基于平滑扩充原理的商业银行信用风险评级模型及实证
|
摘要
本文通过银行的资产质量方面、资本充足率方面、管控效能层面、盈利状态层面、流动性层面与社会敏感度层面等构建商业银行信用风险评价体系。根据平滑扩充原理模拟生成大样本数据,对评级得分进行扩充,进而根据扩充后的大样本数据划分银行的信用风险等级。解决了由于样本少、无法对信用等级合理划分的难题。通过实证分析可以了解到,本文得出的银行评级信息和标准普尔提供的评价结论存在共同的序关系状态。因此,可根据本模型对大多数未经过国际权威机构评级的银行进行风险评级。
The weights of the index for the six aspects,based on the commercial bank credit risk evaluation index system,including capital adequacy,asset quality,management,earnings,liquidity and social sensitivity,can be determined by the use of coefficient of variation. Based on the principle of smooth expansion,large sample data could be simulated and the rating score could be expanded. According to the credit risk rating of large sample data,the credit risk rating of banks could be divided. The innovations and features of this paper are as follows: First,we simulate the large sample data and expand the rating score according to the principle of smooth expansion,and then according to the credit risk rating we expand the original sample after 20 times to divide the bank's credit risk rating. This solves the problem: the small sample can not be credit grading. Second,by the use of the nonparametric Mann-Whitney U method,the conclusion is obtained that large sample data with the original rating score data distribution is consistent at the 95% confidence level. Third,the empirical study shows that the sequence relation of bank rating is in accordance with Standard & Poor's rating result. Therefore,this model can be used on credit rating by most of banks which have not been rated the international rating authorities.
The weights of the index for the six aspects,based on the commercial bank credit risk evaluation index system,including capital adequacy,asset quality,management,earnings,liquidity and social sensitivity,can be determined by the use of coefficient of variation. Based on the principle of smooth expansion,large sample data could be simulated and the rating score could be expanded. According to the credit risk rating of large sample data,the credit risk rating of banks could be divided. The innovations and features of this paper are as follows: First,we simulate the large sample data and expand the rating score according to the principle of smooth expansion,and then according to the credit risk rating we expand the original sample after 20 times to divide the bank's credit risk rating. This solves the problem: the small sample can not be credit grading. Second,by the use of the nonparametric Mann-Whitney U method,the conclusion is obtained that large sample data with the original rating score data distribution is consistent at the 95% confidence level. Third,the empirical study shows that the sequence relation of bank rating is in accordance with Standard & Poor's rating result. Therefore,this model can be used on credit rating by most of banks which have not been rated the international rating authorities.
引文
[1] Moody’s Investors Service. Global credit research[R].New York:Moody’s Investors Service,2005. 6.
[2] Standard&Poor’s. Standard&poor’s china top 50banks[R]. New York:Standard&Poor’s,2007. 6.
[3]大公国际资信评估有限公司.大公商业银行信用评级方法框架[R].北京:大公国际资信评估有限公司,2005. 6.
[4]中诚信国际.商业银行信用评价方法[R].北京:中诚信国际,2004. 3.
[5]柯孔林,薛锋.基于扩展数据包络判别法的商业银行信用风险评估[J].系统工程理论与实践,2004,(4):117-122.
[6]薛锋,柯孔林.基于混合整数规划法的企业信用风险评估研究[J].中国管理科学,2006,14(2):39-44.
[7]王建新,于立勇.基于信用风险度的商业银行风险评估模型研究[J].管理工程学报,2007,21(4):85-90.
[8] Tsai C,Wu J. Using neural network ensembles for bankruptcy prediction and credit scoring[J]. Expert Systems with Applications,2008,34(4):2639-2649.
[9] Lin S,Shiue Y,Chen S,Cheng H. Applying enhanced data mining approaches in predicting bank performance:a case of taiwanese commercial banks[J]. Expert Systems with Applications,2009,36(9):11543-11551.
[10]徐春红,路正南.商业银行信用风险识别的模型构建与政策建议[J].统计与决策,2010,(2):128-130.
[11]程砚秋.基于违约判别度的小企业信用风险评价研究[J].科研管理,2015,36(S1):510-517.
[12]陈晓红,杨志慧.基于改进模糊综合评价法的信用评估体系研究——以我国中小上市公司为样本的实证研究[J].中国管理科学,2015,23(1):146-153.
[13]张奇,胡蓝艺,王珏.基于Logit与SVM的银行业信用风险预警模型研究[J].系统工程理论与实践,2015,35(7):1784-1790.
[14]衣柏衡,朱建军,李杰.基于改进SMOTE的小额贷款公司客户信用非均衡SVM分类[J].中国管理科学,2016,24(3):24-30.
[15]张大斌,周志刚,刘雯,焦鹏.上市公司信用风险测度的不确定性DE-KMV模型[J].系统工程学报,2015,30(2):165-173.
