The History of the Cross Section of Stock Returns

Abstract

Using data spanning the 20th century, we show that most accounting-based return anomalies are spurious. When examined out-of-sample by moving either backward or forward in time, anomalies' average returns decrease, and volatilities and correlations with other anomalies increase. The data-snooping problem is so severe that even the true asset pricing model is expected to be rejected when tested using in-sample data. Our results suggest that asset pricing models should be tested using out-of-sample data or, when not feasible, by whether a model is able to explain half of the in-sample alpha.

Keywords: Factors; anomalies; asset pricing; data-mining

JEL Classification: G14