發現金融 (DFS) 2021 Q3 法說會逐字稿

Good afternoon. My name is Ashley, and I will be your conference operator today. At this time, I would like to welcome everyone to the Third Quarter 2021 Discover Financial Services Earnings Conference Call. (Operator Instructions) Thank you. And I will now turn the call over to Mr. Eric Wasserstrom, Head of Investor Relations. Please go ahead.

Thank you, Ashley, and good morning, everyone. Welcome to today's call. I'll begin on Slide 2 of our earnings presentation, which you can find in the financial section of our Investor Relations website, investorrelations.discover.com.

Our discussion today contains certain forward-looking statements that are subject to risks and uncertainties that may cause actual results to differ materially. Please refer to our notices regarding forward-looking statements that appear in our third quarter earnings press release and presentation.

Our call today will include remarks from our CEO, Roger Hochschild; and John Greene, our Chief Financial Officer. After we conclude our formal comments, there will be time for a question-and-answer session. (Operator Instructions)

Now it's my pleasure to turn the call over to Roger.

Thank you, Eric, and thanks to our listeners for joining today's call. We had another period of strong financial results in the third quarter with earnings of $1.1 billion after tax or $3.54 per share. In many respects, these results reflected the unique benefits of our integrated digital banking and payments model, which continues to be a source of significant competitive advantage by supporting our value proposition to consumers and merchants and differentiating our brand.

These advantages enabled our continued investment in account acquisition, technology and analytics while generating substantial capital. In an environment characterized by new entrants and intensifying competition, we believe the strengths of our model position us to accelerate our growth.

Underlying our results this quarter were 3 important advancements. The first was our return to year-over-year receivables growth, which is driven by our investment in acquisition and brand marketing and continued strong sales trends. Total sales were up 27% over 2019 levels with strong momentum across all categories. Even travel sales increased. And while they dropped a bit in August due to concerns related to the Delta variant, travel has steadily improved since then.

We also continue to see attractive opportunities for account acquisition and increased our marketing investment to take advantage of this. While the competitive environment has intensified, new accounts are now up 17% over 2019, reflecting the strength of our value proposition. This value proposition remains anchored in our industry-leading onshore customer service model, no annual fees and useful in transparent rewards.

While some of our peers had to reinvigorate their rewards offerings at substantial cost, our rewards costs were up only 6 basis points year-over-year and nearly all of this increase was driven by higher consumer spending as evidenced in our strong discount revenue. Given these dynamics, we will continue investing in new accounts as long as the environment supports profitable opportunities and a robust account growth and our expectations for modest improvement in payment rates supports our view of stronger receivables growth in 2022.

The second key trend was credit, which remained exceptionally strong. Our disciplined approach to credit management and favorable economic trends contributed to a record low net charge-off rate and continued low delinquencies. The delinquency outlook affirmed our expectations that losses will be below last year's levels for the full year and supported additional reserve releases during the quarter.

And third is the continued expansion of our payments business. PULSE saw a meaningful increase in debit volume with 9% growth year-over-year and a 26% increase over the third quarter of 2019, demonstrating both the impact of the recovery and an increase in debit use through the pandemic. Our diners business has also started to see some improvement from the global recovery with volume up 12% from the prior year. As the global economy recovers, we will continue to look for opportunities to expand our international reach.

In summary, our value proposition continues to be attractive, and our integrated digital banking and payments model supports profitable long-term customer relationships and is highly capital generative. I continue to feel very good about our prospects for future growth. I'll now ask John to discuss key aspects of our financial results in more detail.

Thank you Roger, and good morning, everyone. Once again, our results this quarter reflect strong execution and the continued economic recovery.

Looking at our financial summary results on Page 4, there are 3 key things I want to call out. First, our total revenue net of interest expense is up 8% from the prior year, excluding a $167 million unrealized loss due to market adjustments on our equity investments. Including this, revenue was up 2% for the quarter. Second is a continuation of very strong credit performance. Net charge-offs were down $343 million from the prior year, which supported a $165 million reserve release this quarter. Lastly, we continue investing for growth with increased marketing spend. Higher operating expenses in other areas were largely related to the economic recovery. I'll go over the details of our quarterly results and our full year outlook on the following slides.

