Utilizing large-scale insurance data sets to calibrate sub-county level crop yields

Abstract

Crop yields are crucial for research on agricultural risk and productivity but are typically only available at highly aggregated levels. Yield data at more granular levels of observation have the potential to enhance econometric identification and improve statistical power but are typically inaccessible. Crop insurance contracts offered via the US Federal Crop Insurance Program (FCIP) are priced, in part, based on past yields of the farm meaning year-to-year variation in premium rates has the potential to provide insight into how yields vary over time. This paper introduces methods to use observed FCIP rating parameters to calibrate yields for insurance transactions lacking such data. These methods are validated with 148,243 farm-level observations from Kansas for which yields are known. The calibrated yields are applied empirically to examine the impact of asymmetric information in the FCIP via choice of insurance unit structure and the extent to which legislative changes mitigated this effect.