The Influence of Market Conditions on Event-Study Residuals

This paper presents evidence that the mean-adjusted returns and raw-market returns models are misspecified when the event under investigation occurs during either bull or bear markets. To demonstrate this phenomenon, simulation techniques as well as an actual event are employed to examine the reliability of four different return-generating models. When the event occurs during a bull (bear) market, both the mean-adjusted and raw-market returns models produce upwardly (downwardly) biased positive (negative) abnormal returns. This results in statistically significant cumulative abnormal returns over selected preevent and postevent intervals. In contrast, both the market-adjusted and single-index models show far less evidence of any unusual price activity over these same intervals.