This paper examines whether the explosion of proposed investment factors has made traditional statistical significance standards too easy to pass. The authors argue that extensive data mining raises the bar considerably, suggesting many celebrated return predictors may reflect statistical noise rather than genuine market relationships.
… and the Cross-Section of Expected Returns
Yan Liu, Campbell Harvey
Research
101 Pages
Key Takeaways
Higher Statistical Hurdle: A newly proposed factor should clear a t ratio above 3.0, rather than the traditionally accepted threshold of roughly 2.0.
Factor Proliferation Problem: The authors document more than 300 factors proposed across decades of asset pricing research, substantially increasing the risk of false discoveries.
Standards Keep Rising: Using evidence beginning in 1967, the authors project significance thresholds forward 20 years as researchers continue testing additional return factors.
Join our newsletter to have all of this content + Exclusive Newsletter Bonus Content delivered to your inbox every week
