Credit Risk Assessment Based On Small And Medium-sized Manufacturing Enterprises

As my country’s economy turn to a high-quality development stage,it also drives the high-quality development of the financial market.For small and medium-sized enterprises,due to imperfect systems and high management costs,they occupy most of the market share,providing more employment opportunities,so small and mediumsized enterprises are increasingly valued and supported by the government.However,due to the fact that the information of small and medium-sized enterprises is not disclosed or even deliberately concealed,the bank refused to lend to small businesses.In order to solve the financing difficulties problem for small and medium-sized enterprises,we need to establish an effective credit risk assessment system to reduce bank losses and provide better services for small and medium-sized enterprises to achieve a win-win situation.This paper takes China’s manufacturing small and medium-sized enterprises as the research object,from four aspects: the theories and methods of credit risk assessment,the characteristics of manufacturing and credit risk status of small and medium-sized enterprises,selecting credit risk indicators and building model system for manufacturing small and medium-sized enterprises.After reviewing domestic and foreign literatures,it is found that scholars’ research on credit risk includes qualitative analysis and quantitative analysis,and the selected indicators are still mainly financial indicators.In the research of this paper,non-financial indicators related to highquality development and innovation ability are added.At that time,according to the characteristics of China’s manufacturing industry,this paper also added indicators specific to the manufacturing industry,such as labor productivity,solid waste utilization rate,and product quality qualification rate.A total of 32 indicators were selected.First,logistic regression and random forest model were used to analyze the credit risk of the manufacturing small and medium enterprises.Then combine the two methods,first use random forest to filter the characteristics of small and medium enterprises credit data,and then put the filtered data into logistic regression for modeling.The results show that the combination of random forest model and Logistic model significantly improves the accuracy of credit risk assessment of small and medium-sized manufacturing enterprises in China.In addition to financial indicators,this paper adds manufacturing-specific indicators,namely product qualification rate,labor productivity,and Huazheng ESG,an indicator representing the environment,society,and corporate governance.The results show that these indicators are effective in affecting the default of manufacturing SMEs.This indicator has good applicability to the credit risk assessment of China’s manufacturing SMEs at this stage.