The Prediction of Post-Earnings Announcement Drift Using a Measure of Market Underreaction

• In Event: Market reaction to earnings
  In Paper Session: Market Reaction to Earnings

Abstract

In this paper we develop a theory-based empirical measure of underreaction to earnings announcements that results from an asymmetric usage of earnings information. In our model, the squared correlation coefficient between order imbalance and earnings surprise (rho) is the empirical measure of underreaction that one can use to predict the size of the post-earnings announcement drift (PEAD). Specifically, we show that PEAD=k*rho-squared , where k is the information content of earnings. The predictive power of is higher than that of firm attributes prior studies have identified to explain PEAD, including the bid-ask spread, opinion divergence, firm size, analyst following, and institutional ownership. A trading strategy based on generates the Pastor-Stambaugh four-factor alpha of 8.25%.

Authors

• Steve C Lim, Texas Christian University

• Oliver Kim, University of Maryland College Park

• Kee H Chung, University at Buffalo, SUNY

• Sean Yang, University at Buffalo, SUNY