發現金融 (DFS) 2017 Q2 法說會逐字稿

Good afternoon.

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

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

(Operator Instructions)

I would now like to turn the call over to Head of Investor Relations, Tim Schmidt.

You may begin your conference.

Thank you, Julie.

And a sincere thanks to everyone on the call for joining us today.

I'll begin on Slide 2 of our earnings presentation, which you can find in the Financials 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 10-Qs, which are on our website and on file with the SEC.

In the second quarter 2017 earnings materials, we have provided information that compares and reconciles the company's non-GAAP financial measures with GAAP financial information.

And we explain why these measures are useful to management and investors.

We urge you to review that information in conjunction with today's discussion.

Our call today will include formal remarks from David Nelms, our Chairman and Chief Executive Officer; and Mark Graf, our Chief Financial Officer.

After Mark completes his comments, there will be time for a question-and-answer session.

(Operator Instructions)

Now it's my pleasure to turn the call over to David, who will begin his remarks on Page 3 of the presentation.

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

For the second quarter, we delivered net income of $546 million, earnings per share of $1.40 and a return on equity of 19%.

Mark will review the financial results in more detail in a moment, but first, I'd like to update you on our progress in achieving our vision, strategic objectives and shareholder value goal, which are outlined on Slide 3.

We achieved our goal of creating long-term shareholder value by generating profitable, disciplined growth.

In our lead -- lend-centric business model, profitable growth requires diligent management of several key metrics, such as net interest margin, operating expenses, credit cost and loan growth.

We've prudently managed our net interest income and operating expenses as our net interest margin and efficiency ratio rank among the best of the industry and are meeting or outperforming our long-term guidance.

Loan growth over the last few years has contributed to rising charge-offs as those loans season, but it has also generated strong revenue growth that mitigates higher credit costs.

Loan and revenue growth represent our primary near-term opportunity to generate long-term shareholder value.

With that in mind, I'd like to take a moment to examine the current operating environment.

For the first time since 2009, industry-wide card charge-off rates have registered a sustained rise, we believe driven largely by an increasing supply of consumer credit.

Although our charge-off rate remains below the industry average and historical norms, we responded by tightening certain underwriting standards over recent quarters.

Major downturns in the credit cycle usually require an economic recession as a catalyst, however, and microeconomic and macroeconomic conditions remain favorable for U.S. consumers, with a robust labor market, rising housing prices and manageable debt to disposable income levels.

Therefore, we are taking advantage of these favorable conditions to grow loans to prime customers.

Turning now to the quarter's financial highlights on Slide 4, we achieved profitable loan growth and faster payments volume growth, while making prudent investments for the future.

All of our primary lending products contributed meaningfully to our 8% total loan growth, while we increased revenue by 9%.

Notably, we delivered this strong and profitable growth while holding the line on operating expenses and managing our rewards rate slightly lower.

In our view, the level of rewards competition for prime revolvers, our focused customer segment, seems to have reached a plateau.

In our Payment Services segment, volume increased 12% from a year ago, up from the 5% annual increase last quarter.

I'm really pleased that PULSE volume returned to double-digit growth as our network team has secured more favorable routing and expanded our debit relationships.

In keeping with our strategic objectives, we've also deepened and expanded our cardmember relationships by providing excellent products and services that meet our customers' needs with simple and secure interactions.

I'm proud to note that in the recent inaugural studies by J.D. Power, our mobile app ranked highest in customer satisfaction amongst all U.S. credit card companies, and our banking services ranked second highest in customer satisfaction among U.S. direct banks.

With features like Freeze It and instant messaging, our mobile app provides the security and simplicity that our customers seek.

We continually work to find innovative ways to serve and satisfy our customers.

And just last week, we announced a new free feature that helps protect Discover card members from identity theft and fraud.

High customer satisfaction contributes to growth through deep customer loyalty, so these awards validate our focus on the customer as a means to profitable growth and demonstrate our progress towards being the leading direct bank and payments partner.

Looking ahead, we have a strong capital base to support our investments for future growth.

Our strong ROE means we generate more capital than necessary to meet our near-term organic growth needs.

So we continue to return excess capital to shareholders through dividends and share repurchases.

In our recently announced capital plan, we raised both our quarterly common stock dividend and planned share repurchases, positioning Discover to remain a leader in total payout ratio and shareholder yield.

In summary, we continue to make excellent progress in achieving our strategic objectives with a profitable and sustained disciplined approach to generating long-term value for our shareholders.

Before I pass the call to Mark, there's one more piece of information I'd like to share with you.

One of the architects of our disciplined approach to credit management has been Jim Panzarino, our Chief Credit Officer, who is retiring from Discover after 14 years of dedicated service.

I'd like to thank Jim for his many lasting contributions to Discover, one of which was to build a deep and talented credit management team.

This team, which will be led by 10-year Discover veteran, Dan Capozzi, will serve us well by carrying on our strong credit management tradition.

I wish Jim all the best in his retirement.

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

Thanks, David, and good afternoon, everyone.

I'll begin by addressing our summary financial results on Slide 5. For the second quarter, we reported diluted earnings per share of $1.40.

For comparative purposes, I would remind you that last year's second quarter results included a benefit of $0.11 per share related to the resolution of a tax matter with the IRS.

