It’s gonna get worse before it gets better. With what we see with increasing prices because of the tariffs and things that are going on right now, that expense to the consumer is going to go up.
As long as banks and other holders of consumer financial data continue to play fair — a tenuous prospect at best — cash-flow underwriters like Nova Credit offer relatively oracular insights into the state of consumer financial health: providing more systemic and real-time analysis than the FICOs and VantageScores of the world do.
The company’s number crunchers see far from a rosy picture for financial stability. Student debt repayments, tariff-induced price increases, and other policy-induced economic shocks are driving a household liquidity crisis, nudging more consumers to use debt products like BNPL to MacGyver their way through serious cash-flow issues. To complicate matters further, much of that BNPL usage is invisible to lenders, which may affect their ability to make sound and secure underwriting decisions.
In an interview with Fintech Nexus, Nikki Cross, Senior Director of Data Science Solutions at Nova Credit, describes the kinds of data the fintech can access, outlines how data flows are faring in the face of open banking’s uncertain future, and predicts a challenging year ahead for people living paycheck to paycheck.
The following has been edited for length and clarity.
You have been paying attention to changing student loan repayment trends, especially after the Department of Education’s Office of Federal Student Aid (FSA) resumed collections. What kinds of indices are you using to gauge the health of consumers affected by student loan debt?
It’s difficult on the cash flow underwriting side to understand who might have a loan but not be making payments. Cash flow underwriting reports give us a unique position to be able to understand who is making payments, and whether they’re really prepared for them on a more ongoing basis. But, unfortunately, like most things with your credit report, it is a bit of wait and see, because it’s generally difficult to understand who is going to be in that position to resume the payments, and who’s not going to be. That will take time until the bureaus start reporting on that. We’re probably three to six months away for anything very comprehensive there.
Do you have access to seeing things like whether people are prioritizing rent over student loan payments?
It’s going to take a little bit to get through everybody getting in their first payments. So it’s an analysis that we’re figuring out, but not that we’ve actively started.
Everybody has the same question: What’s the best way that we can think about what we can see versus what we can’t see? Because, again, we don’t know who has a student loan that is in deferment that they aren’t making any payments on at this point in time. So there are going to be gaps in consumer-permissioned data. There’ll be gaps in what we can and can’t see, and so trying to create some sort of a semi-optimal view of that — taking the data that we do have to be able to leverage that, and then the appropriate caveats for what we can’t see.
A lot of it comes down to the nitty-gritty of statistical analyses. We look at the things that we do see and try to normalize a similar population based on criteria like payments, signs of distress based on cash balances, outgoing and incoming ratio, things like that. And then you use that as your testing control, regardless of some of the things that we are blind to.
Apart from student loan repayments, are there other shifts that you’ve identified since the start of the year that have affected your view into consumer wellbeing?
BNPL is the thing that everybody asks about, and that we’re doing a lot of research into. FICO’s announcement [that it’s incorporating some BNPL data] has only doubled down people’s interest. There’s a lot to be said for finding ways to include that data in the score. We’ll always push for the idea that as much data as we have, it should go into the score. What I think is harder for a lot of people on the street to grasp is the idea that inclusion is only going to be as effective as the representation that feeds it. So your score can only be as good as the data that underlies it. And FICO has said that it has been built with just Affirm’s data, which is more on the prime side. We know that most buy now, pay later lenders aren’t furnishing that data to the bureaus, and so that score is only as good as the data that gets fed into it. And what we know right now is the vast majority of BNPL has no motivation to report to the bureaus. And so sure, you’ve got a score, but how effective that score can be is still to be determined, because there’s no motivation for the totality of the data that should be represented to be fed into that score. And so that’s what we get with the cash flow and deriving data: We see PayPal and Klarna and Affirm, and we see everything that’s coming out of that account, and so when we start to see what the totality of BNPL payments is for consumer, how much they’re using across different providers, we’re starting to generate some really interesting insights that won’t be represented in the credit bureaus until there’s at least far more holistic reporting than what we’re currently seeing.
