China's Commercial Banks Operational Risk Measurement Methods Based On Extreme Value Theory Research

Firstly, this paper made a preliminary elaboration of operational risk from three aspects:definition, classification and feature.Secondly, based on the reality of China's banking industry, considered internal fraud and external fraud as example, this paper explored the major factors which effect the occurrence of operational risk with mixed strategy model of gaming theory. The results showed the major factors as follow:cost and rank of perpetrator, as well as the possible punishment they would suffer; cost and expectations of examiner, as well as the possible reward they would receive; time interval from the incident of the case to exposure.Thirdly, this paper tried to find the most appropriate modelling method to measre operational risk of China's banking industry by comparing existing measurements from the view of regulator and academician respectively. The result showed that the Extreme Value Theory has Comparative advantage among others. On this basis, using Copula function to describe the correlation among different loss events can significantly improve the measurement accuracy.Finally, this paper used POT model and Gumbel Copula to make an empirical study of operational risk of China's banking industry with loss data collected from open source. In this way, the Value at Risk would be remarkable smaller than the result calculated from traditional way. That means banks can save large amount of economical capital which certainly benefit to their daily operation. On the other way, it will also reduce the difficulty for the regulators so that regulatory cost can be saved.