Returning to the current quarter, our 9% revenue growth was driven by a combination of strong loan growth and net interest margin expansion.

The reported increase in provision and reserves is consistent with ongoing supply-driven normalization in the consumer credit industry as well as seasoning of our last several years of loan growth.

Operating expense remained relatively flat year-over-year, which in concert with revenue growth, created strong operating leverage.

David noted that we remain focused on profitable disciplined growth, and you can see evidence of that in our 19% return on equity for the quarter.

Turning to Slide 6. Total loans increased 8% over the prior year, driven by 8% growth in credit card receivables, which represent nearly 80% of total loans.

Growth in standard merchandise revolving balances drove much of the increase in card receivables, spurred by strong sales growth, particularly among revolvers.

We also achieved strong loan growth in our other primary lending products.

Personal loans increased 22% from the prior year, thanks to active customer engagement with our simplified application experience and expanded digital presence.

Given our recent tightening of certain underwriting standards in this category, we do expect slower personal loan growth in the coming quarters.

Private student loan balances arose 2% in aggregate, but our organic portfolio increased 12% year-over-year.

As with personal loans, we leveraged expanded digital content to drive this organic growth.

Moving to the results from our payments network.

On the right-hand side of Slide 6, you can see that proprietary volume rose 5% year-over-year.

This increase was driven primarily by an increase in active accounts.

In our Payment Services segment, PULSE volume grew at a faster rate during the second quarter with an increase of 15%.

Diners Club volume rose 8% from the prior year on strength from newer franchises.

Finally, Network Partners volume declined modestly on lower business-to-business volume.

Moving to revenue on Slide 7. Net interest income increased $187 million or 11% from a year ago, driven by a combination of strong loan growth and higher market rates.

Total noninterest income increased $16 million or 3% year-over-year as higher card sales increased net discount and interchange revenue.

We continue to manage our rewards diligently, lowering the rate 1 basis point year-over-year and 5 basis points sequentially on a decline in promotional rewards.

At 1.2%, our rewards rate remains below our 2017 guidance of 1.26% to 1.28%, although we expect a modest uptick during the remainder of the year as we invest to drive holiday sales.

As shown on Slide 8, our net interest margin rose 17 basis points from the prior year and 4 basis points sequentially, ending the quarter at 10.11%.

Relative to the second quarter of 2016, a higher prime rate and a lower proportion of lower-yielding student loans bolstered the margin, offset in part by higher charge-offs and a more costly funding mix.

Total loan yield increased 26 basis points from a year ago to 11.98%, driven by a 24 basis point increase in card yield and a 32 basis point increase in private student loan yield.

Prime rate increases, partially offset by higher charge-offs, drove card yields higher.

Higher short-term interest rates were the driver of the increase in student loan yields.

On the liability side of the balance sheet, we once again generated good growth in our consumer deposits.

Average balances increased $3.7 billion (sic) [$3.6 billion] or 11% year-over-year.

Consumer deposit rates began to move slightly higher during the second quarter, rising 4 basis points sequentially.

We do expect deposit betas will continue to rise gradually and at a pace which is consistent with our margin guidance.

Turning to Slide 9. Operating expenses rose just $6 million or 1% year-over-year.

Higher employee compensation and benefits, driven primarily by compliance-related headcount and higher average salaries, was largely offset by lower information processing and other expenses.

Marketing expenses declined $6 million or 3% from a year ago.

Card marketing accounted for most of the decline, as deferred acquisition fees and lower brand advertising more than offset higher Internet mail expenses.

In summary, we remain prudent expense managers with an efficiency ratio of just 30% -- 38%, an improvement of more than 300 basis points from the prior year, in part due to the absence of $12 million of look-back expenses in this year's quarter.

This expense discipline, coupled with strong revenue growth, produced very healthy operating leverage in the current period.

I'll now discuss credit results on Slide 10.

Total net charge-offs rose 53 basis points from the prior year and 11 basis points sequentially.

Similarly, credit card net charge-offs rose 55 basis points year-over-year and 10 basis points from the prior quarter.

Private student loan net charge-offs rose 11 basis points year-over-year and 25 basis points sequentially, with the sequential increase in this portfolio being reflective of seasonal trends.

Personal loan net charge-offs increased 80 basis points from the prior year and 2 basis points sequentially.

30-day delinquency rates held relatively stable or declined sequentially across all of our primary lending products.

In looking at total loan receivables, our 30-day delinquency rate declined 4 basis points sequentially, consistent with the seasonal improvement in the second quarter of last year.

As we've previously noted, supply-driven credit normalization, along with the seasoning of loan growth in the past few years, have been the primary factors behind rising charge-offs.

Incidence rates on our seasoned card portfolio have remained relatively flat, reflecting the current benign macroeconomic environment.

However, an increasing supply of consumer credit and an attendant increase in consumer leverage have driven higher loss severities.

Consequently, our net charge-off rate has increased somewhat faster than our original expectations, and as a result, we're revising our full year 2017 guidance on the total loan net charge-off rate to a range of 2.7% to 2.8%.

In light of this revised guidance, I think it's important to remember that we have consistently noted that we have multiple levers available to offset some of the impact of increasing provisions when we deem it appropriate.

For example, as you saw in this quarter, strong revenue growth and expense discipline can temper the impact of rising charge-offs on our earnings.

