Ramiro Ruiz Cárdenas

The absence of relatively long data sets as well as of suitable actuarial methods to price crop insurance contracts are among the main reasons for the poor performance of agricultural risk management programs. In recent years new insurance products are being introduced in order to make crop insurance more popular among producers. One of these is area yield crop insurance, in which farmers collect an indemnity whenever the county average yield falls beneath a yield guarantee, regardless of the farmer's actual yields.

Using a hierarchical Bayesian framework, as well as a dynamic modelling approach, we propose a pricing methodology to this type of contracts based on the space-time modeling of crop yield data. In the Bayesian framework, estimates of the premium rates are directly obtained, capturing inference uncertainties involved in its prediction.

In the dynamic space-time modeling approach the spatial dependence is specified through proper Gaussian Markov random fields, which improves the efficiency of the MCMC procedure. Missing crop yield data for the historical series of some counties excluded from the original analysis are properly imputed and incorporated into the new modeling approach. Suitable covariates directly related to crop yield are also identified and included into the model.