This article evaluates the performance of expected value and expected value-variance criteria in achieving risk efficiency in crop selection. Results indicate that the expected returns criterion achieves risk efficiency in many situations because of constraints. However, in the absence of many constraints the expected returns criterion performs poorly except when highly mean-dominant activities are present. The expected value-variance criterion achieves a high degree of risk efficiency for all situations examined. This result implies that criteria more complex than expected value-variance are not necessary for crop selection analysis, given empirical returns distributions.