A Nested Test for Common Yield Distributions with Application to U.S. Corn

We propose the use of maximum-entropy techniques to nest and test the functional form of commonly used yield densities. We demonstrate how common parametric yield models can be nested within a maximum-entropy framework and subsequently tested for using standard hypothesis tests. We include an empirical application that tests the beta distribution against a more general maximum-entropy alternative using county-level corn yield data for the U.S. Corn Belt and find evidence in favor of the alternative generalized maximum-entropy distribution.