This paper presents a novel approach for dealing with risk in agricultural resource allocation decisions by synthesising the conventional Markowitzean, or MOTAD, methods within a compromise programming model to generate ‘best‐compromise’ solutions which come closest to an ideal point defined in terms of risk minimisation. This approach can be regarded as the compromise‐risk programming model. The purpose here is to show how this ‘hybridisation’ of Markowitz/MOTAD and compromise programming approaches overcomes some of the weaknesses of the traditional approach to handling risk in resource allocation models.