Forecasting Equity Volatility Dynamics with Markov-Switching EGARCH Models

Abstract

Understanding and anticipating stock market volatility enables better portfolio management. We forecast US equity volatility with a Markov-Switching EGARCH model with one high and one low volatility regime. We show that this model contains similar information about future volatility as the VIX Index. It also outperforms single-regime GARCH and EGARCH models. Moreover, the model’s 1-day ahead regime predictions are economically significant: market volatility and kurtosis, equity risk premia, and stock-bond relations shift when the model forecasts a regime change.