Spatial Data-Driven Approach to Crop Insurance Rate Making

Abstract

The agricultural sector in India contributes approximately 18.0% to the GDP and employs 52% of the total workforce. This sector drives a structural transformation process, shifting resources from low productivity to higher-productivity sectors, fostering holistic and inclusive growth. The changing climate and technological advancements impact crop yields, often resulting in nonstationary losses, impoverishment, and indebtedness, particularly among small and marginal farmers. The Government of India's commitment to farmers' welfare is reflected in the substantial budget allocation, which has surpassed 125 crores for the financial year 2023-24. Given the substantial resources dedicated to various schemes, accurate rating of insurance contracts is critical for both government and private insurers. As crop insurance schemes become a cornerstone of Indian agricultural policy, accurately modelling crop yield distributions by incorporating relevant loss information is essential for effective risk management for stakeholders, including farmers, insurers, and governments. This study presents multiple distribution models and introduces a flexible, data-driven spatial model designed to smooth across spatial units, thereby effectively capturing the distributed data. Our findings show that significant information can be gained from spatial data, and such methodological improvements can greatly enhance the accuracy of rate estimation, ultimately benefiting the stakeholders across the agricultural sector.