發現金融 (DFS) 2018 Q3 法說會逐字稿

Good afternoon.

My name is Selissia, and I will be your conference operator today.

At this time, I would like to welcome everyone to the Third Quarter 2018 Discover Financial Services Earnings Conference Call.

(Operator Instructions) Thank you.

I will now turn the call over to Mr. Craig Streem, Head of Investor Relations.

Please go ahead.

Thank you, Selissia.

Welcome, everyone, to this afternoon'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 about the company's future financial performance and business prospects, which are subject to risks and uncertainties and speak only as of today.

Factors that could cause actual results to differ materially from these forward-looking statements are set forth within today's earnings press release, which was provided to the SEC in an 8-K report and in our 10-K and second quarter 10-Q, which are on our website and also on file with the SEC.

In the third quarter '18 earnings materials, we've provided information that compares and reconciles our non-GAAP financial measures with GAAP financial information, and we explain why these measures are useful to management and investors.

And of course, we urge you to review that information in conjunction with today's discussion.

As we announced on August 3, effective October 1, Roger Hochschild succeeded David Nelms as Chief Executive Officer of Discover.

David and Roger worked together in leading Discover for 20 years, with Roger serving as President and Chief Operating Officer since 2004.

Roger has been very active with the investment community in the past, so he should be well-known to many of you.

And the transition in leadership should be as seamless to all of you as it's been to our 16,000 colleagues at Discover.

Our call today will include remarks from Roger, covering third quarter highlights and developments, then Mark Graf, our Chief Financial Officer, will take you through the rest of the earnings presentation.

And as Selissia said, after Mark completes his comments, there'll be time for questions and answers.

(Operator Instructions) Now it's my pleasure to turn the call over to Roger.

Thanks, Craig, and thanks to our listeners for joining today's call.

I'm tremendously honored to succeed David as CEO.

This transition comes with Discover in a great position, generating solid growth and robust returns, and with the Discover brand stronger than ever.

Discover represents a unique combination of assets, built on a culture that places the highest value on serving our customers, whether by introducing unique features in our card business or innovations in payments.

I'm proud to have been part of the team for 20 years, and excited to have the opportunity to build upon that legacy as we pursue our aspiration to be consumers' most trusted financial partner.

Even with a change at the top, our priorities remain the same.

First, the Discover lend-centric business model, integrated with the benefits of our proprietary network, remains at the core.

We will continue to invest in the brand in a differentiated customer experience and in advanced analytics.

And, of course, we'll maintain our consistent, conservative approach to credit risk management.

This strategy has served us well and continues to produce outstanding results.

With that, let's turn to Slide 3 and review those results for the third quarter.

We earned $720 million after tax, or $2.05 per share this quarter, generating a return on equity of 26%.

This performance reflects the strength of the Discover business model, as well as consistent execution on our strategic priorities.

Looking at the fundamentals, total loans grew by 8% in the quarter, led by 9% growth in the card business.

This was above industry average as the Discover brand continues to resonate with our target customers.

New accounts were up 14% year-to-date and we're also seeing higher engagement, with both the number of active card members and spending per cardmember growing nicely.

Our private student loan business shows the strength of the Discover brand beyond the card business, setting us up for another year of record originations.

Student loan originations during the peak season were up 12% as a result of improved brand awareness and greater effectiveness in our marketing efforts.

In personal loans, it's clear the competitive intensity has not lessened, supported by the significant level of capital that has entered the market over the last few years.

Much of this competition is from unbranded players, which means that their principal means of differentiation and growth is on price or credit standards.

As we said, we will not chase growth that does not meet our return objectives, hence our decision to cut back on origination activity, and we continue to expect personal loan receivables at year-end to be flat to down from the prior year-end.

Focusing now on credit.

Overall performance is encouraging.

In our card portfolio, the seasoning of recent years' vintages, along with continued normalization in the back book, are in line with our expectations.

In addition, we continue to invest in enhancements not only in our underwriting processes, but also our servicing platform.

We are increasingly leveraging data, advanced analytics and machine learning to enhance the quality and efficiency of our servicing activities, whether in day-to-day customer service or how we collect past-due accounts.

This is paying off for us in a variety of ways.

It creates a better customer experience, it improves the scalability of servicing operations and it leads to better credit performance.

Turning to Payment Services.

The segment generated a healthy 22% increase in pretax income, driven by strong volume growth at PULSE and in our Network Partners business.

Before I turn the call over to Mark, I want to personally thank David for his leadership over the last 20 years.

I'm enormously proud of the business and the team we built together over that time, and I couldn't be more excited about Discover's prospects for the future.

I'll now ask Mark to discuss our financial results in more detail.

Thanks, Roger, and good afternoon, everyone.

I'll begin by addressing our summary financial results on Slide 4. Looking at key elements of the income statement, revenue growth of 8% this quarter was driven primarily by strong loan growth.

The increase in provision was largely consistent with ongoing supply-driven normalization in the consumer credit industry as well as the seasoning of our strong loan growth.

Operating expenses rose 7% year-over-year as a result of higher compensation expense as well as investments in support of both growth and new capabilities.

The effective tax rate this quarter was 25.5% versus our expectation for about 24%, as a result of adjustments to reserves for certain tax matters, which cost us about $0.02 per share.

We continue to expect that our effective tax rate for the full year will be about 24%.

Turning to Slide 5. Total loans increased 8% over the prior year, led by 9% growth in credit card receivables.

Growth in standard merchandise balances drove much of this increase, spurred by continued strong sales growth.

Promotional balances were a modest contributor to growth this quarter, coming in somewhat above our expectations.

