Estimating temperature impacts on perennial crop losses in California: Insights from insurance data

This study provides a novel empirical framework for estimating the effects of temperature on the production of several tree crops in California by using underused insurance data and a variable selection technique. We utilize a Bayesian variable selection technique to select relevant temperature variables. We then use the selected temperature variables to assess the temperature effects on crop losses. We find that a greater length to freeze exposure increases crop losses and relatively smaller roles of chill hours and heating hours. We also find the heterogeneity of the effects across insurance coverage levels.