51 Pages Posted: 30 Jun 2016 Last revised: 5 Sep 2017
Date Written: September 5, 2017
Due to data mining, the expected returns of stock market predictors may be biased. This bias, however, may be mitigated by the journal review process. We develop an estimator for the net bias and apply it to replications of 172 cross-sectional stock return predictors. Bias-adjusted long-short returns are only 13% smaller than in-sample long-short returns. This small bias comes from the dispersion of t-stats across predictors, which is too large to be accounted for by noise, indicating that many predictors have positive true returns. The bias is too small to account for the deterioration in average returns after publication (p-value = 0.0002), suggesting an important role for mispricing. Among predictors that can survive journal review, a low t-stat hurdle of 1.8 controls for multiple testing using statistics recommended by Harvey, Liu, and Zhu (2015).
Keywords: Stock return anomalies, publication bias, data mining, mispricing
JEL Classification: G10, G12
Suggested Citation: Suggested Citation
Chen, Andrew Y. and Zimmermann, Tom, Publication Bias and the Cross-Section of Stock Returns (September 5, 2017). Available at SSRN: https://ssrn.com/abstract=2802357