Dynamic Hedging Effectiveness: An empirical M-GARCH comparison

Absztrakt (kivonat)

This study utilizes M-GARCH models, specifically DCC-GARCH and GO-GARCH, to examine the volatilities and conditional covariances among the SP 500, bond prices, oil prices, gold prices, and EUR/USD. A rolling window approach is employed to construct out-of-sample one-step-ahead forecasts of dynamic conditional covariances to assess time-varying dynamic hedge ratios. The purpose of this paper is to assess the hedging effectiveness (HE) of dynamic hedge ratios for four assets, which are constructed based on the forecasted conditional covariances of M-GARCH models. The findings suggest that EUR/USD and bonds are the most effective for hedging purposes, while oil and gold have relatively low HE values. DCC-GARCH out-performs GO-GARCH for all assets, except for oil, where the latter yields slightly higher HE value. While the F-test is a recently introduced method in hedging literature for evaluating HE, its power is low when the underlying variable is not normally distributed, therefore, the Brown-Forsythe test, which is better suited for t-distributed time series, is used in this study. Based on the F-test, only the EUR/USD hedge is statistically significant at a 10% significance level, while according to the Brown-Forsythe test, none of the hedges proved significant. The results appear robust to different numbers of model refits and distributional assumptions, with DCC-GARCH producing more robust results than GO-GARCH.