Nonstationarity of beta and tests of market efficiency

Conclusions from event type studies are usually supported by data generated from some form of the market model. This study examines the robustness of these conclusions to different static and dynamic estimates of beta when the event under investigation occurs during a period of changing systematic risk. The results indicate that when beta is nonstationary, the findings of market inefficiency (or efficiency) in previous studies may be an artifact of the static beta estimation method.