Predicting the monthly volatility of the EuroStoxx 50 using data sampled at different frequencies

Abstract

This paper analyses the forecastability of the EuroStoxx 50 monthly returns volatil- ity. We consider different proxies for the unobserved volatility variable by using data sampled at di¤erent frequencies, and GARCH and AGARCH models with Normal and Student s t errors for the dynamics of returns conditional variance. We nd that a method based on aggregation of multi step (daily) ahead GARCH-type forecasts provide quite accurate predictions of monthly volatility.