How are you doing that? Are you doing that through credit card amortization, or are the mechanics there?
Anytime those payments are coming out of that checking account, we see the payer and the amount there. In many cases, BNPL is a tool of last resort, but we also see plenty of people who have plenty of money sitting in the bank and are just choosing it for free money. It’s kind of starting to tell the story of two very disparate populations who are using BNPL. There are interesting things like that that just aren’t represented in the traditional data that we’re getting a lot of questions about because it is an interesting and ever-growing blind spot.
Ever-growing being the key term here: Even regulators don’t know how big the market is.
I think there was some expectation in the past several years that eventually there would become a regulatory mandate for them to report in the same way that has been for [other debt products]. But if you think about it, so much of the credit bureau space is built on the idea of gift-get: Unless I report my performance, I can’t buy the data I need to make decisions. In spaces like payday and in spaces like BNPL that doesn’t matter as much, because I might buy a report for the first one, but the 17th time you take out BNPL with me, I already know that you paid back 16 other times, so I’m not bothering to pull. The same thing happens in payday.
So until you get to a space in which there’s some sort of a motivation to report, all I’m doing, if I’m a lender, is sitting on this population of people I know information about that none of my competitors do. Why in the world would I freely give that away? There’s no motivation to report unless you’re obligated to do so from a regulatory perspective. And certainly the current administration doesn’t seem to show any interest in making that a regulation. So until that gap gets resolved, cash flow underwriting is really the only way to create some sort of a complete picture in the BNPL space. And so it’s something we get asked about by almost every lender we talk to, because it is the biggest gap right now. Seven, eight years ago, people might have said it was payday, but the Bureau has found a way to generally track that. So you’re now in a position where BNPL is in that place, and there just isn’t a reason to give away information when it’s driving your revenue.
It’s not just lenders looking to understand BNPL usage — the regulators don’t know how big the industry is. But if I’m a lender, even in the super-prime space, I just don’t know how much of your money is going into BNPL. I could convince myself with a traditional report that I understand your financial situation, I understand your obligations, and then there’s this giant hole. Because in the same way that five to seven years ago, everybody cared about loan stacking in the installment space, BNPL is designed to be stacking. And so we see people have hundreds of dollars in BNPL payments over a two-week period; that’s the equivalent to having a car payment I didn’t know about. Again, for some people, it is just a cash-flow thing: They’re just choosing to do this because they consider it to be free money, and they’ve got the ability to pay it back. But that’s not everybody. When you start to see three- and — in some cases — four-figure BNPL payments that have been made over a month, that’s eye opening and changes the way that you think about a consumer.
What we know more than anything in the financial services space is that we want certainty. At least, if I know you’ve missed this payment, I can price you accordingly. I can make you the right offer at a lower dollar, higher interest. If I know you’re subprime and I know that that is the case, I know what to do with that. And if I know you’re super-primed, I know what to do with that. And in a world where I don’t know, it’s the scariest thing in the world for a lender, because now the risk is all on me, as opposed to being able to use the data and to use that information to be able to make a decision there. And so we look at cash flow underwriting to gain information to fill in this huge gap, this huge point of uncertainty.
How do you gauge who is using BNPL for free money, and who’s using it to make their everyday purchases at a grocery store?
I think the easiest way to just get a gut feeling on that is their current balance and their assets. In a test that we recently ran for a BNPL lender, we saw that about a third of the population had less than $100 in total assets based on what they had connected. That person’s having a liquidity crisis. That person is putting their groceries on BNPL because they need groceries, they just don’t have the money to pay for whatever it is they need at this exact moment in time. There’s a comparably sized population that has more than $2,000 in assets. Those people can afford their groceries if they choose to, and they’re making a different choice there. And so the easiest, immediate way to think about that is just: Could they have paid cash if they wanted?
What happens to cash-flow underwriting if open banking doesn’t materialize?
