基于Brown-Mood中位数检验的小企业债信评级体系  

A Small Enterprise Facility Rating System Based on Brown-Mood Test

作　　者：周颖[1] 沈隆 ZHOU Ying;SHEN Long(School of Economics and Management,Dalian University of Technology,Dalian 116024,Liaoning,China)

机构地区：[1]大连理工大学经济管理学院,辽宁大连116024

出　　处：《系统管理学报》2020年第6期1043-1055,共13页Journal of Systems & Management

基　　金：国家自然科学基金重点项目(71731003);国家自然科学基金面上项目(71873103,71971051,71971034,72071026);国家自然科学基金青年科学基金资助项目(71901055,71903019);国家社会科学基金重大项目(18ZDA095);大连银行小企业信用风险评级系统与贷款定价项目(2012-01)。

摘　　要：债信评级是通过挖掘评级数据与违约风险的关系,评价一笔债务回收的可能性及不同等级违约损失率的大小。基于Brown-Mood中位数检验建立小企业债信评级体系。创新与特色:①根据某个指标的非违约样本中位数显著大于违约样本的中位数,则该指标越能将非违约与违约样本区分开的思路构造Brown-Mood中位数检验统计量值,遴选通过检验、即违约鉴别力显著的指标。改变了绝大多数现有研究未以违约鉴别力显著为指标遴选标准的不足。②在相关系数大的一对指标中,删除Brown-Mood中位数检验统计量值小的指标,在避免了指标反映信息冗余的同时,也避免了指标的误删。③根据指标的违约鉴别力越强、指标权重就应越大的思路对指标赋权,确保了违约鉴别力越大的指标,权重也越大。④根据非违约样本的债信得分中位数显著大于违约样本的债信得分中位数,则债信得分越能显著区分违约状态、债信评分体系越合理的思路,检验债信评分体系的合理性,开拓了债信评分体系检验的新思路。实证表明:包括资产负债率、恩格尔系数等24个指标的小企业债信评级体系具有显著的违约鉴别力,在债信评级时,应重点关注恩格尔系数、行业景气指数等企业外部宏观环境因素以及法人代表汽车及不动产价值、担任该职务时间等法人代表基本情况因素。Debt credit rating is to evaluate the possibility of a debt recovery and the size of different levels of default loss rate by mining the relationship between rating data and default risk.This paper constructs a credit rating system for small enterprises by using Brown-Mood test.First,according to the idea that the median of non-default samples of a certain indicator is significantly larger than that of default samples,the Brown-Mood median test statistics was constructed to select the indicators.The indicators passed testing was retained,which have a great default discriminant ability.It changes the shorting which did not take the default discriminant ability as the criteria to select indicators in majority existing research.Second,in the pair of indicators with large correlation coefficients,the indicator with a small value of the Brown-Mood median test statistic was deleted,which avoided the indicator from reflecting information redundancy as well as the erroneous deletion of the indicator.Third,according to the idea that if the default discrimination power of this indicator is stronger,the indicator weight should be greater,weight was given to the indicator.Fourth,if the median of the credit score for the non-default sample is significantly greater than that for the default sample,the credit score can distinguish the default status,which indicates that the credit rating system is more reasonable.The rationality of the system has opened up a new way for testing the credit rating system.The empirical evidence shows that the debt rating system including 24 indicators,such as“asset-liability ratio”and“Engel coefficient”,has a significant default discriminatory power.For debt rating of small enterprises,attention should be paid to the external macro-environmental factors including“Engel coefficient”and“industry prosperity index”and the legal representative’s basic situation factors including“the value of automobiles and real estate”and“time in this position”.

关 键 词：小企业 信用评级 债信评级 违约鉴别力 Brown-Mood中位数检验 

分 类 号：F830.56[经济管理—金融学] N945.16[自然科学总论—系统科学]