We are seeing higher-than-expected response rates, a FICO profile north of the portfolio average and strong economics from these channels.

As a result, we've made a decision to maintain the mix of promotional balances around their current levels as opposed to our prior expectation for them to decline.

Obviously, if we see a deterioration in these economics, we'll revisit this approach.

Looking at our other primary lending products.

Year-over-year growth in personal loans slowed to 2% and, as Roger said, balances are expected to be flat to slightly down on a year-over-year basis by year-end, reflecting the continued impact of credit tightening.

Private student loan balances rose 2% in aggregate, but our organic portfolio increased 10% year-over-year.

Moving to the results from our payments segment.

On the right-hand side of Slide 5, you can see that proprietary volume rose 9% year-over-year.

In Payment Services, PULSE volume growth was strong again this quarter, with 14% higher volume compared to the prior year, driven by both new issuers as well as incremental volume from existing issuers.

Diners Club volume rose 5% from the prior year on strength from newer franchises, and Network Partners volume increased 34%, driven by AribaPay pay volume.

Moving to revenue on Slide 6. Net interest income increased $173 million, or 8% from a year ago, driven by higher loan balances.

Total noninterest income grew $26 million, primarily from a 12% increase in discount and interchange revenue.

Sales volume also grew by 12%, led by growth in active cardmembers and, to a lesser extent, higher spend per account.

Adjusted for the number of processing days, sales growth was 9%.

Our rewards rate for the quarter was 131 basis points, up 1 basis point year-over-year.

As shown on Slide 7, our net interest margin was 10.28% for the quarter, flat compared to the prior year and up 7 basis points sequentially.

Relative to the third quarter of last year, the net benefit of a higher prime rate is offset by an increase in promotional balances and higher interest charge-offs.

Relative to the second quarter, the net benefit of a higher prime rate and modest sequential improvement in interest charge-offs was partially offset by higher deposit costs in both brokered and direct-to-consumer.

As I mentioned a moment ago, we are seeing better-than-expected production levels, credit quality and economics in our promotional channels.

In addition, we are seeing month-over-month decreases in the portfolio revolve rate as transactor engagement continues to pick up.

For these reasons, we now expect NIM for the full year to come in modestly below our prior guidance of 10.3% to 10.4%.

Total loan yield increased 30 basis points from a year ago to 12.45%, as a result of a 23 basis point increase in card yield and a 63 basis point increase in private student loan yield.

Prime rate increases and a modest increase in the revolve rate led card yield higher, partially offset by a higher mix of promotional balances and higher charge-offs.

In student loans, where about 2/3 of the portfolio has a variable rate, higher short-term interest rates drove the 63 basis point increase in yield.

On the liabilities side of the balance sheet, consumer deposits grew 12% as we continued to shift our funding mix toward more stable and cost-effective sources.

Consumer deposit rates rose during the quarter, increasing 17 basis points sequentially and 51 basis points year-over-year.

Deposit betas have increased, although cumulative betas continue to be better than historic norms and our expectations.

Turning to Slide 8, operating expenses rose $67 million from the prior year.

Employee compensation and benefits was higher, the result of higher average salaries as well as increased volume-related and technology headcount to support business growth.

Marketing expenses were up as a result of greater investment in generating new loans, and these investments are yielding great returns.

For example, in credit card, while total marketing dollars are up, acquisition costs per account are down on a year-to-date basis, and we expect that trend to continue through the fourth quarter.

As Roger talked about earlier, we continue to invest in infrastructure and analytic capabilities, which led to an increase in information processing costs.

I'll now discuss credit results on Slide 9. Total net charge-offs rose 34 basis points from the prior year, but are down 14 basis points sequentially, in line with normal seasonality.

Supply-driven credit normalization, along with the seasoning of loan growth from the past few years, continue to be the primary drivers of the year-over-year increase in charge-offs.

Credit card net charge-offs rose 34 basis points year-over-year.

From a sequential perspective, this quarter represented the fourth consecutive quarter of slowing year-over-year increases in card charge-offs.

The credit card 30-plus delinquency rate was up 18 basis points year-over-year and 16 basis points sequentially.

Credit performance in the card business has been very solid.

We've also mitigated a degree of future exposure by selectively closing long-term inactive accounts, eliminating nearly $30 billion in contingent risk since the beginning of 2017.

Private student loan net charge-offs fell 18 basis points year-over-year, but were up 4 basis points sequentially.

Personal loan net charge-offs increased 90 basis points from the prior year and 12 basis points sequentially.

The 30-plus delinquency rate was up 30 basis points year-over-year and 15 basis points sequentially.

There are a number of factors driving personal loan credit performance.

First is the seasoning and normalization phenomena that is also driving losses in card.

This trend continues to develop as we expected.

Second is the loss experience in certain origination channels we started discussing with you in the third quarter of last year.

We cut off all origination activities in these subsegments as soon as the problems became apparent, but the 24 month seasoning curve in personal loans means that the related losses will peak in the next few quarters.

This will drive dollars of charge-off higher until the pig is through the python.

Third, I would point out that slowing loan growth will also contribute to higher charge-off rates, because of the denominator effect.

And finally, there is a seasonality aspect to our personal loan charge-offs in that we typically see sequential increases in both the fourth and first quarters.

Putting all this together, we would expect to see about a 60, 6-0, basis point increase sequentially in the personal loan charge-off rate in the fourth quarter and a full year 2019 charge-off rate of around 5%.

Looking at capital on Slide 10.

Our common equity Tier 1 ratio decreased 20 basis points sequentially as loan balances grew.

Our payout ratio for the last 12 months was 109%.

To sum up the quarter on Slide 11, we generated 8% total loan growth and a 26% return on equity.

