Inferences of default risk and borrower characteristics on P2P lending
© 2019 Elsevier Inc. This paper employs data from China's online peer-to-peer (P2P) lending platform to assess the probability of default as well as the significant impact variables. The research provides some key advantages as follows: (i) we use variable selection methods to identify a parsim...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Journal |
| Published: |
2019
|
| Subjects: | |
| Online Access: | https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=85068068271&origin=inward http://cmuir.cmu.ac.th/jspui/handle/6653943832/65569 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Chiang Mai University |
| Summary: | © 2019 Elsevier Inc. This paper employs data from China's online peer-to-peer (P2P) lending platform to assess the probability of default as well as the significant impact variables. The research provides some key advantages as follows: (i) we use variable selection methods to identify a parsimonious and descriptive model with relatively few parameters that could help predict the default risk of a P2P platform; (ii) employing the logistic quantile regression (LQR) model, we find how those selected variables can affect the default risk in different quantile levels; and (iii) through the predicting evaluation methods, we prove that our selected variables are efficient and bring out the best forecasting performance compared to different variable selection methods. The variables we finally decide to use include periods, loan periods (contract time of the loan), interest due, interest rate, loan type, and regulation change. The LQR estimates show that some variables increase the probability of default and exhibit a significant turnaround on a particular quantile level. The results point out that the new regulation actually brings out more default risk in this dataset than before despite the government's efforts in tightening market control. Checking for robustness by adopting stratified random sampling suggests an easier analysis technique for investors or platform managers. |
|---|