Title: Forecasting volatility in stock market: The gains from using intraday data
Authors: Valeriy Zakamulin - University of Agder (Norway)
Xingyi Li - University of Agder (Norway) [presenting]
Abstract: The aim is to comprehensively assess the gains in volatility forecasting accuracy provided by realized measures of daily volatility versus that provided by using daily returns. We extend previous studies on forecasting stock market volatility in several directions. First, we use an extensive set of intraday data on prices of single stocks and stock market indices. Second, we assess the gains in forecast accuracy provided by using intraday data over multiple horizons, ranging from 1 day to 6 months. Third, we compare forecasting abilities of several competing models. Our results indicate that there are marginal differences between forecast accuracies provided by alternative models. The major finding of our empirical study is that, regardless of the length of the forecasting horizon, intraday data allow one to reduce the mean squared forecasting error up to 35\%. Thus, the gains from using intraday data are highly economically significant.