Implications of within county yield heterogeneity for modeling crop insurance premiums

Purpose - Farm level data are essential to accurate setting of crop insurance premium rates, but their time series tends to be too short to allow them to be the sole data source. County level data are available in longer time series, however. The purpose of this paper is to present a methodology to make full use of the information inherent in each of these data sets.Design methodology approach - The paper uses a novel application of statistical tools for using farm and county level yield data to generate farm level yield densities that explicitly incorporate within county yield heterogeneity while accounting for systemic risk and other spatial or intertemporal correlations among farms within the county.Findings - The empirical analysis shows that current approaches used by the Risk Management Agency to individualize premiums for a farm result in substantial mispricing of crop insurance premiums because they do not adequately capture farm yield variability and yield correlations between farms. The new premium setting method is empirically shown to substantially reduce government subsidies for crop insurance premiums.Originality value - The paper demonstrates how to extract more information from available data when setting crop insurance premiums, which allows the government to more closely tailor premiums to the farm than do current approaches.