Stock Recommendations as a Source of Bias in Earnings Forecasts

Abstract

Recent scandals and controversies have focused substantial attention on the behavior of financial analysts. Responses such as the Sarbanes-Oxley Act, new regulations at securities exchanges, and massive legal settlements are consistent with the perception that analysts' research and stock recommendations exhibit significant self-serving bias. While anecdotal and legal evidence support the allegations that some analysts have intentionally mislead the investing public, recent archival research suggests unintentional cognitive processes also contribute to systematic bias in analysts' forecasts (Eames et al. 2002). However, studies based on stock-market data cannot distinguish between unintentional cognitive processes and intentional bias stemming from economic incentives (e.g., trade boosting). In a laboratory experiment we eliminate economic incentives and find that cognitive processes unintentionally lead to earnings forecast bias. Our results suggest that recent regulations and policy changes by Congress, the Securities and Exchange Commission, exchange markets, and brokerage firms will not totally eliminate bias in analysts' earnings forecasts.