How Social Data Might Reinvent the Credit Score

All too often, one’s creditworthiness is mistaken for one’s trustworthiness. Our credit scores are used to determine 90 percent of credit decisions, and can affect whether we’re able to rent an apartment, buy a car, or even get a job. And for an awful lot of Americans, that’s a really big problem.

According to the FDIC, approximately 28 percent of U.S. households are un-or underbanked. That means they don’t have access to FDIC-insured institutions, and must attain their financial services from non-insured organizations like a non-bank lender or check casher (the home of nefarious “payday loans”). While some families or individuals may have opted out of the banking system entirely, the majority of the underbanked never had a choice in the first place: They simply don’t have the credit history to demonstrate that they are reliable borrowers.

Lately, however, there has been a push to find a way to replace or reinvent the current credit score. Even a young man named Louis Beryl—a graduate of Princeton, former Morgan Stanley employee, and, as he puts it, “just about the most low-risk person possible”—was flat-out denied by several private loan companies while he was attending Harvard business school. He needed a co-signer to guarantee his loan before he was approved.

Beryl went on to launch a company called Earnest in 2013 to help lower the high costs and barriers to credit faced by otherwise financially responsible people. Rather than viewing applicants as numbers, Beryl says Earnest assesses them as individuals—reviewing LinkedIn and other social profiles, education level, work history, utility payments, real-time payment behavior, cash flow, savings over time, salary, and earning potential—to determine whether someone might be a trustworthy borrower. Beryl describes Earnest as “more of a first resort than a last resort” for loans and credit; many loans are typically approved and fulfilled in under two days.

Earnest isn’t alone: LendUp similarly mines borrowers’ Facebook and Twitter accounts for factors that indicate whether they’ll be likely to default, as does Affirm, which also gleans information from cellphone accounts. Beryl says this kind of data analysis will become more and more common, especially as more young people approach financial maturity. Without a long credit history, even the most successful millennial may struggle with the asymmetrical gap between their online reputations and what banks traditionally care about.

For people who have less access than a former Ivy Leaguer, FactorTrust might be an option. This credit bureau was developed specifically for those without a financial institution, using a proprietary database of underbanked consumer loan performance and third-party data sources to determine creditworthiness.

FactorTrust CEO Greg Rable says that though “non-prime consumers” have been overlooked traditionally, they are finally able to get some attention through alternatives to the credit score. Rable is quick to add, however, that though social media data is an intriguing way to assess one’s social standing and professional connections, regulatory and compliance frameworks simply haven’t caught up with these kinds of innovations. Essentially, people who rely too much on the dream of the “social” credit score are as stranded as ever.

For more traditional financial stalwarts, there may be yet another way to assess credit newbies: Since November 2014, a new pilot program—still unnammed and developed by the Fair Isaac Corporation, Equifax, and LexisNexis Risk Solutions—has been trying to dig up the true value and meaning of our current credit score system, assessing whether there’s a better way to paint a complete, accurate picture of borrowers.

This new solution, though still in its infancy, could potentially upend our current socieconomic system, making auto loans, mortgages, credit cards, and jobs accessible to as many as 53 million new people by taking financial institutions out of the equation. Instead, the program will calculate a score based on consumers’ payment history when it comes to cable, cellphone, electric and gas bills, as well as how often he or she changes addresses, among other factors.

According to Jason Flemish, Vice President of Consumer Credit and Risk Products at Equifax, this kind of data is valuable because of two factors: the breadth of coverage, and predictability of the data source. “Real-time payment behavior data—such as how timely a consumer is when paying their cell phone bill—is a very desirable data source,” he said. “This is because it is a good indicator of how [well] the consumer would treat a future line of credit.”

It may seem like alternative data is all the rage right now, but Chi Chi Wu, an attorney at the National Consumer Law Center, warns that not all data is created equal. She says that data from utility companies in particular is problematic, as these bills can fluctuate wildly (think of a winter heating bill in Boston, for example), which may leave people falling short at certain times of the year—and even negatively impact their ability to get credit in the future.

Wu also points out there is a vicious cycle in this kind of credit reporting. If something bad happens—say an illness or accident—and leaves someone in a bad spot, instant real-time financial data may make it harder for borrowers to regain control of their personal and professional lives. Nearly fifty percent of employers use credit scores when making hiring decisions for some positions, says Wu.

“Using alternative data doesn’t get rid of those problems,” she says. “It’s a different type of data, but if you have a consumer who [faces unpredictable obstacles], they’re as likely not to pay their phone bill as anything else.”

Wu suggests the answer may lie in a combination of returning to fundamentals of underwriting—like getting to know customers’ ability to pay—alongside new methods of analysis. Earnest’s Beryl sees a rosier future. “Ultimately, we’re talking about people who are financially responsible paying far less through a more fluid system.”