Do Analysts Under-React to Bad News and Over-React to Good News?
29 Pages Posted: 26 Dec 2000
Date Written: December 2000
Easterwood and Nutt (1999) show that analysts under-react to bad news in past earnings changes (or forecast errors) but over-react to good news in past earnings (or forecast errors) consistent with analysts exhibiting systematic optimism. We find that their results are sensitive to the cutoff used to define outliers.
More specifically, when we define outliers as observations with the absolute value of earnings-related variables exceeding 100% of stock price, we obtain results similar to those reported in their study. However, when we define (and exclude) outliers as observations with the absolute value of earnings-related variables exceeding 20% of stock price, we find evidence of under-reaction to the information in past earnings and forecast errors. However, the under-reaction to bad news is not significantly greater (in magnitude) than the average under-reaction. Furthermore, we find some evidence of overreaction to good news in past earnings changes but no evidence of overreaction to the information in past forecast errors.
With a 5% cutoff, there is no evidence of over-reaction to good news in past earnings changes or forecast errors. Furthermore, there is no evidence of over or under-reaction to the information in past returns. The last result suggests that the evidence of under-reaction to the information in past returns observed in previous studies may have been overstated. Overall, our results cast doubt on the hypothesis that analysts exhibit systematic optimism with respect to past information.
Keywords: Analysts' forecasts, systematic optimism, under-reaction
JEL Classification: G14, G29, M41
Suggested Citation: Suggested Citation