Bread Financial Holdings, Inc. (BFH) Q2 2018 Earnings Report, Transcript and Summary

Good morning, and welcome to the Alliance Data Second Quarter 2018 Earnings Conference Call. [Operator Instructions] In order to view the company's presentation on their website, please remember to turn off your pop-up blocker on your computer. It is now my pleasure to introduce your host, [ Ms. Vicki Necla ] of Advisory Partners. Thank you. The floor is yours.

Thank you, operator. By now, you should have received a copy of the company's second quarter 2018 earnings release. If you haven't, please call Advisory Partners at (212) 750-5800. On the call today, we have Ed Heffernan, President and Chief Executive Officer of Alliance Data; and Charles Horn, Chief Financial Officer of Alliance Data. Before we begin, I would like to remind you that some of the comments made on today's call and some of the responses to your questions may contain forward-looking statements. These statements are subject to the risks and the uncertainties described in the company's earnings release and other filings with the SEC. Alliance Data has no obligation to update the information presented on the call. Also on today's call, our speakers will reference certain non-GAAP financial measures, which we believe will provide useful information for investors. Reconciliation of those measures to GAAP will be posted on the Investor Relations website at alliancedata.com. With that, I would like to turn the call over to Ed Heffernan. Ed?

Great. Thanks, [ Vicki ], and good morning, everyone. Joining me today, as always, is Charles Horn, our CFO, and he'll provide an update on the second quarter results. And then I'll add some color and update our 2018 outlook. We plan to keep the prepared remarks to about 20 minutes so we can dedicate more time to Q&A. Charles?

Thanks, Ed. Let's turn to Slide 4 and talk about the consolidated results. Pro forma revenue increased 8% to $1.97 billion for the second quarter of 2018, led by double-digit growth at LoyaltyOne and Card Services. EPS increased 59% to $3.93 for the second quarter of 2018, aided by a onetime foreign tax benefit, which reduced the GAAP effective tax rate to approximately 14% for the second quarter. Core EPS increased 31% to $5.01 using a core effective tax rate of approximately 24%, which excludes the onetime foreign tax benefit. Lastly, we spent approximately $100 million on share repurchases during the second quarter while dropping our corporate leverage ratio to 2.5x compared to our covenant of 3.5x and the end of Q1 2018 of 2.7x. Let's go to the next slide and talk about the hot topic of the last few days: delinquencies. Even though delinquency trends improved during the quarter, as shown on the chart, the delinquency rate of 5.5% at quarter-end was above our expectations. The miss was primarily due to the influence of TDRs, modified accounts added during the first quarter of 2018, primarily from FEMA-designated disaster areas. What is a TDR? It is essentially our hardship program, which is designed to help cardholders with extenuating circumstances get back on their feet after an event such as a hurricane. The program reduces minimum payment rates, lowers interest rates and suspends late fees. The maximum time frame an account can stay in the program is 12 months, and the accounts are not allowed to bounce into and out of the program. As you saw in the first quarter 10-Q, we added a number of accounts into the program following the natural disasters late last year. It is important to note that TDR accounts are not removed from delinquency calculations, which can create noise in the delinquency trends. Why? Because there are 3 outcomes: the first, the account makes a payment and stays in the existing aging bucket; second, the account makes 3 consecutive payments and returns to current status; and third, the account misses 2 consecutive payments or 3 payments over 12 months, is removed from the program and ages to charge-off. As accounts were added to the TDR program, a higher reserve rate is applied to these balances assuming there's a higher probability of later default. The reserve rate for TDRs is currently at 28%. We feel good that our allowance for bad debt reserve is adequate. For June, we were wrong about the timing of the recidivism. It was a little faster than expected. TDR accounts in the mid to late stages of delinquency who have not met the payment requirements have begun to roll to charge-off and will do so in the next several months. Past program experience suggests the majority of the drag should resolve by the end of the third quarter. In summary, there are 2 important takeaways: the first, the high reserve rate for TDRs covers off any future P&L risk; and second, any pressure to net loss rates from recidivism in our TDR program will be mitigated by improving recovery rates. Accordingly, we remain confident that we are tracking to approximately a 6% net loss rate for 2018. Let's go to the next slide and talk about our segment results. Starting with LoyaltyOne, pro forma revenue increased 12% to $315 million. Breaking down the quarter, AIR MILES pro forma revenue increased 1% from the second quarter of 2017. Importantly, AIR MILES issued increased 2% compared to the second quarter of 2017, benefiting from increased promotional activity by the Bank of Montreal. BrandLoyalty's revenue increased 34% from the second quarter of 2017 as customer bookings were up double digits from the prior year. As we have talked about before, some bookings are event-driven and the FIFA World Cup helped the second quarter. Turning to Epsilon. Revenue decreased 5% to $514 million. Similar to the first quarter, agency revenue dropped more than 20% due to client losses in the CPG sector from consolidation and budget cuts. Importantly, strong expense management and the favorable shift in revenue mix allowed us to increase EBITDA margins and maintain the same level of adjusted EBITDA as the second quarter of 2017. Lastly, Card Services revenue increased 14% to $1.15 billion, consistent with the growth rate in average card receivables, and adjusted EBITDA net increased 10% to $336 million. Net loss rates improved sequentially by 30 basis points as the recovery rates improved from high single digits in the first quarter to mid-teens in the second quarter. The second quarter reserve rate remains at 6.8% and should drop over the remainder of 2018 as the net loss rates improve to the high 5% range in the third quarter and to the mid-5% range in the fourth quarter. With that, I will turn it over to Ed.

