A New, More Accurate Model of Default for P2P Loans

New Accurate Model Default P2p Loans
BadCredit.org Staff
By: BadCredit.org Staff
Updated: June 23, 2020
BadCredit.org publishes personal finance studies on the latest trends in the subprime marketplace. Our articles follow strict editorial guidelines.

Thanks to the unprecedented convenience of the Internet, peer-to-peer lending has become a more popular alternative for consumers to find new access to capital. Sites like Prosper and Kiva have been a go-to for borrowers who have been turned down by banks.

However, as with traditional loans, there’s always a chance the borrower will default on his or her loans.

In “Modeling Default for Peer-to-Peer Loans,” Don Carmichael, a second-year University of Houston C.T. Bauer College of Business Ph.D. candidate, claims to have developed a new — and more accurate — model for default.

The peer-to-peer lending process

For his research, Carmichael analyzed thousands of 36-month and 60-month loans issued through Lending Club between June 2007 and October 2013.

When a borrower applies for a loan through Lending Club, the company uses his or her credit score and report to generate a risk subgrade to indicate potential credit risk.

But unlike traditional loans, ordinary people fund the loans and want to see more information about where their money’s going.

One of the great benefits of the peer-to-peer lending process is lenders tend to have more information available beyond the traditional credit score and credit report — borrowers can self-report income, length of employment, homeownership (own, mortgage or rent), the purpose of the loan, and a loan description.

“We can check, for example, to see if a borrower writes in complete sentences or makes particular claims about himself in his loan application,” Carmichael writes. “[This] allows us to consider borrower characteristics not available for other kinds of loans.”

Explaining default in peer-to-peer loans

While you’ll need to see his paper for the finer details, Carmichael developed an equation based around a myriad of details including, but not limited to, the borrower’s income, revolving credit utilization and even whether or not the user-generated loan description contains end marks (periods, questions marks and the like).

According to Carmichael, Lending Club borrowers are “better credit risks if they write in complete sentences,” claiming that such a detail could serve as an indicator of education level and, by extension, creditworthiness.

Another fun fact in his research — “Generally people are honest when they claim creditworthiness … when people go out of their way to claim they’re credit-worthy, that’s actually a good sign.”

The takeaway

Carmichael drew the following conclusions from his research:

  • Income and inquiries are highly significant determinants of default.
  • Income verification or an interaction between income verification and income is insignificant.
  • Loans for debt repayment should be evaluated separately from other loans.
  • Borrowers submitting loan descriptions lacking complete sentences are more likely to default.
  • Borrowers’ claims to creditworthiness in their loan descriptions are believable.
  • Home ownership, recent delinquencies, derogatory public records and length of employment are insignificant for explaining default.
  • For a median-risk loan, probability of default rises rapidly to a maximum at 13 months and then falls 21 percent from 13 months to maturity.

Just a few things to consider the next time you fund a peer-to-peer loan!

Thinking about applying for a personal loan yourself? Check out our trusted list of personal lenders.

Photo credit: ghostwrittenebooks.com