So to assist in properly assessing the impact of this revised charge-off guidance, I would break from our custom of not commenting on EPS expectations and note that we believe that The Street's full year 2017 consensus estimate of $5.98 per share is in the correct ZIP code.

Looking at capital on Slide 11.

Our Tier 1 common equity ratio declined 40 basis points sequentially as risk-weighted assets increased and we returned more capital to shareholders in the form of common stock dividends and share repurchases.

Year-over-year, our CET1 ratio declined 130 basis points as we recorded strong loan growth and returned more than $2.4 billion of capital to our shareholders.

On June 28, in conjunction with receipt of our CCAR non-objection, we announced our capital plan for the 4 quarters ending June 30, 2018.

Specifically, we disclosed planned share repurchases of $2.23 billion over the next 4 quarters as well as a $0.05 per share increase in our quarterly common stock dividend.

One additional item to note is that beginning this quarter, the payment date for the common stock dividend will be moved from late in the second month to early in the third month of each quarter.

In short, we continue to actively deploy our shareholders' capital through profitable and disciplined asset growth, while at the same time, returning a healthy portion in the form of dividends and share repurchases.

To sum up the quarter on Slide 12, we're very pleased with our operating performance.

With half the year behind us, we're delivering solid financial results with 9% revenue growth, a double-digit NIM and a healthy 19% return on equity.

In addition, expenses remain well managed as evidenced by one of the best efficiency ratios in the industry.

With regard to the balance sheet, we generated strong 8% total loan growth, with significant contributions from all 3 of our primary lending products.

We continue to expand our consumer deposit base to fund these loans while realizing the benefits of relatively low deposit betas.

With respect to credit, while our charge-off rates have risen as credit conditions normalize and loan season, they remained below both historical norms and the industry averages.

Finally, our robust capital position enables both profitable asset growth and a leading payout ratio and total yield for our shareholders.

That concludes our formal remarks.

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

(Operator Instructions) Your first question comes from the line of Sanjay Sakhrani with KBW.

I guess, the first question is on credit quality.

Obviously, it's been a bit more of a moving target this year for the worst.

Could you just talk about sort of what specifically is driving this trend?

And maybe just what drives your comfort level on the revised expectations?

And what it might mean for next year?

Sure, Sanjay.

I guess, I'd say a couple different things.

First of all, I just reinforced that by historical standards and compared to industry norms, credit remains exceptionally well below what we'd expect in either of those categories normally.

I would say, the -- what we're seeing is a combination of both seasoning of new account growth as well as an increase in severity when losses occur in the back book.

Up to this quarter, it's been principally driven by the seasoning effect.

In this past quarter, we also saw an uptick in the severity elements associated with the back book.

In terms of confidence in terms of where we stand right now, I would say delinquency rates have largely stabilized.

And as you know, they're a very reliable predictor of what we think's coming at us over the course of the next 6 months, the remainder of the year.

The macro fundamentals remain strong, too, right?

You've got a very strong labor market, HPI is in good shape, bankruptcy filings in good shape, debt-to-disposable income in good shape.

So on balance, we continue to believe it's 2 factors driving this, the first of which is the seasoning of new accounts, and the second of which really is the increase in leverage that we've seen out there across consumers.

So while incidence rates aren't -- on the back book aren't moving, we are seeing movement in severities.

Okay.

And then as far as asset sensitivity is concerned, obviously, we haven't seen a whole lot of it year-to-date.

Could you just talk about how we should think about further rate rises and its impact on the NIM?

Yes.

I think in terms of what we're seeing flow through, the market rate element in this last quarter remain pretty consistent with what we've guided to before, Sanjay.

So if you did the walk on a sequential basis from the first to the second quarter, market rates got you about a 13 basis point increase in NIM.

The receivables rate was about a 10 basis point bad guy on margin.

That was due to higher promotional balances in the card book.

The funding rate was about a 1 basis point detractor; and then receivables mix, due to a higher revolve rate in the card book, got you about a 2 basis point good guy.

So that's where your 4 basis points walk on a sequential basis comes from.

So really, as you think about it going forward, I'd say that rough 15 basis point market rate movement for every 25 basis point movement, rates remains intact.

The other things that affect it though, obviously, that you got to keep an eye on are portfolio mix, promotional activity, charge-offs of accrued interest and deposit betas.

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

Mark, thank you for the -- reiterating the back half of the year guidance.

I guess, when we think forward, thinking about credit, I guess on the back of Sanjay's question, loan growth accelerated the past 10 to 12 months, which I'm assuming we haven't really started to see the seasoning component of that.

If we were to assume no change in the operating environment, how should we think about the pace of credit losses as we start to look beyond the next 2 quarters, given that we have accelerated and we all understand the fact that there is a seasoning component, especially when you bring on a much bigger vintage?

Understood.

Ryan, what I would say is I'm reluctant to guide for next year until our January call, which is when we typically give guidance beyond this year.

But I guess what I would say is, if I think about the type of credit we're originating, we continue to remain very disciplined in terms of what we're putting on the books.

And all the vintages continue to season essentially in line with or better than our expectation.

So we continue to be very disciplined in terms of how we're approaching the credit space.

