Not long ago, bankers tallied creditworthiness with a fat envelope full of papers they assessed over the course of weeks. Then electronic reports took hold, with the three major credit bureaus (Experian, Equifax and TransUnion) ruling the roost and brandishing that all-important FICO score. To borrow from FICO’s paradigm, its importance to borrowers might as well be ranked a top-tier 850.
Now, the industry buzz is about instant mortgages, with loans granted in minutes—often based on alternative data. Granted, FICO isn’t about to go the way of the dodo. Yet as they adopt other metrics, fintechs and digital lenders are sending a message: The FICO (named after the Fair Isaac Corporation that created it) no longer serves as the sole traffic light in Credit Town.
Who’s underserved by this system? Millennials, says Steve Smith, co-founder and CEO of Finicity, a financial data aggregation and insights platform. Last year, Finicity launched a credit decisioning product suite and signed partnerships with major players that include Quicken Loans and Experian.
“A high percentage have bank accounts, compared to the general population in the U.S., but they have a lower instance of credit cards,” he said of millennials. “You've got a generation of people saying they don’t want credit cards: ‘I’m going to live off my bank account, my debit card, my Venmo and my PayPal.’”
And because of that lifestyle choice—along with their young age—millennials often lack FICO scores that reflect their true creditworthiness. With the FICO score, payment history on traditional bills such as mortgages carries the greatest weight—followed by available credit, credit history and types, and number of credit inquiries.
While millennials could represent your best buyers and borrowers, Smith says, they might have a “no file” or “thin file” with the credit bureaus. What’s more, someone fully or partially on the "gig economy" might also lack a credit file.
So what could work instead? Smith suggests that by looking at transaction data—cell phone payment records, utility or rent records—a more complete credit picture can emerge.
Predicting the future of predictive models
Given a FICO score’s limitations, “We spend a tremendous amount of time on this: How do you build the most predictive model?” says Ken Meiser, vice president of identity solutions at ID Analytics. “And how do you build the fairest one? And how do you build assurance into this that you can share data with your client in a way that makes them feel comfortable?”
Telecommunications data represents a great add-on to traditional scoring, he says.
“You might have a long relationship with that telecommunications provider that isn’t visible to a traditional credit bureau but is visible to us,” says Meiser, whose company uses analytics to combine traditional and nontraditional scoring that builds fraud and credit risk models.
The telco bill “might be the second largest monthly expense you have, after rent or a car payment,” so a tidy payment history can illuminate patterns not visible to the big credit bureaus, he notes.
Marketplace and online lenders offer another opportunity, says Jason Heil, ID Analytics’ head of credit risk solutions.
According to Heil, “One of the major bureaus plays somewhat heavily in that space: They have 65 percent coverage. Others have less. We’re covering around 85 percent of that market. We see those transactions, which give an additional lift in better understanding.”
Other alternative data sources include social media, says Leo Loomie, senior vice president at Digital Risk. For example, a LinkedIn profile can show current and past employment and where you earned your college degree.
“There’s more impetus for innovation internationally because there are a lot more underbanked, credit-invisible people,” says Loomie, whose company works in risk, compliance and tech services to offer solutions in mortgage, consumer lending and similar industries. Of the estimated three billion credit-invisible people globally, he contends many can pay a loan but lack a traditional data file.
He adds that psychometrics and personality assessments online have become a tool for alternative credit assessment, though generally outside the U.S. FICO is partnering internationally to create online assessments that ask, among other questions, whether someone feels she has the resources to repay.
SoFi—originally a purveyor of alumni-funded student loans—made a splash in lending based on alternative data such as income verification, accepted job offers and “ability to repay.” SoFi even leveraged alumni networks to help those trying to refinance student loans find new jobs—a fintech wrinkle traditional banks have yet to adopt.
Certainly, not all nontraditional lenders have had a smooth go of it. Lending Club had a very visible scandal over corporate governance and loan standards. But momentum stems from the sheer number of new players in this space. And a very young startup, Petal, just raised $13 million to use alternative assessments to give credit cards to people who lack credit history.
The growth of computing power and artificial intelligence have also played huge roles in adopting new models, Smith says. “In an era of big data and more capability around modeling, and more computer power, we’re in a place where we can use data much more effectively than 15 years ago.”
So profound is the influence that even the traditional credit bureaus are exploring alternative techniques. A FICO blog post discusses the use of alternative data in risk modeling, and new products such as this from Transunion and this from FICO carry things further.
Regulation, tradition apply the brakes
But don’t call out the hearse for old-school FICO just yet. As pressure mounts for better credit assessment models, there’s counterpressure to preserve traditional standards.
“When you have years and years of risk modeling, there’s some comfort in that,” Smith notes. “But traditional credit scoring is ‘behind’: It waits until you’re late on a payment and then impacts your score. When you use transaction data, it accelerates your view of risk.”
Regulation—more stringent in the U.S. than elsewhere—also hits the brakes on change, thus quick-moving worldwide trends may grow slower here. And yet, the Federal Housing Finance Authority, parent company of Fannie Mae and Freddie Mac, along with the Consumer Financial Protection Bureau, have sought comment on alternative credit scores. That signals a clear call for change—particularly in equalizing economic opportunity.
Fairness and accountability are also key. “If I am turned down for credit, I have a right to understand why, and what data was used, and whether that data was accurate,” Meiser says.
Yet it wasn’t that long ago—1970 in fact—that fears over excessive information collection and use at Retail Credit Company led to Congressional hearings and the passing of the Fair Credit Reporting Act. (Soon after, Retail Credit changed its name to Equifax.)
Putting it all together: Faster, broader, stronger
Looking ahead, several trends emerge: faster transactions and broader knowledge of course, though that won’t make human judgment obsolete, says Barbara Carrollo-Loeffler, senior vice president and director of residential and consumer lending at Provident Bank, in business since 1839.
“In the next five to 10 years, we’ll see heightened technology, faster turnaround times and the potential ability for consumers to work directly with credit reporting agencies to correct errors and omissions,” Loeffler predicts. “This is still a very personal transaction and while many enjoy the ability to do as much as possible electronically … there has to be a human element.”
Which raises a fascinating prospect: the layering of data and artificial intelligence might just, in the long run, reveal more about the human behind the FICO score.
Want more Banking Strategies? Sign up for our free newsletter!
Jeanne Pinder is founder of ClearHealthCosts.com, an award-winning startup bringing transparency to the health care marketplace.
For more articles like this, check out our recent Executive Report: "New trends in lending and mortgage". With the right information delivered to the right people, you and your team members will learn more efficiently. That's why BAI compliance experts, including former examiners, have designed the BAI Mortgage Series specifically around the needs of lenders.