Capital One Financial Corporation (COF) Earnings Call Transcript & Summary

In this morning's slot of companies, very pleased Capital One -- to have Capital One Financial with us. Capital One's been a very, very long time supporter of this event. So, we appreciate the participation. From the company we have Jeff Norris, who is the Senior Vice President of Global Finance. Jeff, thanks for being with us, again.

Thanks, Jason. It's good to be here.

Obviously, I think most of the audience know Capital One, one of the largest credit card issuers, one of the largest auto issuers, both a branch-based and online depository gatherer really across the country. I guess given that, I suspect you have some unique insights into the consumer. So maybe just start big picture or your kind of normal impressions of the U.S. consumer and just the outlook against the backdrop of very low but potentially rising unemployment. Also they tell me elevated inflation and just still higher than normal payment rates.

Sure. So I'm not sure how unique our insights are, but our view of the consumer at the moment is that they continue to be a source of strength in an otherwise highly uncertain economic backdrop. We see labor markets, probably the most important driver of consumer health is still being quite strong. They've softened a little bit, but are still quite strong. I think there's some excess savings that consumers garnered through the pandemic that there's still a source of strength, although high inflation is kind of offsetting that and driving real wages down. So there's some pressure. But I think on the whole, consumers are still very much a source of strength. The debt burdens are low by historical standards and so forth. The one thing that we've got our eye on, though, in addition to continuing inflation, what that's doing to real wages, is the period we've just been through in the pandemic, which had unusually low and unsustainably low charge-offs and delinquencies. And it's our experience that on the other side of really good times like that, there's some catching up that happens on the other side. Just as like through the great recession, there was some elevated levels of charge-offs that sort of pull forward through the great recession and that had the effect of a couple of years in the aftermath of that a very benign credit, because it accelerated some charge-offs. We have an intuition that we can't really quantify that the opposite will happen as we leave the pandemic kind of in the rearview mirror to the event of widespread industry forbearance and stimulus payments and so forth that bridge consumers through that period of time. That could very well have the impact of delaying some charge-offs. And we might have to see some catching up on the other side. So with a couple of headwinds, we still find the consumers to be a relative source of strength.

I guess maybe we could just stick with that kind of credit quality comment. If we look 30-plus delinquencies are kind of back to pre-pandemic levels, losses still below, which something you alluded to. And you talked about monthly charge-offs returning to that 2019 level around the middle of this year. Can you maybe just talk to how much is that tied to normalization versus kind of deterioration? And just after you get to that point in midyear, does it stop or just continue to go up? And just how should we think about that trajectory?

So as we said, delinquencies and particularly the early formation of delinquencies in the domestic card business for us in the U.S. is already back to pre-pandemic or sort of 2019 levels on a monthly basis. And charge-offs aren't there yet, but we expect them to catch up fairly quickly. And we've said probably by the middle of this year. Normalization is kind of a funny term, right? As we've seen delinquencies and charge-offs rise, but remain below pre-pandemic or 2019 levels, that sort of increasing trend we and everybody in the industry is called normalization. But once you reach those 2019 levels, if things continue to worsen, you can't really say that's normalizing anymore, right? So I think we're about to drop that word from our vocabulary. We're assuming some continued pressure on delinquencies and charge-offs, driven by the persistent effects of inflation that we just talked about and just looking at the roll forward of our delinquency trends. So I think they're going to go a little bit higher. We haven't been specific about when or how high. But I expect that the worsening trends will continue for a little bit. We've assumed that in our most recent allowance build. We see the unemployment rate probably moving from its current levels to something above 5%. What we're assuming is that, that trend happens kind of by the end of 2023, early 2024. And all the factors I just talked about that are some emerging headwinds for consumers, I think, are what's in our thinking when we see some continued rising in losses. Now on the other hand, when we look at sort of early vintage performance, we're seeing a couple of encouraging trends. The first one is that if we look at early vintage performance of the accounts and balances we're originating today, it looks very similar to the early vintage performance from back in 2017, 2018, let's say, before the pandemic. So those vintage curves are sort of right on top of each other. And if we look at sort of monthly vintages each successive month, the early performance is right on top of the prior months. So there are some signs that some positive indicators. But against that backdrop, I think our macro view and just rolling forward current delinquency trends, we're assuming we're in store for a little bit more increase to go.

And I guess that's on the card front. I guess on the auto front, we've seen losses normalize quicker than expected there or not so quick or just quicker not than expected in card. And I guess, more recently, it's actually behaved consistent with seasonal patterns. Perhaps used car prices have been fairly elevated to the first 3 quarters -- first 3 months of this year, will pull back in April. Just kind of maybe -- just kind of delve into what your expectations are there.

