Table 7. Logistic regression analysis for potentially insolvent status of P2P borrowers, showing B coefficients and significance levels.
Model 1 | Model 2 | Model 3 | Model 4 | Model 5 | Model 6 | Model 7 | |
---|---|---|---|---|---|---|---|
Borrower Assessment | |||||||
Subgrade | 0.086*** | 0.095*** | 0.089*** | 0.087*** | 0.090*** | -0.008 | |
Interest Rate | -26.102*** | -32.924*** | |||||
Purpose | |||||||
Car | -2.742* | -2.383 | |||||
Credit Card | -1.717 | -1.605 | |||||
Debt Consolidation | -1.731 | -1.527 | |||||
Educational | -2.113 | -1.646 | |||||
Home Improvement | -1.588 | -1.416 | |||||
House | -2.150 | -1.980 | |||||
Major Purchase | -2.555* | -2.265 | |||||
Medical | -2.708 | -2.264 | |||||
Moving | 19.148 | 19.461 | |||||
Small Business | -1.878 | -1.372 | |||||
Other | -1.957* | -1.966 | |||||
Borrower Characteristics | |||||||
Housing Situation: Own | -20.784 | -20.826 | |||||
Housing Situation: Mortgage | -20.592 | -20.327 | |||||
Housing Situation: Rent | -20.498 | -20.334 | |||||
Housing Situation: Other | -20.843 | -20.609 | |||||
Annual Income | 0.001 | 0.001 | |||||
Credit history | |||||||
Inquiries Last 6 Months | -0.327*** | -0.337** | |||||
Delinquency 2 Years | 0.357 | 0.438 | |||||
Public Records | 0.413 | 0.428 | |||||
Revolving Utilization | -0.331 | -0.430 | |||||
Indebtedness | |||||||
Loan Amount to Annual Income | -11.227** | -11.685** | |||||
Annual Instalment to Income | 24.219** | 24.172** | |||||
Hosmer–Lemeshow test | 0.730 | 0.942 | 0.449 | 0.766 | 0.236 | 0.168 | 0.505 |
Nagelkerke R Square | 0.076 | 0.078 | 0.114 | 0.088 | 0.128 | 0.124 | 0.212 |
Correctly predicted (primary sample) | 58.8% | 58.0% | 59.7% | 60.2% | 60.1% | 62.0% | 64.6% |
Correctly predicted (holdout sample) | 75.2% | 62.0% | 72.8% | 76.0% | 71.9% | 80.6% | 65.1% |
Primary sample comprises 274 loans funded in 2008 first semester, where 137 are defaulted and 137 non-defaulted. Test sample comprises all the 3,788 loans funded in 2011 third trimester, where 401 are defaulted and 3,387 are non-defaulted.
*** significant at the 1% level
** significant at 5% the level
* significant at the 10% level.