This paper examines whether hundreds of widely cited market anomalies actually survive replication under stricter testing standards. After reviewing 447 anomalies, the authors argue many disappear once microcaps and data mining biases are addressed, challenging parts of the investing ecosystem and raising questions on “alpha”.
Replicating Anomalies
Alpha Architect
Wesley Gray
Research
6 Pages
Key Takeaways
Replication Failure Scale: 286 of 447 anomalies, including 95 of 102 liquidity signals, failed conventional 5% significance thresholds after controlling for microcap distortions.
Microcap Dependence Exposed: Microcaps represented 61% of stocks but just 3.28% of market capitalization, helping explain why equal weighted anomaly strategies often looked stronger historically.
Factor Models Absorb Alpha: The q factor model rendered 115 of 161 previously significant anomalies statistically insignificant, suggesting many “alpha” signals may reflect broader systematic exposures.
Join our newsletter to have all of this content + Exclusive Newsletter Bonus Content delivered to your inbox every week
Related Content
