Portfolio Optimization of Energy Commodity Futures Returns with Minimum Information Copula

Payap Tarkhamtham, Jirakom Sirisrisakulchai, Roengchai Tansuchat

Abstract

Energy commodity futures returns are modeled using GARCH and EGARCH processes. Their dependence structures are constructed from vine copula approach. Minimum information methods were applied to approximate bivariate copulas for all pairs in vine structure. A copula that satisfies a set of data constraints and that has minimum relative entropy (with respect to the independence copula) among the class of all copulas satisfying those constraints is called minimum information copula. The vine copula built from minimum information copulas is used to quantify the risks of energy commodity portfolio. Optimal portfolio that minimizes the risks with a given expected return are obtained for 4 energy commodity products (crude oil, natural gas, gasoline, and heating oil) using data from the New York Mercantile Exchange (NYMEX).