The Bayesian Decision Model with more than One Predictor—An Application to the Stocking Rate Problem

The variables a manager faces in making decisions may be divided into two broad categories—those which are determined by the manager and those which are outside of his control. Agricultural economists have made many efforts to develop expectation models for one or more of the uncontrollable variables facing farmers and have suggested procedures for utilizing the resulting expectations. Recent developments in statistical decision theory provide a logically consistent framework for incorporating the predictions of expectation models [4, pp. 192-196]. Applications of Bayesian analysis utilizing predictions of one uncontrollable variable have been reported in the literature [1, 3]. However, many decision problems logically require expectations of two uncontrollable variables (such as price and yield) or more. This article illustrates a method of including predictors for more than one uncontrollable variable in the Bayesian framework, and reports some empirical results of an application to a stocking rate problem.