Publication Bias and the Cross-Section of Stock Returns
65 Pages Posted: 30 Jun 2016 Last revised: 2 Nov 2018
Date Written: November 1, 2018
We develop an estimator for publication bias and apply it to 156 replications of published hedge portfolios. Bias adjusted returns are only 12.3% smaller than sample returns with a standard error of 1.7 percentage points. The small bias comes from the dispersion of returns across predictors, which is too large to be accounted for by data-mined noise. Our estimate uses only in-sample data and is free of economic effects that cloud out-of-sample tests. An alternative dataset of 77 hand-collected portfolios leads to a 9.9% adjustment. The bias is much smaller than post-publication decay (p-value < 0.0001), suggesting mispricing is important.
Keywords: Stock return anomalies, publication bias, data mining, mispricing
JEL Classification: G10, G12
Suggested Citation: Suggested Citation