Management of unexpected project risk events by the icss and dummy garch model

Unexpected project risk events are important for managers to evaluate and control large-scale projects. In the past, these issues are widely handled by event studies. However, event studies have been criticized for its irrational assumption, such as normal distribution and constant variance. In this paper, we consider a dummy GARCH model, which releases the assumption of the normal distribution, to deal with the impact of unexpected project risk events. In addition, the important events are detected automatically by using the ICSS, instead of determining by decision makers. From the result of our empirical studies, it can be seen that the flexibility of the proposed method is better than that of traditional event studies.