A safety-first approach to dynamic cropping decisions

A risk-sensitive optimal flexible cropping plan for spring barley in southeastern Washington state is identified by stochastic dynamic programming. State transformation equations are determined from data generated by a process simulation model. Katoaka's safety-first criterion is proposed as an operational procedure for discounting payoff values for differential riskiness. Stronger short-run risk aversion leads to the use of more conservative optimal flexcrop strategies which require higher threshold spring soil moisture levels for cropping. The dynamically optimal barley flexcrop strategies generate similar expected net returns but decrease soil loss relative to a fixed winter wheat–fallow rotation; however, annual income risk increases.