Speaker 1Elevated household debt levels, high inflation, and economic uncertainty. As recession fears mount, analysts are digesting an array of data for signals to gauge the financial health of the American consumer. DVO1's In the Trotches of Structured Finance podcast dives into consumer data to explore macroeconomic trends, loan performance, and monetary policies. Here's your host, Vadim Verhoglad.
Speaker 2Hello, everyone, and welcome to another episode of the In the Trotches of Structured Finance podcast. Back with you after the summer break, and let's get right into it. I'm your host, Vadim Verhoglad, and in this publication, in this podcast, we'll be doing two things. First, we'll be taking some questions that we didn't get to answer on our latest consumer credit webinar, and after that, we'll be doing a brief update on credit performance and origination trends across the major three sectors that we track. As usual, please like, rate, subscribe, tell your friends. Tell everyone. Come listen to the pod, and with that, let's get into it. So starting with questions, the one first question we've gotten is, has the consumer gotten more prime with the continued rally in equity prices, and how do you manage the risk that markets don't go to the sky? The answer to that question is, the narrative of consumers being more prime over time isn't new and isn't really limited to the post-COVID period. It's essentially been getting more prime every single year since 2010, and it's not really driven by equities. Instead, it's more focused and more reflective of a longer-term shift. I think this is the best response to the common trope that we hear about, which in my mind is a completely meaningless term, because inflation really just means that there are a lot more healthy balance sheets and a lot more healthy consumer behavior out there. I think that's the best response to the common trope that we hear about ICO inflation, which in my mind is a completely meaningless term, because ICO inflation really just means that there are a lot more healthy balance sheets and a lot more healthy consumer behavior out there. I think this is the best response to the common trope that we hear about ICO inflation, which in my mind is a completely meaningless term, because ICO inflation really just means that there are a lot more healthy consumer behavior out there. happen much earlier than it has historically. Now, by several months, this will not fully last. Some of this will reverse because we're entering into the second half of the year. We're performing tends to deteriorate slightly. And there's kind of a seasonal catch-up loan factor that occurs that will shift these curves over time. We don't think it will shift them materially. So there is still a very positive sentiment. But, you know, not as great as just the charts alone tell you today. But even with that improvement, right, this is the vintages that saw some of those really high early impairment surges. So later vintages like 25, Q3, Q4, maybe the first quarter of 26 tend to be looking quite a bit better. So they won't even see that high ramp up early. Now, where you do still see bifurcation in performance is ROI. And ROI is really bifurcated in one specific category. That's great. More so than anything else. And we've talked about this still. The cows come home. Grade is the singular best attribute to track for performance in the consumer unsecured space if you have nothing else to track. Certainly, you're paid to do your homework, underwrite through all the variables that you have access to. But if you're tracking just one, it's great. It's great by a country mile. FICO doesn't come close. And what we've been seeing is, on the whole, if you take the 24 and 25 vintages, on the whole, their ROIs are kind of in line with maybe slightly below historical comps. And that's not a performance thing. That's really a prepayment thing. Because you have quite a substantially faster prepayment rates today, which takes out the best loans early. And because ROI is a percentage of original balance, it removes kind of interest proceeds on those loans. On an IRR basis, you wouldn't see the same flattening out. They'd still be quite a bit better than historical comps. But that's the case on ROI. But even there, even in that ROI, you have significantly different behavior between top grades. Which, even on an ROI basis, are several hundred bips above comps. So, in some cases, 150. In some cases, as high as 250 to 300 bips above. For a product that isn't the kind of ROIs or in the low double digits. So, 200 basis points to 300 basis points. Those are big, big differences. And that continues to be the case in top grades across the board. Middle and bottom, what we've seen is a really kind of abridged period of outperformance, especially in the bottom grades. So, that was really complicated. Concentrated in the late 23 to maybe early, mid-2024 vintages, where you saw really, really strong ROI growth. But we saw those numbers come down quite a bit. So, we talked earlier about the improvements in performance of the 25 vintage. That does not extend to bottom grade ROIs. Not really to middle grade ROIs either. The kind of outperformance proved to be maybe a year long, maybe five quarters. Because by, again, early 25, the ROIs were definitely coming down already. So, if you look at the bottom grade ROIs, we don't see any outperformance, but the bottom grade ROIs is a very small, small portion of originations. And as we talked about, what's happening in top grades is really kind of, because they're, again, they account for such a large volume, they account for more than 50% of originations. What happens in top grades is really, really important. Why we talk about ROI grade being so important is because if you look at ROI on a FICO basis, right, the story is quite a bit different. The story is that over the past two years, the ROI. Returns and lower FICO have significantly outperformed versus higher FICO significantly FICO has