基于PROMETHEE-II的商户小额贷款信用评级模型及实证

发布时间：2018-02-10 04:05

本文关键词： 信用评级信用风险 PROMETHEE-II 商户小额贷款　出处：《运筹与管理》2017年09期 　论文类型：期刊论文

【摘要】：截至2014年底,中国注册个体工商户为4984.06万户,个体私营经济吸纳社会从业人员已达2.5亿人,加上中国商户小额贷款对象的分散性、财务信息不健全等特点和难点,商户小额贷款信用评级体系极不完善,甚至绝大多数银行都没有建立这个体系。本文通过相关分析剔除反映信息重复的指标,通过显著性判别遴选对商户违约状态影响显著的指标,建立了能显著区分商户违约状态的小额贷款信用评级指标体系。在此基础上,结合PROMETHEE-II(偏好顺序结构)和聚类分析方法,构建了商户小额贷款信用评级模型,并对中国某国有商业银行2157个商户小额贷款样本进行了实证。本文创新与特色:一是通过将偏好顺序结构评估法(PROMETHEE-II)引入商户小额贷款信用评级,构建了基于PROMETHEE-II的小额贷款信用评分模型,求解商户的净流量信用得分Φ(a),揭示了商户a与其余商户、评价指标间的相互作用对评价结果的影响,避免了现有研究由于评价指标之间的相互替代性、严重影响评价结果可靠性的不足。二是借鉴模糊聚类"数据越集中、越应该被分为一类"的思想,采用R聚类对商户信用得分进行分类;进而采用K-W检验,对分类数目l进行非参数检验,确定商户的信用等级。既保证了不同等级商户在信用得分数值上存在显著差异,也确保了不同等级商户能反映不同的信用特征;同时,也避免了现有利用信用得分区间、违约概率阈值或客户数分布方法划分信用等级时,得分区间、违约概率阈值或客户数分布分位点人为主观确定的不足。三是实证研究表明,影响商户小额贷款信用风险的重要性排序依次为:X_3偿债能力X_1基本情况X_6宏观环境X_5营运能力X_2保证联保X_4盈利能力。
[Abstract]:As of end of 2014, there were forty-nine million eight hundred and forty thousand and six hundred registered individual industrial and commercial households in China, and 250 million people were employed in the private and individual sectors of the economy. In addition, there are many characteristics and difficulties, such as the dispersal of the small loan objects of Chinese merchants, the imperfection of financial information, and so on. The credit rating system of small loans for merchants is extremely imperfect, and even most banks have not established such a system. A credit rating index system for small loans is established by using significant discriminant selection indexes which have a significant impact on the default status of merchants. On this basis, the index system is combined with PROMETHEE-II (preference sequence structure) and clustering analysis. The credit rating model of small loans is constructed. An empirical study has been made on 2157 samples of small loans by a state-owned commercial bank in China. This paper has some innovations and features: firstly, the credit rating of small loans of commercial households is introduced by introducing the method of preference sequence structure evaluation (PROMETHEE-II). In this paper, a small loan credit scoring model based on PROMETHEE-II is constructed, and the net flow credit score of merchants is solved, which reveals the influence of the interaction between merchant a and other merchants on the evaluation results. It avoids the deficiency of the reliability of evaluation results which is seriously affected by the substitution of evaluation indexes in existing research. Second, the more centralized the data is, the more should it be divided into one category by using fuzzy clustering as a reference. Using R clustering to classify the credit score of merchant, then using K-W test, carrying on nonparametric test to the number of classification l, determining the credit grade of merchant, which ensures that there are significant differences in the value of credit score between different grades of merchants. It also ensures that different grades of merchants can reflect different credit characteristics. At the same time, it also avoids the existing use of credit score interval, when the default probability threshold or customer number distribution method is used to divide credit rating, Default probability threshold or customer number distribution sub-locus is not determined subjectively. Third, empirical research shows that, The order of importance of influencing the credit risk of small loans is: XX3 debt repayment ability X1 basic situation X6 macro environment X5 operation ability X2 guarantee the profitability of UNPROFOR X4.
【作者单位】：西北农林科技大学经济管理学院;东北财经大学管理科学与工程学院;大连理工大学管理与经济学部;
【基金】：国家自然科学基金重点项目(71731003);国家自然科学基金青年项目(71503199,71502026);国家自然科学基金面上项目(71471027,71373207,71672019) 中国博士后科学基金资助项目(2015M572608,2016T90957) 中央高校基本科研业务费人文社科专项(2015RWYB09) 西北农林科技大学“青年英才培育计划”(Z109021717) 陕西省博士后科学基金资助项目(K3380216032) 中国邮政储蓄银行总行小额贷款信用风险评价与贷款定价资助项目(2009V07)
【分类号】：F275;F832.4

【相似文献】