The Distribution Approach Of Banks' Operational Risk Measurement Based On Stable Distribution

The serious loss in operational risk incidents attracts widely attention of the whole banking. In the New Basel Capital Accord operational risk is brought into the risk management framework for minimum capital supervision, which pushed the development of operational risk management. Because the study on operational risk is late in our country and the management level is lagged, operational risk incidents happened frequently. So how to improve operational risk measurement technique is the urgent task of our banking.Firstly, the paper expounds the commutation, the sorts and the么measurement approaches of operational risk. The measurement approaches include Basic Indicator Approach, Standardized Approaches and Advanced Measurement Approach. And Advanced Measurement Approaches also include Internal Measurement Approach,Extreme Value Theory Approach and Loss Distribution Approach. Especially, the Loss Distribution Approach (LDA), one of the advanced measurement approaches is researched here.LDA is a statistical method for calculating Capital-at Risk using a Value-at-Risk measure. It models the loss distribution based on two hypothesizes: loss frequency distribution hypothesis and loss severity distribution hypothesis. The paper aims at improve LDA using stable distribution for simulating loss severity distribution. It has been proved that the loss severity distribution always has a fat tail and the stable distribution can simulate the distribution with fat tail better than others.At last, the paper use a loss events sample of operational risk published on the Internet between 2000 and 2004 to calculate the Capital-at Risk of state banks.