New rules for new lending

On the web, those who make microloans prosper with curious criteria

CAMBRIDGE, Mass., December 1, 2011—When is a bad credit rating a potential boon to the borrower? When the loan is being sought on the microlending website Prosper.com, an MIT Sloan researcher has discovered.

Juanjuan Zhang, an associate professor marketing, and her colleague, Peng Liu of Cornell University, analyzed over two and a-half years of data on the behavior of lenders and borrowers on Prosper, the largest of the microlending websites and also one of the oldest. Lenders look most favorably on borrowers who are attracting loans despite poor credit ratings, the researchers found.

“They infer that there must be something really good about a borrower who is getting a loan in spite of having a poor credit rating,” Zhang said. “Similarly, if a borrower’s proposal gets an endorsement from a friend, lenders will tend to discount the amount of funding that the borrower has raised, attributing much of it to friends’ support.”

With the turmoil in credit markets in recent years, web-based microlending, also known as peer-to-peer lending or crowd-funding, has emerged as an important alternative to traditional lenders. On Prosper.com, individuals who want to borrow money state their case in a public posting. Borrowers typically rely on multiple lenders, each of whom contributes a portion of the loan.

The seemingly counterintuitive calculations of Prosper lenders are not the only surprise that Zhang and Liu discovered. The researchers tracked the performances of the loans to determine which ones were more likely to default.

Their conclusion: Loans with the lower credit ratings favored by lenders tended to perform better than those that were equally popular among lenders but had higher ratings.

“The reassuring news from this study is that people on microlending websites tend to be savvy, entrepreneurial investors who learn from their peer lenders,” Zhang said.

While banks and mortgage companies may have better access to the credit histories of borrowers, microlending websites have a source of information traditional lenders do not have: the decisions of other lenders.

The loan approval process on Prosper is transparent. Potential lenders see what a borrower has offered and whether other lenders found the borrower trustworthy. “Loan officers at banks can go through your credit file and gauge your credit-worthiness, but they don’t always see the decisions of a dozen other banks,” Zhang said.

In their study, Zhang and Liu set out to understand the dynamics of lender behavior. One possibility they considered was that lenders simply imitated the behavior of earlier lenders, a phenomenon known as “irrational herding.” If irrational herding was at work, then a loan proposal that received a strong early response would attract more and more lenders—even if friends’ endorsements accounted for the strong early response.

“In irrational herding, you passively mimic what others did without considering the reason behind their choices. You don’t second guess the motivations or the circumstances behind their choices,” Zhang said.

What the researchers found instead was evidence of what they call rational herding. “In rational herding, a borrower’s funding momentum becomes modified based on loan attributes,” Zhang said. “Rational herding happens when you see what people do and you consider the environment, the context, and the reasons for what you see.”

An expert in social influences, Zhang is a leading researcher into the phenomenon known as observational learning, or the tendency of individuals to watch the behavior of others, learn from it, and act accordingly. Observational learning has important implications for marketers, consumers, and anyone involved in the exchange of goods and services in a social environment.

In addition to explaining the workings of microloans, Zhang has used the concepts of observational learning to understand why patients waiting for transplant kidneys are reluctant to accept an organ that was turned down by other patients, why homes that languish on the real estate market after their initial listing become increasingly difficult to sell, and why website visitors are drawn to contents that have received large volumes of clicks. People assume that peer patients, earlier home buyers, and other website visitors have made their choices for good reasons.

For Media Inquiries

Paul Denning
Director of Media Relations
Tel: 617-253-0576
Fax: 617-253-5875
E-mail: denning@mit.edu

Patricia Favreau
Associate Director of Media Relations
Tel: 617-253-3492
Fax: 617-253-5875
E-mail: pfavreau@mit.edu