基于K-S检验与距离相关分析的网络借贷信用评价指标体系构建

网络借贷作为一种新型互联网金融模式,提升了金融资源使用效率,缓解了小企业融资难的困局。构建合理的网络借贷信用评价指标体系,从而对网络借贷的潜在风险及时甄别与预防,对互联网金融健康持续发展意义重大。本文根据K-S检验与距离相关分析相结合,筛选对借款客户违约状态甄别能力强的指标,建立了网络借贷信用评价指标体系,通过P2P网络借贷(peer to peer lending,个人对个人借贷)平台LendingClub交易数据进行实证研究,结果表明:不仅借款金额、借款利率等借款标的特征对借贷者违约具有显著相关性,借款者年龄等个人特征、借款者年收入等财务特征以及借款者违约次数等信用特征均对借贷者违约风险产生显著影响。投资者在出借资金时,往往青睐于已婚、年龄适中、具有一定工作经历、历史违约次数较少的借款人。因此,风险监管部门应构建网络借贷违约风险评估模型,对P2P平台进行风险监测,同时建立关键信息共享机制,融合多源数据,明确审查范围,实现P2P网络借贷行业健康有序发展。

Construction of Evaluation Index System of Network Credit Based on Significant Differentiation of Default

As a new and vigorous Internet financial model, Internet lending improves the efficiency of the use of financial resources and alleviates the financing difficulties of small enterprises. However, compared with traditional credit methods, the information asymmetry of online lending is more serious due to the low threshold of loans and the lack of real contact between borrowers and lenders, which leads to frequent platform defaults and increased credit risk. It is of great significance for the healthy and sustainable development of Internet finance to construct a set of reasonable credit evaluation index system and scientifically evaluate its credit risk status, so as to screen and prevent the potential risk of Internet lending in time. Based on K-S test and rank correlation analysis, this paper screens out indicators with strong ability to identify default status of borrowers, establishes an evaluation index system of online lending credit, and conducts an empirical study on the actual transaction data of Lending Club, the largest P2P network lending platform in the world. The results show that the amount of borrowing, the occupation of borrowers and the unemployment rate in the evaluation index system of this study. The 12 indicators have a significant impact on distinguishing default status. One of the characteristics and innovations of this paper is that the larger the statistical value of K-S test is, the larger the deviation between the distribution function of corresponding default samples and the distribution function of non-default samples is. It shows that the stronger the ability of the evaluation index to discriminate the default status of borrowers, the better the evaluation index can be selected to distinguish the default status from the non-default status. Foot. The second is to use K-S test and rank correlation analysis, which have no requirement on the overall distribution of evaluation indicators and are suitable for the non-parametric statistical method with unknown specific distribution to screen the evaluation indicators of network lending credit, to overcome the drawbacks of existing index screening methods that require the evaluation indicators to obey the strict requirement of normal distribution, and to expand the scope of application of index screening methods.