基于似然比检验的工业小企业债信评级研究

发布时间：2018-01-18 02:31

本文关键词：基于似然比检验的工业小企业债信评级研究　出处：《中国管理科学》2017年01期 　论文类型：期刊论文

【摘要】：债信评级是评价一笔债务偿还的可能性或违约损失率。由于工业小企业贷款存在风险高、额度小、财务数据不真实等特点,使商业银行无法准确对工业小企业贷款的信用风险进行科学评估。因此,构建一套合理的债信评级体系成为亟待解决的问题。本文一是构造某一个指标与违约状态之间的逻辑回归方程,通过对比仅含常数项的零模型的对数似然值与仅含有某一个指标的完整模型的对数似然值,构造χ~2统计量,若有、无某指标时的两个对数似然值偏差越大,则该指标对区分违约与非违约状态的贡献越大,该指标越易保留的思路对指标进行遴选,保证遴选出的指标都对违约状态具有显著的区分能力,弥补现有研究不以能否区分违约状态对指标进行筛选的不足。二是通过计算同一准则层内任意两个指标的相关系数,确定这两个指标反映信息的重复程度,在相关系数大于某一阈值的两个指标中,删除χ2统计量小、即对违约状态区分程度小的指标,既避免指标体系的信息冗余、又避免误删对违约状态判别能力强指标。改变现有研究在相关系数大的指标中人为主观删除一个的弊端。三是通过提取中国某区域性商业银行分布在全国28个城市分支行的贷款数据进行实证,建立了由资产负债率、成本利润率、近三年企业授信情况等26个指标构成的适用于工业小企业信用风险评价的指标体系。四是通过对工业小企业进行债信评级,不仅得到每个小企业的信用等级,还得到每个贷款小企业对应的违约损失率,改变现有的信用评级研究仅仅计算贷款客户的信用得分和进行评分排序的不足。
[Abstract]:Credit rating is the assessment of the possibility of a debt repayment or default loss rate. Because of small industrial enterprises in the loan risk is high, the small amount of false financial data, etc., so that commercial banks can not be accurately for small industrial enterprises in the credit risk evaluation. Therefore, constructing a reasonable credit rating system to become the problem to be solved. This paper is a structure of a between the index and the default logic regression equation, the log likelihood of the log likelihood model compared with only zero constant values of the complete model and contains only a single index value of the structure of X ~2 statistics, if there is no two, a log likelihood the index value of the deviation is greater, the index to distinguish between default and non default state contribution to the greater the index is easy to retain ideas for index selection, ensure the selection of the indicators are the default state Distinguish ability is significant, make up the existing research on default could distinguish screening for less than two. The correlation coefficient index is calculated with a standard layer of any of the two indicators, determine the degree of repetition of these two indicators to reflect the information of the correlation coefficient is greater than two indicators in a certain threshold, delete x 2 Statistics small, to distinguish the degree of default of small targets, both to avoid the redundant information of the index system, and to avoid default discriminant ability deleted index. To change the existing research on index related coefficient of the subjective shortcomings. Three a delete is empirically by extracting a regional business loan data China the bank located in the country's 28 City branch, established by the asset liability ratio, cost profit rate, which is nearly three years of business credit 26 indicators such as the credit risk in small industrial enterprises The evaluation index system. Four is the credit rating of small industrial enterprises, not only by every small business credit rating, but also get each corresponding to the small business loan default loss rate change, lack of research on credit rating of existing only calculate the loan customer credit scoring and ranking for a score.

【作者单位】：大连理工大学管理与经济学部;
【基金】：国家自然科学基金资助项目(71171031，，71471027) 国家社科基金资助项目(16BTJ017) 辽宁省社科规划基金资助项目(L16BJY016) 教育部科学技术研究项目(2011-10) 大连银行小企业信用风险评级系统与贷款定价项目(2012-01) 辽宁经济社会发展重点课题(2015lslktzdian-05)
【分类号】：F832.4;F425
【正文快照】： 1引言债信评级不仅是对某一笔贷款进行评级,同时确定该笔贷款对应的违约损失率。如何控制众多工业小企业的信用风险是商业银行重点关注的问题之一。小企业在国民经济中扮演重要的角色。目前中国中小企业总数已经达到全国企业数量的99.9%,创造的产品及服务价值相当于GDP的6

【相似文献】