Forecasting Stock Volatility: The Gains from Using Intraday Data

Abstract

There is evidence that volatility forecasting models that use intraday data provide better forecast accuracy as compared with that delivered by the models that use daily data. Exactly how much better is still unknown. The present paper fills this gap in the literature and extends previous studies on forecasting stock market volatility in several important directions. First, we employ an extensive set of intraday data on 31 individual stocks over a sample period of 19 years. Second, we use forecast horizons ranging from 1 day to 6 months. Third, we evaluate the precision of volatility forecast provided by various competing models. Fourth, we conduct several robustness checks to assess the sensitivity of our results to various alternative choices. The major finding of our empirical study is that the gains from using intraday data are rather significant and persist over longer forecast horizons. Depending on the forecast horizon, the improvement in forecast precision varies from 30 to 50 percent. We demonstrate that our main results on the forecast accuracy gains are robust to the choice of intraday data frequency and the choice of measure of realized daily volatility.

Keywords: stock markets, volatility forecasting, intraday data, measures of realized daily volatility, forecast accuracy, out-of-sample forecasting, model comparison

JEL Classification: C22, C53, C58, G17