coverage would be a “top-up” cov- 2) Strengthening of the CCEs for area yield erage for the farmer to insure against yield shortfall estimation. As noted in Chapter 5, areas for from approximately 80 percent of the area yield down government support include: (1) significantly to 50 percent when CLIS would come in to ensure that increasing the CCEs, to permit the UAI to be set the bank was protected for yield loss below 50 percent of at the Union Council or eventually at the individ- insured yield. If such a proposal were to be adopted, the ual village level, and (2) introducing mobile phone main changes that would need to be agreed with SBP or electronic tablet technology to record the CCE and the participating insurers and GoPunjab authori- data and to transmit this in real-time to insurance ties is that rather than declaring a calamity declaration underwriters and other stakeholders. This tech- to trigger payouts on the CLIS, the latter would agree nology has already been developed, tested, and is to follow the terms and conditions of the AYII Policy now under large-scale implementation in India as including (1) the definition of the Unit Area of Insurance part of the Fasal Bima Yojana program. (UAI), (2) the average yield index for that UAI, and (3) to 3) Investment in farmer awareness, edu- base any payouts on the objective CCEs that the CRS is cation, and training in the role of crop conducting at the time of harvest and to only make pay- insurance and the operation of the vari- outs if the actual average yield falls short of 50 percent. ous insurance products and programs. Farmer xxii A Feasibility Study insurance literacy creation is a key pillar to the GoPunjab will need to establish an annual budget sustainability of the Punjab crop insurance pro- to cover the premium subsidies and contributions gram under the SMART Punjab program. to start-up and operating costs, and appoint an 4) Monitoring and Evaluation (M&E). It institution that will be responsible for adminis- will be critical to implement a M&E system to tering the premium subsidy regime on its behalf. assess the program’s inputs and outputs, time- The norm in most subsidized agricultural programs is liness, effectiveness, and impact over time on that (1) the farmer is only charged the unsubsidized por- ­ semicommercial/progressive farmers input pur- tion of the premium, and (2) the insurer then reclaims chasing decisions and their crop production and the premium subsidy amount from the entity appointed yields and incomes. For subsistence farmers, M&E by the government, which is responsible for auditing and should focus on measuring whether the program processing and repaying the premium subsidies. enables them to maintain their consumption levels following major floods or droughts, and whether they are able to get back into production for the following season. CROP INSURANCE FIVE-YEAR BUILD-UP PLAN In addition, GoPunjab support to providing crop (and livestock) premium subsidies for small AND FINANCIAL BUDGET farmers will be very important in determining Chapter 6 of this report presents a five-year the demand for and scale-up of the program. (FY2018/19 to FY2022/23) crop insurance GoPunjab has indicated its intention to support the intro- buildup plan and an indicative financial budget duction of crop insurance in Punjab through the provi- for GoPunjab to consider starting in the Kharif sion of premium subsidies. The feasibility study suggests season 2018. The purpose of presenting the crop insur- different premium subsidy levels to enable GoPunjab to ance buildup plan and budget (numbers of insured farm- establish its own budget for premium subsidies as per ers, insured area, sums insured and premium projections) each of the three crop insurance programs: is to assist GoPunjab to develop its own five-year crop Program 1: AYII for semicommercial/progressive insurance plan and budget and to assess the fiscal costs farmers: 50 percent premium subsidy. The ratio- of premium subsidy support and financial support for nale is that small semicommercial/progressive other operational activities. The buildup plan presents farmers can afford to contribute toward the costs physical projections for the three crop insurance pro- of their crop insurance premiums. grams (1) AYII for semicommercial/progressive farmers, Program 2: AYII social protection program for sub- (2) AYII for subsistence farmers, and (3) NPCI coverage sistence farmers: 100 percent premium subsidies. for fruit and vegetable growers, of the number of insured GoPunjab would fully fund the premiums in rec- farmers by season and by year and the corresponding ognition of the fact that poor subsistence farm- estimates of insured area, sums insured, and premium ers are unlikely to be able to afford to fund crop income based on a series of assumptions detailed in this insurance premiums. chapter. A key variable is the pricing of the various crop Program 3: NPCI for horticulture producers: 50 per- insurance programs—for this budget exercise target aver- cent premium subsidy. These farmers tend to be age commercial premium rates of 5.0 percent for Kharif larger commercial farmers producing high value crops and 3.5 percent for Rabi crops are assumed.6 The cash crops for sale, and they can afford to contrib- costings presented in this report will need to be scruti- ute toward their crop insurance premiums. nized and confirmed by GoPunjab, and in the case of the indicative or target commercial premiums these will For the 2018 Kharif Pilot Crop Insurance Pro- need to be calculated and confirmed by the insurance gram  1 (AYII coverage for semicommercial/­ companies and their lead reinsurers. The main five-year progressive farmers), GoPunjab has subsequently insurance financial plan and budget for GoPunjab are advised the following farm size categories and presented in Table ES3, and in Chapter 6 these projec- premium subsidy levels: tions have been subjected to a further series of sensitivity a) Owner Farmers with <5 acres (100 percent pre- analyses. mium subsidy); and b) Owner Farmers between 5–25 acres (50 percent 6Chapter 6 includes a sensitivity analysis assuming higher average premium subsidy) rates of 7.5 percent for Kharif crops and 5.0 percent for Rabi crops. Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan xxiii TABLE ES3:  PUNJAB FIVE-YEAR CROP INSURANCE PORTFOLIO SHOWING ESTIMATED NUMBER INSURED FARMERS, INSURED CROP AREA, SUM INSURED AND PREMIUM INCOME (US$) Program / Item FY2018/19 FY2019/20 FY2020/21 FY2021/22 FY 2022/23 Total Number of Insured Farmers: Program 1. Crop AYII for Commercial Farmers > 2.5 Acres 600,000 1,000,000 1,250,000 1,425,000 1,500,000 5,775,000 Program 2. Crop AYII Catastrophe Cover Small Farmers < 2.5 Ac. 750,000 1,750,000 2,750,000 3,500,000 8,750,000 Program 3. Crop Insurance for Tree Fruit (Mango, Citrus) 2,500 5,000 7,500 10,000 25,000 Total Insured Farmers 600,000 1,752,500 3,005,000 4,182,500 5,010,000 14,550,000 Insured Area (Acres) Program 1. Crop AYII for Commercial Farmers > 2.5 Acres 2,100,000 3,500,000 4,375,000 4,987,500 5,250,000 20,212,500 Program 2. Crop AYII Catastrophe Cover Small Farmers < 2.5 Ac. 750,000 1,750,000 2,750,000 3,500,000 8,750,000 Program 3. Crop Insurance for Tree Fruit (Mango, Citrus) 6,250 12,500 18,750 25,000 62,500 Total Insured Area (Acres) 2,100,000 4,256,250 6,137,500 7,756,250 8,775,000 29,025,000 Sum Insured (US$) Program 1. Crop AYII for Commercial Farmers > 2.5 Acres 717,500,000 1,207,500,000 1,522,500,000 1,741,250,000 1,837,500,000 7,026,250,000 Program 2. Crop AYII Catastrophe Cover Small Farmers < 2.5 Ac. 125,000,000 300,000,000 475,000,000 612,500,000 1,512,500,000 Program 3. Crop Insurance for Tree Fruit (Mango, Citrus) 6,250,000 12,500,000 18,750,000 25,000,000 62,500,000 Total Sum Insured (US$) 717,500,000 1,338,750,000 1,835,000,000 2,235,000,000 2,475,000,000 8,601,250,000 Premium Income (US$) Program 1. Crop AYII for Commercial Farmers > 2.5 Acres 30,362,500 51,712,500 65,887,500 75,643,750 80,062,500 303,668,750 Program 2. Crop AYII Catastrophe Cover Small Farmers < 2.5 Ac. 5,125,000 12,750,000 20,375,000 26,687,500 64,937,500 Program 3. Crop Insurance for Tree Fruit (Mango, Citrus) 625,000 1,250,000 1,875,000 2,500,000 6,250,000 Total Premium Income (US$) 30,362,500 57,462,500 79,887,500 97,893,750 109,250,000 374,856,250 The proposed five-year crop insurance program The estimated costs of GoPunjab financial sup- plan for Punjab is very ambitious. For the AYII port to the crop insurance program is estimated program for semicommercial/progressive farmers, it is at US$239 million over five years (Table ES4). This assumed that at full-scale implementation by year five, includes premium subsidies of US$220 million and subsi- about 750,000 farmers would be insured in both the dies on start-up and operating costs of US$19 million. By Kharif and Rabi seasons respectively (or 1.5  million year five (FY2022/23) which is assumed to be full-scale farmers in total per year) (Table ES3). This represents implementation, the annual cost of government support a penetration (uptake rate) of about one in every four to the program will be about US$73 million per year. (26 percent) of all semicommercial/progressive farmers. For Program 2, the macro-level fully funded AYII pro- gram for subsistence farmers, the assumed uptake rate by year five would be 1.75 million farmers per season WORK PLAN AND TIMETABLE (Kharif and Rabi) or 3.5 million farmers per year, equiv- alent to an uptake rate of nearly 80 percent of this group FOR AYII PROGRAM of farmers, each with less than 2.5 acres. Finally, it is esti- FOR SEMICOMMERCIAL/ mated that 10,000 fruit and vegetable farmers might be insured by year five. PROGRESSIVE FARMERS IN KHARIF 2018 The crop insurance budget shows that by year The final Chapter 7 sets out a detailed work plan five (FY 2022/23) an estimated 8.75 million hect- and timetable for the major activities that need ares will be insured on an annual basis (including to be carried out in the design and planning of both Kharif and Rabi seasons) with an estimated the AYII program between August 2017 and total sum insured of US$2.475  billion and esti- March 2018, the planned launch date of the pro- mated premium income of US$109 million. This gram. The work plan identifies responsibilities for each represents a very significant requirement for underwrit- organization or stakeholder which will be involved in this ing capacity, and the local insurers or pool will need to public-private partnership agricultural insurance initia- attract significant support from international reinsurers. tive for Punjab. The six-month timeframe to complete all xxiv A Feasibility Study TABLE ES4:  INDICATIVE COSTS OF GOPUNJAB PREMIUM SUBSIDY SUPPORT AND OTHER FINANCIAL SUPPORT TO CROP INSURANCE PROGRAMS 2018/19 TO 2022/23 (US$) Program / Item FY 2018/19 FY 2019/20 FY 2020/21 FY 2021/22 FY 2022/23 Total Premium Subsidies (US$) Program 1. Crop AYII for Commercial Farmers > 2.5 Acres 15,181,250 25,856,250 32,943,750 37,821,875 40,031,250 151,834,375 Program 2. Crop AYII Catastrophe Cover Small Farmers < 2.5 Ac. 5,125,000 12,750,000 20,375,000 26,687,500 64,937,500 Program 3. Crop Insurance for Tree Fruit (Mango, Citrus) 312,500 625,000 937,500 1,250,000 3,125,000 Sub-Total Premium Subsidies 15,181,250 31,293,750 46,318,750 59,134,375 67,968,750 219,896,875 Other Financial Costs borne by Government (US$) Data strenthening for Crop Insurance 1,500,000 1,000,000 750,000 500,000 500,000 4,250,000 Strenthen Crop Cutting Experiments (mobile phone system) 300,000 1,200,000 1,500,000 1,800,000 2,100,000 6,900,000 Farmer insurance awareneness, education and training 1,200,000 1,000,000 1,000,000 1,000,000 1,000,000 5,200,000 Monitoring and Evaluation 250,000 400,000 400,000 400,000 1,000,000 2,450,000 Sub-Total Other costs 3,250,000 3,200,000 3,250,000 3,300,000 4,600,000 18,800,000 Total Budgeted Costs to Government of Punjab 18,431,250 34,493,750 49,568,750 62,434,375 72,568,750 238,696,875 Cost per insured farmer 30.7 19.7 16.5 14.9 14.5 16.4 design and planning tasks and to put in place all insur- Chapter 7 also highlights some of the operational issues ance operating systems and procedures is very tight and and challenges that have arisen during the implementa- will require that all stakeholders ensure that the tasks tion of the Kharif 2018 pilot crop insurance program. and activities allocated to them are completed on time. Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan xxv CHAPTER 1 INTRODUCTION AND OBJECTIVES OF THE STUDY This report presents the findings and recommendations of a feasibility study for the introduction of large-scale crop and livestock insurance in Punjab, which is an essential component of the GoPunjab’s strategy to transform agriculture in the province within the next five years. This introductory chapter sum- marizes the origins and objectives of the feasibility study. It describes the critical local and national importance of agriculture in Punjab and the factors that suppress growth in agricultural productivity, including the highly destabilizing effects of climatic risk. It moves on to discuss the importance of making crop and livestock insurance avail- able on a wider scale for Punjab’s farmers (who are mostly small-scale farmers). Better access to insurance could shield farmers from climatic risks, bolster agricultural pro- ductivity and investment, and contribute to the government’s strategy to foster agri- cultural transformation. The chapter concludes by outlining the topics and analysis presented in subsequent chapters of the report. THE IMPORTANCE OF AGRICULTURE 1.1.  AND AGRICULTURAL GROWTH IN PUNJAB Punjab—the most densely populated province of Pakistan—consists of 36 administrative districts and covers 205,344 square kilometers. A large pro- portion of this area is arable, and because several major tributaries of the Indus River traverse the province from North to South, Punjab is one of the world’s most heavily irri- gated crop-producing areas. The total cultivated area in Punjab is 16.5 million hectares, of which 14.3 million hectares (87 percent) are irrigated (Government of Punjab 2015). Agriculture is a key economic sector in Punjab. The agricultural crop, live- stock, fisheries, and forestry sectors in Punjab account for 26 percent of Punjab’s gross domestic product (GDP) and 40 percent of employment. Punjab delivers more than half of Pakistan’s total GDP and population and, according to the 2010 census, has 5.2 million farming households—63.5 percent of Pakistan’s farmers. It is important to note that the average farm size in the province is only 6.5 acres, and nearly two-thirds of farmers own or cultivate less than 2 hectares (5 acres). Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 1 BOX 1.1: AGRICULTURE’S LACKLUSTER PERFORMANCE IN PUNJAB Nationally, agriculture accounts for 21 percent of GDP, employs 44 percent of the labor force, and directly and indirectly delivers nearly 80 percent of the total value of Pakistan’s exports. Yet as noted, growth in Pakistan’s agricultural sector fell from 3.3 per- cent over the last decade to nearly zero in FY2016, before recovering in FY 2017. Crop and livestock productivity are lower than in other Asian countries. Except for maize, crop yields have barely risen in decades. In Punjab, growth in agriculture has been similarly low and highly erratic as well. This performance has occurred even though Punjab has a vast area of fertile land, about 14.3 million hectares under irrigation, and a wide array of natural resources and cli- matic conditions capable of supporting diversified and productive agriculture. Still, 90 percent of cultivated land remains used for under five major crops (wheat, rice, cotton, sugarcane, and maize), leaving only about 10 percent for horticulture and other high- value crops. Punjab has approximately 73 percent of Pakistan’s national cropped area and 78 percent of national irrigated area. Approximately 60 percent of the province’s cultivated area lies within the Indus Basin Water System. The lack of progress in agriculture in Punjab has numerous causes. Perhaps most fundamentally, it reflects low growth in pro- ductivity at the farm level, which leads to high unit production costs and a lack of competitiveness, distorted cropping patterns, limited diversification into higher-value crop and livestock activities, and expanding populations of animals that are relatively unproductive. In crop production, large gaps exist between average yields, the yields obtained by progressive farmers, Punjab’s potential yields, and the world’s best averages. Agricultural growth is also held back by poor adoption of modern technologies, poor service delivery, and poorly functioning agricultural markets. Punjab could restore its agricultural competitiveness through innovations that renew growth in on-farm productivity and improve efficiency and quality throughout the post-harvest value chain. At 0.18 percent of agricultural gross domestic product (AgGDP), Pakistan’s public expenditures on agricultural research are the lowest in a region that is already lagging behind others. Most agricultural research expenditures still go to food grains, sugarcane, and cotton, rather than to high-value crops and livestock products. Few resources are dedicated to post-harvest man- agement, including value addition, quality, food safety, and nutrition. A high payoff could be gained by redirecting public expen- ditures and associated policies toward the best potential investments for outcomes, with a focus on reforms in wheat, irrigation, subsidies, and marketing, and concomitant investments to improve service delivery, agricultural research and development, and insurance. Punjab is Pakistan’s leading agricultural prov- per year over the last decade to below 3 percent during ince, a major producer of a wide range of food, 2011–15, and it fell to zero in 2016. Both crop and live- industrial, and horticultural crops, as well as stock productivity are low compared to levels achieved milk and meat. In 2016–17, Punjab produced 19.6 mil- by semicommercial/progressive farmers and farmers in lion tons of wheat, equivalent to 74 percent of wheat pro- other Asian countries, and growth in yields is largely flat duction in Pakistan. It is also a major producer of maize for major crops such as wheat and rice.7 (81 percent of national production), rice (51 percent of total production), and other food crops such as jowar (sor- Agricultural growth in Punjab is held back pri- ghum), bajra (millet), and gram. In that same year, Punjab marily by significant intervention from the Gov- also produced most of the nation’s cotton crop (6.9 mil- ernment of Punjab (GoPunjab) in both input and lion bales, or two-thirds of national production), sugar- output markets, and a pattern of public spending cane (65  percent of national production), and rapeseed on agriculture that is dominated by subsidies, and mustard (72 percent of national production). The estimated at US$1.25 billion in Punjab for fis- province is a very important producer of horticultural cal year (FY) 2017. Other reasons for low agricultural crops (fruits and vegetables), accounting for 97 percent of growth include limited adoption of modern technologies, total citrus production, 76 percent of all guava produc- poor service delivery, inefficient irrigation water delivery tion, and 75 percent of all mango production. Livestock and pricing, and poorly functioning agricultural markets production for both milk and meat is very important in (World Bank 2017). Climatic risks also threaten Punjab’s Punjab. Virtually all farming families own some livestock: capacity to increase agricultural growth in significant ways, there are about 14 million cattle (49 percent of the total as the next section will show. in Pakistan), 16 million buffaloes (65 percent of the total), 29 million sheep (24 percent of the total) and 68 million goats (37 percent of the total) (GoPunjab 2015). 7Maize is one exception: cultivated area, production, and yields have all increased significantly in recent years owing to strong demand for maize from At the same time, agricultural growth is lagging the animal and poultry feed industries and the adoption of high-yielding in Punjab (Box 1.1). Growth declined from 3.3 percent hybrid maize varieties. 2 A Feasibility Study 1.2. CLIMATIC RISKS Sind and Punjab provinces suffered the most from the 2010 floods. Damages in Sind were val- TO CROP AND ued at PKR 371.3 billion (US$4.4 billion) or 43.6 per- cent of the total value of damage (ADB and World LIVESTOCK PRODUCTION Bank 2010). Damages in Punjab were estimated at IN PUNJAB PKR 219.3 billion (US$2.6 billion) or 25.7 percent of total damage. The costs of reconstruction were esti- Punjab, like the whole of Pakistan, remains mated at between PKR 93.5 billion (US$1.1 billion) highly exposed to climatic risks. Pakistan is ranked and PKR  117.6 billion (US$2.1 billion) (ADB and among the top ten most climate vulnerable countries in World Bank 2010). the world in the Global Climate Risk Index. The country has witnessed an increase in the frequency and intensity Current government programs operate in the of extreme weather events and natural disasters, which wake of natural disasters to compensate the have caused significant agricultural losses. A recent study population that has been affected, but budget by the World Bank established that the melting of the constraints mean that those programs pro- Hindu Kush-Karakoram-Himalayan glaciers could affect vide compensation for a mere fraction of the water flows into the Indus River system, with implications agricultural production that is lost. Natural for agricultural production.8 These disturbances—which disaster management is coordinated at the national expose the agricultural sector to greater uncertainty and level by the National Disaster Management Author- risk, and whose effects are being exacerbated by rapid ity (NDMA) and implemented in the provinces by population growth, increasing water scarcity, and uncon- the Provincial Disaster Management Authorities trolled urbanization—call for concerted efforts to help (PDMAs). An example from Punjab demonstrates agriculture adapt and remain resilient. the limitations on disaster relief compensation in agriculture. PDMA Punjab reported that 3.3 million Several major climate-related risks affect Pun- acres of crops were damaged in the severe floods of jab and unleash natural disasters. Risks include 2010–13. These crop losses were estimated at about drought in the rain-fed (barani) areas of the province, US$1.6 billion by the World Bank, yet over that same localized damage from windstorms and hail, and incur- period, compensation payments to farmers in Punjab sions by crop and livestock pests and diseases. Monsoon amounted to only US$67 million10 or just over 4 per- flooding can result in costly major losses of crops and cent of the total flood losses. livestock. Severe floods not only drown animals but can cause them to starve when fodder supplies and grazing areas are destroyed and animal diseases break out.9 The 2010 floods in Pakistan affected more than 20 million 1.3. AGRICULTURAL people (over one-tenth of Pakistan’s population), with over 1,980 reported dead and nearly 2,946 injured; about INSURANCE 1.6 million homes were destroyed, and thousands of acres FOR CROPS of crops and agricultural lands were damaged. The total direct and indirect flood damage was estimated at PKR AND LIVESTOCK 854.8 billion (US$10.1 billion at the exchange rate pre- vailing at the time), and half of that damage was incurred IN PAKISTAN by agriculture, livestock, and fisheries. Costs of recon- Private commercial agricultural crop and struction were estimated at between PKR  577.9 billion livestock insurance is poorly developed in (US$6.8 billion) and PKR 757.8 billion (US$8.9 billion), Pakistan, and few farmers in Punjab are while costs of reconstruction in the agricultural sector insured against the loss of crops and live- were estimated at between PKR 21.9 billion (US$257 stock to major floods. The largest crop insurance million) and PKR 89.1 billion (US$1,049 billion). program in Pakistan is the Crop Loan Insurance Scheme (CLIS), introduced in 2008 by the State Bank of Pakistan (SBP) in conjunction with a group of pri- vate insurers. The CLIS is a “catastrophe loss of yield 8Yu et al. (2013). 9The earthquake risk is significant but mainly causes loss of life and damages 10Dataprovided by the Provincial Disaster Management Authority of urban private and public infrastructure and property rather than agriculture. Punjab (PDMA Punjab). Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 3 coverage” that is triggered when a disaster is declared by The “Strengthening Markets for Agricul- the provincial and/or district authorities and when crop ture and Rural Transformation in Punjab” losses exceed 50 percent of normal expected production ­ Program-for-Results (SMART Punjab PforR) is and yields. Small-scale farmers11 cannot obtain seasonal a US$300 million loan from 2018 to 2022 by the crop loans unless they participate in CLIS. SBP fully sub- World Bank to the GoPunjab to assist its pro- sidizes the premiums for these farmers up to a maximum grams to transform agriculture for the prov- premium rate of 2.0 percent of the sum insured. The ince’s 5.2 million mainly smallholder farmers. main criticism of CLIS is that although it protects the The SMART Punjab PforR assists the GoPunjab’s pro- lending institutions against farmers who default on their gram aimed at increasing crop and livestock productivity, loans in times of catastrophic losses, the insurance pro- promoting diversification, adding value, increasing pri- gram does not directly benefit small farmers themselves. vate sector engagement in farm and nonfarm businesses, deregulating crop and livestock markets, and enhanc- In Punjab about 700,000 farmers (loanees) are ing resilience. The World Bank’s Board of Directors insured under CLIS, representing about 13 percent approved the SMART Punjab PforR on December 15, of the 5.2 million farmers in the province. Since 2017, and it became effective on February 2, 2018. 2011 SBP has offered a livestock investment mortality cov- erage that also attracts premium subsidies. Finally, there By focusing on three results areas, the SMART have been a few private sector crop index insurance pilot Punjab PforR will contribute to the following programs and livestock insurance initiatives. outcomes: »» Results Area 1 (increased on-farm produc- tivity and value of crops and livestock) • Increased crop productivity. THE SMART PUNJAB 1.4.  • Increased livestock productivity. PROGRAM • Improved functioning of the agriculture research system. TO TRANSFORM • Removal of wheat market distortions and shift AGRICULTURE from wheat to high-value agriculture. »» Results Area 2 (increased value addition GoPunjab aims to stimulate growth in the agri- and competitiveness of crops and livestock) cultural sector by facilitating increases in crop and • Increased value addition of agricultural products. livestock productivity, enhancing resilience, and increas- • Improved employment opportunities in value ing competitiveness in agricultural marketing and trade, adding. by providing a conducive climate for private investment • Deregulation of crop and livestock markets. and improving supply chains and value addition. • Improved food safety. »» Results Area 3 (enhanced resilience of To achieve these aims, GoPunjab has allocated a smallholder farmers to climate change US$3.9 billion capital budget for agriculture and and natural disasters) irrigation over five years (FY 2018 to FY 2022), • Improved sustainability of irrigation systems. out of which US$1.6 billion is allocated to agriculture. • Improved access to crop and livestock insurance. In addition, GoPunjab spends about US$1.25  billion »» Enhanced resilience of farmers to climate change. per year on subsidies in the agricultural sector, including input subsidies on the cost of seed and fertilizer, subsi- dies on irrigation water pricing, subsidies on credit and power, and a major program for subsidized wheat pro- 1.5. GOVERNMENT curement based on minimum prices that are well above OF PUNJAB REQUEST world market prices for wheat (World Bank 2017). TO THE WORLD BANK GROUP FOR TECHNICAL 11Defined as farmers who own up to 25 acres of land (increased in 2015 ASSISTANCE from 12.5 acres) and who borrow seasonal production loans from a financial institution to cultivate any of five major crops (wheat, rice, maize, sugarcane, As part of its commitment to transform the cotton). agricultural and rural sectors, GoPunjab seeks 4 A Feasibility Study to provide better access to improved agricultural (3) provide options and recommendation on the ways in crop and livestock insurance products and pro- which GoPunjab could support the implementation of grams for the province’s 5.2 million farmers, at crop and livestock insurance in Punjab; (4) identify a five- affordable premium rates. The GoPunjab is seek- year implementation plan for agricultural crop and live- ing solutions for all segments of the farming population, stock insurance in Punjab and provide estimates of the including the important group of semicommercial/­ financial costs of GoPunjab premium subsidies and other progressive farmers operating on a small scale (with financial support to these programs; and (5) develop pro- 2.5 acres to 25 acres), who account for about 56 percent posals for a pilot crop insurance program for implemen- of all farmers in Punjab. This group could potentially tation early in the course of the SMART Punjab PforR. benefit from a strategy to link crop insurance with the pro- vision of credit, which will enable them to invest in fertil- izer and improved seed of high-yielding varieties, which in turn should increase their production and incomes and 1.7. ORGANIZATION further the government’s aim of transforming agriculture in Punjab. GoPunjab is also seeking solutions for small OF THIS REPORT subsistence farmers who have less than 2.5 acres and The remainder of this report is structured account for 42 percent of the farmers in Punjab, and for around the objectives of the feasibility study that commercial fruit and vegetable farmers. Finally, GoPun- have just been cited. Chapter 2 provides an overview jab is seeking improved livestock insurance products for of crop and livestock production in Punjab Province and producers of meat and dairy products. presents an assessment of the main risk exposures faced by farmers in the province. Chapter 3 reviews the crop and livestock insurance products and schemes that are GoPunjab aims to achieve rapid development currently implemented in Pakistan, highlighting their and massive uptake of crop and livestock insur- issues and challenges. That chapter also contains an ance in Punjab. GoPunjab has signaled its willingness analysis of the flood disaster compensation payments to provide major financial support for this large-scale made to farmers by PDMA Punjab. Chapter 4 presents agricultural insurance program by subsidizing premi- options and proposals based on international best prac- ums and providing financial support for the program’s tice for new crop and livestock insurance products, and start-up costs and ongoing operational costs. programs for the GoPunjab to consider. A review of the legal, institutional, and operational considerations and options for Punjab follows in Chapter 5. Chapter 6 out- lines a five-year plan and budget to build up a crop and 1.6. SCOPE AND OBJECTIVES livestock insurance program; it also presents estimates of OF THIS FEASIBILITY the potential costs of GoPunjab financial support to this program for premium subsidies and subsidies on insur- STUDY ance start-up and operating costs. Finally, Chapter 7 deals with the next steps in designing and implementing This agricultural insurance feasibility study was agricultural insurance in Punjab, starting with the launch conducted as part of the process for preparing the in Kharif season 2018 of a crop area-yield index insur- SMART Punjab PforR. The objectives of the feasibil- ance (AYII) program for small-scale semicommercial/ ity study, which started with a diagnostic mission focusing progressive farmers, as well as a review of some of the on agricultural insurance in Punjab in April 2017, were key challenges identified in launching and implement- to: (1) assess the existing crop and livestock insurance ing the pilot program in Kharif 2018. Several technical programs in Punjab and the potential for improvement; annexes provide extensive background data for reference (2) identify suitable crop and livestock insurance programs as well as information on international experience with to meet the needs of each segment of the farming popu- agricultural insurance programs. lation according to the priorities identified by GoPunjab; Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 5 6 A Feasibility Study CHAPTER 2 KEY FEATURES OF AGRICULTURE IN PUNJAB AND THE AGRICULTURAL IMPACTS OF CLIMATIC AND NATURAL DISASTERS This chapter delineates the key features of agriculture in Punjab to provide an understanding of the context in which the feasibility of a large-scale crop and livestock insurance initiative is being explored. For the same reason, the discussion also focuses on the major types of climatic and natural disasters that lead to agricultural losses in the province, along with data on their magnitude and impacts. A DENSELY POPULATED PROVINCE 2.1.  WHERE SMALL FARMS PREDOMINATE Punjab, with approximately 110 million people, is the most densely popu- lated of Pakistan’s six provinces. Lahore is the provincial capital. For administra- tive purposes, Punjab is divided into nine divisions and 36 districts (Figure 2.1). Each district is divided into smaller subdivisions or tehsils: there are 127 tehsils in Punjab.12 The next administrative tier below the tehsil is the union council (sherwan), and below that is the village. Punjab has 5.25 million farms (63.6 percent of the 8.26 million farms in Pakistan), and most of them are small.13The 2010 Agricultural Census reported an average farm size in Punjab of 5.6 acres: the small size of farms is illustrated by the fact that 91 percent of farms are under 12.5 acres (5 hectares), and 41  percent are smaller than 2.5 acres (1 hectare) (Table 2.1). According to the same census, 82 percent of the 5.25 million farms in Punjab are owner occupied; a further 9 percent are oper- ated by owners-cum-tenants, and another 9 percent are operated by tenants. 12Source: https://www.punjab.gov.pk. 13The latest Agricultural Census (2010) is available at http://www.pbs.gov.pk/content/agricultural-census- 2010-pakistan-report Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 7 FIGURE 2.1: PUNJAB: DISTRICT MAP Source: PDMA. In Punjab, the fact that so many farms are Chapter 4 presents a review and proposals of the types small presents major challenges for identifying of crop insurance products and programs that may be (1)  cost-effective delivery channels for credit, suitable for most of the very small farmers in Punjab, and farm inputs, and crop insurance and (2) suit- Chapter 5 examines the options for delivering these crop able crop insurance products for small farmers. insurance programs to small farmers cost effectively. 8 A Feasibility Study TABLE 2.1: PUNJAB: NUMBER AND AREA OF FARMS BY SIZE OF FARM Number Percent Percent of Average farm Farm size (acres) of farms of farms Farm area (acres) farm area size (acres) <2.5 2,203,102 42% 2,602,187 9% 1.2 2.5 < 5.0 1,144,394 22% 3,906,900 13% 3.4 5.0 < 12.5 1,411,353 27% 10,458,728 35% 7.4 12.5 < 25 360,467 7% 6,315,269 21% 17.5 25 < 100 121,741 2% 4,536,823 15% 37.3 100 & > 8,771 0% 1,986,621 7% 226.5 Total 5,249,828 100% 29,806,528 100% 5.6 Source: 2010 Census data reproduced in Agricultural Statistics of Pakistan 2014–15. Note: Farm area has been converted by the authors from hectares into acres (1 hectare = 2.47 acres). 2.2. CROP AND LIVESTOCK Punjab’s irrigation infrastructure consists of 14 headworks and barrages that feed 21 differ- PRODUCTION IN PUNJAB ent main canals. These canals, with their branches, run almost 4,000 miles to deliver water to more than MAJOR FOOD AND CASH 2.2.1.  2,000 distributaries and minor canals. This vast network CROPS AND TRENDS IN CROP channels water to 20 million acres of irrigable land in the province. The fact that agriculture in Punjab is PRODUCTION AND YIELDS mainly irrigated means that the exposure to drought is Punjab has two main crop growing seasons, the relatively lower in Punjab than in other provinces. On Kharif summer monsoon season, followed by the other hand, crops and livestock in much of Punjab the winter dry Rabi winter season. The sowing are highly exposed to floods during the Kharif mon- of summer (Kharif) crops starts in February for sugar- soon season. cane, March–May for cotton, June–July for rice, and July–August for maize. The harvesting of Kharif crops starts in September and continues to December, except Wheat is the most important crop grown in for sugarcane, which can be harvested until March or even Punjab during the Rabi season, with total cul- beyond. Rabi crops, including most importantly wheat tivated area of close to 17 million acres, fol- and barley, are sown during October–­ December and lowed by cotton (5.7 million acres) and rice harvested during March–April. The planting of orchards (4.5 million acres). The other main crops (based and other trees is carried out in spring (­ February–March) on area cultivated in 2015–16) include gram (2.1 mil- or during the monsoon (July–August). Wheat, cotton, lion acres), maize (1.8  million acres) and sugarcane rice, sugarcane, and maize occupy most of the cropped (1.7 million acres). Wheat area has increased by about land and are categorized as major crops. The remaining 0.5 million acres since 2010–11. The maize area has crops, grown on smaller areas, are categorized as minor also increased significantly by about 0.4 million acres crops. because of demand from the animal and poultry feed industry. Gram and rice area declined slightly over the ten years from 2006–07 to 2015–16 (Figure 2.2 and Rabi is the principle cropping season, when Annex 1). 9.4  million hectares (of which 83 percent is irrigated) are cultivated. Wheat is the main Rabi crop, accounting for nearly three-quarters of this cultivated area. In the Except for maize, agricultural production and Kharif monsoon season, 6.4 million hectares are cultivat- yields have by and large stagnated over the past ed,14 and supplementary irrigation is available on nearly 10 years or more in Punjab. Annex 1 contains statis- 94 percent of the cropped area. tics on the 10-year (2006–07 to 2015–16) annual average yields in kilograms per acre at provincial and district lev- els in Punjab for the five major crops of wheat, cotton, 14The main Kharif crops include cotton (34 percent of cultivated area), rice rice, maize, and sugarcane, along with measures of the (20 percent of area), and fodder (15 percent of area). variation in mean annual yields as given by the standard Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 9 deviation and coefficient of variation (CV).15 In the case For cotton and maize, however, yields have been of wheat, the average provincial production is about much more variable at the district level over the 18.4 million tons over the past 10 years, and yields have past ten years in Punjab. In the case of cotton, the remained at an average of 1.1 metric ton (MT) per acre CV in provincial-level annual yields has been 12.6 per- over this period. Rice yields are also low at 0.78 MT/acre cent between 2006–07 and 2015–16, yet in individual and again have not increased significantly over the past districts, yields have been very variable, with CVs rang- decade, while average maize yields increased from about ing from a low of 9.3 percent in Layyah to a high of 2.0 MT/acre to 2.4 MT/acre over the corresponding 52  percent in Jhelum. In 2015–16, when the cotton period (Figure 2.2 and Annex 1). Reasons for this stag- crop in Punjab was severely affected by adverse climatic nation in the production of traditional food crops such conditions, the average yield for the province was only as wheat and rice appear to include the land tenure sys- 549 kg/­ acre—about one-third lower than in the previous tem, which features a very high proportion of absentee year. Maize yields have also been more variable, as shown landowners and land leased to tenant farmers and share- by the CV of 10.4 percent in provincial yields and CVs croppers; declining soil fertility; lack of access to credit; ranging from a low of 4.6 percent in Sahiwal District to suboptimal use of fertilizer and plant protection chemi- a high of 70.6 percent in Bahawalpur District. The main cals; and government policies focused on subsidies rather reason for the higher variability in maize yields is the than investments in research, extension, irrigation main- introduction of improved high-yielding varieties (mainly tenance, and other productivity enhancing activities. hybrids) and the significant increase in average yields over the past five years. In the case of Rabi winter wheat, which is mainly irrigated, the provincial CV is only 4.9  percent, and the range in district level CVs is from a low of 2.2.2. LIVESTOCK PRODUCTION 4.2 percent in M.B.Din District to a high of 27.5 percent IN PUNJAB in Chakwal District (Figure 2.3 and Annex 1). For wheat, Livestock production is very important to the economy the highest CVs are encountered in the most northerly of Pakistan. The agricultural sector contributes 21 per- parts of Punjab Province in Rawalpindi Division, where cent of the GDP of Pakistan: the relative contribution of only 11 percent of the sown area is irrigated and which the livestock sector is 56 percent compared to 44 percent includes the districts of Rawalpindi (4 percent area for the crop sector. While the crop sector experienced irrigated), Chakwal (6 percent area irrigated), Attock major contractions in FY16 with a negative growth (12 percent area irrigated), and Jhelum (33 percent area (–6.25 ­percent)—mainly because of major losses in cot- irrigated) (Figure 2.4 and Annex 1). In most districts with ton caused by adverse weather (–21.26 percent growth)— assured irrigation, however, the CV in annual wheat the livestock sector has consistently grown over the past yields is extremely low, between 5 percent and 10 per- seven years and recorded growth exceeding 3 percent cent. There is a high degree of correlation (Pearson R² = percent in most years (Table 2.2). 73 percent) between higher variability in district wheat yields and the percentage of sown area which is unirri- Livestock production is very important in Pun- gated (rain-fed) in the district (Annex 1). The differences jab. Punjab Province has a cattle herd of 14.4 million between crop yield production and variability over time head, or nearly half (49 percent) of all cattle in Pakistan. under irrigated and rain-fed conditions have major impli- Punjab also accounts for 65 percent of all buffaloes, cations for the design and rating of an area-yield index 24 percent of sheep, 37 percent of goats, and 35 percent insurance product, as explained in detail in Chapter 4. of poultry in Pakistan (Table 2.3). Rice yields have also been very stable over the Historically, productivity in the livestock sector past 10 years with a CV of 4.2 percent in provin- of Punjab has been low in terms of meat and cial yields. At the district level the range in CVs is from milk output. Because GoPunjab recognizes the impor- a minimum of 4.1 percent in Kasur to a maximum of tance of livestock in the livelihoods of its farmers, and 14.8 percent in Lodhran. especially of poor landless households, it is implementing key initiatives to raise the sector’s productivity. These ini- tiatives include the “Save Buffalo-Calf ” program, a calf fattening program, the provision of poultry units, free 15The CV is equal to the standard deviation divided by the average (mean) livestock vaccination programs, and registration of cat- and expressed as a percentage. tle farmers and their cattle, among others. In addition, 10 A Feasibility Study FIGURE 2.2:  PUNJAB: AREA, PRODUCTION, AND YIELDS FOR MAJOR CROPS, 2006/07–2015/16 Punjab: Cultivated area major crops (“000” acres) 20,000 18,000 16,000 Wheat 14,000 Gram 12,000 Cotton 10,000 Rice 8,000 Maize 6,000 Sugarcane 4,000 2,000 0 1 2 3 4 5 6 7 8 9 10 Punjab: Production major crops (“000” tonnes) 50,000 45,000 40,000 Wheat 35,000 Gram 30,000 Cotton 25,000 Rice 20,000 Maize 15,000 Sugarcane 10,000 5,000 0 1 2 3 4 5 6 7 8 9 10 Punjab: Average yields major crops (kilograms/acre) 3,000 2,500 Wheat 2,000 Gram 1,500 Cotton Rice 1,000 Maize 500 0 1 2 3 4 5 6 7 8 9 10 Source: CRS-DoA, GoPunjab, April 2017. Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 11 Sown area and irrigated area (hectares) Average yield and standard deviation (kg/acre) 12 1,600 0 200 400 600 800 1,000 1,200 1,400 Th e Pu 0 500 1,000 1,500 2,000 2,500 3,000 3,500 nj C ab ah kw Bahawalpur Division 4.9% al At . Bahawalpur R to aw ck . 27.5% Bahawalnagar al pi Is nd la i. FIGURE 2.3:  FIGURE 2.4:  R. Y. Khan m 24.7% 21.0% ab ad D.G. Khan Division Jh . el 15.8% D. G. Khan um 21.0% Si . Layyah al ko M t. Muzaffargarh ia 15.5% nw R N ali. Rajanpur ah a 12.5% im row Source: CRS-DoA, GoPunjab, April 2017. Faisalabad Division Ya al. TO 2015–16 rK 11.6% Source: GoPunjab (Bureau of Statistics), 2014. Faisalabad ha n G Chiniot Sh ei uj ra kh t. Jhang up ur 10.3% 10.1% a. Toba Tek Singh M ul ta Gujranwala Division R IRRIGATED, 2013–14 aj n. an pu Gujranwala Sown area (000 hectares) r 9.0% 8.6% 8.5% La . Gujrat ho Average yield (kg/acre) Kh re. Hafizabad us h D .G ab. 8.3% 8.1% Mandi Baha-ud-Din .K ha Narowal Lo n. dh Sialkot r Sa an. rg Lahore Division od T. ha T. . Lahore Si ng h 7.4% 7.2% 7.2% 7.1% Kasur La . yy ah Nankana Sahib Ve . G ha Standard deviation (kg/acre) Sheikhupura uj ra ri. N 6.9% 6.6% Multan Division an nw ka al na a. Multan Ba Sah ha ib Khanewal w Irrigated area (000 hectares) a H lpu Lodhran af iz r. a Vehari Fa ba is d. Ba al Rawalpindi Division ha aba w al d. 6.4% 6.4% 6.4% 6.3% 6.0% Rawalpindi na ga r. Attock M Jh Cofficient of variation (%) uz an Chakwal af g. fa rg Jhelum Kh arh 5.6% 5.6% 5.3% an . Sahiwal Division e Pa wa kp l. Sahiwal at t Bh an. Okara ak ka Pakpattan r C . hi ni Irrigated area (% of total) Sargodha Division Sa ot Sargodha hi w al Bhakkar Ka . su r. 5.1% 5.1% 5.0% 5.0% 4.9% 4.9% Khushab O WHEAT IN PUNJAB: DISTRICT-LEVEL AVERAGE YIELD AND STANDARD ka Mianwali M r a .B . .D DEVIATION (KG/ACRE), AND COEFFICIENT OF VARIATION FROM 2005–06 Islamabad in PUNJAB: SOWN AREA, IRRIGATED AREA, AND PERCENTAGE SOWN AREA . 4.5% 4.2% 0 0 5 20 40 60 80 30 10 15 20 25 100 120 Percent of sown area irrigated (%) Coefficient of variation (%) A Feasibility Study TABLE 2.2: PAKISTAN: AGRICULTURAL GROWTH PERCENTAGES (BASE = 2005–06) Sector 2009–10 2010–11 2011–12 2012–13 2013–14 2014–15 2015–16 2016–17 Agriculture 0.23 1.96 3.62 2.68 2.5 2.53 0.15 2.07 Crops –4.16 0.99 3.22 1.54 2.64 1.04 –5.27 0.91   i) Important crops –3.74 1.5 7.87 0.17 7.22 –0.52 –5.86 2.18   ii) Other crops –7.24 2.27 –7.52 5.58 –76 3.09 0.4 –2.66   iii) Cotton ginning 7.29 –8.48 13.38 –2.9 –1.33 7.24 –22.12 5.58 Livestock 3.8 3.39 3.99 3.45 2.48 3.99 3.36 2.99 Forestry –0.07 4.76 1.79 6.58 1.88 –10.43 14.31 –2.37 Fishing 1.4 –15.2 3.77 0.65 0.98 5.75 3.25 1.23 Source: Pakistan Bureau of Statistics. TABLE 2.3: PUNJAB: LIVESTOCK AND POULTRY NUMBERS, 1960–2006 2006 as a percentage of total Animal 1960 1972 1976 1986 1996 2006 livestock, Pakistan Cattle 9,673 8,226 8,108 8,817 9,382 14,412 49% Buffaloes 6,129 7,413 7,979 11,150 13,101 17,748 65% Sheep 5,583 6,280 8,037 6,686 6,142 6,362 24% Goats 2,973 5,943 7,767 10,755 15,301 19,831 37% Camels 266 365 338 321 187 199 22% Horses 226 264 286 245 181 163 47% Asses 897 1,063 1,139 1,657 1,948 2,232 52% Mules 23 20 29 36 57 63 36% Poultry 6,440 8,688 13,783 27,848 24,511 25,906 35% Source: Agricultural Statistics of Pakistan 2014-15. SMART Punjab contains a number of livestock-related domestic private banks17 with a 21 percent market share. policy reforms, such as the removal of price caps on milk The former national agricultural development bank and meat, the shifting of public funding for animal health (Zarai Taraqiati Bank Limited, ZTBL) had a 15 percent away from curative care and toward preventive care, and market share, followed by microfinance institutions. The an increased investment in breeding stocks of local ani- Punjab Provincial Cooperative Bank Limited (PPCBL) mal breeds. disbursed loans of PKR 10.3 billion in 2015–16 (1.7 per- cent share). 2.3. ACCESS TO According to ZTBL, in Pakistan the gap between the demand for agricultural credit and the sup- AGRICULTURAL CREDIT ply of credit has remained substantial over the years. This gap was about PKR 438 million in 2012–13. In Pakistan, public and private bank lending to In 2015–16, ZTBL estimated that the total demand for the agricultural sector has tripled over the past agricultural credit was PKR 1,100 billion, against an eight years, and in FY2015/16 lending stood at PKR 598.3 billion. Table 2.4 shows that the main lend- ers were five commercial banks16 accounting for 52 per- 17The 14 domestic private commercial banks are: (1) Askari Commercial cent of all agricultural loans in 2015/16, followed by 14 Bank, (2) Bank Al-Habib, (3) Bank Al-Falah, (4) My Bank, (5) Faysal Bank, (6) Habib Metropolitan Bank, (7) PICIC Commercial Bank, (8) KASB Bank, (9) Prime Commercial Bank, (10) Saudi Pak Commercial Bank, (11) Soneri 16Allied Bank Limited (ABL), Habib Bank Ltd (HBL), MCB Bank Limited Bank, (12) Bank of Khyber, (13) Bank of Punjab, and (14) Standard (MCB), National Bank of Pakistan (NBP), and United Bank Ltd. (UBL). Chartered Bank (Pakistan). Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 13 TABLE 2.4:  PAKISTAN: SUPPLY OF AGRICULTURAL CREDIT BY LENDING INSTITUTIONS (PKR BILLION) 5 major commercial Islamic ZTBL banks 14 DPBs PPCBL MFBs banks Year [1] [2] [3] [4] [5] [6] Total 2007–08 66.94 94.75 43.94 5.93 211.56 2008–09 75.14 110.67 41.63 5.58 233.01 2009–10 79.01 119.61 43.78 5.72 248.12 2010–11 65.36 140.31 50.19 7.16 263.02 2011–12 66.07 146.27 60.88 8.52 12.11 293.85 2012–13 67.07 172.83 69.27 8.30 18.77 336.25 2013–14 77.92 195.49 84.81 8.81 22.80 1.53 391.35 2014–15 95.83 262.91 108.71 10.49 32.95 4.99 515.87 2015–16 90.98 311.40 123.10 10.33 53.94 8.54 598.29 % 2015–16 15.2% 52.0% 20.6% 1.7% 9.0% 1.4% 100.0% Source: State Bank of Pakistan, Karachi. [1] ZTBL—Zarai Taraqiati Bank Limited. [2] Includes ABL, HBL, MCB, NBP, and UBL. [3] DPB—Domestic Private Banks. [4] PPCBL—Punjab Provincial Cooperative Bank Limited. [5] Microfinance Bank included since July 2011. [6] Three Islamic banks included since July 2013. actual credit supply of PKR 598 billion (creating a gap other years, heavy flooding and rains were the major of PKR 502 billion or 46 percent of total demand). reason for rescheduling. A record number of bank offices had to reschedule loans because of floods in Limited access to production credit appears to ­ 2010—135  branches in five of Pakistan’s six provinces. be a major constraint on increasing farm invest- The highest value of agricultural loans that banks had to ments and thus increasing productivity and reschedule was in 2012 (Table 2.5). yield gains in Punjab. According to SBP, currently about 1 million small farmers with less than 25 acres— only 12 percent of all farmers in Pakistan—borrow from banks and are insured on a compulsory basis under 2.4. EXPOSURE OF CLIS. About 70 percent (700,000) of these farmers are AGRICULTURE TO estimated to be in Punjab. Several commercial banks reported that they were reluctant to lend to farmers CLIMATIC AND NATURAL because of the high historical default rates. DISASTERS Between 2007–16, because of natural calami- MAIN CLIMATIC AND NATURAL 2.4.1.  ties, 440 bank branches across Pakistan’s six RISK EXPOSURES provinces had to reschedule agricultural loans Punjab is susceptible to a variety of natural with a total value of PKR 11,632.5 billion (about disasters. Tornadoes, tropical cyclones, and earth- US$116 million). The two provinces that were most quakes all occur, although not frequently. The Murree severely affected were Punjab, where 44 percent of all Hills and parts of Islamabad and Rawalpindi are located reported branches had to reschedule loans, and Sind, on or close to fault lines that can cause earthquakes, but where 35 percent had to do so. In Baluchistan the major the rest of Punjab is considered relatively safe from earth- natural disasters causing loans to be rescheduled were quake threats. drought (in 2007) and a tropical cyclone (2008); in all 14 A Feasibility Study TABLE 2.5:  AGRICULTURAL LOAN RESCHEDULING: NUMBER OF BRANCHES RESCHEDULING BY PROVINCE AND VALUE OF RESCHEDULED LOANS (2007–16) Loan Amount Baluch- Gilgit- S. No. Year Punjab Sind KP AJK Total Rescheduled Reason for Calamity istan Baltistan (PKR Million) 1 2007 1 2 20 23 711.333 Flood, drought 2 2008 7 7 14 815.150 Flood, Tropical Cyclone 3 2009 1 2 9 12 1,788.701 Frost, Cloudy weather, rains 4 2010 38 49 14 25 9 135 1,305.995 Heavy Flood, Rains 5 2011 2 56 4 62 1,187.358 Heavy Flood, Rains 6 2012 10 27 7 44 2,336.852 Heavy Flood, Rains 7 2013 56 2 58 771.232 Heavy Flood, Rains 8 2014 55 10 65 714.280 Heavy Flood, Rains 9 2015 17 3 20 381.512 Heavy Flood, Rains 10 2016 7 7 1,621.049 Heavy Flood, Rains Total 194 156 45 36 9 0 440 11,633.462 Heavy Flood, Rains % of Total 44% 35% 10% 8% 2% 0% 100% Source: ZTBL 2017. Note: KP = Khyber Pakhtunkhwa; AJK = Azad Jammu and Kashmir. Punjab’s geographic location, climate, and major between April and June also cause lodging in the wheat river network make it very vulnerable to mon- crop, grain shedding at maturity, and damage to tree fruit soon flash floods and riverine floods. Punjab is the such as mangoes. Localized hailstorms also occur and basin for several major rivers that run North to South can severely damage cereal and horticultural crops. and originate in the Himalayas, including the Indus, Jhe- lum, Chenab, Ravi, and Sutlej rivers. Figure  2.5 shows the exposure in Punjab to flash flooding and riverine 2.4.2. HISTORICAL EXPOSURE flooding. The likelihood that floods will occur, as well TO NATURAL AND CLIMATIC as their intensity, have increased significantly in the past DISASTERS IN PAKISTAN decade or so. Punjab and other provinces of Pakistan experienced severe flooding in 2010, 2011, 2012, and AND PUNJAB 2014. The increased frequency and severity of flooding This section reviews the frequency and sever- is believed to be linked to climate change. ity of natural and climatic disasters in Punjab. The analysis draws on two data sources: (1) NDMA and PDMA Punjab data for the 40 years from 1973 to 2012 Drought has become a frequent phenomenon in (presented in World Bank 2015c) and (2) data collected by Pakistan. Drought is common throughout Pakistan; if the Centre for Research on the Epidemiology of Disas- the monsoon season fails to deliver rains, drought emerges. ters (CRED) for 1990–2017. The drought of 1998–2002, considered to be the worst in 50 years, was identified by the Economic Survey of Pakistan as one of the factors precipitating poor growth performance during that period. Baluchistan, especially NDMA/PDMA Punjab 40-year Data in its western and central areas, is very prone to drought. Between 1973 and 2012 in Pakistan, every year Punjab has two major sandy deserts, the Thal and Cho- an average of approximately 3 million people lishtan, which are susceptible to drought. Agriculture in were affected by natural catastrophes, predomi- the rest of the province is relatively secure against drought nantly floods. Floods affected 76.7 percent of all peo- because of its irrigation infrastructure. ple affected by natural catastrophes, followed in order of severity by droughts (13.5 percent of all people affected), Other climatic sources of crop losses include earthquakes (4.0 percent), windstorms (2.3 percent), and frost, wind, and hail. Frost in Punjab in December other perils such as avalanches and landslides (0.01 per- and January may induce crop losses and severely damage cent) (World Bank 2015c). fruit and other horticultural crops. Localized wind storms Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 15 FIGURE 2.5: PAKISTAN: FLOOD MAP A very high proportion (67 percent) of all people between 1900 and 2017, and floods were by far affected by natural disasters between 1973 and the most frequent type of disaster. Floods occurred 2012 were in Punjab (Figure 2.6). Punjab accounts 94 times (representing 47 percent of all occurrences of for approximately 56 percent of the population of Paki- natural disasters) over this period, and affected 79.3 mil- stan, which suggests that people in this province are more lion people (98 percent of all people affected by natural exposed to and affected by natural catastrophes than peo- disasters). Earthquakes were the second most frequent ple in the other provinces. event, with 31 occurrences (15 percent of total), but it was by far the largest cause of death, accounting for 143,734  fatalities (81 percent of all recorded deaths). The CRED EM-DAT International Disaster Windstorms accounted for 12 percent of the events and landslides 11 percent. Only one drought was reported Database in Pakistan over this 117-year period; in this regard, it According to CRED EM-DAT data, 203 occur- is important to note that drought is a slowly developing rences of natural disasters affected Pakistan 16 A Feasibility Study FIGURE 2.6:  PAKISTAN: GEOGRAPHIC DISTRIBUTION OF PEOPLE AFFECTED BY NATURAL DISASTERS, 1973–2012 AJ&K 1.5% FATA 0.2% Balochistan 4.3% KPK 7.3% Sindh 20.1% Punjab 66.6% Source: World Bank 2015c. Data from National and Provincial Disaster Management Authorities. Note: KPK = Khyber Pakhtunkhwa; AJ&K = Azad Jammu and Kashmir; FATA= Federally Administered Tribal Areas. peril that affects people’s livelihood over a longer period IMPACT OF NATURAL DISASTERS 2.4.3.  