Holding Size While Improving Power in Tests of Long-Run Abnormal Stock Returns

Barber and Lyon (1996a) and Kothari and Warner (1996) document conventional tests of long-run abnormal returns are misspecified. In this research, we propose alternative methods to test for long-run abnormal returns. Our methods have two key characteristics. First, long-run abnormal returns are calculated using reference portfolios that yield an abnormal return measure with a population mean that is identically zero. Second, our methods control for the documented positive skewness in long-run abnormal returns calculated using reference portfolios. We control for the positive skewness by either (1) adjusting conventional t statistics using well-documented statistical methods, or (2) generating the empirical distribution of mean long-run abnormal returns via simulation. In addition to yielding reasonably well-specified test statistics in a variety of sampling situations, we document that these two methods are more powerful than the control firm approach analyzed by Barber and Lyon.