Public information arrival and volatility persistence in financial markets

This paper explores the relationship between daily market volatility and the arrival of public information in four different financial markets. Public information is measured as the daily number of economic news headlines, divided in six categories of news. Statistical analysis of the news data suggests the presence of particular seasonality effects, as well as a strong degree of autocorrelation. Over the period 1994-1998, significant effects of specific news categories on the volatility of US stocks, treasury bills, bonds and dollar were detected. However, the effects - in size and duration - vary by news category and by financial market. It is demonstrated that most of the volatility persistence, as observed by GARCH models, tends to disappear when news is included in the conditional variance equation.