and is seldom reported as an isolated natural disaster (Figure 2.7 and Annex 2). ON THE AGRICULTURAL SECTOR IN PUNJAB The total estimated value of damage over the In Punjab, monsoon floods cause severe damage 117-year period was US$28.3 billion, and flood to agricultural crops grown in the Kharif sea- damage accounted for 74 percent of this value. son. Between 2010 and 2013, PDMA Punjab estimates The second most costly source of damage was earth- that floods affected 3.25 million acres of cropped area or quakes (19 percent), followed by storms (6 percent) and an average of nearly 815,000 acres per year. In the worst droughts (1 percent) (Figure 2.7 and Annex 2). flood year, 2010, monsoon floods affected 1.91 million acres or about 10.3 percent of all Kharif cultivated area in Punjab (Table 2.6).18 Figure 2.9 shows districts and The earthquake of 2005 was the single worst numbers of people affected by the 2010 floods in Punjab. event in terms of loss of life, with 73,338 reported deaths and economic losses of US$5.2 billion. The 2010 flooding was the worst event in terms of total The World Bank has estimated the value of lost crop pro- numbers of people affected (20.4 million) and was also duction (gross revenue) in Punjab due to flooding at an associated with the highest economic losses, estimated at average of PKR 40.7 billion (US$407 million) per year US$9.5 billion (Annex 2). between 2010 and 2013, with a maximum loss in 2010 of PKR 95.7 billion (US$957 million). This analysis assumes that (1) the crop area affected is 100 percent damaged, There is clear evidence that the frequency and and (2) the average gross revenue for all affected crops, severity of natural disasters are increasing including rice, maize, cotton, sugarcane, vegetable crops, over time in Punjab. Figure 2.8 shows that between and tree fruit, is PKR 50,000 per acre. Over this four- 1990–99 and 2000–09 the number of events dramat- ­ year period, total flood damage to crops is estimated at ically increased, along with the number of people affected, which rose to over 35 million in the most recent decade. The value of damage has also risen dramatically 18InPunjab the total cropped area in the Kharif season is about 7.5 million in recent decades. hectares or 18.5 million acres (Bureau of Statistics, GoPunjab 2015). Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 17 FIGURE 2.7:  PAKISTAN: RECORD OF DAMAGE BY TYPE OF NATURAL AND CLIMATIC EVENT, 1900–2017 Number of events Total number of people affected by type of event 100 94 90,000,000 90 79,332,048 80,000,000 80 70,000,000 70 60,000,000 60 50,000,000 50 40 40,000,000 31 30,000,000 30 25 22 20 17 20,000,000 2,604,586 10 10 10,000,000 7,275,388 80,574 2 1 1 34,154 16,486 0 2,200,000 0 0 d ke m e re ic y) ht n d ke m e re ic y) ht n oo id io oo id io em em or dr ug or dr ug tu tu ua ua sl at sl at Fl Fl St ra t( St ra t( ro ro nd nd id id q st q st pe en pe en rth rth D D Ep Ep fe fe La La m em m em in in Ea Ea te te ct ct ov ov se se e e m m m m In In tre tre s s as as Ex Ex M M Total number deaths Total value of damage (US$000) 160,000 143,734 25,000,000 140,000 20,969,178 120,000 20,000,000 100,000 15,000,000 80,000 60,000 10,000,000 40,000 5,329,755 5,000,000 20,000 17,248 11,969 1,715,036 789 2,774 283 63 143 18,000 18,000 0 247,000 0 0 d e m e re ic y) ht n d e m e re ic y) ht n oo ak id io oo ak id io em em or dr ug or dr ug tu tu sl at sl at qu qu Fl Fl St t( St t( ra ra ro ro nd nd id id st st en en pe pe rth rth D D Ep Ep fe fe La La em em m m in in Ea Ea te te ct ct ov ov se se e e m m m m In In tre tre s s as as Ex Ex M M Source: CRED EM-DAT. PKR 163 billion (US$1.6 billion). This loss is very sub- EFFECTS OF CLIMATE CHANGE 2.4.4.  stantial for the province and its farmers, especially small ON AGRICULTURAL CROP subsistence farmers (Table 2.7). Chapter 3 presents an analysis of the actual flood compensation paid to farmers PRODUCTION AND YIELDS by GoPunjab during the corresponding period. IN PUNJAB It is now widely accepted that climate change threatens the The floods of 2010 caused livestock producers through- stability of agricultural productivity and output as well as out Pakistan to lose about 1.2 million animals, including the increases in productivity and output required to meet poultry, and severely affected producers in Punjab (see future demand. Between 2010 and 2050, it is estimated Figure 2.9 showing the extent of the 2010 floods in Pun- that the world’s population will increase from 6.7  bil- jab). In 2011, flooding mainly in Baluchistan and Sind lion to 9 billion, mostly in South Asia and Sub-Saharan provinces led directly to the death of 115,500 livestock Africa. As a result, total agricultural production will need but adversely affected a further 5 million animals due to increase by an estimated 70 percent over that period. to migration from the flood-affected areas, disease out- The long-term changes in patterns of temperature and breaks, and other related events. precipitation that are part of climate change are expected 18 A Feasibility Study FIGURE 2.8:  PAKISTAN: ANALYSIS OF NATURAL AND CLIMATIC DISASTER DAMAGE RECORDS BY DECADE, 1990–2017 Number of occurrences Total number people affected 80 40,000,000 70 35,000,000 60 30,000,000 50 25,000,000 40 20,000,000 30 15,000,000 20 10,000,000 10 5,000,000 0 0 9 9 30 29 40 39 50 49 60 59 70 69 80 79 90 89 00 99 10 09 7 9 9 30 29 40 39 50 49 60 59 70 69 80 79 90 89 00 99 10 09 7 90 91 01 90 91 01 19 –19 19 –19 19 19 19 –19 19 –19 19 19 19 –19 20 19 20 20 19 –19 19 –19 19 19 19 –19 19 –19 19 19 19 –19 20 19 20 20 –1 –1 –2 –1 –1 –2 – – – – – – – – 00 10 20 00 10 20 19 19 19 19 19 19 Total deaths Total damage ($,000) 90,000 20,000,000 80,000 18,000,000 70,000 16,000,000 60,000 14,000,000 50,000 12,000,000 40,000 10,000,000 8,000,000 30,000 6,000,000 20,000 4,000,000 10,000 2,000,000 0 0 9 9 30 29 40 39 50 49 60 59 70 69 80 79 90 89 00 99 10 09 7 9 9 30 29 40 39 50 49 60 59 70 69 80 79 90 89 00 99 10 09 7 90 91 01 90 91 01 19 –19 19 –19 19 19 19 –19 19 –19 19 19 19 –19 20 19 20 20 19 –19 19 –19 19 19 19 –19 19 –19 19 19 19 –19 20 19 20 20 –1 –1 –2 –1 –1 –2 – – – – – – – – 00 10 20 00 10 20 19 19 19 19 19 19 Source: CRED EM-DAT. TABLE 2.6: PUNJAB: LOSSES AND DAMAGES CAUSED BY FLOODS, 2010–13 Cropped Area No. of District Villages Area Affected Houses Year Persons Affected Affected Person Died Affected Affected (Acres) Damaged (Acres) 2010 11 1,810 5,038,992 3,471,109 1,914,104 379,520 258 2011 12 335 26,393 136,758 125,513 1,284 4 2012 12 1,271 887,345 1,490,827 473,998 67,324 60 2013 22 2,994 184,147 945,541 745,655 20,411 111 Total 57 6,410 6,136,877 6,044,235 3,259,270 89,019 175 Annual Average 14 1,603 1,534,219 1,511,059 814,818 29,673 58 Source: PDMA Punjab 2014. TABLE 2.7: PUNJAB: ESTIMATED VALUE OF FLOOD DAMAGE TO AGRICULTURE 2010–13 Cropped area Value of crop loss Total value of crop Value of crop Year affected (acres) (PKR/acre) losses (PKR) losses (US$) 2010 1,914,104 50,000 95,705,200,000 957,052,000 2011 125,513 50,000 6,275,650,000 62,756,500 2012 473,998 50,000 23,699,900,000 236,999,000 2013 745,655 50,000 37,282,750,000 372,827,500 Total 3,259,270   162,963,500,000 1,629,635,000 Annual Average 814,818   40,740,875,000 407,408,750 Source: Authors’ analysis of PDMA Punjab flood damage data. Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 19 FIGURE 2.9: PUNJAB: DISTRICTS FLOODED IN 2010 AND NUMBER OF PEOPLE AFFECTED to shift production seasons, alter pest and disease patterns, wheat yields. In the case of rice and sugarcane, increas- and modify the set of crops that can be produced. All of ing temperature and relative humidity are associated these changes will affect production, prices, incomes, and with higher yields in these crops. Overall, climate change ultimately livelihoods and lives (FAO 2010). has adverse impacts on the yields of major food crops. The authors note that almost 60 percent of Pakistan’s Pakistan is vulnerable to climate change, includ- population is living below the poverty line, and as the ing increased temperatures and more extreme population is growing rapidly, the country may face food droughts, rainfall, and flooding. There have been security challenges in the near future. several studies on the effects of climate change on crop production and yields. Ali et al. (2017) examined the Siddiqui et al. (2012) conducted a separate study effects of climate change (such as maximum and mini- of how changes in climate change indicators may mum temperatures, levels of rainfall and relative humid- affect production of four major crops in Punjab. ity, and sunshine hours) on the major crops grown in Their results show that in the short run the increase in Pakistan, including Rabi wheat, Kharif rice, maize, and temperature is expected to reduce wheat yields, but in sugarcane. Their findings differ for each crop type. For the long term the increase in temperature has a positive wheat, they found that increasing temperature leads to effect on wheat productivity. The increase in precipita- a significant yield reduction, and that excessive rainfall tion, however, has a negative impact on wheat yields in and relative humidity are also negatively correlated with both the short and long terms. A rise in temperature is 20 A Feasibility Study beneficial for rice production initially, but beyond a cer- climate-smart agriculture are likely to lower emissions in ­ tain optimal temperature, further increases in tempera- from agriculture. ture become harmful for rice production. Interestingly, the increase in precipitation does not seem to harm rice The role of climate risk insurance, including agricultural productivity. It has been evident that the change in cli- insurance, as part of both a small farmer development mate variables (temperature, precipitation) has a signif- strategy and a climate smart adaptation strategy has been icant negative impact on production of cotton. Finally, enshrined under the 2016 Paris Agreement on Climate the increase in temperature also harms sugarcane pro- Change. The Paris Agreement identified a number of ductivity in the long term. areas of cooperation between the international commu- nity of nations, including: early warning systems; emer- gency preparedness; slow onset events; comprehensive POTENTIAL ROLE FOR 2.4.5.  risk assessment and management; and risk insurance AGRICULTURAL INSURANCE facilities, climate risk pooling, and other insurance solu- IN A CLIMATE CHANGE tions. Climate risk insurance is seen as a tool that can ADAPTATION STRATEGY help the rural poor, including small farmers, to address loss and damage from the extreme weather events (such SMART Punjab is expected to generate consid- as storms, floods, or droughts) that are increasing in erable climate adaptation co-benefits. In Results frequency and severity due to climate change. Climate Area 1, reorienting and increasing funding to agricultural risk insurance can also contribute to building resilience research should result in more funding for agricultural (or adaptation), as resilience measures can be incorpo- research oriented toward climate resilience. Improve- rated into the design of the insurance, for instance by ments in livestock health should increase animals’ resil- providing incentives such as lower premiums for under- ience to heat stress and diseases. The modernization taking activities such as planting trees or using seeds of of the wheat market includes improved wheat storage drought-resistant varieties. Specific initiatives, such as the facilities, which should protect emergency wheat stocks G7 InsuResilience initiative, have recently been launched from climatic effects, compared to the dilapidated stor- to increase by up to 400 million the number of people age facilities where stocks are currently stored. A shift to in the most vulnerable developing countries with access high-value agriculture will involve crop diversification, to direct or indirect insurance coverage against climate shorter growing cycles, and more efficient use of irriga- change hazards by 2020 (RESULTS 2016). tion, which should reduce farmers’ vulnerability to cli- mate change. Results Area 2 of SMART Punjab, which provides incentives to agribusiness for investment in value Linking or bundling agricultural insurance with credit addition, may finance improved storage facilities, which can improve small farmers’ access to loans to enable them should improve the climate resilience of stored agricul- to invest in productivity-enhancing and climate-smart tural products. Results Area 3, which improves the sus- agricultural technology. Agricultural insurance can be tainability of irrigation,19 should help producers adapt to a win-win for both the farmer and the lending institu- the impacts of climate change on water resources. The tion. Many lending institutions are reluctant to lend to introduction of agricultural insurance and climate-smart small farmers, whom they regard as poor risks; however, agriculture assists farmers to make their operations more when credit is bundled with a crop or livestock insurance resilient to adverse weather events. coverage, bank loans are protected against default in the event of major climate-induced crop failure or death of the animal. Where bundling is adopted, banks are gen- In addition, SMART Punjab is expected to gen- erally more willing to extend loans to small farmers (as erate climate mitigation co-benefits. Improved seen in India, Pakistan, Malawi, and Kenya). Farmers, in livestock breeding would reduce pressure on rangelands turn, benefit by gaining access to credit for investing in and benefit maintenance of carbon pools in rangeland often riskier but higher yielding seed and fertilizer tech- areas. A shift to high-value agriculture is expected to nology or in livestock breeds that can produce more milk, lead to reduced fertilizer use, and increased investments thereby benefiting from production and income gains, as well as the security that loans can be repaid if a crop 19 Irrigation is one of the most effective measures for enhancing crop fails or livestock die. Governments in many countries production and yields, leading to gains of up to 130 percent in crop actively promote compulsory crop or livestock insurance productivity over rain-fed cropping (FAO 2010). Irrigation can also help to for farmers who borrow from formal lending institutions. reduce such adverse effects of climate change as reduced precipitation and higher variability in rainfall. One example is India, where the National Agricultural Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 21 Insurance Scheme (NAIS) is mandatory for all such bor- invest in and adopt climate-smart technologies and farm- rowers (loanees), and another is Pakistan, where CLIS is ing practices that reduce risk, as they are insured against compulsory for small farmers to obtain seasonal loans climatic disasters. Little evidence in the insurance liter- (Chapter 3 reviews the performance of CLIS, and Chap- ature supports the argument that agricultural insurance ter 4 reviews that of NAIS). is a disincentive to the adoption of climate-smart tech- nology, but it will still be important that in Punjab the It will be very important to ensure that the proposed expansion of agricultural insurance is accompanied by expansion of agricultural crop and livestock insurance in insurance literacy campaigns and training and educa- Punjab does not act as a disincentive for government and tion about the role of insurance. It is also important for farmers alike to invest in climate-smart technology and GoPunjab to recognize that agricultural insurance is not practices. The potential downside of introducing agricul- a substitute for social protection systems nor for invest- tural insurance, especially where it is supply driven and ment in disaster risk reduction and climate change adap- heavily subsidized, is that farmers may be less willing to tation strategies. 22 A Feasibility Study CHAPTER 3 AGRICULTURAL INSURANCE PROVISION AND NATURAL DISASTER RELIEF PROGRAMS IN PUNJAB The private sector insurance market is very small in Pakistan compared to neighboring countries. In 2014 the total life and non-life insurance premium in Pakistan was PKR 180 billion (US$1.8 billion), of which the life insurance market contributed 69 percent of premiums compared to the non-life market of only 31 per- cent. In Pakistan in 2014, the total expenditure on life and non-life insurance was US$9.60 per capita, which was equivalent to a market penetration of 0.71 percent of GDP, compared to an expenditure of US$488.96 per capita in Malaysia (4.34 per- cent of GDP), US$50.97 (3.14 percent of GDP) in India, and US$35.67 per capita (1.02 percent of GDP) in Sri Lanka (AXCO 2017). Agricultural insurance is relatively undeveloped in Pakistan. Livestock insurance was introduced on a pilot basis in 1983 by two private insurers, Adamjee Insurance Company and the Eastern Federal Union Insurance Company. Crop insur- ance is relatively new, dating from 2008 under the public-private partnership (PPP) that established CLIS as a scheme with national scope. 3.1.  CROP INSURANCE 3.1.1.  CROP LOAN INSURANCE SCHEME: KEY FEATURES Launched in the 2008 Kharif season, CLIS is an initiative of SBP. It is struc- tured as a PPP between SBP, 22 of 36 public and commercial banks and microfi- nance institutions (MFIs) lending to farmers, and a group of 14 insurance companies, including (among others) New Jubilee, EFU General, National Insurance Company, UBL, Adamjee, United, Silver Star, Atlas, and Alfalah. CLIS is currently implemented throughout Pakistan but is concentrated in Punjab Province, because of its leading role in agriculture. Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 23 As mentioned, CLIS is a federal crop-credit insur- The CLIS policy carries a maximum 2 percent ance scheme that is compulsory for all farmers commercial premium rate, which is fully subsi- who obtain seasonal production loans from a dized by the Government of Pakistan (GoP) for commercial bank to cultivate any of five major small and marginal farmers with up to 25 acres crops (wheat, rice, maize, sugarcane, and cot- of land. The insurance companies are free to set their ton). The sum insured is based on the value of the sea- own premium rates for the different crops in different sonal crop loan provided to the farmer. In 2016–17, the regions for damages/losses which exceed 50 percent of maximum loan amount was fixed at PKR 40,000 per acre the expected crop production and yields, subject to a for Kharif food crops (rice, maize) and PKR 30,000 per maximum cap of 2 percent for the operation of federal acre for Rabi crops (wheat). government premium subsidies. The GoP, through SBP, pays 100 percent of the premiums for small and mar- The CLIS policy insures again multiple perils ginal farmers (defined as farmers owning or cultivating that cause yield losses in crops: excessive rain, flood, up to 25 acres; this limit was raised from 12.5 acres in drought, hailstorm, frost, crop pests (such as locusts), and 2015). Farmers with more than 25 acres who take out crop viral and bacterial diseases. seasonal loans do not qualify for any government pre- mium subsidies. The lending banks are responsible for paying the premiums to the insurance companies at the The CLIS is also a catastrophic crop insurance time of inception of coverage. The banks then reclaim product that pays out only if damage or losses the premium subsidies from SBP against the provision of as declared by the government are in excess of premium bordereaux providing evidence of each loanee, 50 percent of crop production and yields in the their sum insured, premium rate, and amount of pre- defined area (the district, tehsil, or village, for mium paid by the bank to the insurance company(ies). example). The scheme carries a two-stage indem- nity procedure: first the Provincial Government/Board of Revenue20 have to declare an agricultural disaster A unique feature of CLIS is that it caps insurers’ (Calamity Declaration) where localized crop losses at and their reinsurers’ liability at a loss ratio of 300 a village level or subdistrict level exceed 50 percent of percent per cropping season—in other words, in the reference crop yield21 for that area. This declaration Kharif and Rabi seasons separately. The loss ratio must be notified in the Gazette. If a calamity is declared, is equal to the value of claims paid divided by the value of the lending banks and insurers are then responsible for written premium expressed as a percentage. The 300 per- assessing actual damage at the individual farmer level for cent loss ratio cap on the CLIS means that the maximum their own loanees/insured farmers. It is also important claims liability borne by an insurer and its reinsurer(s) to note that the CLIS policy is a Constructive Total Loss would be equal to three times the value of the premium Policy: if the assessed damage exceeds the 50 percent they have received. The government’s rationale for cap- trigger, the sum insured (amount of outstanding crop ping losses is (1) to encourage insurance and reinsurance loan owed by the loanee to the lending institution) is paid companies to participate in this catastrophic crop insur- in full to the lending institution, and the farmer’s out- ance program, and where the covariate risk exposure is standing loan is written off. This procedure contrasts with very high, especially for the perils of flood and drought; the more conventional approach taken by crop insurance and (2) given that the premium rate is capped at 2 percent, policies, which is to indemnify losses on a proportional the government decided to cap the losses as well. Further basis according to the actual amount of yield loss up to details of the CLIS are summarized in Table 3.1. a 100 percent loss, when a full payout or indemnity pay- ment is due. CROP LOAN INSURANCE SCHEME 3.1.2.  UPTAKE AND RESULTS Scaling Up the Scheme 20According to SBJ, up to 13 different government departments and the The CLIS has been operational for about a Disaster Relief Commissioner are responsible for conducting the estimation decade, and in 2016–17 the scheme insured about of post-event losses to assess whether a calamity should be declared. This area-based assessment takes into account the cause of loss and the loss of 1 million crop-credit recipients of whom about lives, damage to public property and private dwellings, and also damage to 70 percent (700,000 insured loanee farmers) were crops and livestock. in Punjab.22 Table 3.2 and Figure 3.1 show that over 21The reference yield for each crop is based on the average of annual yields in the middle three of the past five years, discounting the years with the highest and lowest yields (SBP 2008). 22Estimate provided by SBP, April 2017. 24 A Feasibility Study TABLE 3.1: SBP TASK FORCE: CROP LOAN INSURANCE FRAMEWORK Objective To providing insurance coverage to farmers in the event of failure of crops as a result of natural calamities, floods, rains, pests, and diseases Beneficiaries All borrowers availing Agriculture Production Loans from banks/MFBs Crops covered Wheat, rice, sugarcane, maize, cotton Coverage period The insurance coverage would be for the period from sowing/transplanting of the crop to its harvesting, except in the case of sugarcane Premium rate Maximum 2% per crop per season inclusive of standard levies Sum insured Sum insured will be based on the per acre borrowing limits prescribed by the State Bank of Pakistan, subject to a maximum amount agreed between the banks/MFBs and insurance company. Perils covered Indemnity would be payable on the occurrence of production loss due to: (a) natural calamities like excessive rain, hail-storm, frost, cyclone, flood, and drought, etc.; (b) crop diseases like viral and bacterial attacks, or any other damage caused to the produce by infestation like locust attacks, etc. Indemnification A valid claim (as mutually agreed between the bank and the insurance company) under the scope of coverage will be payable subject to the following: (i)  The insured crop is situated in an area declared as a calamity affected by the respective provincial government or revenue authority Damage to the crop was due to any of the insured perils (ii)  Main exclusions »» War, civil war, strikes, riots, terrorism, etc. »» No utilisation/sowing »» Earthquake or volcanic eruption »» Loss before risk declaration or after harvesting »» Price fluctuations and loss of market Premium payment The premium shall be paid up front by the banks in respect of farmers (with subsistence land holding or with land holdings up to 25 acres) at the time of disbursement of production loans. Payment of the claims Claims shall be payable to the banks/MFBs by the insurers for credit to the insured borrower’s loan account. Insurance companies to ensure payment to banks/MFBs within 30 days after notification of calamity for ultimate credit to the loan accounts of the borrowers. Source: SBP 2017. the eight-year period from 2008/09 to 2015/16 the pro- rates charged by CLIS insurers ranged between 1.5 per- gram insured 3.6 million farmers and generated a total cent and 1.7 percent.25 premium of PKR 3,661 million, or an average premium of PKR 1,102 per insured farmer. In 2015–16 the CLIS portfolio increased very significantly to insure 885,852 Underwriting Results for the CLIS insured farmers for a total premium of PKR 1,152 mil- In the eight years from 2008–09 to 2015–16, CLIS lion (about US$11.5 million). In 2016/17, SBP estimated paid claims valued at PKR 3.0 billion, with an that the number of CLIS insured farmers had risen to implied long-term average loss ratio of 81 per- 1 million, and that 70 percent (700,000) were in Pun- cent.26 In other words, once the insurance companies jab—in other words, about 13 percent of all farmers in have added their operating expenses, the CLIS only Punjab were insured under CLIS.23 Over this eight-year breaks even or is marginally profitable for them. So far, period the total sum insured (TSI) has amounted to PKR the worst losses occurred in 2010, when very serve mon- 242 billion, with an implied long-term average premium soon floods in Khyber Pakhtunkhwa spread southward rate of 1.53 percent and average sum insured per insured farmer of about PKR 67,000.24 The 2016–17 premium 25For example, ZTBL insures its crop production loan portfolio of about PKR 60 billion with three CLIS insurers, Adamjee Insurance Company 23Adamjee Insurance Company estimates that CLIS farmers are distributed Limited, United Insurance Company Limited, and Asia Insurance Company as follows: Punjab (70–75%), Sind (10–12%), KPK (5–6%), Federally Limited, with an average rate of 1.7 percent for the five crops (wheat, maize, Administered Tribal Areas (2%), and Baluchistan (<1%). rice, sugarcane and cotton) (ZTBL 2017). 24SBP figures reported by DoA, GoPunjab, June 2017. 26SBP data reported by DoA, GoPunjab June 2017. Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 25 TABLE 3.2:  CLIS: NUMBER OF INSURED BORROWERS, PREMIUM, AND AVERAGE PREMIUM PER LOANEE PER SEASON AND YEAR, 2008/09–2015/16 Year/season No. of insured borrowers Premium amount (PKR) Average premium/borrower (PKR) Kharif 2008 94,386 44,385,009 470 Rabi 2009 200,928 112,534,187 560 Total 2008/09 295,314 156,919,196 531 Kharif 2009 191,442 135,795,437 709 Rabi 2010 246,026 143,714,756 584 Total 2009/10 437,468 279,510,193 639 Kharif 2010 214,157 147,222,573 687 Rabi 2011 277,414 203,073,769 732 Total 2010/11 491,571 350,296,342 713 Kharif 2011 207,873 148,750,526 716 Rabi 2012 246,538 208,084,877 844 Total 2011/12 454,411 356,835,403 785 Kharif 2012 188,148 132,698,994 705 Rabi 2013 224,317 211,779,219 944 Total 2012/13 412,465 344,478,213 835 Kharif 2013 165,678 145,942,057 881 Rabi 2014 176,324 225,697,113 1,280 Total 2013/14 342,002 371,639,170 1,087 Kharif 2014 141,048 207,326,337 1,470 Rabi 2015 188,104 442,248,104 2,351 Total 2014/15 329,152 649,574,441 1,973 Kharif 2015 351,741 394,031,203 1,120 Rabi 2016 504,111 757,800,665 1,503 Total 2015/16 855,852 1,151,831,868 1,346 Total Kharif 1,554,473 1,356,152,136 872 Total Rabi 2,063,762 2,304,932,690 1,117 Total all 8 years 3,618,235 3,661,084,826 1,012 Source: SBP data provided by DoA, GoPunjab, June 2017. through Punjab, Baluchistan, and Sind. At its worst, In Punjab between 2008 and 2015, CLIS has made the flooding covered about one-fifth of the land area payouts to the banks for a total of 54,445 farm- of Pakistan; 18 million people were affected, 12 million ers (loanees) with a total value of claims of PKR houses were damaged or destroyed, 2.2 million hectares 2,205 million. The average size of a claim over these of crops were damaged or destroyed, and 450,000 head eight years was PKR 40,508 per claimant (loanee farmer), of livestock were lost.27 The 2010 floods mainly damaged but Figure 3.2 shows the variation across years. In the crops grown in southern Punjab and Sind provinces. The worst year (2010), claims valued at PKR 711 million extent to which individual lending institutions incurred were settled for 20,723 loanees; other challenging years loss ratios in excess of 300 percent on their loan portfolios were 2011 (10,865 claims valued at PKR 430 million) in the 2010 floods is not known. Flood losses were also and 2012 (11,380 claims valued at PKR 377 million), fol- incurred in 2011, 2012, and 2014. lowed by 2013 (4,654 claims valued at PKR 344 million) and 2014 (5,625 claims valued at PKR 272 million). Since its inception, CLIS has been reinsured with 27See https://www.dec.org.uk/articles/pakistan-floods-facts-and-figures international reinsurers. HannoverRe and SwissRe 26 A Feasibility Study FIGURE 3.1:  CLIS: NUMBER OF INSURED BORROWERS AND PREMIUM INCOME, 2008–09 TO 2015–16 Number of insured farmers (borrowers) 900,000 1,400,000,000 800,000 1,200,000,000 700,000 1,000,000,000 600,000 Premium (PKR) 500,000 800,000,000 400,000 600,000,000 300,000 400,000,000 200,000 200,000,000 100,000 0 0 ta i 2 8 Kh 008 9 9 ta i 2 9 Kh 00 0 0 To ab 10 Kh 01 1 1 To ab 11 Kh 01 2 2 To ab 12 Kh 01 3 3 ta i 2 3 Kh 013 4 4 To ab 14 Kh 01 5 5 5 01 6 16 To ab 0 l 2 00 ar /0 To ab 0 l 2 01 ar 9/1 l 2 01 ar 0/1 l 2 01 ar 1/1 l 2 01 ar 2/1 To ab 1 l 2 01 ar /1 l 2 01 ar 4/1 To ab 1 l 2 01 R f 20 R f 20 R f 20 R f 20 R f 20 R f 20 R f 20 R f 20 5/ ta 2 ta i 2 ta i 2 ta i 2 ta i 2 i i i i i i i i i ar Kh Number of insured borrowers Premium amount (PKR) Source: SBP data provided by DoA, GoPunjab June 2017. FIGURE 3.2: NUMBER AND VALUE OF CLIS CLAIMS IN PUNJAB, 2008–15 25,000 800,000,000 20,723 700,000,000 20,000 600,000,000 Value of claims (PKR) Number of claims 500,000,000 15,000 11,380 400,000,000 10,865 10,000 300,000,000 5,625 200,000,000 4,654 5,000 1,060 100,000,000 70 68 0 0 09 10 11 12 13 14 15 08 20 20 20 20 20 20 20 20 ec ec ec ec ec ec ec ec D D D D D D D D n– n– n– n– n– n– n– l– Ju Ja Ja Ja Ja Ja Ja Ja Number of claims Claims value (PKR) Source: DoA, GoPunjab June 2017. are the main international reinsurers supporting CLIS, Pakistan’s 7 million farming households using and they again enjoy the protection of the capped liabil- formal credit were insured, in 2011 the GoP con- ity of a 300 percent loss ratio. stituted a task force under the SBP to formulate a National Agricultural Insurance Scheme (NAIS). Following the major 2010 floods, and recogniz- The NAIS was intended to extend CLIS coverage to ing that only a relatively small proportion of provide automatic catastrophe protection to all farmers Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 27 in Pakistan, whether they were loanees or non-loanees; e) The declaration by local government of insured crops would include wheat, rice, maize, cotton a disaster (calamity) appears to be sub- and sugarcane (PPAF 2012). The estimated cost of pre- jective and open to interpretation. The miums for the NAIS was about PKR 22 billion (US$220 original SBP framework for CLIS recommended million). The NAIS has not been implemented. that a calamity should be declared where actual yield losses in the defined area exceed 50 percent of the actual average yield for that area in the ISSUES AND CHALLENGES FACING 3.1.3.  past three out of five years. This actual objective THE CROP LOAN INSURANCE assessment of area yields does not appear to be carried out consistently, however. SCHEME f) Standardized or objective loss assessment The drawbacks of the CLIS include: procedures for the in-field measurement a) The direct beneficiaries of CLIS are the of actual damage by insurers and banks lending institutions, which receive the also appears to be lacking. The costs of payouts to offset loans to farmers. There- individual farmer field-level yield loss assessment fore, although small-scale farmers remain credit on very small farms is extremely high and time worthy, they have no protection for the loss of the consuming. Indeed, one MFI reported it did not additional costs they have incurred in growing adjust losses in the field, as it was cheaper to pay the crop or for the loss of crop income from the the full value of the sum of insured loans to the sale of their crop. bank. b) CLIS is a catastrophe insurance product g) Disputes appear to result in major delays that pays claims only when crop damage in settling some claims. According to SBP exceeds 50 percent of the normal or aver- data, over the eight years from 2008–09 to age yield in a defined area (such as the dis- 2015–16, the number of outstanding claims was trict, subdistrict, tehsil, or village) and when the 5,167 (9.5 percent of total claims), valued at PKR government formally declares that a calamity has 220 million (10.0 percent of the total value of occurred. If crop damage exceeds 50 percent, claims).28 the policy pays out the sum insured in full (as per h) Because losses are capped at a 300 per- a Constructive Total Loss policy). Where area cent loss ratio each season, the lending damage is estimated at less than 50 percent of institutions are highly exposed to losses crop production and yield, neither the bank nor arising from catastrophes. This point applies the insured farmer receives any form of compen- particularly to regional banks or MFIs which, sation. The potentially very significant risk with unlike larger lending institutions, cannot spread this provision is that bankers may shy away from their risk geographically across the entire country. crop loans. If losses exceed the 300 percent loss ratio, these c) The crop insurance policy does not lenders must bear the excess losses themselves. indemnify the farmer (loanee) for crop The lending institutions consulted in the course losses. Farmers who lose up to nearly half their of this feasibility study expressed their desire expected crop still have to repay their loans to the that government amend the CLIS to ensure that banks out of the sale of their remaining harvest, losses surpassing the 300 percent loss ratio would which leaves farmers with little or no surplus to be indemnified in full. ZTBL argues that the loss feed their families or to purchase seed and other limit should be raised to either 400  percent or inputs for the next crop season. 500 percent; they calculate that doing so over the d) The scheme insures losses only in five eight years from 2008–09 to 2015–16 would have major Rabi and Kharif food and cash added PKR 572 million to the actual CLIS claims crops and does not protect banks/farm- cost (in case of a 400 percent loss ratio cap) or ers for losses in high-value vegetable and PKR 1.146 billion (in case of a 500 percent loss tree fruit crops. In Punjab, commercial banks ratio cap). This analysis suggests that (1) many of lending to farmers growing high-value crops such the banks incurred uninsured losses in excess of as tobacco, potatoes, bananas, mangoes, or citrus cannot purchase coverage against excess rain and 28According to DoA-GoPunjab, these outstanding payments belong to NIC floods, frost, and hail in these crops. clients. 28 A Feasibility Study TABLE 3.3:  CLIS: INCREASED CLAIMS COSTS FOR 400 PERCENT OR 500 PERCENT LOSS RATIO CAP Additional Additional Number Claim funds required Claim funds required of borrowers Claim amount if indemnity if indemnity if indemnity if indemnity with claims (2009 to date) is 400 percent is 400 percent is 500 percent is 500 percent (2009 to date) (PKR million) (PKR million) (PKR million) (PKR million) (PKR million) 54,587 1,931 2,504 572 3,077 1,146 Source: ZTBL 2017. the 300 percent loss ratio limit over the eight-year lending institutions (providing livestock investment loans period and (2) premium rates would have to be to small-scale cattle, buffalo, and dairy producers) and significantly increased if the higher 400 percent the national government (providing premium subsidies or 500 percent loss limit were introduced to cover to those small-scale livestock producers). the increased claims cost (Table 3.3). i) The lending institutions are required to The GoP approved financial support to LIBS pre-finance the CLIS premiums and to in the form of a 100 percent premium subsidy pay these to the insurance companies at for a maximum premium rate of 4 percent for the time of coverage inception. The lend- small farmers, financing the purchase of up to ing institutions then have to reclaim the 100 per- ten cattle (or buffaloes). The maximum loan size or cent premium subsidies from SBP/GoP. In sum insured is PKR 5 million per producer. Loans above practice, however, the lending institutions have the threshold of 10 animals do not qualify for any SBP/­ experienced delays of between one and two years government premium subsidies, and the livestock pro- in being reimbursed the CLIS premiums by the ducer must pay the full premium. government, which is a major bone of contention for these lenders. Banks are implicitly subsidiz- ing the crop insurance program from their own The LIBS Policy is a standard individual animal funds, over and above the government’s premium indemnity-based product that insures against subsidy. the death of the animal due to named natural and climatic perils, accidental death, and vac- cination failure leading to death by disease. The policy insures adult cows, bulls, and buffaloes from the 3.2. LIVESTOCK INSURANCE ages of nine months to seven years. Animals must be properly tagged and vaccinated, and be in clean health SCHEME FOR to be insurable. The policy is an annual policy; Table 3.4 BORROWERS summarizes its key features. 3.2.1.  KEY FEATURES OF THE SCHEME The sum insured is based on the market price of Given the importance in Pakistan of livestock the animal or the amount of credit loaned by the beef and dairy production (which together repre- bank to purchase the animal. As noted, the maximum sent 55 percent of agricultural GDP and 11.4 per- sum insured is PKR 5 million (US$50,000) per beneficiary. cent of GDP),29 and to promote increased access Therefore, assuming an insurer charges the maximum for producers to livestock investment credit, 4 percent premium rate, the maximum premium subsidy in 2014 SBP launched a Livestock Insurance for a livestock producer with a PKR 5 million sum insured Scheme for Borrowers (LIBS) backed by govern- would be PKR 200,000 (US$2,000) per year. ZTBL (one ment premium subsidies. SBP adopted the same of the leading Livestock Insurance Scheme for Borrow- PPP model as for the CLIS, with private sector insurers ers [LISB] participants) is charged premium rates by the assuming the underwriting risk, working closely with insurers (United Insurance, SPI Insurance, and Asia Insur- ance) of 2.15 percent for local breeds and 4 percent for 29AXCO imported/hybrid animals (ZTBL 2017). 2017. Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 29 TABLE 3.4: FRAMEWORK OF LIVESTOCK INSURANCE SCHEME FOR BORROWERS Objective To improve access to finance the livestock and dairy sector and to mitigate risk of losses to farmers in case of death or disease of animals due to natural calamities and accidents Participants All banks involved in livestock lending and all insurance companies interested to participate Loan covered All livestock loans up to PKR 5 million for the purchase of animals Period of insurance Yearly renewal insurance Animals covered »» Cows, buffaloes, and bulls (age from 9 months to 7 years old) »» All imported animals as per the terms and conditions of underwriting guidelines of the participating insurance companies Insured perils »» Death due to disease/natural »» Death due to floods, heavy rains, windstorms, and drought »» Accidental death Indemnity »» Up to the insured amount of loan or individual price of animal as declared by bank »» Maximum sum insured is PRK 5,000,000 per borrower »» 20 percent compulsory deductible each and every claim Main exclusions »» Death due to Rinderpest, Blackquarter, Haemorrhagic Septicemia, Anthrax and Foot & Mouth disease if animal is not inoculated »» Pre-existing diseases or injury »» Change of location without prior permission in case of transport of animal by land vehicle beyond 25 km from the place of farming and others Premium rate »» Up to a maximum of 4% per annum of amount insured »» Bank will be responsible for collection and payment of premium to the insurer Claim process »» Insured/branch will inform immediately to the company via e-mail, phone call, SMS, writing, etc., and will wait for at least 24 hours before disposing of the carcass »» Insurance company shall arrange a veterinary doctor approved by Pakistan Veterinary Medical Council to investigate the cause of loss and issue a death certificate »» The insured/branch will submit the claim form duly stamped and signed within 14 days »» Insurance company shall settle the claim within 30 days of claim lodgment Payment of the claims Claims shall be payable to the bank by the insurers for credit to the insured borrowers’ loan account Source: SBP 2017. In the event of a claim, the insured/bank is which is extremely low for an individual animal mortal- responsible for notifying the insurer within ity insurance program. The TSI was PKR 28.7 million, 24 hours, following which the dead animal must against paid premiums of PKR 609.1 million, with an be inspected by an approved veterinary officer to average premium rate of 2.1 percent, fully subsidized by confirm the cause of death was due to an insured SBP/GoP. The paid claims (PKR 1.6 million) represent peril(s). All claims carry a deductible of 20 percent of an extremely low loss ratio of 0.27 percent. the sum insured, which is borne by the insured livestock producer. The deductible is set at this level to minimize The extremely favorable LISB underwriting moral hazard. results are not encountered anywhere else in the world on individual animal mortality insurance. While the number of natural disasters (floods, droughts, 3.2.2. UNDERWRITING RESULTS and so on) was low during 2014–16, the LISB’s extremely FOR THE SCHEME low losses are unique in the world of livestock insurance The LISB is one of the world’s most profitable and merit further investigation and validation. For com- livestock insurance programs. Data provided by parison, Table 3.6 presents data from five individual ani- ZTBL show that LISB, in its first three years of oper- mal livestock accident and mortality insurance schemes ation (2014–16), insured 258,800 animals (Table 3.5). in Bangladesh. Under the best-performing scheme During that period, claims amounted to only 32 ani- (PKSF, beef fattening livestock investment-credit insur- mals, representing a mortality rate of 0.012 percent, ance), the livestock mortality rate was 0.3 percent and 30 A Feasibility Study TABLE 3.5:  LIVESTOCK INSURANCE SCHEME FOR BORROWERS: UNDERWRITING RESULTS, 2014–16 Number Average of sum Average animals Number insured premium for which Loss insured Sum insured per animal Premium rate claim Paid claim ratio Year animals (PKR) (PKR) paid (PKR) (%) paid (PKR) (%) 2014 32,497 2,815,009,929 86,624 61,400,135 2.2%  8 210,662 0.34% 2015 112,576 12,755,781,262 113,308 270,104,752 2.1% 16 838,248 0.31% 2016 113,727 13,138,369,309 115,526 277,589,561 2.1%  8 576,466 0.21% Total 258,800 28,709,160,500 110,932 609,094,448 2.1% 32 1,625,376 0.27% Source: ZTBL 2017. TABLE 3.6:  BANGLADESH: RESULTS OF FIVE FORMAL AND INFORMAL LIVESTOCK (CATTLE) INSURANCE PROGRAMS Notes: [1] Traditional cattle accident and mortality insurance policy. [2] Livestock-credit insurance policy for cattle and shoats and poultry (up to 2 years). [3] Livestock-credit insurance policy for dairy cattle (up to 2 years). [4] Livestock-credit insurance for beef cattle fattening program (6 months cover). Source: World Bank 2015a. the loss ratio was 37 percent. For the other programs, crop insurance as a micro-insurance retail product tar- the livestock mortality rates ranged between 1.2 percent geted toward small farmers in many developing countries (SBC and Sajida) and a high of 3.5 percent (Proshika) of in Asia (including China, India, Indonesia, the Philip- all insured animals, and the loss ratios range from 58 per- pines, and Sri Lanka) and Africa (e.g., Kenya, Ghana). cent for SBC to 161 percent for the Sajida scheme. This section briefly reviews some of the limited number of private crop and livestock initiatives that have been piloted since 2012 in Pakistan. 3.3. INNOVATIONS IN CROP AND LIVESTOCK PILOTS OF CROP WEATHER INDEX 3.3.1.  INSURANCE AND AREA-YIELD INSURANCE INDEX INSURANCE The past decade saw very little innovation into In 2011, the Pakistan Poverty Alleviation Fund new crop or livestock indemnity-based or index- (PPAF) and the International Fund for Agricul- based insurance products offered by the private tural Development (IFAD) assisted local stake- sector in Pakistan. This lack of new products con- holders to design two projects to pilot crop trasts with the major expansion of weather index-based Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 31 micro-insurance products for small and mar- loss of yield policy designed to protect marginal farm- ginal farmers. These products included (1) a weather ers with less than 4 acres of land. The insurance product index-based crop insurance (WII) product for wheat and was introduced in two phases, starting with wheat and (2) an area-yield index insurance (AYII) product for wheat followed by cotton. The premium rate was PKR 1,080 and cotton. These were the first pilots of micro-level or per acre for wheat and PKR 1,533 per acre for cotton. individual farmer crop index insurance in Pakistan. Donors also provided a 70 percent premium subsidy to make the AYII policy more affordable to marginal farm- The PPAF WII product for producers of rain- ers. The pilot ultimately insured 14,095 farmers with a fed wheat was piloted in collaboration with the total insured area of 38,239 acres. Hailstorms and heavy World Food Programme (WFP) in two locations rainfall during the monsoon season resulted in 238 claims, in Punjab: Talang (Chakwal District) and Soon with payments valued at PKR 14.62 million.31 The cur- Valley (Khushab District). The wheat policy was a rent status of the AYII program is not known. rainfall deficit index coverage that extended over three vegetative phases from sowing (November/December) to crop maturity and harvest (March/April). The index PILOTS OF LIVE-WEIGHT 3.3.2.  was constructed on more than 30 years of daily rainfall LIVESTOCK INSURANCE data for single weather stations cited in each of the two AND INSURANCE FOR MILKING pilot locations. Two insurers agreed to underwrite the ANIMALS WII pilot, Alfalah and United Insurance, and SwissRe PPAF and IFAD developed two livestock insurance provided technical support and reinsurance capacity products: (1) Live-weight Livestock Insurance (PPAF 2012). and (2) Micro-insurance for Milking Animals. The live-weight livestock insurance, which was designed The wheat WII coverage was promoted and mar- for livestock fattening programs, was the first of its kind keted on a voluntary basis in the 2012 Rabi sea- in the world; its unique feature was that it linked the sum son to about 500 farmers in each of the locations insured and insurance payout in the event of loss to an by two local partners: the National Rural Support Pro- index of the weight gained by the animal during the cov- gram (NRSP) in Talang and the Soon Valley Develop- erage period. The coverage used historical data from the ment Program (SVDP) in Soon Valley. The total wheat government-owned livestock research institutes to serve area insured was 2,376 acres. The product carried a as the basis for determining growth rates (live-weight premium rate of PKR 531 per acre in Soon Valley and gain) for different species under different feeding regimes PKR 610 per acre in Talang. The donor (IFAD) agreed or forage-based management (PPAF 2012). The policy to subsidize 70 percent of the cost of premiums in the was offered for beef cattle, sheep, and goats. It carried a pilot start-up (NRSP 2017).30 6 percent premium rate and covered livestock mortality due to accidents, natural perils, poisoning, and diseases. In 2012 the wheat WII pilots were claims free, and Premiums cost PKR 800–1,250 per animal for cattle and farmers declined to renew coverage the next sea- 300–600 per animal for sheep and goats. The compensa- son. Based on discussions with PPAF, Alfalah Insurance, tion was linked to the actual weight of the animal at the and NRSP, it appears that the index product was difficult time of loss. The program was launched in 2013, and for farmers and communities to understand. Some farm- even though it was popular with farmers, it incurred high ers did not fully trust that data recorded by the weather losses32 and was withdrawn from the market after one station would accurately reflect their own rainfall condi- year. tions, and they preferred a conventional indemnity-based crop insurance product that would cover losses on their A specialized and modified livestock accident own farms (NRSP 2017). and mortality insurance product was also devel- oped for milking animals, along with a conventional The AYII pilot for wheat and cotton producers in dairy cattle insurance product for poor herders located Bahawalpur and Sanghar districts of southern in Sind and Punjab. The average premium charged on Punjab had more success. The AYII product, which was also launched in Rabi 2012, was a multiple peril 31The source for this information is an unnamed and undated Microinsurance Briefing Note. 32According to NRSP, 167 claims were processed, of which 152 were settled 30Unnamed and undated MicroInsurance Briefing Note. for a total payout to farmers of PKR 3.2 million. 32 A Feasibility Study milking animal insurance was about PKR 3,200 per ani- The next section briefly reviews the roles of the mal. PPAF supported by IFAD initially funded 70 percent National Disaster Management Authority (NDMA) of the premium subsidies (later reduced to 50 percent).33 and the Provincial Disaster Management Author- ities (PDMAs) in disaster risk management and The live-weight milking animal, and conven- relief in Pakistan. An understanding of these agencies tional livestock micro-insurance products were and their roles is essential to the discussion in the remain- piloted in 10 districts by about 15,000 insured der of this report of the role of agricultural insurance. livestock producers (of whom 40 percent were female herders) with a total of 70,528 insured ani- mals. Claims payouts (including claims related to flood 3.4.1. DISASTER MANAGEMENT and drought losses) amounted to PKR 10.73 million.34 AT THE NATIONAL LEVEL As noted, the live-weight gain product was withdrawn The NDMA is a federal agency created to manage after one year due to high claims. It is not known whether and coordinate disaster risk management in Paki- the milking animal insurance program and conventional stan. NDMA was established under the National Disaster livestock insurance program were also terminated. Management Act, 2010, and functions under the supervi- sion of the National Disaster Management Commission (NDMC), which is headed by the Prime Minister of the 3.4. NATIONAL AND Islamic Republic of Pakistan. NDMA manages the whole disaster management cycle, which includes preparedness, PROVINCIAL DISASTER mitigation, risk reduction, relief, and rehabilitation. MANAGEMENT The Global Fund for Disaster Reduction and PROGRAMS Recovery (GFDRR) of the World Bank Group has actively assisted NDMA in recent years to In Pakistan the catastrophe insurance market develop a national disaster risk financing strat- for risks like earthquakes and floods is relatively egy for Pakistan. This strategy follows an operational underdeveloped. According to AXCO35 2017, about framework of (1) assessing risk, (2) arranging finan- 70 percent of property fire policies also include coverage cial solutions, and (3) delivering funds to beneficiaries. for earthquakes but few commercial or private properties GFDRR has identified a series of seven strategic options are insured. Flood is perhaps the greatest natural haz- for designing a National Disaster Risk Financing strategy ard in Pakistan, causing considerable damage to property for Pakistan; they are listed in Table 3.7. and loss of life, but in most cases, losses resulting from floods are uninsured (Axco 2017; GFDRR 2015). 3.4.2. DISASTER MANAGEMENT As a consequence of the very poorly developed IN PUNJAB PROVINCE catastrophe insurance market in Pakistan, after GoPunjab established the PDMA in 2010. PDMA any disaster, the government has to bear the Punjab specializes in mitigation, preparedness, and orga- major share of the financial liability for relief nized response to a disaster. The most important role of and recovery activities, including private losses PDMA lies in providing a platform for all provincial depart- as well as public ones. Usually the GoP has to real- ments to come together and strategize management and locate development budgets for disaster management response to disasters and calamities. PDMA also acts as the and also has to depend on foreign aid for covering losses authority that coordinates at post-disaster rescue and rehabil- and reconstruction. After high-magnitude disasters, the itation operations at the provincial, district, and local levels. government faces major financial challenges in provid- ing assistance to the most vulnerable and poor urban and rural people in a timely fashion to restore infrastructure In the past 10 years, the GoPunjab has declared and services and their livelihoods. 70 disasters in Punjab, affecting 10,000 villages. In comparison, 10 disasters were declared in areas of Sind Province and 22 in Kyber Pakhtunkwa. Baluchistan was affected by drought for seven years (Saeed 2014).36 33Unnamed and undated Microinsurance Briefing Note. 34Unnamed and undated Microinsurance Briefing Note. 35AXCO Insurance Information Services (https://www.axcoinfo.com/). 36See http://news.trust.org/item/20140704055948-vj87s/ Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 33 TABLE 3.7: PAKISTAN: SEVEN OPTIONS FOR A DISASTER RISK FINANCING STRATEGY Time frame Options for disaster risk financing Short term Develop a central database of disaster losses and expenditures to better predict future 1.  financial costs of disaster Short term Operationalize the National and Provincial Disaster Management Funds 2.  Short term Clarify contingent liability associated with post-disaster cash transfer programs and 3.  enhance the programs’ financing sources to ensure efficient access to funds after a disaster Develop financial disaster risk assessment tools, including financial catastrophe risk Short to medium term 4.  