Scary as it sounds, deriving metrics like a credit score from behavioral data could be attractive for both parties involved.
Daniel Björkegren, an economist at Brown University in Providence, Rhode Island, endeavored to predict how likely someone was to pay back a loan based on call metadata. In results presented at the NetMob conference in Cambridge, Massachusetts earlier this month, he analyzed the records of 3,000 borrowers from a bank in Haiti — who they called, for how long, and how much money they spent. After applying an algorithm, Björkegren found the bank could’ve reduced defaults by 43 percent if it’d taken his approach to selecting trustworthy applicants.
How’d he figure? Information like phone metadata can be easily extrapolated. Prompt bill payers are typically more reliable than those who hold off, and people who make frequent calls far outside a bank’s network are more likely to have trouble making deposits. However, even the esoteric information can factor in — researchers at Cignifi, a Cambridge-based firm studying the predictive capabilities of mobile data on loan repayment and savings, found that the time of day and neighborhoods from which calls are placed can be indicators, too.
Two startups, Inventure and Lenddo, are shopping the idea of “data mining for credit” to lenders. As part of a pilot in South Africa last year, Inventure gave smartphones to a small group of people and collected phone and demographic data as they paid the company back. Lenddo — which monitors interactions, connections, and activity on social networks– already conducts business in the Philippines and Colombia.
Think the implications are worrisome? You’re definitely not alone, but there’s an argument to be made in favor of companies like Lenddo. Kyle Mead, innovation director of credit-scoring company Entrepreneurial Finance Lab (EFL), told New Scientist about his experience trying to get a loan after moving to the U.S. from Canada. “I had no debt, no previous information on me in the U.S.,” he said. “The highest I was offered was $300, because I’m credit thin.”
“I had no debt, no previous information on me in the U.S. The highest I was offered was $300.”
Millions of people in developing countries are in the same boat. They lack any kind of financial history, making institutions hesitant to lend. That’s where behavioral monitoring services can really come in handy — Lenddo has been used to approve loans for “phones and motorcycles,” said a spokesperson.
Scary as it sounds, the idea of deriving metrics like a credit score from behavioral data is an attractive one for both parties involved — banks get better snapshots of customers, and first-time borrowers encounter way less resistance than they otherwise would. The process may be invasive, but as our dependence on Internet-connected devices grows, it’s only likely to become more pervasive.
Next time you apply for a loan, just read the fine print carefully. Very, very carefully.
