Influence Factors And Prediction Of Default Risk On P2P Borrower In China Based On Survival Analysis

As a typical Internet financial model,the stability of P2P online loan service industry is related to the security and stability of the national Internet financial industry.With the rapid development of Internet finance,how to assess the credit risk of individual borrowers in China's P2P online loan service industry properly has become an important issue.In the context of big data,this paper uses the survival analysis method to evaluate the default risk of P2P platform borrowers from a dynamic perspective.The survival analysis method provides a framework for incorporating macroeconomic variables into the model as time-dependent variables.However,these variables are difficult to incorporate into the traditional credit risk assessment method logistic regression model.Firstly,based on the theoretical analysis,this paper establishes the credit risk assessment index system based on the macro and micro levels to select the index that affects the default rate of borrowers,and incorporates several macroeconomic variables closely related to the credit of individual borrowers.Including the broad money supply,the consumer price index,the national housing boom index,etc.Secondly,the paper evaluates the default probability of individual borrowers on the P2P lending platform through COX PH survival analysis method,and verifies that the default rate of P2P platform borrowers will change with macroeconomic fluctuations.And the survival analysis method can model the macroeconomic variables that change with time,which constitutes the basis of stress testing.Therefore,the paper finally uses the discrete-time survival analysis model to stress test the overall default rate of P2P platform borrowers.The paper uses the Monte Carlo method to simulate the extreme macroeconomic situation,and estimates the loss distribution of the borrower's default probability,and uses the risk value and the expected gap as indicators to evaluate the stress test,thus improves the subjectively set situation generation method.The results show that:(1)Compared with logistic regression method of the traditional personal credit risk assessment,the survival analysis method significantly improves the prediction accuracy of the borrower's default rate and the second type of error by including dynamic macroeconomic variables;(2)Macroeconomic variables are important influence factors of personal credit risk of P2P platform.Among them,broad money supply,national housing prosperity index,consumer confidence index and economic sentiment index are relatively important macroeconomic indicators;(3)In one percent extreme economic situation the overall expected default rate of P2P platform borrowers is 5.88 times under nornal circumstances,which is higher than the overall default level of retail borrowers in developed countries.According to the research conclusions,this paper proposes counter?easures to standardize China's P2P online loan service industry fUrther.First,combine the traditional credit evaluation method and survival analysis method to improve the domestic retail credit scoring system further.Secondly,incorporate macroeconomic indicators into the credit risk assessment indicator system of domestic individual borrowers;finally,the stress test is used as a tool to regularly assess and predict the performance of P2P platform borrowers1 defeult risk in extreme situations,which can effectively prevent the default risk of borrowers.