RISK ANALYSIS BASED ON STOCHASTIC-DOMINAN CE - A CASE-STUDY WITH SELECTED CASH CROPS

A ''new'' crop is favourable for a risk-neutral decision maker (farmer ) if its expected gross margin is higher than the expected gross margi n of an competitive crop. A decision maker who is averse to risk is in terested additionally in the range of the expected gross margins. Assu ming specified restrictions on the decision maker's preferences the ri sk efficiency analysis offers a possibility to examine the risk of pla nting a crop. In this study the stochastic efficiency criteria is appl ied. The results of this technique are compared with those using Bayes theorem und mu-sigma-criterion. The yields for calculating the cross revenue of the various crops are obtained from Bundesanstalt fur Pflan zenbau based on field trials carried out in Fuchsenbigl (Lower Austria ) over a period of eleven years (1981 to 1991). Prices and costs are b ased on the 1992 farm level. Four conventional crops (corn, winter whe at, winter barley und winter rye) and two ''new'' crops (soybeans and sunflowers) are examined in this study. Assuming prices and costs on t he 1992 level, the ''new'' crops - specially soybeans - are competitiv e regarding expected gross margin as well as risk efficiency using the techniques named above. In this case the contribution of the area bas ed premium, the so called ''Flachenpramie'', to the gross margin is a decisive factor for the risk of planting a ''new'' crop, which means t hat the risk decreases by increasing contribution of the ''Flachenpram ie''. The rotations including soybeans and sunflowers turn out to be f avourable in comparison of ten hypothetical crop rotations. The analys is using criteria of first-degree stochastic dominance, where the deci sion makers prefer more to less at all outcome levels, results in an e fficient set which contains rotations including soybeans and/or sunflo wers. A further procedure assuming decision makers to be averse to ris k (second-degree stochastic dominance) identifies only a slightly smal ler efficient set. The analysis using criteria of third-degree stochas tic dominance does not reduce the number of rotations in the efficient set.