Dispersion, Coverage, and the Rational Investor: Using Analysts ’ Earnings Forecasts to Predict Stock Returns

Researchers have often used intrinsic properties of stocks such as earnings per share ratios and dividend yields to predict stocks ’ future performances. In their paper Differences of Opinion, Diether et al. (2002) propose a new, non-intrinsic predictive variable based on analyst earnings forecasts. Using their methods, we validate the claim that greater differences of opinion in analyst predictions of future stock returns tracks poorer future returns of a stock and that small stocks and poorly performing stocks are especially affected. Going further, we also found that the dispersion effect was greatest for mid-level analyst coverage and that market performance was negatively related to the dispersion effect. Also, by considering more realistic and rational investor behavior, we are able to more confidently reject the hypothesis that dispersion effects can be explained by Fama-French factors. However, inclusion of more realistic rational investor behavior led to different portfolio behavior, and decreased explanatory power of dispersion, suggesting that the dispersion effect is not as robust as previously thought