Looking at loan growth on Slide 5. We saw the return to growth this quarter with ending loans up 1% over the prior year and up 2% sequentially. Card loans were the primary driver and were also up 1% year-over-year and 2% over the prior quarter. The year-over-year increase in card receivables was driven by strong sales volume and robust account acquisitions.

Sales growth continued to accelerate and was up 27% over the third quarter of 2019. Year-to-date, new accounts were up 27% from the prior year and up 17% over 2019 levels. The contribution from these factors was mostly offset by the ongoing high payment rates as household savings and cash flows remain elevated. The payment rate was approximately 500 basis points over pre-pandemic levels. We anticipate that the payment rate will moderate a bit as most federal COVID support programs have ended and consumer savings rates have started to decrease. That said, we expect payment rates to remain above historical levels through 2022.

Looking at our other lending products. Organic student loans increased 4% from the prior year with originations up 7% as most schools have returned to the normal in-person learning model. Personal loans decreased 4%, driven by high payment rates. Our underwriting criteria have returned to pre-pandemic levels, and we expect a return to growth in this product in future periods.

Moving to Slide 6. Net interest margin was 10.8%, up 61 basis points from the prior year and 12 basis points from the prior quarter. Compared to the prior quarter, the increase in net interest margin was primarily driven by lower interest charge-offs and lower funding costs. This was partially offset by a higher mix of promo rate balances.

Card loan yield was up 1 basis point sequentially as lower interest charge-offs were offset by the increased promotional balance mix. Yield on personal loans declined 15 basis points sequentially due to lower pricing. The margin continued to benefit from lower funding costs, primarily driven by maturities of higher rate CDs and an increased mix of lower rate savings and money market balances.

Average consumer deposits were flat year-over-year and declined 1% from the prior quarter. The quarter-over-quarter decline was largely driven by consumer CDs. We also saw a slight decline in savings and money market deposits as consumers continued to spend excess levels of liquidity.

We also continue to optimize our funding stack. Late in September, we executed our first ABS issuance since October 2019, consisting of $1.2 billion security with a 3-year fixed rate coupon of 58 basis points and a 5-year $600 million security with a fixed coupon of 103 basis points. These were our lowest ABS coupons ever and show good execution and timing by our treasury team.

Looking at revenue on Slide 7. Total noninterest income increased $90 million or 20% over the prior year, excluding the unrealized loss on equity investments. Net discount and interchange revenue was up $61 million or 26%, driven by strong sales volume. This was partially offset by increased rewards costs due to high sales in the 5% category, which were restaurants and PayPal, both this year and last. We continue to benefit from strong sales through our partnership with PayPal while restaurant sales were up 62% year-over-year as dining activity recovered. Loan fee income was up $21 million or 21%, primarily driven by lower late fee charge-offs and higher nonsufficient funds and cash advance fees..

Looking at Slide 8. Total operating expenses were up $185 million or 18% from the prior year. The details reflect our focus on investing for future growth while managing our operating costs. Employee compensation increased $12 million, driven by a higher bonus accrual in the current year. Excluding bonuses, employee compensation was down 3% from the prior year from lower headcount.

Marketing expense increased $70 million, supporting another quarter of strong new account growth. Other expense included a $50 million legal accrual. Professional fees were up $47 million, primarily due to higher recovery fees. Courts reopening, combined with strong credit and economic conditions, have driven an increase in recoveries and their associated fees. Year-to-date, recoveries were up 20% compared to the prior year. The benefits of these costs is reflected in lower credit losses.

Moving to Slide 9. The trend of sustained strong credit performance continued. Total net charge-offs were a record low at 1.46%, down 154 basis points year-over-year and 66 basis points sequentially. Total net charge-off dollars decreased $343 million from the prior year and were down $131 million quarter-over-quarter. Credit performance was strong across all products as evidenced by the net charge-off rates on card, private student loans and personal loans.

