TransUnion Live: Affordability Impacts to Card Payments

Watch our on-demand TransUnion Live for a candid conversation between Josh Turnbull and Craig LaChapelle, industry experts and hosts of the card and banking Extra Credit podcast, as they explore the link between understanding consumer affordability and maintaining steady growth. Based on recent TransUnion research, the discussion will cover everything from how consumer card origination has shifted since before the pandemic to tailoring offers that accommodate affordability concerns while managing risk.

Full Transcript

Craig LaChapelle:

Welcome to TransUnion's card and banking LinkedIn live session. We're very excited to have folks here with us today. We like to think of this as our special live version of our Extra Credit podcast, which can be found on Spotify, Apple Podcasts or wherever you get your podcasts.

We are Josh Turnbull and Craig LaChapelle, hopefully you can get our background from LinkedIn, and we're excited to be here. I'm very excited to grill Josh today on the study that he and his team led. This study explored credit card risk and marketing dynamics in today's uncertain environment, highlighting key observations, conclusions, but we also address the implications for card issuers. By the way, you can further explore the so-called affordability study by downloading our market brief, which I believe is linked in the event description.

For today, we're going to try to stick to our theme and of our Extra Credit podcast. We try to stay within 30 minutes. We do welcome questions related to today's conversation. Please put them in the chat area, and if they're applicable and we don't get to them we'll do our best to respond afterwards. Before we turn it over to Josh and sharing the findings of the study, want to refresh everyone on market background of the situation for both consumers and issuers. If you remember, record originations for both bank card and retail really occurred in 2021 and 2022 as folks came out of the pandemic and returned to spend.

This led to balances growing, but additionally inflation for a lot of other reasons increased and made these purchases more expensive. This yielded higher balances coupled with higher rates that we're starting to see as the Fed tries to tackle that inflation meant and mean that minimum payments due are significantly higher than pre-pandemic. Additionally, as TransUnion forecast and others forecast delinquency rates increased, which are the logical consequence of these conditions. Issuers are sitting here in 2023, they're still recovering from the 2022 retreat, they've seen great originations in '21 and '22 so growth expectations are positive or higher in '23, but there is a desire to safely grow.

The question is how do you do that not only with the increased demand, but with the increased delinquencies and increased I would say consumer affordability challenges? Issuers are discussing with us and asking us our advice about how they should consider risk strategy refinements. In particular after the strong '21 and '22 drove growth expectations, the prime vintages, the prime risk segment in particular, vintages in those time periods perform worse than they did historically. Those vintages perform worse than other historical vintages, so our customers really asked us quite pointedly, "What is going on? Are scores no longer effective? How do I navigate this environment?" This was the genesis for the TU study, which I'm going to turn to Josh and begin quizzing him on right now. But before I do that, Josh, pretty good setup. Did I miss anything?

Josh Turnbull:

No, we're good. Thanks, Craig.,

Craig LaChapelle:

Great. Great. By the way, you are not muted, so that's good.

Josh Turnbull:

Good. Good. Good first step. Quiz away.

Craig LaChapelle:

Yeah. Josh, how did we approach answering these questions? How did we structure the study?

Josh Turnbull:

Yeah, no, good start there, Craig. All those questions that you mentioned, we were asking ourselves, customers were asking us, and so what we did was we took two national samples. We looked at consumers who had originated cards in the period pre-pandemic and then a sample of consumers who had originated cards in I believe it was late 2021. We really wanted to dig into those and see what did we find in terms of segments, what was the charge off behavior in those, but then what distinguished different segments? Were they performing the same, differently than they had been before? Most importantly then based on that, "If I'm an issuer, what might I do knowing that to address origination strategies or to manage those accounts that I'd booked, to your point, that may be same credit score but are performing worse than I would've expected based on how things have been pre-COVID in 2019 and before?"

Craig LaChapelle:

Great. Great. That's how we structured the study. Let's jump to it. What did we find in our study? What are at least the headlines?

