The Research Of Model And Empirical Study On Credit Risk Of Commercial Bank

Since the 20th century, the international commercial bank has made the considerable development, the scale has expanded and the strength has strengthened continuously. The commercial banking business multiplies gradually, the proportion of business different of saving and loan enhances gradually, it is advantage to reduction of credit risk. However, the essence depends upon lendings does not have any change. Therefore, with the business expanding and the new service opening gradually, the commercial bank also even more takes the credit risks.The credit risk is the main risk which the current banking industry faces, the Basel Committee proposed the stricter calculated tool for the risk capital, moreover, has also done the research and updated the tool to control credit risk effectively. Therefore, the " Basel Accord" proposed by the Basel Committee in 1988 declared that the international bank should take this"Accord"as the standard, and the "New Basel Accord " in 2004 proposed the more flexible risk capital computation principle. The domestic and foreign academic uses their own data separately according to the request to study the credit risks discrimination models, mainly concentrates in estimating an important target Default of Probability (PD) in the primary method of"Basel Accord"and make it into use. In addition, " New Basel Accord" propose that the commercial bank may use the higher method to estimate their own risk, if they have sufficient historical data and have enough ability to estimate their risk. It requests the commercial bank to compute an important target loss given default (LGD). The domestic and foreign acadmic mainly focus on analyzing its influent factors and conclude the data about the target. Nowadays, there is not a mature model for estimating the loss given default(LGD).This article has analyzed the influence of default of probability (PD) in credit risks, summarized the advantages and disadvantages of existing model, and proposes the non-linear double combination distinction model to estimate the default of probability (PD). In addition, this article has analyzed the influencing factor of loss given default (LGD), and carries on the classical statistical method as well as the combination estimate method to estimate by using the influencing factor. Based on, this article comfirms these models and do the contrast by the near 5 years finance data of public company and puts forward the corresponding policy proposal.