Abstract
This study compares the performance of the ISD, the GARCH(1,1), the historical volatility estimates and of two lagged trading volume measures for predicting the Swiss Stock Market Index's (SMI) volatility. The ISD has a superior daily informational content than the GARCH(1,1) estimate and retains unbiased but decreasing explanatory power over up to 20 days ahead horizons. Mean and spread daily volume measures play a significant correcting role when forecasting stock market volatility over daily and longer intervals respectively and clearly dominate the GARCH(1,1) forecasts. Their significance emphasises heterogeneous horizon traders' influence on the SMI volatility time series properties