been is when you look at ROIs on a FICO basis and you come different ranges, what you typically have is the weaker FICO scores will under outperform for the first 12 to 18 months because again, how much higher coupons and then as you start to get into the heavy loss periods, you're going to start to see that ROI converge and in some cases kind of invert entirely. You saw lower FICO scores have higher ROIs and higher FICO scores, mainly for like a risk adjusted basis. It'll again, start high and converge over time towards zero, but historically maybe stayed above zero. Then in 22 and 23, completely inverted top grades, sorry, high FICO is definitely outperformed low FICO is and kind of just that was what was going on in those vintages. But by the first half of 2021, you saw not only that lower FICO scores. Significantly, you have significant higher ROIs there and they haven't hit their inflection point yet where the ROI start to converge and this has already been six or 10 months past the point where they should have started and we think the story there is that high FICO doesn't mean top grade and low FICO doesn't mean bottom grade. It is showing that lenders are finding ways to underwrite low FICO in an effective way because if they categorize as a top grade. We talked about what top grade performance is looking like and it's certainly not what's going on in in the FICO sphere now that's the performance end of it what's been more interesting has been originations so we talked about the 30% 40% of volumes the where that's coming from is that origination growth has been absolutely on fire not just this year this is a story that's that date backs dates back to 2024 let's remember consumer unsecured is a fairly established asset class this is not small. Small amount of money that's being lent. Right certainly doesn't compare to 1.2 trillion of credit card debt outstanding but consumer unsecured space does have several hundred billion dollars outstanding and we've seen a material and continued. Concentration of that lending in the fintech space and away from kind of the traditional banks that used to occupy it or some of the legacy lenders that have been. Had around in this space but why that matters the fact that this is established as class. Is that year-on-year origination growth at least in terms of the fintech lenders where TV on has most perfect data has exceeded 30% for nearly 3 years. 30% growth in consumer lending is kind of unheard of in today's environment in any kind of. Asset class and especially one as large as this what was major in over the past two months is originations crossed 7 billion dollars on a monthly basis in both June and July of 2026 7 billion eclipse the prior monthly origination record. By more than 1.2 billion dollars the prior high before June 26 was five and three quarters now the natural anecdote is going to be is that driven by two factors. One is a credit loosening to is it a coupon compression to the first one there's no evidence of credit loosening in fact actually credit standards tightened slightly over those same two months even as origination standards group. You still have top grades at the higher end of their kind of 48 to 54 55% range. That have been established over the past three years. The axis are 47 to 54 and in June and July was around 52 so very much at the high end of the range. Coupon compression has been going on we mentioned this earlier for the past 18 months but interestingly you didn't actually see that much compression over the past two months with those record numbers. If anything coupons are kind of flat or slightly of especially at those top and middle grades. So I mean it's not material it isn't. Like or the changes more than a couple of basis points but it's still kind of telling that it that really wasn't the driver the other interesting anecdote is and this is something we'll get into another podcast episode shortly. Is that those record volumes come against the backdrop of significant slow down in credit card originations previewing a future podcast episode credit card debt growth or credit card growth outstanding for the Federal Reserve's quarterly consumer credit report. So let's get into that. So the first thing is that the credit card growth grew at its slowest pace year on year since 2019. And its third slowest pace if you strip out kind of like one off examples like code in possibly as much as 20 years maybe going back to the pre GFC days. Because you know post GFC a lot of the decline was due to a lot of loans being written off because the charge offs and obviously you materially changing credit tightening but you didn't see this kind of low kind of growth and that's exactly what what what occurred over the second quarter. So you've got this significant growth against the backdrop of a slowdown in credit card debt and the only thing we can conclude from that honestly is people are really thinking about their credit card debt and they're really trying to think about continuing to improve their balance sheet continuing to have their finances in the best possible position now shifting to the yellow sector that's not you know actually let's do auto first because that's what I'd say is a red light the auto pictures still look strained in subprime universe but generally speaking outside of subprime universe. I don't think subprime it is kind of stabilizing 30 plus impairments or 30 plus delinquency rates are still very very high in subprime in some cases there are 700 several hundred basis points above 2019 comes which really is not a good comp because 2019 is what compares to 2008. So anytime you're talking about impairments in context to 08 something not great is going on that said the dynamic in prime is quite a bit different where you do have high delinquency rates but you're not approaching 2019 charge offs and you're certainly not approaching 2019 charge offs. Meanwhile taking a step back from the ABS side if you look at performance and bank balance sheets which again