A Re-Examination of the Importance of Analyst and Forecast Characteristics for Forecast Accuracy: The Effects of Analyst Disagreement, Stale Forecasts, and the Information Environment

Abstract:     
In this study, we investigate whether the importance of previously identified analyst and forecast characteristics for analyst forecast accuracy varies with the level of disagreement among analysts. We find a large variation in analyst forecast accuracy when analyst disagreement is high, suggesting that investors can form better expectations of earnings if they can identify analysts whose forecasts are likely to be more accurate. We also find that analyst forecast accuracy is positively associated with their prior year accuracy and negatively associated with forecast horizon. The importance of prior year accuracy and forecast horizon increase as the disagreement among analysts increases. Unlike prior studies suggesting that forecast horizon is the single most important factor explaining analyst forecast accuracy, we find that analysts’ prior year accuracy is the most important factor for analyst forecast accuracy in the two highest quintiles of forecast dispersion. In the lowest two dispersion quintiles, prior year accuracy and forecast horizon have similar effects. The inconsistencies between our findings on the importance of forecast horizon and those of prior studies are due to the effect of the information environment (i.e., Regulation Fair Disclosure (Reg FD)) and the exclusion of stale forecasts. Specifically, we find that a substantial decrease in the importance of forecast horizon for forecast accuracy post-Reg FD. Furthermore, experience and brokerage house size are not positively associated with forecast accuracy post-Reg FD. This suggests that analysts’ access to private information was correlated with these characteristics in the pre-Reg FD environment and that access to private information affects their importance for forecast accuracy.

Number of Pages in PDF File: 46

Keywords: analyst forecast accuracy, analyst disagreement, stale analyst forecasts, information environment

JEL Classification: M4, M41

working papers series