Optimal hedge ratio estimation and effectiveness using ARCD

This paper examines the importance of higher moments for optimal hedge ratio estimation by using Autoregressive Conditional Density (ARCD) models which allow for time varying skewness and kurtosis. The performance of ARCD is evaluated against that of conventional hedge ratio estimation methodologies based on OLS, error-correction, exponentially weighted moving averages, and, univariate and multivariate GARCH, respectively. An empirical application using spot and futures data on the DJI, FTSE and DAX indices, compares the in-sample and out-of-sample hedging effectiveness of each approach in terms of risk minimization. The results show that the ARCD has the best performance thus suggesting that higher moments are of practical importance in futures hedging.