Our consumer deposit business also posted robust growth to 12%, while deposit rates increased 51 basis points year-over-year.

With respect to credit, our total company charge-off rate, below 3%, reflects very positive underlying trends in card and student loan.

Finally, we're continuing to execute on our capital plan, with strong loan growth and capital returns helping to bring our capital ratio closer to target levels.

In conclusion, we're pleased with our performance this quarter, characterized by strong receivables growth, good credit performance and very strong returns.

That concludes our formal remarks.

So now I'll turn the call back to our operator, Selissia, to open the line for Q&A.

(Operator Instructions) We will take our first question from Ken Bruce with Bank of America Merrill Lynch.

I guess, my first question relates to the NIM.

It's been fairly stable, all things being equal and, obviously, you're kind of ratcheting down the guidance.

Mark, can you just roll through maybe some of the puts and takes to what is impacting that outlook for the rest of the year, please?

Sure, absolutely.

Ken, I mean, I guess I'll just tackle it head on and say it's a conscious decision on our part.

When we had our call last quarter, we were looking at a forecast that was the lower end of our guidance range and we were planning on taking down our level of promotional activity as a percentage of our total balances.

And we made some changes, and we saw incredibly strong response rates and economics start flowing through.

So we made a conscious decision that we could either manage to a NIM rate or we could manage to the long-term shareholder value accretion by booking business that was demonstrating extremely strong return profiles.

We made the decision that we would sacrifice the NIM in the near term to book that good long-term business.

So that's the underlying thought process behind it all, I would say.

If you think about it in context, quarter-over-quarter, market rates got you about a 9 basis point good guy for NIM.

The funding rate took away about 5 basis points.

Credit, with charge-offs of accrued interest being lower, added back about 2 basis points, and then the receivables rate was about a point or 2 with promo rates being a little bit higher.

So net-net, that gets you your 7 basis points sequentially in NIM.

And we continue to be positioned to be modestly asset sensitive.

The bigger driver of NIM trajectories going forward is going to more likely be that mix of promo balances that we now expect to be flat.

So we would continue to expect some NIM accretion in the fourth quarter, but it will not get us all the way to that prior guidance range.

And firstly, Roger, welcome to the call, and congratulations on your elevated role.

I would -- I'm interested in kind of a bigger picture question.

Obviously, you've mentioned in the prepared remarks the lend-centric model that Discover's been well-known for.

Obviously, you've got a payments business, both within the Discover card in a way that is more payments-related and obviously is geared towards transactors.

Can you maybe just kind of think -- or help us think through what -- how you want to be kind of targeting these different markets?

We have just juxtaposed it with like an American Express, who's very transaction-oriented and has been moving into lending.

They think that they have an easy way to pick up balances.

You're more of a lend-centric business and maybe you're trying to pick up transactors.

Could you maybe just help us think through that strategic decision that you make?

Yes.

Thanks, Ken.

I wouldn't confuse our payments business with a focus on transactors.

And in fact, the biggest part of our payments profitability comes from PULSE, which is a debit business.

So payments business is really working with third parties, whether they are networks around the world or Diners Club franchisees or the banks that issue over PULSE.

In terms of our card-issuing business, we remain very focused on our lend-centric model, that's something we've had consistently, and so while, again, we're seeing great transaction growth and net income growth in Payments, a lot of that is actually driven by the debit market.

Okay.

Maybe I can just -- I realize I only have one follow up, but maybe, within the Discover card though, you're clearly trying to bring in transactors just through the -- effectively the rewards that you -- the cash back reward, just more driven towards transactors.

But you need to convert those to borrowers to really drive the earnings model, I guess, that's really what I was going with.

Yes.

I mean, we've had rewards since we invented rewards over 30 years ago.

So I wouldn't point to that as a change from our lend-centric model.

Our next question comes from the line of Sanjay Sakhrani with KBW.

Congratulations, Roger.

So first question, Mark, for you is on the personal loan provisions.

Is it fair to assume most of the provisioning related to that pig is behind us?

Or should we expect more pressure as it relates to the losses there?

Yes.

If you look, our reserve rate was flat quarter-over-quarter, Sanjay.

So what you saw going on over the prior year since we started talking about this, you saw big increases in our reserve rate for the personal loan business.

Now what you're seeing is those things for which we built the reserve starting to migrate through to charge-off.

So you'll still have that normalization phenomena we talked about impacting personal loans and you'll still have, obviously, whatever happens to the economy between now and then, but that vintage of loans that we've been talking about that was driving reserves is now just flowing through to charge-off.

Okay.

Great.

My follow-up question, I guess, is for Roger.

It's sort of like the second question previously.

Maybe you could just talk about how your approach might be different than David's was.

I know you guys have seen generally eye-to-eye.

But maybe you could just talk about any changes, even if nuanced, that you'd propose?

Yes.

Thanks, Sanjay.

For 14 years as President, I ran all the operating areas of Discover.

So got to drive a lot of the growth, whether it's in the payment segment, whether it's around our technology strategy or the new features we launched on the card side.

So I think really it's continuing the path we're on.

We have a great set of assets, we've been very successful in diversifying our Direct Banking business from beyond card, and we're incredibly excited about the growth opportunities we see in the payments business as well.

I'd point to some of the partnerships we're striking with networks around the world.

So again, I would not expect any big changes in strategy, but continued focus around execution and driving growth.

Your next question comes from the line of Ryan Nash with Goldman Sachs.

Roger, congrats again.

Maybe I could start with the promotional activity, Mark, that you talked about, keeping it the same level.