We are still seeing lenders who are choosing to move forward as if 1033 were in place, and consumers who are adopting in those same ways. Like I said before, lenders like certainty. And I think the approach that we’re seeing from most lenders is it’s still the law of the land, regardless of if it’s being actively enforced. And so as quickly as quickly as this administration has chosen not to, the next administration could say that they’re going to enforce it. And because big ships turn very, very slowly, it behooves them to continue to move forward, assuming that it will become enforced at some point in time. Once there is some momentum, once there is some traction, it just doesn’t make sense for most lenders to try to put the brakes on that. What we’re seeing is that maybe people are moving a bit more slowly than they might have, there’s not a specific deadline that they’re trying to hit in the same ways, or maybe with the same urgency. But what we generally see is that people are still moving forward.
The thing I will put my finger on is, right now, in a non-1033 world, there are some banks and lenders that limit the amount of data that goes out. It’s hard, because some services are going to do screen scraping and things like that. So you can’t really limit it. We can just make it less secure, which is a poor choice to make. But what we do see is that there are people who are choosing to limit the amount of data that goes out. So for many banks and lenders, you can get up to 24 months of transaction history. There’s a much smaller subset for which you can only get, like, 90 days. Those, I think, fall into a couple of categories, and I think one of those are banks and lenders who are notoriously tight with their information. Realistically, if we peel it back, I think there are also some for whom it’s just a technology choice. It’s like, that’s what I’ve enabled my consumers to see, and those are generally, probably, I would say more regional or smaller banks, some of whom have homegrown systems. So there are two groups who are sharing less data for slightly different reasons.
Where do you foresee consumer debt, or the state of affairs, heading if you look at things like BNPL usage and these other indices you can access?
It’s gonna get worse before it gets better. With what we see with increasing prices because of the tariffs and things that are going on right now, that expense to the consumer is going to go up. And what economists say about what, what’s likely to happen because of the One Big Beautiful Bill and the implications that it has for food scarcity and things like that, what we’ll expect to see with health care costs, it appears that of these impacts are going to hit the people who have the least ability to be resilient to them. We know half of American households live paycheck to paycheck, including a third of high-income households. It doesn’t take that big a shock to the system for the majority of households to be unable to make the recurring payments that they would expect. It’s one thing when it’s a flat tire or a medical bill [affecting individual households], but when you add in an increasingly impactful systemic issue, it doesn’t bode well for the average American consumer, and it doesn’t bode well for the overall financial health of the nation.
For fintech operators, I imagine that means the people who have paid off their BNPL debts 17 times in a row might fail their 18th time, kicking off a liquidity crisis for individual households, and then a domino effect from there.
Because I’ve spent most of my career on the lender side in the subprime space, what I will always say is that there’s sometimes this idea that people are subprime and have subprime credit ratings because they aren’t paying attention and they’re not educated. And I say to those people, you have never listened to a call center and heard somebody in a subprime space — like the payday space that I was in for a bit — who’s telling you, I need you to take $2.17 out of this checking account, and then I’m gonna make a $17 payment out of this checking account. The privilege I have to not worry about the pennies that are in my account is not one that many consumers have. And so it’s not an education issue in many cases — it is in some — but just there’s not enough money to pay everybody, and you make the most strategic choices that you can for yourself and for your family, and to keep a roof over your head, to put food on the table. It’s sometimes easy in a world where we do have privilege, and we do have a cushion, to forget that you have to make those very tactical choices sometimes. So yes, the people who are in those positions, who are making the most strategic choices they can with the limited means that they have, are going to figure out how to make that work for them. And if that’s the payday loan that gets them through the next week, or if that’s the BNPL that gets them the groceries that they need, they find ways to make all those dominoes work. And you hope, and you pray, for a long period of time that those all still hold. But it’s a house of cards at some point in time, and if you’re not able to keep all of that together, it’s devastating for that family.
One last question: What would that kind of contagion look like quantitatively on Nova Credit’s side? Is it collapses in bank balances, discontinuation of BNPL activities, et cetera?
You will start to see those issues coming out in credit reports in a few months, but the leading indicators that we look at are going to be things like bank balances and incoming and outgoing ratios. People for whom we see funds just aren’t coming in at the same rate they’re going out.
