Hedging industrial metals with stochastic volatility models

The financialization of commodities documented in [Tang and Xiong (2012) Financial Analyst Journal, 68:54–74]has led commodity prices to exhibit not only time-varying volatility, but also price and volatility jumps. Using the class of stochastic volatility (SV) models, we incorporate such extreme price movements to generate out-of-sample hedge ratios. In-sample estimation on China’s copper (CU) and aluminum (AL) spot and futures markets confirms the presence of price jumps and price-volatility jump correlations. Out-of-sample hedge ratios from the [Bates (1996) Review of Financial Studies, 9:69–107] SV with price jumps model deliver the greatest risk reduction on the unhedged positions at 59.55% for CU and 49.85% for AL. But it is the [Duffie, Pan, and Singleton (2000) Econometrica, 68:1343–1376] SV model with correlated price and volatility jumps that produces hedge ratios which yield the largest Sharpe Ratios of 0.644 for CU and 0.886 for AL.