Forecasting Stock Return Volatility: A Comparison of GARCH, Implied Volatility, and Realized Volatility Models

Abstract

We investigate the information content of implied volatility forecasts for stock index return volatility. Using different autoregressive models, we examine whether implied volatility forecasts contain information for future volatility beyond that in GARCH and realized volatility models. Results show implied volatility follows a predictable pattern and confirm the existence of a contemporaneous relationship between implied volatility and index returns. Individually, implied volatility performs worse than alternate forecasts, however, a model that combines an asymmetric GARCH model with implied and realized volatility through (asymmetric) ARMA models is preferred model for forecasting volatility. This evidence is further supported by consideration of value-at-risk.