Rating on a behavioral curve

Abstract

Sell-side analysts rate on a particular type of behavioral curve: recency. Although they claim to use objective criteria (like expected raw, market-adjusted, or industry-adjusted returns), we find that, even after controlling for these claims, their recommendations on a particular stock are negatively influenced by their assessment of the quality of the few other stocks they have rated that month. This recency bias has price implications. The next day's alpha of a sophisticated trading strategy that incorporates this bias is about 40 % higher compared to the alpha of an unsophisticated strategy that uses rating information only.