基于泰尔指数修正的ELECTRE III小企业信用评价模型

doi:10.16381/j.cnki.issn1003-207x.2019.10.003

中国管理科学 ›› 2019, Vol. 27 ›› Issue (10): 22-33.doi: 10.16381/j.cnki.issn1003-207x.2019.10.003

基于泰尔指数修正的ELECTRE III小企业信用评价模型

程砚秋^1,2, 徐占东³

1. 东北财经大学会计学院, 辽宁大连 116025;
2. 中国内部控制研究中心, 辽宁大连 116025;
3. 东北财经大学经济学院, 辽宁大连 116025

收稿日期:2018-01-23 修回日期:2018-05-17 出版日期:2019-10-20 发布日期:2019-10-25
通讯作者: 程砚秋(1981-),女(汉族),山西灵石人,东北财经大学会计学院,副教授,中国内部控制研究中心研究员,博士,研究方向:信用风险管理、复杂系统评价,E-mail:chengyanqiu_happy@163.com. E-mail:chengyanqiu_happy@163.com
基金资助:
教育部人文社会科学研究青年基金资助项目（18YJC790017）；国家自然科学基金资助项目（71201018，71571034，71731003，71873103，71971051）

Credit Risk Evaluation of Small Enterprises Based on Revised ELECTRE III by Theil Index

CHENG Yan-qiu^1,2, XU Zhan-dong³

1. School of Accounting, Dongbei University of Finance and Economics, Dalian 116025, China;
2. China Internal Control Research Center, Dalian 116025, China;
3. School of Economics, Dongbei University of Finance and Economics, Dalian 116025, China

Received:2018-01-23 Revised:2018-05-17 Online:2019-10-20 Published:2019-10-25

摘要/Abstract

摘要： 针对现有评价方法存在的贷款客户在某些评价指标下的低分可以完全由其他评价指标下高分来补偿的问题，本文对基于ELECTRE III的小企业信用风险评价模型进行了研究。首先，借鉴ELECTRE III的评价原理，根据新增贷款客户优于所有历史贷款客户的净可信程度，计算新增贷款客户的信用风险评价得分。不仅解决了新增贷款客户信用风险评价的问题，而且保证评价模型具有从历史数据学习的能力。其次，借鉴泰尔指数不仅可以反映收入总体差异、而且可以将总体差异分解为组内差异和组间差异的特征，对小企业信用风险的各评价指标进行了赋权。体现了"越能区分客户违约状况、指标权重越大"的赋权思想。最后，以违约、非违约样本加权后的组内差异程度为基础，确定ELECTRE III的偏好阈值；以非违约本组内差异程度为基础，确定ELECTRE III的无差别阈值；以违约样本组内差异程度为基础，确定ELECTRE III的否决阈值；不仅反映了不同评价指标数据差异大小对评价结果的影响，而且避免了现有阈值人为主观确定的不足。中国某地方性大型商业银行的研究结果发现：企业近三年授信情况、法人代表贷款违约记录、企业通过本行回笼货款总额占比等评价指标的权重较大，能有效地区分违约客户和非违约客户。而从不同指标的总体差异程度来讲，现金比率、净资产与年末贷款余额比率、企业成立年限等评价指标的差异程度较大。

关键词: 信用风险评价, 小企业贷款, ELECTRE III, 泰尔指数

Abstract: One of the key tasks of credit risk evaluation is to determine the reasonable evaluation model. However, many existing credit risk evaluation models are complete compensatory methods. Those complete compensatory methods have a problem that high scores under certain evaluation indicators can fully compensate low scores under other evaluation indicators. Furthermore, in the credit loans, commercial bank will not grant loans to customer who's the specific indicator is too low and other indicators perform well. To fill in the above gap, a credit risk evaluation model for small enterprises is constructed based on ELECTRE III. This model can avoid the complete compensatory problem.
First of all, using the net credit score flow of the new loan customer to decide whether this new loan customer loans or not. This method can evaluate the credit risk of the new loan customer and make ELECTRE III has the ability of learning from history loan customers. Secondly, the Theil index can not only reflect income differences, but also break the differences into difference within groups and difference between groups. Then, by using the Theil index, the index weights are determined based on the idea that "the greater the influence on default is, the bigger the weight will be". Finally, the preference thresholds of ELECTRE III are determined based on Theil index within groups. And the indifference thresholds of ELECTRE III are determined based on Theil index of non-default samples;the veto thresholds of ELECTRE III are determined based on Theil index of default samples. The influences of different evaluation index on evaluation result can reflect through those thresholds.Also the subjective determination of three thresholds could be eliminated.
The constructed model has been verified using the samples of a Chinese national commercial bank. The empirical results show that the key inductors of affecting the credit risk of small enterprises are enterprise's credit extension of the past three years, loan default record of legal representative, the total amount of payment through our bank, and so on. Also, the indicators which have big differences between default and non default customers are found, such as cash ratio, the ratio of net assets to loans in the end of year, the age of the enterprise, and so on.
In conclusion, this research aims at establishing a credit risk evaluation model for small enterprises based on ELECTRE III and extending a new vision on small enterprises credit risk evaluation theory and models.

Key words: credit risk evaluation, small enterprises' loans, ELECTRE III, Theil index

中图分类号:

程砚秋, 徐占东. 基于泰尔指数修正的ELECTRE III小企业信用评价模型[J]. 中国管理科学, 2019, 27(10): 22-33.

