Quantitative Management Of Credit Risk Of Commercial Bank On Logit Regression Model

As the deepening reform of China’s financial system, growing level of opening-up financial industry,and innovative products becoming more and more sophisticated.The risks faced by domestic commercial banks tend to diversify during the business management process. Risks of such complex forms put forward higher requirements on the management level of domestic banks.Currently, operating process of domestic commercial banks is confronted with credit risk, market risk, interest rate risk, liquidity risk and operational risk, etc, among that credit risk is the most important and one of the oldest type of the risks. In financial history,there are numerous instances of bank failures due to improper credit risk management.With the rapid development of domestic financial markets, gradually increased pressure of international regulatory, increasingly fierce competition between our banks, commercial banks must carry out a more flexible, positive and proactive management of their own credit risk.Based on Finance, Investment, Quantitative Economics, this paper carry out an in-depth analysis of a common model in Credit Risk Measurement Models, a commonly used model—logit model. This paper details the model’s theoretical analysis, software operation, and the use of detailed models. This paper selects a large number of listed companies as samples, extracts the number of each sample’s depth financial data statistical analysis, the final logit model was constructed based on commercial bank credit risk measurement models, test results show that the model the credit risk of commercial banks prediction accuracy rate is very high, it is worth to promote the use in practice.In this paper, there are still some shortcomings, for example the model which used is just a binary choices logit model, so the classification of the credit risk is more rough, but if we use multiple logit model instead of the model that used, the quantification of credit risk will be more accurate. However, firstly multiple logit model requires more groups of listed companies classified, and it is extremely difficult to collect the current data. So this article can not be using multiple logit model of corporate credit risk. There are many issues worth exploring, I will be in future academic career to these issues to the relevant experts and scholars to study, to deepen understanding of these issues.