In the days of community banks, a person’s upstanding reputation around town gave him access to a reasonable loan. But global financial institutions can’t trust strangers, so the credit rating was born.
A startup called Lenddo hopes to return lending to that community bank era, but with a modern twist. The company gauges a person’s creditworthiness using his or her online reputation, as assessed through sites such as Facebook, Twitter, and LinkedIn, to grant loans. To secure repayment, it forgoes collateral and instead relies on peer pressure through the same social networks.
The target market is a demographic often ignored by banks today: the 1.2 billion people, largely in developing countries, who are part of the world’s emerging middle class but who still struggle to access credit because they lack a documented financial history and strong identity records. “Our theory is, we could duplicate the social dynamics of microfinance, but instead do it online,” says CEO Jeff Stewart, referring to the practice of making small cash loans to the world’s poorest people and relying on peer accountability to ensure low default rates.
For now, Lenddo is lending several hundred dollars at a time—the equivalent of one month’s salary—to applicants in the Philippines and Colombia. In May, it raised $8 million dollars from investors to add engineers and expand into new countries.
Lenddo sees a big prospective market. Mexico alone will graduate 800,000 engineers in the next four years, and many will want to access credit for the first time, Stewart says. Money is only loaned to pay for “life changing” expenses like health or education bills.
Stewart says Lenddo is ultimately not a lending institution but a technology company—and he hopes that large financial institutions will use its platform to expand worldwide access to many financial services, from credit to insurance.
The heart of Lenddo’s underwriting platform is the technology it uses to calculate a credit score after a user grants access to his or her accounts, including Facebook, Gmail, Twitter, LinkedIn, and Yahoo.
The company relies on three classes of algorithms to gauge a person’s likelihood of loan repayment. One validates truthfulness; for example, it would be statistically odd if a supposed engineering student in Bogota had few friends at school or never wrote e-mails containing certain words. Another looks for behavioral and demographic clues that predict the probability of repayment, similar to how online ads are targeted based on Web surfing patterns today.
The last element Stewart calls a “PageRank for people,” referring to Google’s method for returning high-quality search results by examining the credibility of incoming hyperlinks. Lenddo encourages loan applicants to invite their most reliable friends to sign up themselves for a Lenddo score and become part of the user’s “trusted network.” A higher-quality network yields the applicant a higher Lenddo score.
That’s also where accountability comes in. Once a loan is granted, Lenddo will inform a person’s network about a late payment, and friends’ scores will also drop. Today, Stewart says, applicants are sometimes denied loans when they have not yet built a strong network, an activity that will become easier as people in a country repay more Lenddo loans.
It’s not hard to imagine a scenario in which a sophisticated scammer could manipulate the system, for example by creating fake accounts or compromising real ones. However, even traditional credit scores are not perfect. During the financial crisis, many people chose to default on their loans in ways their past history and FICO scores did not predict. Stewart says that, so far, repayment rates have been higher than 95 percent.
Rodney Nelsestuen, senior research director at CEB Tower Group, says that credit-reporting firms and other financial institutions in developed markets are also taking a “tremendous interest” in exploring how social and online data could improve credit scores and loan decisions. But the work is still in its early stages, and there are privacy and regulatory concerns, he says. Many people unwittingly share more publicly online than they realize. And while the value of the data might be well established in retail and marketing sectors, some financial institutions want to see even stronger behavioral correlations before using the approach.
Nelsestuen is skeptical himself. “Everybody is a little smarter, a little richer, and a little better looking on Facebook,” he says. “In a credit environment … we need to see what happens when people aren’t putting their best foot forward.”
A horrifying new AI app swaps women into porn videos with a click
Deepfake researchers have long feared the day this would arrive.
We can’t afford to stop solar geoengineering research
It is the wrong time to take this strategy for combating climate change off the table.
Meet Altos Labs, Silicon Valley’s latest wild bet on living forever
Funders of a deep-pocketed new "rejuvenation" startup are said to include Jeff Bezos and Yuri Milner.
The new version of GPT-3 is much better behaved (and should be less toxic)
OpenAI has trained its flagship language model to follow instructions, making it spit out less unwanted text—but there's still a way to go.
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.