Font Size: a A A

A Fuzzy Clustering Customer Credit-Rating Model And Its Sensibility Analysis For Commercial Banks

Posted on:2004-01-25Degree:DoctorType:Dissertation
Country:ChinaCandidate:X Y WangFull Text:PDF
GTID:1116360092475022Subject:Management Science and Engineering
Abstract/Summary:PDF Full Text Request
Risk management is the key for successful operations of commercial banks, and is based on customer credit analysis. This paper is aimed to discuss how to generally evaluate customer credit in China, where credit is relatively low for individuals and firms.Based on the analysis of types and sources of risks that confront commercial banks, the paper first determines the methodology for its research, i.e. mathematical statistics for quantity factors and the fuzzy discriminating analysis for the quality factors.Combining the evaluation approaches of the banks in Germany and China, the paper gets 13 common indices, and with mathematical statistical method, chooses 4 factors that will influence customer credits: equity capital/total asset, velocity of stock in trade, velocity of total assets and payoff rate of total sale. The four factors with liquidity factor reflect the customer's financial characteristics, such as capital structure, operation, earnings and liquidity.Some quality factors are also introduced in order to get a more comprehensive picture for customer's credit. From the aspects of customer characteristics (including general, financial and preferred features), bank-firm relationships and external characteristics (including industrial features and government-firm relationships), the paper gets 20 quality factors that influence customer credit. Through an experts grading system, banks can get weights for each factor and an initial evaluation for the customer on each factor. A comprehensive credit evaluation can be then got in a multi-level fuzzy evaluation model. In the normal fuzzy evaluation model, the rule of absolute dominance ensures only the majority can prevail, and the concept of "diversification degree" reflects how the initial evaluation is diversified among the experts and makes improvements in the evaluation system.The paper also analyzes the stability of the initial evaluation to the comprehensive evaluation, gets the tolerance interval where the result of the model remains, and ensuresthe reliability of the model's application.In the end, the paper demonstrates the application of the model with two listed companies in China, Younger and Kelong. It makes some calculation based on the initial experts grading, and gets basic status of the two companies by the analysis of weights, relative importance and effect of each factor. The results from the model are not contradictious to those from professional credit rating companies, who usually evaluate firms through financial analysis and on-site investigation.Based on the fuzzy comprehensive analysis, the model in the paper provides a scientific approach to combine the initial experts grading for each factor, and lowers the professional requirements for individual expert. Therefore, customer credit evaluation can be developed more broadly in China, and it will help to solve the bad debt problem in large commercial banks.
Keywords/Search Tags:risk management for commercial banks, credit evaluation for customers, quality and quantity factors, multi-valuate statistical analysis, multi-level fuzzy discriminating analysis, fuzzy evaluation model, rule of absolute dominance
PDF Full Text Request
Related items