Outliers in Finance Research
55 Pages Posted: 21 Oct 2014 Last revised: 11 May 2021
Date Written: May 18, 2015
Articles in the top finance journals largely ignore the potential of outlier-induced bias in empirical research. When finance researchers do address outliers they use techniques that tend to cause additional problems. We illustrate the problems via simulations as well as replications of studies recently published in the top journals. To address the issues, we present new technology that mitigates outlier influence in models and data structures common in finance.
Keywords: Outliers; Winsorizing, Trimming, Outlier-robust regression, Good leverage, Bad leverage,Vertical outlier, MM-estimators, L-estimators, S-estimators, Bias, Influence function, Breakdown point
JEL Classification: C31, C52, C87, G31, G32, G34, G38
Suggested Citation: Suggested Citation