Moving to the allowance for credit losses on Slide 10. This quarter, we released $165 million from the reserves, and our reserve rate dropped 35 basis points to 7.7%. The reserve release reflects continued strong credit performance and a largely stable macroeconomic outlook. The impact of these was partially offset by a 2% increase in loans from the prior quarter. Our economic assumptions include an unemployment rate of approximately 5.5% by year-end and GDP growth of just over 6%. These assumptions were slightly less positive than those used in the second quarter but still reflect a strong economic outlook.

Looking at Slide 11. Our Common Equity Tier 1 for the period was 15.5%, well above our 10.5% target. We repurchased $815 million of common stock and, as we had previously announced, increased our dividend payable by 14% to $0.50 per share. These actions reflect our commitment to returning capital to our shareholders. On funding, we continue to make progress towards our goals of having deposits to be 70% to 80% of our funding mix. Deposits now make up 68% of total funding, up from 62% in the prior year.

Wrapping up on Slide 12. Our outlook for 2021 has not changed and reflects continued strong execution against our financial and strategic objectives. In summary, we remain well positioned for profitable growth from improving loan trends. Credit performance trends remained favorable, reflecting positive macroeconomic conditions and our approach to underwriting and credit management. Investments for growth have supported a significant increase in new accounts while we've contained operating expenses. Lastly, our integrated digital banking and payments model is highly capital generative, allowing us to invest for growth and return capital to shareholders. We look forward to providing our outlook for 2022 on our conference call in January.

With that, I'll turn the call back to our operator, Ashley, to open the line for Q&A.

(Operator Instructions) And we'll take our first question from John Pancari with Evercore ISI.

I want to see if you can give us a little more color on your payment rate expectations. I know you're expecting to remain elevated through 2022. But I just wanted to see if you're going to see signs of accounts that typically revolve, start to see some declines in payment rates there? And then I guess, how material of an inflection you think we can see through 2022? Or do you think that's not likely until '23?

Great. All right. Thanks for the question. So the payment rate has been persistently high. And what we did see in the month of September was that ticked down mildly. It did increase from July to August, which was, frankly, a bit of a surprise. But as I look at the data we're seeing here in terms of revolve and the forecasted trends, my expectation is that in the fourth quarter, it will continue to step down. Some of that has to do with government support programs ending in September, and some of it has to do with the holiday season.

And then as we look on the holiday season into 2022, I do expect that it will continue to step towards a normalized rate. But frankly, I don't think that will happen until 2023. So how do we think about the implications from that? Certainly, the payment rate is a bit of a headwind to growth. But what we've seen is really strong account acquisition and strong sales, which to date has helped offset some of that payment rate impact. So overall, we feel very, very comfortable that the balance of '21 and then '22 will have a bit of tailwind related to both payment rates declining and then strong execution from the new accounts and sales growth.

Got it. Okay. And then on the account acquisition front, I know you indicated that you expect marketing costs to be higher in the second half. So if you can give us a little bit more detail around the expectation there and how they could trend for the fourth quarter? And then does that imply that you could see some continued upside pressure into 2022 on marketing as you drive account acquisition in light of pressured payment rates?

Yes. Yes. So from a marketing standpoint, we spend the money as we see opportunity to drive profitable new accounts. And frankly, we've had a great quarter and a great year with that. The third quarter spend actually came in mildly lower than what we originally anticipated. And the guidance I had provided on the last call was that we would approximate the 2019 levels for marketing expense. I think it will be a little bit under that -- a bit under that, largely not because of opportunities, but basically kind of some process-oriented stuff in terms of account targeting. So we feel like the money we will spend in the fourth quarter will certainly generate positive new account growth. It will tick up from the third quarter certainly and provide us a good trajectory for 2022.

We'll take our next question from Ryan Nash with Goldman Sachs.

Roger, you talked about intense competition and the impact of new entrants. Can you maybe just expand on those points about what you're seeing competitively? How you think Discover is positioned for it? And where do you believe this is having the biggest impact on your business? And how are you responding for it?

