Contemporary event study test: event-induced variance and cross correlation among abnormal returns in dividend

Although many literatures related to event studies have reported the problem of event-induced variance and cross correlation among abnormal returns, a lot of researchers still employ conventional event-study methods which tend to reject the null hypothesis of zero average abnormal returns too frequent when it is true (higher type I error). In this paper, we applied a more advanced event-study method, namely the adjusted Boehmer, Mucumeci, and Poulsen (Adj-BMP) test, to provide a remedy to the issue of event-induced variance and cross correlation among abnormal returns. Using cash dividend increase to evaluate a battery of both statistical tests, the empirical results found the presence of the cross-correlation among abnormal returns. Consequently, the Adj-BMP test produces four significant abnormal returns from day 10 before the event to day 30 after the event while the BMP test generates eight significant abnormal returns. The BMP test exhibits 100% over-rejection of null hypothesis. At the same time, the level of significance has been decreased from 5% to 1% in the BMP test to 10% to 5% in the Adj-BMP test. Thus, we show that the Adj-BMP test is a robust test in presence of cross correlation among abnormal returns. According to the Adj-BMP test, this study found that there is an impact of cash dividend increase events on the average abnormal returns. This study makes a major contribution to research on providing an empirical comparison between BMP test and Adj-BMP test to resolve event-induced variance and cross correlation among abnormal returns in event studies of emerging market.

Preview Text 2) Contemporary Event Study Test.pdf Download (1MB) | Preview Official URL or Download Paper: http://www.ijem.upm.edu.my/vol12_noS2/2)%20Contemp...

Actions (login required)

View Item