I would say from the standpoint of -- to the standpoint of pulling -- to the point of pulling back a few years ago when we saw some of the channels we were originating in weren't generating the kind of returns we were before returning to the kind of growth we're seeing today over the course really of that last 10 to 12 months, as you noted.

The other thing I'd point out is, as we noted in our prepared remarks, we are pulling back on personal loan growth now, particularly in the unsolicited portfolio where, candidly, we're seeing some development that's not consistent with what we would like to see.

So I think we feel very good that we're booking prudent, profitable loans that, as we've said before in a lend-centric business model, are really key to compounding value over time.

So I know that's not a direct response to what's the pace for 2018.

We'll make you wait for that until January.

What I would say, though, is we feel very good about the growth we're booking.

Got it.

And maybe a 2-part question for David.

One, the reward costs came in better than we expected.

I think you talked about plateauing.

You talked about it inter quarter, that competition's started to level off.

Can you maybe just talk about what you're seeing competitively?

And then second, you guys did note that you were -- that you tightened.

Can you maybe talk about where you tightened?

And what that could mean for expectations for loan growth moving ahead from here?

Sure, Ryan.

Well, I certainly think that if you saw the modest reduction in our reward rate this quarter, even with accelerating growth, that would suggest that we're not having to pay up more on rewards to generate that growth.

And so that, I think, is one indicator that I feel like, over the last 12 months, that rewards competition after a big ramp-up has sort of plateaued, still at a high level.

But I think that we see that in those companies that report their number of new accounts and marketing expenses, that the first half of this year, a number of people have pulled back a bit.

And I believe that our discipline from a year ago, when we saw some cost per accounts, that it got too high and the economics didn't make sense to us, other people read some of those results and maybe made adjustments, which have allowed us to return to growth while growing revenue even faster and keeping our expenses remarkably flat.

In terms of where we've tightened, Mark mentioned personal loans.

I'd say in the credit card side, generally, because most of what we're now seeing, we think is more environmentally.

It's not that we had credit issues.

It's -- I would characterize it as pretty modest adjustments that we've made, and it's really been on specific segments where we've seen some stress.

And Mark talked about the higher credit expansion out there.

We've seen certain segments where there was -- maybe competitors have gone a bit far, and then that affects our credit as well.

And the place I'd point to -- one place I'd point to is in personal loans.

We've seen a lot of new players that don't have 10 years of data, haven't been through a cycle, lot of expansion in credit.

And so we have a relatively small number of credit card customers, who have personal loans with someone else.

But those people have tended to underperform and underperform what our expectations are, and underperform what their FICO scores would otherwise suggest.

And so we've adjusted our models to help compensate for what we think is a bubble of outside credit that is affecting us.

Your next question comes from the line of Brian Foran with Autonomous.

I wonder on the credit comments, so 2 things: One, just a clarification.

Mark, I think you mentioned you were seeing the delinquencies stabilize.

Is that -- were you referring to the year-over-year comps or delinquency by vintage?

Or what delinquency metric were you referring to?

And then two, when you do look at the vintages, are there any early signs that second half '16 or even first -- I guess it's a little early for first half '17.

But are there any early signs that the most recent vintages have started doing a little bit better than maybe the vintages from '14 and '15?

Yes, so a bunch in there.

I'll try and hit it all.

If I miss something, let me know.

I would say the stabilization and delinquencies comment that I made was specific to sequentially, because we've seen the delinquency formation really begin to stabilize, which gives us a high degree of comfort in the guidance that we offer at this point in time.

We are seeing some stabilization in some of the newer vintages as well.

So I would say, on balance, it feels like we're hitting somewhat of a new plateau at this point in time.

That being said, the wildcard is, as David noted, we got really good insight into our own growth-driven provisioning or our own growth-driven credit delinquency statistics.

It's some of the environmental stuff where we're seeing the severities that has been the unexpected piece, quite honestly, over the course of the year, this year.

If you think back to the end of the first quarter, the old guidance we had out there for charge-offs, we were currently forecasting to be right at the upper end of that range.

Over the course of this quarter, we've seen the severity in the back book pick up and really attributable to the overall leverage levels consumers are carrying.

It seems to be managed.

It seems to be controlled that the income ratios are great.

The incidence rates are very stable.

So it's just a matter of when folks do go bad, and they have more credit outstanding, and so what's available to each creditor is a little bit lower.

So it feels like normalization-type activity like we saw back in the mid-90s, the last time we saw credit really normalize without a cyclical turn, and it feels like that's what we see, Brian.

And we continue to feel confident it's a good environment in which to be originating new loans.

One follow-up, if I could.

I think over time, you've been clear about some of the challenges with the aggregator channel that developed, I guess, some time a year ago or maybe a little bit more.

And I think you've mostly cited the cost to acquire accounts.

As you kind of continued to read the data, has the aggregator channel produced any adverse credit results as well?

So I would say the -- at the margin, the aggregator channels have gotten a little bit more cost-effective.

But also, we've moved away from as heavy of a mix of aggregators.

So some of our digital channels, in particular, have been very cost-effective.

So I'd say, in terms of the credit performance, I haven't seen any particular trend.

It's been more of a cost-per-account issue, as far as I know.

Although, the one thing I would call out, Brian, is clearly on a prequalified basis, credit continues to perform better than on an unsolicited basis.

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

A couple of questions.

