A Novel Credit Rating Model: Empirical Analysis from Chinese Small Enterprises

ABSTRACT

This article establishes a novel credit rating model for small enterprises, thereby solving the problem that commercial banks cannot accurately obtain financial information about small enterprises or reasonably evaluate the credit risk of small enterprise loans. Through the identification of default status and removal of redundant information for indicator screening, this article adopts a weighting method that can be used to classify small enterprises. The empirical results show that the discriminant precision of default status by the credit rating system of China’s small enterprises, constructed by this article, is up to 91.9%. The weighting results show that in the credit rating of small enterprises, financial indicators cannot reflect all the liabilities, and the role of qualitative indicators in credit ratings is more important. This article empowers 30 indicators based on the principle of distinguishing different types of customers, classifies customers into nine different levels, and avoids the unreasonable phenomenon that default customers have higher credit scores than non-default customers. The results can help commercial banks to distinguish customers of different significance levels.

KEYWORDS:

• Credit rating model
• small enterprises
• neural network
• partial correlation coefficient

Disclosure Statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

The research was supported by The Liberal Arts and Social Sciences Foundation of the Ministry of Education in China [No. 21YJCZH107]; The National Natural Science Foundation of China [No. 71731003; 71831002; 71873103; 71971034; 72072018]; and Credit Risks Rating System and Loan Pricing Project of Small Enterprises for Bank of Dalian [No. 2012-01].