Josh Turnbull:

Yeah, so a couple of headlines. One, to your first question or point, are credit scores still working, yes, they are. They are very effective rank orders of risk that we saw and we continue to see in analysis that we do. What we did find and what many of our customers also found is that... I'll just pick a score of 720. If I originated a tranche of consumers with a credit score of 720 in 2021 and the same in a tranche in 2019, that 2021 cohort was going to perform materially worse than the pre-COVID cohort. That's where we dug in and really wanted to understand what was different between those two groups.

The first thing we did was try and segment people out or segment people and look at what credit scores had been. There are four segments that emerged. There's what I would call the Steady Eddies, people who've always kind of had 720, plus or minus a couple points. There's a group of people whose scores were on the decline to 720, and I'm using 720 arbitrarily, when they opened the credit card. There's a group of people whose scores had risen fairly dramatically to 720 on that point which they originated a card. Then there's a group of people who were relatively new to the bank card scene who had originated. Those four segments of folks, kind of the decliners, the Steady Eddies, the people whose scores had gone up, and the people who were new to the scene.

What we found when we dug into that is, and this makes sense when you think about it logically, if someone has always been about a 720, if they were a 720 in 2015 and in 2019 and if they still are in 2022, their performance really hasn't changed. They perform today just like they did a couple of years ago. The group where we did see a big change are those peoples whose scores had increased. Even in 2019, those folks whose scores had gone up a good amount, they performed worse than some of the other segments, the Steady Eddies or the peoples whose scores had gone down, but one of the big changes that we saw is their performance. It was almost twice as worse in the '21 sample as it was in the 2019 sample. That's one big difference, is that it's the people whose scores had gone up dramatically, they were the ones that are performing materially worse than the rest of the population and then worse than those people that were on the same trajectory several years ago.

Craig LaChapelle:

Understood. Let me echo that back and make sure I got the key points from that. So scores still rank order risk, understood, but the sort of like for like are riskier than in past vintages because the key segment, I'll call them the high migrators, there are more of them than traditionally and that means... Traditionally they do perform worse in some cases more worse, significantly worse than others or slightly worse than the other segments, but not only are there more of them, they are performing worse. That segment is performing worse than traditionally. If I have that right, why is that high migrator segment performing worse than traditionally?

Josh Turnbull:

Yeah. I just want to underscore one point that you made there that I did not, which is so all that stuff about the performance has changed and they're performing worse than they have been historically, but to the point you made, we also in '21, 2022, we originated a ton of cards, the industry did. You made that point in your opening comments. We're closing in on $4 trillion in balances. There's more plastic out there and there are more dollars out there than there's ever been, so that pie is larger and then the slice of the pie that are those high migrators, that's a lot larger of a slice of the population than it has been historically. That contributes to the overall degradation and performance. Now that I've said that, I forgot your question, which was...

Craig LaChapelle:

That particular segment, the high migrators-

Josh Turnbull:

Why are they performing.

Craig LaChapelle:

... why are they performing worse than traditionally?

Josh Turnbull:

Yeah. I think you think about what happened in '21 versus pre-pandemic conditions, and so if my... Let's wind it back to 2018 or 2019, call it. If I'm someone whose score has gone up fairly significantly, 40 points, call it, over the course of a year, now maybe there was some windfall that I got, an inheritance or something or other that allowed me to pay off debt and that increased my score, but a lot of folks would do that through, "I'm adjusting my spending, I'm making changes to my life and paying down debts, and that's what's allowing me to gradually over time increase my score, but I'm making changes to my day-to-day financial behaviors to do that."

What we all experienced in '21 and '22 was there's a huge group of consumers who all of a sudden had this confluence of events, which was stimulus coming in, so cash that was being thrown at me from government programs or elsewhere, compounded with not having to pay on student loans or other obligations and not spending as much as I had on trips or commuting or what have you. There are all these external factors that now allow me to pay down debts, to do things that increase my credit score by that fairly dramatic amount, but I haven't necessarily changed my underlying spending habits, I haven't done some of the things that the same consumer would've done in 2018 which would've led to their score going up. Then when those things go away, when the cash stops coming in, when I start to have to pay my obligations again, when I start to travel again, I haven't made those changes. I'm not necessarily going to perform as well as that score would suggest historically.

