Optimal Futures Hedging Under Distinct Daytime and Overnight GARCH Processes: Leveraging Information FROM Opening, High, Low, and Closing Prices

Abstract

In this paper, novel two-component generalized autoregressive conditional heteroskedasticity (2COMP) models for optimal futures hedging are proposed. These models leverage the distinct dynamic patterns of daytime and overnight returns and valuable information regarding covariance prediction that is contained by opening, high, low, and closing price data. Compared with standard models, the proposed 2COMP models have higher predictive capability, achieving lower out-of-sample realized hedged portfolio variances when applied to US equity indices. These performance advantages can translate to economic benefits for hedgers, particularly those with higher risk aversion, and the findings illustrate the value of 2COMP models for effective hedging.