in auto those are really significant because bank balance sheets hold about one and a half times more auto loans and then then the ABS market and those are generally speaking been coming down over the past six quarters so peak for bank balance sheets was meant to Q3 Q4 2024 and they've really come down quite a bit since then. Now in that regard what I think is important to remember is that we don't think the subprime. auto story is actually just one of pure negativity. Instead, what we think is that an unappreciated part of what's happened in the subprime auto market dynamics writ large is a significant decline in overall auto lending that's happening in the BHPH space, meaning buy here, pay here. That channel accounted for about 14 or 15% of all originations in the middle of last decade, so 2016, 2015 time period. Today, it's nine. It doesn't seem like a lot, but keep in mind that only about 20, depending on how you cut it, either 15% or 25% of all auto originations are subprime and near prime. That's basically the entirety of the spaces, all near prime and subprime. That's what they lend to. Not only is it a 40% relative share of decline, it's also a significant part of the subprime lending landscape writ large. Where are those loans going to go? They're not going in. It's the OEM channel. That's mainly prime. It's going to go into finance cos. It could go into bank balance sheets. It could go into credit unions, but let's remember on bank balance sheets, about 88% of all auto bank balance sheet loans are in the bigger banks. I can't imagine, nor has it been said that any of the large banks have significantly increased their subprime penetration over the past nine years. That would have been a lot more market commentary. We think most of those loans, they're making their way into subprime auto APS. That market is holistically looking at a somewhat different mixture of collateral that originates than it used to, and a different dealer share than it was used to. There are definitely companies that became more aggressive with relationships that they pursued on the dealer side. Those relationships come with cost, and it comes with cost of performance. Now, what remains true in subprime, getting back to the actual performance, is that there's a different relationship between impairments and charge-offs. By the way, I'm not going to go into too much detail on that. I'm going to go into a little bit more detail on that. And by the way, as you guys know, for long-term listeners, we use the term impairment versus delinquencies because impairment captures every loan that isn't making their full payment, whether it's delinquent or it's modified. And we learned from COVID, and we learned in certain periods that impairments are where you see loans that are being worked out. Usually during natural disasters, this thing spikes and then comes down quickly. But if there's an aggressive kind of attempt to modify borrowers, to stave off defaults and all that, then you're going to see mod activity much harder. And I think that's a good way to look at it. I think that's a good way to look at it. And you don't want to kind of exclude those from delinquencies just because they're not reported there. So we use the term impairment. We think it's more comprehensive. And so impairments today, again, as we mentioned, are on 400 bps above pre-COVID levels, but charge-offs kind of in line. The reason is kind of two factors. First is you still have record high cure rates and pretty close to record high made payment rates. And that's saying, because there was a great quote by a subprime auto lender in the last earnings call where they said, "We have a lot of borrowers that are 60 days delinquent, but far less that are 120 days." Meaning they can't make every payment. They're managing, but they don't want to lose the car. This is subprime, so the credit score to this borrower is not as important as it would be to someone over 700. So it's very much a, "Hey, we're managing our payments here." On the flip side of that, those loans that do make it the charge-offs are delivering much, much higher losses by way of severities. And severities are rapidly approaching 60% in the subprime auto space. Which compares to pre-COVID levels of 45. It's a very big difference. It means recoveries are much, much lower. The other factor that's been really important in subprime auto has been tracking the difference in performance by LTV and PTI. FICO is obviously the bellwether for auto space. And it's the main attribute we use in pricing. It's a very, very significant attribute, but it's something that lenders understand here. This is subprime auto. They're used to lending to lower FICO. That's the business model. So FICO, they understand. I think LTV has been less understood. Not entirely misunderstood, of course, but it's less a factor in pricing, which has been to the detriment for investors. Because we found that the sweet spot with the fulcrum is right around 105 to 110 LTV. If you're below that threshold, then your impairment growth since COVID is something like 100 basis points. Fairly de minimis. And your charge-offs are about 30% to 40% lower. So you've been quite happy with your performance. Meanwhile, if you're above that number, especially above 115 on an LTV basis, you're not only looking at record high default delinquency rates and impairment rates, you're looking at them close to 900 basis points above pre-COVID levels, and you're also looking at record high charge-offs. Not just record high from COVID, but just record high, period. Eclipsing their 2022 levels, just really not in good shape. There's a similar, but slightly less dramatic, dynamic with regards to payment to income ratio, which actually, surprisingly, in auto is more predictive than income alone. And payment to income, again, you're really, your sweet spot in performance is below 11 PTI, payment to income, where you've seen some modest outperformance really over the past three to four years. Not everything