Intra -year strategic planning of crop production under embedded risks

The objective of this study was to investigate the possibility of improving dryland crop production decisions that entail embedded risks. Dryland cropping systems are dependent on precipitation use efficiency of the crops either grown as continuous crops or in rotation. Optimal decisions depend on ensuing states of nature. Experimental data from seven cropping systems were used in the analysis to ascertain whether economic improvements can be made using discrete stochastic programming as opposed to often used deterministic methods. Three different versions of a DSP model that used three different information sets (i.e., preseason moisture, April output prices and preseason moisture, and futures prices and preseason moisture) were solved under the assumption of decision maker risk neutrality. The model that used preseason moisture levels and April output prices generated consistently positive improvements over a fixed strategy. The expected value of additional information averaged $73.00 per acre or 43,800 per farm over a 15-year period for the model generated strategy. Nevertheless, robust inferences were not possible as confidence intervals were not established for these values. Such a procedure would require a model validation that show that optimal strategies are not sensitive to approximated distribution of random parameters when random vector is only known through a sample. Most importantly, use of DSP models for decision analysis under embedded risks should be justified in terms of the value of improved decisions