ARE FUTURES HEDGE RATIOS STATIONARY?

There are techniques available for reducing and managing financial risk; the most widely used is hedging with futures contracts. A considerable amount of research has focused on modelling the distribution of spot and futures prices and applies the results to estimate the optimal hedge ratio (HR) using OLS, GARCH, ECM and VECM (Floros and Vougas, 2004; Degiannakis and Floros, 2010). Several studies have investigated the optimal HR using stock index futures estimating a constant and a time-varying HR. According to Sutcliffe (2003), stock index futures can be used to hedge market risk caused by spot price fluctuations. Optimal HRs are estimated by modelling the distribution of stock index and futures price changes within the GARCH methodology which allows for time-dependent conditional variances in the unconditional distribution of price changes (Park and Switzer, 1995). Estimation of optimal HRs has also been approached with time-dependent conditional variance models. To estimate a HR, early works used the slope of an OLS regression of stock on futures prices, while an improvement has been made by adopting a bivariate GARCH framework (see Floros and Vougas, 2004). Although these studies are successful in capturing the time varying covariance-variances, almost all of them focus on the HR estimation only. The main purpose of this paper is to examine if HR is stationary over time, as we know little about the stationarity of HRs. Previous studies such as Ederington (1979) and Anderson and Dathine (1981) assume that optimal HR is constant when it can be obtained as a slope coefficient of an OLS regression. When optimal HRs depend on the conditional distributions of stock index futures price movements, then HRs vary over time as this distribution changes.