One, Mark, earlier you mentioned that there's some offsets that you have to keep EPS in the $5.90 range.

Can you just go through some of what those are?

Yes.

So we've talked about levers over the last couple of years that we have available when these things happen.

I mean, 2 of them, I think you saw some evidence of in the last quarter, that being our ability to drive revenue growth, but really also our ability to be extremely disciplined around the expense line.

I would say beyond that, Betsy, we'd debated giving a truckload of guidance or a little bit of guidance.

And we really went to kind of highlighting that, that $5.98 a share number of The Street out there felt like it was in the right ZIP code.

What I would say is all the levers that we refer to are operating in nature.

Nothing cute intended in there, like one-time tax items up our sleeve or anything like that.

Really, our operating levers we have available on both, I guess, the volume side, the margin side, the expense side, any number of different drivers.

And I would just add, Betsy, that if overall, if you looked at this quarter at a 9% revenue growth and only a 1% expense growth, that's the operating margin that Mark's referring to.

And that higher profit covered a lot of the provision increase.

Yes.

Yes, I know.

Definitely saw that.

And that's part of the reason I was asking the question because not sure how much more operating leverage, like the kind we saw this quarter, is in the back pocket, so to speak.

Yes.

I guess, what I'd say in that regard is, we will see a little bit of drift up in expenses in the back half of the year from where we are right now, just because of the seasonal spend that you have in the back half of the year to drive holiday spending and sales.

That being said, I think a wise person would assume that expenses are one of those levers that, relative to our plan for this year for expenses, we will probably be pulling.

Okay.

Okay.

And then just one other on the credit.

You mentioned the bubble of credit that's impacting right now.

Is this -- are we 1 quarter in, 2 quarters in?

How far in do you think you are?

Well, I talked about one specific piece.

But if you look at credit cards, if you were to look across the credit bureaus at total credit card lines outstanding, there was a contraction up until about 2013.

And then there's been a pretty sizable increase since then.

It's disproportionately gone to transactors, so it hasn't fully shown up in -- a lot of it hasn't shown up in balances.

But one estimate I saw was there's about 30% higher credit lines available today than in 2013 in the card business, and some of that will show up in additional risk.

It's, on average, to less risky people, but there's more available out there.

And at the margin, you get some people that could get into too much debt.

And so higher credit availability affects all creditors.

And that's one of the things that we talk about, supply-driven normalization.

To some degree, we've had a couple of years of returning to normal credit availability postcrisis that's going to drive a move towards more normal credit charge-offs for everyone.

Yes.

And Betsy, the one thing I would say is, just to give you some comfort, I'd reiterate my earlier comment that we saw -- we've seen delinquency rates on a sequential basis really kind of level out, and the delinquency formation seems to have stabilized.

In addition to that, the macro factors all remain very positive.

And incidence rates, which is a critical piece of the puzzle, right, incidence rates in the back book are relatively flat.

So it doesn't need to -- appear to be anything cyclical, anything embedded.

It's really, on a percentage basis, you'd -- just the people who are going bad are carrying more balances.

The other thing I think that's relative to this is, I think we've been pretty transparent over the course of the last couple of years that we chose not to participate in the rewards war for the high-spend transactor.

And as a result, I think we've been pretty transparent.

We've lost some transactor volume.

That, clearly, will show up in the credit-related rates that we discussed, delinquency charge-off, all those things, just because you don't have those balances in the denominator any longer.

But the one thing I would point out is transactors also don't tend to drive a lot of losses.

So one thing I'd encourage everybody to think about is there is a difference between the rate and the dollars of loss, a bit of a disconnect there that having made a decision to let some of those high-spend transactors attrite, I don't think the dollars of loss we face in our P&L will be radically different.

The rates will look different, again, just because of that denominator effect.

But they also don't drive net revenue.

Correct, absolutely correct.

Okay.

And then just last, Dave, can you just give us a sense as to the kinds of things that you've done to tighten up as you saw this happening?

What measures did you take on either underwriting or line size?

Or did you freeze lines?

Or just trying to understand what your strategy had been.

Yes, I -- we didn't because we kept the discipline 2 years ago, 3 years ago, we really -- and because we're a prime competitor, we saw with the industry, what my observation is the people that are in sub-prime business maybe had more loosening than more tightening.

And so I think ours is relatively less, and it's just on the margin.

Customers that had certain characteristics, we might tweak our models for who qualifies for a line increase at the margin.

We would adjust who we're marketing to, where our cutoffs are for new accounts.

And so there's been no major change.

There's been tweaking of models for both the portfolio and for new accounts.

Your next question comes from the line of Ken Bruce with Bank of America Merrill Lynch.

My question -- my first question relates to credit.

And I'm sorry, I know this is kind of getting in deep.

But you'd mentioned that historically, you needed an economic catalyst of sorts to kind of lead the worsening credit.

Our work would suggest that growth itself can actually kind of lead to worsening credit, just with the asset growth.

And you're kind of seeing that now.

I would like to kind of get a sense as to how comfortable you are that you're not going to wake up and 2 years from now, you're not going to see materially worse credit than what you've got currently.

And if you could kind of give us some thoughts around that, that would be very helpful.

Well, I'm not going to give 2-year forward guidance, but I will react to some thoughts a little bit here.

