Using Non-Contemporaneous Data to Specify Risk Programming Models

Specification of the variance-covariance matrix holds continuing interest for agricultural economists considering risk programming applications. This research examines alternative expected value-variance (E-V) frontiers constructed using contemporaneous and non-contemporaneous data and two statistical assumptions concerning crop prices and yields. Empirical examples from two locations for different crops illustrate the various assumptions. Considerable differences in the E-V efficient frontiers occur in both empirical settings.