models for the Ministry of Finance Develop models for improving financial response capacity to disasters Short to medium term 5.  Medium term Establish a robust catastrophe risk insurance program for public assets 6.  Medium to long term 7. Promoto property catastrophe risk insurance for private dwellings Source: World Bank 2015c. TABLE 3.8:  PUNJAB: PAYMENTS (PKR) TO DISTRICTS TO COMPENSATE FLOOD VICTIMS FOR LOSS OF LIVELIHOODS, HOUSING, AND CROPS, 2010–15 House Districts Deaths Injuries Livelihood damage Crops Total Total 13,311,350,000 21,133,860,000 6,704,201,163 41,149,411,163 Percent of total 32% 51% 16% 100% US$ equivalents 133,113,500 211,338,600 67,042,012 411,494,112 Annual average US$ 22,185,583 35,223,100 11,173,669 68,582,352 Source: Punjab Natural Disaster Management Authority 2017 (file: “Data.xls”). Between 2010 and 2015, PDMA Punjab paid PKR Under the compensation programs, farmers typ- 41.1 billion (US$411 million) in disaster compen- ically receive a cash payment to enable them to sation to people affected by flooding (Table  3.8). purchase seed and fertilizer and resume crop The bulk of the compensation—PKR 21.1 billion production, but the value of the compensation (51  percent)—was allocated to rebuild damaged hous- is much lower than the value of the lost crop ing. PKR13.3 billion (32 percent) was paid in compensa- revenue. For example, as of October 1, 2014, the tion for lost livelihoods, and farmers were compensated Khadim-e-Punjab Imdadi Package (KPIP) paid PKR ­ PKR 6.7 billion (16 percent) for the loss of their crops 10,000 per acre for damaged crop area, up to a max- (an average of PKR 1.12 billion (US$11.2 million) per imum of PKR  125,000 for 12.5 acres for any individ- year). Over the six-year period, GoPunjab issued com- ual farmer.37 Compare this payment to the World Bank pensation payments in 22 of Punjab’s 36 districts. Three estimate of the average value of crop output in Punjab districts together received more than half of those pay- (about PKR 50,000 per acre) in Chapter 2. ments: Muzaffargarh (30 percent), Rajanpur (17 per- cent), and Jhahg (11 percent) (Figure 3.3 and Annex 3). The gap between the budget for NDMA and Compensation specifically for crop losses went to farm- PDMA to compensate producers for crop dam- ers in 16  districts, with the highest payments going to ages/losses and the actual value of the damage Jhang District (24 percent of total crop compensation), followed by Muzaffargarh (19 percent) and Sargodha (10 percent). 37See PDMA Punjab website: http://pdma.punjab.gov.pk 34 A Feasibility Study FIGURE 3.3:  PUNJAB: DISTRIBUTION OF CROP COMPENSATION PAYMENTS (PKR) BY DISTRICT, 2010–15 1,800,000,000 30 1,600,000,000 25 24% 1,400,000,000 Compensation (PKR) 1,200,000,000 20 Percent of total 19% 1,000,000,000 15 800,000,000 600,000,000 10% 10% 10 400,000,000 6% 5% 5 4% 4% 4% 200,000,000 3% 1% 2% 2% 2% 3% 1% 0% 0% 0% 0% 0% 0% 0 0 r ra ot G a iz t Jh d Jh g an m di hu l ha ab M M in fa n N arh Sh go al kh a Si a a t r an M yah li aj n r an K ewa af a Bh ko pu D kka pu R wa al a an af lta ei dh ur R ha d Sa row uj ni H ujr Kh elu ab Ba sh Kh nw ud al al rg up an G hi uz u y K n La w ia C a G Y r ha Ba M Source: Data provided to World Bank by Punjab Disaster Management Authority April 2017. incurred by farmers is extremely wide. As noted Crop insurance may be one option for GoPunjab to con- in Chapter 2, floods affected 3.26 million acres of crops sider in trying to close this funding gap. in Punjab in 2010–13. GoPunjab would have needed to allocate PKR 32.6 billion (US$326 million) to com- No data are available on the GoPunjab expendi- pensate farmers for that flood-damaged area at the level ture on livestock in the aftermath of a natural of PKR 10,000 per acre (the compensation paid under disaster. Following a disaster, the Department of Live- KPIP). The actual compensation payments shown in stock and Dairy Development (DL&DD) conducts a Table 3.8 of PKR 6.7 billion (US$67.0 million) were rapid assessment of damages to livestock and provides equivalent to only 18.5 percent of the required amount.38 financial and technical resources to district livestock department offices for the immediate provision of medi- cal and material relief. The DL&DD coordinates the pro- 38This disaster-relief compensation funding gap is very much larger if one vision of livestock feed and fodder at subsidized rates, as uses a full-value estimate for the value of crop losses of PKR 50,000 per well as de-worming medicines and vaccines for animals acre (US$500 per acre) presented in Section 2.4.3, which suggests that the true value of the flood losses in the 3.26 million acres could be nearer PKR in disaster areas. 163 billion (US$1.63 billion). In this case, the actual compensation payments of PKR 6.7 billion amount only to 4.1 percent of the estimated full value of losses in agriculture. Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 35 CHAPTER 4 AGRICULTURAL CROP AND LIVESTOCK INSURANCE OPPORTUNITIES FOR PUNJAB This chapter provides an overview of the different types of crop and livestock insurance products (coverage types) that are commercially available, along with an assessment of their potential suitability for introduction in Punjab over the next five years under the SMART Punjab PforR. The proposals set out here focus largely on short- and medium-term opportunities for providing crop insurance, with a brief review of livestock insurance opportunities for the government to consider. 4.1. CROP INSURANCE TYPES, OPPORTUNITIES, AND DATA NEEDS TWO DISTINCT TYPES OF CROP INSURANCE 4.1.1.  AND THE NEED FOR TAILORED SOLUTIONS Two distinct types of crop insurance product might be considered for farmers in the Punjab. The first type—traditional indemnity-based crop insurance—protects the individual farmer against actual physical damage or loss of yield to the crops he/she grows in his/her fields. This product involves in-field assess- ment of the losses. Three indemnity-based products are listed in Table 4.1—named peril crop insurance (NPCI), multiple peril crop insurance (MPCI), and crop revenue insurance—along with details on their availability and potential suitability for Punjab. The second type—new index-based insurance—makes use of a proxy index that correlates closely with crop yield, such as the amount of rainfall as measured at a local weather station, and payouts are triggered when the actual amount of precipitation recorded during the crop growing season falls short of a previously agreed thresh- old. As such, an index coverage does not make payouts according to the actual crop losses experienced by individual farmers on their own fields. Crop index insurance products include WII, which makes use of ground-located weather stations and is typ- ically designed to insure against a shortage or excess of rainfall; satellite-based weather indexes; AYII; and specialist indexes such as flood index insurance (Table 4.1). Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 37 TABLE 4.1:  TYPES OF CROP INSURANCE PRODUCT AVAILABLE AND POTENTIAL SUITABILITY FOR PUNJAB Type of agricultural Basis of insurance insurance product and indemnity Availability Suitability for Punjab a) Indemnity-based crop insurance 1. Named Peril Crop Insurance Percent damage Widespread Possible (e.g., hail, frost, wind) (NPCI) 2. Multiple Peril Crop Insurance Yield loss Widespread Only for large growers >40 ha cereals (MPCI) 3. Crop revenue insurance Yield loss and price loss Very restricted (USA) Not available b) Index-based crop insurance 4. Crop Weather Index Weather index payout scale Widespread Limited weather station density Insurance (WII) based on (39 UNMA for all Uganda). Not best ground weather stations suited to micro-level insurance for small cereal farmers <2 ha. Possible applications for horticulture and fruit crops. 5. Crop Weather Index Weather index payout scale Fairly widespread Satellite data freely available. Not best Insurance (WII) based on suited to micro-level insurance for synthetic satellite rainfall small farmers <2 ha. 6. Crop Area Yield Index Area yield loss Fairly widespread Potential for small farmers (cereals, Insurance (AYII) cotton, sugarcane) using Department of Agriculture/Crop Reporting Services CCE yield data 7. Specialist indexes, e.g., flood Flood Index Payout Scale Very restricted Major research required to launch index, Bangladesh coverage Source: Authors. An important consideration is that one size does progressive farmers, who typically own between not fit all: in other words, crop insurance products 2.5 and 25 acres and who account for 55 percent must be tailored to the risk transfer needs of differ- of all farmers in Punjab. These farmers produce ent types of farmers. Figure 4.1 presents the three main crops both for family consumption and for sale; they categories of farmers in Punjab, classified by size of farm increasingly use seasonal crop loans to invest in improved or cultivated landholding, alongside the suitability of tradi- high-yielding seed and fertilizer technology; and they face tional indemnity-based and or index insurance products. a financial exposure in the event of crop loss.39 For these farmers, an index-based insurance product such as WII For commercial farmers in Punjab, who account or AYII, which do not require costly pre-inspections or for just 3 percent of all farm households and individual field-by-field loss assessment, may offer solu- who have more than 25 acres, individual grower tions to their risk transfer needs. MPCI or NPCI may be suitable products for the largest of these farmers—those with more than Since the early 2000s, WII has been heavily pro- 100 acres (40 hectares) of insured crops. Insurers moted by development agencies, NGOs, and aca- can offer MPCI coverage to large farmers because the demics as a low-cost insurance product that is premium generated by each risk is adequate to cover suitable as a retail product for resource-poor the costs of pre-acceptance risk inspections, mid-season small-scale farmers. The authors of this report monitoring inspections, and end-of-season crop yield argue however that individual farmer-based insurance assessments. (Figure 4.1). 39Thiscategory of progressive farmer also includes some very small farmers Individual grower MPCI is not, however, a with <2.5 acres, who receive seasonal crop credit from MFIs under the suitable product for small semicommercial/ GoPunjab e-Kissan credit scheme. 38 A Feasibility Study FIGURE 4.1:  SUITABILITY OF CROP AND LIVESTOCK INSURANCE PRODUCTS FOR DIFFERENT TYPES OF FARMER • Large farm units All (> 25 Ac, 3% HHs) perils Commercial • Access to credit (MPCI) Named farmers • High levels input use peril • Produce for sale Index insurance • Medium and smallholder Semicommercial (2.5–25 Ac, 55% HHs) “progressive” • Some assets farmers • Some access to credit • Part consumption/part sale Social safety-net programs: • Very few assets Small (< 2.5 Ac, 42% HHs) Government purchases subsistence • Subsistence farming insurance on behalf of pre- farmers identified producers • Very vulnerable to climatic shocks Source: Authors. Note: HH = households; MPCI = multiple peril crop insurance. is not suitable to the risk management needs of subsis- the field, less a deductible expressed as a percentage, is tence farmers, whose priority is to smooth consumption applied to the pre-agreed sum insured. The sum insured and who do not face a financial exposure in the form of may be based on production costs or on the expected crop loans for growing their crops. In Punjab, subsistence value of crop output (yield 3 sales price). Where damage farmers, defined as those with less than 2.5 acres of cul- cannot be measured accurately immediately after the loss tivated land, account for 42 percent of all households.40 occurrence, the assessment may be deferred until later in Rather than selling often costly individual crop insurance the crop season. Damage-based indemnity insurance is to these farmers, governments can develop social safety best known for hail, but is also used for other named peril net programs in the form of conventional ex-post natu- insurance products (such as excess rainfall or wind). ral and climatic disaster compensation programs, which could be insured/protected through the purchase of In Punjab, there may be scope to develop NPCI to ex-ante index insurance at the macro-level—that is, by protect against specific perils such as frost, hail, the provincial or local government (Figure 4.1). and excess rain in high-value horticultural crops such as potatoes, or to protect against frost, hail, and wind damage in tree crops, including mango OPPORTUNITIES FOR INDEMNITY- 4.1.2.  and citrus. Potentially, hail and wind damage insurance BASED CROP INSURANCE could also be offered for cereals if there is a significantly PRODUCTS FOR PUNJAB high exposure in these crops, especially at the time of grain maturity and harvest. If NPCI is to be successfully Named Peril Damage-based Policies developed in Punjab for fruit and horticultural crops, key Named peril crop insurance (NPCI) is an indi- considerations will include: vidual farmer damage-based indemnity crop a) The availability of time-series meteorological insurance policy; under this kind of policy, the data to establish the frequency and severity of insurance claim is calculated by measuring the occurrence of each peril. While rainfall and tem- percentage damage in the field, soon after the perature data are available from the network of damage occurs. The percentage damage measured in meteorological weather stations in Punjab, hail occurrence and wind data are lacking, and both 40Note that for crop insurance purposes, farmers with <2.5 acres who of these perils are very localized. obtain seasonal crop credit from MFIs and invest in improved or hybrid seed b) The availability of time-series historical crop loss and fertilizer technology would not be considered under this category of subsistence farmers. and damage data for each peril and each crop Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 39 at a localized level (such as the district or tehsil). extremely poor. Common problems include low According to the Punjab DoA, such data may uptake,41 high levels of anti-selection and moral hazard, be recorded at a local level by field extension and high administration and operating costs. Under- officers, but it is not systematically analyzed or writing results are usually negative. Most individual reported. grower MPCI programs that are voluntary suffer from c) Given the small average size of horticultural very high levels of anti-selection and moral hazard. farms and fruit farms in Punjab, it would be For example, farmers may purchase coverage only for necessary to assess ways of designing low-cost low-lying areas that are subject to flooding and fields in ­ procedures for assessing damage in the field. waterlogging (anti-selection), or they may apply subop- Otherwise such a coverage cannot be commer- timal levels of husbandry and pest, disease, and weed cially viable. control (moral hazard) in the expectation of then claim- ing the yield loss on their crop insurance policy. MPCI programs are usually very exposed to systemic drought, Recommendation for Punjab frost, and windstorm losses that correlate at the regional It is recommended that a feasibility study for and national levels, and the premium rates that have to the development of NPCI for horticultural be charged to cover the combination of high losses and and tree crops be implemented in the second high administrative costs are often in excess of 10 percent phase of the Punjab agricultural insurance to 15 percent of the sum insured. Nearly all individual program, starting in 2019/20. grower MPCI programs operate at a financial loss (neg- ative underwriting results) and depend on government premium subsidies to make the coverage more affordable and acceptable to farmers and/or depend on govern- Yield-based Crop Insurance ment subsidies of excess claims.42 (multi-peril crop insurance) A loss-of-yield policy is an individual farmer pol- Furthermore, as noted in section 4.1.1, individ- icy that protects the farmer against crop produc- ual grower MPCI is a product that is generally tion and yield losses on his or her own farm. An only offered by insurance companies to large insured yield (expressed, for instance, in tons per hectare) commercial farmers because the high costs of is established, as a percentage of the historical average administering such a coverage make it prohibi- yield of the insured farmer: the insured yield is typically tively expensive to operate with small farmers. set at between 50 percent and 75 percent of the average MPCI is the product that is most widely marketed to large yield on the farm. For irrigated agriculture, however, in cereal and oilseed producers in the USA and Canada. In which yields tend to be much more stable, insurers may Brazil, MPCI is offered to medium and large cereal and be willing to insure up to 90 percent of the farmer’s aver- oilseed farmers (usually linked to credit provision), and in age yield. The actual yield is measured by an independent China several regional MPCI programs are implemented assessor at the time of harvest, and if the actual yield is for smallholder farmers, but only as mandatory schemes less than the insured yield, an indemnity is paid equal to in which all farmers are insured. All of these MPCI pro- the difference between the actual yield and insured yield, grams carry very high levels of premium subsidy. multiplied by a pre-agreed value of sum insured per unit of yield. The most common form of a loss-of-yield policy is an MPCI policy that provides comprehensive protection against all unavoidable natural, cli- matic, and biological perils that may cause yield 41A major exception is the U.S. Federal Crop Insurance Program (FCIP), loss. This coverage is widely demanded by farmers both for which uptake rates for major grain crops such as wheat and maize are in developed and in developing countries because it pro- extremely high and exceed 85 percent to 90 percent of all eligible sown vides all-risk protection against loss of crop production acreage. High uptake rates also apply to soy and sunflower. One major and yields. reason for the high uptake rates of the FCIP for these crops is the very high premium subsidy rates offered by the government to insuring farmers. For minor crops, however, uptake rates are much lower in spite of the high Even so, the international experience with indi- premium subsidies. 42For a comprehensive review of the performance of public-sector MPCI, vidual farmer MPCI is, with few exceptions, refer to Hazell et al. (1986) and Mahul and Stutley (2010). 40 A Feasibility Study Recommendation for Punjab rainfall deficit or drought, where rainfall measurements Insurance companies offer individual grower are made at a reference weather station(s), during defined MPCI only to medium and large commercial period(s), and insurance payouts are made based on a farmers because it is prohibitively expensive preestablished indemnity scale set out in the insurance to administer such a coverage on small farm policy. The sum insured is normally based on the produc- units. Because commercial farmers account for only tion costs for the selected crop, and indemnity payments 3 percent of farmers in Punjab, and because DoA’s are made when actual rainfall in the current cropping stated priorities are to address the needs of the other season as measured at the selected weather station falls 97 percent—the small semicommercial/progressive below predefined threshold levels. and subsistence farmers—no plans exist to develop MPCI under the SMART Punjab program. This The main advantage of WII is the elimination does not, however, prevent private insurance compa- of the adverse selection and moral hazard prob- nies from developing and promoting MPCI in Pun- lems common to MPCI. Since payouts are made jab (and the rest of Pakistan). based on an objective measurement at the reference weather station, there are few information asymmetries to be exploited, and the behavior of the insured cannot Crop Revenue Insurance influence the extent of payouts. In addition, WII reduces administration costs (particularly because it does not A crop revenue policy combines conventional require in-field inspections or loss adjustment) for the multiple peril insurance based on loss of crop insurer, which could make premiums more affordable. yield (MPCI) with protection against loss of mar- Indexed products are also likely to facilitate risk trans- ket price at the time of sale of the crop. Currently, fer to the international reinsurance markets. However, this product is marketed only on a commercial basis in although WII offers opportunities for reduced admin- the USA for grains and oilseeds quoted on commodity istration and operating costs, the development phase markets (the Chicago Board of Trade) and where future requires intensive technical inputs and ongoing technical price contracts can be combined into the revenue policy. inputs are required to refine products over time. Currently, crop revenue insurance could not be offered to small semicommercial/progressive and subsistence farm- ers in Punjab. The most important challenge for WII is basis risk, which significantly limits the applicability of index instruments. Basis risk is the difference between the pay- 4.1.3. OPPORTUNITIES FOR out as measured by the index, and the actual loss incurred INDEX-BASED CROP INSURANCE by the insured farmer(s). Because no field loss assessment is made under index insurance, the payout may either be PRODUCTS FOR PUNJAB higher, or lower, than the actual loss of crop suffered by Weather Index Insurance (using ground the farmer(s). Basis risk is lower when the risk is highly weather stations) correlated—that is, the risk affects a large geographical area to relatively the same extent and simultaneously. The Crop WII is an alternative approach that aims to level of basis risk can be somewhat mitigated by careful overcome many of the drawbacks of traditional index design and by the installation of new weather sta- individual grower indemnity-based crop insur- tions, thereby increasing the density of weather stations ance. The key feature of WII products is that they do and data points (typically one weather station with a cov- not indemnify crop yield losses at the individual field or erage area of 15–20 km) and providing more localized grower level, but rather use a proxy variable (the index) precision in the measured climatic peril. Other challenges such as the amount of rainfall or temperature or wind for WII include the need for high-quality weather data speed to trigger insurance payouts to farmers. and infrastructure and the currently limited product options, with most applications in developing countries so WII is a simplified form of insurance in which far concentrating on rainfall indexes. payments are made based on a weather index, rather than measurement of crop loss in the WII is being developed at different levels of field. The index is selected to represent, as closely as aggregation, starting with individual farmers possible, the crop yield loss likely to be experienced by the (micro-indexes) and then at a regional level—examples farmer. The most common application of WII is against Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 41 including input suppliers or banks providing lending credit small-scale farmers and aims to overcome many in a specified area (meso-indexes)—and then finally at a drawbacks of traditional individual farmer national level as a food security instrument (macro-level MPCI. The key feature of this product is that it does not weather indexes). The first country to introduce micro- indemnify crop yield losses at the individual farmer or level WII was India in 2003, and many programs have field level. Instead. it makes indemnity payments to farm- been launched since. Mexico was the first country to ers according to yield loss or shortfall against an average develop macro-level WII index coverages that offer state area yield (the index) in a defined geographical area (such governments catastrophe drought, rainfall, and windstorm as a district, tehsil, union council, or village), commonly index protection for small vulnerable subsistence farmers. referred to as the Unit Area of Insurance (UAI). In Punjab, the density of weather stations is cur- The key advantages of the area-yield approach rently too low to support a province-wide WII are that it minimizes moral hazard and anti-se- program for major crops grown by the majority lection and significantly reduces the costs of of the province’s 5.2 million farmers. The Punjab administering the policy, making an area-yield Meteorological Agency has a network of only 30 official product much more suitable to offer to small- synoptic weather stations or less than one weather sta- scale farmers. Under an AYII policy, yield losses are tion per district. This density is inadequate to develop settled against the area average yield index, as opposed a large-scale WII programs for millions of farmers. An to settling losses on individual farmers’ fields under an additional limitation to introducing WII in Punjab is that MPCI policy. As explained previously, individual farmers 85–90 percent of cultivated area is irrigated, whereas cannot influence the yield outcome by purchasing cov- WII is best suited to rain-fed agriculture, with indexes erage only for fields subject to harm such as drought or typically designed to protect against rainfall deficits (lead- flooding (anti-selection) or by neglecting husbandry and ing to drought) or excesses (leading to flooding). Alterna- plant health practices, only to claim the resulting yield tives exist to develop satellite indexes, including synthetic loss on their crop insurance policy. The costs of operating rainfall, evapotranspiration, or Normalized Difference AYII are much lower than for MPCI, especially because Vegetation Index (NDVI) in certain crops. AYII requires no prior field inspections or in-field crop loss assessments on individual farms, which means that an AYII product with lower premium costs can poten- Recommendation for Punjab tially be marketed to small- and medium-sized farmers. We recommend that a feasibility study for the Table 4.2 lists still other advantages of AYII. development of WII for horticultural and tree crops be implemented in the second phase of As with WII, the main disadvantage of an AYII the Punjab agricultural insurance program policy is basis risk—which in this case is the differ- starting in 2019–20, but only in areas served ence in actual yield outcomes achieved by individual by a local weather station. farmers on their own fields and the average area yield. For example, an individual farmer may incur severe crop We also recommend that the development production and yield losses due to localized perils such of WII based on satellite indexes such as the as hail or flooding by a nearby river, but because these Standardized Precipitation Index and/­ or localized losses do not affect the average yield across a Standardized Precipitation Evapotranspi- larger area (the tehsil, village, or other area defined as the ration Index as a “macro-level” insurance UAI), the grower receives no indemnity. Other problems protection be examined in a feasibility study. include the need for an accurate procedure to measure The insured in this case would be the GoPunjab. The the average area yields in the UAI (Table 4.2). insurance could be used to support the social protec- tion coverage by GoPunjab to poor subsistence farm- To operate an AYII coverage, it is necessary to ers with less than 2.5 acres. have (1) accurate historical yield data at local area levels to provide a sound basis to construct a yield index, and (2) an objective and accurate Area-Yield Index Insurance method of establishing the actual average yield in the insured growing season to determine if a pay- Area-yield index insurance (AYII) is a loss-of- out is due or not. In Punjab, the Crop Reporting Ser- crop-yield policy that is suited to the needs of vices (CRS) of the DoA has for many years been involved 42 A Feasibility Study TABLE 4.2:  PRECONDITIONS, ADVANTAGES, AND DISADVANTAGES OF AREA-YIELD INDEX INSURANCE Preconditions Advantages Disadvantages Homogeneous cropping systems in the No need for time-series data on yields of Basis risk (but lower than for weather defined geographical area (e.g., region, individual growers index insurance) district) that forms the Unit Area of Insurance (UAI) Accuracy of historical regional yield data Data are likely to be available because Not suitable for localized perils (e.g., hail) most countries record regional yield statistics Timely, accurate, and impartial Lower cost of delivering insurance Problems of accurately measuring procedures for estimating “actual” product to growers “actual” average yields in UAI average yield in the UAI Special insurance regulation may be Suited to systemic risk (e.g., drought) Farmers’ acceptance required Minimizes adverse selection and moral hazard Requires no in-field loss assessment Cost of loss assessment reduced Because the product is based on yields, it picks up all weather risks and other causes of yield shortfalls Source: Authors. in implementing seasonal crop yield surveys based on semicommercial/progressive farmers (typi- random selection of farmers and fields, which are then cally operating 2.5–25 acres), who grow wheat, subjected to randomly placed crop cutting experiments rice, maize, cotton, and sugarcane in Pun- (CCEs) to estimate crop yields. The Punjab government jab. This coverage could also be extended to semi­ uses the data from the CCEs to estimate average crop commercial/progressive farmers with <2.5 acres who yields and crop production (yield times planted area) at receive credit for seasonal crop production from MFIs the district level for all major cereal crops, including Rabi and under the e-Kissan program in Punjab. For AYII wheat and Kharif paddy and maize, and also for cash to be successfully implemented and scaled up, it would crops such as cotton and sugarcane. Currently, the CRS be necessary to agree on a program with the CRS to adopts a sample frame of 5 percent or about 1,250  vil- increase the density of the CCEs over time. The objec- lages in Punjab to conduct farmer surveys and CCEs. tive would be to conduct CCEs in all villages where the crop AYII program is implemented. The current level Given that CRS in Punjab has estimated district of reporting area yields (the district level) is too large, area yields for many years based on objective and the UAI for the operation of AYII must be reduced CCEs, the CRS data could form the basis for over time to the tehsil level and then eventually down to operating a large-scale AYII program in Punjab. the markaz or even union council level.43 Because the current density of village sampling is too low, however, it would need to be increased to support an Furthermore, GoPunjab could offer AYII as AYII program. a social protection coverage to the very poor 43In India under the PMFBY AYII program, the government in 2015–16 Recommendation for Punjab reduced the size of the insured unit from a subdistrict, block, or teluka to the The CRS CCE yield estimation methodology gram panchayet or individual village level. This change necessitated a huge increase in the number of CCEs that are conducted. In Punjab, the CRS has appears to offer major potential to develop advised authorities that it cannot implement AYII at the individual village and implement a large-scale AYII program for level given the high fiscal costs of conducting CCEs at this level. Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 43 subsistence farmers with less than 2.5 acres. also for industrial crops such as cotton and sugar- AYII could also be used as an ex-ante crop insurance cane (crops for which the CRS conducts CCEs). coverage to trigger objective payouts to the large num- »» AYII coverage could be particularly ber of subsistence farmers (who do not borrow seasonal appropriate for small semicommercial/ crop credit from financial institutions)44 in Punjab and progressive farmers with less than five to operate as a disaster risk financing and insurance acres as an automatic crop-credit insur- coverage to the GoPunjab’s existing natural disaster ance coverage bundled with loans through compensation program, which is operated through the noncommercial banks such as MFIs and Punjab Disaster Risk Management Agency. through GoPunjab’s e-Kissan program. In these cases, the programs are not insured under CLIS and a full-value “ground-up” policy would CONCLUSIONS ON CROP 4.1.4.  be issued to the loanee farmers, covering an area INSURANCE OPPORTUNITIES yield shortfall from say 80 percent down to 0 per- cent of the expected area yield. Any non-loanee FOR PUNJAB farmer could also purchase ground-up coverage Based on the analysis in this feasibility study, on a purely voluntary basis. the following conclusions and recommendations »» For farmers borrowing from commer- are drawn with respect to crop insurance oppor- cial banks who are already automatically tunities for the GoPunjab to consider under the insured under CLIS for catastrophe losses SMART Punjab program: exceeding 50 percent of expected produc- »» Indemnity-based MPCI can be cost effectively tion and yield, it may be possible to offer implemented only with farmers who cultivate AYII “top-up” coverage. These farmers would more than 100 acres (40 hectares) under a single then be covered for area-yield shortfalls from say crop. This coverage would therefore be restricted 80 percent down to 50 percent of expected area to a relatively small number of large-scale com- yield (see the next section for further discussion of mercial farmers in Punjab. this possibility). »» Crop revenue insurance coverage providing »» There may also be scope in Punjab to protection for both crop production and yield loss develop NPCI coverage (both parametric and and also market price loss is available only in the nonparametric) against specific perils such as frost, United States for a few globally traded commod- wind, and hail for vegetable crops (such as pota- ities (wheat, maize, soybeans), and in the start-up toes) and tree fruit (citrus, mangoes, and others). phase in Punjab it would not be available. »» WII coverage will also have limited applica- tion in the short term because of the low density of weather stations (only 30 weather stations in BUILDING ON THE CROP 4.2.  the province) and because 85–90 percent of cul- tivated area is irrigated. In addition, WII is nor- LOAN INSURANCE mally bested suited to rain-fed cropping and insures against rainfall deficit or excess rainfall. WII has SCHEME not yet successfully developed solutions for flood- LINKAGES BETWEEN THE CROP 4.2.1.  prone areas, the major concern for many farmers LOAN INSURANCE SCHEME in Punjab. The feasibility of WII based on satellite indexes at the macro-level may be examined for AND A NEW CROP AREA-YIELD protecting/­insuring the social protection scheme INDEX INSURANCE PROGRAM to support subsistence farmers with holdings of less FOR SEMICOMMERCIAL/ than 2.5 acres. PROGRESSIVE FARMERS »» AYII appears to offer considerable potential for By building on the current CLIS, GoPunjab could “semicommercial/ insuring “subsistence” as well as ­ consider developing commercial crop insurance progressive” small-scale farmers in Punjab for major for semicommercial/progressive farmers who cereal crops, including wheat, rice, and maize, and access seasonal production credit through com- mercial banks (Figure 4.2). Chapter 3 highlighted the 44Thissocial protection coverage would not, however, be provided to small fact that the CLIS product is primarily a catastrophic progressive farmers with <2.5 acres who access seasonal crop loans. insurance coverage, paying out when crop yields drop 44 A Feasibility Study FIGURE 4.2:  OPTION FOR LINKING CLIS AND GOPUNJAB AYII TOP-UP COVERAGE FOR SEMICOMMERCIAL/PROGRESSIVE FARMERS Basmati Paddy 1 Acre Model: Normal average yield 1,500 kg/acre Yield Loanee farmer loss Coverage level (%) Insured yield (kg/acre) 100 1,500 Farmer retention 90 1,350 80% 80 1,200 Top-up area-yield 70 1,050 Punjab Crop Insurance index insurance Program 60 900 (top-up AYII cover) 50 750 50% Catastrophe loan insurance scheme CLIS (federally funded) 0 (premium up to 2% of Sl, i.e., loan amount) 0% Source: Authors. below 50 percent of the historical yield for the area. business potential to the participating insurers for scaling For that reason, farmers incurring more frequent yield up and for spreading risk. losses have no insurance protection. This gap in coverage exposes both the farmer and credit provider (that is, the If a direct linkage is to be promoted between the financial institutions) to significant credit risk. CLIS and a new AYII program for commercial crops, SBP would need to agree to adopt the sea- GoPunjab could support an AYII program for sonal CCEs to establish the actual average yield crops providing “top-up” coverage in excess of in each defined UAI and to settle payouts on this the CLIS 50 percent yield coverage level, up to basis. This would be in place of the current procedure, approximately 80 percent of the area average which requires the local provincial authorities to declare yield. This coverage could be marketed either on a vol- a natural calamity when they estimate crop damage and untary or mandatory basis to farmers accessing seasonal losses to exceed 50 percent of expected crop production. crop loans through the existing bank channels and who In addition, the CRS would need to significantly increase are protected by CLIS. In some cases, insurers may be the density of its CCEs throughout the districts and willing to offer higher coverage to farmers in areas with tehsils of Punjab where the CLIS is currently being sold. assured irrigation and where yield variation is low for area-yield losses, up to a maximum of 90 percent of the area average yield. OPTIONS FOR OFFERING CROP 4.2.2.  AYII TO SEMICOMMERCIAL/ Currently the CLIS is insuring approximately PROGRESSIVE FARMERS WHO 1 million loanee farmers, of whom about 70 per- ARE NOT INSURED UNDER CLIS cent (700,000) are located in Punjab (see Chapter 3 As mentioned, the AYII coverage could also be for details). If the GoPunjab crop AYII program for offered to semicommercial/progressive farmers semicommercial/progressive farmers was linked to the who are not insured under CLIS as a full-value loss- CLIS on a mandatory basis, it would offer considerable of-yield coverage from approximately 80  percent Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 45 FIGURE 4.3:  FULL-VALUE OR “GROUND-UP” AYII COVERAGE FOR SEMICOMMERCIAL/ PROGRESSIVE FARMERS NOT INSURED UNDER CLIS AYII coverage for small farmers borrowing credit from MFts Basmati Paddy 1 Acre Model: Normal average yield 1,500 kg/acre Yield Non-CLIS farmers loss Coverage level (%) Insured yield (kg/acre) 100 1,500 Farmer retention 90 1,350 80% 80 1,200 70 1,050 60 900 50 750 50% Punjab Crop Insurance Program 40 Area yield 600 (full value or ground-up index insurance AAII coverage) 30 (proportional payouts) 450 20 300 10 150 0% 0 0 Source: Authors. down to 0 percent yield (ground-up coverage), as other countries like Brazil, Ghana, Kenya, Mexico, illustrated in Figure 4.3. Two types of semicommercial/ Morocco, Peru, the Philippines, and Sudan have devel- progressive farmer could be offered full-value AYII cov- oped AYII programs specifically for small farmers. erage. The first group consists of small farmers (<5 acres) who are beneficiaries of loans distributed through MFIs India is the world’s largest crop insurance mar- under the GoPunjab e-Kissan credit program, but who are ket, insuring about 32 million farmers in 2012– not insured by CLIS. In this instance, GoPunjab would 13; more than 65 percent of those farmers are make crop insurance compulsory for the Kissan e-credit insured under AYII programs, which have oper- recipients. The second group would be non-loanee farm- ated there for 37 years. The Agricultural Insurance ers wishing to purchase crop AYII insurance on a purely Company of India (AICI) was responsible for imple- voluntary basis. In that case, the challenge would be to menting AYII under the National Agricultural Insurance identify suitable marketing, promotion, and delivery chan- Scheme (NAIS) between 2000 and 2015. Key features of nels for the voluntary coverage for non-loanee farmers. the NAIS included: »» The program targeted small and marginal farm- ers (<2 hectares), who are highly dependent on 4.3. INTERNATIONAL access to seasonal crop credit. »» Crop insurance was compulsory for borrowing EXPERIENCE WITH farmers and voluntary for non-borrowing farmers. »» The program covered a wide range of food, oil- CROP AREA YIELD INDEX seed, pulse, and cash crops. INSURANCE »» The program was heavily subsidized in two ways. First, the maximum insurance charges payable by Crop AYII is implemented in a wide range of farmers were capped at 2.0 percent (Kharif) and developed and developing countries. India intro- 1.5 percent (Rabi) of the sum insured or actuarial duced AYII in the late 1970s and the USA and Canada rate, whichever was less, for food and oilseeds crops introduced this product in the early 1990s. More recently, (all cereals, millets, oilseeds, pulses); and 5 percent 46 A Feasibility Study for annual commercial/horticultural crops. Sec- The key market-based principles that differentiate the ond, in the case of catastrophic losses computed mNAIS relative to the NAIS are that crop premium rates at the national level for an agricultural crop sea- are actuarially determined rather than being capped, the son, the liability of insurance companies was up central and state governments pay direct premium subsi- to 350 percent of the total premium collected (the dies to make the coverage affordable to farmers, and the farmer share plus the government subsidy), or program was reinsured with General Insurance Corpo- 35 percent of the TSI of all the insurance com- ration (GIC) of India, which is the national reinsurer and panies combined, whichever was higher. Losses at international reinsurer, rather than by the federal and the national level in a crop season beyond this ceil- national government. The technical refinements were a ing were met by equal contributions (that is, on a reduction in the size of the UAI to the village level to 50:50 basis) from the central government and the reduce basis risk, and steps to strengthen the crop-cutting concerned state governments. procedure to make it more transparent, accurate, and »» The UAI was normally the taluka or block, but rapid to complete. Other changes included raising the in some cases was as low as the gram panchayat minimum insured yield from 60 percent to 70 percent (village council) or individual village. while maintaining the 80 percent and 90 percent maxi- »» Levels of insured yield coverage were set for mum yield coverage levels. Also, the average or expected each district and UAI according to the degree yield was calculated as the average of the last seven years, of variability in historical area crop yields and excluding two years of declared calamities. actuarial experience. Three levels of threshold or insured yield were offered under NAIS—60 per- Under mNAIS, the move to an actuarial regime cent, 80 percent, and 90 percent of the historical led to major increases in premium rates, as evi- annual area average yield (or “expected yield”)— denced by the average commercial premium according to the degree of yield variability of rate from 2011 to 2014 of 11.1 percent. Table 4.3 each insured crop in each UAI. Under NAIS the summarizes the main features of mNAIS by state for expected yield was calculated as the average of the the four-year period, including the number of insured middle three out of the past five years, excluding farmers in each season, and in total, the average pre- the years with the highest and lowest yields. mium rates charged for Kharif and Rabi crops, and the »» The program was implemented and administered four-year long-term average loss ratio for each season. through the bank branch network, including Agri- The scale of the mNAIS program varied considerably culture Credit Cooperatives in each state, district, between states, with the largest number of insured farm- and block (group of villages). Insurers paid the ers in Andhra Pradesh, Bihar, and Rajasthan. Average banks a commission for managing the scheme on premium rates charged by commercial insurers also var- their behalf, including crop insurance premium ied widely between states. The lowest average rates in financing, which is included as part of the loan of the Kharif season were in Uttar Pradesh (2.70 percent, the borrower. with a loss ratio of 15.4 percent); Uttrakhand (3.28 per- »» Actual area yields were measured by the State cent, loss ratio 99.3 percent); Orissa (3.59 percent, but Department of Agriculture extension officers for the incredibly high loss ratio of 996.3 percent, indicates each crop through sample CCEs at the time of that premium rates need to be much higher) and Hary- harvest. This major and costly exercise suffers ana (5.05 percent, loss ratio 75.3 percent). At the other from delays in processing the results, and for that extreme, in large states such as Bihar, the Kharif aver- reason, indemnity payments under NAIS were age premium rates were extremely high at 21.7 percent often delayed for six months or more. Recently, (loss ratio 41 percent), probably due to the very high governments have been involving the private sec- flood risk exposure in the Kharif season, and in Rajas- tor in conducting CCEs, including steps such as than average rates were 15.10 percent (62.7 percent outsourcing CCEs and using technology (such loss ratio). In the Rabi season, the lowest average pre- as mobile smartphones) in the CCE process to mium rates were seen in Haryana (1.20 percent, with reduce downtime significantly. a loss ratio of 87.1 percent) and Uttrakhand (2.80 per- cent, 10.1 percent loss ratio). States with high aver- For Rabi 2010, India launched the modified NAIS age premium rates in the Rabi season included West (mNAIS), which adopted market-based princi- Bengal (15.26 percent, but with a very low loss ratio of ples technical refinements. Between 2010 and 2014, 19.8 percent) and Karnataka (12.9 percent, loss ratio nearly 10 million farmers were insured under mNAIS. 45 percent). Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 47 TABLE 4.3:  INDIA: INSURED FARMERS, AVERAGE PREMIUM RATES, AND AVERAGE LOSS RATIOS BY STATE AND SEASON FOR THE AREA-YIELD INDEX INSURANCE PROGRAM OF THE MODIFIED NATIONAL AGRICULTURAL INSURANCE SCHEME, 2001–14 Number insured farmers Average premium rates Average loss ratios State Kharif Rabi Overall Kharif Rabi Overall Kharif Rabi Overall Haryana 73,813 27,841 101,654 5.05% 1.20% 4.01% 75.3% 87.1% 76.3% Rajasthan 560,682 1,824,946 2,385,628 15.10% 8.02% 8.85% 62.7% 46.7% 49.9% Uttrakhand 36,469 9,421 45,890 3.28% 2.80% 3.18% 99.3% 10.1% 82.3% Gujarat 231 201 432 19.01% 5.31% 10.81% 0.0% 0.0% 0.0% Uttar Pradesh 77,293 610,802 688,095 2.70% 5.15% 4.71% 15.4% 309.4% 279.1% Madhya Pradesh 6,709 72,228 78,937 5.81% 3.58% 3.85% 1.5% 3.1% 2.8% Bihar 758,927 77,483 836,410 21.70% 9.79% 20.39% 41.0% 24.2% 40.1% Jharkhand 2,895 8,794 11,689 12.99% 6.69% 9.05% 0.0% 5.7% 2.7% West Bengal 861,655 861,655 15.26% 15.26% 19.8% 19.8% Assam 9,310 7,676 16,986 4.49% 3.78% 4.21% 34.4% 22.5% 30.2% Orissa 69,489 50,209 119,698 3.59% 4.11% 3.81% 996.3% 79.9% 575.3% Andhra Pradesh 1,007,946 170,071 1,178,017 9.47% 4.57% 8.66% 184.2% 48.1% 172.3% Karnataka 500,578 62,064 562,642 11.76% 12.90% 11.84% 42.3% 45.0% 42.5% Tamil Nadu 301,070 130,114 431,184 13.78% 7.75% 11.74% 205.0% 14.1% 162.4% Source: AIC www.aicofindia.com/ To help small and marginal farmers afford the premium rates charged to farmers because of its concern mNAIS coverage, the Federal Government of that mNAIS is too expensive for farmers. Farmers will in India accepted to provide a significantly higher the future pay a uniform premium rate of 2.0 percent for premium subsidy, equivalent to 61 percent of Kharif crops and 1.5 percent for Rabi crops, while the total premiums. The main benefit to the govern- rate for commercial and horticultural crops will be 5 per- ment was that it no longer assumed responsibility for cent. The rest of the premium will be paid by the gov- an unknown and unbudgeted level of excess claims, as ernment with no upper limit on the subsidy amount. In 100 percent of the program liability was transferred from other words, rates will be actuarially determined, and the the government’s balance sheet to local insurers and to government will settle the difference between the flat rate the state and international reinsurers. The average loss paid by the farmer and the rate charged by the insurer. ratio for mNAIS was 73 percent,45 indicating commercial A further major change, introduced to reduce basic risk, viability for the participating insurers and reinsurers. is the reduction in the size of the UAI from the block to the gram panchayat or village level. With the move to the In 2016 the Government of India announced a rad- village level, the minimum number of CCEs has been ical plan to replace NAIS and mNAIS with a single set at four per village for cereals and major crops, which program, the Pradan Mantri Fasal Bima Yogana means that the number of CCEs will need to increase (Prime Minister’s Crop Insurance Scheme, from about 2 million per year to about 8 to 9 million per PMFBY). The main aim of PMFBY is to increase the year—a major challenge for the participating states. penetration of crop insurance in India to 50 percent of all farmers (about 65 million farmers) by 2020. The main Further information on the experience with AYII change is that government has reverted to flat or capped in India, the USA, and Brazil is contained in Annex 4. 45www.aicofindia.com 48 A Feasibility Study 4.4. AREA-YIELD INDEX The operation of an AYII policy providing full-value, ground-up coverage is illustrated for a hypothetical crop INSURANCE FOR of maize over two growing seasons in Figure 4.4. In this example, Village X has a farming population of about SEMICOMMERCIAL/ 10,000 small farmers and an average maize area of about PROGRESSIVE FARMERS 50,000 acres. The average or normal expected yield of maize in Village X is 2,500 kg/acre, which is similar to IN PUNJAB the average yield of maize in Punjab. This forms the area-yield index. AYII insurers typically offer insured The sections that follow present the central fea- yield coverage levels (termed the “coverage” level) that tures of an AYII coverage for semicommercial/ are between a minimum of 50 percent and a maximum progressive farmers in Punjab. The discussion of 90 percent of the average area yield. In the exam- begins by describing the basis of insurance and indem- ple in Figure 4.4, the insured yield (or threshold yield) nity of an AYII product providing (1) full-value or is set at 80 percent of the average, which in this case is “ground-up” yield coverage for non-CLIS farmers and 2,000 kg/acre. The insured yield represents a guarantee (2) layered or “top-up” coverage for farmers already of the yield level, such that if the actual area yield, as insured under CLIS. It moves on to review the precondi- measured at the time of harvest in Village X, falls below tions for operating an AYII program and to describe how an average of 2,000 kg/­ acre, insurers will pay all insured the insured yield coverage level and the sums insured are farmers the amount of yield shortfall or loss per acre set. The next section focuses on methods for rating an times the agreed value (the “sum insured”) times each AYII coverage. The discussion draws on actual sample farmer’s acreage of the insured crop. In this instance it crop yield data from three tehsils in Lodhran District, is assumed that the agreed sum insured is US$500 per Punjab, which were provided by CRS-DoA-GoPunjab. acre. This is the maximum amount the insurer will pay out if there is zero recorded yield (total crop failure) in Village X in the forthcoming season. The example also BASIS OF INSURANCE 4.4.1.  assumes that 2,000 farmers in Village X elect to buy the AND INDEMNITY (PAYOUTS) AYII at the 80 percent coverage level, and that the DoA ON AN AYII POLICY has for many years conducted CCEs at harvest time on a The key principle of a crop AYII coverage is that statistically selected sample of maize farms in Village X it insures and indemnifies farmers for losses to determine actual average yield at that location. against the average area yield in a defined geo- graphic location, such as the tehsil or village In the first crop season, climatic conditions are where they farm, and is therefore not an indi- better than normal for growing maize, and the vidual farmer yield policy that insures them actual average maize yield in Village X as mea- against losses on their own farms and fields. sured by the DoA is 2,800 kg/acre, which exceeds FIGURE 4.4:  EXAMPLE OF AN AYII CONTRACT PROVIDING GROUND-UP COVERAGE FOR MAIZE GROWN IN VILLAGE X 3,000 0 Average yield 2,500 kg/acre) 2,500 Yield (kilograms per acre) 80% insured yield, 2,000 kg/acre 2,000 1,500 1,000 Yield shortfall 2,800 Actual yield 1,000 500 1,000 0 Crop season with no payout Crop season with payout Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 49 the insured yield of 2,000 kg/acre, so no pay- The CRS crop sampling frame will need to be out is due to the 2,000 insured maize farmers. modified, however. The CRS sampling frame is set up In the second season, however, severe drought causes to provide accurate estimates of cropped area, produc- the actual average yield in Village X to fall to only tion, and average yields for all major crops at the district 1,000  kg/acre—a yield shortfall of 1,000 kg/acre level, so official statistics on these variables are reported or 50  percent of the insured yield. All 2,000 farmers only at the district level. The district is, however, too big receive a payout based on this area-yield shortfall of a geographic area for the operation of an AYII coverage, 1,000 kg/acre, irrespective of the actual yields on their and ideally the sampling frame should be adapted over own farms. The per acre payout each farmer receives time to the area of a tehsil or union council. is calculated as the percentage yield shortfall (1,000 kg divided by 2,000 kg = 50 percent) times the sum insured For the feasibility study team to analyze the (US$500), for a payout of US$250/acre. Each farmer design of an AYII coverage for Punjab, CRS pro- is compensated according to the area of maize that the vided area-yield data at the tehsil level. The data farmer has grown and insured. For example, Farmer 1 covered five main crops (Rabi wheat, Kharif rice, maize, has 5 acres of insured maize and therefore receives a pay- sugarcane, and cotton) in three tehsils in Lodhran Dis- ment of 5 acres 3 US$250/acre = US$1,250. Farmer 2 trict (Kehror Placa, Dunyapur, and Lodhran). Lodhran is has 2 acres insured and receives 2 acres 3 US$250/acre = an important wheat and cotton growing area of Punjab, US$500/acre. with a high percentage of irrigated area. Yields and vari- ation in crop yields in Lodhran over time are very similar to those in other districts of Punjab that have assured irri- 4.4.2. PRECONDITIONS gation (see Annex 1 for 10-year crop yields at the district FOR THE DESIGN level). Lodhran is therefore considered to be a “represen- AND OPERATION OF AYII tative” district for the purposes of this preliminary AYII The preconditions for operating AYII for Kharif coverage design and rating analysis. and Rabi crops grown in Punjab include: »» Definable, homogeneous crop producing zones (UAI) with low yield variation between farmers in 4.4.3.  