The Role of Analysts in Differentially Interpreting Earnings Surprises

Abstract

It is commonly believed that relatively large absolute earnings surprises trigger more differential interpretations compared to small earnings surprises (e.g., Bamber 1987; and Morse et al. 1991). However, prior research provides no direct support for this belief. This study tests the prediction that, when forecasting future earnings, the weight the average analyst places on their own idiosyncratic information processing skills (or the degree to which they differentially interpret earnings) increases more after relatively large absolute earnings surprises. This prediction is based on the argument that less persistent earnings surprises (in this case relatively large earnings surprises) cause analysts to rely less heavily on the information commonly observed in earnings surprises, and to rely more heavily on their own idiosyncratic information processing skills. Using forecasts of quarterly earnings, we examine changes in the relative weight analysts place on their private information around prior quarterly earnings announcements. Consistent with our expectations, we show that the weight analysts place on their private information increases more after relatively large earnings surprises compared to small earnings surprises. We also show that the weight analysts place on their private information increases more after negative earnings surprises, which are also less persistent. Our evidence suggests that the relative importance of analysts' interpretive informational role varies across earnings announcements, and is more prevalent when there is likely to be more demand for analysts' differential interpretations. We show that this effect serves to improve the accuracy of the mean forecast.