I guess, Ken, I liken the period we're in right now, from what we're seeing, most to a period we last saw in the mid-90s, right?

So you have supply-side, credit-driven phenomena, as opposed to a macroeconomic or cyclical-driven phenomena that's causing a normalization of losses, right?

So the macro fundamentals all remain very strong.

The labor market, HPI, [BKs], all the rest of those factors continue to be very well behaved.

But what you seen is the cycle has persisted for longer, and as we have come off of loss rates that were just lower than we've ever seen for longer than we've ever seen, I think you saw an increase of consumer credit come into the marketplace.

And I think, as folks are going bad, what you're seeing is, with that increased leverage, particularly in personal loans and auto loans, and a few other categories, you're seeing fewer dollars available to cover that increased debt burden.

So it really is what I would call a supply-driven as opposed to a macroeconomic, cyclical-type phenomenon.

So that's why we're comfortable continuing to underwrite credit to the extent we are.

Obviously, we have not suspended the fact that we compete in a cyclical industry, and at some point in time, those macroeconomic factors will kick in and affect everybody.

Not going to speculate as to when that would be.

We're not seeing evidence of that at this point in time though, is what I'd say.

No, I understand, and I agree with your points.

I guess, the fact is that there is continuing to be an expansion of credit into the sector.

I think, without putting too precise of a time on it, by the time that the industry figures it out, it's probably going to be too late.

We're going to be whistling past the graveyard, so to speak.

And we're going to wake up with much higher losses, and some of these books are not going to perform the way that everybody had expecting them to.

So let me get this -- the follow-up question is, what do you need to see to start to pull back more aggressively?

Let me try my hand at it, too, Ken.

I think so far, the losses, the new vintages are performing per our expectations.

And so if they started going beyond that, then we would start taking action.

But to some degree, I have been in the industry when there's times when people, consumers and creditors overdo it.

I don't think we're in that kind of time.

I think that consumers and creditors were overly conservative a few years back, which is natural postcrisis.

And frankly, if you're in the prime credit card business, we never shopped for nor assumed that we were in a 2% charge-off industry.

In fact, we would be leaving a lot of money on the table by shooting for that low.

And the industry was abnormally low.

So what I'm seeing so far is a sort of return to what you would normally expect would sort of maximize the profitability, where you're not taking too much risk, but not way underdoing it with risk as well.

And to some degree, this is the first normalization post-CARD Act, post cycle.

So there is some uncertainty as to that new normal.

But I think whether it's us or anyone else in the industry, people have been talking about 3% or 4% as being more normal, and it seems like we're just moving towards that.

We're not even to that level yet.

So what I'm seeing so far is normal.

I'm just surprised it took this long, frankly.

Well, I think we were all kind of wondering when it was going to show up.

And yes, I realize I'm being a little critical here.

It's not meant to necessarily second-guess your actions.

It just feels like there's just been a big move.

And every time we get essentially surprised by a worsening in credit, that the market reaction is a negative one.

So we're just trying to kind of best understand it.

Your next question comes from the line of Henry [Cyganovich] with Wedbush.

It's Henry Coffey.

And my good colleague, hopefully, will follow.

So in very simple terms, when we look at the reserve build that occurred in the quarter and you look at the revised charge-off guidance, how much of a jump was there because of the revised charge-off guidance?

Or is that kind of higher reserve component going to work its way into earnings over the next several quarters?

So I guess, what I would say, Henry, is that when we set those reserves, that's a -- basically, it's an actuarial model that establishes the reserves.

So it takes into account the development we expect to see over the coming 12-month period of time.

And then it takes into account the number of macroeconomic factors that come in from Moody's, macro advisers, any one of a number of different providers.

I would -- I guess, the way I'd answer the question fundamentally is to underscore the fact that I'd take you back to my earlier comment, that we've seen delinquency formation moderate to a point where it's, on a sequential basis, pretty much flat to slightly down across most of the product.

So I think what that should probably imply is that, based on what we see right now, we wouldn't expect a reserve build of the same magnitude in the next quarter.

But I would just point out that's going to, obviously, be impacted by any changes that we would see over the course of this coming quarter in terms of development.

It would also be impacted by anything we would see in terms of changes to the macroeconomic forecast that we take in as part of that process as well.

But I think the development we've seen is reflected in the reserves we've established for that 12-month, forward-looking period.

And we feel that we are adequately reserved and reserved well in accordance with GAAP.

And then, Henry, if I understood you, though, you were asking if it could get reversed.

And I wouldn't be counting.

No, no, no.

I just wondered is it going to be another -- no.

Are we going to see another jump like that in the next couple of quarters as well?

Okay, okay.

Again, it's impossible to fully predict because it's dependent upon macroeconomic indicators that are forecast by other firms that we don't control.

But the delinquency formation piece of it, all the parts and pieces we see right now would tell us there's a stabilization taking place.

On a unrelated subject, student lending, you're #3.

Is -- we're looking for signs of this, though.

We certainly -- the winds keep -- seem to point to this.

Somebody is gaining a lot of share on Wells Fargo.

Where do you think you'll end up in the league tables by the end of the year?

We're going into the growth season.

It seems like you're very much committed to this as a growth business.

But what is the outlook?

What are your thoughts going into the growth season, et cetera?

We feel good about the season so far.

