Directional Spatial Dependence and Its Implications for Modeling Systemic Yield Risk

The objective of this study is to evaluate and model the spatial dependence of systemic yield risk. Various spatial autoregressive models are explored to account for county level dependence of crop yields. The results show that the time trend parameters of yields are correlated across spaces and the spatial correlations are changing with time. In addition, the spatial correlation of neighborhood in west/east direction is stronger than that of north/south direction. The information of the spatial dependence of yield risk will help the construction of better risk management programs for protecting producers from systemic yield risks.