Maybe just expand a little bit what drove the decision to keep it at current levels, particularly what changed over the course of this quarter relative to your view last quarter, to run it down?

Are you seeing it more just turning into revolve?

What drove the changes over the short term?

So a couple of things.

Great question, Ryan.

A couple of things drove the change to the outlook.

First of all, we did reduce the marketing-eligible populations.

We expected that would result in a response rate that would be lower.

What we actually saw is when we reduced those marketing-eligible populations and refocused on the remaining populations, we actually saw the response rates go up meaningfully.

So the return on the invested dollars ended up being higher than we expected to see it.

Number two, I would say, is the engagement we're getting out of these folks.

So these aren't just balance transfers.

We're talking sales promo as well, so new account promotional activity as well, right?

And we're seeing very strong engagement -- actually an uptick in engagement across a number of those different subsegments relative to what we had seen previously, so that was a piece of it, and then I had mentioned the FICO score has actually migrated north.

It's actually -- the FICO score that we're getting in these accounts right now is actually well north of the portfolio FICO for the card accounts, and the marginal ROEs actually ticked up meaningfully from what we'd been seeing.

So when you put all that in a blender and hit go, we really sat down and said, okay, again, we can manage to a NIM rate, because we're not insensitive to when we give you guys guidance, obviously.

By the same token, managing to the NIM rate would have been walking away from fabulous business that's going to be really accretive to shareholder value over time.

So we made what we thought was the right decision.

So that was kind of the facts and the thought process.

Got it.

And maybe just a question for Roger.

Roger, we're hearing a lot of your competitors talking about accelerating marketing -- American Express, Capital One talked about it on their call last night, and I just wanted to get your view on the competitive intensity in the card market.

You guys have talked about loan growth slowing, pulling back from various channels and -- just wanted to get your sense for the overall competitive landscape, given that, obviously, given where the stocks are trading on P/E, the market is obviously concerned that there could be a downturn at some point in the not so distant future, so any color you have there will be appreciated.

Yes.

I think first, the comments on loan growth slowing and pulling back from certain channels are really about the personal loan business.

New accounts are up strongly on the card side.

We're seeing great productivity on our new account investments, so even with significantly more new accounts, cost per account is down.

My view is the card business is always competitive.

There might have been a lull in '08, '09, but the key is putting a great value proposition out there.

We continue to differentiate in terms of customer experience and the functionality we provide.

So again, we feel very good about the growth we're putting up and we're continuing to invest in driving that growth.

Your next question comes from the line of Don Fandetti with Wells Fargo.

Yes, I had a question about card delinquency rate, 30-plus day.

I was curious, it's been relatively stable on a year-over-year basis, arguably has been a tiny bit elevated more recently.

I was just curious how you see that trend playing out.

Do we stay steady?

Do we have a bias to the upside because of the growth math?

And then also, Mark, on the reserve build.

You've been sort of hanging out this $100 million range.

How do you think about that on a going-forward basis, just directionally?

So in terms of the delinquency rate and the charge-offs, it was the fourth consecutive quarter, Don, where we've seen a slowdown in the rate of increase in charge-off formation.

So I think that migration has stopped delinquencies.

I agree, they came down from, call it, 37 middle of last year, plus or minus -- the increase, 37 basis points -- to more or less flat at about 16, 18 basis points, something like that over the course of the last few quarters.

So they are down meaningfully.

I think one of the big drivers is the growth math piece, right?

I mean, as Roger mentioned, we had very strong growth in new accounts.

We continue to target new card members selectively in this environment.

We think it's -- the window for growth remains open with the current macro backdrop.

So if you think about the reserve piece as opposed to the delinquency piece for a second, roughly, call it, 35%, 40% of what you see for reserving is new account related, with the remainder being more seasoned back book related, right?

So in that context, you do have a big chunk of it that's being driven by the growth math to use my competitor's terminology.

So -- but we think it's manageable.

We feel very good about where it is, and we love the economics.

In terms of the overall reserve build, at roughly $99 million, the bulk of that was related to card and really reflects what we see coming at us in that seasoning and in those pieces of the puzzle.

It's down -- the card reserve build is down, call it, $20 million over the same quarter last year, despite generating a bigger amount of new accounts.

So that probably tells you something about the quality of the accounts and what we expect to see going on there.

So it continues to feel like we're booking really high-end, high-quality business right now, and I feel good about what we're putting on the books.

Reserve rate going forward is also going to be impacted by macroeconomic factors and everything else, so I don't want to prognosticate, other than to say, we like what we're putting on the books.

Your next question comes from the line of Mark DeVries with Barclays.

I have a question about loan growth for Mark.

Is the implication in the commentary around NIM and the decision to do more of the promotional balances than you originally thought, is part of the implication that maybe loan growth comes in at the higher end or above the higher end of expectations?

So I think if you think about it in terms of our guidance range for the full year for total loans of 7% to 9%, I certainly think you would come in north of the midpoint of that range, as we sit here right now.

I think that's a reasonable expectation.

I think one of the things we're encouraged by, again, is not just of the level of growth we're seeing, but again, I'll reemphasize the quality of that growth.

The FICO profile that's coming in from this incremental business that we have elected not to run off because of what we've seen since making some of those changes -- or not to run down since making some of those changes I alluded to earlier.

Feel good not just about the level of growth, but of the quality of the growth we're finding right now.

And to Roger's point earlier as well, cost-effective acquisition of these accounts right now.

Marketing expenses are up, but costs per account are down.

Okay, great.

And then for my follow up.

I think during the quarter, we heard a lot of incremental commentary from other players in the personal loan space that they were seeing credit, at least from loans purchased from some of the marketplace lenders, that was coming in worse than expected; economics that were looking worse.