Sure. Thanks for the question. I think my actual comment was good growth in the face of intense competition, and I think it really sums it up well. The competition in the card business is always intense. We were lucky enough to have a lapse in 2020, and so we saw extraordinarily good cost per account, but it's what we're used to facing. It will vary based on which issuers are refreshing their cards or have more of a desire for growth -- for growth than others. But I think our focus on a clear differentiated value proposition has resulted in continued strong generation of new accounts, and I'm excited about what I see.

In terms of the new entrants, we see that less in the core credit card space, but I would say very, very aggressive competition around personal loans. And that's an area where we focus on the long term, remain disciplined in our underwriting and what we're willing to invest in new accounts and are used to seeing competitors come and go.

Got it. And if I could ask a follow-up question to the question John asked prior. So, John, in terms of marketing, the level you'd be hitting in the fourth quarter as it steps up a little bit, should we think about that as a go-forward run rate? Or are there more investments that needed to be made? And then second, there was comments about BT's increasing and -- as it relates to marketing. Can you maybe just help us understand where are we today versus pre-pandemic levels? And what is the strategy to continue to increase balance transfer activity on a go-forward basis?

Sure. On the marketing spend, that will determined based on the opportunities we see. As we look at '22, we do see continued benign credit environment and frankly, a very strong opportunity to drive profitable new account growth. So as that opportunity continues to persist, we'll continue to spend marketing dollars. So what I would suggest is we concentrate here on 2021. And then in the January call, we'll talk about '22 in terms of the opportunities there. But I'll leave you with this point that the fourth quarter trajectory should help inform what we intend to spend in 2022.

In terms of balance transfers, we did see an impact from some of the increased balance transfers that we executed in early part of '21 in the third quarter in terms of a mild impacts to margin. We'll continue to take a look at that space and see what we can generate profitably. It has been a good source of account acquisition for us historically, and we'll continue to remain disciplined and put those prime revolver accounts on the balance transfer.

We'll take our next question from Don Fandetti with Wells Fargo.

I guess, Roger, a little bit of a longer-term question around your debit business. It seems like we're going to have more direct to bank payments in the United States over time. And there's a lot going on PayPal potentially buying Pinterest. Can you talk about how you see debit evolving over the long term? And are there any implications given that you own a network?

Sure. Great question. There are always, I would say, new payment schemes and methods out there. A lot of them though tend to rely on existing payments. So you think about when Apple launched their wallet -- even PayPal, the vast majority of their volume is processed through existing payment networks. And debit processing is incredibly complex and, quite frankly, very low margin and very efficient as you look at how sort of the 3 major debit networks, us, Mastercard and Visa operate.

And right now, given the interchange caps that a lot of banks have, the economics for them are relatively thin. So while we look at what the Fed might be proposing, what goes on in other markets, I don't see anything in the near or even medium term. It looks like it has the potential. And quite frankly, you're starting to see an increase in volume on the debit side, even a lot of the buy now pay later players are leveraging debit for their payments.

You don't see debit really going away per se and being replaced by direct-to-account payments in the U.S.?

No. The risk management of payments is quite complex in terms of both fraud and identity. There are a lot of processes and controls that the major networks have in place. There's a robust ecosystem, if you think about point-of-sale devices, merchant acquirers. And so to think that, that will change suddenly, I see as a low probability.

And we will take our next question from Betsy Graseck with Morgan Stanley.

Roger, just following up on that, is there an opportunity for you to take your debit network and debit capabilities and expand it into that new -- I wouldn't call it new, I guess, but into that revitalized interest in debit that we're seeing?

Yes. Great question, Betsy. Certainly, we have a unique set of network assets and can provide connectivity to merchants, whether through proxy card numbers or a series of other technologies and work closely with a number of fintechs around that. So that's at the core of some of what we do with Sezzle. Actually, that was the beginning of our relationship with Marqeta many, many years ago, and there are a lot of others that are either in the market now or that we're in discussions with. So I think that the pulse and the debit assets we have, combined with our signature network are an advantage and one that we look to monetize.

Okay. And then as we're thinking about the dance that's coming over the next couple of years between loan growth and credit normalization. Can you help us understand how you're thinking about managing that trajectory? Because what I heard earlier in the call is you've been getting more and more efficient at account acquisition, right? Part of your marketing spend coming in lower than expected or at least lower than consensus than we expected is a function of you doing something. I don't know if it's cloud or technology or what to get more efficient at account targeting and account acquisition.

