System Design Of Regression Analysis Based Scenario Prediction For Banks’ Operational Risk

In recent years, more and more banks have started to implement new Basel Accord of risk capital of bank. The first pillar of accord states that all banks must implement risk calculation with respect to theirs’ size & regulators’ request, then reserve the capital derived from its risk calculation. Financial institutions may choose anyone as long as regulators think it makes sense. As a result, the risk capital results are historical oriented; hence the future oriented forecast capability is need. Usually loss of bank varies along with the changes of revenue, business scale and risk control level of the bank. The regression equation of the above risk factors with loss can be derived by analyzing the data which is available. Then business & risk manager can adjust the risk factors according to current situation or his judgment and get future oriented risk capital.Firstly, Basel Accord & Operational Risk (OpRisk) is discussed. Then the drawbacks of AMA approach of RongTong Bank is pointed out and the necessity of improving the current AMA approach by regression analysis based the scenario prediction is confirmed. Regional OpRisk prediction & business line OpRisk prediction are discussed. Main flows of entitlement application, classified loss report & risk prediction are stated. After that detail design of core subsystems:loss entry & approval scrubbing, data pre-handling & risk prediction modeling are stated. "Risk prediction modeling" is elaborated in detail which utilized transformed multi-variable nonlinear regression analysis to set up the connection between OpRisk and key risk factors. Then the correctness is verified by mathematical means. At last from the comparison of RongTong bank with other banks, pros & cons of OpRisk measurement are made.