in auto is a red light, though. I would say there's definitely some signs of optimism. Most notably, performance on loan age is nine to 14 months, which is kind of the first significant period that you see of performance after loans been securitized have gotten better. They've gotten better over the past two years, quite substantially. What's been more troubling is that hasn't translated as effectively as one for one into performance on loan ages 15 to 24 months, meaning that perhaps out of the gate, the loans are originated a little bit better, so you're not seeing as high delinquency rates. But once you get into that next tier, you're starting to see those borrowers come under strain again. So again, subprime auto, if consumer unsecured is green, subprime auto is kind of red with a kind of asterisk that it's not all bad. And against that, non-QM is squarely yellow. It's not consistent improvement as in consumer unsecured, and it's not deterioration as it is in auto. Non-QM is kind of in between. The market showed some significant performance strains from its peak of summer 22, post-COVID peak, to all the way through Q2, Q3, 25. There, you saw a kind of nonstop impairment growth, really significant deterioration in credit performance, and you started to see some defaults come through. Again, still very, very low defaults, but some. Since 2025, it's found some stability. Impairments today are lower than they were at their peak in 2025, and most of the past year has seen impairment growth be negative year on year. It's no longer negative. Now it's around zero. So the decline has probably subsided, but you're not seeing growth again, which is good. And that decline hasn't come with a decline. We're seeing a significant increase in defaults, because defaults are still measured, and CDRs are still measured in the well below one. They're in the 10 to 20 bps range. Some deals are a little bit higher. Some deals are a little lower. But in an EVO on benchmark, you're very, very low, 10 to 20 basis points, even though worst month. So charge-offs are rare. Losses are rare still, because a lot of the loans that do get charged off at substantial performance are getting charged off at zero severity, mainly because LTVs are still fairly low, and home price adjusted LTVs are lower. So that's a big difference. The usual story holds that weaker attributes have borne the brunt of the impairment increase. It's been below 700 FICO, and especially above LTV, and the worst combination, of course, are the two of them. What's been different over the past year is the diversions across doc types. The impairment gap between self-employed doc types, so your 1224 bank statements, your CPA, P&L loans, compared to full doc and DSCR is the widest level ever. And technically slightly wider, you know, right in April 2020. As a result of COVID, a lot of that is just functional, procedural stuff rather than fundamental performance. But aside from that, really the widest gap ever. There's a 350 basis point impairment gap between full doc DSCR loans and self-employed loans on an impairment basis. That gap was 80 basis points in 2024 at the lowest, and even around 50 or 60 in 22 and 23. So nearly 300 basis points gap entirely, which considering the whole sector's impairment ranges in the single digits, that's a very, very vast gap. What really makes non-QM yellow, and in some cases probably pointing to green in the future, is that there's been substantial positive performance improvements in the 2025 H1 vintage, so first half of 25. The first kind of 25-inch where you really see enough performance to talk about behavior. And what we've seen is that both impairments and the more kind of seriously like 60-90 day delinquencies are significantly outperforming 2023 and 2024 comps, when you isolate to what I would call non-underwriting economic non-underwriting factors, namely look at 25 purchase loans versus 23 purchase loans, or 25 cash outs versus 23 cash outs. Why I say that's non-underwriting driven, because the loan purpose isn't like, you're not expanding the credit box by lending to a different loan purpose type. Loan purposes is a function of rates and mortgage activity broadly, rather than an indication of credit standards. And against that, you're not expanding the credit box by lending to a different loan purpose type. And against that, And against that, you're not expanding the credit box by lending to a different loan purpose type. And against that, you actually see the performance of purchase loans or cash outs significantly underperforming, outperforming in 25, especially against what you would see on aggregate vintage. So meaning in aggregate, the 25 vintage looks a little closer to 23 and 24. But if you isolate to the each of the different types, it looks much, much better. And the difference is that when you combine the two, because of different concentrations, a little bit more cash outs, and actually underperformance in rate term refis, which in a small category, but isn't something we ever really thought we'd say, because that tends to be the best performing mortgage type. But in reality, the three kind of non-economic, like purchase loans cash outs definitely saying there's not performance in 25 and and with that we conclude today's podcast we've discussed a lot of things we've been discussing a lot of our performance reports the webinars again please rate it share it subscribe it rating us takes three seconds on your preferred podcast platform of choice sharing it takes one second with that i've been your host thank you very much for listening and until next time keep well
Speaker 1thank you for listening to dvo1's in the tranches of structured finance podcast stay up to date on the latest trends by subscribing to our podcast and our market research if you have any questions about this episode or want to learn how dvo1 data can help you understand what's going on in the market contact us at sales at dv01.co you can also learn more
PodcastTranscript AI