生存分析、宏观经济变量与违约概率  

Survival Analysis, Macro-economic Variables and Default Probability

作　　者：莫易娴[1] 

机构地区：[1]华南农业大学,广东广州510642

出　　处：《上海立信会计金融学院学报》2018年第1期5-19,共15页Journal of Shanghai Lixin University of Accounting and Finance

基　　金：广东省哲学社会科学"十二五"规划2014年学科共建项目(GD14XYJ17):P2P网络借贷;生存分析与宏观经济风险;广东省哲学社会科学课题:信贷资金投向农村的金融战略;2013;国家留学基金委出国访学(201508440475);华南农业大学研究生2016重点课题:网商银行PK传统商业银行案例分析16YJAL-1

摘　　要：在研究个人贷款违约风险中,传统的研究往往只单纯地将宏观经济指标作为协变量,并未考虑宏观经济指标的时变交互特征(vandell,1993;Zandi,1998)。国外研究达成共识的是以Logistic回归的传统模型不能给出违约概率的动态预测值,且反映经济形势的宏观经济变量也不能纳入模型中。论文论述了生存分析与Logistic模型的理论机理,试图在借贷违约风险中加入系统性风险对违约的影响因素,克服了以Logistic回归模型为代表的传统模型在度量信贷违约概率时仅考虑个体非系统性风险的局限。研究结论说明宏观经济变量确实对违约风险有影响,对网贷违约风险来说,Cox模型更优于Logistic函数。In the research of personal loan default risk, the traditional ways can＇t consider time-varying interaction characteristics of macroeconomie indicators （vandell, 1993; Zandi, 1998）.The tradition of the Logistic regression model can＇t give the predicted value of the probability of default dynamic, and reflect the economic situation of the macro economic variables. Thesis discusses the theory of survival analysis and Logistic model mechanism, tries to join the factors that affect thesystematic risk of default, and overcomes the limitations of individual of systemic risk that the traditional model represented by Logistic regression model brings in measuring credit default probability. The result shows that macroeconomic does have an impact on the risk of default and Cox model is more better than Logistic model.

关 键 词：生存分析 LOGISTIC模型 违约概率 

分 类 号：F830.5[经济管理—金融学]