Craig LaChapelle:

Great. Thank you. We mentioned this high migrator segment. What about the other segments? Do those segments perform similarly, worse or even better than in the '21, '22 vintages versus previous vintages?

Josh Turnbull:

Yeah. No, great question. Well, the folks... Again, my term, the Steady Eddies, almost identical. Someone who was a 720 five years ago, three years ago, today, the bad rate on a credit card is about the same. The people whose scores have started at 750 and they originated a 720, their performance? Almost identical to what it's been. The people who are new to credit card, their performances degraded about the same way that the high migrators have done.

Craig LaChapelle:

Okay.

Josh Turnbull:

It's worse now than it was historically. I think that leads us to, "Well, what do I... That's interesting. I understand what's going on now. I understand that there are different segments and it's certain segments who are kind of dragging the overall performance down. What do I do with that information?" That kind of led us to thinking about, "If you're originating credit cards, what I don't want to do is punish those people or punish myself by not originating for those people whose scores who are in that group where they have kind of the constant state score who are going to perform exactly as I would expect them to or the people who've come down a bit who would perform exactly as I'd expect them to. How do I get a little more surgical in my origination, in my account management approaches so that I'm able to grow, I'm able to extend credit, I'm able to help consumers meet their needs, and maybe I still want to originate the whole pie just to have a better understanding of risk and price it appropriately?"

Craig LaChapelle:

Let me make sure I understood and it didn't go in one ear and out the other. What we're saying is differentiated risk strategies to really identify folks who are in those riskier segments so we don't throw out the baby with the bathwater by just raising score thresholds arbitrarily.

Josh Turnbull:

Yeah.

Craig LaChapelle:

It's a more nuanced and sophisticated strategy where we're suggesting disaggregating the score risk bands by using different types of data. Is that what I heard?

Josh Turnbull:

Yeah. Let me make it real, Craig. If you are sitting on top of an origination policy at a bank and you've got a, I'm going to make this number up, a 2% bad rate and you say, "I don't want my bad rate to go above 2%. I want to maintain it where it's always been." What we've seen in the prime risk tier is that it's about a 35 point difference in terms of the performance, so if you want to get the same 2% bad rate that you've had the whole time, you would basically need to raise your score cutoff by 35 points in order to achieve that.

Now, when you do that you're going to knock out about 40% of the people with a prime credit score who are applying for a card with you, and that's a huge chunk of the population. That's what we try to do, is say, "What can you do to keep your credit score where it is, but what can you layer on top of that that allows you to be a little more surgical and still serve those people that you've always wanted to serve, still serve those people that are going to meet the bad rate goal that you have, but really hone in on those folks who are much more likely to have a worse than expected historically bad rate?"

Craig LaChapelle:

Well, what's our advice on how to do that?

Josh Turnbull:

Yeah. It's really... We work with customers day in, day out to understand how different data perform against their populations and can do that, but there are things that I think anyone... They're logical, that you'd want to key on, and looking at liquidity signals, so is a consumer paying more than they owe every month and how much more, but not just the absolute measure of that. How has that changed over time? If you see someone who a year ago they were paying off their cards in full, they were paying more than they owe on other obligations, and that's slowly dwindled or that's very quickly dwindled, that's a clear sign that someone may not be in as liquid a position as they were.

Craig LaChapelle:

For example, you're suggesting trends and excess payment.

Josh Turnbull:

Correct. Correct. Or the converse of that balance magnitude, so if my credit card balance or other kind of line of credit balances are slowly creeping up, that's a good signal that I'm relying on revolving debt a little more heavily than I have and might be in that category. Now, both of those things depending on which credit score you use may already be in the credit score, but then calling those out specifically and layering them on top, that's where we're able to see results that are much better than just whacking away at the 40% that we mentioned by a much more kind of... Just going to your odds chart and figuring out what you want to raise your score cutoff to.

Craig LaChapelle:

Yeah. What we're saying is it's really trended data, without being too salesy, and additional data outside of the traditional credit report to really identify those folks who are more likely to be bad than just a simple move of the risk score threshold.

Josh Turnbull:

Yeah. Yeah. I would say not necessarily outside of the traditional credit report, on the credit report but not in the traditional credit score maybe.

