Nonparametric estimation of crop insurance rates revisited

With the crop insurance program becoming the cornerstone of U.S. agricultur al policy, recovering accurate rates is of paramount interest. Lack of yiel d data presents, by far, the most fundamental obstacle to recovery of accur ate rates. This article employs new methodology to estimate conditional yie ld densities and derive the insurance rates. In our application, we find th e nonparametric kernel density estimator requires an additional twenty-six years of yield data to estimate the shape of the conditional yield densitie s as accurately as the recently developed empirical Bayes nonparametric ker nel density estimator. Such methodological improvements can significantly a id in ameliorating the data problem.