So on the one hand, you've got that running in a very positive direction. You talked about the payment rates being a little bit of a headwind to monetizing that, but at the same time, we've got credit normalizing. So is there an opportunity for you to pull levers on marketing to generate some more loan growth? Or is it more likely to come from the personal loan side or student loan side? As credit is normalizing, how should we think about how you're managing that?

Yes. Betsy, I'll take this one. So there's a lot to that question. So let me just give a view in terms of what we're seeing as we closed out the third quarter. So as I mentioned, the payment rate is stepping down. And from August to September, we expect that to continue. That will help certainly drive loan growth. My sense is today that there's a relationship between payment rate and sales activity. So the persistently high payment rate, I think, has driven sales activity across the industry. Now we benefited, I would say, slightly disproportionately in terms of driving incremental sales, partly due to the acquisition point you mentioned. So there's 2 dynamics.

The third dynamic is credit normalizing. And my sense is that credit normalization will continue through '22 into '23. So today, where I said I'm positive on credit. So those factors in the aggregate, in addition to the point you mentioned in terms of account targeting, I think position us pretty well in '22 for positive loan growth. Not the specifics will come a little bit later in January when we give our view, but we're all positive on that front today.

Okay. And could you touch on what was driving that account acquisition being more efficient? You mentioned the process-oriented improvement. So what specifically are you talking about there?

So Betsy, it's Roger again. A lot of the enhancements that we're seeing are leveraging the advanced analytics. And so that's really helping both on the underwriting side with sort of swap-in, swap-outs as well as better targeting combined with investments we're making from beginning to end in the Marchex stack. And we think those are already serving us well, but there's also plenty of upside as we continue to focus on that area.

We'll take our next question from Bill Carcache with Wolfe Research.

Roger and John, I wanted to follow up on your conviction in more robust loan growth in 2022. If we were to get even just modest mid-single-digit spending growth next year, do you think the incremental tailwind from payment rate normalization could be enough to support double-digit loan growth? And just to clarify on this point, did you say earlier, John, that you think we could see payment rates get back to sort of normalize to maybe 2019 levels by the end of 2023? Is that a reasonable expectation?

So 2023 is the horizon that we're looking at for payment rate to normalize. And when that happens, we expect savings rate to return to normalized levels. So one view is continue to watch the savings rate and that should help inform payment rate. That's frankly one of the metrics we look at.

In terms of robust loan growth in 2022, let me just put it this way here because we're going to hold off until January on providing explicit details. But the new account originations have been important for growth. Sales activity has been important for growth. The payment rate, as I mentioned in the prepared remarks, has been a headwind, not completely unanticipated. We did see it mildly higher than what we had modeled, but we're still on track for delivering the loan growth that we talked about for '21, '22, it will be informed by all those factors and share it, Bill with more -- in more detail in January.

Understood. If I can follow up on a separate question on BNPL risk. I guess there's a group on one side that thinks that it's largely customers who can't qualify for credit and are just using the BNPL to turn their debit cards into credit cards. We've heard the CEO of Sezzle, which we know is one of your partners sort of speak to that point. And then there's this other group that is more concerned about the competitive threat posed by certain BNPL players that do longer-term installment lending and the risk that they'll eat into margins and pricing over time. Can you frame for us how you're thinking about the competitive threat? And now that you've had a little bit more time to study what is happening with the different BNPL players?

Sure. So I'll start with . We've had more time to study it. I'd say the market is not yet mature. And I think market clearing economics have yet to be established. Part of the challenge is there are many segments within Buy Now Pay Later from the people financing a multithousand-dollar purchase, which, by the way, they've done for years in terms of traditional sales finance versus sort of more spreading out payments on $60 worth of cosmetics. So those segments all have different characteristics.