AYII COVERAGE DESIGN the insured unit (IU). Initial Crops Selected for Coverage »» For the defined UAI, historical crop sown area, AYII can, in principle, be designed for any field production, and average yield data for the past 15 crop for which area-yield data are available at years or more to provide the basis for establishing harvest, including cereals, oilseeds, pulses, fibers (such as the insured yield and technical premium rates for cotton), and industrial crops such as sugarcane. AYII is not, an AYII policy. however, suitable for short-duration horticultural and veg- »» An independent and statistically accurate sys- etable crops, which tend to be grown on a very small scale. tem of measuring actual average area-yields in the defined IU to provide the basis for triggering For the start-up phase of the Punjab agricultural claim payments when actual yields fall short of the insurance program, it is proposed to develop AYII insured yield(s). for the five major crops (wheat, rice, maize, sug- arcane, and cotton). The three reasons for focusing As noted in section 4.1, the CRS in Punjab can on these crops initially is that they are the most important meet these data preconditions for an AYII pro- crops grown throughout the province by small-scale farm- gram in Punjab. CRS has for many years been ers, CRS can provide historical yield data on these crops to involved in field survey work and CCEs to measure the construct the AYII yield index, and CRS conducts CCEs at cultivated area and actual average yields of all major harvest to form the basis for determining claims settlements. crops grown throughout the 36 districts of Punjab. The CRS samples 5 percent of Punjab’s 26,000 villages (about 1,250 villages) to conduct the farmer surveys and CCEs. In each sample village, CRS randomly selects Unit Area of Insurance and Minimum three farmers and takes two crop cuts per farmer/field Cultivated Area for the selected crop, for a total of six CCEs per crop Ideally, the UAI should be defined as a homogeneous per sample village. micro-agroclimatic zone where farmers grow the 50 A Feasibility Study same varieties of the insured crop and use simi- Expected Yields and Insured Yield lar husbandry practices (including inputs) so that Coverage Levels normal average yields are similar for all farmers. In reality, UAIs are typically defined based on administra- For the insured unit (in this case the tehsil), it tive units for which crop area, production, and yield statis- is necessary to establish the normal average or tics are collected and reported. As mentioned, the UAI in “expected yield” for the insured crop. AYII pro- India under the NAIS was formerly based on the subdistrict grams conventionally adopt one of two approaches for block (tehsil/taluka), but farmers maintained that this UAI establishing the expected yield: was too large and that rainfall, crop conditions, and yield »» The simplest approach is to take an average of the outcomes were not uniform across the block. Subsequently, past three to five to seven years’ actual area yields. under the mNAIS and now the PMFBY, the UAI is defined As noted, this approach was adopted by the NAIS at the village (gram panchayat) level. in India, which used the average of the middle three of the past five years (after eliminating years with the highest and lowest annual yields) to cal- At this stage it is not possible to determine whether culate the expected yield. The successor program the tehsil or a smaller area should be consid- (PMFBY), in recognition that the relatively short ered as the UAI. To give an idea of the considerations period of five years did not always represent the involved, note that the three sample tehsils in Lodhran Dis- average yield, now uses the average of seven years trict are quite large geographic areas where Rabi wheat by eliminating two bad years. and Kharif cotton are the predominant crops. Together »» The alternative is to detrend the time-series these tehsils produce an average of 125,000–185,000 yields using appropriate statistical curve fitting acres of these two crops each year, on an area equivalent procedures and to extend the detrended yields to to 22–27 square kilometers. Conversely maize, rice, and calculate the expected yield in the forthcoming cotton are cultivated on very small areas in these tehsils insurance season. In the USA under the Group (see Table 4.4 and Annex 5). It is not clear how homoge- Risk Plan (GRP) program, Skees et al. (1997) rec- neous crop yields are between farmers across these tehsils. ommended the use of linear spline regression to Typically, the tehsil-level annual crop yield data compiled detrend county average yield data. Conversely in by CRS are based on the average of about 60 CCEs, and Romania, Varangis et al. (2003) recommended therefore there may be scope to use these data to define a the use of LOESS econometric procedures in smaller administrative area to serve as the UAI. It will be SAS software to adjust area yields for trends. The important to validate these points with CRS during the reasons for detrending yields are discussed in the detailed design phase of the AYII program. following section. Given these circumstances, it is likely that the Average or expected yield can be calculated in crop AYII program in Punjab will start by using different ways, as illustrated in Table 4.5 for the five the tehsil as the UAI. crops grown in the three sample tehsils. The first column of the table shows the long-term average yield, which TABLE 4.4:  LODHRAN DISTRICT, PUNJAB: is the average of the yields over ten years (2007–08 to AVERAGE CULTIVATED AREA 2016–17), along with the standard deviation and coeffi- OF MAIN CROPS IN KEHROR cient of variation (CV). It is immediately noticeable that PLACA, DUNYAPUR, average wheat yields are very similar across the three AND LODHRAN TEHSILS tehsils and that yields exhibit very low variation over the 10-year period, as shown by CVs in mean yield of 11–13 Kehror percent. Conversely maize and cotton yields are much Pacca Dunyapur Lodhran more variable, as shown by CVs of 27–34 percent. Three Crop (acres) (acres) (acres) other estimates of expected yield are shown in Table 4.5, Wheat 126,961 163,053 178,121 including the most recent three-year average yield, the Cotton 136,320 156,960 186,786 average of the middle three of the past five years (dis- Rice 3,678 4,074 8,635 carding the years with the lowest and highest yields), and the long-term average detrended yield. For wheat, Maize 7,861 3,414 2,158 all three estimators of average yield are very similar, Sugarcane 2,104 1,456 2,048 because wheat yields have been stable over time. In com- Source: CRS data for the 10 years from 2007–08 to 2016–17 (see also Annex 5). parison, for maize in Kehror Pacca the 10-year long-term Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 51 TABLE 4.5:  PUNJAB: COMPARISON OF METHODS FOR CALCULATING AVERAGE OR EXPECTED YIELDS FOR SAMPLE TEHSILS Methods of calculating the average or expected yield Average Long term Coefficient 3-year of middle Long term average LTA standard of variation average 3 years out of average Crop/tehsil (LTA) yield deviation percent yield 5-year yields detrended yield Wheat Kehror Pacca 1,387 152 11% 1,364 1,432 1,460 Dunyapur 1,383 163 11.8% 1,418 1,433 1,504 Lodhran 1,408 180 12.8% 1,491 1,521 1,639 Rice Kehror Pacca 1,285 189 14.7% 1,454 1,434 1,506 Dunyapur 1,257 240 19.1% 1,460 1,422 1,503 Lodhran 1,331 228 17.1% 1,550 1,428 1,599 Maize Kehror Pacca 2,060 561 27.2% 2,831 2,447 2,827 Dunyapur 1,889 601 31.8% 2,609 2,246 2,579 Lodhran 1,778 532 29.9% 2,423 2,093 2,645 Cotton Kehror Pacca 688 186 27.0% 613 737 677 Dunyapur 696 235 33.8% 638 768 758 Lodhran 758 238 31.4% 703 846 840 Sugarcane Kehror Pacca 24,606 2,910 11.8% 27,947 27,463 28,844 Dunyapur 21,063 2,652 12.6% 23,777 23,100 22,893 Lodhran 25,501 2,776 10.9% 27,232 27,232 28,758 average yield is 2,060 kg/acre, whereas the average yield level will apply in a UAI is based on the CV around the over the past three years is very much higher (2,381 kg/ mean yield: in UAIs with low-yield CVs, the maximum acre) because farmers in recent years have been using an 90 percent coverage level will be applied, and in UAIs improved technology that increased yields. Rice and sug- with high CVs, only 60 percent coverage is offered. arcane also exhibit small increases in yield over time (see Annex 5 for full details on annual yields for these crops In the USA under the GRP , farmers may select and tehsils). Where crop yields are increasing (or decreas- from optional coverage levels of between 50 per- ing) over time, it becomes necessary to detrend yields to cent and 90 percent of the county average yield. design and rate crop insurance coverages. However, recognizing that some farmers achieve much higher average yields than the maximum insured yield (90 percent of the county average), the GRP allows farmers Insured Yield Coverage Levels to insure their crop at up to 150 percent of the reference AYII policies typically offer optional insured value.46 yield coverage levels of between a maximum of 90 percent and a minimum of 50 percent of the In Punjab it is recommended that coverage lev- average area yield. In India, the PMFBY offers three els of between 50 percent and 90 percent of the coverage levels: 60 percent, 80 percent, or a maximum of 90 percent of the average yield in five of the last seven years in each UAI. The decision over which coverage 46Skees et al. (1997). 52 A Feasibility Study TABLE 4.6:  DUNYAPUR TEHSIL, PUNJAB: LEVELS OF INSURED YIELD COVERAGE FOR MAIZE (KG/ACRE) Insured yield coverage level Average of middle Long-term (percent of average Long-term average 3-year average 3 years out average detrended yield) (LTA) yield yield of 5-year yields yield Average yield (100%) 1,889 2,609 2,246 2,579 90 1,700 2,349 2,022 2,321 85 1,606 2,218 1,909 2,192 80 1,511 2,088 1,797 2,063 75 1,417 1,957 1,685 1,934 70 1,322 1,827 1,572 1,805 65 1,228 1,696 1,460 1,676 60 1,134 1,566 1,348 1,547 55 1,039 1,435 1,235 1,418 50   945 1,305 1,123 1,290 average expected yield be considered for each between PKR 38,000 per acre (hybrid maize) and PKR crop in each UAI, subject to the level of yield vari- 53,000 per acre (sugarcane). ability and price.47 The principles of setting the levels of insured yield coverage are illustrated in Table 4.6 for Under the AYII program for semicommercial/ maize grown in Dunyapur tehsil. progressive farmers in Punjab, key stakehold- ers will need to decide whether to base the sum insured per acre on the value of the loan given to Basis of Valuation and Sum Insured the farmer, or on a higher value representing the Under an AYII policy, the insured crop yields can full costs of production per acre or the expected be valued either on a “costs of production basis” value of output (yield times price). Table 4.7 shows through to a “farm-gate sale price” or revenue the typical average costs per acre for the five crops, basis. In India the NAIS commonly set the sum insured which in the case of sugarcane and maize are 48 percent according to the amount of seasonal production credit higher than the maximum loan size, but only 11 percent provided to the farmer. In the USA, the GRP permits for rice. It is not known why the maximum loan size for farmers to set their yield coverage level up to as much as wheat is in fact higher than the reported average costs 150 percent of the reference price. of production. For budgeting purposes in Chapter 6, the average sums insured for Rabi crops are assumed to be In Pakistan, the CLIS establishes the sum insured PKR 30,000 per acre, and for Kharif crops for full-value based on the loan amount per acre, subject to the or “ground-up” AYII coverage, they are assumed to be maximum permitted limits recommended by PKR 40,000 per acre. the SBP for agricultural loans. These per acre val- ues are shown in Table 4.7 for the main field crops to be Note that subsequent to the submission of the insured in the start-up phase of the Punjab AYII pro- draft of this feasibility study in July 2017, the gram for semicommercial/progressive farmers. For Rabi sums insured for the Kharif 2018 pilot AYII pro- wheat, the maximum loan size is PKR 30,000 per acre, gram were set by project management for both whereas for most Kharif crops the maximum loan size is cotton and rice at the same levels of (1) PKR 50,000 per acre for “ground-up” AYII coverage for non-CLIS farmers, proving yield shortfall protection from 80 per- cent down to 0 percent of expected yield, and (2) PKR 47Note that for the Kharif 2018 Pilot AYII program, insurers and their 20,000 per acre for “top-up” AYII coverage for CLIS reinsurers advised the GoPunjab Crop Insurance Team that they would agree farmers, providing layered protection from 80 percent only to offer a maximum 80 percent insured yield coverage level. down to 50 percent of expected yield. Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 53 TABLE 4.7:  CROP LOAN INSURANCE SCHEME, PAKISTAN: LIMITS ON SUM INSURED PER ACRE (MAXIMUM LOAN SIZE) CLIS maximum sum CLIS maximum 2016–17 production cost insured PKR/acre sum insured PKR/acre Crop PKR/acre [1] for Rabi crops [2] for Kharif crops [2] Wheat 26,723 30,000 Cotton 51,138 40,000 Rice (Basmati) 43,180 39,000 Sugarcane 78,280 53,000 Maize (hybrid) 56,290 38,000 Source: [1] DoA Punjab; [2] SBP 2014. 4.4.4. AYII RATING ANALYSIS: calculating the pure loss cost premium rates, it is extremely important to adjust for any yield METHODOLOGY, PURE RATES, trends that are identified by detrending the data. AND INDICATIVE COMMERCIAL This concept is illustrated in Figure 4.5 using actual PREMIUM RATES maize yields in Dunyapur Tehsil, Lodhran District, from This section presents indicative results of a pre- 2007–08 to 2016–17. Since 2014–15, average maize liminary historical burning cost rating analysis yields in Dunyapur Tehsil increased dramatically from (HBA), which forms the basis of all actuarial rat- 1,600  kg/acre to about 2,600 kg/acre, a change that ing and pricing, based on the sample yield data DoA attributes to the introduction of high-yielding from the three tehsils. The HBA results are presented hybrid maize and improved practices, along with higher (1) for the actual 10-year historical yield data provided by levels of fertilizer use. The data show a very definite CRS and (2) for the detrended 10-year yield data. It not increasing yield trend over the last five years in the series, only demonstrates the principles of AYII rating analysis interrupted by a major yield reduction in 2013–14, pre- but shows the importance of detrending historical yield sumably due to adverse climatic conditions. The upper data before calculating coverage levels and premium red graph in the figure shows the effect of detrending rates. Full results of the HBA for wheat, rice, maize, sug- the yield data using linear trending: the effect is to adjust arcane, and cotton are presented in Annex 5. the historical yields in earlier years upwards, while main- taining years in which major yield shortfalls occur, such Ultimately, all final insurance rating and pricing as 2013–14. decisions for the AYII program will be made by insurers and reinsurers in the detailed design phase. This section merely illustrates pricing methodol- Recommendation for Punjab ogy and presents indicative rates. Under the design of the AYII program for Punjab, it is recommended that all crop yield data be detrended for calculating the expected Importance of Adjusting Time-series Yield yield (and thus coverage levels), the pure risk rates, and commercial premium rates. Data for Trends It is very important to check the historical yield data for trends over time. Yields typically show an increasing trend over time as farmers switch to new Historical Burning Cost Rating Analysis improved varieties and technology, but yields may also for “Ground-up” AYII Coverage show a declining trend where the impacts of soil degra- Annex 4 presents the full details of the HBA for dation or climate change (for example) are severe. the two options: (1) HBA applied to the actual 10-year yield data and (2) HBA applied to the 10-year detrended On a loss-of-yield insurance policy, before yields using linear trending. Results are summarized here setting the Insured Yield Coverage level and for the five main crops. 54 A Feasibility Study FIGURE 4.5:  DUNYAPUR TEHSIL, PUNJAB: ACTUAL HISTORICAL MAIZE YIELDS AND DETRENDED MAIZE YIELDS (KG/ACRE) 4,000 3,500 3,482 3,106 3,000 2,972 2,624 2,889 2,621 2,643 2,638 2,500 2,314 2,163 2,016 2,210 2,000 2,084 Actual yield 1,718 1,872 1,872 Detrended yield 1,492 1,500 1,285 1,370 Linear (actual yield) 1,310 1,000 y = 125.42x + 1199.4 R2 = 0.3991 500 0 2007–08 2008–09 2009–10 2010–11 2011–12 2012–13 2013–14 2014–15 2015–16 2016–17 Rabi wheat Kharif sugarcane The 10-year wheat yields in the three sample Sugarcane yields are also very stable in these tehsils in Lodhran District are very stable and three tehsils over the 10-year period and exhibit a exhibit only a small increasing trend over time, modest increasing trend over time. The HBA sug- with no years of severe yield loss, probably due to gests that for ground-up AYII coverage, 90 percent cover- the twin facts that the crop is irrigated and flooding is not age levels could be marketed with indicative commercial an issue in the Rabi season. The HBA applied to actual premium levels of about 7.5 percent, and for 80 percent yields suggests that for ground-up AYII coverage in these coverage, rates would be much lower at about 1.75 per- tehsils, a high level of coverage of 70 percent to 80 per- cent. The HBA applied to the detrended yields suggest cent could be provided, with indicative commercial pre- that very high coverage levels of 80–90 percent could mium rates of 3.5–5 percent applied to the sum insured. be offered, with average commercial premium rates of Results of the HBA applied to the detrended yields sug- 2.0 percent or less. gest that very high coverage levels of 80–90 percent could be offered, with average commercial premium rates of 3.5 percent or less. Kharif maize As discussed, the 10-year maize yields in the three tehsils exhibit much higher variability compared Kharif rice to the other crops, with a marked increasing The 10-year rice yields are also fairly stable trend over the last five years in the series. These across the three tehsils, but they show a higher characteristics have a major influence on the HBA rating increasing yield trend than wheat. The HBA results applied (1) to the 10-year actual yields and (2) the applied to the actual yields suggests that for ground-up detrended yields. AYII coverage and with coverage levels of 80 percent, indicative commercial premium rates might vary from Table 4.8 presents results of the HBA of actual about 2.5 percent to 7.5 percent. However, the HBA 10-year annual average area yields for maize in applied to the detrended rice yields suggests that very Dunyapur. In Dunyapur, the actual long-term average high coverage levels of 80–90 percent could be offered, yield for maize is 1,889 kg/acre. As current maize yields with average commercial premium rates of 3.5 percent are much higher, however, the conventional approach is or less. The difference in these rates again stresses the for an AYII program to calculate the average yield index importance of detrending time-series yield data. Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 55 TABLE 4.8:  DUNYAPUR TEHSIL, PUNJAB: HISTORICAL BURNING COST RATING ANALYSIS FOR GROUND-UP AYII COVER FOR ACTUAL MAIZE YIELDS FROM 2007–08 TO 2016–17 (KG/ACRE) 90% Insured Yield 80% Insured Yield 70% Insured Yield 60% Insured Yield Trigger Yield Percent Trigger Yield Percent Trigger Yield Percent Trigger Yield Percent Year Actual Yield Yield Shortfall loss % Yield Shortfall loss % Yield Shortfall loss % Yield Shortfall loss % 2007-08 1,718 2,022 303.5 15.01% 1,797 78.9 4.39% 1,572 0.0 0.00% 1,348 0.0 0.00% 2008-09 1,492 2,022 529.3 26.18% 1,797 304.7 16.96% 1,572 80.1 5.09% 1,348 0.0 0.00% 2009-10 1,310 2,022 711.1 35.18% 1,797 486.5 27.07% 1,572 261.9 16.65% 1,348 37.2 2.76% 2010-11 1,285 2,022 736.5 36.43% 1,797 511.9 28.49% 1,572 287.2 18.27% 1,348 62.6 4.65% 2011-12 1,872 2,022 149.7 7.41% 1,797 0.0 0.00% 1,572 0.0 0.00% 1,348 0.0 0.00% 2012-13 2,016 2,022 5.7 0.28% 1,797 0.0 0.00% 1,572 0.0 0.00% 1,348 0.0 0.00% 2013-14 1,370 2,022 651.4 32.22% 1,797 426.8 23.75% 1,572 202.1 12.86% 1,348 0.0 0.00% 2014-15 3,106 2,022 0.0 0.00% 1,797 0.0 0.00% 1,572 0.0 0.00% 1,348 0.0 0.00% 2015-16 2,638 2,022 0.0 0.00% 1,797 0.0 0.00% 1,572 0.0 0.00% 1,348 0.0 0.00% 2016-17 2,084 2,022 0.0 0.00% 1,797 0.0 0.00% 1,572 0.0 0.00% 1,348 0.0 0.00% LTA Average Yield 1,889 Stdev 570.25 COV% 25.4% Average for AYII Insurance# 2,246 Annual Average Loss (AAL) % 15.27% 10.07% 5.29% 0.74% Indicative Commercial Premium Rate (%) 22.22% 14.94% 8.06% 1.27% # average middle 3 out of last 5 years either as the average of the yields for the past three to This analysis is potentially misleading, however, five years, or—as in this example—the middle three of because of the major increasing yield trend for the last five years, eliminating the lowest yield year and maize grown in Dunyapur in recent years. the highest yield year, which produces an average yield of 2,246 kg/acre. Table 4.9 shows the same HBA applied to the Dunyapur detrended maize yields. In this case, The analysis shows that with a 90 percent insured at the 90 percent coverage level with an insured yield yield coverage level or 2,022 kg/acre, actual yields (detrended) of 2,321 kg/acre, there would have been would have fallen short of this guarantee yield small payouts in four years only. The year with the highest level in seven years out of ten. Shortfalls would have payouts would have been 2013–14, with an average yield occurred in all years from 2007–08 to 2013–14, with the shortfall of 449.3 kg/acre, a loss cost of 19.36 percent, worst losses in 2010–11. In that year, the yield shortfall and a 10-year AAL of only 3.13 percent and indicative would have been 736.5 kg/acre, equivalent to a percent- commercial premium rate of 5.33 percent. At the 80 per- age yield shortfall or loss cost of 36.43 percent of the cent coverage level, with an insured yield (detrended) of insured yield of 2,022 kg/acre. For the 90 percent cov- 2,063 kg/acre, the only year with yield losses would have erage level, the annual average loss (AAL) over 10 years been 2013–14; the AAL would be 0.93 percent and the would have been 15.27 percent, which is also termed the indicative commercial premium rate 1.81 percent. These “pure risk rate.” Once loadings are added to cover data indicative rates for 90 percent or 80 percent yield cover- uncertainties and insurers’ operating costs and profit mar- age obviously represent much more affordable crop AYII gin, the illustrative commercial premium rate might be in premium rates for a maize farmer to pay. Table 4.9 also the order of about 22 percent for 90 percent coverage, shows that for this 10-year analysis, there would have which would be prohibitively expensive for any farmer been no losses at the 70 percent insured yield coverage to pay. At the 80 percent coverage level (an insured yield level or lower coverage levels. of 1,792 kg/acre), the number of yield shortfall years would have been five, and the yield loss would have been smaller in each year, as shown by the reduced AAL of Kharif cotton 10.07 percent, and the indicative commercial premium Cotton yields in the three sample tehsils are rate would be 15 percent. At the 70 percent coverage highly variable owing to two years of severe yield level (an insured yield of 1,572 kg/­acre), the number of loss: 2009–10 and 2015–16, when actual yields loss years would have been further reduced to four, with were less than 50 percent of the LTA (Figure 4.6). an AAL of 5.29 percent, and the indicative commercial Cotton shows no significant yield trends over time. The premium rate would be 8.06 percent. 56 A Feasibility Study TABLE 4.9:  DUNYAPUR TEHSIL, PUNJAB: HISTORICAL BURNING COST RATING ANALYSIS FOR GROUND-UP AYII COVER FOR DETRENDED MAIZE YIELDS FROM 2007–08 TO 2016–17 (KG/ACRE) 90% Insured Yield 80% Insured Yield 70% Insured Yield 60% Insured Yield Detrended Trigger Yield Percent Trigger Yield Percent Trigger Yield Percent Trigger Yield Percent Year Yields Yield Shortfall loss % Yield Shortfall loss % Yield Shortfall loss % Yield Shortfall loss % 2007-08 2,972 2,321 0.0 0.00% 2,063 0.0 0.00% 1,805 0.0 0.00% 1,547 0.0 0.00% 2008-09 2,621 2,321 0.0 0.00% 2,063 0.0 0.00% 1,805 0.0 0.00% 1,547 0.0 0.00% 2009-10 2,314 2,321 7.3 0.31% 2,063 0.0 0.00% 1,805 0.0 0.00% 1,547 0.0 0.00% 2010-11 2,163 2,321 158.1 6.81% 2,063 0.0 0.00% 1,805 0.0 0.00% 1,547 0.0 0.00% 2011-12 2,624 2,321 0.0 0.00% 2,063 0.0 0.00% 1,805 0.0 0.00% 1,547 0.0 0.00% 2012-13 2,643 2,321 0.0 0.00% 2,063 0.0 0.00% 1,805 0.0 0.00% 1,547 0.0 0.00% 2013-14 1,872 2,321 449.3 19.36% 2,063 191.4 9.27% 1,805 0.0 0.00% 1,547 0.0 0.00% 2014-15 3,482 2,321 0.0 0.00% 2,063 0.0 0.00% 1,805 0.0 0.00% 1,547 0.0 0.00% 2015-16 2,889 2,321 0.0 0.00% 2,063 0.0 0.00% 1,805 0.0 0.00% 1,547 0.0 0.00% 2016-17 2,210 2,321 111.5 4.80% 2,063 0.0 0.00% 1,805 0.0 0.00% 1,547 0.0 0.00% LTA Detrended 2,579 yield# Stdev 465.96 COV% 18.1% Annual Average Loss (AAL) % 3.13% 0.93% 0.00% 0.00% Indicative Commercial Premium Rate (%) 5.33% 1.81% 0.00% 0.00% # Long Term Average Detrended Yield FIGURE 4.6:  KEHROR PACCA, DUNYAPUR, AND LODHRAN TEHSILS, PUNJAB: 10-YEAR AVERAGE COTTON YIELDS (KG/ACRE) 1,200 1,000 800 600 400 200 0 2007–08 2008–09 2009–10 2010–11 2011–12 2012–13 2013–14 2014–15 2015–16 2016–17 Kehror Pacca Dunyapur Lodhran Source: CRS data. See Annex 5 for full analysis. HBA applied to both actual and detrended yields sug- methodology is used as for the ground-up cover: the only gests that for ground-up AYII cover and a maximum difference is that the policy has both a threshold insured 80 percent insured yield, indicative commercial premium yield, which opens the policy for a payout, and an exit rates would have to be between 10 percent and 15 per- yield, which sets the maximum payout amount. For this cent, and that for coverage of between 60 percent and analysis, the threshold yield is set at 80 percent of the 70 percent, indicative commercial premiums would need area average yield, and the exit yield is set at 50 percent to be on the order of 7.5 percent. of the area average yield. The operation of the top-up AYII cover is illus- Historical Burning Cost Rating Analysis trated in Figure 4.7 for cotton grown in Keh- for “Top-up Cover” AYII Cover for CLIS ror Pacca Tehsil, using the original historical Farmers (untrended) 12-year yields.48 The average yield This section briefly illustrates the application of HBA to a layered or “top-up” AYII cover, which 48This updated analysis for cotton in Kehror Pacca is based on 12 years of historical is proposed for CLIS farmers. The same rating crop yields provided by CRS, as opposed to the original 10-year yield data. Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 57 (middle three of the past five years) is 757 kg/acre, so exit trigger of 379 kg/acre, so the yield shortfall paid to the 80 percent threshold yield is 606 kg/acre, while the farmers would have been capped at 227 kg/acre, equiva- 50  percent exit yield is 379 kg/acre. Reference to Fig- lent to a full payout. ure 4.7 shows that over the 12 years, payouts would have been due in three years: 2009, when the actual cotton Table 4.10 presents a comparison of the HBA for yield was 579 kg/acre, resulting in a payout of 27 kg/acre (1) the AYII top-up cover option linked to CLIS and to all insured farmers in this tehsil; 2010, when the actual (2) the ground-up cover option for cotton grown in cotton yield was 550 kg/acre, resulting in a payout of Kehror Pacca. For the top-up cover option (trigger yield 56 kg/acre to all insured farmers; and 2015, when the 80 percent; exit yield 50 percent) and insured yield layer average cotton yield was a very low 311 kg/acre. In this of 227 kg/acre, the 2009 yield shortfall of 27  kg/acre last case, the actual yield of 311 kg/acre was below the FIGURE 4.7:  KEHROR PAKKA TEHSIL, PUNJAB: EXAMPLE OF HBA FOR TOP-UP AYII COVER WITH AN 80 PERCENT TRIGGER YIELD AND 50 PERCENT EXIT YIELD FOR COTTON, BASED ON ACTUAL 12-YEAR HISTORICAL YIELDS (KG/ACRE) The Historical Yield Shortfalls Inform This Method’s Rating Calucation 1,200 1,000 800 Yield (in kg/ac) 600 400 200 0 2005 2006 2007 2008 2009 2010 20011 2012 2013 2014 2015 2016 Yield shortfall 0 0 0 0 27 56 0 0 0 0 295 0 Historical yield 821 815 736 664 579 550 959 821 743 773 311 755 Guaranteed yield 606 606 606 606 606 606 606 606 606 606 606 606 “Exit yield” 379 379 379 379 379 379 379 379 379 379 379 379 Source: World Bank. TABLE 4.10:  KEHROR PACCA TEHSIL, PUNJAB: COMPARISON OF HBA RESULTS FOR (1) TOP-UP AYII COVERAGE AND (2) GROUND-UP AYII COVERAGE FOR COTTON (1) Layered or Top Up Cover for Yield Loss from 80%to 50% (2) Ground Up Cover for Yield Loss from 80%to 0% Year Historical Trigger Exit Insured Yield Yield Percentage Year Historical Trigger Exit Yield Percentage Yield Yield (80%) Yield (50%) Layer Shortfall Payout Yield Yield Yield Shortfall Payout 2005 821 606 379 227 0 0% 2005 821 606 0 0 0 2006 815 606 379 227 0 0% 2006 815 606 0 0 0 2007 736 606 379 227 0 0% 2007 736 606 0 0 0 2008 664 606 379 227 0 0% 2008 664 606 0 0 0 2009 579 606 379 227 27 12% 2009 579 606 0 27 0 2010 550 606 379 227 56 24% 2010 550 606 0 56 0 2011 959 606 379 227 0 0% 2011 959 606 0 0 0 2012 821 606 379 227 0 0% 2012 821 606 0 0 0 2013 743 606 379 227 0 0% 2013 743 606 0 0 0 2014 773 606 379 227 0 0% 2014 773 606 0 0 0 2015 311 606 379 227 295 100% 2015 311 606 0 295 0 2016 755 606 379 227 0 0% 2016 755 606 0 0 0 Average Yield 758 Average Yield 758 Annual Average Loss (AAL-%) 11.35% Annual Average Loss (AAL-%) 5.18% Indicative Commercial Premium Rate (%) 21.25% Indicative Commercial Premium Rate (%) 9.90% Sum Insured (PKR/Acre) 20,000 Sum Insured (PKR/Acre) 50,000 Indicative Commercial Premium (PKR/Acre) 4,250 Indicative Commercial Premium (PKR/Acre) 4,950 Source: Authors. 58 A Feasibility Study represents a payout of 11.7 percent, the 2010 shortfall of India, assuming that coverage levels are similar at about 56 kg/acre represents a 24.5 percent payout, and the 2015 70–80 percent. Climatic and farming conditions in Hary- shortfall, capped at 227 kg/acre, represents a 100  per- ana State in India are possibly similar to those in Punjab cent payout. The 12-year average annual payout (pure province in Pakistan. In Haryana the average premium loss cost) would have been equivalent to 11.35 percent of rate charged for ground-up coverage is 5.05 percent for the insured yield per year. Using standard rating assump- Kharif crops and 1.2 percent for Rabi crops. The Kharif tions presented in Annex 5, the indicative commercial loss ratio suggests this average rate of 5.05 percent is premium rate for top-up coverage for cotton in this tehsil actuarially adequate, but in the case of Rabi crops the would be on the order of 21.25 percent, which if applied rate may need to be higher (Table 4.3). to the indicative sum insured of PKR 20,000/acre produces an indicative premium of PKR 4,251/acre for Chapter 6, which outlines a five-year plan for top-up cover. For the ground-up cover option (trigger building up a crop insurance program for Punjab, yield of 80 percent, exit yield 0 percent) and insured yield presents two main scenarios for target commer- of 606 kg/acre, the 2009 yield shortfall of 27 kg/acre cial premium rates for ground-up AYII coverage. represents a 4.4 percent payout, the 2010 shortfall of Under the low rate option, the target average commercial 56  kg/acre represents a 9.2 percent payout, and the premium rates are 5.0 percent for the Kharif season and 2015 shortfall of 295 kg/acre represents a payout of 3.5 percent for the Rabi season. Under the higher rate 48.6 percent, with an average payout of 5.2 percent for option, the target average commercial premium rates are the 12-year period. In this case the corresponding indic- 7.50 percent for the Kharif season and 5.0 percent for ative commercial premium rate would be 9.9 percent, Rabi season. and this rate, applied to the higher sum insured of PKR 50,000/acre for ground-up coverage, produces an indic- ative commercial premium of PKR 4,951/acre. To achieve these target commercial premium rates for ground-up AYII coverage, the insured yield coverage levels must be adjusted accord- This comparative analysis clearly illustrates ingly. The preliminary analysis presented in this chapter the first loss nature of a top-up loss of yield pol- for the three tehsils shows that rates are highly influenced icy and very much higher payout rate than for a by lowering the coverage level from say 80 percent to ground-up loss of yield policy. For cotton grown in 70 percent. Kehror Pacca Tehsil, the average percentage premium rate that would have to be charged for the AYII top-up coverage is more than double the premium rate that would have to The preliminary rating analysis also shows that be charged for ground-up coverage, because of the very for top-up AYII coverage linked to CLIS, average much higher payout rate of the top-up coverage. The cal- pure loss cost rates, and therefore commercial culated premium rate for a top-up coverage will always be premium rates, are likely to be considerably much higher than for a ground-up policy. It is most import- higher, as this is a first loss or layered protection. ant that insurers in Pakistan bear these rating principles in mind when they rate the two AYII coverage options. If GoPunjab elects to proceed with the launch of AYII coverage for semicommercial/progressive farmers in Kharif 2018, it will be necessary to Preliminary Conclusions on AYII Rating conduct a full actuarial rating analysis. Key steps in this crop AYII rating exercise would include: Analysis and Next Steps in Rating Under 1. The CRS will need to provide tehsil-level annual Phase II Design Study in Punjab crop yield data for all five main crops for all In summary, with very limited sample data from tehsils in the 36 districts of Punjab. Ideally CRS three tehsils in one district and for 10–12  years will extend the historical data series to at least 15 only, it is currently not possible to predict with years to ensure that loss years are included in the any degree of confidence the likely premium time series. rates that will need to be charged on a pro- 2) CRS should advise whether the tehsil is the most gram for ground-up AYII and coverage levels of appropriate level of UAI or whether the density approximately 70–80 percent of expected yield. of CCEs will permit the UAI to be defined at the Some guidance can be taken from the four years of pub- union council level. lished results (2011–14) from the mNAIS program in Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 59 3. The approach to rating should use standard sta- other inputs to get back into production the following tistical procedures to (1) detrend all yield data for season. The discussion starts by looking at international each crop and UAI; (2) then, through curve fitting experience—specifically at how the Government of and simulation, extend the analysis of yields to Mexico has used index insurance as a social protection say 5,000–10,000 iterations (years); and (3) calcu- coverage for subsistence farmers—and then moves on late the yield shortfall for both policy options— to outline the key features of a comparable program for ground-up coverage and top-up coverage—for Punjab. insured yield trigger levels of say 70 percent up to a maximum of 90 percent of expected yield and It is proposed to design this program for subsis- thus the average pure loss cost rates for each coverage tence farmers in 2018 for launch in Kharif 2019. level and crop and UAI. Chapter 6 presents further details on the proposed tim- 4. Insurers will need to have access to the output of ing for developing and launching each crop and live- the AYII rating study (namely, the calculated pure stock insurance program in Punjab under the five-year rates for each coverage level). It is likely that “rate SMART Punjab program. smoothing” will need to be done to ensure consis- tency in rates between neighboring UAIs in each tehsil and district. 4.5.1. MEXICO’S EXPERIENCE 5. Insurers will need to add an uncertainty load to cover data quality issues and catastrophic events WITH CROP AND LIVESTOCK that have not occurred to date to derive their tech- INDEX INSURANCE nical rates for each AYII product option (ground-up AS A MACRO-LEVEL SOCIAL coverage and top-up coverage) for trigger yield SAFETY NET COVER FOR POOR levels from 70 percent to a maximum of 90 per- SUBSISTENCE FARMERS cent of expected yield for each insured crop in each UAI. Mexico is unique in having a national- and state- 6. Finally, insurers will need to add their loadings to level catastrophe climatic parametric insurance the technical rates for each AYII product option program for subsistence farmers. The program— to derive the final commercial premium rates to Componente Atención a Desastres Naturales (CADENA, be charged to farmers: the loadings are to cover Natural Disaster Response Component Ministry of Agri- their business acquisition costs and own operat- culture, Livestock, Rural Development, Fisheries, and ing costs and reasonable profit margins. Food)—is designed to provide social safety net protection for the large numbers of small, semi-subsistence rural farming households. The program was introduced in 2003 under a public-private partnership between the fed- 4.5. CROP AYII INSURANCE eral government (Ministries of Finance and Agriculture), local state governments, Mexican insurance compa- FOR SMALL SUBSISTENCE nies, and the national agricultural reinsurance company FARMERS IN PUNJAB (Agroasemex), which provides both technical design and underwriting capacity for CADENA. GoPunjab has indicated that it is also interested in receiving proposals outlining how crop insur- Mexico uses the CADENA ex-ante insurance ance could be linked with current provincial instruments to replace traditional ex-post nat- and/or federal natural disaster relief schemes ural disaster compensation programs for the to protect the 2.2 million mainly subsistence rural poor. The program targets crop and livestock pro- farmers in the province with operations of less ducers who are deemed too poor to purchase commercial than 2.5 acres. This section presents a preliminary agricultural insurance and who were beneficiaries of the proposal for GoPunjab to consider for developing crop direct natural disaster compensation programs operated AYII as a macro-level “social protection” insurance cov- by federal and state governments. The insurance pay- erage that the GoPunjab would purchase on behalf of outs are designed to tide farmers over until the next crop this target group of subsistence farmers. The key aim season and enable them to purchase inputs. The state of this coverage would be to enable those farmers to governments are responsible for identifying and register- smooth their consumption and incomes following a ing subsistence crop and livestock farmers using criteria major climatic shock, as well as to purchase seed and 60 A Feasibility Study based on farm size for irrigated and nonirrigated hold- farmers in Punjab. GoPunjab itself would be the ings and numbers of livestock owned (see Annex 6 for insured and the policyholder, and it would buy the cover- further details). age on behalf of subsistence farmers who are landown- ers, tenants, or sharecroppers with fewer than 2.5 acres. For crops, CADENA uses two types of index insur- It is understood that all land is registered in Punjab by ance policy: (1) WII based on ground weather sta- the Board of Revenue, and this database could form the tions and/or satellite data and (2) AYII, where the basis for a registry of the targeted subsistence farmers in municipality forms the UAI and the index is based on his- each tehsil (UAI). torical municipality-yield data provided by the Ministry of Agriculture. The AYII program provides catastrophe The proposal is to use the same AYII product and yield shortfall coverage only for yield losses that exceed program that is being designed for semicommer- 70 percent of normal average yield. The affected farmers cial/progressive farmers (with 2.5–25 acres) to receive a fixed level of compensation, which is currently insure the target group of subsistence farmers set at about US$100 per hectare for rainfed crops and (with less than 2.5 acres). The rationale for using US$200 per hectare for tree crops and irrigated crops the same AYII product—in this case, it is likely that the (see Annex 6). ground-up coverage option would be adopted—is that subsistence farmers live in the same villages and com- Mexico’s federal and state governments fund munities as semicommercial/progressive farmers and CADENA based on a ratio of about 85 percent to grow the same main food crops to be insured under the 15 percent, and during 2003–11 the cost of pre- AYII program. All of these farmers are affected by cli- miums totaled about 5.01 billion Mexican pesos matic risk, which in turn affects area yields. Using the (MXN) (US$375  million). CADENA beneficiaries do same AYII product to insure and indemnify subsistence not make any contributions to the crop and livestock farmers will also ensure that the approach to settling crop insurance premiums, as they have been assessed to be too production and yield losses in these communities is uni- poor to afford coverage. fied and consistent. The same premium rates will apply for both programs, and there will be major cost savings in using the same UAIs, yield indexes, and CCEs to trigger Over the past 13 years, the government has mas- losses and payouts. sively scaled up the CADENA crop and livestock macro-level index insurance programs. These programs now reach about 2.5 million small vulnerable For subsistence farmers it is proposed to simplify crop and livestock producers (about 56 percent of all eli- the AYII program by selecting the main Rabi and gible farmers) in 31 states. Kharif crops grown in the UAI where they reside to serve as reference crops for triggering payouts. The semicommercial/progressive farmer AYII program Evaluation results show that the CADENA pro- will be linked to seasonal crop credit, and the insurance is gram not only helps to put small farmers back in therefore linked to a specific crop and area for which the business after a disaster but reduces their need semicommercial/progressive farmer is obtaining credit. to sell productive assets and leads to higher sown For subsistence farmers, however, it is not recommended area compared to non-beneficiaries. A recent study that GoPunjab register the specific crop(s) they grow and by de Janvry et al. (2016) shows that CADENA payouts the area of each crop they plan to grow on a seasonal increase expenditures by about 27 percent and incomes basis and to insure these crop areas accordingly. Such by about 38 percent for beneficiaries, and the benefits an exercise would be a major and costly undertaking. of the program exceed its costs under a wide range of Instead, it is recommended that in each UAI the main estimates (Annex 6). crop grown in the Rabi season (such as wheat) and the Kharif season (such as rice) be selected as the reference crop to trigger payouts to subsistence farmers. 4.5.2. PROPOSED GOPUNJAB AYII COVERAGE DESIGN A further simplification for the operation of this FOR SUBSISTENCE FARMERS social protection AYII coverage for subsistence It is proposed that GoPunjab consider purchasing farmers would be to insure them all for a fixed a macro-level AYII insurance coverage on behalf crop area in each crop season (Kharif and Rabi). of the approximately 2.2 million subsistence According to the 2010 census, the average farm size for Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 61 subsistence farmers with < 2.5 acres is 1.2 acres. There- LINKAGES BETWEEN CROP 4.5.3.  fore one option for GoPunjab to consider would be to INSURANCE FOR SUBSISTENCE agree to a fixed insured cropped area for each beneficiary in each UAI of one acre per season. The advantage is FARMERS AND GOPUNJAB that there would be no need to collect information on DISASTER COMPENSATION the actual area each subsistence farmer has planted each PROGRAMS FOR FARMERS season. Crop insurance could be used to complement and eventually substitute for disaster compensation In principle, subsistence farmers should be programs for subsistence farmers. In the short term, offered the same yield coverage levels as semi- crop insurance could complement the GoPunjab budget commercial/progressive farmers for each for disaster compensation for subsistence farmers, and in insured crop in each UAI. Therefore, if semi­ the medium term, timelier and more cost-­ effective ex-ante commercial/progressive farmers purchase ground-up crop insurance could become a substitute for ex-post disas- AYII for 80  percent–0 percent area yield coverage in a ter relief, as in Mexico. For governments there are several given village, subsistence farmers should also be provided key advantages of using an ex-ante macro-level insurance with exactly the same 80 percent trigger yield coverage product to finance natural disaster payments, including: level and 0 percent exit yield. This approach would differ 1) For the payment of a pre-agreed premium, the from the CADENA model, which is a Constructive Total maximum liability can be quantified in advance Loss Policy that insures only against catastrophe yield and transferred out of the fiscal budget to local and losses that exceed 70 percent of average production and international insurance and reinsurance markets. yields, or a 30 percent coverage level.49 2) Insurance payouts under an index program can be made very rapidly to state governments (and Subsistence farmers are likely to use lower lev- to farmers when there is an ex-ante farmer reg- els of purchased inputs than semicommercial/ istry), as under weather index programs there is progressive farmers, and it is recommended no need for infield assessments, and under area that they should be provided with a lower fixed yield-based index programs there is a reduced sum insured per acre in the Kharif and Rabi sea- need for such assessments. sons. In the context of this study, it is suggested that the 3) Insurance brings transparency and standardization SBP should set an agreed fixed sum insured per acre of of payout rules to disaster compensation payments. 50  percent of the maximum credit guidelines, namely PKR 15,000 per acre for the Rabi reference crop and For that reason, in planning and designing the PKR 20,000/acre for the Kharif reference crop. proposed social protection AYII coverage for sub- sistence farmers, the key stakeholders should The indemnity payment formula would be exactly work closely with PDMA Punjab to review ways of the same for AYII programs for s ­ emicommercial/ coordinating both the insurance and disaster manage- progressive farmers and for subsistence farm- ment programs. In particular, they should ensure that the ers. A hypothetical example for Kharif rice in UAI 1 programs complement each other and do not cause any can illustrate this approach. Assuming that the 80 percent beneficiaries to be indemnified twice in the event of a insured yield of rice is 1,500 kg/acre and the actual yield declared disaster or a triggered AYII insurance payout. for UAI 1 is only 1,000 kg/acre, all semicommercial/pro- gressive and subsistence farmers would be indemnified for a shortfall of 500 kg/acre (33.33 percent). The dif- 4.5.4. OPERATIONAL CONSIDERATIONS ference would be that the semicommercial/progressive FOR AN AYII PROGRAM FOR farmer with a sum insured of approximately PKR 40,000 SUBSISTENCE FARMERS per acre would be paid PKR 13,333 per acre (40,000 3 In designing the GoPunjab macro-level social protection 33.33 percent), and the subsistence farmer would be paid coverage for subsistence farmers, it will be important to PKR 6,667 per acre (PKR 20,000 3 33.33 percent). address key issues relating to the operation of the pro- gram, including: 49This »» How to register the subsistence farmers who will coverage might, however, be modified to include both proportional payouts and Constructive Total Loss payouts once actual area yields fall become the beneficiaries of the program in each below a specified threshold. district and UAI. 62 A Feasibility Study »» How to ensure that the beneficiaries receive damage caused to each type of fruit and vegetable by awareness education on the objectives of the gov- different climatic perils (including excess rain, flooding, ernment-funded AYII scheme and training to and so on)—is a major constraint to the design and rat- understand how the compensation payments will ing of a traditional named peril damage-based policy for be triggered and the amount of compensation cal- these crops. In the short to medium term, the only way to culated in their UAI. address this problem would be to try to conduct a farm- »» How to ensure that beneficiaries either have level risk assessment survey with key fruit and vegetable a bank account or an Easy-paisa mobile bank producers in Punjab and to attempt to evaluate their loss account into which the AYII insurance payment histories over the past 5–10 years for key selected crops. can be made directly. It will be essential to register each beneficiary’s bank account details at the time There may be opportunities to develop specific of enrolment into the program. WII coverages for tree crops (such as mango and citrus crops) and for vegetable crops (such as potatoes) in Punjab if the crops are located in CROP INSURANCE FOR 4.6.  areas served by an official weather station of the Punjab Meteorological Department. India has con- CASH CROPS (FRUITS siderable experience in the design of WII coverages for these crops, including insurance coverage for mango pro- AND VEGETABLES) duction against wind, excess humidity, and temperature, A recommendation under Component 3 of the designed by AICI. SMART Punjab program is to research and develop suitable crop insurance products for In 2007, ICICI Lombard helped Pepsico to design commercial horticultural farmers in Punjab in a WII coverage to protect Pepsico’s large-scale 2018, with a view to launching these products contract growers of potatoes in Punjab, India. into the market in 2019. As noted in section 4.1, This WII coverage is designed to protect against late options exist to develop both indemnity-based NPCI cov- blight disease in potatoes, and the index was constructed erages and WII coverages. according to high humidity and temperature, both of which are conducive to potato blight. The program Fruit and vegetable crops are complex to insure offers sums insured of US$500–600 per acre based on because damage or loss to the crop usually potato production costs, and it carries average premium involves a combination of quantitative (physical) rates of between 3–5 percent of the sum insured. This damage and qualitative damage that reduce the program has operated successfully for nearly a decade. price the crop can command. There are major chal- Useful lessons from the Pepsico experience and prod- lenges to design: (1) an insurance and indemnity payout uct design could possibly be adopted in Punjab under a system that will cater both for physical losses and for a potato insurance program for commercial farmers (IFAD reduction in quality in fruits or vegetables, and (2) field- and WFP 2010). based loss assessment procedures for measuring physical damage and qualitative losses. A further complication is that for many fruits and vegetables, the crop matures and is harvested over a period of weeks or even months, and 4.7. LIVESTOCK INSURANCE when losses occur it is necessary to adjust the policy for OPPORTUNITIES the amount that has already been harvested. FOR PUNJAB NPCI coverages are most suited to insuring 4.7.1. LIVESTOCK INSURANCE against perils such as hail or windstorm where direct damage to the crop can be assessed at PRODUCTS the time of loss. To design such a policy for farmers A preliminary assessment of opportunities to in Punjab, it will be necessary to have localized time develop livestock insurance was conducted as series damage data for each insured peril in each crop. part of this feasibility study. Punjab Province is an Often the lack of data and statistics on fruit and vege- important producer of dairy cattle and milk. GoPunjab table production—and especially historical data on the sees major potential for increasing the productivity of Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 63 TABLE 4.11:  TYPES OF TRADITIONAL INDEMNITY AND INDEX-BASED LIVESTOCK INSURANCE PRODUCTS AND THEIR SUITABILITY FOR PUNJAB Type of livestock Basis insurance insurance product and indemnity Availability Suitable for Punjab a) Indemnity-based livestock insurance 1. Accidental death Individual animal mortality Widespread Currently being offered by (named peril) local insurers 2. All risk accident and Individual animal mortality Widespread Only for livestock producers mortality with good animal husbandry and sanitation 3. Business interruption e.g., loss of milk production Very restricted (e.g., Currently not available. Only and income from sales Germany) applicable to large commercial dairy herds. 4. Bloodstock insurance Individual animal mortality Restricted Only applicable to high value breeding/stud animals b) Index-based livestock insurance 5. Livestock named peril Livestock mortality index for Very restricted (Mongolia) Lack of mortaility data to mortality index defined area construct such an index 6. Satellite pasture NDVI index Fairly widespread Satellite data freely available. Only drought NDVI index suitable for large rangeland areas. insurance dairy farmers by introducing improved cattle breeds cou- DEVELOPMENT OF INDEMNITY- 4.7.2.  pled with modern animal husbandry and health prac- BASED ACCIDENT AND tices, as well as improved milk marketing systems. MORTALITY COVERAGE FOR There may be opportunities for GoPunjab to pro- DAIRY CATTLE IN 2019 mote individual animal accident and mortality Under the SMART Punjab program, it is pro- coverage for dairy cattle through the banks, dairy posed to assist the GoPunjab in 2018 to research cooperatives, or (fresh) milk processors. For large and develop dairy cattle insurance with a view to commercial dairy herds, insurers may be willing to offer launching coverage in 2019. This initiative will aim All Risk Mortality coverages (see Table 4.11). At pres- to build on the existing SBP-promoted LISB Program ent, however, there appear to be no major opportunities reviewed in Chapter 3 and other private sector livestock to develop livestock index-based insurance (for instance, insurance initiatives. pasture drought NDVI coverage). 64 A Feasibility Study CHAPTER 5 LEGAL, INSTITUTIONAL, AND OPERATIONAL CONSIDERATIONS FOR A PUNJAB AGRICULTURAL INSURANCE PROGRAM 5.1. LEGAL AND REGULATORY CONSIDERATIONS In some countries, crop and livestock index insurance is not a recognized class of insurance but rather falls under the category of a derivative cover- age. Prior to designing the crop AYII program for individual semicommercial/­progressive farmers, it will be important to verify with the government insurance regulator—in this instance the Insurance Division of the Securities and Exchange Commission of Pakistan (based in Islamabad)—that the AYII coverage is permitted under current nonlife insur- ance legislation in Pakistan. Given that Pakistan already has experience with the oper- ation of crop index insurance, including both WII and AYII coverages, it is anticipated that approval to introduce such coverages into the market has already been provided by the insurance regulator, the Securities and Exchange Commission of Pakistan. It may also be useful to present the proposed macro-level policy for subsis- tence farmers to the insurance regulator for approval. The insurance sector in Pakistan has no previous experience with macro-level crop insurance policies, which are issued to the government; the government itself is thus the insured policy holder, acting on behalf of large numbers of beneficiaries (subsistence farmers) and responsi- ble for paying the premium. 5.2. INSTITUTIONAL AND ORGANIZATIONAL OPTIONS Pakistan has a group of about 12 dedicated crop and livestock insurers that operate separately and compete for business under the SBP publicly subsidized CLIS and LISB programs, which operate as public-private part- nerships. Each company therefore has established its own agricultural insurance marketing, underwriting, and claims adjusting departments, with separate reinsurance arrangements. Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 65 GoPunjab and private insurance companies issuance, premium collection, and claims settlement and interested in participating in large-scale crop and processing. Another option would be to appoint the lead livestock insurance programs for Punjab may insurer to conduct these activities on behalf of all of the have another option for organizing their opera- other companies in the pool, which would then contrib- tions. As an alternative to registering their interest with ute to the lead insurer’s operating expenses under an GoPunjab to underwrite the programs separately, as they agreed formula. In Tukey’s Tarsim pool, the lead insurer currently do with CLIS, private insurance companies is selected via a competitive process. could consider some form of coinsurance pool agree- ment, under which they agree to pool the business and to purchase common account reinsurance coverage. 