Understanding Alpha Decay

30 Pages Posted: 18 Apr 2017 Last revised: 20 Nov 2020

Date Written: October 20, 2020


I argue that academic research often inadequately accounts for alpha decay. As an anomaly’s alpha (i.e., the risk-adjusted expected excess return) and realized returns are negatively related, alpha decay coincides with positive realized returns. If the alpha decays at publication, observers may thus misinterpret these repricing returns as evidence that the anomaly will persist in the future, when the anomaly is instead going extinct. Ignoring this negative relationship between alpha and realized returns is also problematic for asset pricing tests. Most tests make a stationarity assumption that rules out any variation in the unconditional alpha. Even if this assumption holds, decay in the conditional alpha can bias the unconditional alpha’s estimate upward. Using results in the literature, I find that the bias is about 1.4% per year for a typical anomaly. I provide a simple formula to remove the bias and show how to incorporate alpha decay tests into the asset pricing toolkit.

Keywords: Anomalies, Cross-Sectional Return Predictability, Market Efficiency

JEL Classification: G00, G12, G14

Suggested Citation

Pénasse, Julien, Understanding Alpha Decay (October 20, 2020). Available at SSRN: https://ssrn.com/abstract=2953614 or http://dx.doi.org/10.2139/ssrn.2953614

Julien Pénasse (Contact Author)

University of Luxembourg ( email )

4 Rue Albert Borschette
Luxembourg, L-1246

Do you have a job opening that you would like to promote on SSRN?

Paper statistics

Abstract Views
PlumX Metrics