Disagreement, Underreaction, and Stock Returns

Abstract

We explore analysts’ earnings forecast data to improve upon one popular disagreement measure — the analyst forecast dispersion measure — proposed by Diether, Malloy, and Scherbina (2002). Our analysis suggests that changes in the standard deviations of forecasted earnings can work as a complementary disagreement measure that is comparable across stocks and immune from other return-predictive information contained in the normalization scalars of analyst forecast dispersion measures. We also document evidence that the change-based disagreement measure predicts future cross-sectional returns significantly only when changes in the mean forecasts are negative. This finding suggests that the interaction between disagreement and underreaction to earnings news affects asset prices.

Keywords: Disagreement, short-sale constraints, underreaction, cross-section of stock returns

JEL Classification: G02, G12, G14