Markov Switching Dynamic Correlation: An Empirical Study of Hedging in Crude Oil and Natural Gas Markets

O. Rossarin, S. Worrawat, A. Kittawit, Y. Woraphon

Abstract

This article studies how to achieve risk-minimization through hedging strategy in crude oil and natural gas markets. In this study, the covariance of spot and futures returns is computed through the Markov Switching Dynamic Conditional Correlation GARCH model. The model is compared to single regime DCC-GARCH for both oil and gas spot/futures pairs in order to examine the presence of the structural change in the dynamic correlation. The result confirms the superiority of two-regime model in terms of log-likelihood, AIC, and BIC. Then, the obtained conditional volatility and correlation are further used to compute the hedge ratio and optimal portfolio weight for oil and gas spot/futures pairs. The results show that the risk-minimizing hedge ratio of oil and that of gas are averagely 0.844 and 0.32, respectively, and investors should hold only half of oil and gas futures contracts (52 and 48 percent) to lowest their risk. Investors should be careful about the situation in the markets before making investment or changing policy.