Good News, Bad News and Garch Effects in Stock Return Data

Abstract

It is shown that the volume of trade can be decomposed into proportional proxies for stochastic flows of good news and bad news into the market. Positive (good) information flows are assumed to increase the price of a financial vehicle while negative (bad) information flows decrease the price. For the majority of a sample of ten split-stocks it is shown that the proposed decomposition explains more GARCH than volume itself. Using the proposed decomposition, the variance of returns for younger split stocks reacts asymmetrically to good news flowing into the market, while the variance for older split-stocks reacts symmetrically to good news and bad news.

JEL classification codes:

• C32
• G14

Key words:

• information flows
• autocorrelation