P2P平台贷款决策研究——利用收益评分取代违约概率评分  

Study on P2P Platform Loan Decision-Replace Default Probability Score with Earnings Score

作　　者：陈卓 蒋炜[1] CHEN Zhuo;YlANG Wei(Anti College of Economics and Management,Shanghai Jiao Tong University,Shanghai 200030,China)

机构地区：[1]上海交通大学安泰经济与管理学院,上海200030

出　　处：《上海管理科学》2018年第5期16-19,共4页Shanghai Management Science

摘　　要：传统的信用评分系统估计贷款违约概率,并将违约概率作为是否放贷的标准。与通常试图确定违约概率的研究不同,课题专注于预测投资P2P贷款的预期收益率,并尝试以此作为贷款人选择的标准。总体而言,课题检查了213 731条P2P贷款记录,并利用CHAID模型分析确定了影响贷款盈利能力的因素,实现了对贷款盈利能力的预测。在分析的样本中发现,通过应用CHAID-IRR评分系统筛选借款人,投资人所获得的收益优于通过使用传统信用评分获得的结果。The traditional credit scoring system estimates the loan default probability and considers the default probability as the main reference for lending.Unlike previous studies that usually try to determine the probability of default,this article focuses on predicting the expected rate of return on investing in P2 P loans and attempts to use this as the lender＇s chosen criterion.Overall,this study examined 213,731 P2 P loan records and used the CHAID model to analyze and determine the factors that affect the profitability of the loan and achieved a forecast of the profitability of the loan.The result of CHAID model shows that by using the CHAID-IRR scoring system to screen borrowers,the income obtained by the investor is better than that obtained by using the traditional credit score.

关 键 词：P2P 决策树 贷款评估 IRR 

分 类 号：F830[经济管理—金融学]