Bias in estimating the systematic risk of extreme performers: Implications for financial analysis, the leverage effect, and long-run reversals

► Explains how bias can arise systematically in beta estimates of extreme performers. ► Evidence shows this bias explains the large shifts in these systematic-risk estimates. ► Methods for conditioning betas on out-of-sample returns are lacking. ► Reconciles disparate results of earlier papers on the risk of extreme performers. ► Provides recommendations for identifying this bias and correcting for it.