Mixed-Copula Based Extreme Dependence Analysis: A Case Study of Food and Energy Price Comovements

Rich empirical literature has investigated the price transmission among spatially separated and vertically linked markets. In this study, we fill a gap in the price transmission literature by investigating extreme dependence that allows varying general dependence structure between extreme and non-extreme market conditions (through mixed copula functions), and changing degree of co-movements (through time-varying dependence parameters for any given copula functions). Our work is a combination and generalization of time-varying attributes with the mixture model idea. The data used for the analysis are weekly prices for US crude oil, ethanol, and corn from Jan 2000 through December 2013. Our results demonstrate that time-varying attributes in extreme price co-movements can result from many reasons such as government interventions, financial contagion, disease outbreaks, and altering consumer tastes. It is thus a useful extension and generalization of existing approaches for modeling price transmission that has appeared in the literature.