Texas dryland upland cotton yields have historically exhibited greater variation and more distributional irregularities than the yields of other crops, raising concerns that conventional parametric distribution models may generate biased or otherwise inaccurate crop insurance premium rate estimates. Here, we formulate and estimate regime-switching models for Texas dryland cotton yields in which the distribution of yield is conditioned on local drought conditions. Our results indicate that drought-conditioned regime-switching models provide a better fit to Texas county-level dryland cotton yields than conventional parametric distribution models. They do not, however, generate significantly different Group Risk Plan crop insurance premium rate estimates.

Three index-based crop insurance contracts are evaluated for representative south Georgia corn farms. The insurance contracts considered are based on indexes of historical county yields, yields predicted from a cooling degree-day production model, and yields predicted from a crop-simulation model. For some of the representative farms, the predicted yield index contracts provide yield risk protection comparable to the contract based on historical county yields, especially at lower levels of risk aversion. The impact of constraints on index insurance choice variables is considered and important interactions among constrained, conditionally optimized, choice variables are analyzed.