Did you see anything during the quarter that made you incrementally more concerned and maybe has you close to the guidance of being down year-over-year versus potentially flat?

No.

I mean, I think, we haven't really seen anything new over the course of the last -- since we spoke to you last in terms of the development there.

We did start pulling back a full year ago.

It was the third quarter of last year when we really started talking to you about what we started seeing in some of these segments.

So I think in consumer lending, you've got to have very -- you got to have continuous monitoring and you've got to have toll gates set up for if you see certain things start to happen, they need to drive actions.

So I think we probably saw some of that start to emerge early and took action relatively more quickly.

But there's no question, as we've said before, just the pile-on effect of every day you open your financial press and there's somebody else talking about entering the personal loan space.

We've said all along, there's a time when you want to get on the accelerator in this business and a time when it makes more sense to -- there's still a window of opportunity for personal loan growth.

It's just nowhere near as wide as it once was and you've got to be more selective.

Your next question comes from the line of Moshe Orenbuch with Credit Suisse.

I was sort of wondering, given your comments about continued strong balanced growth, but improving quality, could you kind of put that into context of -- you -- I mean this was sort of addressed partially in a previous question, but when you think about the reserve needs for that portfolio, again keeping the economy constant as we go into '19, I mean, how should we think about that for the card business?

So one of the big swing factors in an actuarial portfolio, as you know, Moshe, is the macro factors.

So if you control for those and say, hold all the macro factors exactly constant -- right?

-- it would really be more a function of just how much of the book is going through seasoning at that point in time, so it would be: what percentage of the book is new accounts?

would be the driver, and then the second piece would be: what is going on with consumer leverage out there, more broadly?

So what's happening to loss severities given a default?

Those would really become the 2 swing factors that would drive that.

We have seen consumers kind of reapproach that normalized level of leverage.

It's kind of what we believe is part of what's driving this normalization and losses beginning to slow, right?

The normalization in charge-offs beginning to approach some type of a normalized rate.

So I would say, based on what we see right now, we don't see overleveraging being something that I would say driving it right now, but that could change tomorrow, just to be clear, but we don't see it right now.

So I would say, going forward, those reserves are going to be far more affected by what percentage of the book is new accounts at that point in time would really be the swing factor, if you neuter all the macroeconomic effects.

And maybe a second question.

Can -- I mean, I'm very impressed with the declining acquisition costs, and you've alluded to it before, but it seems like it's a more significant factor now.

And I guess I'm trying to understand, just given that this is a continued competitive industry and you do see large upfront bonuses that many issuers are offering, what would you attribute that to?

And anything you can kind of tell us about the persistence of your revenue streams post promo pricing?

And how that's faring in this environment?

Yes.

So in terms of new account pricing and promotions, we've tried to steer away from those kind of big ticket $300, $500.

We do double the Cashback Bonus for the first year and have found that to work very, very well, and don't see a significant decline in spending once that promotional period expires, because it's really driving long-term usage over an entire year.

A lot of the benefits in terms of costs per account come from product differentiation, customer experience differentiation and just the strength of the brand.

So by leveraging our proprietary network, the advertising we have is working well, and then some of the investments we've been making around analytics.

And so optimizing whether it's direct mail, whether it's the digital channels, optimizing conversion rate, those are really paying off and letting us achieve those lower costs per account, even as we book more accounts and have been tightening our credit box.

Your next question comes from the line of Eric Wasserstrom with UBS.

I'll add my congratulations as well.

Just to follow up on some of these topics.

The efficacy of the marketing spend, as you've talked about it, is very apparent in the flat level of marketing dollars spent, and yet growth remains very robust.

But I guess then, like, is the offset in the income statement, therefore, coming primarily from the rewards rate and in terms of where the cost of this promotional activity is being captured?

To some degree, I would say, Eric, that's a fair statement.

We kind of think of marketing dollars and rewards, to some degree, as flip sides of the same coin.

So either way, you're trying to attract and engage customers to engage with the brand, right?

So sometimes, that's more effectively done via marketing dollars.

Sometimes, more effectively done via rewards.

Last quarter, rewards were a big driver not only on the doubles promotion on the new accounts, if you will -- the sales promo piece that we saw -- but also our 5% rotating cashback category that we do.

Last quarter, it was restaurants, and we saw the highest level of engagement with the 5% rotating category we have ever seen in our history.

So I think, yes, we are targeting in on some of those rewards pieces of the puzzle as a trade-off.

But even if you think about it that way, rewards rate year-over-year increases, pretty darn modest.

And you look at the growth in new accounts we're getting, as well as the growth in balances from existing card members, I think the marketing team has done a great job dialing in, getting that mix right and really kind of driving great new account acquisition.

And Mark, if I can just follow up on one other expense line item, which is the information processing.

What is the -- what are the analytical capabilities that you guys are investing in?

And is it a function -- and what's motivating that?

Is it a perceived lack of capability relative to peers?

Or is it -- what's driving that investment at this stage?

So a portion of it is directly related to where that development is taking place.

Historically, when you did legacy development for in-house systems, the development cost was capitalized.

Today, we are pivoting much more into the cloud and a lot of the development work in the cloud gets expensed.

So that is a big piece of the driver of what's going on.

In terms of exactly where we're investing, I'll kick over to my partner, Roger, here, to kind of give you his thoughts on that one.

Yes.

I'd start by saying it's not being driven by a feeling that we're lagging our peers.

It's being driven by the opportunities we see, truly over a multiyear horizon.

If you look at the tremendous changes that are going on, and I hate throwing around the term, sort of, big data or machine learning, but you are seeing fundamental changes in the power of analytics.

