Factors affecting SMEs credit risk and credit risk assessment based on blockchain-driven supply chain finance

The purpose of this research is to examine the factors affecting SMEs’credit risk and credit risk assessment based on blockchain-driven supply chain finance. This research mainly includes three objectives: The first objective is to examine whether the financing enterprises, core enterprises, assets position under financing, blockchain platform and supply chain operation have significant impacts on credit risk by using logistic regression and entropy method. The panel data were collected from CSMAR on fifty-six SMEs, eight core enterprises and twenty-six blockchain enterprises in the period of 2016-2020. The second objective is to establish a credit risk evaluation index system and used factor analysis to extract the principal factors, then 11 factors are extracted as the variable sources for credit risk assessment modeling. The third objective is to build a credit risk assessment model by using five methods: Classification Tree, Bagging algorithm, AdaBoost algorithm, Random Forest and Logistic Regression to construct the credit risk assessment model. Then, according to the model evaluation criteria, this research found out the credit risk assessment model with the best prediction classification performance. The findings show that the financing enterprises, core enterprises, assets position under finance, blockchain platform, and supply chain operation have significant impacts on SMEs’credit risk when the confidence level is 90%. In general, the performance of AdaBoost algorithm model is the best. It has the strongest ability to distinguish between enterprises with credit risk and without credit risk, and has strong stability. The research not only enriches the theories and method of credit risk assessment of SMEs, but also provides assistance in solving the problem of financing difficulties for SMEs due to its ability to accurately assess credit risk.