Craig LaChapelle:

Yeah.

Josh Turnbull:

Yeah.

Craig LaChapelle:

Are there any other examples besides trended excess payments that you'd like to highlight to give folks a sense of a key attribute that we're talking about?

Josh Turnbull:

I mean, I think those are the... There are a couple others that pop and, again, it really depends on the portfolio and some of your goals, but maybe more than attributes that are popping. The other thing to think about beyond origination is account management. We're having those conversations with a lot of customers now who said, "Okay, we either... We didn't do anything with our origination strategy over the past three years and so we've got a portfolio now where our more recent vintages are in this situation. What do we do from a line management or other type of account management strategy? Or yeah, we did make adjustments, but maybe not as much as we should have done. What do we do?" So a lot of conversations now, a lot of interest from issuers in how do you really take the same insights and apply that to line management strategies, so figuring out who do I want to withhold from credit line increases as people think about starting proactive credit line decrease programs up again, which we haven't really seen over the past couple of years, how does that play into it?

Craig LaChapelle:

Is the interest from our customers... Is it as significant in account management as in origination? Does it tend to be more origination or account management? I'm curious on that.

Josh Turnbull:

I would say there's interest in both, and it depends on your goals. Folks who were... Within the last couple of months is engaging with customers, just trying to get people's sentiment about 2024, people are optimistic or cautiously optimistic about the economic situation, about consumer performance, so maybe want to be a little more acquisitive than they have been in '23 but are trying to figure out, "What do I do in terms of my acquisition strategies, my underwriting strategies to make some of those tweaks so that I can feel good myself about getting back in or convince my risk counterparts that it's the right time to get back in a little more intentionally?" But at the same time, as you talked about in your opening comments, we've got people looking at these vintage curves that they're not great. I mean, they're not bad, but they certainly see the difference in performance with these recent vintages and people want to take active steps to manage that.

Craig LaChapelle:

Where are folks starting? Or maybe I'm going to phrase this a different way. Let's say somebody, an issuer either in charge of risk for acquisition or account management, line management, they're listening and they're saying, "Gosh, where do I start? How do I figure out how to use this data to, I would say, tweak my strategies in a more nuanced way?" What's your recommendation for one of their first steps?

Josh Turnbull:

Yeah. I think first step is let's sit down and talk about it and let's... We do these studies with customers or customers will supply the data so customers can do the studies on their own and really look at, very similar to what we did, look at pre-pandemic and more recent vintages and make sure that they can create a model that's working for both, so create it that works for the current or the recent vintages, validate that, though, against a pre-pandemic cohort to make sure that you've not just corrected for the past couple of years. But through that process you'll identify those attributes that pop, you'll identify what we did. I don't know if it's visible in the brief that we've got linked on here but can certainly share it with customers, is an example model that breaks people into I think six or seven nodes based on performance and based on certain attributes that get people into those nodes.

Craig LaChapelle:

Got it. Thank you. I got a note while we were going through this. If the link to the brief isn't there in the comments, it will be shortly. The team is working to put that up while we're concluding or shortly after we conclude. To that point, we have reached the midpoint of our target length. That's 25 minutes. I hope you-

Josh Turnbull:

I thought you were 30 minutes.

Craig LaChapelle:

Well, I said 20 to 30.

Josh Turnbull:

It's not the midpoint. Oh, okay. All right.

Craig LaChapelle:

So 25 is the midpoint. I really hope you folks enjoyed our insights. I know the team worked long and hard on this study and we've presented it in a few different forms and gotten a lot of engagement. I do think it is time to wrap. Again, I want to remind folks or encourage folks to look for our for the most part monthly Extra Credit podcasts, again on all the podcasts platforms.

Josh Turnbull:

Just to... Sorry to cut you off, but if this is of interest and if this is something that you're thinking about in your institution or just wanted to learn more about, let us know. We'd be happy to run through the study in more depth with anyone and talk about what we're seeing with these trends as well.

Craig LaChapelle:

That's great. Time to wrap. Thanks again, appreciate the time and the opportunity.

Josh Turnbull:

Thanks very much.