Certainly, at the lower end, there are many customers who are either debit preferring or do not have access to significant amounts of credit. But I think you'll continue to see it evolve. You're starting to see some pressure from merchants who are unwilling to pay take rates above what they pay in card. So right now, we certainly see opportunity on the payment side, as I talked about earlier, leveraging our assets. Over time, we think there may be opportunity for us as an issuer and again, partially leveraging our unique model and our proprietary network. But right now, I would say we're not seeing any noticeable impact on revolving loans and believe that we are well positioned to respond if it does emerge.

We'll take our next question from Sanjay Sakhrani with KBW.

I wanted to follow up on some of Ryan's questions on competition. I guess, Roger, when you think about the competitive landscape on a go-forward basis. If you look at the post financial crisis period, a lot of that competition was transactor-oriented, like people going up transactor customers. Do you think this time it's going to be a little bit different given what unfolded last time? Or do you think that it's sort of going to be the same? I'm just curious as you think about how you're setting up competitively?

So Sanjay, I am lucky enough to be able to look back over multiple cycles of competition, and it has varied everything from sort of an intense focus across the board on prime revolvers to transactors. I think certainly, American Express will stay focused on that spend-based model. A lot of banks focus on those transactors because they are very profitable customers not so much in card, but in other segments of the business. The prime revolver space has historically been the most challenging in terms of what it requires from a value proposition and underwriting. For transactors, it's really about rolling out a risk rewards program. So we expect to see competition across the spectrum. But what impacts us most will be that in the prime revolver space.

And you don't see any of that really intensifying relative to what you anticipated, it's pretty consistent?

Yes. I would say it's intensified significantly since sort of the lull during the pandemic, but it's now getting back to, I'd say, more normalized levels.

Got it. And then, John, just 1 follow-up on NIM. I know we're not talking about next year's guidance until next quarter. But just broadly speaking, as we sit here today with your NIM pretty high levels relative to history. As the growth comes, do you envision that to be a tailwind for the NIM? Or does that start counteracting some of this because the BT rates increased? I'm just curious sort of how we should think about the progression going forward?

Yes. Thanks, Sanjay. So if you go back and look at NIM historically, you'll see that it was subjected to, frankly, higher funding costs from unsecured debt. What we've tried to do is shore up the liability side of the balance sheet. So targeting 70%, 80% based on -- or from deposits. And then the rest, a combination of secured and unsecured security offerings. So that will create stability in terms of the funding cost.

In terms of pluses and minuses to the revenue line, what we'll see is some BT-related impacts to NIM. You'll see credit, if it does normalize, a bit of credit normalization in terms of NIM. But you'll also see that sustained impact from funding I just talked about. So we're looking at NIM to be higher than what we've experienced historically. And we're going to compare '21, how we ended end the year in terms of the balance sheet position and see how that impacts '22. But overall, I would say the funding position has created more stability and an economic benefit to investors.

We'll take our next question from Mihir Bhatia with Bank of America.

Maybe to start, if we could just clarify a little bit on the marketing expense guidance, I guess, maybe for the fourth quarter? I understand you don't want to talk about 2022 yet. But just for the fourth quarter, when you talk about marketing expense relative to 2019 levels, are you talking on a quarterly basis or fourth quarters, we're looking at $235-ish million? I understand you may not want to give exact numbers, but should we be thinking on a quarterly basis or on a full year basis? Because just given the backup in the first 3 quarters of the year, that would be a pretty big increase versus on a quarterly basis? I just want to make sure we clarify that.

So thanks. So we -- when we referenced 2019, we're talking a full year basis. So then that would force us to do a little bit of math on the third quarter and fourth quarter. We came in lower in the third quarter than we anticipated. Fourth quarter, we expect it will increase from the $210 million. So to frankly, remove some ambiguity on this. So it will be somewhere in the range of $220 million to, call it, $280 million. And then I hesitate to give that level of specificity, but given the confusion, I just wanted to kind of put it away.

Sure. And we appreciate that. Just maybe just taking a step back on the credit side and just longer-term consumer credit performance. As you underwrite new accounts, can you just talk about what kind of through-the-cycle loss rate assumptions you're making? I guess what I'm really trying to understand is what is the normalized loss rate for your credit card portfolio as we think about longer-term economics?

