Detecting Information Content of Corporate Announcements Using Variance Increases: A Methodological Study

Abstract

This paper evaluates various methodologies used in event studies to detect the information content of corporate announcements using the increase in the variance of the security returns. Simulation with daily stock return data show that the Patell procedure, the most commonly used method to detect information content, rejects the null hypothesis of no increase in event-induced return variance too frequently when the null is true. A non-parametric cross-sectional rank test of squared abnormal returns similar to that in Corrado (1989) is proposed that is better specified under the null hypothesis and more powerful under the alternative hypothesis than the parametric tests.