So again, we don't really make forward-looking statements or give specific guidance about where we see charge-offs and delinquencies going. It's true that auto, in our experience, tends to move up faster and move down faster as cycles change. And so the fact that they -- the credit metrics in the auto business reached pre-pandemic levels sooner than card is kind of consistent with our expectations. I think in addition to all the things that we've been talking about in terms of the consumer, there's the overlay of used vehicle prices, which remain elevated compared to long-term historical levels but have been softening somewhat. We expect them to continue to soften and we actually assume a faster pace of deterioration in our underwriting choices, try and provide some buffer and some resilience even in the face of declining auto prices. But we expect that to continue and to continue to put some pressure on those metrics. But I think at the end of the day, we'll have to see where the used vehicle prices go and how that affects recovery values. We'll have to see about new card inventories and how those trends play out to have a better sense of where they're going. But again, similar to card, we've crossed out of the normalization. And now we're into sort of just seeing how the cycle plays out. But I'd point out that our card and auto businesses, we always assume things are going to get way worse than the actual forecast would suggest to build some resilience into our underwriting. So I think we're pretty well positioned even if the credit metrics continue to deteriorate for a little while.

Right. So I guess maybe delve into that, right? So you see higher unemployment. You talked about potential for lower used card prices kind of forecasting charges going up. And we saw you add over the last, I think, 3 quarters like almost $3 billion the allowance for loan loss -- allowance for loan losses. So clearly, kind of looking ahead, I think the card allowance is like 7.7%, so a full number. So from here, I guess, how should we think about the need to kind of build reserves? Or how do you think about the need to build reserves from here? I think it was $1.1 billion in the most recent quarter.

Right. Predicting the reserve build is a tricky business. Let me just talk a little bit about the elements that kind of go into our allowance for credit losses, calculation each quarter. You start with balanced growth, and that can be a tricky thing because we had pretty significant balance growth in the fourth quarter. But there's a nuance that you have to sort of work through, and that is that a lot of the balance build in the fourth quarter is seasonal. And those credit card balances tend to pay down before they have a chance to charge off, so they don't get a lot of allowance coverage in our methodology. Contrast that with the first quarter where growth appears on the surface to be essentially zero, but there's underlying growth. If you take a look at the year-over-year growth trajectory, it was 21% year-over-year balance growth. Unlike the fourth quarter, most of those balances are going to stick with us. And so they get a high allowance coverage because, as you know, credit card charge-offs are front-loaded, even more front loaded in the allowance methodology under CECL. So that growth in the first quarter gets a higher coverage ratio put against it. Then you look at our current economic forecast, and we've already said we assume unemployment rate is going to be above 5% by the end of the year and with a pretty conservative overlay for inflation and a recoveries effect that is kind of unique to Capital One. I'll come back to that in a minute. Just dropping the lost content from the first quarter of 2023 out of the 12 -- first 12 months of the allowance window, and adding the first quarter of 2024 drives an increase in allowance. And then you have the whole sort of notion that our economic forecast in any given quarter could get worse or better. What we said in the first quarter is that it pretty much stayed the same, but that assumption for losses to be higher in the first quarter, a year from now impacted our reserve build. Now I'll get to the recoveries impact. This is kind of a unique Capital One thing. In that, we work our own recoveries of charged-off debt and tend to do better. Historically, we can average as much as 2x the sort of recovery rate, on charged off debt because we worked our own recoveries as opposed to selling them. And it extends the period of time over which we're collecting against those accounts, but we tend to get more in our efforts. That's a good thing in most seasons. But at the moment, after going through the pandemic and the sort of unusually low charge-offs we saw for the last couple of years, the inventory of charged-off debt is lower. There's less for us to recover against and that has a little bit of an outsized impact on our allowance. And when you put all that into the mix, you see us sort of increasing allowance. Now it's also been true through history that our credit trends tend to move a quarter or 2 earlier than the industry. And so that -- I think that's part of the explanation for why our allowance looked like a larger build than others. And then the final thing I'll say is, I'm not sure how others allowance methodologies work, but ours sometimes can appear to be somewhat unique. One of the things is, I think, a lot of folks draw a fairly tight correlation between the level of unemployment rate and the level of credit card charge-off rate. Our models are actually particularly sensitive to the change in unemployment rate. One intuitive way to think about that is it's kind of the formation of new unemployment that is the bigger driver of card charge-offs. When you think about if somebody loses their job, it will take them 6 months to roll through the delinquency buckets and charge-off at 180 days. So if we work back from our assumption of the unemployment rate going from 3.5% to those something higher than 5% in like 2 quarters, that's a pretty rapid rise in unemployment and since our models are very sensitive to that change, I think that drives some of our build versus some others in the industry.

Helpful. And then you mentioned pretty strong card growth influencing reserves. So maybe we can kind of shift gears to that, I think we'll get April data next week. But if we look through March, I think you've had now 12 straight years of -- 12 straight months of 19% plus growth. Sales volumes been double digits. Just maybe talk to in terms of kind of what you're seeing, what are the drivers, any kind of shifting in what people are spending money on? And just how you think about growing loans into a rising unemployment rate backdrop?