And if you think about everything that is model-driven in our environment, whether it's fraud, whether it's credit, whether it's marketing, and the ability of sort of next-generation models and technologies to drive significant results, let alone sort of automation, whether it's in the call centers or robotic process automation, those are some of the investments we're making.

And again, we believe this will be a multiyear journey.

Your next question comes from the line of Chris Brendler with Buckingham.

I want to talk about the brand campaign for a little bit.

Your advertising budget has been relatively flat, it looks like, from our perspective.

But still see relatively high levels of advertising on TV.

I'm just wondering if you can measure, or discuss at all, the lift you're getting from that branding.

The numbers look like they're consistently gaining share and I think you mentioned this for the positive trends you're seeing in the portfolio.

How much can you attribute that to the brand?

And then a related question, when are we going to see advertising at the cashback checking product?

That seems like a no-brainer as a pretty successful product.

Yes.

So in terms of overall advertising, I think, cross-channel attribution over a complex consumer decision journey, such as getting a credit card, is one of the biggest challenges in marketing.

We've been investing heavily in that area as well and the ability to track sort of where customers see our advertising and what drove them to apply for a card, some of that is proprietary.

But the team in marketing does a great job in terms of making sure that we are getting value for our advertising, both in terms of how it drives decisions but also rotating the placements we have.

And so there's been a lot more science than I think in prior years as people look at advertising.

In terms of the checking product, we are very excited about that.

We have been disciplined in terms of ramping it up, just to make sure that we provide the same great customer experience in that product as we do across all of our others.

We have some actual radio spots that are in market now and we're seeing those drive higher search activity.

So while it will take some time for this to become a meaningful part of our funding, we're very excited about the product, the differentiation it has, and our ability to grow it.

And one quick follow up on the personal loan trajectories.

It seemed like the seasonal increase in the fourth quarter seems relatively normal and doesn't -- 50 basis points sounds reasonable.

We're actually a little higher than that, actually, in our model.

But the 5% for next year is a little higher than I was thinking.

And just wondering, can you just talk about the seasoning curve, when would that peak in 2019?

Is it towards the first half of the year, hopefully?

Yes.

I think that's a reasonable way to model it.

At the end of the day, a big driver next year is going to be that denominator effect piece of the puzzle as well, right?

So otherwise, you would see it fall.

In a stable portfolio, or if you were originating just enough to keep the portfolio flat, you would see it come down a little bit from that 5% level.

So I think the denominator effect will be a piece of it.

Your next question comes from the line of Bill Carcache with Nomura.

First question I wanted to ask is on -- Mark, I wanted to follow up on your reserve commentary.

We see that the year-over-year growth in your loan loss reserves has been slowing since the third quarter of 2017, and it looks like it's on a path to converging with your loan growth.

And it also looks like the year-over-year change in your reserve rate peaked in the third quarter at around 40 basis points and it's been on a downward trajectory since.

So two-part question: first part is, is it reasonable to expect that these trajectories will continue?

And then the second part is, if we are indeed approaching a point where credit headwinds are abating and these trends hold and you sustain your industry-leading loan growth and revenue growth, then is it reasonable to conclude that, that's going to translate into incremental positive operating leverage?

Or would you guys reinvest some of that of the business?

Just some color around those points will be great.

And, congratulations, Roger.

Thanks, Bill.

There's a whole bunch in there.

I'm going to try to hit all of it.

Feel free to use your follow-up to the extent I miss something.

I think at the end of the day, in response to your first question about the declining rate of increase in the reserve growth, yes, I think that is directly reflective of the trends we're seeing as we're approaching that normalized level of losses.

So I kind of think about it a little bit like that old calculus class that used to drive me crazy with limit calculations.

You just kind of turn it on its side and you're kind of -- the limit curve is approaching the asymptote.

And so as that's happening, charge-offs are normalizing and you're seeing the impact to reserve builds in the card account -- in the card portfolio normalize as well.

So I think that's a reasonable way to think about it.

The one thing I would remind you -- or the couple of things I'd remind you in that context is it will be affected by relative growth rates in the loan balances as well.

Right?

So if we get a disproportionate amount of new accounts or disproportionate amount of loan balances coming from things that are less seasoned, that would obviously have bearing and could cause hiccups in that and everything.

But I think as a general rule, you're thinking about it the right way.

In terms of operating leverage, in terms of how to think about that, I would pick up on Roger's earlier commentary.

Well, first of all, I would say we are very committed as a company to positive operating leverage.

So let's put that out there.

There is a lot of leverage in the model that we perhaps aren't fully availing ourselves of right now because we are investing very heavily in that digital and analytics side of the equation that we spoke of.

And we're also taking advantage of originating quality accounts while the market -- while the market is ripe for us to do so.

So I think there's more leverage that exists in the model when we choose to pull those levers.

Right now, the levers we're choosing to pull are the investments in the digital and technology investments because we're seeing tremendous payback on this.

Your next question comes from the line of Rick Shane with JPMorgan.

Roger, congratulations.

Collectively, we've spent a lot of time over the last year, thinking about the profile -- the timeline and profile for a cohort of profitability, thinking about growing up that cohort, reserving for it, experiencing the losses.

You guys are focused now, in part, on growing the portfolio through some promotional balances.

I'm curious, as we've spend so much time refining our thought process on what that profile looks like, if it will be different, if it's delayed for balances to grow through promotional balances, in terms of profitability?

So I would say, it's going to, Rick, depend to some degree on what type of promotional balance you're talking about.

Sales promos, I would say no, right?

So if you're doubling the rewards in the first year for a new card member -- those type things, I would say no, those aren't going to have delaying effects and mitigating effects and all those things.

