The Cross-Sectional Anchoring of Forecasted Earnings per Share and Expected Stock Returns

摘要

In this paper, we document that stock returns are significantly higher for firms with high industry median-adjusted forecasted earnings per share (FEPS) than for firms with low industry medianadjusted FEPS. We call this phenomenon the cross-sectional anchoring measure of FEPS effect or the AMFEPS effect. The AMFEPS effect cannot be explained by risk factors, the book-tomarket and earnings-to-price effects, and price and earnings momentum. We hypothesize that the AMFEPS effect is attributed to analysts' anchoring biases. The anchoring bias hypothesis suggests three immediate testable implications. First, the AMFEPS effect is more profound for firms with greater information uncertainty. Second, analysts' FEPS is more overestimated for low AMFEPS firms than high AMFEPS firms. Third, earnings surprises are relative more positive for high AMFEPS firms than for low AMFEPS firms. Our results support all these hypotheses. Finally, we find that analysts revise their earnings forecasts upward more aggressively for stock-split firms, leading to larger earnings forecast errors and more negative changes of earnings surprises for these firms than for their non-stock-split counterparts. These results are also consistent with the anchoring bias hypothesis.