Evidence that Extreme Volatility in Stock Prices is Associated with Reported News Items

Abstract

We examine extreme volatility events for 1987-99 in five S&P sub-aggregate indices: financial, industrial, mid-cap, transport, and utility. Identified event days are those for which the ratio of subsequent to prior return variance falls into the 1% critical range of a two-tailed F test. In sharp contrast to prior literature, 63.5% of extreme volatility increases and 53.4% of decreases are associated with extraordinary reported news items. More than half are coincident for two or more sectors, but only financials and industrials are positively correlated in event timing. Economically and statistically significant price deterioration accompanies volatility increases, consistent with investors raising required returns. Price and return patterns around volatility decreases are muted.

Keywords: Volatility, variance ratio, stock returns, economic factors, event study

JEL Classification: E44, G12, G14, N22