Assessing and Predicting Small Enterprises’ Credit Ratings: A Multicriteria Approach

Abstract

AbstractCredit ratings play a key role in helping financial institutions to make loan decisions and to reduce the financial constraints on small and medium-sized enterprises. However, small enterprises have made it difficult for financial institutions such as commercial banks to accurately determine their credit risk, creating salient loan difficulties, due to the short duration, high frequency, urgent demand for credit, and small amount of their loans. In order to alleviate the difficulties of financing small businesses, this paper develops a new approach for the assessment of credit risk in small enterprises by combining high-dimensional attribute reduction methods with fuzzy decision-making methods. Based on 687 small enterprises in a regional commercial bank of China, we find 17 indicators that have a significant impact on the default risk of small enterprises. Then, it utilizes TOPSIS together with fuzzy C-means to grade the credit ratings of enterprises requesting loans. The standard discrimination and ROC curve dual tests resulted in the prediction accuracy of the standard indicator system reaching 85.40 percent and 90.09 percent, respectively, indicating the strong default discrimination of this rating system and its practicability in commercial banks and other financial institutions.KeywordsCredit ratingDefault riskFuzzy C-meansSmall enterprises