We -- looking at the numbers, I mean, Sallie Mae has, by far, the larger market share.

We grew a bit faster than they did in the first half of this year.

And so I would think that -- I'm hopeful that we will pick up some market share from both of our key competitors, versus Sallie Mae and Wells Fargo.

Your next question comes from the line of Arren Cyganovich with D.A. Davidson.

Maybe on the deposit side.

We have seen some of the high-yield savings rates rise to 1 3, 1 4-type levels.

You're still well below that, from what I can tell.

What are you seeing in terms of inflows and outflows?

And how much risk do you see from that going forward?

Yes, so I would say, Arren, we see on the deposit side is we don't intend to lead a price war by any stretch of the imagination.

We went through a lengthy process over the course of the last 4, 5 years to really move ourselves down in the league table standings and really focus on cross-selling to our existing customer base.

So we're now at a point where greater than 60% of our depositors have a relationship with us on the asset side of the balance sheet.

Over 80% of last year's new depositors have a relationship with us on the asset side of the balance sheet.

So we think our betas can be sustainably lower than what I would call a typical online bank that doesn't have that synergy between the right- and left-hand sides of its balance sheet.

What I would say is we have noted that betas have come back into the mix.

We're actually kind of surprised they really took as long as they did to begin to creep back into the mix.

But they're still well below where we expect them to be normalized, and well below the levels we bake into our margin guidance for the year as well.

That's helpful.

And I guess, from the standpoint that you had a moderate miss this quarter from The Street, that you're kind of saying that The Street's in the right ballpark for the full year, is it mostly on the expense lever that you can make that up?

Or I don't know if you want to get any more granular on where the difference is there.

But you have very strong revenue growth.

I'm just wondering where The Street might be missing things from their second half of the year.

Yes, we decided to be a little bit more pithy in our response than to go line item by line item in terms of the specific guidance.

Again, I would -- Arren, I think I did say earlier and I'd reiterate, clearly, the expense lever is one we will be pulling.

They still will trend up in the back half of the year a little bit because of the seasonal spend, but not necessarily to the levels we had planned for by any stretch of the imagination either.

So I think there's a -- that's definitely one of those levers.

There's other -- there's volume levers.

There's margin levers.

There's all kinds of different other ones we can be pulling.

But we chose, just to again be pretty pithy and stick to just the number we felt really mattered because of that EPS line.

Your next question comes from the line of Jason Arnold with RBC Capital Markets.

We talked about, I guess, the low -- the loan seasoning and kind of growth in card contributing to the higher charge-off rates.

But are there any other drivers of the severity elements in card that we haven't discussed, that you could kind of get into in a little bit more detail?

One thing I would just say is that we have also been doing line increases over the last 5, 8 years.

And so cumulatively, we see a growth in average balance.

And so that naturally contributes as well.

The only other thing I would say there, Jason, is it's not really -- if you think about it holistically, it's not necessarily what happens in card that drives losses in card, right?

So it's really what's the consumer's overall debt burden?

Because we're not seeing -- again, we're not seeing real increase in incidence rates to any meaningful degree.

It's really just that consumers are levering, and we're seeing that leverage occur in the federal student loan program.

We're seeing it in auto loans.

We're seeing it in personal loans.

Those would be the 3 categories that if you index back to precrisis, would show they've kind of driven the relevering of the consumer.

So if that consumer encounters an event, which typically tends to be an event, because those incidence rates haven't moved and debt-to-income remains manageable, they simply are carrying a lot more leverage than they were.

It doesn't really matter what product it occurred in.

If they get over-levered, they have a problem, all the creditors tend to feel a little bit of that pain.

So I wouldn't specifically link card to card or personal loans to personal loans per se.

Okay.

Helpful color there.

And then second one I would ask is just on the appeal of the balance transfer market, kind of what are you seeing competitively there?

And kind of what's the appeal of that market right now?

So it remains a very attractive channel for us right now.

We are utilizing it, particularly on the portfolio side of the equation, as opposed to the new accounts side of the equation, to reengage consumers.

I think I, in my sequential NIM walk earlier, pointed out there was about 10 basis points of detrimental effect to NIM as a result of increased promo balances.

And a good chunk of that was the balance transfer activity.

I'd say they remain very high FICOs.

You know that 730-ish average FICO on the card book is what you should still be thinking about us -- the scope in which we're generally playing, and even higher than that for personal loans and for student loans.

But BTs are a healthy channel right now.

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

So most of my questions have been asked and answered, some more than once.

So we'll try to get on a different tack.

Just thinking about the earnings guidance that you gave, it certainly means that the reserve build moderates in the back half.

And I know you don't want to give a forecast for '18.

But I guess, the question is, assuming that the macroeconomic factors that you kind of outlined are relatively stable and the growth rate is kind of plus or minus where we are, are there any other reasons to assume that there would be a variation going into '18 from where we were in the back half of '17 for the amount that you'd be building reserves?

No, I don't think that's all been necessarily asked and answered.

There's some new stuff in there.

I guess, what I would say, Moshe, is that it's going to depend on a number of different factors.

It's going to depend, obviously, on the level of competitive -- what competitors are willing to do from a credit perspective.

It would, obviously, depend on macroeconomics.

It would depend on any number of different things, would depend on what we choose to do with our own underwritings.

