基于信息解释能力的小企业信用评级体系构建研究

以建立商业银行小企业客户信用评级指标体系为目的,提出一种系统的小企业信用评级指标体系构建方法.检验并剔除对企业违约影响不显著的信用评级指标,提出新的指标筛选方法以剔除综合信用风险识别能力弱的评级指标.其次,通过相关分析法降低评级指标集的整体信息重叠.再次,根据信用评级指标体系识别违约状态的准确率检验小企业信用评级指标体系的合理性.最后,基于中国一家商业银行3111笔小企业信贷历史数据进行了实证分析,研究表明本文构建的小企业信用评级指标体系识别违约风险的准确率高而且稳定.

Credit rating indicator system for small businesses based on the ability to interpret information

This paper puts forward a method to construct the credit rating indicator system for small business customers in commercial banks.Firstly,this paper uses F-test to delete insignificant credit rating indicators for default risk identification.Secondly,it proposes a new indicator screening method to delete the indicators with weak comprehensive risk identification ability.Thirdly,the overall information overlap degree of the indicator system is reduced by deleting indicators with weaker risk identification ability within any highly correlated pairs.The indicator system obtained according to this method is evaluated according to its identification accuracy on default risk.Empirical analysis based on 3111 historical small business credit data from a commercial bank is performed to test the system.The result suggests that this system has high accuracy and stable recognition on default risk.