Zaydoon Munir emigrated to the U.S. from Baghdad, Iraq, so he knows the difference between an authoritarian regime and a free country. But, despite loving our free-market economic system, he criticizes one major aspect of it: credit scoring.
Credit scoring is a system orchestrated by three national credit bureaus (Experian, Equifax, and TransUnion) that track our payment histories–whether we pay our utility or credit cards bills on time (as well as our taxes, car loans, and so on) and, if not, how badly we’re late. They mostly use the FICO system (developed by the company Fair Isaac in 1989), and their numbers are highly determinative, in some cases life-changing. Whether we’re in the 500s, 600s, or 700s makes the difference between buying a house and renting for the rest of our lives, or, perhaps, going to college or not. And it certainly dictates if we can buy a $50,000 BMW 20 minutes after walking into a showroom, and what interest we’ll pay.
Make no mistake, credit scoring is one of the privileges of living in America, Munir says. Lots of other countries don’t keep efficient records and accessing a loan (and selling products and services) is thus all the harder. But, there’s one aspect of the credit scoring system that’s unfair–even un-American, Munir says. It looks only at our past financial history (24 months of it) not at our fundamental creditworthiness. It’s a measure of our past failings, not our potential; of what we’ve become, not what we could be, if we worked at it.
“It doesn’t feel like a 21st-century model,” he says. “I was born and raised in Baghdad. It looks like a system that Saddam would have designed where everyone is required by government to submit their files and you have no say in it.”
Munir’s New York startup, RevolutionCredit, has a tagline of “Be More Than a Score”–which has an anti-Orwellian ring to it. His big idea is to use behavioral science to predict how someone might behave in their personal finances, augmenting the past-facing number kept by the credit bureaus, all other indicators be damned. “Most of the credit scoring models that exist today, in the U.S. or outside, are mainly transaction-data-based,” he says. “This means they are mostly backward-looking and negative selection. The RevolutionCredit model is both forward-looking and positive-selection based.”
RevolutionCredit’s method might appear a little flimsy or unsophisticated at first. It develops “credit clinics”–online puzzles and quizzes that appear before, during, and after transactions, for instance when you’re applying for increased credit with a card company. The clinics are online modules that gather data about us and also educate at the same time. For example, one teaches budgeting, asking users to distinguish between fixed and variable expenses. From how you answer, the module gathers about 200 data points, which can begin to determine a person’s financial “aptitude,” “intent,” and “commitment,” Munir says. Depending on how you answer a quiz question about whether a rental payment is a fixed or variable expense, or how you try to square a theoretical budget, or whether you choose to do the quiz at all, may dictate whether you’re approved a credit increase, or denied it.
I meet Munir at the company’s Spartan offices in the Fashion District, near Times Square. He says he’s just had a big pasta lunch, but he doesn’t come across as post-lunch groggy; he has an excitable amiability the whole hour or more I meet with him. The pasta seems to weigh on his mind, though, because he analogizes about food and fitness often. RevolutionCredit’s modules are like fitness schedules, he says—something some people will choose to do, and stick with; something others will choose, but not finish; and something many people will choose just to ignore. That choice is the most important data point. It indicates whether we’re prepared to improve our financial standing with a company, or remain on the precipice of credit reliability.
Munir, who previously worked at Experian, has an origin story for RevolutionCredit–the sort of folksy tale journalists appreciate even if we know they’re not always completely authentic, or even true. Most ideas aren’t born as lightbulb moments; they’re the slow accretion of thoughts over months of experiences. Munir’s story though involves a burrito from his favorite Mexican restaurant in Laguna Beach, California. He was driving along one day eating it with one hand, while talking into a cell phone with the other. His knees were guiding the steering wheel. Suddenly he sees a police car in the rear mirror and soon he’s stopped for dangerous driving. He’s forced to put the burrito aside, so it’s lying greasily on the side seat as the officer approaches.
Later, Munir had the choice of taking a traffic course as a way of cleaning up his driving record. And it was then that he imagined financial education as a way to purge someone’s financial record–only the financial education would be at the time of the transaction, not some time after the fact when it was less relevant. After two years or so of looking at how to systematize the idea into online modules, he hit upon something workable in September, 2013. Since then, RevolutionCredit has signed up 18 customers, a mix of marketplace (peer-to-peer) lenders, and credit card companies; recently one electric utility has started using the modules as well. (Unfortunately, Munir won’t reveal any company names–for confidentiality reasons he says. But there seems no reason to doubt RevolutionCredit’s success.)
The modules are used in a variety of ways, but always in a positive direction, Munir says. They’re designed to expand credit access or increase engagement between company and customer, not to deny someone service. “We might say ‘you qualify for loans at this rate, however, if you complete one or two modules, you get a lower rate.’ It’s always related to approval or the rate, because those are related to your credit score,” Munir says.
The most common applications are collections. If you’re late on a bill, you might get a text or email asking you to carry out a survey and be rewarded for it–say, with a longer grace period for paying the bill. By working with customers this way, the company increases the rate of collections by up to 37% compared to standard collection processes. “People who do RevolutionCredit have a 30% lower delinquency rate than their peer group. They write off [debt at a] 20% lower rate than their peer group,” Munir says. “We are identifying people who want to help themselves. Today [most of the time] there’s no other option. You’re just stuck with the credit score.”
Another use case is so-called “marginal declines.” That’s when someone is refused credit but the decision was close. Perhaps the product calls for a 650 score, but you only have 640. By asking customers to go through one of six RevolutionCredit modules, lenders can understand which of the rejected customers might actually be good bets as credit risks. And, in so doing, financial providers can expand their customer bases. They can include people with 640 scores, but at little or no extra risk than taking someone at 650, Munir says.
Though it’s still relatively early for the company, RevolutionCredit is a potentially big idea because it shifts creditworthiness away from a simple numbers game. It becomes an activity that people participate in, not just something an elusive company keeps a record about. Munir sees potential to expand into less developed markets where financial exclusion is an even more serious issue, and where the backward-looking data is less sophisticated than it is in this country. Behavioral science, however sophisticated, will never replace scoring completely. But it can offer greater perspective and help fill in some gaps in traditional data, he says.
“In 2011, when I was first shopping my business around, I got turned down because people said it’s the era of big data. They said: FICO will be replaced with a better score with more data,” Munir says, still smiling. “I said, ‘No, you may get better, but you’re not transforming anything.’ And I was right. Let’s be honest, FICO or VantageScore are big data models already. To build an accurate and fair credit scoring model, one should balance data from the past and the future.”