Research On P2P Risk Based On Borrowers' Credit Score And Problematic Platforms' Identification

In recent years,internet technologies represented by big data,artificial intelligence,etc.have developed rapidly.Various emerging industries such as P2 P and third-party payment have facilitated people's lives,but they have also brought many potential dangers.In 2007,China's earliest P2 P network lending platform “PPDAI” was established.In these years,the online lending platforms have entered an outbreak period.According to the third-party organization “p2peye”,by the end of December 2018,there were 6,590 P2 P lending platforms in China.While the number of platforms and transaction volumes skyrocketed,the P2 P platforms also faced many challenges.Platform fraud and running events occurred frequently,and the cumulative problematic platforms reached 4,970,which seriously affected the healthy development of the internet financial industry.This paper empirically explores the borrowers' credit score system and the identification elements of the problematic platforms from the perspectives of the main sources of P2 P risk--borrowers and platforms.From the perspective of the borrowers,using the data of the borrower's transaction data on a P2 P platform,it was found that the historical credit rating,age,availability of housing,loan type,credit card quota,salary level,and marital status have significant impacts on the borrowers' performances.Among them,borrower with a higher level of historical credit rating has a lower probability of default;borrower with an elder age has a lower probability of default;borrower with a house has a lower probability of default than borrower without a house;the types of borrowing were divided into 5 groups:(1)short-term turnover;(2)personal consumption;(3)car loan;purchase loan;medical expenses;decoration loan;(4)wedding preparation;education and training;investment and entrepreneurship;(5)other loans,the risk of default of each group increases;borrower with a larger credit card quota has a lower probability of default;borrower with a higher salary level has a lower probability of default;the risk of default in the four states of marriage(married,unmarried,divorced,widowed)increases.And at last the weight of evidence conversion was used to convert the result into a credit scorecard form.From the perspective of the platform,using the data of P2 P platforms in Shanghai to explore the identification elements of the problematic platforms,it was found that the registered capital,operation time,platform's background(Whether state-owned enterprises,banks,listed companies or venture capital background),comprehensive rate of return,whether the funds are third-party custody,and the number of target types are significant for identifying the problematic platforms.Platform with more registered capital has a lower probability of having problems;Platform with a longer operation time has a lower probability of having problems;platform with a special background has a lower probability of having problems than the platform without a special background;platform with a higher comprehensive rate of return has a higher probability of having problems;the funds are third-party custody,compared with the funds are not third-party custody,the probability of having problems is higher;platform has more target types,the probability of having problems is higher.Based on this,the scenario setting is as follows:(1)When the operation time is 3 years,there is no special background,the comprehensive rate of return is 10%,the funds have third-party custody,and the number of targets is 2,the registered capital gradually changes from 10 million yuan to 100 million yuan,the probability of having problems gradually decreases from 27.61% to 17.36%.(2)When the registered capital is 50 million,there is no special background,the comprehensive rate of return is 10%,the funds have third-party custody,and the number of targets is 2,the operation time gradually changes from 1 year to 10 years,the probability of having problems gradually decreases from gradually decreased from 57.36% to 0.14%.(3)When the registered capital is 50 million,the operation time is 3 years,the comprehensive rate of return is 10%,the funds are third-party custody,and the number of targets is 2,the probabilities of having problems with private platforms and platforms with state-owned enterprises,banks,listed companies or venture capital backgrounds are 22.64% and 10.04% respectively.(4)When the registered capital is 50 million,the operation time is 3 years,there is no special background,the funds have third-party custody,and the number of targets is 2,the comprehensive rate of return gradually changes from 1% to 25%,the probability of having problems gradually increased from 3.74% to 89.45%.(5)When the registered capital is 50 million,the operation time is 3 years,there is no special background,the comprehensive rate of return is 10%,and the number of the targets is 2,the probabilities of having problems of the platforms which funds with third-party custody and funds without third-party custody are 22.64% and 54.79%,respectively.(6)When the registered capital is 50 million,the operation time is 3 years,there is no special background,the comprehensive rate of return is 10%,and the funds have third-party custody,the number of targets gradually changes from 1 to 7,the probability of having problems gradually increases from 18.05% to 54.76%.