Balance transfers, the answer is, it depends on the nature of that balance transfer.

If you do a $5,000 line assignment and somebody transfers $4,995 over the day you do the line assignment, you've probably made a mistake.

You're probably dealing with somebody who doesn't have incremental borrowing capacity and you're dealing with a surfer and so you try and minimize for that to the greatest extent you can under, obviously, fair credit laws which you're going to comply with at all costs, obviously.

So I'd say that's a piece of the puzzle.

The other balance transfers, the more normalized balance transfers, I would say no.

There might be a delay by a couple of months, but not anything that's dramatically significant or that I would say would affect your modeling in a huge way.

Your next question comes from the line of Bob Napoli with William Blair.

Roger, congratulations on outlasting David.

The one thing, Roger, that you sounded a little more incrementally excited about for Discover is the growth of the payments business.

You did have very nice growth and there's probably nothing that could help your valuation multiple more than if that became a bigger piece of the pie.

And it seems like a significant opportunity to do that with this -- the network that you have.

Is there an expectation or a focus incrementally that -- on that payment business, on having it become a bigger part of the pie?

And would you look at acquisitions?

And I know they're expensive in the payments space, but I think even a short term dilutive, long term-accretive acquisition probably would be well received.

Yes.

I mean, if you think about payments as a percent of total earnings, we would never slow down the earnings growth on the Direct Banking side, and you can see, while earnings are growing nicely on the payments side of the business, we're also focused on driving loan growth and growth in profitability on the Direct Banking side, too.

Some of our biggest acquisitions have been on the payments side.

So if you look at PULSE, that got us into the debit business, if you look at Diners Club, that brought us global acceptance around the world, both of those were transformational.

Traditionally, we have not done huge-scale acquisitions, so may look at bolt-ons or things that add capabilities.

I would echo your comment around valuations on the payments side.

But growing that business is a big part of our strategy.

We have a unique set of assets.

We're really excited about some of the partnerships we have, whether it's working with PayPal or with Apple or with networks all around the world.

So it is a key focus of ours.

My follow-up question.

Just on the student loan business, the -- there have been a couple of businesses, like SoFi or maybe Navient, I know, focusing very aggressively on refinancing the highest-quality borrowers within that portfolio.

Is that a -- are you seeing incremental -- an increase in the loan payoffs?

And is that something that concerns you about being able to grow that business longer term?

Yes.

So we are seeing pressure on payment rates from student loan consolidators, and that's having an impact on our loan growth.

I wouldn't say it's necessarily impacting charge-offs as much, but really the impact is on the loan growth.

We still have been growing at 10% year-over-year, we expect another record year of originations, and so I'd say it's something we're monitoring.

It'll be interesting to see how those consolidators do in a rising rate environment, given the amount of fixed rate loans that are out there.

So I'd say something we're watching.

We're seeing an impact, but nothing that's changing our overall strategic view on the student loan business.

Your next question comes from the line of Betsy Graseck with Morgan Stanley.

Congratulations, Roger.

Just a couple of follow-ups.

One, on the NIM outlook, I know that you talked about, earlier in the call, at the coming in for the full year below the range of 10.3% to 10.4%.

Could you just give us a sense as to -- that's like ever-so-slightly modestly below, because it feels to me like it would be hard to go really south of 10.2%, but maybe I'm wrong on that.

Yes.

I know you'll see accretion in the fourth quarter from the third quarter's level.

And the way you're describing it is, by far away, the right way to think about it, Betsy.

So it's not going to -- it's -- we're talking a handful of basis points.

We're not talking below the guidance range.

We're not talking a precipitous decline here.

And then separately, I just want to get a little bit of a sense from you on how you think your customers are doing financially?

Obviously, good loan growth year-to-date and expecting to continue to have good loan growth.

So I want to understand, because I know you see your customer's full picture -- or pretty close to their full financial picture -- do you see them having a lot more capacity to add on leverage?

Do you see them taking the leverage that you provide them with and paying down other parts of the stack?

And how do you see their cash flows?

Maybe give us some color on that.

Yes.

So clearly, it's a very robust economic environment.

If you look at unemployment, if you look at wage growth starting to come into the marketplace.

So, given the breadth of our book, that's benefiting our customers as well.

In terms of whether that opens up a huge amount of capacity, we've been pretty clear that we've been tightening the credit box over the last couple of years, and don't really plan on shifting that focus at this point, given how late we are in the cycle.

But again, we feel good about where our customers are, and that's showing up in the numbers.

Yes.

And then I would, kind of, pile on to that, Betsy, and just say one of the things that we watch very closely is obviously not just the overall level of leverage but, as we're also in a rising rate environment, what is the cost of that leverage, right?

And one of the things that's somewhat encouraging, or maybe less troubling about the relevering in the consumer, might be a better way to say it -- is the vast majority of that incremental leverage they put on has been fixed rate.

So unlike going into the crisis, when everybody had a giant ARM and when rates started to rise, the payments exploded, here, it's principally auto loans, student loans and personal loans that have driven the bulk of the increase in leverage since the low point.

And all those have the combination -- could be characteristic of being largely fixed rate in nature.

So, while leverage is back to normalized leverage, the rate gearing implications of that are a little better than historically we have seen them.

So keeping an eye on it, watching it like a hawk.

But right now, it feels manageable.

Okay.

And then just lastly, CECL -- we haven't talked much about that and I know, I think parallel run starts next year.

Is there anything that you would be able to do with regard to the reserve ratio to prepare for that?

Or do you just have to wait until 1Q '20?

Maybe give us some color and updates on how you're thinking about that?

