Myth Busted: Stock Return Anomalies Revisited

Sammanfattning: Research has uncovered over 450 anomaly factors that exhibit stock return predictability. However, after anomalies are published and studied in successive literature, the return predictability often seems to attenuate or disappear. This raises the question of whether return predictability existed in the past, but have been arbitraged away, or whether published anomalies simply is an artifact of p-hacking. Using out-of-sample analysis, we study 21 eminent accounting-based stock return anomalies and show that a majority of the cross-sectional return predictability can be attributed to p-hacking.