A Comparison of Forecasting Volatility Strategies into ARCH Class through Option Pricing

Daily data on the German market index return are used to consider multiple issues in a forecasting comparison of ARCH-type specifications. first, attention is paid to the impact of different sample sizez, different horizons and fitting of historical versus implied data. Secondly, the issue of volatility transmission is addressed by modelling French and Germany market indexes into simultaneous conditionally heteroskedasticity framework. Errors obtained by updating the Black and Scholes formula with the different volatility forecasts are compared. The findings support, if no implied volatility is available, the use of the simplest GARCH specification estimated on short recent sample