The False Discovery Rate in Finance: Identification Failure and Search-Adjusted Estimation

ADIA Lab Research Paper Series

Authors: Marcos López de Prado, Frank Fabozzi

Date Published: March 2026

Recent empirical studies conclude that publication bias in financial economics is modest and that most published factors in the asset pricing literature represent genuine discoveries. These conclusions are obtained by inferring the false discovery rate (FDR) from the cross-sectional distribution of reported in-sample test statistics. This paper shows that such inference faces a fundamental identification problem. We prove that FDR cannot be identified from in-sample statistics alone when reported results arise from a search-and-selection process.