Endogenous risk in rational-expectations commodity models: A multivariate generalized ARCH-M approach

The feasibility of including endogenous risk in a rational-expectations model is examined by using a multivariate GARCH-M setup. Unlike previous attempts to model risk under the rational-expectations hypothesis (REH), the model's full covariance structure is allowed to be time varying. The application is with a market model of the U.S. broiler industry; results indicate broiler production is responsive to time-varying price volatility, although the estimated effects are not large. The GARCH-M model is compared with one that uses a two-step approach. Evidence indicates the informationally efficient GARCH-M approach is superior.