An econometric model of Manitoba crop supply response under risk aversion and price uncertainty

Abstract

Restricted and unrestricted distributed lag models are used to investigate and establish probable lag structures in specific crop supply response to market and Canadian wheat board prices. The identified lag structures are in some respects consistent with other studies showing different lag structures for crop yield and acreage response and also give an indication of model sensitivity to numeraire choice in normalized quadratic functional forms. Lag structures form the basis of expectation and variance proxies employed in risk neutral, linear mean-variance and non-linear mean-variance econometric model specification and estimation using duality. Linear SUR was used to estimate duality models in the cases of risk neutrality and linear mean-variance. Both a linear reduced form and an non-linear duality model was estimated in the case of non-linear mean-variance. Price variances and expected prices are more significant in the risk aversion models which imply gains in model specification. Symmetry and CRTS are rejected in all models. Homogeneity conditions corresponding to risk neutrality is rejected in all models. Constant absolute risk aversion and Constant relative risk aversion are also rejected.