Yes. So what I would confirm is we do use a through-cycle loss rate, so do not use current losses, but that's not something we disclose.

And we'll take our next question from Arren Cyganovich with Citi.

The pace of spending by your cardholders accelerated a little bit in 3Q relative to 2019. Is that consistent through the quarter? Or did it show any signs of rising or falling in terms of the cadence there?

We've seen ups and downs. In particular, we talked about a bit of a softening around travel, that occurred in August, and I think that was driven by a lot of what was going on with the pandemic. But we continue to see very robust sales volume even continuing into October.

Okay. And then the comments on the personal loan pricing and competitive pressures there, is that mainly from new entrants in the market? We're seeing a lot of different types of personal loan models coming to market in the past couple of years. Is that something that you would expect to continue as a competitive pressure going forward? Or is this more of a just marking down relative to where the existing book was?

A lot of it is coming into -- from new entrants. Personal loans are probably the easiest to fund outside of a bank and have the easiest servicing requirements. And so it's where a lot of the new entrants will go with new models. We are disciplined both on it as well as pricing and will not go below the targeted returns we want to hit. A lot of times, that will flush its way through when they either don't get the performance that they want to see or have other challenges. You do get some times people who are prioritizing growth over returns, but that tends not to last for too long.

We will take the next question from Mark DeVries with Barclays.

A question about what you're seeing in the student lending space and one of your competitors has noted some headwinds in the quarter. Just how is that shaping up?

So we feel pretty good about peak season. Volumes were up roughly 7% year-over-year. And we believe that we gained share, although we're still sort of processing. So clearly, it's -- we saw one major competitor step back last year. But we're very excited about that market. We underwrite it in a very disciplined way, and it's a great way to get a really good product in our brand in front of next generation of prime cardholders.

Okay. Got it. And then just a follow-up question on PULSE, as you know, you had some pretty strong volumes there. Roger, is there a point at which you start to get some meaningful scale benefits and the earnings contribution from that becomes more meaningful and grows faster than the volumes?

We would love to increase the percent of our earnings that comes out of the payments segment. I think one of the challenges, we also are working pretty hard to increase earnings from our banking segment. So we're not standing still there either. It's -- while we're excited about PULSE, it's also very competitive. We're battling for routing at the merchant level with Visa and Mastercard day in, day out. We've seen a lot of growth from expanding from traditional PIN debit to card not present. And so that's been particularly helpful. But yes, we're very focused on growing that business.

And we'll take our next question from Moshe Orenbuch with Credit Suisse.

Maybe just to come back to the new account acquisition. Is there -- particularly in card, is there a way to talk about either if there are any different channels that you're using? Or what's the nature of the consumer? Are there any differences in the type of consumer that you're seeing?

Not really, Moshe. I don't think we've seen anything dramatic since prior to the pandemic. Clearly, the continued migration to digital channels, direct mail becomes a smaller and smaller piece, but that's been a trend for many years now. So we continually are fine-tuning the digital channels. I guess, what we've seen is a greater ability to measure the results of top of funnel spend. As it shifts more digital, when that spend moves from TV to even video that's online, you can do a better job with tracking and attribution. And so we are repositioning how we spend, and that's one of the areas we're leveraging some of our advanced analytics.

Got you. And as you mentioned kind of using balance transfer as part of this, can you talk a little bit about anything that you can share with us in terms of whether they're all at 0, whether there some that are at yields above 0? And any kind of update on what you've seen in terms of retention of those balances?

Sure. So probably one of the bigger trends, there was a real collapse in balance transfer demand during the pandemic. Clearly, a lot of issuers we're pulling back on credit line increases. But I think consumers were much more focused on paying down their debt versus moving it around. So you saw a softening in demand for personal loans and balance transfers. That's starting to come back.

For us, a lot of our acquisition offers are at 0%. The portfolio offers tend to either be above 0 or have a fee that provides a pretty effective yield. And as John pointed out, for most of those, we see a very high return from the balance transfer itself given this funding environment as well as a good percent sticking, and that's something we model very carefully.

We'll take the next question from Dominick Gabriele with Oppenheimer.