Thanks, Charles. If everyone could turn to Slide 7. I guess I would comment that the second quarter is probably the best quarter we've posted in terms of operational effectiveness in quite some time. Unfortunately, it was a bit overshadowed by the noise that Charles has talked about on the delinquency front that, hopefully, we can help overcommunicate today and in the future. So first on LoyaltyOne. It's been a long road back ever since we had the issues in the program a couple of years ago, up in Canada, where our parliament decided to change the laws around how we run the program. And so we've been building, building, building really over the past 1.5 years. And finally, in Q2, we saw a return to double-digit pro forma revenue and EBITDA. And LoyaltyOne includes both the Canadian AIR MILES business as well as our European-based BrandLoyalty businesses. And so the first return to double-double in quite some time, and it was really -- probably the big news was the key expansion by our largest client or sponsor in the AIR MILES program in Canada, which is Bank of Montreal, which essentially really decided to put their shoulder into it and added a real kicker in terms of additional incentives for people to use the card and use the program. So that was a big vote of confidence from BMO, and we appreciate it. And then finally, the key metric, AIR MILES issued, since we get paid based on the number of miles that are issued, which has been struggling ever since the model came into question a couple of years ago. And it's been a long battle back, and we finally did in fact break above water in Q2 with a couple of points of positive issuance for the first time in 1.5 years, and that's a very good sign. More importantly, as I look at it, the outlook for Q3 and Q4, the momentum is beginning to pick up, and that suggests that we are past sort of the rebuilding stage and have returned now to growth, which is good news. We also announced an additional sponsor of Telus, one of the big telecom players up in Canada. And then in BrandLoyalty, a deal with Kroger to take a look at seeing if our grocery loyalty programs would be effective in the U.S. Okay. For the second half guide, we don't see things slowing down. We see a continuation of strong results from both AIR MILES and from BrandLoyalty that we saw in the second quarter. And actually, we expect the key metric, AIR MILES issued, to actually accelerate in the back half. So overall, I would suggest that LoyaltyOne is back on its feet and doing quite well. Epsilon, as we've guided to, we knew the first half would be soft. Revenue was down in the first half. But again, as Charles mentioned, this is primarily passed through a low-margin agency business. Frankly, we're deemphasizing those types of areas. And as a result, you saw the softness there. The adjusted EBITDA, however, we held for the first half, up 4%, and tracking to sort of that mid-single-digit guidance that we want for the year. We are using some additional dollars from the tax windfall, as we mentioned last year, for additional development in our various digital channels. Second half, we do have decent visibility on mid-single-digit revenue and adjusted EBITDA growth. Both are probably big growth engines. The Auto and the Conversant CRM revenue are tracking to double-digit growth, and that's approximately half of Epsilon. And the other businesses are relatively stable or flat, and that's where we get our mid-single digits. And so we'll be talking more in Q3 as Epsilon begins to contribute, along with the other 2 businesses. All right, let's turn to Slide 9, Card Services. Obviously, it's been another very strong first half of the year, 14% revenue growth. EBITDA growth started soft due to the higher loss rate, and then moved from a minus 4% in Q1 to plus 10% in Q2, and that's really being driven by the loss rates continuing to tick down as we move throughout the year. We've got almost 20% average receivable growth for active programs and 12% reported. I'll get into that in a little bit. And then we also are spending significant dollars in the innovation fund to support various consumer, as we call them, bells and whistles on both the frictionless mobile initiatives as well as building out our consumer deposit platform. Our second half guide is going to be very familiar in terms of the revenue and receivable growth because it will look very similar to the first half. The adjusted EBITDA, again, which includes the cost of funds and the cost of bad debt, will continue to ramp up. So we did minus 4% in Q1, plus 10% in Q2, and we should be in the 20% range in the second half. The net loss rates, which again is the