A New Approach to Improving Hedging Performance in the OLS Model

Abstract

The sensitivity of VIX futures to market movements changes over time with changes in market risk. Accordingly, in the case of using the OLS (ordinary least squares) model to hedge S&P 500 exposure with VIX futures, hedge ratios are affected by changes in risk appetite, which in turn contributes to the overall hedging performance as well as the asymmetry of the performance distribution. The conventional OLS approach does not effectively reflect this phenomenon in the model. In this study, the authors explore a new approach to improving hedging performance in the OLS model. They introduce an interaction term between the VIX and VIX futures returns into the OLS model. They find that the hedge ratios derived by the new approach provide better hedging results compared to the univariate OLS model in terms of mean return and downside risk protection, and also improve the asymmetry of the performance distribution. They extend their research to compare it with the performance of the dynamic conditional correlation (DCC)-generalized autoregressive conditional heteroskedasticity (GARCH) model. The new approach also shows better results than the DCC-GARCH approach. They obtain the same results in case studies of the Global Financial Crisis and the COVID-19 pandemic, and also in applying a trading strategy to each hedging methodology.