5.3. OPERATIONAL Coinsurance pools are fairly common features of CONSIDERATIONS major national or regional agricultural insurance programs based on PPPs, including the Agroseguro FOR COMMERCIAL Program in Spain, the Tarsim pool program in Turkey, CROP INSURANCE FOR and various regional coinsurance pools in China. Key features of the Spanish and Turkish agricultural insur- SEMICOMMERCIAL/ ance pool programs are presented in Annex 8. Similarly, several developing countries in Africa, including Ghana, PROGRESSIVE FARMERS Kenya, Malawi, and Senegal, have formed agricultural BUNDLING WITH CROP CREDIT: 5.3.1.  coinsurance pools in recent years. THE KISSAN PROGRAM Chapter 4 recommended bundling the AYII pro- The potential advantages of coinsurance pools gram with the GoPunjab Kissan seasonal credit include: (1) cost sharing in the research and develop- program, so that farmers who are not already ment and start-up stages; (2) cost savings in establishing insured under the CLIS would be targeted to a single underwriting unit, staff, and equipment, either receive ground-up AYII coverage. International within the lead coinsurer or as a separate underwriting experience shows that bundling crop credit with crop entity (namely, a Special Purpose Vehicle); (3) the ability insurance on an automatic or compulsory basis can bring for each company to select a share according to its risk many advantages for: appetite; and (4) major cost savings in purchasing pooled 1) Farmers. Bundling makes it easier for farmers reinsurance (common account) protection (Mahul and to access crop loans for purchasing production Stutley 2010). Further information on the advantages inputs and offers a better value proposition than and disadvantages of coinsurance pools is contained in stand-alone crop insurance. The advantages of Box 5.1. this approach are particularly strong if the credit and insurance packages are also linked to input It will be important to seek guidance from the supply (for instance, bulk deliveries to farm- insurance regulator on the formation of any pool ers’ villages) and the provision of training and insurance agreement in Pakistan. In the short term extension advice on how to use improved seed/­ it is unlikely that the participating insurers would want to fertilizer technology. create and incorporate a new pool insurance company 2) Insurance companies. Mandatory bundling for the specific purpose of insuring crops and livestock of credit and insurance significantly increases the in Punjab. Rather, they are more likely to seek a simple potential for an insurance company to achieve coinsurance agreement which would allow each of them scale and a proper spread of risk, as well as sav- to take up an agreed share of the risk. In this case, as the ing on the operating costs entailed in marketing pool would not be a legal entity, it is likely that one com- and promoting the coverage, issuing policies, col- pany would be appointed to lead the pool and to issue lecting premiums, and settling claims, because policies on their own paper. The pool insurers would the business can largely be administered through also need to agree on how they would manage the busi- bank branch offices and crop insurance poli- ness. One option would be to share the workload among cies marketed at the same time that the farmer themselves for the key functions of marketing and pro- receives the loan. Here it is reasonable for the motion, education and training, underwriting and policy bank to charge a commission to manage the crop 66 A Feasibility Study BOX 5.1:  BENEFITS AND LIMITATIONS OF COINSURANCE POOL ARRANGEMENTS Given that agricultural insurance is a specialized class of insurance, and given the catastrophic nature of the risk, many insurance companies will not venture into agricultural insurance on their own, without either backing or support from the government and/ or some sort of collaboration by a host of insurance companies within a country or province. Coinsurance pool arrangements have benefits that can encourage private insurance companies to participate in offering agricultural insurance, although they have limitations as well. Benefits »» Economies of scale through operating as a single entity with shared (pooled) administration and operat- ing functions leading to costs savings due to: • Reduced staffing requirements (fixed costs). • Shared costs of product research and development, actuarial rating, and pricing. • Reduced costs of underwriting, claims control, and loss adjustment. »» Cost advantages in purchasing common account (pooled) reinsurance protection rather than having each company trying to put its own reinsurance program into place. The advantages of pooled reinsurance protection are due to: • Stronger negotiating position with reinsurers. • Larger and more balanced portfolio and better spread of risk. • Reduced costs of reinsurance due to pooled risk exposure. • Reduced transaction costs (reinsurance brokerage, and so on). »» No competition on rates in a soft market and ability to maintain technically set rates. Most pools operate as the sole insurance provider or monopoly (as in Austria, Senegal, Spain, and Turkey), and therefore there is no competition on pricing, but significant competition on service delivery (quality). »» Ability to maintain underwriting and loss adjustment standards. Under a pool monopoly arrangement, the pool manager can ensure that common and high standards are maintained in the underwriting of crop and livestock insurance and in the adjusting of claims. Where companies are competing against each other for standard crop insurance business, there is often a problem of varying loss adjustment standards between companies. Limitations »» A pool may act as the sole agricultural insurer, resulting in lack of competition in the market in terms of the: • Range of products and services offered by the monopoly pool underwriter. • Restrictions on the range of perils that are insured. • Restrictions on the regions where agricultural insurance is offered or the type of farmer insured. • Lack of competitiveness in premium rates charged by the pool. Source: Mahul and Stutley 2010. insurance business on behalf of the insurance insurance premium if the bank prefinanced the company. premium. 3) Financial institutions lending to farmers. These lenders are protected against catastrophic For that reason, it is recommended that in design- crop failure of the kind that leads large numbers ing the AYII program for semicommercial/­ of farmers to default on loans. It is common progressive farmers, interested insurers should under a crop-credit insurance program serving actively engage from the very start with the lend- individual farmers for the bank to be named in ing institutions involved in the Kissan Program the crop insurance policy as the first beneficiary in Punjab (but are not already insured under CLIS) to for its respective rights and interests—namely the agree on bundling crop credit with crop insurance and to loan amount, any interest due, and also the crop define their respective roles and responsibilities. Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 67 LINKING THE CROP AYII PROGRAM 5.3.2.  methodology and ensure maximum transpar- ency, accuracy, and timeliness in conducting FOR SEMICOMMERCIAL/ CCEs and recording and transmitting sample PROGRESSIVE FARMERS yield results from the field to CRS headquarters WITH CLIS and the participating companies offering AYII. Chapter 4 discussed the option of linking top-up Two of the key technological alternatives that CRS may AYII insurance for semicommercial/progressive wish to consider are moisture meters and wireless data farmers with CLIS. As mentioned, a top-up cover entry and transmission using smartphones or tablets. would enable a farmer to insure against a yield shortfall, say from 80 percent of the area yield down to 50 percent, Grain moisture meters could speed the process when CLIS would cut in and ensure that the bank is pro- of deriving a dry weight sample yield. The cur- tected for a loss below 50 percent of the insured yield. If rent CRS methodology for CCEs involves two steps in such a proposal were to be adopted, the main changes the field. First, grain is harvested from the sampled CCE that would need to be agreed with SBP, participating plot and weighed when it is wet; second, the grain is dried insurers, and GoPunjab authorities are that rather than for up to two weeks and then weighed again to obtain the declaring a calamity to trigger payouts on CLIS, the gov- final yield. From a crop insurance view, it is extremely ernment authorities would agree to follow the terms and important to pay claims in a timely fashion. Grain mois- conditions of the AYII policy, including (1) the definition ture meters make it possible to electronically measure of the UAI, (2) the average yield index for that UAI, and the wet and dry weights of a grain sample from a CCE (3) to base any payouts on the objective CCEs that the plot at the same time, speeding the generation of yield CRS is conducting at the time of harvest, and to make data by at least two weeks. It would not be necessary to payouts only if the actual average area yield falls short of issue a moisture meter (which costs US$250–350) to each 50 percent. CCE field team. Rather, the regional supervisors could be issued a moisture meter, collect a small sample of the grain (carefully bagged and labeled) obtained from each SALES OF AYII COVERAGE 5.3.3.  CCE conducted by the teams under their supervision, TO SEMICOMMERCIAL/ measure the moisture content of each sample the day it is PROGRESSIVE FARMERS collected, and convert that measurement to a dry weight using standardized conversion tables. WHO ARE NON-LOANEES GoPunjab wishes to promote crop insurance to all farmers, including both loanees and non-loanees. Simple smartphones or tablets could facili- In designing the AYII program for semicommercial/­ tate electronic data entry for CCEs and trans- progressive farmers, it will be necessary to consider options mit data in real-time to CRS and participating for promoting, marketing, and administering crop insur- insurance companies. The largely paper-based pro- ance for these “non-borrowing” farmers. One option cess for recording CCE field data is both cumbersome would be for the insurance companies (or pool) to mar- and liable to errors, as well as to data losses arising from ket AYII ground-up coverage to non-loanee farmers on a mailing the data to CRS headquarters. By introducing a voluntary individual basis through their networks of sales simple short message system (SMS) based app that can agents. International experience usually shows, however, be loaded into a low-cost smartphone (with GPS50 and that it is prohibitively expensive for insurers to retail cov- video recording capability), CCE results can be recorded erage to individual small-scale farmers often located in faster, with greater accuracy, and texted to CRS. Phones remote rural areas, and that it is necessary to seek to dis- equipped with GPS can record the actual location of the tribute coverage through a regional risk aggregator such as CCE, which is useful not only for auditing CCE results a lending institution, farmer cooperative, or input dealer. if necessary, but also for reducing the number of crop cuts taken at the tehsil or village levels over the medium term. A phone with video capabilities can record crop 5.3.4.  TECHNOLOGY SOLUTIONS FOR cuts while they are conducted, which again can be useful for auditing. The World Bank Group has worked with the CROP-CUTTING EXPERIMENTS insurance industry in India to test the use of smartphone The SMART Punjab program proposes to work closely with CRS-DoA to identify cost-effective ways of using technology to strengthen the CCE 50Global Positioning System, GPS. 68 A Feasibility Study technology in CCEs, and the methodology has proved first necessary to seek CRS-DoA’s assistance in very popular with the field survey teams because it speeds amassing the date required. CRS will need to pro- the process and reduces the work load. cess time series CCE yield data for the five major crops in up to 36 districts to recalculate average yields for up to 127 tehsils in Punjab. The task of recalculating average crop yields at the tehsil level represents a major undertak- KEY ROLES OF PUBLIC- 5.4.  ing, in which the insurers will need to seek GoPunjab’s PRIVATE PARTNERSHIP assistance. Similarly, the task of designing and rating an AYII coverage requires specialized actuarial skills and PLAYERS experience with designing crop insurance coverage. Here the insurers may need to request technical assistance 5.4.1.  ROLE OF PRIVATE INSURERS from the SMART Punjab program to design and rate this The full participation of private agricultural new AYII crop cover. insurers will be critical for successfully imple- menting the very ambitious GoPunjab program to develop suitable crop and livestock insur- Risk Acceptance and Underwriting ance products and programs for each segment Under the proposal to link the AYII program for of the farming population. The following insurance semicommercial/progressive farmers to the Kis- functions are considered to be principally private sector san crop credit program on a mandatory basis, functions: the insurers will need to agree on the terms and 1) Product design and rating. conditions for risk acceptance with the partici- 2) Data collection for risk assessment and product pating banks, which will ultimately be responsible for design and rating. processing the crop insurance coverage along with the 3) Risk acceptance and underwriting. loan application for each farmer. Equally important, 4) Decisions about risk retention and reinsurance the insurers will need to agree on policy issuance to the strategies. farmers and premium payments and collection with the 5) Marketing, promotion, and farmer insurance banks. education and training. 6) Distribution of crop and livestock insurance products. 7) End-of-season results declaration and claims set- Risk Retention and Reinsurance tlement strategy. Insurers usually assume full responsibility for decisions over how much risk they will retain (subject to solvency requirements set by the In practice, many functions are shared by the insurance regulator) and how much risk they private and public sectors. The public sector often will cede to local and international reinsurers. In plays a role in both risk financing and data collection; many countries, however, governments also elect to par- and although the private sector is responsible for prod- ticipate in risk financing and reinsurance, either through uct design and rating, the government will have a strong a specialist national reinsurer (as in Spain and Mexico) interest in the price of the product—and therefore in the or where federal and provincial governments assume the product’s rating—where it provides significant subsidies role of a catastrophe reinsurer (as in India under the old (World Bank 2015b). For example, in the case of CLIS NAIS program). In some start-up programs which are the GoP fully subsidizes the premium and also caps the very small, insurers may have difficulty placing their busi- losses (at 300 percent loss ratio), and hence would be ness with international reinsurers at competitive prices, very keen to understand the pricing model and should be and here governments may step in to reinsure the pro- involved in vetting the process. The functions listed above gram in its early years and until it has scaled up. are reviewed in more detail below for the planned AYII program for semicommercial/progressive farmers. In Punjab, the insurance and reinsurance capac- ity requirements of the large-scale crop and live- Data for Product Design and Rating stock insurance programs will be substantial when they are fully implemented, and the par- For insurers to design and rate the AYII program ticipating local insurers are likely to need major for semicommercial/progressive farmers, it is Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 69 support from international reinsurers. The next ROLE OF LENDING 5.4.2.  chapter provides estimates that under a scenario in INSTITUTIONS/BANKS which insurance is widely adopted (a high-uptake sce- nario), the total sum insured for the AYII program for The lending institutions (banks) will play a cen- semicommercial/­ progressive farmers in Year 1 may be tral role in implementing and managing the crop on the order of US$718 million, rising to US$2.475 bil- AYII program for semicommercial/progressive lion by Year 5 for all three of the crop insurance pro- farmers. If the banks agree to the bundling of crop grams assessed in this report. The insurers will need to credit and crop insurance and to act as the distribution engage with international reinsurers at an early stage in channel for the AYII program, their initial roles will the design of this ambitious crop insurance program for be to agree on coverage terms and conditions with the Punjab. insurers, including the coverage levels that will be offered to farmers in each UAI, the basis of the sum insured, and maximum sums insured for each crop. Then they will need to adopt the premium rates set by insurers for Marketing, Promotion, and Farmer the agreed coverage levels. At the time of negotiating the Insurance Education and Training seasonal loan, the banks will also need to process each If the AYII program for semicommercial/progres- insured farmer’s insurance application according to the sive farmers is linked to crop credit on a mandatory planned cultivated area of the crop, the sum insured, basis, the need to promote and market the policy will and premium, and then either to collect the premium up be much reduced. Even so, it is very important that front or to add it to the loan amount (prefinancing), and farmers in Punjab are made aware of the insurance finally to issue the farmer with some form of insurance program and receive education and training about how cover certificate. The banks will also play a very import- it works. It is likely that the insurers will need to seek ant role when claims are settled by distributing payments support from the financial institutions and GoPunjab to loanees’ bank accounts. to design and implement such training and communi- cation programs. ROLE OF THE GOVERNMENT 5.4.3.  OF PUNJAB Distribution Channels International experience51 shows that govern- Subject to the approval of the banks, it is proposed to ments can support agricultural crop and live- distribute the AYII coverage through the banks as part of stock programs in a number of ways. For example a bundled package with credit. governments can create an enabling legal and regulatory framework; strengthen data collection and information systems; provide technical assistance for risk assessment End-of-Season Results Declaration and product design; fund communication efforts to create awareness about the insurance products and programs and Claims Settlement Strategy to educate and train farmers to use them; make insur- The success or failure of the AYII program for ance more affordable for small farmers by subsidizing semicommercial/progressive farmers will hinge premiums; and provide risk financing (catastrophe layer on the ability of the CRS to conduct the random reinsurance). This section highlights key ways in which CCEs in each UAI in a timely, transparent, and GoPunjab can potentially support the successful imple- accurate fashion to derive the actual average mentation and upscaling of large-scale crop and livestock yield for the insured crop in each UAI. The sys- insurance programs. tem must be one in which insured farmers and their local representative have full trust. The insurers in conjunction with CRS-DoA and the supporting banks will need to There are four main areas where GoPunjab’s design an end-of-season results declaration strategy for financial support to the crop insurance start-up each UAI, including the publication of CCE yield results in each UAI whether a claim is due or not. Where claims 51For a review of government support to agricultural insurance, see Mahul payouts are due, the procedure for settling the claims and and Stutley (2010), which presents the findings for a survey of public and repayment of farmer loans must be agreed by the insur- private agricultural insurance programs in 65 countries and types of support ers with the banks. provided by government. 70 A Feasibility Study and annual operating costs would be critical to In addition, GoPunjab support in the form of the successful implementation of this program, subsidies for crop (and livestock) insurance pre- including: miums for small farmers will be very import- 1) Data strengthening for crop insurance, ant in determining the demand for insurance including designing and implementing a farmer programs and their capacity to scale up. The electronic registration and database system, and following premium subsidy levels are recommended for providing insurers with time series yield data at GoPunjab to consider for the four programs outlined in the tehsil level for the major crops. this report: 2) Strengthening the CCEs for area yield »» Program 1: AYII for semicommercial/progressive estimation. As noted, the government can sup- farmers: 50 percent premium subsidy. port this effort by: (1) significantly increasing the »» Program 2: AYII social protection program for density of CCEs to permit the UAI to be set at subsistence farmers: 100 percent premium subsidy. the union council or eventually even the individ- »» Program 3: NPCI for tree fruit producers: 50 per- ual village level, and (2) introducing technology cent premium subsidy. (moisture meters, as well as smartphone or tablet »» Program 4: Dairy cattle insurance: 50 percent pre- technology) to rapidly obtain and record CCE mium subsidy. data and transmit it in real-time to underwrit- ers and other stakeholders. This technology has GoPunjab will need to establish an annual been developed, tested, and widely implemented budget to cover the premium subsidies and in India’s PMFBY program. contributions to start-up and operating costs, 3) Investing in farmer awareness, educa- and appoint an institution that will be respon- tion, and training in the role of crop sible for administering the premium subsidy insurance and the operation of the var- regime on its behalf. The norm in most subsidized ious insurance products and programs. agricultural insurance programs is that (1) the farmer Building insurance literacy among farmers is a is charged only the unsubsidized portion of the pre- key pillar of a sustainable crop insurance pro- mium, and (2) the insurer then reclaims the amount of gram under SMART Punjab. the premium subsidy from the entity appointed by the 4) Monitoring and evaluation (M&E). It is crit- government to audit, process, and repay the premium ical to implement an M&E system to assess the subsidies. insurance programs’ inputs and outputs, timeliness, and effectiveness, as well as their impacts over time on the input purchasing decisions, crop yields, and The next chapter presents a five-year plan for incomes of semicommercial/­ progressive farm- building up a crop insurance program, complete ers. For subsistence farmers, M&E should focus with physical and financial projections for Pun- on measuring whether insurance enables them to jab Province, based on the World Bank team’s maintain their consumption levels following major best estimates for the types of crop insurance floods or droughts and whether they are able to envisioned. return to production in the following season. Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 71 CHAPTER 6 A FIVE-YEAR PLAN AND BUDGET FOR BUILDING AND SCALING UP A LARGE-SCALE CROP INSURANCE PROGRAM IN PUNJAB For the consideration of GoPunjab, this chapter presents a five-year (FY2018/19 to FY2022/23) plan to build up a crop insurance program along the lines discussed in previous chapters, starting in Kharif 2018. Four sce- narios distinguished by contrasting uptake and premium rates are used to develop projections for scaling up the program (numbers of insured farmers, insured area, indicative sums insured, and premiums) and of the fiscal load for GoPunjab. A FIVE-YEAR BUILD-UP PLAN, PORTFOLIO 6.1.  PROJECTIONS, AND FINANCIAL REQUIREMENTS Figure 6.1 shows a proposal for GoPunjab, in conjunction with its private sector partners, to introduce three large-scale crop insurance programs in two phases over FY2018/19 and FY2019/20. The first phase, starting in Kharif 2018, would introduce an AYII program for semicommercial/progressive farmers with 2.5–25 acres. The second phase, starting in Kharif 2019, would introduce two addi- tional crop insurance programs: AYII for subsistence farmers with less than 2.5 acres, and NPCI for tree fruit and vegetable farmers. The three main crop insurance programs to be developed and launched over 2018–19 to 2022–23 were presented in detail in Chapter 4. To recapitulate: 1) AYII Program 1: AYII for semicommercial/progressive farmers with 2.5–25 acres. This group of 2.9 million farmers represents 56 per- cent of all farmers in Punjab (Table 6.3). For the reasons discussed in Chap- ter 4, AYII Program 1 will be explicitly linked to two crop-credit/seasonal loan schemes in Punjab, the federal CLIS and the GoPunjab Kissan credit program. This seasonal AYII program will initially insure Kharif rice, maize, cotton, and Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 73 FIGURE 6.1:  PUNJAB: PROPOSED PHASING OF NEW CROP INSURANCE PROGRAMS, FY2018/19 TO FY2022/23 Financial year FY 2018–19 FY2019–20 FY2020–21 FY2021–22 FY2022–23 Crop season Kharif Rabi Kharif Rabi Kharif Rabi Kharif Rabi Kharif Kharif Crop insurance programs Area yield index insurance Launch Kharif 2018 1.  for progressive farmers >2.5 Ac <25 Ac Area yield index insurance 2.  Launch Kharif 2019 for subsistence farmers <2.5 Ac Named peril crop 3.  Launch Kharif 2019 insurance for tree fruit and vegetable farmers Livestock insurance programs Dairy cattle insurance 1.  Launch in FY2019–20 (indemnity-based accident and mortality cover) sugarcane, and Rabi wheat. The CRS will pro- TABLE 6.1:  UPTAKE AND PRICING vide seasonal area yield estimates for these crops SCENARIOS ANALYZED based on objective CCEs to enable the area yield index approach. Premium pricing/ Target high Medium 2) AYII Program 2: AYII for subsistence uptake scenarios uptake rate uptake rate farmers with less than 2.5 acres. GoPun- Target (low) average Scenario 1 Scenario 3 jab will implement this fully subsidized social commercial premium protection program in conjunction with suitable rates local risk aggregators/distributors. AYII Program Higher average Scenario 2 Scenario 4 2 is also seasonal; it will insure an agreed sum per premium rates acre in the Kharif and Rabi seasons. 3) Program 3: Tree fruit NPCI for mango and citrus farmers. This program would pro- rates (Kharif season 7.5 percent, Rabi season 5.0 per- tect producers from specific perils such as frost, cent). The target commercial premium rates are based on hail, and wind. international experience as well as the preliminary anal- ysis presented in Chapter 4 and Annex 5 for ground-up AYII crop insurance. The targets will need to be refined 6.1.1  FOUR SCENARIOS FOR THE and confirmed by insurers and their reinsurers in the design phase of the program. Premium rates will need PROPOSED CROP INSURANCE to be established for each crop in each UAI according to PROGRAMS the type of AYII policy that applies—namely, ground-up Four scenarios for the proposed crop insurance cover for non-CLIS farmers and top-up cover for CLIS programs are presented in Table 6.1. Each scenario farmers—and for the agreed coverage level(s). reflects a combination of two premium pricing rates (a target low rate and a higher rate) and two uptake levels The two uptake levels in Table 6.1 consist of a high (a target high uptake level and a medium uptake level). uptake level, in which the assumed penetration rates by Year 5 are 25 percent for semicommercial/progressive More specifically, the premium pricing rates in farmers and 80 percent for subsistence farmers, and a Table 6.1 consist of the target (low) average commercial medium uptake level, estimated at 50 percent of the premium rates (Kharif season 5.0 percent, Rabi season Year 5 uptake (number of farms and insured area) for all 3.5 percent) and higher average commercial premium three crop insurance programs. 74 A Feasibility Study 6.1.2. PORTFOLIO PROJECTIONS: projected to rise from 2.1 million acres in FY2018/19 to 5.25 million acres in FY2022/23 (Table 6.2). NUMBERS OF INSURED FARMERS, INSURED CROPS, AND INSURED For AYII Program 2 (the macro-level fully funded AREA program for subsistence farmers), the assumed For AYII Program 1 (focusing on semicommer- uptake rate by Year 5 would be 1.75 million farm- cial/progressive farmers), it is assumed that ers per season (Kharif and Rabi), or 3.5 million at full scale implementation by Year 5, about farmers per year, equivalent to an uptake rate of 750,000 farmers would be insured in both the nearly 80 percent (Table 6.2 and Figure 6.2). The rea- Rabi and Kharif season, respectively (or 1.5 mil- son for the very high planned uptake rate is that GoPun- lion farmers in total per year) (Table 6.2 and Fig- jab would provide this cover on an automatic basis for ure 6.2). This level of participation represents an uptake all subsistence farmers who are eligible and who register (penetration) rate of about 1 in every 4 (26 percent) of for it. Subsistence farmers have an average farm size of all semicommercial/progressive farmers. This assump- 1.2 acres, and this analysis assumes that each will receive tion is based both on the numbers of CLIS and Kissan automatic AYII coverage for a fixed area of 1 acre per farmers, as well as on international experience. Accord- season (Table 6.3). It is planned to launch AYII Pro- ing to the 2010 census data, the average farm size for gram 2 for subsistence farmers in FY2019/20 with a total semi­commercial/­ progressive farmers is 6.9 acres, and insured Kharif and Rabi area of 0.75 million acres, ris- the portfolio is modeled on the basis that the average ing to 3.5 million acres by FY2022/23 (Table 6.2). farmer cultivates and insures 50 percent of his/her farm area each season under any of the five insurable crops, For Program 3, NPCI for citrus and mango farm- giving an insured area of 3.5 acres per semicommercial/­ ers, the uptake rate by Year 5 is assumed to be progressive farmer per season. Based on these assump- 10,000 farmers with a total of 25,000 insured tions, about 13 percent of the total farm area of this acres in Punjab. (Table 6.3 and Figure 6.1). This esti- group of farmers would be insured under AYII Program 1 mate may need to be revised upward if the government each season (Table 6.2). Total insured area of Rabi and makes this program a priority. Kharif crops for semicommercial/progressive farmers is TABLE 6.2:  PORTFOLIO PROJECTIONS, FY2018/19 TO FY2022/23: NUMBER OF INSURED FARMERS, INSURED AREA, AND SUM INSURED FOR SCENARIO 1 (AVERAGE COMMERCIAL PREMIUM RATES 5.0 PERCENT IN KHARIF SEASON AND 3.5 PERCENT IN RABI SEASON) Program / Item FY2018/19 FY2019/20 FY2020/21 FY2021/22 FY 2022/23 Total Number of Insured Farmers: Program 1. Crop AYII for Commercial Farmers > 2.5 Acres 600,000 1,000,000 1,250,000 1,425,000 1,500,000 5,775,000 Program 2. Crop AYII Catastrophe Cover Small Farmers < 2.5 Ac. 750,000 1,750,000 2,750,000 3,500,000 8,750,000 Program 3. Crop Insurance for Tree Fruit (Mango, Citrus) 2,500 5,000 7,500 10,000 25,000 Total Insured Farmers 600,000 1,752,500 3,005,000 4,182,500 5,010,000 14,550,000 Insured Area (Acres) Program 1. Crop AYII for Commercial Farmers > 2.5 Acres 2,100,000 3,500,000 4,375,000 4,987,500 5,250,000 20,212,500 Program 2. Crop AYII Catastrophe Cover Small Farmers < 2.5 Ac. 750,000 1,750,000 2,750,000 3,500,000 8,750,000 Program 3. Crop Insurance for Tree Fruit (Mango, Citrus) 6,250 12,500 18,750 25,000 62,500 Total Insured Area (Acres) 2,100,000 4,256,250 6,137,500 7,756,250 8,775,000 29,025,000 Sum Insured (US$) Program 1. Crop AYII for Commercial Farmers > 2.5 Acres 717,500,000 1,207,500,000 1,522,500,000 1,741,250,000 1,837,500,000 7,026,250,000 Program 2. Crop AYII Catastrophe Cover Small Farmers < 2.5 Ac. 125,000,000 300,000,000 475,000,000 612,500,000 1,512,500,000 Program 3. Crop Insurance for Tree Fruit (Mango, Citrus) 6,250,000 12,500,000 18,750,000 25,000,000 62,500,000 Total Sum Insured (US$) 717,500,000 1,338,750,000 1,835,000,000 2,235,000,000 2,475,000,000 8,601,250,000 Premium Income (US$) Program 1. Crop AYII for Commercial Farmers > 2.5 Acres 30,362,500 51,712,500 65,887,500 75,643,750 80,062,500 303,668,750 Program 2. Crop AYII Catastrophe Cover Small Farmers < 2.5 Ac. 5,125,000 12,750,000 20,375,000 26,687,500 64,937,500 Program 3. Crop Insurance for Tree Fruit (Mango, Citrus) 625,000 1,250,000 1,875,000 2,500,000 6,250,000 Total Premium Income (US$) 30,362,500 57,462,500 79,887,500 97,893,750 109,250,000 374,856,250 Source: Authors’ calculations. Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 75 FIGURE 6.2:  NUMBER OF INSURED FARMERS BY YEAR AND PROGRAM TYPE, 2018/19–2022/23 6,000,000 5,040,000 5,000,000 Number of insured farmers 4,202,500 4,000,000 3,500,000 3,015,000 3,000,000 2,750,000 2,000,000 1,752,500 1,750,000 1,425,000 1,500,000 1,250,000 1,000,000 1,000,000 750,000 600,000 600,000 0 0 2018/19 2019/20 2020/21 2021/22 2022/23 Program 1. Crop AYII for commercial farmers > 2.5 acres Program 2. Crop AYII catastrophe protection for small farmers < 2.5 acres Program 3. Crop insurance for fruit and vegetables Total insured farmers Source: Authors. Note: Total number of insured farmers per year is sum for Kharif and Rabi seasons. TABLE 6.3:  UPTAKE PROJECTIONS FOR AREA YIELD INDEX INSURANCE FOR SEMICOMMERCIAL/PROGRESSIVE FARMERS AND SUBSISTENCE FARMERS (AYII PROGRAMS 1 AND 2) Number Percent Percent of Average farm size Farm size (acres) of farms of farms Farm area (acres) farm area (acres) < 2.5 2,203,102 42% 2,602,187 9% 1.2 2.5 to < 25.0 2,916,214 56% 20,186,927 69% 6.9 >25.0 130,512 2% 6,525,444 22% 50.0 Total 5,249,828 100% 29,314,558 100% 5.6 Number insured farmers per Percent Insured area per Percent of Average insured Crop insurance season by of farms season by year 5 farm area area/season program year 5 insured (acres) insured (acres) 2) Subsistence 1,750,000 79% 1,750,000 67% 1.0 farmers < 2.5 acres 1) Progressive farmers 750,000 26% 2,625,000 13% 3.5 2.5 to < 25 acres Source: Farm size data based on 2010 Census. Note: With two cropping seasons, the total number of insured farmers for AYII Program 1 is 750,000 3 2 = 1,500,000 per year. For AYII Program 2, the total number of insured farmers per year is 1,750,000 3 2 = 3,500,000. The portfolio modeling assumes the same farmers are insured in both seasons. 76 A Feasibility Study 6.1.3.  SUMS INSURED Under Scenario 1, the target average commercial For AYII Program 1 (ground-up AYII for semi- premium rates identified for AYII Programs 1 commercial/progressive farmers linked to crop and 2 are 5 percent for Kharif crops and 3.5 per- credit), the sum insured is likely to be linked to cent for Rabi crops, and an indicative commer- the amount of the loan extended by the financial cial premium rate of 10.0 percent is estimated for institution, although farmers wishing to obtain Program 3 (tree fruit). Based on these assumptions, this coverage with a higher sum insured could Table 6.2 shows that in Year 1 (FY2018/19) the commer- request to do so. The following per acre dollar sums cial premium income is estimated at US$30.4 million, ris- insured are used in this budgeting exercise for AYII Pro- ing by Year 5 (FY2022/23) to US$109.2 million. Given gram 1: that AYII Program 1 is by far the largest, it accounts for »» Rabi crop: US$300 (PKR 30,000) per acre nearly three-quarters of the annual premium (US$80 mil- »» Kharif crops: US$400 (PKR 40,000) per acre lion) by Year 5. Full details are presented by cropping sea- son in Annex 9. For AYII Program 2 (ground-up AYII as a social protection coverage for subsistence farmers), The above five-year crop insurance physical this budgeting exercise assumes a flat rate sum uptake and financial projections are intended to insured of US$200 per acre for Kharif crops and assist GoPunjab to prepare its own five-year crop US$150 per acre for Rabi crops. These sums insured insurance business plan and financial budget to are equal to 50 percent of the indicative sums insured cover (1) premium subsidies and (2) financial for semicommercial/progressive farmers and reflect the support to other start-up and ongoing operating fact that subsistence farmers are likely to use lower lev- costs as identified in Chapter 5. It is noted that at els of purchased inputs. In the design phase, these esti- the time of finalizing this report, the Crop Insurance mates of sums insured should be refined with program Team attached to CRS-DOA-GoPunjab is in the process management. of finalizing its own five-year crop insurance projections and costed business plan and budget. For Program 3 (tree crops), a sum insured of US$1,000 (PKR 100,000) per acre is assumed. 6.2. COSTS TO GOVERNMENT In Year 1 of the crop insurance program, for Scenarios 1 and 2 (the high uptake scenarios), OF CROP INSURANCE total sum insured (TSI) is estimated at US$717.5 PREMIUM FINANCING million, rising by Year 5 to US$2,475.0 million. AYII Program 1 for semicommercial/progressive farm- AND OTHER PROGRAM ers is the largest program, with an estimated TSI of OPERATING COSTS US$1,837.5 million by Year 5 (FY2022/23), followed by AYII Program 2 for subsistence farmers, with a TSI of GoPunjab has indicated its commitment to pro- US$612.5 million, and Program 3 for tree fruit, with a viding financial support to the crop insurance TSI of US$25 million (Table 6.2). programs both in the form of (1) support to start-up and operating costs, and (2) premium subsidies. This section presents the indicative costs of 6.1.4 INDICATIVE COMMERCIAL government financial support to the crop insurance pro- gram over 5 years for Scenario 1, high uptake rates and PREMIUMS target premiums on the 2 AYII programs of 5 percent Indicative commercial crop insurance premi- for the Kharif season and 3.5 percent for the Rabi sea- ums are presented here so that GoPunjab can son. Note, however, that these cost estimates are based assess the possible annual cost of premiums and on the analyses in the previous chapters, and therefore plan the premium subsidy program accordingly. potentially subject to substantial change depending on At the same time, it is vital to stress that all crop insurance the GoPunjab’s policies and decisions. pricing decisions will be made by local insurers and their reinsurers. Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 77 COST OF CROP INSURANCE 6.2.1.  3. Program 3: NPCI for tree fruit producers: 50 per- cent premium subsidy. PREMIUM SUBSIDIES The costings are based on the following pre- mium subsidy levels expressed as a percentage Using these assumptions, the cost to GoPunjab of the commercial premium rate for each crop of premium subsidies could be in the order of program: US$15.2 million in year 1, rising to US$68.0 mil- 1. Program 1: AYII for semicommercial/progres- lion by year 5 at full-scale implementation of the sive farmers: 50 percent premium subsidy. program. Over the five-year life of the program, the 2. Program 2: AYII social protection program for total cost of premium subsidies may be in the order of subsistence farmers: 100 percent premium subsidy. about US$220 million (Table 6.4 and Figure 6.3). TABLE 6.4:  PUNJAB: COSTS OF GOVERNMENT SUPPORT TO CROP INSURANCE PREMIUM SUBSIDIES AND PROGRAM IMPLEMENTATION COSTS, SCENARIO 1 (HIGH UPTAKE RATES AND AVERAGE PREMIUM RATES OF 5.0 PERCENT IN KHARIF SEASON AND 3.5 PERCENT IN RABI SEASON) Program / Item FY 2018/19 FY 2019/20 FY 2020/21 FY 2021/22 FY 2022/23 Total Premium Subsidies (US$) Program 1. Crop AYII for Commercial Farmers > 2.5 Acres 15,181,250 25,856,250 32,943,750 37,821,875 40,031,250 151,834,375 Program 2. Crop AYII Catastrophe Cover Small Farmers < 2.5 Ac. 5,125,000 12,750,000 20,375,000 26,687,500 64,937,500 Program 3. Crop Insurance for Tree Fruit (Mango, Citrus) 312,500 625,000 937,500 1,250,000 3,125,000 Sub-Total Premium Subsidies 15,181,250 31,293,750 46,318,750 59,134,375 67,968,750 219,896,875 Other Financial Costs borne by Government (US$) Data strenthening for Crop Insurance 1,500,000 1,000,000 750,000 500,000 500,000 4,250,000 Strenthen Crop Cutting Experiments (mobile phone system) 300,000 1,200,000 1,500,000 1,800,000 2,100,000 6,900,000 Farmer insurance awareneness, education and training 1,200,000 1,000,000 1,000,000 1,000,000 1,000,000 5,200,000 Monitoring and Evaluation 250,000 400,000 400,000 400,000 1,000,000 2,450,000 Sub-Total Other costs 3,250,000 3,200,000 3,250,000 3,300,000 4,600,000 18,800,000 Total Budgeted Costs to Government of Punjab 18,431,250 34,493,750 49,568,750 62,434,375 72,568,750 238,696,875 Cost per insured farmer 30.7 19.7 16.5 14.9 14.5 16.4 Source: Authors’ calculations. FIGURE 6.3:  ESTIMATED COSTS OF CROP INSURANCE PREMIUM SUBSIDIES FY2017/18 TO FY2021/22 (US$) 80,000,000 70,000,000 67,968,750 60,000,000 59,134,375 50,000,000 46,318,750 40,000,000 31,293,750 30,000,000 20,000,000 15,181,250 10,000,000 0 2017–18 2018–19 2019–20 2020–21 2021–22 Program 1. Crop AYII for commercial farmers > 2.5 acres Program 2. Crop AYII catastrophe coverage for small farmers < 2.5 acres Program 3. Crop insurance for fruit and vegetables Subtotal premium subsidies Source: Authors’ calculations. 78 A Feasibility Study OTHER CROP INSURANCE 6.2.2.  strengthening, the Year 1 budget of US$1.5  million includes the design of an electronic registration system, PROGRAM COSTS BORNE the purchase of hardware and software, and subsequent BY GOPUNJAB allocations for the field-level costs of registering farm- Chapter 5 identified four ways in which financial ers, for a total budget over five years of US$4.25 mil- support from GoPunjab is critical for success- lion. For the CCEs, the budget includes an allocation fully starting up and implementing a crop insur- for purchasing equipment (smartphones and design of a ance program on a large scale: suitable SMS-based app, grain moisture meters, weigh- 1) Data strengthening for crop insurance. ing scales, and other items), along with a contribution 2) Strengthening the CCEs for area yield estimation. toward the costs of conducting CCEs. The budget for 3) Investing in farmer awareness, education, and strengthening the CCEs over five years is estimated at training in crop insurance. US$6.9 million. The cost of farmer awareness, educa- 4) M&E. tion, and training is estimated at US$5.2 million over five years and includes the initial costs of designing training The sections that follow present preliminary estimates of materials and of training the trainers and farmers in each the program support costs involved with (1)–(4) above. district over time. Finally, the M&E budget is estimated at Note, however, that like the estimates of premium subsidy US$2.45 million over five years. costs, these estimates are based purely on the analyses in the previous chapters, and therefore they are potentially subject to substantial change depending on GoPunjab TOTAL COSTS TO GOPUNJAB OF 6.2.3.  policies and decisions. SUPPORT TO CROP INSURANCE PROGRAMS Under Scenario 1 (high uptake), these other pro- Under Scenario 1 (high uptake and target pre- gram support costs are estimated at US$3.25 mil- mium rates of 5.0 percent Kharif and 3.5 percent lion in Year 1, rising to US$4.6 million by Year 5, Rabi) the total budgeted costs to GoPunjab of the with a total estimated cost to GoPunjab of premium subsidy, start-up, and operation of the US$18.8 million (Figure 6.4 and Table 6.4). For data FIGURE 6.4:  ESTIMATED COSTS OF GOVERNMENT SUPPORT TO CROP INSURANCE PROGRAM START-UP AND OPERATING COSTS, FY2017/18–FY2021/22 (US$) 5,000,000 4,600,000 4,500,000 4,000,000 3,500,000 3,300,000 3,250,000 3,200,000 3,250,000 3,000,000 2,500,000 2,000,000 1,500,000 1,000,000 500,000 0 2017–18 2018–19 2019–20 2020–21 2021–22 Data strengthening for crop and livestock insurance Strengthen crop cutting experiments (mobile phone system) Farmer insurance awareness, education, and training Subtotal other costs Source: Authors’ calculations. Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 79 FIGURE 6.5:  TOTAL COSTS OF GOPUNJAB SUPPORT TO CROP INSURANCE (PREMIUM SUBSIDIES AND SUBSIDIES ON OPERATING COSTS), FY2017/18–FY2021/22 (US$) 250,000,000 238,696,875 200,000,000 150,000,000 100,000,000 72,568,750 62,434,375 49,568,750 50,000,000 34,493,750 18,431,250 0 2017–18 2018–19 2019–20 2020–21 2021–22 Total Premium subsidies Operating expenses subsidies Total costs to Punjab government Source: Authors’ calculations. three crop insurance programs are estimated to to be charged by insurers and their reinsurers to provide reach US$72.6 million per year by Year 5 (or at full the desired levels of protection to farmers. GoPunjab implementation). Over the five years, the total cost of will need to budget for the event that higher rates are government support is estimated at US$238.7 million, required. which works out at US$16.4 of crop insurance subsidy support per farmer beneficiary (Table 6.4, Figure 6.5). Under Scenario 2, the total cost of crop insur- ance premiums will be about 47 percent higher than under Scenario 1. In other words, at Year 5 6.3. CROP INSURANCE (full-scale implementation) the premium bill will rise from US$109.2 million per year to US$160 million per SENSITIVITY ANALYSIS year, and the total premium over five years will rise from US$375 million to US$549 million (Table 6.5). SCENARIO 2: HIGH UPTAKE 6.3.1.  RATES, BUT HIGHER AVERAGE Under Scenario 2, the costs to GoPunjab of pre- CROP INSURANCE PREMIUM mium subsidies at Year 5 will increase from RATES US$68  million to nearly US$100 million per year, Scenario 2 is based on the same high uptake rates and total premium subsidies over five years will rise to as Scenario 1, but the average premium rates US$322.5 million. The costs of government support for both AYII Programs 1 and 2 are higher, at for start-up and operating costs would remain the same 7.5 percent for Kharif season and 5.0 percent for under Scenario 2, at US$18.8 million over five years. The Rabi season. During the detailed design and planning total costs to GoPunjab of financial support to the crop stage of this crop insurance program, a granular actu- insurance programs would be about US$104.3 million at arial analysis at the tehsil level in all districts of Punjab Year 5, full-scale implementation (Table 6.5). may show that higher average premium rates will need 80 A Feasibility Study TABLE 6.5:  SCENARIO 2: FIVE-YEAR CROP INSURANCE PORTFOLIO PROJECTIONS FOR HIGH UPTAKE AND HIGHER AVERAGE PREMIUM COSTS Program / Item FY2018/19 FY2019/20 FY2020/21 FY2021/22 FY 2022/23 Total Number of Insured Farmers: Program 1. Crop AYII for Commercial Farmers > 2.5 Acres 600,000 1,000,000 1,250,000 1,425,000 1,500,000 5,775,000 Program 2. Crop AYII Catastrophe Cover Small Farmers < 2.5 Ac. 750,000 1,750,000 2,750,000 3,500,000 8,750,000 Program 3. Crop Insurance for Tree Fruit (Mango, Citrus) 2,500 5,000 7,500 10,000 25,000 Total Insured Farmers 600,000 1,752,500 3,005,000 4,182,500 5,010,000 14,550,000 Insured Area (Acres) Program 1. Crop AYII for Commercial Farmers > 2.5 Acres 2,100,000 3,500,000 4,375,000 4,987,500 5,250,000 20,212,500 Program 2. Crop AYII Catastrophe Cover Small Farmers < 2.5 Ac. 750,000 1,750,000 2,750,000 3,500,000 8,750,000 Program 3. Crop Insurance for Tree Fruit (Mango, Citrus) 6,250 12,500 18,750 25,000 62,500 Total Insured Area (Acres) 2,100,000 4,256,250 6,137,500 7,756,250 8,775,000 29,025,000 Sum Insured (US$) Program 1. Crop AYII for Commercial Farmers > 2.5 Acres 717,500,000 1,207,500,000 1,522,500,000 1,741,250,000 1,837,500,000 7,026,250,000 Program 2. Crop AYII Catastrophe Cover Small Farmers < 2.5 Ac. 125,000,000 300,000,000 475,000,000 612,500,000 1,512,500,000 Program 3. Crop Insurance for Tree Fruit (Mango, Citrus) 6,250,000 12,500,000 18,750,000 25,000,000 62,500,000 Total Sum Insured (US$) 717,500,000 1,338,750,000 1,835,000,000 2,235,000,000 2,475,000,000 8,601,250,000 Premium Income (US$) Program 1. Crop AYII for Commercial Farmers > 2.5 Acres 44,625,000 76,125,000 97,125,000 111,562,500 118,125,000 447,562,500 Program 2. Crop AYII Catastrophe Cover Small Farmers < 2.5 Ac. 7,500,000 18,750,000 30,000,000 39,375,000 95,625,000 Program 3. Crop Insurance for Tree Fruit (Mango, Citrus) 625,000 1,250,000 1,875,000 2,500,000 6,250,000 Total Premium Income (US$) 44,625,000 84,250,000 117,125,000 143,437,500 160,000,000 549,437,500 Program / Item FY 2018/19 FY 2019/20 FY 2020/21 FY 2021/22 FY 2022/23 Total Premium Subsidies (US$) Program 1. Crop AYII for Commercial Farmers > 2.5 Acres 22,312,500 38,062,500 48,562,500 55,781,250 59,062,500 223,781,250 Program 2. Crop AYII Catastrophe Cover Small Farmers < 2.5 Ac. 7,500,000 18,750,000 30,000,000 39,375,000 95,625,000 Program 3. Crop Insurance for Tree Fruit (Mango, Citrus) 312,500 625,000 937,500 1,250,000 3,125,000 Sub-Total Premium Subsidies 22,312,500 45,875,000 67,937,500 86,718,750 99,687,500 322,531,250 Other Financial Costs borne by Government (US$) Data strenthening for Crop Insurance 1,500,000 1,000,000 750,000 500,000 500,000 4,250,000 Strenthen Crop Cutting Experiments (mobile phone system) 300,000 1,200,000 1,500,000 1,800,000 2,100,000 6,900,000 Farmer insurance awareneness, education and training 1,200,000 1,000,000 1,000,000 1,000,000 1,000,000 5,200,000 Monitoring and Evaluation 250,000 400,000 400,000 400,000 1,000,000 2,450,000 Sub-Total Other costs 3,250,000 3,200,000 3,250,000 3,300,000 4,600,000 18,800,000 Total Budgeted Costs to Government of Punjab 25,562,500 49,075,000 71,187,500 90,018,750 104,287,500 341,331,250 Cost per insured farmer 42.6 28.0 23.7 21.5 20.8 23.5 Source: Authors’ calculations. SCENARIO 3: MEDIUM CROP 6.3.2.  full-scale implementation in Year 5, the total annual cost of premiums would be about US$55 million per year INSURANCE UPTAKE AND LOW (Table 6.6). Whereas the costs of GoPunjab premium PREMIUM RATES subsidies in Year 5 are US$68 million under Scenario Under Scenario 3, with medium uptake, the num- 1, they decline to US$34 million under Scenario 3. The ber of insured farmers and the insured area over total costs of premium subsidy support over five years five years would be exactly half of the estimates would be US$110.9 million (Table 6.6). under the high-uptake Scenarios 1 and 2, for a total of about 2.5 million farmers insured in the Kharif and Rabi seasons by Year 5. This number SCENARIO 4: MEDIUM CROP 6.3.3.  of insured farmers is still very large, bearing in mind INSURANCE UPTAKE AND that CLIS, which is compulsory for borrowing farmers, currently lends to only about 1 million farmers per year HIGHER AVERAGE PREMIUM throughout Pakistan. RATES The only difference under Scenario 4 relative to Under Scenario 3, then, the sum insured and Scenario 1 is that higher average costs of crop premium costs would be reduced by 50 percent, insurance premiums apply: 7.5 percent for as would the costs of the premium subsidy. At Kharif crops and 5.0 percent for Rabi crops. Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 81 TABLE 6.6:  SCENARIO 3: FIVE-YEAR CROP INSURANCE PORTFOLIO PROJECTIONS FOR MEDIUM UPTAKE AND LOW AVERAGE PREMIUM COSTS Program / Item FY2018/19 FY2019/20 FY2020/21 FY2021/22 FY 2022/23 Total Number of Insured Farmers: Program 1. Crop AYII for Commercial Farmers >2.5 Acres 300,000 500,000 625,000 712,500 750,000 2,887,500 Program 2. Crop AYII Catastrophe Cover Small Farmers <2.5 Ac. 375,000 875,000 1,375,000 1,750,000 4,375,000 Program 3. Crop Insurance for Tree Fruit (Mango, Citrus) 2,500 5,000 7,500 5,000 20,000 Total Insured Farmers 300,000 877,500 1,505,000 2,095,000 2,505,000 7,282,500 Insured Area (Acres) Program 1. Crop AYII for Commercial Farmers >2.5 Acres 1,050,000 1,750,000 2,187,500 2,493,750 2,625,000 10,106,250 Program 2. Crop AYII Catastrophe Cover Small Farmers <2.5 Ac. 375,000 875,000 1,375,000 1,750,000 4,375,000 Program 3. Crop Insurance for Tree Fruit (Mango, Citrus) 6,250 12,500 18,750 12,500 50,000 Total Insured Area (Acres) 1,050,000 2,131,250 3,075,000 3,887,500 4,387,500 14,531,250 Sum Insured (US$) Program 1. Crop AYII for Commercial Farmers >2.5 Acres 358,750,000 603,750,000 761,250,000 870,625,000 918,750,000 3,513,125,000 Program 2. Crop AYII Catastrophe Cover Small Farmers <2.5 Ac. 62,500,000 150,000,000 237,500,000 306,250,000 756,250,000 Program 3. Crop Insurance for Tree Fruit (Mango, Citrus) 3,125,000 6,250,000 9,375,000 12,500,000 31,250,000 Total Sum Insured (US$) 358,750,000 669,375,000 917,500,000 1,117,500,000 1,237,500,000 4,300,625,000 Premium Income (US$) Program 1. Crop AYII for Commercial Farmers >2.5 Acres 15,181,250 25,856,250 32,943,750 37,821,875 40,031,250 151,834,375 Program 2. Crop AYII Catastrophe Cover Small Farmers <2.5 Ac. 2,562,500 6,375,000 10,187,500 13,343,750 32,468,750 Program 3. Crop Insurance for Tree Fruit (Mango, Citrus) 312,500 625,000 937,500 1,250,000 3,125,000 Total Premium Income (US$) 15,181,250 28,731,250 39,943,750 48,946,875 54,625,000 187,428,125 Program / Item FY 2018/19 FY 2019/20 FY 2020/21 FY 2021/22 FY 2022/23 Total Premium Subsidies (US$) Program 1. Crop AYII for Commercial Farmers >2.5 Acres 7,590,625 12,928,125 16,471,875 18,910,938 20,015,625 75,917,188 Program 2. Crop AYII Catastrophe Cover Small Farmers <2.5 Ac. 2,562,500 6,375,000 10,187,500 13,343,750 32,468,750 Program 3. Crop Insurance for Tree Fruit (Mango, Citrus) 156,250 312,500 468,750 625,000 1,562,500 Sub-Total Premium Subsidies 7,590,625 15,646,875 23,159,375 29,567,188 33,984,375 109,948,438 Other Financial Costs borne by Government (US$) Data strenthening for Crop Insurance 1,500,000 1,000,000 750,000 500,000 500,000 4,250,000 Strenthen Crop Cutting Experiments (mobile phone system) 300,000 1,200,000 1,500,000 1,800,000 2,100,000 6,900,000 Farmer insurance awareneness, education and training 1,200,000 1,000,000 1,000,000 1,000,000 1,000,000 5,200,000 Monitoring and Evaluation 250,000 400,000 400,000 400,000 1,000,000 2,450,000 Sub-Total Other costs 3,250,000 3,200,000 3,250,000 3,300,000 4,600,000 18,800,000 Total Budgeted Costs to Government of Punjab 10,840,625 18,846,875 26,409,375 32,867,188 38,584,375 128,748,438 Cost per insured farmer 36.1 21.5 17.5 15.7 15.4 17.7 Source: Authors’ calculations. Under this Scenario, the total annual premium US$275  million (Table 6.7). The costs to GoPunjab of income would be about US$80 million at full- premium subsidies would rise to nearly US$48 million at scale implementation (Year 5). Over the five Year 5, and over the five years the total cost of premium years, the total estimated premiums would be about subsidies would be about US$155 million (Table 6.7). 82 A Feasibility Study TABLE 6.7:  SCENARIO 4: FIVE-YEAR CROP INSURANCE PORTFOLIO PROJECTIONS FOR LOW UPTAKE AND HIGHER PREMIUM COSTS Program / Item FY2018/19 FY2019/20 FY2020/21 FY2021/22 FY 2022/23 Total Number of Insured Farmers: Program 1. Crop AYII for Commercial Farmers >2.5 Acres 300,000 500,000 625,000 712,500 750,000 2,887,500 Program 2. Crop AYII Catastrophe Cover Small Farmers <2.5 Ac. 375,000 875,000 1,375,000 1,750,000 4,375,000 Program 3. Crop Insurance for Tree Fruit (Mango, Citrus) 2,500 5,000 7,500 5,000 20,000 Total Insured Farmers 300,000 877,500 1,505,000 2,095,000 2,505,000 7,282,500 Insured Area (Acres) Program 1. Crop AYII for Commercial Farmers >2.5 Acres 1,050,000 1,750,000 2,187,500 2,493,750 2,625,000 10,106,250 Program 2. Crop AYII Catastrophe Cover Small Farmers <2.5 Ac. 375,000 875,000 1,375,000 1,750,000 4,375,000 Program 3. Crop Insurance for Tree Fruit (Mango, Citrus) 6,250 12,500 18,750 12,500 50,000 Total Insured Area (Acres) 1,050,000 2,131,250 3,075,000 3,887,500 4,387,500 14,531,250 Sum Insured (US$) Program 1. Crop AYII for Commercial Farmers >2.5 Acres 358,750,000 603,750,000 761,250,000 870,625,000 918,750,000 3,513,125,000 Program 2. Crop AYII Catastrophe Cover Small Farmers <2.5 Ac. 62,500,000 150,000,000 237,500,000 306,250,000 756,250,000 Program 3. Crop Insurance for Tree Fruit (Mango, Citrus) 3,125,000 6,250,000 9,375,000 12,500,000 31,250,000 Total Sum Insured (US$) 358,750,000 669,375,000 917,500,000 1,117,500,000 1,237,500,000 4,300,625,000 Premium Income (US$) Program 1. Crop AYII for Commercial Farmers >2.5 Acres 22,312,500 38,062,500 48,562,500 55,781,250 59,062,500 223,781,250 Program 2. Crop AYII Catastrophe Cover Small Farmers <2.5 Ac. 3,750,000 9,375,000 15,000,000 19,687,500 47,812,500 Program 3. Crop Insurance for Tree Fruit (Mango, Citrus) 312,500 625,000 937,500 1,250,000 3,125,000 Total Premium Income (US$) 22,312,500 42,125,000 58,562,500 71,718,750 80,000,000 274,718,750 Program / Item FY 2018/19 FY 2019/20 FY 2020/21 FY 2021/22 FY 2022/23 Total Premium Subsidies (US$) Program 1. Crop AYII for Commercial Farmers >2.5 Acres 11,156,250 19,031,250 24,281,250 27,890,625 29,531,250 111,890,625 Program 2. Crop AYII Catastrophe Cover Small Farmers <2.5 Ac. 3,750,000 9,375,000 15,000,000 19,687,500 47,812,500 Program 3. Crop Insurance for Tree Fruit (Mango, Citrus) 156,250 312,500 468,750 625,000 1,562,500 Sub-Total Premium Subsidies 11,156,250 22,937,500 33,968,750 43,359,375 49,843,750 161,265,625 Other Financial Costs borne by Government (US$) Data strenthening for Crop Insurance 1,500,000 1,000,000 750,000 500,000 500,000 4,250,000 Strenthen Crop Cutting Experiments (mobile phone system) 300,000 1,200,000 1,500,000 1,800,000 2,100,000 6,900,000 Farmer insurance awareneness, education and training 1,200,000 1,000,000 1,000,000 1,000,000 1,000,000 5,200,000 Monitoring and Evaluation 250,000 400,000 400,000 400,000 1,000,000 2,450,000 Sub-Total Other costs 3,250,000 3,200,000 3,250,000 3,300,000 4,600,000 18,800,000 Total Budgeted Costs to Government of Punjab 14,406,250 26,137,500 37,218,750 46,659,375 54,443,750 180,065,625 Cost per insured farmer 48.0 29.8 24.7 22.3 21.7 24.7 Source: Authors’ calculations. Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 83 CHAPTER 7 LAUNCH OF PUNJAB AGRICULTURAL CROP INSURANCE PROGRAM IN KHARIF SEASON 2018: PLANNING CONSIDERATIONS AND IMPLEMENTATION CHALLENGES This final chapter, which originally detailed the steps involved in prepar- ing to launch a pilot AYII program in Kharif 2018, has been updated with new information on issues encountered in implementing the pilot through May 2018. The first section, prepared in July 2017, describes activities undertaken in the second half of 2017 and first quarter of 2018 to prepare the way for a pilot of AYII Program 1 (for semicommercial/progressive farmers) in Kharif 2018 in selected districts and tehsils of Punjab. The second section provides an update on the pilot pro- gram and highlights key issues and challenges experienced up to May 2018. STEPS AND TIMETABLE FOR LAUNCHING 7.1.  A CROP INSURANCE PILOT IN KHARIF 2018 The provisional work plan and timetable detailed here was presented for the consideration of GoPunjab in planning and designing a pilot of AYII Program 1 for a proposed launch date in Kharif 2018. GoPunjab had origi- nally hoped to launch a pilot in Rabi 2017/18 for wheat, but with planting occurring between late September and the end of December, a policy inception date of Sep- tember 2017 would have been required. This timeframe was considered inadequate to conduct all of the implementation planning and design tasks, so the World Bank team recommend deferring the pilot launch to Kharif 2018. The cover inception date would be March 1, 2018, to coincide with the start of the sugarcane growing season, followed by the sowing of cotton in April–May 2018 and finally the planting of rice and maize in June–August 2018. This schedule would leave a window of six months (from September 1, 2017, through the end of February 2018) to design and rate the AYII product, plan all operating systems and procedures, and put them into place. This timeframe was still very tight and required all stakeholders to complete the tasks and activities allocated to them on time. Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 85 The detailed work plan and timetable leading up The Steering Committee is chaired by the Chairman to the proposed launch of the pilot on March 1, P&D Board, and the Secretary DoA figures as the Secre- 2018, is presented at the end of this chapter in tary General. The Steering Committee comprises senior Table 7.1. The 19 major activities outlined in the work decision makers for the private insurance sector, the insur- plan are discussed next, concluding with an update on ance regulator, SBP, financial lending institutions, and implementation progress for each activity as at May 2018. key public sector organizations in Punjab, including the Ministry of Finance, Meteorological Agency, Irrigation (1) Preparation of diagnostic report (lead entity: World Department, Punjab Information Technology Board, Bank) National Agricultural Research Centre, and PDMA, among others. The Steering Committee is responsible The draft feasibility study was submitted to for overall policy and planning and implementation and GoPunjab in July 2017. Based on the results, GoPun- financial decisions on the large-scale crop and livestock jab decided to launch a pilot crop AYII program for semi- insurance programs. commercial/progressive farmers in Kharif 2018. The Technical Design and Implementation Com- (2) Approval of five-year plan by GoPunjab and alloca- mittee, under the leadership of the Secretary, DoA– tion of financial resources GoPunjab, consists of representatives from CRS and For the Kharif 2018 pilot, GoPunjab approved a Agriculture Extension Division–GoPunjab, Agriculture budget of PKR 170 million for two kharif crops Credit Unit–SBP, participating insurance companies (cotton and rice) in four selected districts.52 A and financial (lending) institutions, and the Punjab Infor- sum of PKR 100 million was allocated to premium subsi- mation Technology Board. The Technical Design and dies and PKR 70 million to implementation support and Implementation Committee is responsible for planning operating costs, including the design of public awareness all operating systems and procedures and for implement- programs for farmers, media campaigns, and farmer ing the program, including the design and rating of crop training programs. The FY2019 Annual Development insurance products. Plan contains an allocation of PKR 1,000 million by the Planning and Development Department of the Govern- (4) Setting up a Technical Support Unit ment of Punjab to cover the following three seasons (Rabi 2018/19, Kharif 2019, Rabi 2019/20). The Punjab crop Due to the highly specialized, technical nature of insurance team prepared a five-year crop insurance busi- crop insurance, and also given the high level of ness plan and budget, which is now with the provincial fiscal support (or cost) on the part of GoPunjab, parliament for approval. the World Bank team recommended that GoPun- jab should establish a Technical Support Unit (TSU). The team recommended that the TSU be man- (3) Engagement with public and private stakeholders to dated to provide objective technical analysis and over- support the Punjab Agricultural Insurance Initiative sight of the crop insurance market to GoPunjab, offer and formation of a Steering Committee and Tech- objective actuarial analysis of crop insurance product nical Implementation Committee proposals (thus acting like a quality control unit), ensure The 2017 feasibility study recommended that that farmers are receiving appropriate, good value prod- GoPunjab engage with key stakeholders, includ- ucts, develop technical standards, test and disseminate ing the Department of Agriculture (DoA), private market innovations, and manage scaled-up implementa- sector insurers, the insurance regulator, and the tion of successful innovations. lending institutions to secure their agreement to participate in a PPP for the proposed five-year In October 2017, GoPunjab established the Pun- crop and livestock insurance initiative. In October jab Crop Insurance Team, headed by the director 2017, a Steering Committee and a Technical Design and CRS–DoA, GoPunjab, to serve as the TSU. Three Implementation Committee were established and have Crop Insurance Implementation Teams (CIITs) were met subsequently as needed to guide policy and planning established under the supervision of the Project Director, and offer technical guidance, respectively, to launch the CRS: a Crop Insurance Policy Team, Crop Insurance Kharif 2018 pilot AYII program. Operation Team, and Crop Insurance Regional Team. These teams work closely with the implementing agen- cies (financial sector, insurers, crop cutting and extension 52These districts are Sheikhupura, Lodhran, R.Y. Khan, and Sahiwal. services, and others). 86 A Feasibility Study (5) Processing historical crop yield data at the teh- to underwrite the program. For the main Rabi sil or union council level (lead entity: CRS-DoA, 2018/19 launch, GoPunjab plans to issue a new tender. GoPunjab) This report recommends that CRS conduct a (9) Insurance and reinsurance planning and finalization major exercise to review its historical crop area, of coverage levels and commercial premium rates production, and crop cut yield data (officially The projections in Chapter 6 for FY2018/19 for published at the district level each season) and number of farmers and sum insured under a rework the data at the tehsil or preferably union large-scale crop insurance program differ from council level for all 36 districts. The analysis would the number of farmers and sum insured antic- be conducted for the five major crops for which CRS has ipated under the pilot. The Chapter 6 projections 10 years or more of CCE data: Rabi wheat and Kharif indicated that insurers would need to place coverage cotton, rice, maize, and sugarcane. It recommended that for up to 600,000 semicommercial/progressive farmers CRS provide guidance on (1) which departments have (250,000 in Kharif 2018 and 350,000 in Rabi 2018/19), the most comprehensive historical yield data for a min- for a TSI over both seasons of US$717 million. Under imum of the last 10 years (and hopefully 15 years) and the much smaller Kharif 2018 pilot approved by GoPun- where the program should be launched in Kharif 2018, jab, however, the FY2018 budget for premium subsidies and (2) whether the density of CCEs that CRS uses to may enable 30,000–50,000 farmers to be insured under estimate the actual average yield at the departmental the crop AYII program for loanee farmers (ground-up level is adequate to support establishing UAIs at the tehsil and top-up cover options). The appointed insurer has level or possibly the lower union council level. confirmed its terms and conditions and premium rates; it has also confirmed that reinsurance protection is in place As noted, CRS selected four districts for the Kharif for Kharif 2018. 2018 pilot: Lodhran, Rahim Yar Khan, Sahiwal, and Sheikhupura. In each district the UAI has been defined (10) Agree distribution channels through the lending at the tehsil level for 2018, and CRS has provided the institutions (lead entities: insurers and banks) appointed insurer with 12 years of historical crop yield The feasibility study recommended establish- data ­ (2005–16) for the selected tehsils. ing a mandatory linkage between the AYII pro- gram for semicommercial/progressive farmers (6) Crop area-yield product design and rating in Punjab and seasonal crop loans offered under The TSU used a tender process to select and appoint various programs, including top-up coverage one insurance company to underwrite the Kharif for the CLIS and coverage for the GoPunjab Kis- 2018 pilot AYII program. The appointed insurer sub- san loan scheme. As noted in Chapter 5, wherever mitted its winning bid and premium rates to GoPunjab possible, the AYII coverage should be implemented in February 2018. The World Bank has assisted the TSU-­ though the financial institutions to reduce the adminis- GoPunjab by developing an Excel-based AYII Crop Insur- trative costs of marketing and issuing policies, charging ance Contract Design and Training Tool and has provided and collecting premiums, and settling the payment of initial training for GoPunjab and interested insurance com- claims through the banks. The Kharif 2018 pilot offers panies (including the appointed insurer) for Kharif 2018. ground-up coverage (80–0 percent of expected yield) on an automatic basis for farmers who are beneficiaries (7) Product approval of the Kissan e-credit program and who are borrowing through the MFIs. The pilot also offers top-up cover- GoPunjab and the appointed insurer have entered age (80–50 percent of expected yield) for loanee farmers into a formal services agreement for provision insured under CLIS. of AYII crop insurance in the Kharif 2018 sea- son. The World Bank prepared a draft crop AYII Pol- icy Wording which was shared with GoPunjab and the (11) Design operating systems and procedures (farmer appointed insurer in May 2018. enrolment, policy issuance, premium collection, and settlement of claims) (8) Insurance planning To launch the pilot, the insurers, in conjunc- tion with the lending institutions, needed to put For the Kharif 2018 pilot, GoPunjab used com- in place all operating systems and procedures petitive bidding to select one insurance company required for underwriting and settling claims Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 87 on the AYII Program 1 for semicommercial/ selected districts on March 1, 2018. The AYII pol- progressive farmers. Key tasks included agreement icy is designed to cover farmers in each UAI (in this case, on procedures for issuing farmers who apply for crop tehsil) during the Kharif season from the time of plant- loans with an individual insurance policy, arriving at the ing to completion of the harvest and determination of agreed coverage levels and premium rates that apply, the actual area yield. and determining procedures for collecting premiums, for reimbursing them on loan expiry, and for settling claims (14) Design and implement strengthened sample where they are due. The parties also needed to decide area-yield measurement based on crop-cutting on a system whereby the banks submit premium bor- experiments dereau53 on a weekly or monthly basis to the insurers so The CRS Regional and Operations Teams are that they can track the number of farmers who have been currently planning for increased CCE coverage issued with AYII coverage and for which crops, as well as at the cotton and rice harvests in the UAIs in the their insured acreage, sum insured, and premium due. four pilot districts. The recommendations for CRS Responding to these needs, the GoPunjab Crop Insur- in Chapter 5 were to (1) increase the number of villages ance Team invested heavily in the design of a web-based sampled and number of CCEs over time to facilitate crop insurance portal to enable the financial institutions implementation of the AYII program at the tehsil or (lending banks) and insurance company(ies) to input all union council levels; (2) simplify and speed up procedures crop loans and crop insurance-related information on a by introducing grain moisture meters to determine wet routine or daily basis. and dry grain weights after one field visit; and (3) intro- duce smartphone or tablet SMS technology to record (12) Crop insurance marketing and sales and farmer and transmit CCE results from the field at the time of awareness creation and education the crop cuts. For the Kharif 2018 Pilot, GoPunjab agreed that ground-up coverage will be compulsory for all (15) Train CRS field extension staff in new CCE e-Kissan crop credit recipients, and that for CLIS procedures farmers, top-up coverage will be offered on a vol- The feasibility study’s recommendation is to untary basis with the sales window closing one start training field extension staff in November month before the expected crop harvest.54 Given 2017 and for training to run into 2018, up to the the proposal to link crop AYII coverage on a mandatory time the Kharif harvest begins. It is important that basis with crop loans, insurers will not need to market and all staff receive equipment, are trained to use it, and learn promote coverage to individual farmers, yet it is still very the procedures for electronic data entry and transmis- important that the insurers agree with the banks on the sion. This activity is ongoing as of June 2018. The New approach to create farmer awareness and educate them Crop Insurance Portal has a module for uploading CCE about the crop insurance coverage for which they are data that is transmitted electronically by smartphones or signing up as part of their loan package. GoPunjab has tablets. invested heavily in designing and implementing farmer awareness and training programs, including media and materials. (16) Implement end of 2018 Kharif season CCEs The timing of CCEs is critical. Kharif crops are (13) Initiate crop insurance coverage for Kharif crops, harvested mainly between September and the end of starting with sugarcane in March–April 2018 November, with the timing of the harvest varying by crop, district, and the prevailing climatic conditions The Kharif 2018 AYII program was launched during each season. CCEs must be completed in a timely for cotton and rice loanee farmers in the four fashion so that claim payments (where they are due) are also made as quickly as possible after the harvest. 53A premium bordereau contains a detailed list of policies insured under an insurance contract during the reporting period, reflecting such information (17) Settle claim payouts as the name and address of the (primary) insureds, the amount and location of the risk, the effective and termination dates of the insurance, the amount The insurers will need to agree in advance with insured, and the insurance premium applicable. 54The World Bank has advised that the convention for multiple peril crop the lending institutions on the procedures for set- insurance policies is to close sales at the time the crop is planted. It is tling claims. It is likely that the banks will require that recommended that GoPunjab reconsider the cut-off dates for selling policies claim payments are settled through the bank to enable prior to the main launch in Rabi 2018–19 season. 88 A Feasibility Study the bank to recover the loan, plus interest and premium IMPORTANCE OF INVESTMENT 7.2.2.  payments, and to then transfer any remaining money IN A WEB-BASED CROP from the claim to the insured farmer. INSURANCE PORTAL GoPunjab has invested significantly in the design (18) Design and implement a monitoring and evaluation of a web-based crop insurance portal to enroll system farmers in the crop insurance program and facilitate easy The design of an M&E system started in Novem- communication and rapid transfer of data between finan- ber 2017, with a view to launching the system in the cial institutions, insurance companies, and GoPunjab. 2018 Kharif season. The portal also enables farmers to register online for crop insurance. The system is designed to ensure maximum (19) World Bank technical support to the Punjab Agri- transparency. For example, access to CCE yield results is cultural Insurance Programs essential for insurers and reinsurers, and the system per- The World Bank fielded a number of missions mits CCE yield results to be uploaded in real-time as they to provide technical support to GoPunjab, the become available in the field. The crop insurance portal proposed TSU, and the insurance sector in planning came on line in May 2018 and will greatly assist the day- and designing the AYII program for semicommercial/ to-day management and implementation of the Punjab progressive farmers. As of June 2018, technical support crop insurance project, especially if it is scaled up as missions had visited Punjab in October/November 2017, planned, to underwrite several million policies by Year 5. February 2018, and May 2018. THE NEED FOR TRAINING 7.2.3.  IN AREA-YIELD INDEX INSURANCE 7.2. KEY LESSONS CONTRACT DESIGN AND RATING AND CHALLENGES AYII is a new crop insurance product in Paki- EXPERIENCED IN stan, and insurance companies lack detailed knowledge and expertise in designing and rat- ROLLING OUT CROP AYII ing this product. Insurers are not familiar with the concepts of crop loss of yield insurance and the need to IN KHARIF 2018 detrend time series crop yield data before establishing the The key lessons and challenges encountered as average or expected yield and setting insured yields and Crop AYII has been rolled out in Punjab in Kharif (where applicable) exit yields. Aside from these issues of 2018 reflect a number of technical and logistical AYII contract design, local insurers have no experience concerns, as well as stakeholders’ perceptions. of methods for rating (pricing) AYII products. For the The sections that follow summarize the major lessons and Kharif 2018 pilot, it appears that most tendering insur- challenges arising to date. ers based their pricing decisions on the rates charged on CLIS (maximum 2 percent rate) rather than on system- atic loss of yield rating analysis for the AYII coverage. THE IMPORTANT COORDINATING 7.2.1.  ROLE PERFORMED BY THE TSU The lack of knowledge on designing and rating Early formation of the Punjab Crop Insurance AYII contracts in the local insurance market Team as the TSU for the insurance program was complicated the tender and bidding processes for key to the successful launch of the AYII pilot in the Kharif 2018 pilot. This included specific problems Kharif 2018. This TSU has been responsible for coor- related to understanding the differences between top-up dinating the planning, design, and implementation of the AYII coverage for farmers who are already insured under Punjab crop insurance program from the very beginning, CLIS, and ground-up coverage for farmers who are not starting with the Kharif 2018 pilot. Specialized staff were insured under CLIS. This insufficient understanding was recruited to complement the team, including a project reflected in the commercial premium rates of the win- manager and a local crop insurance specialist, who for- ning bid, which did not adequately reflect the underlying merly headed the SBP CLIS initiative. The World Bank risk exposures, particularly for cotton in several UAIs in has channeled its technical assistance for the design of the pilot districts. It seems that insurers submitted bids the AYII product and program through the TSU. without clearly establishing the expected yields and Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 89 insured yields that would apply in Kharif 2018 for each farmer insurance certificates that should be easily insured crop in each UAI. Furthermore, the original ten- understood by farmers, many of whom have no insurance der was issued with a 300 percent loss ratio cap, which knowledge or literacy. Under the CLIS program, insur- subsequently had to be withdrawn, and the tender was ers and financial institutions enter into insurance agree- reissued. ments, but no form of insurance wording or insurance certificate is issued to the individual farmers protected To address these issues, the World Bank team under this scheme. In Punjab, the World Bank team developed an Area-Yield Index Insurance Con- strongly recommended to the TSU and the appointed tract Design and Rating Tool, as well as a series insurer that (1) each financial institution should be issued of modular training programs aimed at key pub- with a Master Policy Wording and Schedule for the new lic and private sector stakeholders in Punjab. AYII product and (2) that each insured farmer should The Crop AYII Contract Design and Rating Tool was receive a Certificate of Insurance. During the May 2018 delivered to the Punjab Crop Insurance Team at the time mission, the World Bank team shared specimen copies of of the February 2018 World Bank mission. Since the ten- an AYII policy wording and farmer certificate with the der process was under way, the Punjab Crop Insurance TSU and the appointed insurer. Team decided it was not appropriate to share the tool with the insurance companies at that time. Some pre- liminary training in the use of the tool was provided by 7.2.5. FINANCIAL INSTITUTIONS’ VIEW the World Bank team to the insurance companies during ON COMPULSORY VERSUS its May 2018 mission. The current tool enables the user VOLUNTARY CROP INSURANCE to design an AYII contract for the selected Kharif 2018 For the Kharif 2018 pilot, GoPunjab decided that pilot crops and tehsils/districts under either Option 1: any small farmer who wished to access e-Kis- Ground-up coverage using the agreed terms for 2018, san crop loans (e-credit) will be insured on a i.e., fixed (for all crops/tehsils) threshold or trigger yield compulsory basis. This principle has been agreed of 80 percent of expected yield and exit yield of 0 per- by the financial institutions for e-credit farmers who are cent of expected yield; and Option 2: Top-up coverage not already insured under CLIS. However, for farmers for farmers insured under CLIS with fixed threshold or already insured under CLIS (which is free to the farmer), trigger yield of 80 percent of expected yield and exit the financial institutions were very reluctant to impel the yield of 50 percent of expected yield. The tool can gen- farmers to purchase top-up AYII coverage on a compul- erate expected average loss costs (pure rates) using actual sory basis and then require them to pay 50 percent of the historical yield data provided by CRS for each crop and cost of the top-up coverage premium. Therefore, under tehsil or detrended yields. The tool also allows users to the Kharif 2018 pilot it has been agreed that top-up AYII generate technical rates and indicative commercial pre- coverage will be offered on a voluntary basis to farmers mium rates using their own assumptions. It is stressed who are already insured under CLIS. The sale of vol- that the tool is a training tool and that all final pricing untary crop insurance will require the identification of decisions rest with the insurers and their reinsurers. In suitable distribution channels to market and promote August 2018, the World Bank plans to deliver to GoPun- voluntary coverage, as well as the design of mechanisms jab an updated Crop AYII Contract Design and Rating to collect premiums from farmers. This decision will be Tool (version 2), which incorporates a more sophisticated reviewed prior to the Rabi 2018–19 launch of the crop rating methodology that conforms to best practices used insurance program in Punjab. by international crop reinsurers. THE NEED TO INCLUDE ALL TYPES 7.2.6.  THE NEED TO ASSIST THE 7.2.4.  OF FARMER IN THE PUNJAB CROP PAKISTAN INSURANCE MARKET INSURANCE INITIATIVE TO DEVELOP CROP INSURANCE The Kharif 2018 pilot specifically targets small POLICY WORDINGS cotton and rice farmers who are e-credit recipi- Under this commercial crop insurance initia- ents. GoPunjab would like to see the program rapidly tive, it is recommended that insurers observe start to develop crop insurance products that meet the international norms by designing tailor-made needs of medium and large farmers as well, however, crop insurance policy wordings and individual and to provide incentives for those farmers to purchase 90 A Feasibility Study crop insurance. Although some larger farmers may want convey the concept to farmers. At the same time, results to purchase AYII coverage, the crop insurance build-up of the feasibility study had clearly shown that yields of plan identifies individual farmer MPCI and NPCI prod- cotton in Lodhran District exhibit extreme variability ucts that are designed for larger farmers and planned to year-on-year and that an actuarially rated AYII coverage become available in 2019–20. with 80 percent insured yield would be very expensive in those tehsils. Conversely, the 12-year average yield data provided by CRS show that yields of kharif paddy rice 7.2.7. CONCERNS ABOUT CAPPING are extremely stable. With few exceptions, actual annual INSURED YIELDS AT 80 PERCENT yields have never fallen short of 90 percent of average yields at the district and tehsil level. During the launch of OF EXPECTED YIELD the Kharif pilot program, farmers reported to the TSU For the Kharif 2018 pilot, the AYII insurers and and the insurers that they did not believe it was fair to their reinsurers expressed a desire to cap the offer AYII for paddy with an 80 percent insured yield, maximum insured yield (threshold yield) at because they believe that there will never be a payout. For 80 percent of the expected yield. Their concern was that reason, for the main Rabi 2018/19 launch, it is rec- that as the program was new and untested in Punjab, they ommended that the insured yield for wheat in each tehsil wish to avoid over-insuring average area yields and would (UAI) should be closely related to the historical yield vari- not want to offer more than an 80 percent insurance ability and that insurers should be requested to consider a yield coverage level in the start-up phase. The GoPunjab higher maximum insured yield of 90 percent of expected crop insurance team therefore elected to adopt a fixed yield in UAIs with low yield variability. 80 percent insured yield for both cotton and rice in all tehsils for the 2018 pilot, because it would be simple to Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 91 AREA-YIELD INDEX INSURANCE PROGRAM 1 (FOR SEMICOMMERCIAL/PROGRESSIVE FARMERS LINKED 92 TABLE 7.1:  TO CROP CREDIT): IMPLEMENTATION WORK PLAN AND TIMETABLE LEADING UP TO LAUNCH IN KHARIF SEASON 2018 2017 2018 Key Tasking Lead Entity Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec 1. Prepare Diagnostic Report WBG 2. Approval by Govt. Punjab & Allocation of GoPunjab Financial Resources 3. Agreement by Public and Private stakeholders to support the PPP agricultural insurance initiative and form Steering Committee and Technical committees 4. Establishment of Technical Support Unit (TSU) 5. Historical Crop Area Yield Data Provision at Crop Reporting Tehsil / Union Council level Services (CRS) 6. Crop Insurance Product Design & Rating WBG / Insurers 7. Product Approval (Insurance Regulator) Insurers / Regulator 8. Insurance Planning (Consortium, Pool, Special Purpose Vehicle) Insurers / GoPunjab 9. Insurance & Reinsurance Planning and approval of coverage levels and premium rates Insurers & reinsurers 10. Agree Distribution Channels through Banks Insurers / Banks & (Linkage with KISAN Credit Scheme) MFIs 11. Design Operating Systems & Procedures (farmer enrolment, policy issuance, premium Insurers / Banks & collection) MFIs Kisan Banks / 12. Crop Insurance Policy Sales and Farmer Insurers / Crop Sugar Cane / Cotton Rice Maize Eduction Extension Dept 13. Incept Crop Insurance Cover for Kharif Season Sugar Cotton Rice Maize 2018 starting in April with Sugar Cane & Cotton Cane Insurers 14. Design and Test Strengthened Sample Yield WBG / CRS Measurement (Crop Cutting Experiements) 15. Train CRS Extension staff in new CCE Procedures 16. Implement end of Kharif 2018 season CCE's CRS 17. Settle Claims Payouts Insurers 18. Design Monitoring & Evaluation Systems and WBG / GoPunjab / Design Implement M&E Implement Insurers 19. WBG Technical Support Missions A Feasibility Study BIBLIOGRAPHY ADB (Asian Development Bank) and World Bank. 2010. “Pakistan Floods 2010: Pre- liminary Damage and Needs Assessment.” Islamabad, November. Agroasemex. 2006. La Experiencia Mexicana en el Desarrollo y Operación de Seguros Paramétricos Orientados a la Agricultura. Agroseguro. 2012. Informe Annual 2012. Ali, S., Y. Lui, M. Ishaq, T. Shah, Abdullah, A. Ilyas, and I.U. Din. 2017. “Climate Change and Its Impact on the Yield of Major Food Crops: Evidence from Paki- stan.” Foods 6 (6): 39, doi:10.3390/floods6060039 Antón, J. and S. Kimura. 2011. “Risk Management in Agriculture in Spain,” OECD Food, Agriculture and Fisheries Papers, No. 43, OECD Publishing, Paris. https://doi .org/10.1787/5kgj0d57w0wd-en Arias, D., A. de la Fuente, J. Escamilla, and C. Stutley. 2014. Insuring Resilience: Mex- ico Case Study. The World Bank, January 2014. AXCO. 2017. Pakistan Non-Life (P&C) Report. Bora, B. 2010. Subsidised Agricultural Insurance in Turkey (TARSIM). Available from: http://www.mapfre.com/mapfrere/docs/html/revistas/trebol/n54/docs/ Articulo2en.pdf Clarke, D. J., O. Mahul, K. N. Rao, and N. Verma. 2012. Weather-based Crop Insur- ance in India. Policy Research Working paper 5985. The World Bank, Washing- ton, DC. De Janvry, Alain. 2015. Quantifying through Ex Post Assessments the Micro-Level Impacts of Sovereign Disaster Risk Financing and Insurance Programs. Policy Research Working Paper 7356.World Bank, Washington, DC. De Janvry, A., E. Ramirez Ritchie, and E. Sadoulet. 2016. Weather Index Insurance and Shock Coping: Evidence from Mexico’s CADENA Program. World Bank Pol- icy Research Working Paper 7715. Washington, DC.: World Bank. FAO. 2010. Climate-Smart Agriculture: Policies Practices and Financing for Food Security, Adaptation and Mitigation. Food and Agriculture Organisation of the United Nations (FAO). 2010 Fuchs, Alan and Hendrik Wolff. 2010. Concept and unintended consequences of weather index insurance: the case of Mexico, Working Paper, 2010. Fuchs, A., and H. Wolff. 2011. Concept and Unintended Consequences of Weather Index Insurance: The Case of Mexico. Discussion Paper No. 6234 December 2011. https://d-nb.info/1020196564/34 GFDRR. 2011. National Agricultural Insurance Scheme in India. Making Insurance Markets Work for Farmers. www.gfdrr.org.drfi GFDRR (Global Facility for Disaster Reduction and Recovery). 2015. “Fiscal Disaster Risk Assessment: Options for Consideration—Pakistan.” World Bank, Washing- ton, DC. June. Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 93 Government of Punjab. 2015. Punjab Development Sta- Development Review 51(4) Part II: 261–76. http://www tistics 2015. Lahore: Bureau of Statistics, Govern- .pide.org.pk/pdf/PDR/2012/Volume4/261-276. ment of Punjab. December. pdf Hazell, P., C. Pomareda, and A. Valdez. 1986. Crop Skees, J. R., J. R. Black, and B. J. Barnett. 1997. Design- Insurance for Agricultural Development. Baltimore, ing and Rating an Area Yield Crop Insurance Con- MD: Johns Hopkins University Press. tract. American Journal of Agricultural Economics. 79 IFAD and WFP (International Fund for Agricultural (May 1997): 430–438. Development and World Food Programme). 2010. State Bank of Pakistan. 2017. CLIS Crop and Livestock Potential for Scale and Sustainability in Weather Index Insur- Insurance Frameworks (SBP 2017). ance for Agriculture and Rural Livelihoods. Rome. SwissRe. 2013. Partnering for Food Security in Emerging Mahul, O., and C. Stutley. 2010. Government Support Markets. SIGMA No 1/2013. to Agricultural Insurance—Challenges and Options Varangis P., J. Skees and H. B. Price. 2003. Alternatives for Developing Countries. Main Report, plus for Financing Catastrophic Yield Risks in Romanian Annex E: International Experience with Agricultural Agriculture. Insurance: Findings from a World Bank Survey of World Bank. 2011. Enhancing Crop Insurance in India. 65 Countries. The World Bank, Washington, DC. Report No 61491-IM. India Country management NRSP (National Rural Support Programme). 2017. Unit, South Asia Finance and Private Sector Unit, “NRSP’s experience in Crop and Livestock Insur- Global Capital Market Non Banking Unit, The ance.” Islamabad. Word Bank, Washington, DC. PPAF (Pakistan Poverty Alleviation Fund). 2012. World Bank. 2013. “Mexico Agriculture Insurance Mar- “Reducing Vulnerabilities of the Poor through ket Review.” Washington, DC. http://documents. Micro-­ Insurance.” Good Practice Note. October. worldbank.org/curated/en/891631468280471285/ Islamabad. pdf/880970BRI0P1300ance04Pager0Overview.pdf PDMA (Provincial Disaster Management Authority), World Bank. 2015a. “Bangladesh: Agricultural Insurance Punjab. 2014. “Punjab Disaster Response Plan.” Solutions Appraisal Technical Report.” Washington, Lahore. DC. August. Rao, K. N. 2015. Climate Risk and Agriculture—Case ———. 2015b. “Kenya—Towards a National Crop and of Indian Crop Insurance. Keynote 3 presentation, Livestock Insurance Program: Background Report.” Climate Change and Food Security, CCAFS Work- Washington, DC. October. shop 2015. ———. 2015c. “Fiscal Disaster Risk Assessment: RESULTS UK. 2016. “Ensuring Climate Risk Insur- Options for Consideration.” Report No. 103648. ance Works for the Poor.” London. Global Facility for Disaster Reduction and Recovery Saeed, A. 2014. Corruption and Poor Design Threaten (GFDRR), World Bank, Washington, DC. Pakistan Crop Insurance Scheme. Thomas Reu- ———. 2017. Program-for-Results Concept Note on a ters Foundation News. http://news.trust.org/item/ Proposed Loan in the Amount of 300 Million USD 20140704055948-vj87s/ to the Islamic Republic of Pakistan for Strengthening SBP (State Bank of Pakistan). 2008. “SBP Task Force Markets for Agriculture and Rural Transformation Report on Crop Loan Insurance Framework.” Agri- in Punjab. World Bank, Washington, DC. cultural Credit Department, State Bank of Pakistan, Yu Winston, Yi-Chen Yang, Andre Savitsky, Donald Karachi. Alford, Casey Brown, James Wescoat, Dario Debo- ———. 2014. “Indicative Credit Limits and Eligibil- wicz, and Sherman Robinson. 2013. “The Indus ity for Agri. Financing.” Agricultural Credit and Basin of Pakistan: The Impacts of Climate Risks on Microfinance Department, State Bank of Pakistan, Water and Agriculture.” World Bank, Washington, Karachi. DC. Siddiqui, R., G. Samad, M. Nasir, and H. H. Jalil. 2012. ZTBL. 2017. Zarai Taraqiati Bank Limited (data pro- “The Impact of Climate Change on Major Agricul- vided for feasibility study). tural Crops: Evidence from Punjab, Pakistan.” Pakistan 94 A Feasibility Study ANNEX 1 AREA, PRODUCTION, AND YIELDS OF MAJOR CROPS IN PUNJAB, PAKISTAN, 2006/07–2015/16 Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 95 TABLE A1.1: PUNJAB: AREA, PRODUCTION, AND YIELDS OF MAJOR CROPS, 2006/07–2015/16 96 Average Average 10 Year 2006-07 to 2011-12 to Year 2006-07 2007-08 2008-09 2009-10 2010-11 2011-12 2012-13 2013-14 2014-15 2015-16 Average 2010-11 2015-16 Punjab : Cultivated Area by year ("000" Acres) Wheat 15,896 15,820 16,893 17,084 16,534 16,020 16,090 17,054 17,247 17,085 16,572 16,445 16,699 Gram 2,251 2,444 2,395 2,388 2,384 2,274 2,244 2,120 2,136 2,113 2,275 2,372 2,177 Cotton 6,086 5,992 5,495 6,019 5,438 6,261 5,705 5,434 5,740 5,542 5,771 5,806 5,736 Rice 4,271 4,259 4,887 4,773 4,366 4,236 4,229 4,470 4,640 4,399 4,453 4,511 4,395 Maize 1,217 1,321 1,321 1,248 1,343 1,492 1,452 1,704 1,662 1,770 1,453 1,290 1,616 Sugar cane 1,759 2,044 1,647 1,501 1,661 1,881 1,897 1,870 1,756 1,743 1,776 1,722 1,829 Total 31,480 31,880 32,637 33,013 31,726 32,164 31,617 32,651 33,181 32,651 32,300 32,147 32,453 Punjab : Crop Production ("000" Tonnes) Wheat 17,853 15,607 18,420 17,919 19,041 17,739 18,587 19,739 19,282 19,527 18,371 17,768 18,975 Gram 728 388 658 488 429 225 691 331 322 227 449 538 359 Cotton 4,968 4,350 4,200 4,104 3,769 5,342 4,572 4,389 4,932 3,045 4,367 4,278 4,456 Rice 3,076 3,286 3,643 3,713 3,384 3,277 3,478 3,481 3,648 3,502 3,449 3,420 3,477 Maize 2,162 2,694 2,627 2,502 2,959 3,442 3,353 4,021 4,020 4,391 3,217 2,589 3,845 Sugar cane 37,542 40,306 32,295 31,324 37,481 42,893 42,982 43,704 41,074 41,968 39,157 35,790 42,524 Total 66,329 66,631 61,844 60,050 67,063 72,917 73,663 75,664 73,278 72,660 69,010 64,383 73,637 Punjab: Crop Yields (Kilograms/Acre) 37 Wheat 1,123 986 1,090 1,049 1,151 1,107 1,155 1,157 1,118 1,143 1,108 1,080 1,136 Gram 324 159 275 204 180 99 308 156 151 107 196 228 164 Cotton 816 726 764 682 693 853 801 808 859 549 755 736 774 Rice 720 771 745 778 775 774 822 779 786 796 775 758 791 Maize 1,776 2,040 1,989 2,005 2,203 2,306 2,309 2,360 2,418 2,481 2,189 2,003 2,375 Sugar cane 21,343 19,719 19,608 20,869 22,565 22,803 22,658 23,371 23,390 24,074 22,040 20,821 23,259 Source: CRS-DoA, GoPunjab. A Feasibility Study TABLE A1.2: PUNJAB: AVERAGE WHEAT YIELDS (KG/ACRE) BY DISTRICT, 2006/07–2015/16 NAME 2006-07 2007-08 2008-09 2009-10 2010-11 2011-12 2012-13 2013-14 2014-15 2015-16 average stdev cov min max THE PUNJAB 1,123 987 1,090 1,049 1,152 1,107 1,155 1,157 1,118 1,143 1,108 54 4.9% 987 1,157 Attock. 821 516 690 395 656 435 427 556 701 722 592 146 24.7% 395 821 Rawalpindi. 806 649 732 360 557 499 739 568 653 676 624 131 21.0% 360 806 Islamabad. 816 616 673 340 597 491 680 560 667 668 611 128 21.0% 340 816 Jhelum. 854 657 826 587 766 562 728 660 803 910 735 116 15.8% 562 910 Cahkwal. 646 455 530 250 532 375 658 531 670 729 538 148 27.5% 250 729 Sargodha. 1,067 958 1,079 908 1,045 1,077 1,110 937 934 1,008 1,012 73 7.2% 908 1,110 Khushab. 798 642 721 779 815 683 775 670 749 798 743 60 8.1% 642 815 Mianwali. 841 742 767 849 969 892 1,014 1,013 719 824 863 108 12.5% 719 1,014 Bhakkar. 959 867 920 935 990 904 986 1,019 901 923 940 47 5.0% 867 1,019 Faisalabad. 1,255 1,062 1,183 1,150 1,280 1,231 1,263 1,275 1,295 1,272 1,227 74 6.0% 1,062 1,295 T.T. Singh. 1,190 1,175 1,217 1,134 1,392 1,262 1,331 1,340 1,337 1,367 1,275 91 7.1% 1,134 1,392 Jhang. 1,250 1,049 1,133 1,103 1,187 1,147 1,202 1,235 1,244 1,170 1,172 65 5.6% 1,049 1,250 Chiniot 1,109 1,259 1,245 1,261 1,273 1,175 1,186 1,215 61 5.0% 1,109 1,273 Gujrat. 803 656 752 586 771 714 780 723 622 686 709 71 10.1% 586 803 M.B.Din. 1,067 1,037 1,095 1,041 1,056 1,103 1,149 1,052 987 1,042 1,063 44 4.2% 987 1,149 Sialkot. 1,030 1,015 1,041 878 1,069 1,176 1,212 1,221 693 991 1,033 160 15.5% 693 1,221 Narowal. 937 824 793 743 950 894 985 1,001 713 843 868 101 11.6% 713 1,001 Gujranwala. 1,217 1,225 1,293 1,302 1,272 1,410 1,280 1,284 1,100 1,197 1,258 81 6.4% 1,100 1,410 Hafizabad. 1,131 1,162 1,165 1,036 1,288 1,250 1,231 1,249 1,238 1,242 1,199 75 6.3% 1,036 1,288 Sheikhupura. 1,076 997 1,081 1,014 1,230 1,162 1,234 1,240 1,300 1,165 1,150 103 9.0% 997 1,300 Nankana Sahib 1,206 1,148 1,201 1,071 1,213 1,265 1,255 1,257 1,347 1,306 1,227 79 6.4% 1,071 1,347 Lahore. 1,063 992 1,011 1,036 1,162 1,214 1,248 1,184 1,109 1,207 1,122 93 8.3% 992 1,248 Kasur. 1,225 1,093 1,203 1,096 1,249 1,254 1,213 1,218 1,242 1,172 1,197 59 4.9% 1,093 1,254 Okara. 1,419 1,283 1,380 1,376 1,450 1,436 1,444 1,452 1,403 1,287 1,393 63 4.5% 1,283 1,452 Sahiwal. 1,234 1,088 1,247 1,219 1,313 1,258 1,264 1,295 1,233 1,269 1,242 61 4.9% 1,088 1,313 Pakpattan. 1,369 1,258 1,384 1,461 1,525 1,360 1,414 1,437 1,362 1,384 1,396 71 5.1% 1,258 1,525 Multan. 1,153 940 1,195 1,014 1,137 1,052 1,176 1,195 1,223 1,225 1,131 97 8.6% 940 1,225 Lodhran. 1,199 1,044 1,302 1,198 1,265 1,218 1,284 1,325 1,263 1,371 1,247 90 7.2% 1,044 1,371 Khanewal. 1,276 1,085 1,260 1,221 1,250 1,188 1,256 1,283 1,276 1,302 1,240 63 5.1% 1,085 1,302 Vehari. 1,188 1,122 1,167 1,168 1,277 1,299 1,325 1,339 1,331 1,324 1,254 83 6.6% 1,122 1,339 Muzaffargarh. 1,099 960 1,096 1,099 1,170 1,114 1,135 1,138 1,130 1,167 1,111 59 5.3% 960 1,170 Layyah. 1,079 882 1,026 1,035 1,140 964 1,058 1,097 1,029 1,044 1,035 71 6.9% 882 1,140 D.G.khan. 1,086 1,028 940 1,020 1,071 1,077 1,153 1,180 1,208 1,118 1,088 81 7.4% 940 1,208 Rajanpur. 1,161 944 1,037 1,092 1,107 956 1,049 1,055 1,143 1,246 1,079 92 8.5% 944 1,246 Bahawalpur. 1,178 1,043 1,212 1,143 1,273 1,291 1,214 1,224 1,233 1,306 1,212 77 6.4% 1,043 1,306 Rahim Yar Khan 1,221 971 1,099 1,111 1,263 1,257 1,348 1,364 1,277 1,330 1,224 126 10.3% 971 1,364 Bahawalnagar. 1,225 1,037 1,157 1,212 1,264 1,190 1,207 1,228 1,273 1,235 1,203 67 5.6% 1,037 1,273 Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 97 TABLE A1.3: PUNJAB: AVERAGE COTTON YIELDS (KG/ACRE) BY DISTRICT, 2006/07–2015/16 98 NAME 2006-07 2007-08 2008-09 2009-10 2010-11 2011-12 2012-13 2013-14 2014-15 2015-16 average stdev cov min max THE PUNJAB 816 726 764 682 693 853 801 808 859 549 755 95 12.6% 549 859 Attock. Rawalpindi. Islamabad. Jhelum. 206 206 225 94 125 245 187 442 480 187 240 125 52.2% 94 480 Cahkwal. 197 187 197 461 0 208 164 78.8% 0 461 Sargodha. 324 294 421 474 409 491 466 234 348 253 371 94 25.3% 234 491 Khushab. 230 235 269 590 614 420 367 362 291 394 377 135 35.9% 230 614 Mianwali. 637 711 830 642 477 530 517 597 584 456 598 114 19.0% 456 830 Bhakkar. 563 456 640 497 426 657 689 568 644 377 552 108 19.6% 377 689 Faisalabad. 554 550 716 562 495 543 507 477 642 417 546 84 15.4% 417 716 T.T. Singh. 593 518 741 505 452 632 623 656 898 471 609 136 22.4% 452 898 Jhang. 563 585 723 421 490 561 492 494 536 405 527 91 17.3% 405 723 Chiniot 498 440 485 370 349 354 336 405 68 16.9% 336 498 Gujrat. M.B.Din. 304 299 302 246 263 291 255 205 383 224 277 50 18.2% 205 383 Sialkot. Narowal. Gujranwala. Hafizabad. Sheikhupura. 537 537 537 537 Nankana Sahib 264 225 355 282 67 23.7% 225 355 Lahore. Kasur. 381 365 492 417 422 501 482 461 474 334 433 58 13.4% 334 501 Okara. 553 582 660 551 698 832 964 853 891 577 716 156 21.8% 551 964 Sahiwal. 587 546 845 532 691 895 910 777 845 419 705 175 24.8% 419 910 Pakpattan. 758 732 895 610 786 938 839 846 854 474 773 140 18.1% 474 938 Multan. 848 781 780 694 708 941 889 847 885 420 779 149 19.1% 420 941 Lodhran. 902 803 625 723 652 895 818 798 886 467 757 141 18.6% 467 902 Khanewal. 840 749 885 623 729 981 911 946 908 445 802 167 20.9% 445 981 Vehari. 752 558 737 816 731 936 912 873 883 457 765 156 20.4% 457 936 Muzaffargarh. 681 725 539 539 793 750 721 696 866 424 673 134 19.9% 424 866 Layyah. 601 631 662 508 493 598 563 549 622 605 583 54 9.3% 493 662 D.G.khan. 947 847 701 704 626 777 646 701 838 502 729 128 17.5% 502 947 Rajanpur. 978 844 669 859 623 742 684 900 945 873 812 124 15.2% 623 978 Bahawalpur. 926 772 849 667 785 933 836 835 894 643 814 99 12.2% 643 933 Rahim Yar Khan 875 787 905 803 689 896 801 858 916 699 823 82 9.9% 689 916 Bahawalnagar. 863 739 846 752 796 957 919 924 910 717 842 86 10.3% 717 957 A Feasibility Study TABLE A1.4: PUNJAB: AVERAGE RICE YIELDS (KG/ACRE) BY DISTRICT, 2006/07–2015/16 NAME 2006-07 2007-08 2008-09 2009-10 2010-11 2011-12 2012-13 2013-14 2014-15 2015-16 average stdev cov min max THE PUNJAB 720 772 745 778 775 774 822 779 786 796 775 27 3.5% 720 822 Attock. Rawalpindi. Islamabad. Jhelum. 680 680 848 505 640 677 677 727 680 693 681 83 12.3% 505 848 Cahkwal. Sargodha. 651 789 660 680 638 678 657 761 729 643 689 53 7.7% 638 789 Khushab. 750 807 705 534 677 694 584 630 729 710 682 80 11.8% 534 807 Mianwali. 683 796 753 543 781 723 586 674 577 578 669 93 13.9% 543 796 Bhakkar. 690 830 680 620 610 597 600 673 660 700 666 69 10.4% 597 830 Faisalabad. 581 649 643 665 733 687 741 702 730 725 686 51 7.5% 581 741 T.T. Singh. 644 776 684 724 758 714 752 690 837 839 742 64 8.6% 644 839 Jhang. 672 728 716 678 704 653 647 713 719 713 694 29 4.2% 647 728 Chiniot 661 796 775 810 754 691 742 747 54 7.3% 661 810 Gujrat. 682 755 748 686 636 603 712 646 591 590 665 62 9.3% 590 755 M.B.Din. 735 821 688 742 689 755 782 741 696 714 736 43 5.8% 688 821 Sialkot. 738 787 724 837 819 776 813 705 718 734 765 47 6.2% 705 837 Narowal. 623 713 563 643 707 652 681 654 666 701 660 45 6.8% 563 713 Gujranwala. 734 799 803 895 851 885 947 803 802 863 838 62 7.4% 734 947 Hafizabad. 687 732 764 842 837 806 855 827 778 847 797 56 7.0% 687 855 Sheikhupura. 651 680 680 725 676 720 752 725 720 781 711 39 5.5% 651 781 Nankana Sahib 687 718 744 731 725 726 783 814 826 778 753 45 6.0% 687 826 Lahore. 643 749 772 738 733 708 834 691 789 763 742 53 7.2% 643 834 Kasur. 783 824 806 760 756 772 866 804 795 816 798 33 4.1% 756 866 Okara. 893 956 921 906 953 938 995 1,036 1,023 951 957 48 5.0% 893 1,036 Sahiwal. 778 729 707 724 823 700 794 682 811 735 748 49 6.6% 682 823 Pakpattan. 860 893 842 942 987 906 1,041 956 959 904 929 60 6.5% 842 1,041 Multan. 603 725 650 619 665 652 675 698 773 813 687 67 9.7% 603 813 Lodhran. 660 770 653 611 578 609 700 544 772 864 676 100 14.8% 544 864 Khanewal. 645 718 767 662 694 693 739 703 842 732 719 56 7.8% 645 842 Vehari. 720 750 747 646 734 759 805 790 833 756 754 51 6.8% 646 833 Muzaffargarh. 720 823 829 711 692 719 808 817 783 751 765 53 6.9% 692 829 Layyah. 685 806 818 669 718 589 565 694 661 725 693 81 11.7% 565 818 D.G.khan. 851 881 891 970 838 933 895 907 880 890 894 38 4.2% 838 970 Rajanpur. 626 818 808 834 783 788 780 845 695 653 763 77 10.1% 626 845 Bahawalpur. 697 580 691 592 689 646 654 534 773 820 668 87 13.0% 534 820 Rahim Yar Khan 694 768 778 685 617 731 768 773 802 747 736 56 7.6% 617 802 Bahawalnagar. 753 705 728 840 816 820 865 847 846 798 802 55 6.9% 705 865 Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 99 TABLE A1.5: PUNJAB: AVERAGE MAIZE YIELDS (KG/ACRE) BY DISTRICT, 2006/07–2015/16 100 NAME 2006-07 2007-08 2008-09 2009-10 2010-11 2011-12 2012-13 2013-14 2014-15 2015-16 Average Stdev COV Min Max THE PUNJAB 1,776 2,040 1,989 2,005 2,203 2,306 2,309 2,360 2,418 2,481 2,189 227 10.4% 1,776 2,481 Attock. 351 419 423 376 453 470 615 908 875 806 570 216 37.9% 351 908 Rawalpindi. 438 405 361 287 409 407 384 656 479 489 432 98 22.7% 287 656 Islamabad. 436 394 357 294 413 409 381 486 394 472 403 55 13.7% 294 486 Jhelum. 2,265 2,875 2,716 2,984 2,957 2,740 2,327 2,863 2,562 2,868 2,716 254 9.3% 2,265 2,984 Cahkwal. 400 400 424 333 375 360 350 344 333 319 364 35 9.5% 319 424 Sargodha. 1,120 1,217 1,277 1,324 1,379 1,374 1,365 1,334 1,317 1,248 1,296 82 6.3% 1,120 1,379 Khushab. 500 522 625 481 550 750 652 645 628 583 594 82 13.9% 481 750 Mianwali. 533 533 500 1,269 1,188 1,000 800 870 489 500 768 303 39.4% 489 1,269 Bhakkar. 444 360 379 367 400 414 481 462 471 460 424 46 10.7% 360 481 Faisalabad. 1,627 2,166 1,884 1,891 2,012 2,119 2,271 2,256 2,440 2,390 2,106 254 12.1% 1,627 2,440 T.T. Singh. 1,885 2,036 1,948 2,172 2,256 2,263 2,580 2,558 2,545 2,660 2,290 283 12.3% 1,885 2,660 Jhang. 2,348 2,480 2,391 2,462 2,610 2,506 2,523 2,410 1,830 2,380 2,394 213 8.9% 1,830 2,610 Chiniot 2,463 2,609 2,632 2,440 2,177 2,370 2,754 2,492 191 7.7% 2,177 2,754 Gujrat. 692 703 659 692 750 800 757 735 702 673 716 43 6.0% 659 800 M.B.Din. 753 851 916 929 914 950 911 856 819 690 859 84 9.8% 690 950 Sialkot. 1,214 1,935 2,376 2,269 2,167 2,360 2,404 2,288 2,190 1,804 2,101 367 17.5% 1,214 2,404 Narowal. 667 923 1,338 1,375 1,750 1,667 1,600 1,000 630 500 1,145 461 40.3% 500 1,750 Gujranwala. 600 1,261 1,552 1,636 1,750 1,762 1,588 1,444 1,294 1,514 1,440 339 23.6% 600 1,762 Hafizabad. 676 941 945 861 909 1,000 1,000 893 632 621 848 148 17.5% 621 1,000 Sheikhupura. 680 687 688 700 706 714 724 750 752 806 721 39 5.4% 680 806 Nankana Sahib 679 688 703 723 740 747 757 758 730 717 724 28 3.8% 679 758 Lahore. 1,865 1,675 1,551 1,695 1,692 1,734 1,759 725 717 730 1,414 483 34.1% 717 1,865 Kasur. 2,174 2,388 2,641 3,198 2,933 3,083 3,121 3,115 2,824 3,158 2,863 355 12.4% 2,174 3,198 Okara. 1,828 2,586 2,464 2,512 2,731 2,894 2,638 2,720 2,868 3,129 2,637 346 13.1% 1,828 3,129 Sahiwal. 2,665 2,859 2,630 2,626 2,961 2,950 2,837 2,818 2,939 2,799 2,808 129 4.6% 2,626 2,961 Pakpattan. 2,774 2,709 2,727 2,605 2,817 2,773 2,871 2,937 3,064 3,033 2,831 146 5.2% 2,605 3,064 Multan. 937 1,015 952 919 953 1,117 1,664 1,647 1,727 1,775 1,271 378 29.7% 919 1,775 Lodhran. 652 778 1,709 1,685 1,722 1,852 1,851 1,661 2,743 2,424 1,708 633 37.1% 652 2,743 Khanewal. 2,630 2,538 2,558 2,707 2,890 2,548 3,060 2,901 2,929 2,935 2,770 194 7.0% 2,538 3,060 Vehari. 2,304 2,489 2,431 2,351 2,508 2,472 2,747 2,833 3,160 3,025 2,632 293 11.1% 2,304 3,160 Muzaffargarh. 602 612 609 627 608 613 625 610 673 678 626 27 4.4% 602 678 Layyah. 700 727 750 667 667 656 667 667 667 630 680 36 5.2% 630 750 D.G.khan. 690 727 750 737 714 714 593 696 677 672 697 45 6.4% 593 750 Rajanpur. 571 571 400 571 500 571 600 583 548 444 536 66 12.4% 400 600 Bahawalpur. 657 650 648 649 651 1,000 2,000 2,020 3,259 2,876 1,441 1,018 70.6% 648 3,259 Rahim Yar Khan 673 698 690 690 688 957 1,067 1,321 1,638 1,886 1,031 445 43.2% 673 1,886 Bahawalnagar. 1,987 2,154 2,110 2,113 2,301 2,470 2,391 2,429 3,099 3,232 2,429 420 17.3% 1,987 3,232 A Feasibility Study TABLE A1.6: PUNJAB: AVERAGE SUGARCANE YIELDS (KG/ACRE) BY DISTRICT, 2006/07–2015/16 NAME 2006-07 2007-08 2008-09 2009-10 2010-11 2011-12 2012-13 2013-14 2014-15 2015-16 average stdev cov min max THE PUNJAB 21,343 19,719 19,608 20,869 22,565 22,803 22,658 23,371 23,391 24,078 22,041 1,569 7.1% 19,608 24,078 Attock. 7,930 15,860 16,040 14,300 15,720 14,920 14,128 3,107 22.0% 7,930 16,040 Rawalpindi. Islamabad. Jhelum. 8,400 16,800 17,160 15,120 15,300 15,400 15,600 16,090 15,230 14,860 14,996 2,435 16.2% 8,400 17,160 Cahkwal. Sargodha. 20,491 18,886 18,289 18,849 22,992 22,283 20,528 20,342 19,969 20,081 20,271 1,474 7.3% 18,289 22,992 Khushab. 20,678 17,767 17,281 17,916 18,886 18,326 19,446 16,759 20,006 19,036 18,610 1,231 6.6% 16,759 20,678 Mianwali. 17,729 15,938 15,490 21,910 23,141 25,231 24,261 23,216 18,737 20,342 20,599 3,488 16.9% 15,490 25,231 Bhakkar. 17,356 15,490 15,863 18,252 20,155 20,528 21,574 20,566 21,088 22,096 19,297 2,386 12.4% 15,490 22,096 Faisalabad. 22,469 19,409 19,782 19,707 21,126 20,902 20,976 21,462 21,872 22,208 20,991 1,070 5.1% 19,409 22,469 T.T. Singh. 22,731 19,931 20,230 20,753 23,514 23,664 23,514 23,141 23,328 22,768 22,357 1,462 6.5% 19,931 23,664 Jhang. 20,715 18,140 19,036 20,155 22,320 21,760 22,208 23,328 22,693 22,022 21,238 1,679 7.9% 18,140 23,328 Chiniot 20,230 22,731 21,462 20,342 22,022 21,648 21,835 21,467 901 4.2% 20,230 22,731 Gujrat. 17,207 17,729 15,863 15,490 17,767 17,915 18,102 17,990 17,692 16,422 17,218 949 5.5% 15,490 18,102 M.B.Din. 19,969 16,721 16,423 17,543 18,289 19,595 19,409 18,662 18,774 19,035 18,442 1,201 6.5% 16,423 19,969 Sialkot. 11,683 12,242 12,018 11,980 12,690 13,250 13,438 13,250 13,063 12,317 12,593 627 5.0% 11,683 13,438 Narowal. 12,615 12,877 12,318 12,130 15,340 15,750 15,713 15,750 13,623 12,690 13,880 1,566 11.3% 12,130 15,750 Gujranwala. 14,818 15,490 15,303 15,080 19,446 19,408 19,036 18,738 15,788 15,043 16,815 2,042 12.1% 14,818 19,446 Hafizabad. 18,699 15,676 15,378 15,303 15,863 17,916 18,849 18,476 18,625 17,543 17,233 1,501 8.7% 15,303 18,849 Sheikhupura. 18,066 18,290 18,290 17,804 20,977 22,544 18,887 19,036 21,312 19,467 1,702 8.7% 17,804 22,544 Nankana Sahib 17,356 18,103 18,364 18,588 19,483 18,812 18,961 20,715 20,902 19,031 1,167 6.1% 17,356 20,902 Lahore. 16,610 17,540 17,650 17,360 18,290 18,400 19,040 20,300 19,780 18,330 1,198 6.5% 16,610 20,300 Kasur. 20,342 18,289 17,916 20,454 20,491 20,528 18,924 20,342 20,491 20,603 19,838 1,040 5.2% 17,916 20,603 Okara. 18,774 16,759 18,476 18,812 19,782 19,969 19,894 19,484 19,222 19,035 19,021 944 5.0% 16,759 19,969 Sahiwal. 16,348 16,796 16,647 16,050 16,796 19,222 19,110 19,409 19,521 19,969 17,987 1,569 8.7% 16,050 19,969 Pakpattan. 22,470 20,491 19,409 20,193 20,902 21,275 21,089 20,155 20,267 20,342 20,659 831 4.0% 19,409 22,470 Multan. 17,654 17,767 17,356 16,050 20,080 20,529 20,678 19,931 19,856 20,529 19,043 1,666 8.7% 16,050 20,678 Lodhran. 21,275 19,520 18,662 21,275 18,886 20,902 22,768 22,842 23,030 25,754 21,491 2,189 10.2% 18,662 25,754 Khanewal. 21,461 19,819 20,155 20,977 22,021 21,051 21,462 20,902 24,821 25,231 21,790 1,821 8.4% 19,819 25,231 Vehari. 24,261 22,507 21,387 24,672 25,008 24,634 24,821 23,813 23,701 23,888 23,869 1,137 4.8% 21,387 25,008 Muzaffargarh. 22,022 21,648 19,782 21,424 21,350 23,962 22,395 24,261 23,514 25,381 22,574 1,678 7.4% 19,782 25,381 Layyah. 19,931 19,707 20,267 21,499 21,275 21,686 21,648 22,021 21,835 21,723 21,159 855 4.0% 19,707 22,021 D.G.khan. 22,283 20,193 20,230 21,834 22,955 23,328 22,768 23,141 22,880 24,821 22,443 1,408 6.3% 20,193 24,821 Rajanpur. 22,581 21,835 21,648 29,524 27,620 29,113 28,740 29,486 27,434 30,233 26,821 3,425 12.8% 21,648 30,233 Bahawalpur. 22,059 21,126 20,902 21,126 26,911 25,157 25,008 25,381 25,008 26,687 23,936 2,376 9.9% 20,902 26,911 Rahim Yar Khan 26,314 26,053 24,298 26,575 27,993 28,927 30,046 30,233 30,121 30,905 28,146 2,239 8.0% 24,298 30,905 Bahawalnagar. 20,528 18,476 18,961 20,118 24,559 22,880 23,328 22,955 23,179 23,291 21,827 2,111 9.7% 18,476 24,559 Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 101 TABLE A1.7:  PUNJAB: SOWN AREA, UNIRRIGATED AREA, AND IRRIGATED AREA BY DIVISION AND DISTRICT, 2013–14 Sown Area Un-irrigated Un-irrigated Irrigated Irrigated (000 Area (000 Area (% of Area (000 Area (% of Division / District Hectares) hectares) total) hectares) total) Bahawalpur Divn. 3,096 25 1% 3,071.00 99% Bahawalpur 1,564 4 0% 1,560.00 100% Bahawalnagar 644 16 2% 628.00 98% R. Y. Khan 888 5 1% 883.00 99% D.G.Khan Divn. 1,892 196 10% 1,696.00 90% D. G. Khan 346 44 13% 302.00 87% Layyah 503 44 9% 459.00 91% Muzaffargarh 668 20 3% 648.00 97% Rajanpur 375 56 15% 319.00 85% Faisalabad Divn. 2,033 79 4% 1,954.00 96% Faisalabad 691 0 0% 691.00 100% Chiniot 287 0 0% 287.00 100% Jhang 682 79 12% 603.00 88% Toba Tek Singh 373 79 21% 294.00 79% Gujranwala Divn. 2,256 213 9% 2,043.00 91% Gujranwala 548 546 100% 2.00 0% Gujrat 296 131 44% 165.00 56% Hafizabad 344 0 0% 344.00 100% Mandi Baha-ud-Din 345 4 1% 341.00 99% Narowal 307 65 21% 242.00 79% Sialkot 416 11 3% 405.00 97% Lahore Divn. 1,484 2 0% 1,482.00 100% Lahore 163 2 1% 161.00 99% Kasur 505 0 0% 505.00 100% Nankana Sahib 299 0 0% 299.00 100% Sheikhupura 517 0 0% 517.00 100% Multan Divn. 2,089 7 0% 2,082.00 100% Multan 451 4 1% 447.00 99% Khanewal 544 2 0% 542.00 100% Lodhran 455 0 0% 455.00 100% Vehari 639 1 0% 638.00 100% Rawalpindi Divn. 811 722 89% 89.00 11% Rawalpindi 231 221 96% 10.00 4% Attock 227 199 88% 28.00 12% Chakwal 248 232 94% 16.00 6% Jhelum 105 70 67% 35.00 33% Sahiwal Divn. 1,424 0 0% 1,424.00 100% Sahiwal 426 0 0% 426.00 100% Okara 617 0 0% 617.00 100% Pakpattan 381 0 0% 381.00 100% Sargodha Divn. 2,109 741 35% 1,368.00 65% Sargodha 561 0 0% 561.00 100% Bhakkar 788 386 49% 402.00 51% Khushab 427 286 67% 141.00 33% Mianwali 333 69 21% 264.00 79% Islamabad 25 25 100% 0.00 0% Punjab 17219 2010 12% 15209 88% Source: Government of Punjab, Bureau of Statistics, 2014. Note: Excludes 485,000 hectares under orchards and 17,000 hectares under tobacco, sown under “Zaid Rabi” crop. 102 A Feasibility Study TABLE A1.8:  PUNJAB: CORRELATION BETWEEN PERCENTAGE OF 2013–14 SOWN AREA UNIRRIGATED PER DISTRICT AND COEFFICIENT OF VARIATION IN WHEAT YIELDS, 2006–07 TO 2015–16 Standard Percent Sown Average Yield Deviation Cofficient of area Un- NAME (Kg/Ha) (Kg/Ha) Variation (%) irrigated (%) THE PUNJAB 1,108.1 54.4 4.9% 12% Cahkwal. 