[16]姚德权,张宏亮,黄学军.基于变结构KMV模型的商业银行风险承担度量研究[J].中国软科学,2015,(11):109-122.
[17]迟国泰,潘明道,齐菲.一个基于小样本的银行信用风险评级模型的设计及应用[J].数量经济技术经济研究,2014,(6):102-116.
[18] Haiyan Bai. A new resampling method to improve quality of research with small samples[D]. University of Cincinnati,2006.
[19]迟国泰,隋聪.基于非线性映射的银行信用风险评价指标体系构建[R].大连:大连理工大学金融风险与系统评价管理研究中心,2010. 6.
[20]郭亚军.综合评价理论、方法及应用[M].北京:科学出版社,2002:44-51.
[21]中国银行业监督管理委员会.股份制商业银行风险评级体系(暂行)[EB/OL]. 2004[2013-6-15].http://baike. baidu. com/view/2486816. htm.
[22]迟国泰,王卫.基于科学发展的综合评价理论、方法与应用[M].北京:科学出版社,2009,9.
[23]盛骤,谢式千,潘承毅.概率论与数理统计[M].第四版.北京:高等教育出版社,2008-6-1:132.
[24]中国银行,中国工商银行,中国建设银行,等.中国银行,中国工商银行,中国建设银行等46家商业银行年报[R].中国银行、中国工商银行、中国建设银行等,2012.
[2] Standard&Poor’s. Standard&poor’s china top 50banks[R]. New York:Standard&Poor’s,2007. 6.
[3]大公国际资信评估有限公司.大公商业银行信用评级方法框架[R].北京:大公国际资信评估有限公司,2005. 6.
[4]中诚信国际.商业银行信用评价方法[R].北京:中诚信国际,2004. 3.
[5]柯孔林,薛锋.基于扩展数据包络判别法的商业银行信用风险评估[J].系统工程理论与实践,2004,(4):117-122.
[6]薛锋,柯孔林.基于混合整数规划法的企业信用风险评估研究[J].中国管理科学,2006,14(2):39-44.
[7]王建新,于立勇.基于信用风险度的商业银行风险评估模型研究[J].管理工程学报,2007,21(4):85-90.
[8] Tsai C,Wu J. Using neural network ensembles for bankruptcy prediction and credit scoring[J]. Expert Systems with Applications,2008,34(4):2639-2649.
[9] Lin S,Shiue Y,Chen S,Cheng H. Applying enhanced data mining approaches in predicting bank performance:a case of taiwanese commercial banks[J]. Expert Systems with Applications,2009,36(9):11543-11551.
[10]徐春红,路正南.商业银行信用风险识别的模型构建与政策建议[J].统计与决策,2010,(2):128-130.
[11]程砚秋.基于违约判别度的小企业信用风险评价研究[J].科研管理,2015,36(S1):510-517.
[12]陈晓红,杨志慧.基于改进模糊综合评价法的信用评估体系研究——以我国中小上市公司为样本的实证研究[J].中国管理科学,2015,23(1):146-153.
[13]张奇,胡蓝艺,王珏.基于Logit与SVM的银行业信用风险预警模型研究[J].系统工程理论与实践,2015,35(7):1784-1790.
[14]衣柏衡,朱建军,李杰.基于改进SMOTE的小额贷款公司客户信用非均衡SVM分类[J].中国管理科学,2016,24(3):24-30.
[15]张大斌,周志刚,刘雯,焦鹏.上市公司信用风险测度的不确定性DE-KMV模型[J].系统工程学报,2015,30(2):165-173.
[16]姚德权,张宏亮,黄学军.基于变结构KMV模型的商业银行风险承担度量研究[J].中国软科学,2015,(11):109-122.
[17]迟国泰,潘明道,齐菲.一个基于小样本的银行信用风险评级模型的设计及应用[J].数量经济技术经济研究,2014,(6):102-116.
[18] Haiyan Bai. A new resampling method to improve quality of research with small samples[D]. University of Cincinnati,2006.
[19]迟国泰,隋聪.基于非线性映射的银行信用风险评价指标体系构建[R].大连:大连理工大学金融风险与系统评价管理研究中心,2010. 6.
[20]郭亚军.综合评价理论、方法及应用[M].北京:科学出版社,2002:44-51.
[21]中国银行业监督管理委员会.股份制商业银行风险评级体系(暂行)[EB/OL]. 2004[2013-6-15].http://baike. baidu. com/view/2486816. htm.
[22]迟国泰,王卫.基于科学发展的综合评价理论、方法与应用[M].北京:科学出版社,2009,9.
[23]盛骤,谢式千,潘承毅.概率论与数理统计[M].第四版.北京:高等教育出版社,2008-6-1:132.
[24]中国银行,中国工商银行,中国建设银行,等.中国银行,中国工商银行,中国建设银行等46家商业银行年报[R].中国银行、中国工商银行、中国建设银行等,2012.