Earnings Surprises and the Cross-Section of Stock Returns

Abstract: Do earnings surprises affect stock prices during the subsequent quarter? If so, what is the estimated impact, and to what extent can it be clearly distinguished from other factors? To answer these questions we build ten dynamic portfolios in which the companies are continuously reallocated according to their latest earnings surprise. A cross-sectional regression based on these portfolios indicates a distinct albeit nonlinear effect of the earnings surprise. To check whether the apparent effect of earnings surprises can be explained by other factors we test whether the renowned Fama-French three-factor model is able to explain the observed variation in returns across the portfolios. We then augment the Fama-French three-factor model with a factor based on earnings surprises and study the explanatory power of this fourth factor. By adding the forth factor the model is slightly improved. We also show that by "spanning" out the factors over more portfolios the model yields more reliable values. However, some issues arise associated to the linearity assumption of the model and thus we start to develop a nonlinear model. Finally, we present a simple trading strategy based on the spread between the positive and negative surprise portfolios.