537.6 147.6 27.5% 94% Attock. 592.1 146.3 24.7% 88% Rawalpindi. 623.9 131.2 21.0% 96% Islamabad. 610.8 128.1 21.0% 100% Jhelum. 735.3 116.5 15.8% 67% Sialkot. 1,032.7 160.2 15.5% 3% Mianwali. 863.2 107.7 12.5% 21% Narowal. 868.2 100.7 11.6% 21% Rahim Yar Khan 1,223.9 126.4 10.3% 1% Gujrat. 709.3 71.4 10.1% 44% Sheikhupura. 1,149.8 103.4 9.0% 0% Multan. 1,131.1 96.8 8.6% 1% Rajanpur. 1,079.1 92.1 8.5% 15% Lahore. 1,122.5 92.7 8.3% 1% Khushab. 742.9 60.5 8.1% 67% D.G.khan. 1,088.0 80.6 7.4% 13% Lodhran. 1,246.9 90.0 7.2% 0% Sargodha. 1,012.4 73.0 7.2% 0% T.T. Singh. 1,274.6 90.8 7.1% 21% Layyah. 1,035.3 71.5 6.9% 9% Vehari. 1,254.0 83.2 6.6% 0% Gujranwala. 1,257.9 81.1 6.4% 100% Nankana Sahib 1,226.9 78.7 6.4% 0% Bahawalpur. 1,211.8 77.2 6.4% 0% Hafizabad. 1,199.0 75.4 6.3% 0% Faisalabad. 1,226.6 73.8 6.0% 0% Bahawalnagar. 1,202.9 67.2 5.6% 2% Jhang. 1,171.9 65.2 5.6% 12% Muzaffargarh. 1,110.8 59.2 5.3% 3% Khanewal. 1,239.6 63.4 5.1% 0% Pakpattan. 1,395.5 71.0 5.1% 0% Bhakkar. 940.3 47.2 5.0% 49% Chiniot 1,215.4 60.6 5.0% 0% Sahiwal. 1,242.0 61.2 4.9% 0% Kasur. 1,196.6 58.9 4.9% 0% Okara. 1,393.0 63.1 4.5% 0% M.B.Din. 1,062.9 44.2 4.2% 1% Correlation 73% Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 103 ANNEX 2 PAKISTAN CRED-EM-DAT DATABASE FOR NATURAL DISASTERS, 1900–2017 TABLE A2.1:  PAKISTAN: TOP 10 DISASTERS BY TOTAL NUMBER OF DEATHS Total Disaster Total Total damage Year type Occurrence deaths Injured Affected Homeless affected (US$000) 2005 Earthquake 1 73,338 128,309 5,000,000 5,128,309 5,200,000 1935 Earthquake 1 60,000 1965 Storm 1 10,000 1974 Earthquake 1 4,700 15,000 30,000 5,200 50,200 3,255 1945 Earthquake 1 4,000 1950 Flood 1 2,900 2010 Flood 4 2,113 2,946 20,360,550 20,363,496 9,500,000 1992 Flood 2 1,446 9,888,553 2,951,315 12,839,868 1,000,230 2015 Extreme 1 1,229 80,000 80,000 temperature 1995 Flood 3 1,063 1,855,000 1,855,000 Total 16 160,789 146,255 32,214,103 7,956,515 40,316,873 15,703,485 Source: http://www.emdat.be/database Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 105 TABLE A2.2: PAKISTAN: DAMAGE RECORD BY TYPE OF EVENT, 1900–2017 Number Total of Total Total damage Type of natural disaster events deaths Injured Affected Homeless affected (US$000) Flood 94 17,248 11,670 75,078,228 4,242,150 79,332,048 20,969,178 Earthquake 31 143,734 150,279 1,937,624 5,187,485 7,275,388 5,329,755 Storm 25 11,969 1,456 2,369,040 234,090 2,604,586 1,715,036 Landslide 22 789 209 30,645 3,300 34,154 18,000 Extreme temperature 17 2,774 324 80,250 0 80,574 18,000 Epidemic 10 283 211 16,275 0 16,486 Mass movement (dry) 2 63 0 0 0 0 0 Drought 1 143 2,200,000 2,200,000 247,000 Insect infestation 1 Total 203 177,003 164,149 81,712,062 9,667,025 91,543,236 28,296,969 Number Total of Total Total damage Type of natural disaster events deaths Injured Affected Homeless affected (US$000) Flood 46% 10% 7% 92% 44% 87% 74% Earthquake 15% 81% 92% 2% 54% 8% 19% Storm 12% 7% 1% 3% 2% 3% 6% Landslide 11% 0% 0% 0% 0% 0% 0% Extreme temperature 8% 2% 0% 0% 0% 0% 0% Epidemic 5% 0% 0% 0% 0% 0% 0% Mass movement (dry) 1% 0% 0% 0% 0% 0% 0% Drought 0% 0% 0% 3% 0% 2% 1% Insect infestation 0% 0% 0% 0% 0% 0% 0% Total 100% 100% 100% 100% 100% 100% 100% Source: http://www.emdat.be/database TABLE A2.3: PAKISTAN: ANALYSIS OF DAMAGE RECORD BY DECADE Total No Total No. of Decade Total deaths Injured Affected Homeless people damage (US$ occurrences affected 000) 1900-1909 1910-1919 1920-1929 1 0 0 0 0 0 0 1930-1939 2 60,000 0 0 0 0 0 1940-1949 2 4,000 0 0 10,000 10,000 0 1950-1959 8 3,850 0 0 0 0 0 1960-1969 5 10,519 0 625,502 0 625,502 7,400 1970-1979 8 6,850 15,000 13,412,200 260,200 13,687,400 1,169,755 1980-1989 20 1,074 3,578 315,875 1,002,000 1,321,453 5,000 1990-1999 44 6,654 1,451 17,881,338 3,233,770 21,116,559 1,361,166 2000-2009 68 77,142 131,479 12,369,076 5,019,420 17,519,975 7,536,648 2010-2017 45 6,914 12,641 37,108,071 141,635 37,262,347 18,217,000 Total 203 177,003 164,149 81,712,062 9,667,025 91,543,236 28,296,969 Source: http://www.emdat.be/database 106 A Feasibility Study ANNEX 3 COMPENSATION PAID TO FLOOD VICTIMS IN PUNJAB, 2010–15 TABLE A3.1:  PUNJAB: COMPENSATION PAID (PKR) TO FLOOD VICTIMS FOR LOSS OF LIVELIHOOD, HOUSING, AND CROPS, 2010–15, BY DISTRICT HOUSE DAMAGE DISTRICT DEATHS INJURIES LIVELIHOOD CROPS % Crops TOTAL % Total (TOTAL) Bahawalpur 31,390,000 94,580,000 137,407,200 2% 263,377,200 1% Chiniot 28,340,000 570,890,000 245,529,000 4% 844,759,000 2% Gujranwala 39,820,000 48,825,000 395,900,750 6% 484,545,750 1% Gujrat 11,000,000 70,630,000 67,415,000 1% 149,045,000 0% Hafizabad 13,450,000 359,155,000 366,131,800 5% 738,736,800 2% Jhang 406,170,000 2,613,805,000 1,625,415,000 24% 4,645,390,000 11% Jhelum 13,310,000 9,185,000 142,310,250 2% 164,805,250 0% Khanewal 20,880,000 97,415,000 223,437,500 3% 341,732,500 1% Khushab 89,340,000 59,960,000 129,290,250 2% 278,590,250 1% Mandi Bahauddin 16,750,000 76,340,000 251,043,000 4% 344,133,000 1% Multan 534,920,000 1,193,455,000 644,680,675 10% 2,373,055,675 6% Muzaffargarh 5,242,110,000 5,947,260,000 1,306,791,600 19% 12,496,161,600 30% Narowal 13,520,000 81,645,000 169,279,800 3% 264,444,800 1% Sargodha 298,190,000 272,110,000 671,862,638 10% 1,242,162,638 3% Sheikhupura 38,430,000 30,475,000 238,809,500 4% 307,714,500 1% Sialkot 29,090,000 204,210,000 88,897,200 1% 322,197,200 1% Bhakkar 223,220,000 175,800,000 0% 399,020,000 1% DG Khan 525,760,000 1,158,580,000 0% 1,684,340,000 4% Layyah 916,860,000 1,606,040,000 0% 2,522,900,000 6% Mianwali 1,297,780,000 951,160,000 0% 2,248,940,000 5% RY Khan 998,480,000 938,120,000 0% 1,936,600,000 5% Rajanpur 2,522,540,000 4,574,220,000 0% 7,096,760,000 17% Total 13,311,350,000 21,133,860,000 6,704,201,163 100% 41,149,411,163 100% Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 107 FIGURE A3.1:  PUNJAB: COMPENSATION PAID (PKR) TO FLOOD VICTIMS FOR LOSS OF LIVELIHOOD, HOUSING, AND CROPS, 2010–15, BY DISTRICT 7,000,000,000 6,000,000,000 5,000,000,000 4,000,000,000 3,000,000,000 2,000,000,000 1,000,000,000 0 r ra ot a af jrat ad Jh g Kh lum an Kh al ha b M Mu n af tan N rh l kh a Si a Bh ot r an M ah i an r Sa wa al pu D ka pu al an Ba sha di ei dh ur ew ni k a nw ab Kh yy Kh nw ud al u l ak al rg up an o hi e Jh Sh rgo G an ar La u w iz ia fa C aj G Y ha R R uj H uz Ba G di M Livelihood House damage (total) Crops Source: PDMA Punjab April 2017. 108 A Feasibility Study ANNEX 4 INTERNATIONAL EXPERIENCE WITH CROP AREA-YIELD INDEX INSURANCE The origins of area-yield index crop insurance date to 1952 in Sweden. India introduced area-based crop insurance in the late 1970s, and the USA and Canada introduced it in the early 1990s. Countries that have developed area-based crop insur- ance in the past decade include Brazil, Mexico, Morocco, Peru, and Sudan.55 In recent years, the World Bank has provided assistance for technical feasibility studies of AYII in Bangladesh (paddy rice), Burkina Faso (cotton), Guyana (paddy rice), Kazakhstan (rainfed spring wheat), Nepal, (rainfed and irrigated food crops and oilseeds), Senegal (rainfed food crops and oilseeds), and most recently Kenya (maize). The experiences of India, the USA, and Brazil are reviewed here. AREA-YIELD INDEX INSURANCE IN INDIA India has the world’s largest public-sector index-based crop insurance pro- gram. National in scope, this program now covers tens of millions of mainly small and marginal farmers. The national program originated in the late 1970s, when AYII was implemented on an experimental basis. Between 1985 and 1999, the government imple- mented AYII under the Comprehensive Crop Insurance Scheme (CCIS), which was formally replaced in the Rabi 1999–2000 season by the National Agricultural Insurance Scheme (NAIS), underwritten by the Agricultural Insurance Company of India Ltd (AIC). NAIS was conceived originally as a public-sector social crop insurance pro- gram. It was designed to provide small and marginal farmers access to seasonal pro- duction credit to invest in high-yielding crop technology and to remain creditworthy in the event of severe crop losses. The program was a partnership between the national and state governments. In recognition that individual farmer MPCI would have been impossible given the large number of very small landholdings, an area yield indexed approach was adopted from the outset. The NAIS program was explicitly linked on a compulsory basis to the provision of seasonal crop production credit through the rural banking network, but non-borrowing farmers were also encouraged to purchase cov- erage on a voluntary basis. The program, which covered a wide range of food, oilseed, 55In Peru, the AYII program covers a range of crop types, from quinoa, potatoes, and cereals (such as barley) to cotton. Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 109 pulse, and cash crops, is heavily subsidized in two ways: an additional premium subsidy on the capped premium (1) premium rates for crops are capped at 1.5–2 percent rates. The scheme used government-implemented sam- for Rabi crops and 2.5–3.5 percent for Kharif crops, and ple yield CCEs to establish the actual average area yield, (2) federal and state governments reinsure (on a 50:50 which was the basis for indemnifying losses if the actual basis) all losses in excess of a 100 percent loss ratio for yield was below the trigger yield. Box A4.1 summarizes food crops and oilseeds and a 150 percent loss ratio for other key features of NAIS. commercial crops. The sum insured for loanees was usu- ally based on the value of the loan, but they were entitled AIC has continued to innovate. In 2007, AIC to purchase additional coverage up to 150 percent of the launched a new series of market-based and actuari- value of threshold yield for the payment of commercial ally rated micro-level WII programs (organized under premium rates. Small and marginal farmers qualified for BOX A4.1:  KEY FEATURES OF INDIA’S AREA-BASED NATIONAL AGRICULTURAL INSURANCE SCHEME FOR CROPS Implementing agencies. The Agricultural Insurance Company of India (AIC), a public-sector insurer specializing in crop insurance, is responsible for implementing area-yield index insurance (AYII) under the National Agricultural Insurance Scheme (NAIS). The program started implementation in 1980. Target audience. The program is targeted at small and marginal farmers (with less than 2 hectares), who are highly dependent on access to seasonal crop credit. Crop insurance is compulsory for borrowing farmers and voluntary for non-borrowing farmers. Insured crops include wheat, paddy rice, maize, other cereals, oilseeds, pulses, and industrial crops such as cotton and sugarcane. Insured unit. The insured unit is normally the block or panchayat, which comprises a group of nearby villages and which may include up to 10,000 hectares or more of a single crop and several thousands of small and marginal farmers. Farmers may select coverage levels of 60 percent, 80 percent, or a maximum of 90 percent of the five-year average area yield. Sum insured. The sum insured for each insured yield coverage level is based on the amount of seasonal crop credit borrowed by the farmer. Premium rates. Premium rates are capped by the government at 2.5–3.5 percent for most food crops, oilseeds, and pulses to make the program affordable to small and marginal farmers. Commercial crops are charged at the full actuarially determined premium rates. Administration and distribution channels. The program is marketed through the rural agricultural bank branch network in each state and department and block (group of villages). AIC maintains a national headquarters staff and a small regional team in each state. It has not, however, attempted to establish branch offices, as there is no need to duplicate the rural bank branch network. The insurers’ administrative costs are kept to a minimum by linking insurance with rural finance. Area-yield measurement. Actual area yields are established through sample crop cutting and weighing of crop yields from randomly selected farms in each insured unit. The crop cut yields are averaged to calculate the actual average area yield in each insured unit. This major and costly exercise can suffer from delays in processing the results; indemnity payments are often delayed for six months or more. Scale and outreach. By virtue of being a mainly compulsory program, the NAIS is the world’s largest crop insurance program, currently insuring about 34 million Indian farmers (representing an insurance uptake rate of about 24 percent percent of all farmers). Government financial and reinsurance support. The program is highly dependent on government subsidies and operates at a major financial loss. The federal and state governments provide excess of loss claims reinsurance protection to AIC and also fund premium subsidies. AIC’s administrative and operating expenses are subsidized by the government. Modified NAIS. Since the Rabi 2010–11 season, AIC has operated a fully commercial modified version of NAIS (mNAIS) in about 10 percent of the departments covered by the program. Under this market-based program, AYII is charged at the full commercial premium rates, and AIC places a combination of proportional and non-proportional reinsurance with international reinsurers. Source: Authors. 110 A Feasibility Study BOX A4.2:  MAIN FEATURES OF INDIA’S MODIFIED NAIS (MNAIS) SCHEME FOR RABI 2010–11 Actuarial regime. The mNAIS scheme operates on an “actuarial regime” in which the government’s financial liability is pre- dominantly in the form of premium subsidies given to AIC and funded ex-ante, thereby reducing the contingent and uncertain ex-post fiscal exposure faced by the government under NAIS and reducing delays in settling claims. Up-front premium subsidies. AIC receives premiums (farmer collections plus premium subsidies from the government) and is responsible for managing the liability of mNAIS through risk transfer to private reinsurance markets and risk retention through its reserves. It is financially able to operate on a sustainable basis. On-account partial payment. The mNAIS product continues to be based on an area yield approach, with a provision for an early partial payment to farmers (in season) based on weather indices. Small insurance units. Crop cutting experiments to estimate crop yields are lowered from the block level to the village level to reduce basis risk (i.e., the mismatch between the actual, individual crop yield losses and the insurance indemnity). Cutoff dates. Adverse selection is reduced through the enforcement of early purchase deadlines ahead of the crop season. Additional benefits. Additional benefits are offered for prevention of sowing, replanting, and post-harvest losses, as well as for localized risk, such as hail losses or landslides. Early settlement of claims. mNAIS combines weather-based indices for on-account partial payment of claims in case of adverse mid-season conditions, while area yield indices are used for final payment of claims. The final estimation of loss is based on area yield measurement at the time of harvest using crop cutting experiments. Source: GFDRR 2011. TABLE A4.1: INDIA: PUBLIC SECTOR CROP INDEX INSURANCE COVERAGE, 2012–13 Insured farmers Insured area Sum insured Premium Program (million) (million hectares) (US$ million) (US$ million) Type of program NAIS 15.45 29.92 5,892.35 195.67 Administered WBCIS 13.23 18.39 4,038.47 379.45 Actuarial MNAIS 2.98 2.97 1,149.20 125.10 Actuarial Total 31.66 51.28 11,080.02 700.22 Source: Rao 2015. the umbrella of AIC’s Weather-Based Crop Insurance Between the start of NAIS in 2000 and 2014, the Scheme), and in 2010 it launched a Modified NAIS program insured nearly 229 million farmers. (mNAIS), which is a market-based actuarially rated pro- Over this period, 59 million farmers (26 percent of total) gram (described in Box A4.2). have received a claim payment, but on account of the capped premium rates (average rate of 3.0 percent), the The Indian crop insurance program has program has operated at a huge financial loss (cost to expanded significantly in recent years. In 2008, government), as shown by the loss ratio of 314 percent NAIS was insuring 18 million farmers (16 percent of all (or a payout of Rs 3.14 for every Rs 1.0 of collected pre- rural households, of which two-thirds were small and mium) (Table  A4.2). While NAIS has successfully pro- marginal farmers with less than 2 hectares) and gener- vided crop credit and insurance for about one in four ated total premiums of Rs 800 crore (US$178 million) Indian farmers, the program has also been criticized at (World Bank 2011). In 2013–14, with the addition of several levels, including: (1) the basis risk associated with the Weather-Based Crop Insurance Scheme and modi- unit areas of insurance that are too large and a num- fied mNAIS, the number of insured farmers had risen ber of CCEs that are too low, (2) the significant delays to 34 million (24 percent of all farming households) with in settling claims (often 6–9 months), and (3) the unpre- premiums of US$700 million (Table A4.1). dictable and unbudgeted financial burden to the state Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 111 TABLE A4.2:  INDIA: SUMMARY OF COVERAGE AND PERFORMANCE OF PUBLIC SECTOR SUBSIDIZED CROP INDEX INSURANCE FOR SMALL FARMERS Number of Average Premium % Loss No. Farmers % farmers Farmers Area Insured Sum Insured Premium Premium Subsidy Premium Claims (Rs. ratio receiving a receiving Program Period Insured (000) (000 Ha) (Rs. Crore) (Rs. Crore) Rate % (Rs.Crore) subsidy Crore) % Payout (000) a payout National Rabi Agricultural 1999/2000 to Insurance Kharif 2014 Scheme NAIS (30 seasons) 229,349 339,674 349,667 10,599 3.0% 1,392 13% 33,329 314% 59,154 26% Rabi 2010/11 to Kharif 2014 Modified NAIS (8 seasons) 9,681 10,836 21,359 2,363 11.1% 1,444 61% 1,719 73% 1,656 17% Weather Based Kharif 2007 to Crop Insurance Kharif 2014 Scheme (WBCIS) (16 Seasons) 34,136 45,987 62,714 5,950 9.5% 3,948 66% 4,079 69% 19,006 56% Source: AIC, http://www.aicofindia.com/ and national governments of compensating excess claims mNAIS is too expensive for farmers to afford. Farmers (World Bank 2011). For these reasons, in 2005 the Gov- in the future will pay a uniform premium rate of 2.0 ernment of India requested assistance from the World percent for Kharif crops and 1.5 percent for Rabi crops, Bank to reform the NAIS. while the rate for commercial and horticultural crops will be 5 percent. The rest of the premium will be paid by the As noted, the Government of India has heav- government with no upper limit on the subsidy amount— ily promoted weather-based crop index insur- in other words, rates will be actuarially determined, and ance since 2007. Many types of WII coverage have government will settle the difference between the flat rate been tested in India over the past decade, including: paid by the farmer and the rate charged by the insurer. total seasonal rainfall indexes; weighted rainfall indexes; The program will be open to all 4 public sector insur- multiple-phase weather indexes where the growing sea- ance companies (including AIC) and to 11 private sector son for a named crop is divided into phases from sow- insurers that are involved in crop insurance provision. ing and germination to the vegetative stage, flowering, The government hopes to roll out PMFBY to 50 percent grain formation, and maturity; specific indexes for excess of farmers over the next two years, which would require rain, high or low temperature, and humidity; and even coverage to increase by roughly 100 million people. weather indices for pests and diseases (Clarke et al. 2012). The Weather-Based Crop Insurance Scheme has proved popular with the state governments of India, several of which have switched from NAIS to a Weather-Based BRAZIL Crop Insurance Scheme linked to credit. Between 2007 Brazil introduced a maize-seed AYII program and 2014, the scheme has insured 34 million farmers. for small farmers in 2001 under a PPP between Actuarially determined premium rates are again high at the government of Rio Grande do Sul State, local 9.5 percent on average, and attract very high levels of insurers, and international reinsurers. The AYII government premium subsidies (66 percent on average). coverage—known as the Grupo de Risco Municipalizado The program operated profitably between 2007 and (GRM, Municipalized Risk Group) program—was linked 2014, with a loss ratio of 69 percent (Table A4.2). The to the state government maize seed swap program.56 The main criticisms of WII programs in India center on spa- seed swap program was aimed at introducing new hybrid tial and product basis risk (Clarke et al. 2012; Rao 2014; maize and was a voluntary crop insurance program for see further discussion below). individual farmers. The insured unit was the munici- pality, and in the first year the program provided a fixed insured yield coverage level of 90 percent of the expected In 2016, the government announced a radical or average maize yield for the municipality, which was plan to replace the NAIS and mNAIS programs reduced to 80 percent coverage in subsequent years. The with a single program termed the Pradan Man- state government provided very high premium subsidies tri Fasal Bima Yogana (PMFBY, Prime Minister’s (around 90 percent of the premium costs) to promote the Crop Insurance Scheme). The main change is that the government will revert to charging farmers flat or capped premium rates, on account of its concern that 56Programa Troca Troca de Sementes (PTTS). 112 A Feasibility Study BOX A4.3: BRAZIL’S MAIZE AYII PROGRAM IN RIO GRANDE DO SUL STATE The Grupo de Municipalized Group Risk (Risco Municipalizado, GRM) plan was a public-private partnership between the State Department of Agriculture and Supply (SSAA), the State Bank of Rio Grande do Sul (BANRISUL), the State Data Processing Company (PROCERGIS), various private local insurers, IRB (the national reinsurer), and international reinsurers. At its launch in 2001, the program was underwritten by Porto Seguro Insurance Company, with reinsurance support from PartnerRe. Agro- Brazil, a private risk management agency based in Rio Grande do Sul, implemented the program. The GRM product used the average maize yield at the municipality level as the index for triggering payouts to insured farmers. The Brazilian Institute of Geography and Statistics (IBGE) provided historical and real-time (i.e., during the insurance coverage period) maize production and yield data in each municipality (the insured unit). The insured yield level was initially set in 2001–02 at 90 percent of the expected yield in each municipality but adjusted downwards to 80 percent coverage across all states in 2002–03 to avoid over-in- surance of actual average yields. The program was marketed on a voluntary basis to farmers participating in the state maize seed swap program (PTTS). The sum insured was based on the cost of maize production and specifically included the cost of hybrid seed provided under PTTS; it varied from a low of Real (R)$ 200 per hectare to a maximum of R$ 1,000 per hectare. Premium rates varied from an average low of 11.1 percent in 2001–02 for 90 percent coverage to a maxim average of 17.1 percent for 80 percent coverage in 2007–08. The program experienced a high loss ratio in Year 1 of 215 percent, largely because expected yields were overestimated (they were subsequently corrected). The loss ratio in 2004–05, a severe drought year in Rio Grande do Sul, was 377 percent. The overall loss ratio was 80.1 percent over the life of the GRM. The main operational drawback of the program was that IBGE publishes crop yield estimates by municipality and state in October, so farmers had to wait 3–6 months to receive payouts. GRM MAIZE RESULTS 2001–02 TO 2007–08 Number Sum Average Number Value of insured insured Premium premium claims payouts Loss ratio Crop year farmers (RS) (R$) rate % payouts (RS) % 2001–02 25,068 17,804,385 1,978,154 11.1 17,590 4,247,742 215 2002–03 38,620 28,445,320 4,174,436 14.7 59 5,550 0 2003–04 20,122 14,993,630 2,278,775 15.2 4,254 1,063,611 47 2004–05 24,151 19,320,800 2,749,323 14.2 23,248 10,364,084 377 2005–06 46,175 36,940,000 6,139,370 16.6 9,547 1,914,202 31 2006–07 25,071 20,056,800 3,343,580 16.7 129 30,461 1 2007–08 14,893 11,914,400 2,037,171 17.1 2,951 593,551 29 Total 194,100 149,475,335 22,700,809 15.2 57,778 18,219,201 80 Source: IFAD and WFP 2010. program and make it more affordable to the farmers, avoid overestimating expected yield potential, and also in defined as those with less than 80 hectares. The Insti- setting the maximum insured yield coverage at a realistic tuto Brasileiro de Geografia e Estatística (IBGE, Brazilian level of no more than 80 percent (IFAD and WFP 2010). Institute of Geography and Statistics) provided historical Further details of the program are included in Box A4.3. data on maize production and yield at the municipality level for contract design and rating analysis. IBGE was also appointed as the official organization responsible for declaring the average maize yield in each municipality. USA Payouts to individual farmers were triggered when the In the USA, AYII is marketed under the Group Risk actual municipality average maize yield fell short of the Plan (GRP). Rather than being based on the yield loss 80 percent trigger yield. The program operated until experienced by each individual farmer, payouts under the 2007–08. Over its lifespan, the program insured 194,100 GRP are based on the actual value of an area yield index maize farmers, issued claim payments to nearly 58,000 in a certain area (the insured unit), which in the USA is farmers, and had an overall loss ratio of 80.3 percent. the county (the average insured unit is 2,500 square kilo- A key lesson from this program was that great caution meters). A farmer is indemnified when the actual yield should be taken in starting up a new AYII program to for the insured crop in the county where the insured is Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 113 situated, as determined by the National Agricultural Sta- county average. Final payments are not determined until tistics Service, falls below the guaranteed yield chosen by six months after the crop is harvested, when the National the farmer. Under the GRP, farmers can choose among Agricultural Statistics Service releases the actual yields different coverage levels (insured yield options): 90 per- for each county. Payments are then made within 30 days. cent, 85 percent, 80 percent, 75 percent, or 70 percent GRP insurance policies are easier to administrate and of the expected county yield. The sum insured for each less costly than the traditional individual grower MPCI crop is based on a percentage of the expected market policy. The GRP policies may not cover individual crop price. The grower may elect an insured value of between losses, however, if the county yield does not suffer a simi- a minimum of 90 percent and a maximum of 150 per- lar level of loss. This type of insurance is most appropri- cent of the expected market price. The justification for ate for farmers whose crop production and yields (and permitting growers to insure at up to 150 percent of the losses) typically follow the county pattern. expected market price is that this affords adequate pro- tection for growers whose own yields are higher than the 114 A Feasibility Study ANNEX 5 RESULTS OF PRELIMINARY AYII COVERAGE AND RATING ANALYSIS FOR LODHRAN DISTRICT, PUNJAB The tables and figures that follow present data on area and yields of the five main crops (wheat, cotton, rice, maize, and sugarcane) over 10 years in Kehror Pacca, Dunyapur, and Lodhran tehsils of Lodhran District, along with the results of AYII rating analysis using actual and detrended yields. For wheat, there is very little evidence of any trend toward increasing yields over the past 10 years in these three tehsils (Table A5.1, Figure A5.1). Table A5.2 shows the AYII rating analysis results with average and detrended wheat yields for the three tehsils. TABLE A5.1:  WHEAT CULTIVATED AREA AND AVERAGE YIELDS, 2007–08 TO 2016–17 (a) Cultivated area by tehsil (acres) (b) Average yield (kg/acre) Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 115 FIGURE A5.1: WHEAT CULTIVATED AREA AND AVERAGE YIELDS, 2007–08 TO 2016–17 (a) Cultivated area by tehsil (acres) (b) Average yield (kg/acre) 250,000 1,900 y = 41.921x + 1177.8 1,700 R2= 0.4987 200,000 1,500 150,000 1,300 1,100 y = 21.992x + 1262.4 y = 13.337x + 1313.2 100,000 R2= 0.1676 R2= 0.0704 900 50,000 700 0 500 8 9 0 1 2 3 4 5 6 8 9 0 1 2 3 4 5 6 7 –0 –0 –1 –1 –1 –1 –1 –1 –1 –0 –0 –1 –1 –1 –1 –1 –1 –1 –1 07 08 09 10 11 12 13 14 15 07 08 09 10 11 12 13 14 15 16 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 Kehror Pacca Dunyapur Lodhran Kehror Pacca Dunyapur Lodhran Linear (Kehror Pacca) Linear (Dunyapur) Linear (Lodhran) 116 A Feasibility Study WHEAT AYII RATING ANALYSIS BASED ON ACTUAL AVERAGE AND DETRENDED YIELDS TABLE A5.2:  5 Year - Average Yield excluding min and max years Detrended Yields Kehror Pacca Average Yield (Kg/Acre) 1,432 Range of Indicative Comm. Premium Kehror Pacca Range of Indicative Comm. Premium Average Yield (Kg/Acre) 1,460 Technical Technical Insured Yield Coverage Insured Yield Annual Worst Soft (15% Medium (30% Hard (50% Insured Yield Coverage Insured Yield Annual Worst Annual Soft (15% Medium Hard (50% Premium (5% Premium (5% level (Kg/Acre) Average Loss Annual Loss load) load) Load) level (Kg/Acre) Average Loss Loss load) (30% load) Load) WAL) WAL) 90% 1289 2.02% 12.94% 2.66% 3.06% 3.46% 4.00% 90% 1,314 1.04% 5.94% 1.33% 1.53% 1.73% 2.00% 80% 1146 0.21% 2.06% 0.31% 0.35% 0.40% 0.46% 80% 1,168 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 70% 1002 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 70% 1,022 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 60% 859 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 60% 876 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 50% 716 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 50% 730 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% Range of Indicative Comm. Premium Range of Indicative Comm. Premium Dunyapur Average Yield (Kg/Acre) 1,433 Dunyapur Average Yield (Kg/Acre) 1,504 Technical Technical Insured Yield Coverage Insured Yield Annual Worst Soft (15% Medium (30% Hard (50% Insured Yield Coverage Insured Yield Annual Worst Annual Soft (15% Medium Hard (50% Premium (5% Premium (5% level (Kg/Acre) Average Loss Annual Loss load) load) Load) level (Kg/Acre) Average Loss Loss load) (30% load) Load) WAL) WAL) 90% 1,290 1.98% 9.70% 2.47% 2.84% 3.21% 3.70% 90% 1,354 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 80% 1,147 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 80% 1,203 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 70% 1,003 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 70% 1,053 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 60% 860 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 60% 902 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 50% 717 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 50% 752 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% Range of Indicative Comm. Premium Range of Indicative Comm. Premium Lodran Average Yield (Kg/Acre) 1,521 Lodran Average Yield (Kg/Acre) 1,639 Technical Technical Insured Yield Coverage Insured Yield Annual Worst Soft (15% Medium (30% Hard (50% Insured Yield Coverage Insured Yield Annual Worst Annual Soft (15% Medium Hard (50% Premium (5% Premium (5% level (Kg/Acre) Average Loss Annual Loss load) load) Load) level (Kg/Acre) Average Loss Loss load) (30% load) Load) WAL) WAL) 90% 3.48% 21.55% 4.56% 5.25% 5.93% 6.84% 90% 1,475 0.46% 4.63% 0.69% 0.80% 0.90% 1.04% 80% 1.17% 11.74% 1.76% 2.03% 2.29% 2.64% 80% 1,311 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 70% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 70% 1,147 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 60% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 60% 983 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 50% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 50% 820 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 117 For cotton, there is no clear trend toward increasing high temperatures, and pest attacks led to a major reduc- yields over time in the three tehsils (Table A5.3, Fig- tion in cotton production and yields in Punjab Province. ure  A5.2). Average yields per acre have been severely Table A5.4 shows the AYII rating analysis results with reduced in two years out of ten, however—namely in average and detrended cotton yields for the three tehsils. 2009–10 and again in 2015–16. In 2015–16 excess rain, TABLE A5.3: COTTON CULTIVATED AREA AND AVERAGE YIELDS, 2007–08 TO 2016–17 (a) Cultivated area by tehsil (acres) (b) Average yield (kg/acre) FIGURE A5.2: COTTON CULTIVATED AREA AND AVERAGE YIELDS, 2007–08 TO 2016–17 (a) Cultivated area by tehsil (acres) (b) Average yield (kg/acre) 250,000 1,200 200,000 1,000 150,000 800 600 100,000 400 50,000 200 0 0 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 –0 –0 –1 –1 –1 –1 –1 –1 –1 –1 –0 –0 –1 –1 –1 –1 –1 –1 –1 –1 07 08 09 10 11 12 13 14 15 16 07 08 09 10 11 12 13 14 15 16 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 Kehror Pacca Dunyapur Lodhran Kehror Pacca Dunyapur Lodhran 118 A Feasibility Study TABLE A5.4: COTTON AYII RATING ANALYSIS BASED ON ACTUAL AVERAGE AND DETRENDED YIELDS 5 Year - Average Yield excluding min and max years Detrended Yields Kehror Pacca Average Yield (Kg/Acre) 737 Range of Indicative Comm. Premium Kehror Pacca Average Yield (Kg/Acre) 677 Range of Indicative Comm. Premium Technical Technical Insured Yield Insured Yield Annual Average Worst Annual Soft (15% Medium Hard (50% Insured Yield Insured Yield Annual Average Worst Annual Soft (15% Medium Hard (50% Premium (5% Premium (5% Coverage level (Kg/Acre) Loss Loss load) (30% load) Load) Coverage level (Kg/Acre) Loss Loss load) (30% load) Load) WAL) WAL) 90% 663 8.87% 53.15% 11.53% 13.26% 14.99% 17.30% 90% 609 7.53% 49.70% 10.02% 11.52% 13.02% 15.03% 80% 590 6.12% 47.30% 8.48% 9.76% 11.03% 12.73% 80% 542 5.24% 43.41% 7.41% 8.52% 9.63% 11.11% 70% 516 4.10% 39.77% 6.09% 7.00% 7.92% 9.14% 70% 474 3.53% 35.33% 5.30% 6.09% 6.89% 7.95% 60% 442 2.97% 29.73% 4.46% 5.13% 5.80% 6.69% 60% 406 2.45% 24.55% 3.68% 4.23% 4.79% 5.52% 50% 369 1.57% 15.68% 2.35% 2.70% 3.06% 3.53% 50% 339 0.95% 9.46% 1.42% 1.63% 1.84% 2.13% Dunyapur Average Yield (Kg/Acre) 768 Range of Indicative Comm. Premium Dunyapur Average Yield (Kg/Acre) 758 Range of Indicative Comm. Premium Technical Technical Insured Yield Insured Yield Annual Average Worst Annual Soft (15% Medium Hard (50% Insured Yield Insured Yield Annual Average Worst Annual Soft (15% Medium Hard (50% Premium (5% Premium (5% Coverage level (Kg/Acre) Loss Loss load) (30% load) Load) Coverage level (Kg/Acre) Loss Loss load) (30% load) Load) WAL) WAL) 90% 691 13.42% 48.83% 15.86% 18.24% 20.61% 23.79% 90% 682 9.08% 44.90% 11.33% 13.02% 14.72% 16.99% 80% 614 10.41% 42.43% 12.54% 14.42% 16.30% 18.80% 80% 606 6.47% 38.02% 8.37% 9.62% 10.88% 12.55% 70% 538 7.62% 34.21% 9.33% 10.73% 12.12% 13.99% 70% 531 4.05% 29.16% 5.51% 6.33% 7.16% 8.26% 60% 461 4.06% 23.25% 5.22% 6.00% 6.79% 7.83% 60% 455 1.74% 17.36% 2.60% 2.99% 3.38% 3.90% 50% 384 0.87% 7.89% 1.26% 1.45% 1.64% 1.90% 50% 379 0.08% 0.83% 0.12% 0.14% 0.16% 0.19% Lodran Average Yield (Kg/Acre) 846 Range of Indicative Comm. Premium Lodran Average Yield (Kg/Acre) 840 Range of Indicative Comm. Premium Technical Technical Insured Yield Insured Yield Annual Average Worst Annual Soft (15% Medium Hard (50% Insured Yield Insured Yield Annual Average Worst Annual Soft (15% Medium Hard (50% Premium (5% Premium (5% Coverage level (Kg/Acre) Loss Loss load) (30% load) Load) Coverage level (Kg/Acre) Loss Loss load) (30% load) Load) WAL) WAL) 90% 761 13.00% 53.03% 15.65% 18.00% 20.35% 23.48% 90% 756 7.35% 48.76% 9.79% 11.26% 12.73% 14.68% 80% 677 9.47% 47.16% 11.82% 13.60% 15.37% 17.74% 80% 672 5.34% 42.36% 7.46% 8.58% 9.70% 11.19% 70% 592 6.39% 39.61% 8.37% 9.63% 10.88% 12.56% 70% 588 3.41% 34.13% 5.12% 5.89% 6.65% 7.68% 60% 508 4.12% 29.55% 5.60% 6.44% 7.28% 8.40% 60% 504 2.31% 23.15% 3.47% 3.99% 4.51% 5.21% 50% 423 1.55% 15.46% 2.32% 2.67% 3.01% 3.48% 50% 420 0.78% 7.78% 1.17% 1.34% 1.52% 1.75% Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 119 For rice, as shown in Table A5.5 and Figure A5.3, there is a slight gradual trend of increasing average annual yields in all three tehsils over the past ten years. Table A5.6 shows the AYII rating analysis results with average and detrended rice yields for the three tehsils. TABLE A5.5: RICE CULTIVATED AREA AND AVERAGE YIELDS, 2007–08 TO 2016–17 (a) Cultivated area by tehsil (acres) (b) Average yield (kg/acre) FIGURE A5.3: RICE CULTIVATED AREA AND AVERAGE YIELDS, 2007–08 TO 2016–17 (a) Cultivated area by tehsil (acres) (b) Average yield (kg/acre) 2,500 25,000 2,000 20,000 1,500 15,000 y = 47.927x + 1083.8 1,000 R² = 0.3192 10,000 500 5,000 0 0 08 09 0 1 2 3 4 5 6 7 –1 –1 –1 –1 –1 –1 –1 –1 – – 07 08 09 10 11 12 13 14 15 16 08 09 0 1 2 3 4 5 6 7 –1 –1 –1 –1 –1 –1 –1 –1 20 20 20 20 20 20 20 20 20 20 – – 07 08 09 10 11 12 13 14 15 16 20 20 20 20 20 20 20 20 20 20 Kehror Pacca Dunyapur Lodhran Kehror Pacca Dunyapur Lodhran Linear (Dunyapur) 120 A Feasibility Study TABLE A5.6: RICE AYII RATING ANALYSIS BASED ON ACTUAL AVERAGE AND DETRENDED YIELDS 5 Year - Average Yield excluding min and max years Detrended Yields Kehror Pacca Average Yield (Kg/ Acre) 1,434 Range of Indicative Comm. Premium Kehror Pacca Average Yield (Kg/ Acre) 1,506 Range of Indicative Comm. Premium Technical Technical Insured Yield Coverage Insured Yield Annual Average Worst Soft (15% Medium (30% Hard (50% Insured Yield Insured Yield Annual Average Worst Annual Soft (15% Medium Hard (50% Premium (5% Premium (5% level (Kg/ Acre) Loss Annual Loss load) load) Load) Coverage level (Kg/ Acre) Loss Loss load) (30% load) Load) WAL) WAL) 90% 1,291 5.96% 20.80% 7.00% 8.05% 9.09% 10.49% 90% 1,355 0.67% 4.23% 0.88% 1.01% 1.14% 1.32% 80% 1,147 1.95% 10.90% 2.50% 2.87% 3.24% 3.74% 80% 1,205 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 70% 1,004 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 70% 1,054 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 60% 860 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 60% 904 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 50% 717 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 50% 753 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% Dunyapur Average Yield (Kg/ Acre) 1,422 Range of Indicative Comm. Premium Dunyapur Average Yield (Kg/ Acre) 1,503 Range of Indicative Comm. Premium Technical Technical Insured Yield Coverage Insured Yield Annual Average Worst Soft (15% Medium (30% Hard (50% Insured Yield Insured Yield Annual Average Worst Annual Soft (15% Medium Hard (50% Premium (5% Premium (5% level (Kg/ Acre) Loss Annual Loss load) load) Load) Coverage level (Kg/ Acre) Loss Loss load) (30% load) Load) WAL) WAL) 90% 1,291 8.27% 25.97% 9.57% 11.00% 12.44% 14.35% 90% 1,355 2.01% 11.16% 2.57% 2.95% 3.34% 3.85% 80% 1,147 4.63% 16.72% 5.47% 6.29% 7.11% 8.20% 80% 1,205 0.01% 0.06% 0.01% 0.01% 0.01% 0.01% 70% 1,004 1.00% 4.82% 1.25% 1.43% 1.62% 1.87% 70% 1,054 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 60% 860 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 60% 904 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 50% 717 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 50% 753 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% Lodran Average Yield (Kg/ Acre) 1,428 Range of Indicative Comm. Premium Lodran Average Yield (Kg/ Acre) 1,599 Range of Indicative Comm. Premium Technical Technical Insured Yield Coverage Insured Yield Annual Average Worst Soft (15% Medium (30% Hard (50% Insured Yield Insured Yield Annual Average Worst Annual Soft (15% Medium Hard (50% Premium (5% Premium (5% level (Kg/ Acre) Loss Annual Loss load) load) Load) Coverage level (Kg/ Acre) Loss Loss load) (30% load) Load) WAL) WAL) 90% 1,291 4.95% 18.83% 5.89% 6.77% 7.65% 8.83% 90% 1,355 1.21% 10.59% 1.74% 2.00% 2.26% 2.61% 80% 1,147 1.18% 8.68% 1.61% 1.86% 2.10% 2.42% 80% 1,205 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 70% 1,004 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 70% 1,054 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 60% 860 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 60% 904 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 50% 717 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 50% 753 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 121 For maize, the cultivated area increased significantly programs for balanced fertilizer use and improved crop in 2016–17, especially in Kehror Pacca Tehsil. Aver- management practices. age maize yields show a marked increasing trend since 2014–15, with an average increase over this period of ­ As shown in the examples for maize grown in Dunyapur about 1  metric ton per acre (Table A5.7, Figure A5.4). (Figures A5.5 and A5.6), it is extremely important to According to DoA-GoPunjab, the increase is the result detrend yields for crop area yield ratings. Table A5.8 of higher yields obtained from introducing new hybrid shows the AYII rating analysis results with average and maize varieties, coupled with government support detrended maize yields for the three tehsils. TABLE A5.7: MAIZE CULTIVATED AREA AND AVERAGE YIELDS, 2007–08 TO 2016–17 (a) Cultivated area by tehsil (acres) (b) Average yield (kg/acre) 122 A Feasibility Study Dunyapur Maize: Historical Burn Analysis Actual yields would have been below the 80 percent insured yield in five out of 10 years (2007–08, 2008–09, Based on Actual 10-year Yields 2009–20, 2010–11, and 2013–14), equal to an annual Average yield in 3 out of last 5 years = 2,246 kg/acre average yield shortfall over 10 years of 10.07 percent of 80 percent yield coverage, insured yield = 1,797 kg/acre the insured yield. FIGURE A5.4: MAIZE CULTIVATED AREA AND AVERAGE YIELDS, 2007–08 TO 2016–17 (a) Cultivated area by tehsil (acres) (b) Average yield (kg/acre) 60,000 50,000 3,500 y = 125.42x + 1199.4 40,000 R² = 0.3991 3,000 2,500 30,000 y = 139.35x + 1294 2,000 R² = 0.565 1,500 20,000 1,000 y = 157.56x + 911.4 500 R² = 0.8032 10,000 0 8 9 0 1 2 3 4 5 6 7 –0 –0 –1 –1 –1 –1 –1 –1 –1 –1 0 07 08 09 10 11 12 13 14 15 16 20 20 20 20 20 20 20 20 20 20 8 9 0 1 2 3 4 5 6 7 –0 –0 –1 –1 –1 –1 –1 –1 –1 –1 07 08 09 10 11 12 13 14 15 16 Kehror Pacca Dunyapur 20 20 20 20 20 20 20 20 20 20 Lodhran Linear (Kehror Pacca) Kehror Pacca Dunyapur Lodhran Linear (Dunyapur) Linear (Lodhran) FIGURE A5.5: MAIZE YIELD AND SHORTFALL BASED ON ACTUAL 10-YEAR YIELDS 3,500 0 3,000 3,106 0 Yield and shortfall (kg/ha) 2,500 2,638 0 0 2,000 0 79 2,084 2,016 305 1,872 1,500 1,718 486 512 427 1,492 1,310 1,285 1,370 1,000 500 0 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 Actual yield (kg/ha) Yield shortfall (kg/ha) Insured yield (kg/ha) Historical performance of key variables Year 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 Insured yield (kg/ha) 1,797 1,797 1,797 1,797 1,797 1,797 1,797 1,797 1,797 1,797 1,797 1,797 Actual yield (kg/ha) 1,718 1,492 1,310 1,285 1,872 2,016 1,370 3,106 2,638 2,084 Average Yield shortfall (kg/ha) 79 305 486 512 0 0 427 0 0 0 shortfall Yield shortfall (% of ins. Y) 4.4% 17.0% 27.1% 28.5% 0.0% 0.0% 23.7% 0.0% 0.0% 0.0% 10.07% Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 123 Dunyapur Maize: Historical Burn Analysis Detrended yields would have been below the 80 per- cent insured yield in only one out of 10 years (2013–14), Based on Detrended 10-year Yields equal to an annual average yield shortfall over 10 years of Average detrended yield = 2,579 kg/acre 0.93 percent of the insured yield. 80 percent yield coverage, insured yield = 2,063 kg/acre FIGURE A5.6: MAIZE YIELD AND SHORTFALL BASED ON DETRENDED 10-YEAR YIELDS 4,000 0 3,500 3,482 Yield and shortfall (kg/ha) 0 0 3,000 0 0 0 2,972 2,889 2,500 0 0 2,621 0 2,624 2,643 2,000 2,314 191 2,163 2,210 1,872 1,500 1,000 500 0 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 Detrended yield (kg/ha) Yield shortfall (kg/ha) Insured yield (kg/ha) Historical performance of key variables Year 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 Insured yield (kg/ha) 2,063 2,063 2,063 2,063 2,063 2,063 2,063 2,063 2,063 2,063 2,063 2,063 Detrended yield (kg/ha) 2,972 2,621 2,314 2,163 2,624 2,643 1,872 3,482 2,889 2,210 Average Yield shortfall (kg/ha) 0 0 0 0 0 0 191 0 0 0 shortfall Yield shortfall (% of Ins. Y) 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 9.3% 0.0% 0.0% 0.0% 0.93% 124 A Feasibility Study TABLE A5.8: MAIZE AYII RATING ANALYSIS BASED ON ACTUAL AVERAGE AND DETRENDED YIELDS 5 Year - Average Yield excluding min and max years Detrended Yields Kehror Pacca Average Yield (Kg/ Acre) 2,447 Range of Indicative Comm. Premium Kehror Pacca Average Yield (Kg/ Acre) 2,827 Range of Indicative Comm. Premium Technical Technical Insured Yield Insured Yield Annual Worst Annual Soft (15% Medium Hard (50% Insured Yield Insured Yield Annual Worst Soft (15% Medium Hard (50% Premium (5% Premium Coverage level (Kg/ Acre) Average Loss Loss load) (30% load) Load) Coverage level (Kg/ Acre) Average Loss Annual Loss load) (30% load) Load) WAL) (5% WAL) 90% 2,202 15.01% 40.63% 17.05% 19.60% 22.16% 25.57% 90% 2,544 1.69% 11.09% 2.25% 2.59% 2.92% 3.37% 80% 1,958 8.14% 33.21% 9.80% 11.27% 12.74% 14.70% 80% 2,262 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 70% 1,713 2.51% 23.67% 3.69% 4.25% 4.80% 5.54% 70% 1,979 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 60% 1,468 1.10% 10.95% 1.64% 1.89% 2.14% 2.46% 60% 1,696 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 50% 1,224 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 50% 1,414 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% Dunyapur Average Yield (Kg/ Acre) 2,246 Range of Indicative Comm. Premium Dunyapur Average Yield (Kg/ Acre) 2,579 Range of Indicative Comm. Premium Technical Technical Insured Yield Insured Yield Annual Worst Annual Soft (15% Medium Hard (50% Insured Yield Insured Yield Annual Worst Soft (15% Medium Hard (50% Premium (5% Premium Coverage level (Kg/ Acre) Average Loss Loss load) (30% load) Load) Coverage level (Kg/ Acre) Average Loss Annual Loss load) (30% load) Load) WAL) (5% WAL) 90% 2,021 15.27% 36.43% 17.09% 19.66% 22.22% 25.64% 90% 2,321 3.13% 19.36% 4.10% 4.71% 5.33% 6.14% 80% 1,797 10.07% 28.49% 11.49% 13.21% 14.94% 17.23% 80% 2,063 0.93% 9.27% 1.39% 1.60% 1.81% 2.09% 70% 1,572 5.29% 18.27% 6.20% 7.13% 8.06% 9.30% 70% 1,805 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 60% 1,348 0.74% 4.65% 0.97% 1.12% 1.27% 1.46% 60% 1,547 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 50% 1,123 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 50% 1,290 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% Lodran Average Yield (Kg/ Acre) 2,093 Range of Indicative Comm. Premium Lodran Average Yield (Kg/ Acre) 2,645 Range of Indicative Comm. Premium Technical Technical Insured Yield Insured Yield Annual Worst Annual Soft (15% Medium Hard (50% Insured Yield Insured Yield Annual Worst Soft (15% Medium Hard (50% Premium (5% Premium Coverage level (Kg/ Acre) Average Loss Loss load) (30% load) Load) Coverage level (Kg/ Acre) Average Loss Annual Loss load) (30% load) Load) WAL) (5% WAL) 90% 1,884 14.21% 39.50% 16.19% 18.62% 21.04% 24.28% 90% 2,381 0.66% 6.64% 1.00% 1.15% 1.29% 1.49% 80% 1,674 9.03% 31.94% 10.63% 12.23% 13.82% 15.95% 80% 2,116 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 70% 1,465 5.47% 22.21% 6.58% 7.57% 8.56% 9.88% 70% 1,852 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 60% 1,256 1.85% 9.25% 2.31% 2.66% 3.01% 3.47% 60% 1,587 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 50% 1,047 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 50% 1,323 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 125 In sugarcane, there is a small trend toward increasing yields per acre over the past 10 years (Table A5.9, Fig- ure A5.7). Table A5.10 shows the AYII rating analysis results with average and detrended sugarcane yields for the three tehsils. TABLE A5.9: SUGARCANE CULTIVATED AREA AND AVERAGE YIELDS, 2007–08 TO 2016–17 (a) Cultivated area by tehsil (acres) (b) Average yield (kg/acre) FIGURE A5.7: SUGARCANE CULTIVATED AREA AND AVERAGE YIELDS, 2007–08 TO 2016–17 (a) Cultivated area by tehsil (acres) (b) Average yield (kg/acre) 35,000 y = 770.58x + 20368 6,000 30,000 y = 592.12x + 22244 R² = 0.643 R² = 0.4169 5,000 25,000 20,000 4,000 15,000 y = 332.88x + 19232 R² = 0.1444 3,000 10,000 2,000 5,000 1,000 0 8 9 0 1 2 3 4 5 6 7 –0 –0 –1 –1 –1 –1 –1 –1 –1 –1 0 07 08 09 10 11 12 13 14 15 16 20 20 20 20 20 20 20 20 20 20 08 09 10 11 12 13 14 15 16 17 7– 8– 9– 0– 1– 2– 3– 4– 5– 6– 0 0 0 1 1 1 1 1 1 1 Kehror Pacca Dunyapur 20 20 20 20 20 20 20 20 20 20 Lodhran Linear (Kehror Pacca) Kehror Pacca Dunyapur Lodhran Linear (Dunyapur) Linear (Lodhran) 126 A Feasibility Study TABLE A5.10: SUGARCANE AYII RATING ANALYSIS BASED ON ACTUAL AVERAGE AND DETRENDED YIELDS 5 Year - Average Yield excluding min and max years Detrended Yields Range of Indicative Comm. Range of Indicative Comm. Kehror Pacca Average Yield (Kg/ Acre) 27,463 Kehror Pacca Average Yield (Kg/ Acre) 28,884 Premium Premium Technical Worst Technical Medium Insured Yield Insured Yield Annual Worst Annual Soft (15% Medium Hard (50% Insured Yield Insured Yield Annual Soft (15% Hard (50% Premium (5% Annual Premium (5% (30% Coverage level (Kg/ Acre) Average Loss Loss load) (30% load) Load) Coverage level (Kg/ Acre) Average Loss load) Load) WAL) Loss WAL) load) 90% 24,717 5.30% 15.66% 6.08% 6.99% 7.90% 9.12% 90% 25,996 0.13% 1.25% 0.19% 0.22% 0.24% 0.28% 80% 21,970 0.95% 5.12% 1.20% 1.39% 1.57% 1.81% 80% 23,107 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 70% 19,224 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 70% 20,219 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 60% 16,478 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 60% 17,330 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 50% 13,732 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 50% 14,442 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% Range of Indicative Comm. Range of Indicative Comm. Dunyapur Average Yield (Kg/ Acre) 23,100 Dunyapur Average Yield (Kg/ Acre) 22,893 Premium Premium Technical Worst Technical Medium Insured Yield Insured Yield Annual Worst Annual Soft (15% Medium Hard (50% Insured Yield Insured Yield Annual Soft (15% Hard (50% Premium (5% Annual Premium (5% (30% Coverage level (Kg/ Acre) Average Loss Loss load) (30% load) Load) Coverage level (Kg/ Acre) Average Loss load) Load) WAL) Loss WAL) load) 90% 20,790 4.44% 18.09% 5.35% 6.15% 6.95% 8.02% 90% 20,604 0.91% 4.43% 1.13% 1.30% 1.47% 1.70% 80% 18,480 0.79% 7.85% 1.18% 1.35% 1.53% 1.77% 80% 18,314 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 70% 16,170 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 70% 16,025 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 60% 13,860 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 60% 13,736 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 50% 11,550 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 50% 11,447 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% Range of Indicative Comm. Range of Indicative Comm. Lodran Average Yield (Kg/ Acre) 27,232 Lodran Average Yield (Kg/ Acre) 28,758 Premium Premium Technical Worst Technical Medium Insured Yield Insured Yield Annual Worst Annual Soft (15% Medium Hard (50% Insured Yield Insured Yield Annual Soft (15% Hard (50% Premium (5% Annual Premium (5% (30% Coverage level (Kg/ Acre) Average Loss Loss load) (30% load) Load) Coverage level (Kg/ Acre) Average Loss load) Load) WAL) Loss WAL) load) 90% 24,509 2.46% 18.05% 3.37% 3.87% 4.38% 5.05% 90% 25,882 0.61% 4.33% 0.83% 0.96% 1.08% 1.25% 80% 21,786 0.78% 7.80% 1.17% 1.35% 1.52% 1.76% 80% 23,006 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 70% 19,062 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 70% 20,131 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 60% 16,339 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 60% 17,255 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 50% 13,616 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 50% 14,379 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 127 ANNEX 6 THE CADENA PROGRAM IN MEXICO: STATE-LEVEL CATASTROPHE INSURANCE AS A SAFETY NET FOR SMALLHOLDERS ORIGINS OF A NATIONAL SAFETY NET PROGRAM FOR POOR FARMERS Mexico is unique in having a national- and state-level parametric insur- ance program (Seguro Catastrófico Agropecuario, SAC) designed specifi- cally to provide social safety net protection for the large numbers of small, semisubsistence farming households in rural areas who experience cli- mate-induced catastrophes but are below the threshold of insurability by the commercial sector.57 In 2003, Mexico was the first country in the world to recognize the potential of replacing traditional ad-hoc post disaster relief schemes with formal parametric crop and livestock insurance solutions at the state level. Since 1995, the federal and state governments had operated an ex-post national scheme under FONDEN,58 a program that provided financial compensation to small rural farming families who had been affected by natural disasters but were not eligible for private crop and livestock insur- ance. Between 1995 and 2003, the federal and state governments paid out US$212 million and US$74 million, respectively, in direct support payments to small rural farmers under FONDEN. In 2003, under FAPRACC59—a fund to support rural peo- ple affected by climatic events—the government contracted Agroasemex, the national agricultural reinsurer, to substitute the ex-post disaster compensation programs with an ex-ante macro-level index insurance for catastrophic climatic perils (Agroasemex 2006). 57Thisannex draws extensively on two reports about CADENA to which the author was a major contributor: World Bank (2013); Arias et al. (2014). 58Fondo de Desastres Naturales (Natural Disaster Fund). 59Fondo para Atender a la Población Rural Afectada por Contingencias Climatológicas (Support Fund for Rural People Affected by Climatic Events). Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 129 In 2003, Agroasemex designed the world’s first macro-level drought index insurance coverage CADENA EVOLUTION for rainfed cereals grown in Guanajuato State, AND FUNDING and since then its range of parametric crop and livestock insurance products for catastrophic Since inception in 2003, the CADENA-SAC pro- perils has expanded into nearly all other states gram has expanded hugely. By 2011, approximately of Mexico. Since 2004, private insurance companies 8 million hectares of crops were insured in 27 states, ben- have also actively provided traditional nonparametric efiting over 2.5 million insured farmers who represented catastrophe crop and livestock coverages to the state gov- about 56 percent of the target group (4.5 million sub- ernments. In 2008, FAPRACC was replaced by PACC,60 sistence farmers operating on 16.5 million hectares). In which operated for three years before being superseded in addition, more than 4.2 million head of livestock were 2011 by CADENA—Componente Atención a Desastres insured throughout Mexico under the livestock catastro- Naturales (the Natural Disaster Response Component of phe program in 2011. In the same year, CADENA crop SAGARPA, which is the Ministry of Agriculture, Live- and livestock insurance programs covered 2,362 munic- stock, Rural Development, Fisheries, and Food). Today ipalities in 30 of Mexico’s 31 states,61 with a total pre- CADENA contains two main elements: (1) the SAC (Cat- mium income of more than MXN 1.5 billion and total astrophic Agricultural Insurance) programs for farmers, sum insured (TSI) of MXN 12 billion (Figure A6.1) livestock producers, aquaculture farmers, and fishers, (World Bank 2013; Arias et al. 2014). and (2) in states where SAC is not provided, continued direct support (apoyo directo) compensation payments to Governments see three key advantages in using farmers for climatic disasters. an ex-ante macro-level insurance product to finance natural disaster payments. First, for the FIGURE A6.1: MEXICO: EVOLUTION OF THE CADENA PROGRAM, 2003–11 (INSURED CROP AREA IN HECTARES AND NUMBER OF INSURED LIVESTOCK) 1 3 18 23 19 30 31 30 30 9,000,000 8,000,000 8,044,332 8,032,872 7,000,000 6,607,475 6,000,000 5,623,433 5,000,000 4,161,581 4,227,187 4,101,801 3,461,305 4,000,000 2,412,991 2,260,991 3,000,000 1,487,175 1,240,891 2,000,000 298,483 261,987 95,415 1,000,000 0 2003 2004 2005 2006 2007 2008 2009 2010 2011 Hectares Animal units Source: World Bank 2013, based on SAGARPA data. 60Programa de Atención a Contingencias Climatológicas (Climatic Event Response Program). 