Comprehensive Use Of Internal And External Data Sources In The Operational Risk Measurement

In recent years, with China's rapid economic growth,China's financial industry has entered the stage of rapid development。However, the rapid development in the financial industry also brought operational risk frequently, and took negative impact for China's economic development and financial stability. Operational risk has become one of the main risks which restricted development of China's financial industry, How to measure the operational risk accurately is challenge that we face. However, because we pay attention to the risks of operating so late in China and collect less data in the loss of operational risk, the lack of data is already become the most serious problem as the constraint to the development of operational risk measure. Although the development of international operational risk measurement model is more rapid, most of the models require a higher loss of data, so most of the model is not suitable for measurement of national operational risk. How to solve the loss of data in operational risk effectively is already become the hot point and problem in financial theory, which is also the aim of study and research purpose.At present, the operational risk measurement methods are top-down and bottom-up. Top-down measurement model includes the basic indicator approach, income volatility method, model of income and expenditure, CAPM, securities factor model, model of the operation lever and so on. Bottom-up operational risk measurement models also include internal measure, loss of sub-step, Bayesian network method, the scorecard method and the extreme value theory. Most of the international financial institutions in their own operational risk measurement choose bottom-up approach.Because POT model could be used as the solution of operational risk loss distribution of the "Fat Tail" features, and more suitable for high and frequency measurement of operational risk losses, this paper selects the POT model to measure operational risk, and introduces reliability theory——Buhlmann credibility model, to measure operational risk in the mixed internal data and external data, in order to obtain solution to the problem of insufficient data in operational risk measurement approach, which is the major innovation of the paper.This paper collects total of 115 cases on Bank of China and Agricultural Bank of China's operational risk loss, through the authority of the national news media and major newspapers, and establishes sample data of the operating losses. This paper proposes hybrid method of external data and takes empirical research by using these data. The empirical results also show that the method can reduce the extraction of BOC's regulatory capital, meanwhile, also shows the feasibility of the proposed method.Finally, the paper also described the inadequacy of this method are and proposed direction for further research.