Abstract A Bayesian risk methodology is outlined for making decisions under uncertainty. A practical example is given for a crop insurance where the insurer decides how to take risks. The insurer's objective is formulated as a goal function whose expected value must equal zero. Risk to the insurer arises from the uncertainty and variation in the input variables of a previously developed deterministic yield model. Monte Carlo simulation provides a cumulative frequency histogram of the goal function from which risks are calculated. The methodology is general and can be used in many situations to determine the risk in a project from uncertain inputs.