The Relative Importance of Overnight Sentiment Versus Trading-Hours’ Sentiment in Volatility Forecasting

The focus of the paper is to examine the relative importance of overnight sentiment versus trading-hours’ sentiment in forecasting volatility. In previous studies on investor sentiment, either overnight sentiment is ignored, or overnight sentiment and trading-hours’ sentiment are aggregated. With the help of Chinese sentiment dictionary, we extract investor sentiment from Chinese internet social forums. Our empirical evidences confirm that investor sentiment significantly affects realized volatility. In particular, we find that overnight sentiment is more informative than that in trading-hours in volatility forecasting, and the predictive power of overnight sentiment also significantly exceeds that of overnight returns, which are widely used to capture overnight information. Our results pass a series of robustness tests, including highly volatile sample, alternative rolling window size and alternative sentiment proxy