A Comparison of Event Study Methodologies Using Daily Stock Returns: A Simulation Approach

Abstract

Accounting research studies commonly use measures of abnormal security returns to test hypotheses about accounting information and policies. Alternative methods are used in these studies to detect abnormal performance and thereby examine the information content of identified events. The ability of these methods to detect abnormal returns using simulation techniques is examined by Brown and Warner [1980; 1984]. Morse [1984] addresses this issue analytically. The analytical approach presumes that the event study data is generated by the process modeled, while the simulation approach uses the security return data generated by the market. Simulation provides a tractable means of dealing with situations where an analytical approach may yield results suggesting direction but not magnitude or where such techniques are unusually cumbersome. Both limitations face us in this study. Brown and Warner [1980] use monthly data to examine the comparative abilities of several combinations of return-generating models and