Overreaction and Underreaction in Analysts' Forecasts: A Pre- Versus Post-Crash Investigation

Abstract

The financial markets crash, which occurred March-April 2000 has aroused suspicion about financial analysts' forecasts accuracy which could have made them more cautious, perhaps pessimistic about future performance. This article investigates the possibility of recent behavioral changes by examining analysts' reactions to good and bad news before and after the crash. We analyze the interaction between over- and under-reaction and the optimistic bias following Easterwood and Nutt (1999)'s methodology. If the crash leads analysts to be less prone to optimism (or even pessimism), we might observe weaker underreaction (or possibly overreaction) to bad news and weaker overreaction (or possibly underreaction) to good news.

Our results indicate a strong decrease in optimism after the crash. The tendency to produce overly positive and moderately negative revisions or forecasted changes diminishes or reverses. Reactions to new information are consistent with a reduction in optimism: analysts' systematic underreaction to negative information strongly decreases after the crash. Overreaction to positive information, less pronounced before the crash, is also weakened.

Note: Downloadable document is in French.

Keywords: Analysts' estimates, accuracy, crash

JEL Classification: G14, D83