61Mexico has 31 states, 1 federal district, and 2,445 municipalities. 130 A Feasibility Study payment of a pre-agreed premium, the maximum liabil- ity can be quantified in advance and transferred out of CADENA INSURANCE the fiscal budget to local and international insurance and PRODUCTS AND ELIGIBILITY reinsurance markets. Second, insurance payouts under an index program can be made very rapidly to state gov- CRITERIA ernments (and to farmers, where there is a prior registry CADENA offers two types of crop macro-level of farmers), because there is no need to assess damage index insurance. The first type is catastrophic para- in individual farmers’ fields under weather index pro- metric or WII policies, which typically use ground-based grams and a reduced need for such assessments under weather stations to insure crops against key perils such area yield-based index programs. Third, insurance as rainfall deficit (drought) or excess rain and other cat- brings transparency and standardization of payout rules astrophic climatic perils such as hurricane windspeeds, to disaster compensation payments. low temperature/freezing, and floods. The second type is AYII policies, which usually operate at the level of a CADENA is funded by the federal and state gov- municipality, agrarian nucleus, or ejido (land that was ernments and underwritten by the national rein- formerly held in common at a locality). The AYII policies surer, Agroasemex and several private sector involve actual infield sampling of crop yields to establish insurance companies. CADENA is administered by the actual average municipality yield and, if applicable, SAGARPA. The state governments separately purchase the amount of yield loss (Table A6.1). The AYII policies macro-level crop and livestock index insurance to finance are designed to insure against catastrophic yield loss at their catastrophe climatic disaster programs for poor the municipality or locality level. For each insured crop, farmers in their states. The costs of the program (includ- the insured yield is set at 30 percent of the municipality ing most importantly premium financing) are shared on a average yield using SAGARPA historical production and ratio of about 90 percent federal government and 10 per- yield data. Thus, the products respond to catastrophic cent state governments. A system of competitive annual crop losses exceeding 70 percent of expected production. tendering is used each year to appoint insurers to under- write the program. TABLE A6.1:  MEXICO: CADENA CROP AND LIVESTOCK INSURANCE PRODUCTS AND PROGRAMS Type of CADENA catastrophe Basis of insurance and insurance program indemnity Insured perils 1. Parametric crop weather index insurance Weather indexes measured at Drought, excess rain, flood, (Seguro Agrícola de Indices Climáticos, SAIC) ground stations hurricane, windstorm 2. Crop area-yield index insurance Area yields measured by infield loss Comprehensive multiple-peril (Seguro Agrícola de Índices de Producción, SAIP) assessment 3. Livestock-pasture Normalized Satellite-measured NDVI index All perils that reduce pasture growth Difference Vegetation Index (NDVI) (mainly drought) (Seguro Pecuario de Índices de Vegetación, SPIV) 4. Traditional livestock insurance Decreased forage and extraordinary Drought (Seguro Pecuario Catastrófico, SPC) weight loss in animals Source: World Bank 2013. Note: The program types listed in the table are SAGARPA’s classifications of the CADENA crop and livestock insurance programs. Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 131 For livestock, CADENA also offers two products. compensation amounts are small, but are designed to tide The first is a livestock loss of pasture/grazing policy the small farmer over until the next season. based on satellite imagery (NDVI), the second is a tradi- tional livestock coverage against loss of forage. In the event of a triggered payout on the CADENA macro-level crop and livestock insurance pro- Mexico applies strict eligibility criteria to define grams, payment is made to the state government poor farmers who are eligible for free protection as the insured, or to SAGARPA in the case where under the CADENA crop and livestock insur- the latter purchases coverage. It is then the respon- ance programs. For farmers the criteria are based on sibility of the state-level governments and SAGARPA to farm size limits and for livestock producers on the max- distribute the benefits to the farmers in the affected areas. imum number of Livestock Units owned. These crite- ria apply both for CADENA Direct Support Payments There are two systems for disbursing insurance and Catastrophe Agricultural Insurance (SAC). The CADENA claims payments. The first method is where CADENA also carries fixed sums insured/compensation the state government purchases aggregate coverage for a payments which are applied throughout the country: particular crop in a municipality, but does not pre-­ register for rain-fed annual crops the 2012 payout was a fixed the eligible farmers (beneficiaries). In the event of a loss value of MXN 1,300/ha or about US$100/ha and for being triggered, the state government receives a lump tree fruit a higher value of MXN 2,200/ha or about sum payout and then uses infield assessment to establish $175/ha. The insurance value for livestock was MXN which farmers have incurred losses and then distributes 600 per or US$45 per livestock unit (Table A6.2). These TABLE A6.2:  ELIGIBILITY CRITERIA FOR CADENA PROGRAMS (DIRECT SUPPORT AND SAC) IN 2012 Maximum Amount of Maximum Amount of Suppor t Support (DIRECT (CATAS TROPHE AGRICULTURAL Amount of Payout per CADENA component SUPPORT) INSURANCE -SAC) Unit (MXN) A. Agriculture. $1,300 pesos per hectare Up to 10 Has. per in rainfed crops I. Annual crops. Up to 10 Has. Per producer producer $2,200 pesos per hectare in irri gation cr ops II. Fr uit tr ees, coffee and prickly Up to 10 Has. per $2,200.00 pesos per Up to 10 Has. Per producer pear (nopal). producer hectare under irrigated and rainfed crops Up to 50 Animal Units in case of feeding $600 pesos per Animal supplement. Up to 50 Animal Units in case of Unit B. Livestock. feeding supplement. Up to 5 Animal Units in $1,500 pesos per Animal case of death. Unit C. Fishi ng One boat per fisherman One boat per fisherman $10,000 pesos per boat D. Fish farming. Up to 2 Has. per $8,000 pesos per I. Extensive or semi intensive. producer. Up to 2 Has. Per producer. hectare. Up to 2 Aquacultur e Up to 2 Aquaculture Units per $8,000 pesos per II. Intensive. Units per producer producer Aquaculture Unit Up to 2 Aquacultur e Up to 2 Aquaculture Units per $1,000 pesos per III. Molluscs. Units per producer producer Aquaculture Unit Source: World Bank 2013. 132 A Feasibility Study the payout accordingly to the affected farmers. The main 14 percent. Benefiting farmers do not make any con- drawback of this two-stage method of distributing pay- tributions toward the costs of the CADENA insurance outs is that conducting farm-level loss assessment is very programs (Figure A6.2). time consuming. The second method involves the a-priori registration using the PROCAMPO lists of the targeted The CADENA program has experienced two beneficiaries in each municipality and establishment of major loss years, i.e., 2009 which was the second the sum insured for each named farmer. In the event of worst drought year in 60 years62 with a loss ratio of 118 payout being triggered on the policy, each registered ben- percent, and 2011 which was both a severe drought year eficiary receives a direct payment in accordance with the and a major freeze year (one in 50-year return period) insured area of the crop (or number of insured animals). with a loss ratio of 129 percent. Over the nine-year The second method is more transparent, and timely pay- period, total claims payouts have amounted to MXN 4.1 ments can be made to each beneficiary. S ­ AGARPA is billion. The fact that the program has been able to sus- actively promoting the registration of CADENA benefi- tain such severe loss years is due to the actuarial basis of ciaries in each state and municipality and conducts sea- rating and the high premium rates charged by the insur- sonal monitoring surveys to ensure farmers are receiving ers, averaging 11.9 percent over the nine years of oper- their correct payouts in a timely fashion. ations. The long-term average loss ratio at end of 2011 was 82.1 percent which represents a break-even position after deduction of operating expenses and underwriting CADENA PREMIUMS margins (Table A6.3). AND CLAIMS AND COSTS Over the same period, the CADENA program has AND BENEFITS paid out a similar amount, or MXN 4.04 billion in direct compensation payments to resource poor Since 2003 the costs of the CADENA agricultural farmers who are not yet included under the CADENA insurance premium subsidies has been MXN catastrophe index insurance programs. (Table A6.4). 5.01 billion (about US$375 million) of which SAG- ARPA (federal government) has covered 86  per- cent and the state governments have subsidized FIGURE A6.2: CADENA COST OF PREMIUM SUBSIDIES TO STATE AND FEDERAL GOVERNMENTS (MXN ’000) 2,000,000 1,500,000 1,000,000 500,000 0 2003 2004 2005 2006 2007 2008 2009 2010 2011 State premium subsidies Federal premium subsidies Source: World Bank 2013 based on SAGARPA data. 62SAGARPA, 30 October 2012. Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 133 TABLE A6.3:  CADENA CONSOLIDATED AGRICULTURAL INSURANCE RESULTS 2003–2011 (MXN ’000) Total sum Average insured Total premium premium Total claims Loss ratio Loss cost Year (MXN ’000) (MXN ’000) rate (%) (MXN ’000) % % 2003 34,445 3,438 10.0   0.0 0.0 2004 229,134 25,896 11.3 1,001 3.9 0.4 2005 906,866 124,327 13.7 110,329 88.7 12.2 2006 1,667,406 200,875 12.0 52,941 26.4 3.2 2007 2,106,128 238,624 11.3 104,093 43.6 4.9 2008 7,617,721 839,488 11.0 311,118 37.1 4.1 2009 8,477,013 917,748 10.8 1,079,160 117.6 12.7 2010 9,025,091 1,136,499 12.6 488,000 42.9 5.4 2011 12,039,010 1,523,137 12.7 1,966,190 129.1 16.3 Total 42,102,815 5,010,031 11.9 4,112,833 82.1 9.8 Source: World Bank 2013 based on SAGARPA data. TABLE A6.4: COST OF DIRECT COMPENSATION PAYMENTS (MXN) Federal Federal State State government government government government Year (MXN) (%) (MXN) (%) Total (MXN) 2003 73,765,783 70 31,591,733 30 105,357,516 2004 195,308,915 70 83,679,029 30 278,987,944 2005 529,104,634 70 227,002,198 30 756,106,832 2006 301,407,030 70 129,176,036 30 430,583,066 2007 242,269,883 70 103,849,063 30 346,118,946 2008 87,531,603 60 58,017,640 40 145,549,243 2009 113,993,830 50 113,993,830 50 227,987,660 2010 80,697,614 50 80,697,614 50 161,395,228 2011 792,113,066 50 792,113,066 50 1,584,226,132 Total, 2003–11 2,416,192,358 60 1,620,120,209 40 4,036,312,567 Annual average 268,465,818   180,013,357   448,479,174 Source: SAGARPA, 30 October 2012. OPPORTUNITY COSTS levels of financial protection against unforeseen climatic contingencies than can be achieved OF CATASTROPHE through budgetary allocations alone. Over the past nine years, the Mexican government has expended a INSURANCE VERSUS DIRECT total of MXN 9.1 billion on a combination of direct sup- COMPENSATION PAYMENTS port payments to small farmers and insurance premium payments to the insurance sector, in return for total lia- A key advantage to federal and state govern- bility (TSI) protection valued at MXN 42.1 billion. In the ments of purchasing CADENA catastrophe underwriting year 2011, agricultural crop and livestock insurance is the ability to leverage much higher insurance premiums amounted to MXN 1.52 billion 134 A Feasibility Study against a TSI of MXN 12.0 billion and insurance pay- have compensated an average of 14 percent of the actual outs of MXN 1.97 billion. The government expensed a insured area. further MXN 1.58 billion in direct support payments for a total financial outlay of MXN 3.5 billion. Had 2011 been an even more severe loss year, the CADENA insur- ance programs would have afforded government protec- CADENA WELFARE IMPACTS tion up to MXN 12 billion (Figure A6.3). There have been several studies to measure the impacts of the CADENA program on vulnerable Another way of analyzing the cost-effectiveness farmers. Agricultural insurance has a direct effect of of the CADENA agricultural insurance programs making payouts in the event of crop failure or death of is to assume a situation in which no insurance livestock, which can help smooth consumption or ensure program was in place and to calculate the ben- sufficient resources for production in subsequent seasons. efits in terms of direct support payments that The risk reduction that this entails can have indirect could have been made using the saved premium effects on economic outcomes by altering farmers’ invest- costs. This analysis has been conducted for the crop ment decisions. Fuchs and Wolff (2010) found that the insurance programs and the results are presented in Fig- CADENA program increased small farmers maize yields ure A6.4. The gray line shows the actual insured area and rural per capita expenditure and income. De Janvry rising to 8.03 million hectares in 2011 for a premium cost (2015) found that CADENA increased the sown area of of MXN 1.33 billion. The red line shows that had no maize in the year after a payout, but did not lead to sig- insurance been in place, the MXN 1.33 billion in saved nificant increases in agricultural income. premiums could have been used in 2011 to fund direct compensation payments for 1.17 million hectares or only A recent study by Arias et al. (2014) found that 15 percent of the actual insured area. Over the full period CADENA WII reduced moderate income poverty 2003 to 2011, the saved insurance premiums could only by 1.78 percentage points, but income inequality FIGURE A6.3:  CADENA: COMPARISON OF INSURANCE COVERAGE PURCHASED (TSI) VERSUS TOTAL COST OF GOVERNMENT FINANCIAL SUPPORT (PREMIUM SUBSIDIES AND DIRECT PAYMENTS) (MXN) 14,000,000,000 12,000,000,000 10,000,000,000 8,000,000,000 6,000,000,000 4,000,000,000 2,000,000,000 0 2003 2004 2005 2006 2007 2008 2009 2010 2011 Total sum insured Total government financial support to CADENA Source: World Bank 2013. Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 135 FIGURE A6.4: HYPOTHETICAL ANALYSIS OF ALTERNATIVE USE OF CROP INSURANCE PREMIUMS TO MAKE DIRECT SUPPORT PAYMENTS 9,000,000 18 As if direct support area as % of actual insured area 8,000,000 16 Insured area/direct support area (ha) 7,000,000 14 6,000,000 12 5,000,000 10 4,000,000 8 3,000,000 6 2,000,000 4 1,000,000 2 0 0 2003 2004 2005 2006 2007 2008 2009 2010 2011 Actual insured area (ha) As if area under direct support payments (ha) As if direct support area as % of actual insured area Source: World Bank 2013. increased marginally. While CADENA WII had no and Wolff (2011) emphasize the potential unin- effect on extreme poverty, the study could not determine tended consequences of the large-scale WII pro- if this was due to ineligibility of the extreme poor under grams under CADENA. WII can create disincentives CADENA rules of operation, or because of ineffective to invest in other noninsured crops leading to potential targeting. overspecialization and monoculture. WII may generate disincentives to invest in irrigation systems because farm- In a separate study, de Janvry et al. (2016) ana- ers are insured only as long as production takes place on lyzed the effects of the CADENA insurance pay- nonirrigated land. Finally, in case of catastrophic events, outs on ex-post investment decisions and coping indemnity payments may contribute to food price infla- mechanisms, providing evidence that index insur- tion at the expense of the uninsured poor. ance can improve welfare for rural households by pro- viding resources to invest in the subsequent planting season. This finding is confirmed in an earlier study by the Autonomous University of Chapingo 2009 (cited by KEY OPERATIONAL ISSUES Arias et al. 2014) which showed that nearly 100 percent AND CHALLENGES FOR of surveyed beneficiaries had continued to remain in agricultural production following the catastrophe event CADENA due to the CADENA payouts they had received. De From an operational and implementation view- Janvry et al. concluded that the benefits of the program point, one of the main areas requiring improve- exceed the costs, even without taking into account the ment on CADENA is in the timeliness of payouts risk management effect which prevents households from reaching the targeted beneficiaries. CADENA first resorting to costly coping mechanisms, such as reducing makes payouts to the state governments, which are then consumption. responsible for distributing the payouts to affected farm- ers. This process requires speeding up: the Autonomous In contrast to the above studies that highlight the University of Chapingo 2009 study indicates that the positive consequences of CADENA WII, Fuchs average time taken post-event for beneficiaries to receive 136 A Feasibility Study their CADENA payouts was 89 days. Overall, 62.1 per- limit. And finally, surveyed beneficiaries considered the cent of surveyed farmers received their payouts between value of the CADENA payouts as too allow to cover the three and six months after the event. However, a signif- costs of their investments in their agricultural enterprises. icant 37.9 percent of payouts were received between Overall, 60 percent of respondents (and as high as 72.2 six and nine months after the event. It is important to percent of crop producers) indicated that the payouts reduce the time for the payouts to reach the beneficia- covered less than 25 percent of the amounts invested ries to no more than 90 days for all CADENA programs up to the time of loss. Overall, only 14.2 percent replied in all states. Moreover, while farmers benefiting from that the payouts exceeded 50 percent of their investment CADENA payouts fell strictly within the farmer size eli- costs. SAGARPA is trying to address this issue by increas- gibility criteria, about a third of all livestock beneficia- ing the payout levels over time. ries owned more livestock than the maximum permitted Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 137 ANNEX 7 EXAMPLES OF AGRICULTURAL INSURANCE POOL PROGRAMS IN SPAIN AND TURKEY CASE STUDY 1: AGROSEGURO POOL, SPAIN Prior to 1980, there was very limited agricultural insurance provision in Spain. In 1980, the Spanish government enacted legislation to create a national agricultural insurance program, termed the Combined Agricultural Insurance (Seguros Agrarios Combinados) program, a public-private partnership (PPP) underwritten by Agroseguro—a private coinsurance pool with a mandate to provide subsidized agricultural insurance to all of Spain’s regions and farmers on a voluntary basis. Today, Agroseguro is Europe’s largest and most comprehensive national agricultural insurance program underwriting over 200 different crop, livestock, aquaculture, and forestry programs and generating total commercial premiums of €676 million in 2012. The key forms of government support to agricultural insurance in Spain include insur- ance legislation, subsidies on agriculture insurance premiums paid by farmers/herders, coinsurance and reinsurance through the National Catastrophe Reinsurance Com- pany (Consorcio de Compensación de Seguro [CCS]), and assistance to data collection and insurance product research and development. The key parties involved in the implementation of the Spanish agricultural insurance PPP include the following (Figure A 7.1): »» National administrator: ENESA (Entidad Estatal de Seguros Agrarios—National Agricultural Insurance Agency) coordinates the system and manages resources for subsidizing insurance premiums. »» Ministry of Agriculture Food and Environment: Responsible for data coordination and information collection for new product research and develop- ment in conjunction with Agroseguro’s insurance specialists. »» Pool coinsurers companies: There are currently 28 coinsurers in the Agroseguro pool, which include both private and mutual insurance companies, including Mapfre Insurance and Reinsurance Company (Spain’s largest insur- ance company) and the Spanish public sector catastrophe reinsurer (Consorcio de Compensación de Seguros). The largest shareholder in the pool is Mapfre, with a shareholding of 20 percent; the smallest coinsurer has less than a 1 percent share in the pool. Each company’s share of annual agricultural insurance premiums Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 139 FIGURE A7.1: AGROSEGURO SPAIN: INSTITUTIONAL FRAMEWORK Regional Regional governments committees Coordinating committee Ministry of the Environment, and Rural General and Marine Affairs ENESA committee Planning and subsidies in the cost of insurance Professional Crop and Private insurers on a Manager of the organizations livestock coinsurance basis coinsurance table and farming farmers INSURED (Agrosecure) cooperatives INSURED Oversight and regulation of the insurance system Insurance Directorate General Compensation Ministry of Finance for Insurance and Consortium (CCS) Pension Funds Source: Antón et al. 2011. and liability is determined according to its per- management company, underwriting more than centage share in the pool during the underwriting 260,000 agricultural insurance policies and a fur- year. Participation in the pool is completely vol- ther 30,000 livestock, forestry, and aquaculture untary and insurance companies are permitted policies generating total commercial premiums of to join and leave the pool after completion of an €676 million in 2012. Agroseguro has a full-time underwriting campaign (year). In order to main- complement of about 75 permanent staff based in tain continuity, companies usually agree to join its headquarters in Madrid and an equal number the pool for a three-year period. based in each of the 14 autonomous regions. It has »» Managing underwriter: Agroseguro, which a general management unit, a legal department, is owned by the 28 shareholders/coinsurers, has and regional branches, as well as core operational been appointed by the coinsurers to underwrite, departments responsible for (1) product research adjust, and settle claims on their collective behalf. and development, (2) production and communica- Agroseguro started with a very small team of tion (underwriting), (3) claims administration and agricultural underwriters, claims managers, loss loss assessment, (4) administration and accounting, assessors, and office support staff; today it has and (5) organization and information technology grown into Europe’s largest agricultural insurance systems. As such, it functions as a very professional 140 A Feasibility Study commercial managing company on behalf of its replace ad hoc ex-post natural disaster relief compensa- coinsurers. Agroseguro’s internal administration tion payments by a comprehensive national agricultural and operating (A&O) costs are financed out of insurance program. Spanish farmers are not eligible earned premiums on the agricultural insurance for disaster payments for perils for which insurance is business it writes on behalf of the pool. Over the offered. For non-covered perils, ad hoc disaster payments past five years, its internal A&O expenses have are available, but only if the producer has already pur- amounted to 3.5 percent of total earned premi- chased agricultural insurance for covered perils. ums (Agroseguro 2012). »» Consorcio de Compensación de Seguro The costs to government of premium subsidies are high (CCS). The national (state) catastrophe rein- as shown in Figure A7.2. Over the past 33 years (1980 surer providing reinsurance to Agroseguro pool to 2012), the total cost of premium subsidy support by coinsurers. federal and autonomous state governments amounted to »» International commercial reinsurers: Pro- €5.98 billion or 56 percent of the total costs of premi- viders of (1) stop-loss reinsurance to pool reinsur- ums while farmers paid the remaining 44 percent of total ers on their viable line retentions and (2) multiyear premium earnings. In 2012, the total premium income catastrophe stop loss to CCS. amounted to €728.3 million and state financed premiums ENESA in conjunction with the Ministry of Agriculture, were €393 million (54 percent of the total). Food and Environment (MAFE) is responsible for devel- oping a three-year rolling agricultural insurance plan in consultation with the state governments, producer orga- The government is responsible for fixing premium sub- nizations, and Agroseguro. ENESA is also responsible sidy levels. A system of differential premium subsidies for drafting the annual implementation plan setting out applies, which provides different levels of premium sub- the premium subsidy levels that will apply to each prod- sidies for each category of crops and livestock and the uct line and program in the current year and the agreed type of insurance product (named-peril, etc.). Additional budget from the government for premium subsidies. For subsidies are provided for collectively purchased policies 2013, the approved state budget for agricultural insur- through associations, for target groups of farmers includ- ance premium subsidies was €205 million. ing young farmers, and for the contracting of multi-crop policies or multiyear coverages. Under the Spanish model, premium subsidies are used as a policy instrument to promote the widest possible volun- In 2012 Agroseguro underwrote almost 485,000 crop tary adoption of agricultural insurance by farmers and to and livestock policies with total premium volume of FIGURE A7.2: AGROSEGURO PREMIUMS AND SHARE PAID BY FARMERS AND BY THE STATE 800 70 State share of total premiums (%) 700 60 Premiums (euro million) 600 50 500 40 400 30 300 20 200 100 10 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 State contributions Farmer contributions State share (%) Source: Agroseguro 2013. Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 141 €728 million. The year 2012 was a very severe year for Prior to the formation of the Tarsim pool in Turkey in frosts, drought, and hail, and total claims amounted to 2005, only 0.5 percent of agricultural areas in Turkey €800 million equivalent to a loss ratio of 118.5 percent were insured (Bora 2010). A number of private insur- (Agroseguro 2013). ance companies provided limited crop and greenhouse insurance mainly against hail, and livestock insurance Agroseguro currently underwrites about 200 viable was poorly developed. The agricultural insurance market and experimental crops, livestock, and marine aquacul- was fragmented; the system operated with limited data ture lines, and forestry insurance covering a wide range on which to design and rate products and programs; and of crop types including cereals, oilseeds, horticultural there was inadequate actuarial expertise, a lack of trans- crops, leaf and fibers, tree fruits and vines, and livestock parency and underfunded research, coordination, and types. The company offers a comprehensive range of monitoring. At the time, the Turkish government did not ­ single-peril hail, named-peril, and multi-peril crop insur- support agricultural insurance but rather provided lim- ance policies. Agroseguro only underwrites two index ited ex-post ad hoc disaster relief to crop and livestock insurance coverages: one for bees and the other a live- producers after a catastrophic loss event. The Tarsim PPP stock NDVI pasture-drought index policy. In 2012, the initiative was promoted to overcome these constraints company retained a national network of 397 crop loss and to create a modern national agricultural insurance adjusters and 123 livestock veterinary inspectors. capability (Bora 2010). Turkey elected to model its new system on the Spanish pool structure with centralized underwriting claims handling and reinsurance purchas- Agroseguro has traditionally purchased stop-loss reinsur- ing (see Box A7.1 for further details). ance protection from the national catastrophe reinsurer, CCS. There are different reinsurance agreements in place for the different insurance lines of A, B, and C according Tarsim’s pool operating company is a joint stock com- to the perils insured and their degree of catastrophe loss pany owned by the 24 insurance companies that partic- potential. The reinsurance protection provided by CCS ipated in the agricultural insurance pool in 2012, each has been a major factor in the financial viability of the with an equal shareholding. As in Spain, the role of Agroseguro pool program over the past 33 years. Tra- each pool insurance company is two-fold: (1) to market ditionally, CCS has purchased multiyear Stop-Loss Ret- Tarsim’s standard policies at approved rates to Turkish rocession protection on its liability. The individual pool crop, greenhouse, livestock, poultry, and aquaculture coinsurers have also been permitted to purchase addi- producers, and (2) to provide insurance capacity to the tional stop-loss reinsurance protection on their retentions pool. The pool operating company is responsible for all from international reinsurers. underwriting and claims management and IT systems and procedures. Tarsim reports to a management board comprised of two representatives from the Ministry of Agriculture; two representatives from the Under Secre- CASE STUDY 2: TARSIM tariat of the Treasury; and one member from each of the insurance and reinsurance companies of Turkey, the POOL, TURKEY union of chambers of agriculture, and from Tarsim (the The Turkish agricultural insurance pool, Tarsim, was operating company) (Figure A7.3). established by Law No. 5365 in 2005. The law covers the establishment of the pool, the risks to be insured, the Under the PPP for agricultural insurance, the Turkish gov- pool’s income and expenses, government support in the ernment provides Tarsim with 50 percent premium subsi- form of premium subsidies and excess of loss reinsurance dies on all classes of agricultural insurance, except for crop support, insurance contracts, the contracting of rein- policies which also include coverage against frost for which surance, and the principle duties of the pool operating the subsidy level is 66 percent. In addition, the government company and the coinsuring members. Additional legis- provides catastrophe excess of loss (stop-loss) protection lation that governs Tarsim’s operations is defined by the to Tarsim. Other benefits include subsidies on Tarsim’s Regulation of the Application of the Agricultural Insur- administration and operating expenses and loss adjustment ance (No. 26172, 18 May 2006) and the Agricultural costs, and sales tax exemption for agricultural insurance Insurance Pool Operating Procedures and the Principles premiums (Mahul & Stutley 2010). Tarsim is responsible of the Agricultural Insurance Regulations (No. 26172, for deciding its risk retention and reinsurance strategy. The 18 May 2006). 142 A Feasibility Study BOX A7.1: OBJECTIVES OF THE TARSIM POOL »» To contribute to the development and generalization of agricultural insurance. »» To provide standard insurance contracts covering the risks falling within the scope of the Act. »» To centralize and standardize loss adjustment activities. »» To have claims processed quickly and paid fairly by a central entity. »» To lay down procedures and principles for the operation of agricultural insurance. »» To provide insurance coverage for catastrophe risks like drought, frost, etc., that could overwhelm an individual insurance company. »» To expand reinsurance capacity and coverage by introducing incentives for participation in reinsurance. »» To make effective, joint use of insurance companies’ information, and human and financial resources. »» To make effective use of government subsidies and the government’s catastrophe stop-loss protection. »» To prevent unfair price competition. »» To encourage participation in insurance. Source: Bora 2010. FIGURE A7.3: TURKEY: INSTITUTIONAL FRAMEWORK OF THE TARSIM AGRICULTURAL INSURANCE POOL State Nongovernmental (4) organizations (1) Republic of Turkey Ministry of Food, Turkish Prime Ministry Union of Turkish Insurance Agriculture and Under Secretariat of Agricultural Chambers Association of Turkey Livestock the Treasury (1) (1) (2) (2) Agricultural Insurance Pool’s Board of Directors (7) Agricultural Insurance Insurance Private sector companies Pool (1) (24) (TARSIM) Operating company (1) Source: Tarsim Annual Report 2012. Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 143 FIGURE A7.4: TARSIM AGRICULTURAL INSURANCE POOL: RISK TRANSFER MECHANISMS Turkish government Premium Stop-loss subsidy Retrocession Insurance TARSIM Reinsurance companies (agricultural insurance pool) companies Premiums Premiums Indemnifications Farmers Source: SwissRe: Sigma No1/2013. FIGURE A7.5: TARSIM: GROWTH IN NUMBER OF POLICIES SOLD, TSI AND PREMIUM INCOME, 2007–2012 Total sum insured (TSI) and premiums (TL ’000) 6,000,000 800,000 700,000 5,000,000 Number of policies insured 600,000 4,000,000 500,000 3,000,000 400,000 300,000 2,000,000 200,000 1,000,000 100,000 0 0 2006 2007 2008 2009 2010 2011 2012 2013 TSI (TL) Premium (TL) Number of policies Source: Tarsim 2012 Annual Report. law permits Tarsim to retrocede business back to the pool Following the establishment of Tarsim, there has been insurers and/or to reinsure through the local reinsurer Mil- a major expansion in the demand by farmers for agri- liRe and international reinsurers (Figure A7.4). cultural insurance in Turkey. This demand has also been stimulated by the close PPP with the government which Since its formation, Tarsim has standardized all agricul- provides a minimum 50 percent premium subsidy on the tural insurance policies and tariffs, increased its range of costs of all agricultural insurance policies. In the five years product lines, and made a major investment in a web- that Tarsim has been operational, the number of policies based centralized national insurance application, under- sold has increased from 218,938 to 744,093 (an increase writing, and claims administration system. In addition, of 240 percent), and premium income has increased the pool operator has established a national crop and from Turkish Lira (TL) 47 million to TL 273 million (a livestock farm inspection and loss assessment capability, 482  percent increase) (Figure A7.5). Over this period which can draw on 536 qualified and registered crop Turkey has grown to be the third largest agricultural inspectors and 568 livestock inspectors. insurance market in Europe by premium volume. 144 A Feasibility Study ANNEX 8 POSSIBLE OPTIONS FOR COINSURANCE POOLS IN PUNJAB This annex is taken from the World Bank 2015 report “Kenya—Towards a National Crop and Livestock Insurance Program.” NON-STATUTORY COINSURANCE POOLS Insurance pools can be statutory (established by specific legislation) or non-statutory (not established by specific legislation). Different structures are commonly used to establish non-statutory insurance pools: 1) A coinsurance pool may be established by the participating insurers as an insurer in its own right, so that it is the pool itself that issues the insurance contracts and assumes the risk on behalf of the insurers. In this case, either the pool would sell its own insurance contracts or the insurers would sell insurance contracts as intermediaries (i.e., agents) on the pool company’s behalf, the risk being underwritten by the pool company. 2) The insurance contracts may be written by the insurer pool members on an individual basis, but with the risk ceded to the pool. In this case, the pool may be either (1) a special pool company established by the insurers; or (2) an arrangement between the insurers whose terms are set out in a pool agreement. 3) The insurance contracts may be written by a lead insurer on behalf of the other insurers that are members of the pool. Again, under this scenario, the pool may be a special company established by the insurers or an arrangement between the insurers set out in a pool agreement. If a coinsurance pool is established as an insurer, the pool company underwrites the risks directly in its own right. A pool company that underwrites risks must, of course, be licensed to write insurance business and must be fully capitalized as an insurer. Other coinsurance pools, whether or not established solely by contract or as a special (non-insurer) company, usually share the following features: 1) Each insurer accepts a pre-agreed share in all the risks that are covered by the pool agreement. 2) All premiums are paid into the pool, less an amount to cover expenses. Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 145 3) The pool manager or administrator assesses and provides for the governance of the pool and sets out the settles claims. pool’s functions. The legislation may also cover other 4) If there is an underwriting gain, the surplus matters, such as the provision of some form of subsidy. (beyond any reserve retained in the pool) is paid Because they are established by legislation, statutory to each insurer in accordance with its agreed pools take many forms and may be structured very differ- share. ently than a typical voluntary pool. 5) If there is an underwriting loss, the insurers con- tribute to the loss in accordance with their agreed The legislation may establish a coinsurance pool, but not share. as a corporate body. For example, the pool may be estab- lished as a contractual arrangement between participat- If a pool is established solely through a contractual ing insurers. In this case, although the legislation would arrangement, the “pool” is not a legal person and does set out the functions of the pool, those functions would not have the power to contract. The pool could not, not usually include acting as an insurer, since the pool is therefore, write insurance contracts. not a legal person. Of course, the legislation may estab- lish a corporate body to act as manager of the pool, but If the insurers enter into their own individual insurance not to write insurance contracts. contracts, the insurance business is conducted under their individual licenses. The capital of the participating insur- The legislation establishing the pool would usually pro- ers supports the risk. The position may be rather more vide the pool with exclusive rights in relation to the busi- complicated if the insurance contracts are underwritten ness underwritten by the pool. This is necessary to prevent by a lead insurer on behalf of the other insurers. non-pool insurers undermining the pool by offering sim- ilar insurance products at a lower, non-sustainable, price. It is important to appreciate that where the insurers write their own insurance contracts and cede the risk to the Statutory coinsurance pools sometimes operate as pool, each participating insurer typically accepts a pre- hybrids, with some limited reinsurance functions. agreed share of all the risks ceded to the pool, not just the risks that the insurer has written. Management of a coinsurance pool, where the pool is BENEFITS OF AN incorporated as a (noninsurance) company, involves the AGRICULTURAL INSURANCE pool company acting as the pool manager or administra- tor. Where a special pool company is not incorporated, POOL the pool may be managed by a lead insurer; by a techni- All coinsurance pools offer benefits but also have limita- cal management unit contracted or employed by, or on tions. These are summarized in Box A8.1. behalf of, the participating insurers; or by a third party such as a broker, another nonparticipating insurer, or a reinsurer. The participating insurers typically share the management costs in accordance with their proportion- INTERNATIONAL ate risk share. PRECEDENTS If a program steering committee is established to address the institutional framework for agricultural insurance, STATUTORY COINSURANCE it could consider a number of precedents: (1) the Turk- POOLS ish agricultural insurance pool (Tarsim); (2) the Spanish agricultural insurance pool (Agroseguro); and (3) the pro- Statutory insurance pools are often, but not neces- posed Mongolian Index-Based Livestock Reinsurance sarily, corporate bodies. Usually, statutory coinsurance Company (which will have features of a pool and a rein- pools are part of a national or regional program and are surance company). established as part of a PPP. Relevant legislation typically 146 A Feasibility Study BOX A8.1: BENEFITS AND LIMITATIONS OF COINSURANCE POOL ARRANGEMENTS Coinsurance pools offer these benefits: »» They achieve economies of scale through operating as a single unit with shared (pooled) administration and operating functions. These lead to costs savings from (1) reduced staffing requirements (fixed costs); (2) shared costs of product research and development, and actuarial services including rating; and (3) reduced costs of underwriting and claims control and loss adjustment. »» There are cost advantages to companies when they purchase common account (pooled) reinsurance protection rather than trying to place their own reinsurance program. The advantages arise from (1) a stron- ger negotiating position with reinsurers; (2) larger and more balanced portfolios and better spread of risk; (3) reduced costs of reinsurance due to pooled risk exposure; and (4) reduced transaction costs (reinsurance brokerage, etc.). »» There is no competition on rates in a soft market, and pools can maintain technically set rates. Most pools operate as the sole insurance provided or monopoly (as in Austria, Senegal, Spain, and Turkey, for example), and there is therefore no competition on pricing. »» Pools are able to maintain underwriting and loss adjustment standards. Under a pool monopoly arrange- ment, the pool manager can ensure that common and high standards are maintained in the underwriting of crop and livestock insurance and in the adjusting of claims. Where companies are competing against each other for standard crop insurance business, there is often a problem of varying loss adjustment standards between companies. »» Within a PPP, governments can more easily coordinate support to a pool than to individual insurers. Governments seeking to coordinate national agricultural insurance policy and planning and specific support functions (e.g., provision of premium subsidies, research and development, education and training) can work more easily with a pool than with individual insurers, each of which may have very different priorities for agricultural insurance. Coinsurance pools have these limitations: »» When a pool acts as the sole agricultural insurer, lack of competition in the market may result. This could (1) limit the range of products and services offered by the monopoly pool underwriter; (2) restrict the range of perils insured; (3) restrict the regions where agricultural insurance is offered and/or the type of farmer insured; and (4) lead to a lack of competitiveness in premium rates charged by the pool. Source: Mahul and Stutley 2010. Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 147 ANNEX 9 DETAILS REGARDING NUMBER OF INSURED FARMERS, INSURED AREA, SUM INSURED AND PREMIUM, UNDER ALTERNATIVE SCENARIOS Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 149 PUNJAB CROP INSURANCE PROGRAM: NUMBER OF INSURED FARMERS, INSURED AREA, SUM TABLE A9.1:  150 INSURED AND PREMIUM: SCENARIO 1, HIGH LEVEL OF UPTAKE PERCENT TARGET AVERAGE PREMIUM RATES: KHARIF 5.0 PERCENT AND RABI 3.5 PERCENT (US$) TOTAL Year/season Kharif 2018 Rabi 2018/19 Kharif 2019 Rabi 2019/20 Kharif 2020 Rabi 2020/21 Kharif 2021 Rabi 2021/22 Kharif 2022 Rabi 2022/23 (cumulative) Crop Program 1. AYII For Progressive Farmers linked to KISAN Credit Package (2.5 Acres to 12.5 Acres) Number of Insured Farmers 250,000 350,000 450,000 550,000 600,000 650,000 700,000 725,000 750,000 750,000 5,775,000 Insured Area per Farmer (Acres/farmer) 3.5 3.5 3.5 3.5 3.5 3.5 3.5 3.5 3.5 3.5 3.5 Total Insured Crop Area (Acres) 875,000 1,225,000 1,575,000 1,925,000 2,100,000 2,275,000 2,450,000 2,537,500 2,625,000 2,625,000 20,212,500 Sum Insured (US$ per Acre) 400 300 400 300 400 300 400 300 400 300 Total Sum Insured (US$) 350,000,000 367,500,000 630,000,000 577,500,000 840,000,000 682,500,000 980,000,000 761,250,000 1,050,000,000 787,500,000 7,026,250,000 Premium Rate (%) 5.0% 3.50% 5.0% 3.50% 5.0% 3.50% 5.0% 3.50% 5.0% 3.50% Total Premium Income(US$) 17,500,000 12,862,500 31,500,000 20,212,500 42,000,000 23,887,500 49,000,000 26,643,750 52,500,000 27,562,500 303,668,750 Crop Program 2. AYII for Subsistence Farmers <2.5 Acres Number of Insured Farmers 250,000 500,000 750,000 1,000,000 1,250,000 1,500,000 1,750,000 1,750,000 8,750,000 Insured Area per Farmer (Acres/farmer) 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 Total Insured Crop Area (Acres) 250,000 500,000 750,000 1,000,000 1,250,000 1,500,000 1,750,000 1,750,000 8,750,000 Sum Insured US$ per Acre 200 150 200 150 200 150 200 150 Total Sum Insured (US$) 50,000,000 75,000,000 150,000,000 150,000,000 250,000,000 225,000,000 350,000,000 262,500,000 1,512,500,000 Premium Rate (%) 5.0% 3.50% 5.0% 3.50% 5.0% 3.50% 5.0% 3.50% Total Premium Income(US$) 2,500,000 2,625,000 7,500,000 5,250,000 12,500,000 7,875,000 17,500,000 9,187,500 64,937,500 Crop Program 3. Tree Crops (Mango, Citrus) 2019/20 2020/21 2021/22 2022/23 Number of Insured Farmers 2,500 5,000 7,500 10,000 25,000 Insured Area per Farmer (Acres/farmer) 2.5 2.5 2.5 2.5 2.5 Total Insured Crop Area (Acres) 6,250 12,500 18,750 25,000 62,500 Sum Insured US$ per Acre 1,000 1,000 1,000 1,000 Total Sum Insured (US$) 6,250,000 12,500,000 18,750,000 25,000,000 62,500,000 Premium Rate (%) 10.00% 10.00% 10.00% 10.00% Total Premium Income US$ 625,000 1,250,000 1,875,000 2,500,000 6,250,000 TOTAL ALL CROP PROGRAMS TOTAL INSURED FARMERS 250,000 350,000 700,000 1,052,500 1,350,000 1,655,000 1,950,000 2,232,500 2,500,000 2,510,000 14,550,000 TOTAL INSURED AREA (ACRES) 875,000 1,225,000 1,825,000 2,431,250 2,850,000 3,287,500 3,700,000 4,056,250 4,375,000 4,400,000 29,025,000 TOTAL SUM INSURED (US$) 350,000,000 367,500,000 680,000,000 658,750,000 990,000,000 845,000,000 1,230,000,000 1,005,000,000 1,400,000,000 1,075,000,000 8,601,250,000 TOTAL PREMIUM INCOME (US$) 17,500,000 12,862,500 34,000,000 23,462,500 49,500,000 30,387,500 61,500,000 36,393,750 70,000,000 39,250,000 374,856,250 A Feasibility Study PUNJAB CROP INSURANCE PROGRAM: NUMBER OF INSURED FARMERS, INSURED AREA, SUM TABLE A9.2:  INSURED, PREMIUM AND PREMIUM SUBSIDIES: SCENARIO 2, HIGH LEVELS OF UPTAKE AND HIGHER TARGET AVERAGE PREMIUM RATES: KHARIF 7.5 PERCENT AND RABI 5.0 PERCENT (US$) TOTAL Year/season Kharif 2018 Rabi 2018/19 Kharif 2019 Rabi 2019/20 Kharif 2020 Rabi 2020/21 Kharif 2021 Rabi 2021/22 Kharif 2022 Rabi 2022/23 (cumulative) Crop Program 1. AYII For Progressive Farmers linked to KISAN Credit Package (2.5 Acres to 12.5 Acres) Number of Insured Farmers 250,000 350,000 450,000 550,000 600,000 650,000 700,000 725,000 750,000 750,000 5,775,000 Insured Area per Farmer (Acres/farmer) 3.5 3.5 3.5 3.5 3.5 3.5 3.5 3.5 3.5 3.5 3.5 Total Insured Crop Area (Acres) 875,000 1,225,000 1,575,000 1,925,000 2,100,000 2,275,000 2,450,000 2,537,500 2,625,000 2,625,000 20,212,500 Sum Insured (US$ per Acre) 400 300 400 300 400 300 400 300 400 300 Total Sum Insured (US$) 350,000,000 367,500,000 630,000,000 577,500,000 840,000,000 682,500,000 980,000,000 761,250,000 1,050,000,000 787,500,000 7,026,250,000 Premium Rate (%) 7.5% 5.00% 7.5% 5.00% 7.5% 5.00% 7.5% 5.00% 7.5% 5.00% Total Premium Income(US$) 26,250,000 18,375,000 47,250,000 28,875,000 63,000,000 34,125,000 73,500,000 38,062,500 78,750,000 39,375,000 447,562,500 Crop Program 2. AYII for Subsistence Farmers <2.5 Acres Number of Insured Farmers 250,000 500,000 750,000 1,000,000 1,250,000 1,500,000 1,750,000 1,750,000 8,750,000 Insured Area per Farmer (Acres/farmer) 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 Total Insured Crop Area (Acres) 250,000 500,000 750,000 1,000,000 1,250,000 1,500,000 1,750,000 1,750,000 8,750,000 Sum Insured US$ per Acre 200 150 200 150 200 150 200 150 Total Sum Insured (US$) 50,000,000 75,000,000 150,000,000 150,000,000 250,000,000 225,000,000 350,000,000 262,500,000 1,512,500,000 Premium Rate (%) 7.5% 5.00% 7.5% 5.00% 7.5% 5.00% 7.5% 5.00% Total Premium Income(US$) 3,750,000 3,750,000 11,250,000 7,500,000 18,750,000 11,250,000 26,250,000 13,125,000 95,625,000 Crop Program 3. Tree Crops (Mango, Citrus) 2019/20 2020/21 2021/22 2022/23 Number of Insured Farmers 2,500 5,000 7,500 10,000 25,000 Insured Area per Farmer (Acres/farmer) 2.5 2.5 2.5 2.5 2.5 Total Insured Crop Area (Acres) 6,250 12,500 18,750 25,000 62,500 Sum Insured US$ per Acre 1,000 1,000 1,000 1,000 Total Sum Insured (US$) 6,250,000 12,500,000 18,750,000 25,000,000 62,500,000 Premium Rate (%) 10.00% 10.00% 10.00% 10.00% Total Premium Income US$ 625,000 1,250,000 1,875,000 2,500,000 6,250,000 TOTAL ALL CROP PROGRAMS TOTAL INSURED FARMERS 250,000 350,000 700,000 1,052,500 1,350,000 1,655,000 1,950,000 2,232,500 2,500,000 2,510,000 14,550,000 TOTAL INSURED AREA (ACRES) 875,000 1,225,000 1,825,000 2,431,250 2,850,000 3,287,500 3,700,000 4,056,250 4,375,000 4,400,000 29,025,000 TOTAL SUM INSURED (US$) 350,000,000 367,500,000 680,000,000 658,750,000 990,000,000 845,000,000 1,230,000,000 1,005,000,000 1,400,000,000 1,075,000,000 8,601,250,000 TOTAL PREMIUM INCOME (US$) 26,250,000 18,375,000 51,000,000 33,250,000 74,250,000 42,875,000 92,250,000 51,187,500 105,000,000 55,000,000 549,437,500 Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 151 PUNJAB CROP INSURANCE PROGRAM: NUMBER OF INSURED FARMERS, INSURED AREA, SUM INSURED, TABLE A9.3:  152 PREMIUM AND PREMIUM SUBSIDIES: SCENARIO 3, MEDIUM LEVELS OF UPTAKE AND LOWER AVERAGE PREMIUM RATES: KHARIF 5.0 PERCENT AND RABI 3.5 PERCENT (US$) TOTAL Year/season Kharif 2018 Rabi 2018/19 Kharif 2019 Rabi 2019/20 Kharif 2020 Rabi 2020/21 Kharif 2021 Rabi 2021/22 Kharif 2022 Rabi 2022/23 (cumulative) Crop Program 1. AYII For Progressive Farmers linked to KISAN Credit Package (2.5 Acres to 12.5 Acres) Number of Insured Farmers 125,000 175,000 225,000 275,000 300,000 325,000 350,000 362,500 375,000 375,000 2,887,500 Insured Area per Farmer (Acres/farmer) 3.5 3.5 3.5 3.5 3.5 3.5 3.5 3.5 3.5 3.5 3.5 Total Insured Crop Area (Acres) 437,500 612,500 787,500 962,500 1,050,000 1,137,500 1,225,000 1,268,750 1,312,500 1,312,500 10,106,250 Sum Insured (US$ per Acre) 400 300 400 300 400 300 400 300 400 300 Total Sum Insured (US$) 175,000,000 183,750,000 315,000,000 288,750,000 420,000,000 341,250,000 490,000,000 380,625,000 525,000,000 393,750,000 3,513,125,000 Premium Rate (%) 5.0% 3.50% 5.0% 3.50% 5.0% 3.50% 5.0% 3.50% 5.0% 3.50% Total Premium Income(US$) 8,750,000 6,431,250 15,750,000 10,106,250 21,000,000 11,943,750 24,500,000 13,321,875 26,250,000 13,781,250 151,834,375 Crop Program 2. AYII for Subsistence Farmers <2.5 Acres Number of Insured Farmers 125,000 250,000 375,000 500,000 625,000 750,000 875,000 875,000 4,375,000 Insured Area per Farmer (Acres/farmer) 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 Total Insured Crop Area (Acres) 125,000 250,000 375,000 500,000 625,000 750,000 875,000 875,000 4,375,000 Sum Insured US$ per Acre 200 150 200 150 200 150 200 150 Total Sum Insured (US$) 25,000,000 37,500,000 75,000,000 75,000,000 125,000,000 112,500,000 175,000,000 131,250,000 756,250,000 Premium Rate (%) 5.0% 3.50% 5.0% 3.50% 5.0% 3.50% 5.0% 3.50% Total Premium Income(US$) 1,250,000 1,312,500 3,750,000 2,625,000 6,250,000 3,937,500 8,750,000 4,593,750 32,468,750 Crop Program 3. Tree Crops (Mango, Citrus) 2019/20 2020/21 2021/22 2022/23 Number of Insured Farmers 1,250 2,500 3,750 5,000 12,500 Insured Area per Farmer (Acres/farmer) 2.5 2.5 2.5 2.5 2.5 Total Insured Crop Area (Acres) 3,125 6,250 9,375 12,500 31,250 Sum Insured US$ per Acre 1,000 1,000 1,000 1,000 Total Sum Insured (US$) 3,125,000 6,250,000 9,375,000 12,500,000 31,250,000 Premium Rate (%) 10.00% 10.00% 10.00% 10.00% Total Premium Income US$ 312,500 625,000 937,500 1,250,000 3,125,000 TOTAL ALL CROP PROGRAMS TOTAL INSURED FARMERS 125,000 175,000 350,000 526,250 675,000 827,500 975,000 1,116,250 1,250,000 1,255,000 7,275,000 TOTAL INSURED AREA (ACRES) 437,500 612,500 912,500 1,215,625 1,425,000 1,643,750 1,850,000 2,028,125 2,187,500 2,200,000 14,512,500 TOTAL SUM INSURED (US$) 175,000,000 183,750,000 340,000,000 329,375,000 495,000,000 422,500,000 615,000,000 502,500,000 700,000,000 537,500,000 4,300,625,000 TOTAL PREMIUM INCOME (US$) 8,750,000 6,431,250 17,000,000 11,731,250 24,750,000 15,193,750 30,750,000 18,196,875 35,000,000 19,625,000 187,428,125 A Feasibility Study PUNJAB CROP INSURANCE PROGRAM: NUMBER OF INSURED FARMERS, INSURED AREA, SUM INSURED, TABLE A9.4:  PREMIUM AND PREMIUM SUBSIDIES: SCENARIO 4, MEDIUM LEVELS OF UPTAKE AND HIGHER AVERAGE PREMIUM RATES: KHARIF 7.5 PERCENT AND RABI 5.0 PERCENT (US$) TOTAL Year/season Kharif 2018 Rabi 2018/19 Kharif 2019 Rabi 2019/20 Kharif 2020 Rabi 2020/21 Kharif 2021 Rabi 2021/22 Kharif 2022 Rabi 2022/23 (cumulative) Crop Program 1. AYII For Progressive Farmers linked to KISAN Credit Package (2.5 Acres to 12.5 Acres) Number of Insured Farmers 125,000 175,000 225,000 275,000 300,000 325,000 350,000 362,500 375,000 375,000 2,887,500 Insured Area per Farmer (Acres/farmer) 3.5 3.5 3.5 3.5 3.5 3.5 3.5 3.5 3.5 3.5 3.5 Total Insured Crop Area (Acres) 437,500 612,500 787,500 962,500 1,050,000 1,137,500 1,225,000 1,268,750 1,312,500 1,312,500 10,106,250 Sum Insured (US$ per Acre) 400 300 400 300 400 300 400 300 400 300 Total Sum Insured (US$) 175,000,000 183,750,000 315,000,000 288,750,000 420,000,000 341,250,000 490,000,000 380,625,000 525,000,000 393,750,000 3,513,125,000 Premium Rate (%) 7.5% 5.00% 7.5% 5.00% 7.5% 5.00% 7.5% 5.00% 7.5% 5.00% Total Premium Income(US$) 13,125,000 9,187,500 23,625,000 14,437,500 31,500,000 17,062,500 36,750,000 19,031,250 39,375,000 19,687,500 223,781,250 Crop Program 2. AYII for Subsistence Farmers <2.5 Acres Number of Insured Farmers 125,000 250,000 375,000 500,000 625,000 750,000 875,000 875,000 4,375,000 Insured Area per Farmer (Acres/farmer) 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 Total Insured Crop Area (Acres) 125,000 250,000 375,000 500,000 625,000 750,000 875,000 875,000 4,375,000 Sum Insured US$ per Acre 200 150 200 150 200 150 200 150 Total Sum Insured (US$) 25,000,000 37,500,000 75,000,000 75,000,000 125,000,000 112,500,000 175,000,000 131,250,000 756,250,000 Premium Rate (%) 7.5% 5.00% 7.5% 5.00% 7.5% 5.00% 7.5% 5.00% Total Premium Income(US$) 1,875,000 1,875,000 5,625,000 3,750,000 9,375,000 5,625,000 13,125,000 6,562,500 47,812,500 Crop Program 3. Tree Crops (Mango, Citrus) 2019/20 2020/21 2021/22 2022/23 Number of Insured Farmers 1,250 2,500 3,750 5,000 12,500 Insured Area per Farmer (Acres/farmer) 2.5 2.5 2.5 2.5 2.5 Total Insured Crop Area (Acres) 3,125 6,250 9,375 12,500 31,250 Sum Insured US$ per Acre 1,000 1,000 1,000 1,000 Total Sum Insured (US$) 3,125,000 6,250,000 9,375,000 12,500,000 31,250,000 Premium Rate (%) 10.00% 10.00% 10.00% 10.00% Total Premium Income US$ 312,500 625,000 937,500 1,250,000 3,125,000 TOTAL ALL CROP PROGRAMS TOTAL INSURED FARMERS 125,000 175,000 350,000 526,250 675,000 827,500 975,000 1,116,250 1,250,000 1,255,000 7,275,000 TOTAL INSURED AREA (ACRES) 437,500 612,500 912,500 1,215,625 1,425,000 1,643,750 1,850,000 2,028,125 2,187,500 2,200,000 14,512,500 TOTAL SUM INSURED (US$) 175,000,000 183,750,000 340,000,000 329,375,000 495,000,000 422,500,000 615,000,000 502,500,000 700,000,000 537,500,000 4,300,625,000 TOTAL PREMIUM INCOME (US$) 13,125,000 9,187,500 25,500,000 16,625,000 37,125,000 21,437,500 46,125,000 25,593,750 52,500,000 27,500,000 274,718,750 Assessing the Potential for Large-Scale Agricultural Crop and Livestock Insurance in Punjab Province, Pakistan 153 Report No. 133553–PK