ultimate driver of the credit expense, we guided to a high 6% in Q1. We came in at 6.7%. We guided to a mid-6% in Q2. We came in a little better at 6.4%, and we are currently tracking nicely to high 5s in Q3 and mid-5s in Q4. And as Charles mentioned, that will give us approximately 6% for the year. We believe the noise from the first half is largely gone. And the delinquency rates, which caused the hubbub over the last week or so, in fact, did decline. They were up 60 basis points in April, and they declined to up only 40 basis points in June. Importantly, the early-stage delinquencies have narrowed to up 25 basis points in June, which bodes well for the future. I do want to take 2 minutes to talk a little bit more and circle back to the June delinquency concern that Charles covered. We did guide to a significant improvement in year-over-year delinquency rates during Q2. The actual results did show improvement from 60 basis points over prior year in April to 40 basis points in June. While that was a significant improvement, we had, in fact, guided to even better results. We did not factor in those hardship accounts that still remain neither cured nor written off. This overstated the June delinquency number by roughly 30 of the 40 basis points, and thus, the normal run rate would have been only about 10% -- 10 basis points higher than the prior year. Why do we make the distinction? Well, probably, the simple reason is, these accounts were placed in hardship back in the first quarter of this year, and importantly, they were fully reserved for in the first quarter of this year at a 28% reserve rate versus our less-than-7% rate for the entire portfolio. So they were completely covered off in Q1. This means that the P&L hit was already taken in Q1. And whenever some portion of these accounts flow to write-off and out of delinquency, it's already been covered and expensed. So internally, when we carve out these accounts when we do our internal analysis, since they're viewed as noise, since it's hard to compare a 28% reserve versus a 7% reserve for the portfolio, you got a little bit of apples and oranges. Our internal analysis continues to support loss rates stepping down nicely from the high 6s as predicted to the mid-6s as predicted to the high 5s and then to the mid-5s in Q4 when recoveries have fully ramped. Frankly, I do apologize for the poor communication, and we'll try to overcommunicate going forward on these complicated items. The bottom line, however, is that the credit quality within the portfolio looks quite good and losses are trending down nicely to plan. All right. Page 10 are a bunch of credit metrics that you can look at, at your leisure. What I'd like to do is get into little bit more detail of what we're trying to do strategically with the business, which is starting on Page 11. Let's talk about the first half credit sales. And when you look at the active clients, those are clients that are not in liquidation or bankruptcy or have been terminated. And if you look at just those active clients, our credit sales are up a very healthy 12% during the quarter. Even though the printed credit sales growth was only 2%, it was masked by the liquidating files of Bon-Ton, Gander Mountain and Virgin America. So what's left is obviously doing quite well. Probably the most important item, frankly, that we focus on here is the whole strategic shift off-mall. And essentially, the new clients, and that's really represented by the new clients or the new vintages, those we signed between 2015 and '18, they are up approximately 70% year-over-year, now represent over 1/4 or roughly 1/4 of the entire credit sales of the business, which if you think of the size of it, is fairly remarkable. And what this means is we have made a strategic decision to shift very heavily to where the growth is, and a lot of that is off-mall. We do think that with this shift in place, that this will account for probably over half of the file within 2 years. And as a result, the concerns over the concentrations in our traditional soft good apparel should dissipate as this strategy unfolds. So we're making for a much stronger business going forward, and you can see it behind the numbers. Turning also now to receivable growth. Same story for those accounts that are not in liquidation or bankruptcy. You'll see our active accounts are almost up 20% year-over-year, and that is from the newer vintages, again, sort of that off-mall strategy. They're actually up 110% versus prior year. So again, this strategy of shifting to where the growth is, I think, is something that you will see an acceleration in that as we move forward. Okay. And so now we'll