CHENG Yan-qiu, XU Zhan-dong. Credit Risk Evaluation of Small Enterprises Based on Revised ELECTRE III by Theil Index[J]. Chinese Journal of Management Science, 2019, 27(10): 22-33.

参考文献

[1] 辛飞, 孙永广, 邓晶.日本中小企业信用风险度量及启示[J]. 金融理论与实践, 2005, (7):80-81.
[2] 中国工商银行.关于印发《中国工商银行小企业法人客户信用等级评定办法》的通知[R].中国工商银行, 2005.
[3] 中国建设银行.中国建设银行小企业客户评价办法[R].中国建设银行,2007.
[4] 赵志冲,迟国泰.基于似然比检验的工业小企业债信评级研究[J].中国管理科学, 2017, 25(1):45-56.
[5] 迟国泰,张亚京,石宝峰.基于Probit回归的小企业债信评级模型及实证[J].管理科学学报,2016,19(6):136-156.
[6] Chi Guotai, Zhang Zhipeng. Multi criteria credit rating model for small enterprise using a nonparametric method[J]. Sustainability, 2017, 9(10):1-22.
[7] 徐晓萍,马文杰.非上市中小企业贷款违约率的定量分析-基于判别分析法和决策树模型的分析[J].金融研究, 2011, (3):111-120.
[8] Shi Baofeng,Wang Jing,Qi Junyan,et al. A novel imbalanced data classification approach based on logistic regression and fisher discriminant[J]. Mathematical Problems in Engineering, 2015, (2015):1-12.
[9] Blanco A,Pino R, Lara J, et al. Credit scoring models for the microfinance industry using neural networks:Evidence from Peru[J]. Expert Systems with Applications,2013,40(1):356-364.
[10] 肖斌卿,杨旸,李心丹,等.基于模糊神经网络的小微企业信用评级研究[J].管理科学学报,2016,19(11):114-126.
[11] 衣柏衡,朱建军,李杰.基于改进SMOTE的小额贷款公司客户信用风险非均衡SVM分类[J].中国管理科学, 2016, 24(3):24-30.
[12] 石宝峰,刘峰,王建军,等.基于PROMETHEE-II的商户小额贷款信用评级模型及实证[J].运筹与管理,2017, 26(9):137-147.
[13] Chavira D A G, Lopez J C L, Noriega J J S, et al. A credit ranking model for a parafinancial company based on the ELECTRE-III method and a multiobjective evolutionary algorithm[J]. Applied Soft Computing, 2017,(60):190-201.
[14] Doumpos M, Figueira J R. A multicriteria outranking approach for modeling corporate credit ratings:An application of the electre tri-nC method[J]. Omega,2019, (82):166-180.
[15] 孙世岩, 邱志明, 王航宇. 级别不低于方法及其发展评述[J].管理工程学报,2008,22(3):41-56.
[16] Govindan K, Jepsen M B. ELECTRE:A comprehensive literature review on methodologies and applications[J].European Journal of Operational Research, 2016,250(1):1-29.
[17] Behzadian M, Kazemzadeh R B, Albadvi A, et al. PROMETHEE:A comprehensive literature review on methodologies and applications[J]. European Journal of Operational Research, 2010,200(1):198-215.
[18] Roy B. ELECTRE III:Un algorithme de classements fonde sur une representation floue des preferences en presence de criteres multiples[J]. Cahiers Centre Etudes Recherche Operationnelle, 1978, 20(1):3-24.
[19] Rogers M, Bruen M. Using ELECTRE III to choose route for Dublin port motorway[J]. Journal of Transportation Engineering, 2000, 126(4):313-323.
[20] 王建军, 杨德礼.ELECTRE III的一种排序新方法[J]. 系统工程, 2005, 23(12):95-98.
[21] Corazza M, Funari S, Gusso R. Creditworthiness evaluation of Italian SMEs at the beginning of the 2007-2008 crisis:An MCDA approach[J]. The North American Journal of Economics and Finance, 2016,38(c):1-26.
[22] Corazza M, Funari S, Siviero F. A MURAME-based technology for bank decision support in creditworthiness assessment[J]. Banks and Bank Systems, 2014,9(3):7-17.
[23] Theil H. Economics and information theory[M]. Chicago:Rand Mcnally and Company, 1967.
[24] Conceicao P, Galbraith J K, Bradford P.The Theil index in sequences of nested and hierarchic grouping structures[J]. Eastern Economic Journal, 2001, 27(4):491-514.
[25] 许治,陈志荣,邓芹凌. 国家级创新型城市技术成就指数俱乐部收敛效应[J].科学学研究,2013,31(5):790-800.
[26] 李刚,李建平,孙晓蕾, 等. 兼顾序信息和强度信息的主客观组合赋权法研究[J].中国管理科学,2017, 25(12):179-187.
[27] 程砚秋.基于不均衡数据的小企业信用风险评价[J].运筹与管理,2016, 25(16):181-189.
[28] 孙世岩, 王航宇, 石章松. ELECTRE-III法权重的敏感性分析方法[J].系统工程与电子技术,2009,31(11):2645-2647.

基于泰尔指数修正的ELECTRE III小企业信用评价模型

Credit Risk Evaluation of Small Enterprises Based on Revised ELECTRE III by Theil Index

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 1

Metrics

本文评价

推荐阅读 10