I guess if we just think about comments around the economic outlook getting -- or the assumptions around your economic outlook becoming slightly worse from the third quarter -- from the second quarter to the third quarter, but then being incrementally positive into '22 credit trends. Can you just square those 2 pieces together?

Yes, happy to, Dominick. So when we modeled in the second quarter, frankly, the economy saw a higher level of GDP growth and employment returning to normalized levels sooner. And when we updated those, the GDP had come down, employment returning to normalized levels pushed a bit. So that was one of multiple inputs. We provide that detail because it's, frankly, the most transparent that people can take a look at to get deal with how we're thinking about the reserves.

Now the other items that are important in terms of our thinking here in the third quarter was, first, the portfolio performance, and it's been super strong. Second was the ending of many of the government support programs related to COVID. So most of those ended in August and September. What we'd like to do is see some seasoning of the impacts of those into the fourth quarter and perhaps even in the first quarter next year.

If what we expect is that the impact of that season will be very, very mild on the portfolio and the broad macros continue to look favorable, there is enough kind of points of reference that would lead us to begin to step the reserves back to day 1. But a lot of things need to happen. And we ended up where we were in the third quarter. We ran through multiple models, multiple scenarios. And we got comfortable with where we were to make sure the balance sheet was fairly stated. But we do see that the broad macros out into the future are positive. So we'll see how they sustain and make a call in the fourth quarter and in 2022.

Great. And if you all think about how the hierarchy of products that you have as well as across the consumer space for normalization on which products could normalize. Perhaps let's just, I don't know, to even maybe 2019 levels, NCO rates. Are there certain products you would expect to normalize either faster than other products? Or any color you can provide along the trajectory of one product versus the other, given what you're seeing in your customer base, that would be awesome?

Yes. That's a difficult thing to do in that it will be subject to the economic environment. So we know historically, right, if there's a period of tougher financial situation, the personal loans will be the ones that get the least priority in terms of payment. Credit cards tend to be high. The student loans given the long duration of those contracts, they don't seem to gyrate as much with the economic outputs plus there is so much of our portfolio is co-signed that it provides additional coverage there.

So you get to kind of personal loans subject to the economic environment. And then the card, my sense is the card has been prioritized in terms of the payment hierarchy because of what has happened in terms of cash in the digital environment. It's really difficult shop without a card these days. So I look for stability there.

And our final question will come from Bill Ryan with Seaport Research.

A couple of questions. One on the personal loans business. I believe you've been in the business since 2007. A lot of new entrants in the marketplace today, you kind of talked about it. But I remember, 2/3 of your book is roughly from the existing Discover credit card file, but the other 1/3 historically coming from the other places out of market or out of your existing file.

As far as the underwriting you're seeing from the new entrants, you kind of talked a little bit about it, but I was wondering if you could take a little bit more granular approach to it and talk about, are they making compromises relative to what you're willing to do on credit payment income, yield verification, all of the above? And second, just kind of going back to the bonus accruals. A lot of companies are stepping up bonuses as employee retention efforts. Is this something we should kind of view as transitory one time? Or might this transfer into a little bit higher compensation expense going into next year?

Yes. I'll start with the bonus accrual. We haven't made any changes to our bonus program. So to the extent there are changes in bonus accrual, it just reflects our stronger financial performance this year versus last year and then how that translates.

In terms of the personal loan space, it is very hard to get line of sight into the practices of these companies and there are many of them. We've seen them come and go over multiple cycles. In general, I would say they tend to have fewer manual or labor-intensive processes. So I'd be surprised if they did the extent of employment verification we do. And again, it might be appropriate if their average ticket is lower, they may have different return profiles. The one thing I would say, the vast majority lend to a much broader spectrum than we do. And so we focus on the prime segment, again, primarily debt consolidation, but it is hard to generalize across the many different new entrants that may have different models.

Excellent. Well, thank you all for joining us. And if you have any further questions, please reach out to the IR team. We'll be around all day to address them. Thanks again, and have a great day.

Thank you. And this does conclude today's program. Thank you for your participation. You may disconnect at any time.