Research On Credit Risk Management Based On Early Warning Model For Public Credit Review

The manufacturing industry plays a pivotal role in China’s national economy,and with the "Made in China 2025" strategy,it is necessary to pay attention to the growth of manufacturing enterprises and their capital borrowing status.For a long time,China’s manufacturing output value has been occupying about 40%of GDP,manufacturing enterprises rely on low labor costs,through large-scale production to maintain long-term profit growth,and has become the basis of China’s real economy,the new engine of rapid economic development.Manufacturing enterprises are mostly in the middle and upper reaches of the industrial chain,and the production transactions are highly correlated.Facing the pressure of the economic down cycle,the business situation is not optimistic,and the default of one party of the transaction will make the other party or multiple parties face the problem of capital shortage,and it touches every link in the field of commodity circulation,so the characteristics of risk transmission in manufacturing enterprises are very extensive.Encouraging,supporting and guiding the high quality and high speed development of manufacturing industry is the necessary path of supply-side reform,and it is also the main direction for commercial banks to expand their business scope,achieve structural adjustment and improve their operation system.This thesis starts from the basic concepts such as connotation and characteristics of credit risk,then elaborates the concept,characteristics,time of formation and reasons of credit risk of manufacturing listed companies,and the current situation at home and abroad,followed by the advantages of big data information interoperability based on financial technology applied to build credit evaluation and early warning model of commercial banks,the experimental results of credit risk metrics and The Lasso+Logit model is selected as the empirical model for this study after analyzing the advantages and disadvantages of related models.In the empirical analysis,the original early warning rules were screened using hierarchical analysis,and then three qualitative indicators and three quantitative indicators related to manufacturing enterprises were finally screened using the Lasso model to construct the Lasso-Logit model for early warning of public credit review.In this study,400 financial statement data of manufacturing enterprises in 2021 from A city merchant bank in Guizhou province were taken as the analysis cross-section,among which 266 were non-defaulting enterprises and 134 were defaulting enterprises.Two groups of samples are established,one for the establishment of the credit review early warning model for manufacturing enterprises;the other group is used to test the feasibility of the model.This thesis conducts a large number of empirical studies on the critical value points of the Logit model and establishes several samples in order to improve the credit assessment of manufacturing enterprises.The validity of the optimal default rate is verified according to the ROC model,and the final value of 0.307 is the optimal probability of default,which is more accurate and better able to identify the problematic companies than the model with a p-value of 0.5,which is of greater concern to Commercial Bank.Therefore,it can be concluded that the model has a good overall fit and high accuracy,which can provide effective reference for credit risk analysis of manufacturing enterprises.