This paper suggests a cross hedging strategy for managing non-energy commodity price risk using both crude oil futures and corresponded non-energy commodity futures. We apply multiple random coefficient autoregressive Markov regime switching models (MRCARRS) for simultaneously estimating the optimal hedge ratios of crude oil futures and non-energy commodity futures. We also envision a more parsimonious partial switching version of MRCARRS (PRCARRS) for multiple futures hedging. Empirical results show that either MRCARRS or PRCARRS is the best performer for all commodities considered. According to the Diebold, Mariano and West (DMW) test statistics, the hedging performance of the multiple futures ordinary least square (MOLS) is statistically no worse than the single futures ordinary least square (OLS). This justifies the superiority of multiple futures hedging over single futures hedging. Moreover, all DMW statistics are positive for the best performer (MRCARRS or PRCARRS) over competing hedging strategies indicating that multivariate state-dependent RCARRS models have a tendency to outperform state-independent and static hedging models.
期貨與選擇權學刊=Journal of Futures and Options 8(1)，頁41-84