Area Response Functions
The results of the area response Equations (8) are listed in Table 3. The model was estimated using seemingly unrelated regression because the null hypothesis that the residuals of the three area response equations are independent was rejected using the Breusch– Pagan test [Chi-square (3) = 18.363; p-value = 0.0004]. Symmetry was also imposed for the three equations in the model. The null hypothesis that the coefficients for the variance, covariance and wealth effects are jointly zero was rejected [Chi-square (21)=130.20; p-value = 0.0000], implying that producers of these crops are not risk-neutral. Similarly, the null hypothesis that that the coefficients of wealth are jointly zero was rejected [Chisquare (3)=42.19; p-value=0.0000], implying non-CARA. These findings are consistent with von Massow and Weersink (1993) for crop farmers in Ontario. The overall fit of the resulting models is very good, with adjusted R2 values of 0.92, 0.94, and 0.56 for corn, soybean, and winter wheat, respectively.