Right. So, that's been a frequent topic of conversation over the last several months. On the one hand, we've got increasing charge-offs and delinquencies and larger than competitors sort of allowance build. On the other hand, we keep saying we're really happy with the growth opportunity and leaning into growth. And that, it's a little bit of cognitive and dissonance how can both of those things feature at the same time. The short answer is, in our experience, the balances that we're laying on now are going to have cash flows and revenue annuities that last a really long time and on a through-the-cycle basis, even in the context of short-term forecast for rising charge-offs and delinquencies. When we look at those things through the cycle, they still create a lot of resilience and a lot of attractive returns over the long term. And in our NPV models come back with very attractive net present value. So let me sort of unpack that a little bit. First, we're seeing pretty strong growth from existing clients as they build balance. These are clients that we know well and have a pretty good handle on credit risk. And the new originations, we're having the most traction in our marketing spend, generating new account relationships. But where we think the customers are at the riskier end of the credit spectrum, we originate them with lower lines. So that mitigates some of the credit risk. We're also seeing spend levels kind of slow down and moderate a little bit, which we think is a good thing. It's evidence of rational consumer behavior against an uncertain backdrop for the economy. If you look at new originations, or about -- or I'm sorry, spend growth, it was about 10%, and that's down from sort of high teens, 20% that we saw sort of a year ago. And if you look at it on a per active account basis, it's actually relatively flat and in some segments, down a little bit. So the entirety of our spend growth on the portfolio is driven by spending on new originations. And spend per account is moderating, which we think is a sign of a healthy consumer and a rationale consumer. So when you put it all together, we're pretty comfortable with the growth. And there's an additional thing that's kind of everything I just talked about is kind of true in our history as well as at the present day. We're also a little bit more comfortable with growth at the moment because we've really transformed our technology. We're operating essentially 100% in the cloud. That enables us to have big data streaming in real time and that enables us to drive in our core credit risk underwriting models, machine learning at scale. And that allows us to -- Rich talks about how we're generally leaning into growth, but trimming around the edges where we find subsegments of business that are performing less well than the rest of the book. With machine learning at scale, we can be a lot more precise and surgical about identifying the pockets of underperformance. So just to draw a very broad sort of hypothetical example, pre-tech transformation, we would have to sort of trim around the edges and I'm just making up the numbers, sort of $25 million, $30 million, $50 million subsegments. With the models and the technology we're employing today, we can sort of -- within a $50 million subsegment that's underperforming, we can generally identify the $10 million of it, that is the primary cause and shut that off and keep going with the other $40 million sort of. So our agility, I think, and precision is in a different place. And when we talk about our growth opportunities being enhanced by technology, that's kind of what we mean.

Got it. I guess on contrasting to the card growth, auto growth, we've seen it actually slow each or last 12 straight months. March, I think, was the slowest pace we've seen in 12 years. I know you've talked about pulling back, I would suspect there's a demand component. Just maybe just talk about why the pullback and then what would cause you to pivot and then maybe to get a little more aggressive.

Sure. There is a demand component, and that has to do with sort of vehicle inventories and the sort of relative confidence level of auto buyers. I think the bigger driver of our pullback by a wide margin, is our own view that industry margins had compressed in a way over the last couple of quarters, in a way that margin coverage for unexpected increases in losses is a key part of how we view resilience. So we need to always feel comfortable that the margin is sufficient to absorb losses that might actually exceed what we model. And we found a couple of quarters ago that credit unions as a whole, and if you sort of look at credit unions as a whole, that's a pretty significant share of prime originations in auto. And then a couple of large competitors that are in through near prime and sub-prime auto as well. We're not, I think, passing along the rising rate-driven funding cost increases into the pricing, and we saw a sort of compression of industry margins, that had us sort of move to the sidelines and pull back pretty sharply on originations. We always thought that, that would be kind of a temporal impact because even if you have a very simple sort of funds transfer pricing methodology, over time, it will work its way into the pricing. And we're seeing that happen. I think over the last few months, we've seen some softening of the -- well, I think softening is not the right word. We've seen some relief of the margin pressures that we've seen across the industry. And that might create some opportunities for us to sort of come back into the originations market. However, well, we've been on the side lines. We've also observed some rising delinquencies and credit losses in the deeper end of sub-prime, which is actually below where we play in the auto business. But in response to seeing those trends below our current origination box, we've actually tightened it to try and keep some distance between where we do originate and where we're seeing the sort of credit pressures. So any re-entry into the originations growth trajectory would be, I think, tempered somewhat by the fact that we'd be re-entering with a tighter credit box. And we'll have to see it if the margin trends sustain in a way that makes us comfortable to come back and something to watch.

You mentioned marketing earlier. And while down in the first quarter, 20% in mid seasonality, and it was also down a little bit year-over-year, are still a big number like $900 million. So we've seen good account growth, good sales growth, and I suspect, obviously, that's driven by marketing spend. As you think about, again, rising unemployment and the like, do you kind of temper marketing. Can you maybe just talk about just more in detail in terms of what you've been spending the money on, why you need to spend the money on in the future? And kind of just your thoughts around that in general.

Sure. So yes, marketing was down in the first quarter by about 20% on a linked quarter basis, which we attributed as sort of fairly typical seasonality. It was kind of flattish year-over-year. It was down like 2%.