Yes, the only way we could do anything would be to early adopt, Betsy, and we are still working through all the modeling activity that we need to be digging through to get this get ready to implement the standard.

We do intend to do parallel runs.

We do intend to, at some point next year, start giving you guys transparency into what we're looking at.

Our clear hope at the end of the day is there has been a growing drumbeat across multiple different industry groups, across many different people, for a delay and study approach to CECL.

We think there are a number of different quantitative impacts here that just haven't been fully considered.

The interaction between CCAR and CECL or whatever the stress testing for the mid-tiers like ourself is going to be in CECL, how does that interact?

The impact on the provision of loans into the economy.

I mean, there's a pro-cyclical element of this sucker that you're going to pull back on extending credit at the beginning of downturns, right when folks want banks to lend, right?

From an SEC standpoint there's going to be comparability challenges in terms of how you think about -- if, in our card assumptions, we assume a short shallow recession 2 years from now and brand X assumes a long, deep recession in 18 months, I don't know how you're going to compare our reserve thoughts to their reserve thoughts and figure that out.

So I really think delay and study is the right answer.

We have commented on that and to the extent those of you on the call think that, that would make a sense as well, we would encourage you to join us and sign on.

I would definitely like to study more.

Your next question comes from the line of Chris Donat with Sandler O'Neill.

I would also like to add my congratulations to Roger.

And then I'll turn and pivot and ask Mark a question from your February 2014 Investor Day, just for fun.

Mark, what I figured I would ask is just on the personal loan business, because I think that was the last time you gave us some disclosure on sort of the complexion of those loans.

The average FICO was 750, typical term was 3 to 5 years.

And then 60% of that portfolio had a separate Discover relationship.

I was just wondering if, in recent years, if it still looked pretty much the same, particularly on that Discover relationship being the majority of the portfolio for personal loans?

Yes.

So it's about 60% of the portfolio still cross-sold has a relationship with -- on the credit card side -- with us as well.

The average FICO is right around that 750 level still, so no meaningful upward or downward drift in FICO.

In terms of the term, the average life we're seeing right now is about 2 years.

We offer 3 to 7, but the average life is running about 2 years, give or take.

So I would say on a -- actually, I think the average life is like 1.8 years, but round it and call it 2. So at the end of the day, I would say no significant drift in the composition of the portfolio at this point in time, broadly.

Got it.

And then another history question for you, I apologize, but just the mix of promotional balances.

Is that around 19%, that's what is was about a year ago.

Just wondering if it's still in that sort of zip code.

No, it's up about 1% over the course of the last year, give or take, in terms of where it's sitting today.

Our next question comes from the line of Jill Shea with Citi.

Just on the credit quality side on the card side and the slowing year-over-year growth and net charge-offs.

Could you just walk us through the dynamics that you're seeing in the front book versus the back book loss curve?

And then, could you also provide some color on the performance of the front book vintage years and how they're performing relative to each other?

Yes.

We don't provide vintage year disclosure, so I'll stay away from that one.

What I can say as you think about the reserve build this quarter, Jill, a roughly give or take about half of it was result of seasoning of new loan growth.

And roughly half of it was due to that normalization phenomena, with increasing severities as customers have carried more leverage.

So that's about what's driving the reserving phenomena, and obviously, reserving expects what we -- is kind of the driver, really, of kind of what we expect to see coming at us in the future.

So that probably gives you a little bit of sense to how we're thinking about that.

And in terms of the composition, then, of where we're seeing some of this come from in a different lens, I'd say roughly -- let's call it 40% is new account related, 2/3 of the total is back book related.

Now those Venn diagrams aren't mutually exclusive, there's some overlap in there, so don't look at them both -- because some of the back book is new loan balances and so there's different pieces of the puzzle there.

But I think that probably gives you a pretty good way to think about it.

And our final question comes from the line of Bill Ryan with Compass Point.

Just to follow up on the promotional balances, last quarter, I think you disclosed it was about 21% of the card book.

If you looked a year ago, roughly what was the percentage of promo balances?

And if you looked at Discover's card book over a long-term period, roughly what percentage of the card book is typically on a type of a promo balance?

And second part of that question, I mean, you've been running the Double Cashback promo for quite a while.

Is there any incremental types of promo programs that you're running at this point?

You bet.

So a couple of different parts to that.

So the promo rates are up about 1% over a year ago.

So that, call it, 21-ish percent number is still the right way to think about it.

They were more like 20% a year ago, give or take.

Rough math in both cases.

We kind of expect to see them flat, and as I mentioned on my prepared remarks, we don't expect to see them ticking up from here, so we'd expect it -- even though we're remaining more active in some of these segments than that we initially thought it might be, that mix will stay flat as far as that goes.

In terms of thinking about how that level has varied over time, I would say it tends to vary pretty dramatically.

As you might imagine when you see a downturn coming -- when you're in the middle of a downturn, you don't do a lot of promo activity so that balance falls off pretty meaningfully.

When you're at a point in time in the cycle where you see great opportunities to acquire accounts and engage accounts via promotional activity, that tends to tick up.

So we are toward the higher end, if you look at the ranges that we've run at from time.

We're not at the absolute high, but we're toward -- up toward the more upper end of that as opposed to the lower-end, given the nature of the environment we're in right now.

In terms of the doubles, I would say now double continues to be the significant sales promo we're running.

We do run other things from time to time, but the real driver of the numbers is going to be the doubles on the sales promotion side and it's going to be the BT activity as well.

Selissia, we're good.

So thank you, everybody for your questions, your interest and you know how to find us for any follow-ups.

Thanks.

This concludes today's call.

You may now disconnect.