Is Overnight Volatility Overlooked

We use both daytime and overnight high-frequency price data of equity index futures to estimate the realized volatility (RV) of S&P500 and NASDAQ-100 indexes. Our empirical results reveal strong inter-correlation between the regular-trading-time and after-hour RVs, as well as a significant predictive power of overnight RV on daytime RV and versa vice. We propose a Day-Night Realized Stochastic Volatility (DN-SV-RV) model, where the daytime and overnight returns are jointly modeled with their RVs and their latent volatilities are correlated. The newly proposed DN-SV-RV model has the best in-sample estimations and out-of-sample return distribution forecasts among benchmark models. Under this innovative framework, the model implications convince that daytime and overnight volatility are two sides of the same coin, and the volatility clustering is persistent throughout during and after regular trading hours.