But I would say that when we think about credit, your perspective on credit is decidedly dependent upon the temporal lens through which you're viewing it.

If you think about charge-offs, we'll readily stipulate that they have been going up.

There's no question about it.

And if you think about it on a 3-quarter graph, it looks like a 45-degree angle sloping upward.

But if you think about it on a 10-year graph, they fell from crazy levels to the lowest levels we've ever seen in a long time, and they've kind of hung in there.

And they've just now started ticking up, really, over the course, say, of the last year or so.

So I mean, from our perspective, it remains an extremely attractive time to be putting new loans on the books because charge-offs, loss rates are well below what we would expect in normalized terms they would be.

So I guess, that's a long-winded way of setting up.

We would expect some degree of normalization in the book will continue just coming off of unsustainably low levels, principally driven by this supply-side phenomena that we kind of referenced a couple times here.

But at the end of the day, it doesn't feel, based on what we're seeing right now, like this is anything like resembling any kind of a runaway train.

And the one thing I would add is we're happy with our growth, and continued growth will tend to drive continued reserving as well.

Right.

And I wanted to come back to one comment that you made actually a couple of times about unsolicited borrowers performing a lot worse.

Could you just talk about the incidence of that, both in your personal loans and if there is any in the credit card?

And how that might have -- you might be managing that going forward?

Well, it's -- I'd say, I think we bought it up specifically on personal loans.

Over half of our new accounts and balances come from cross-selling our existing largely credit card customer base.

And we're seeing a lot more stability there.

In recent years, we have expanded into the broad market, and we've seen a lot.

It's the lowest barriers to entry.

It's relatively quite easy to enter the personal loans business relative to student loans or credit cards or even deposits.

And so we've seen a huge number of new players come in, huge growth in availability.

And a lot of these creditors, they don't have 10 years of data through a cycle.

They talk about the great models.

But you need experience in data to be strong.

So if there's an overcapacity there, it's going to tend to be in the non-solicited.

It's going to tend to be in the broad market.

And so I would cite the really explosion of new entrants in the broad market that are impacting credit, actually, both for our unsolicited personal loans, but it's also impacting credit cards and other asset types because it affects a subset of customers.

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

I'm wondering, you made comments about tightening of underwriting.

You've also talked about expense as a lever into the second half of the year.

I'm wondering how we should think about that in terms of loan growth.

Second quarter was essentially a -- the fourth sequential quarter where loan growth on a year-over-year basis accelerated.

Should we start to see that moderate, given the other 2 factors?

Not necessarily, because you saw us achieve accelerated loan growth this quarter even with the very modest expense growth.

So we're achieving efficiencies.

Our cost per account are very strong.

And so I would say in personal loans, you'll see -- I would expect some slowdown in the growth rate.

For the reasons I mentioned, we'll be pulling back somewhat in the unsolicited channel.

But in credit cards, it's not that we've been spending more money that we can now cut back.

It's that we've been getting better results, and we're pleased.

And frankly, the lever of getting more revenue from good accounts is a bigger lever than cutting marketing expense.

Yes.

And I would also point out -- you didn't ask the question, but I will volunteer one more thing to give you some added comfort on that, Rick, and that is that these new accounts are coming in with much lower average lines, too, right?

So I think we're being very disciplined and not -- so yes, the growth is very solid right now.

But we're very, very disciplined about the nature of that growth.

And again, I'd point to the -- our experience with the aggregator channel a number of years ago, where we pulled back when the economics just didn't make sense.

I'd point to our statement today about pulling back on the unsolicited channel and personal loans where it doesn't make sense.

And if we start to see development in the card book that would warrant doing something, we'd certainly do it.

But to David's point, heretofore, what we're booking continues to be the kind of stuff that compound shareholder value over the long haul.

Mark, that actually saves me my follow-ups and gives me a chance to ask one more, which is, I think the most misunderstood factor in the card industry right now might be the actual seasoning curves of loan losses.

And I'm curious, if you were to look and sort of index to a 5-year -- 5 years out with a loss rate as, let's call that 100, what in year 2 and year 3 losses would look like, so that we can get a real sense of what that peak development -- how that peak development occurs?

That's a -- I have to give some time to think about that one to get to your index graph, Rick.

I guess, what I'd say is, I'd remind everybody that in a prime card book, you tend to take peak losses about 24 months after you originate.

I would remind everyone that the way you reserve for those losses is you reserve on a 12-month, forward-looking basis.

So about 6 to 12 months out, you start really seeing the reserve build associated with what you expect to be losing at the peak of that seasoning curve, right?

And then you kind of fall off that 24-month peak.

And somewhere between 36 and 48 months out, you kind of stabilize at what I would call the portfolio loss rate, is really the way to think about the way the curve plays out.

What I would say is since the crisis, those vintages are -- all those vintages are peaking at lower levels than what we've seen prior to the crisis at those peak levels, so they're peaking lower.

They're inching a little bit higher every year, but still peaking well below what we would've seen in the typical precrisis vintages.

So hopefully, that's helpful for folks to think through it.

Great.

Well, that concludes today's call.

And we'd like to thank everyone for joining us.

If you have any follow-up questions, please call the Investor Relations Department.

Have a good night.

This concludes today's call.

You may now disconnect.