move to, finally, loss rates. Since we're getting near the end here, finally, of this long process of normalizing loss rates, I get the question all the time of, what do you target, what is the right loss rate? And what I would say is, if you went back a long time to sort of pre-Great Recession, we have found that sort of the optimal level to run the portfolio at to generate the sort of 30-plus percent ROEs and above-industry averages in terms of our ROAs, et cetera, that we like to run the file at about a 6% overall loss rate, plus or minus. During the Great Recession era, that's trended up to about 8%. And then post-recession is this sort of over-earnings phase where losses went below trend as there were, frankly, fewer and fewer people who qualified for credit. As people returned to the marketplace, you saw this normalization, which started to take place at the end of '15, '16, '17, and we're finishing it out this year. And all we're doing is we're just returning back to the trend that we've had for the last 13 years. And so we're at the end of that, thankfully. And from a guidance perspective, when we talk about that, we expect pro forma revenue to be up 10% to about $8.5 billion. Core EPS, we expect up somewhere between 16% and 19%. Growth rate by quarters, you'll see the big jump from Q1 to Q2 as earnings move very quickly from 13% to 31% and revenue doubled as well. We expect, as we move into Q3, another very strong performance. If you think of it, earnings in the first quarter were about $4.50. And then we move into second quarter, it has a $5 handle. As we move to the third quarter, it will have a $6 handle. And again, that's the acceleration that we've been talking about. On the revenue side, a couple of things going on there. We're running a bit softer than the original thought, but we expect 0 impact to profits or cash flow. Essentially, 2 factors: Card Services Bon-Ton, which was an $800 million file. Usually, these accounts go into bankruptcy and then they're prepackaged and come back out. Bon-Ton went into liquidation, which means everything must go. And that, severely, obviously, the runoff in the accounts occurs much faster, and that hits our top line. We've already offset it in the expense line, so it shouldn't really change anything from the cash flow or profits. At Epsilon, we are taking a proactive approach there. There are certain verticals in the agency business, frankly, primarily the CPG or the packaged goods vertical that are just real tough verticals right now that we are deemphasizing because there's just no margin there. And as a result, you don't have a cash flow impact, but you do have some low-margin revenue that won't be flowing through. Overall, I think the move at Epsilon is a good move in the sense of we're focusing on where the good margin is. And obviously, from a core EPS perspective, unlike the past couple of years, where we're looking at a really big back end, right now in the first half, our core EPS is already up 22% and our full year guidance is only 16% to 19%. So we don't have the big back-end bet that people get concerned about in prior years. And so I think both Charles and I would say we're increasingly comfortable with the visibility of the guidance, and we'll probably tweak it to the good, hopefully, in Q3. And with that, I think we'll go ahead and open it up for questions.

[Operator Instructions] Your first question comes from the line of Sanjay Sakhrani with KBW.

Charles, on the TDR noise, can you clarify how we're thinking about the provision impact going forward? I respect that you guys said that you've taken most of the provision hit already, but I heard you also talk about the offset from higher recovery rates. So looking ahead, when we think about your charge-off rate averaging about 6% for the year, is it true that then the provisioning impact going forward would be less than that because you've already provisioned for some of those losses?

Yes. So there's a couple of ways to look at it. I'll start off with what you're saying, which is yes, the provision's already set up. As Ed and I both talked about, we put up a 28% reserve against those accounts. So we feel more than comfortable that's going to cover us. The second point to really think about with the TDRs, we have good experience estimating how many ultimately fall out. Where we're sometimes off is how quickly they'll fall out of the program, which is what caught us a little bit off guard. So history would tell us that we're in good shape in terms of our estimates for Q3 loss rates. We're guiding to high 5s for Q3, that's up 40 basis points year-over-year. That should more than accommodate the noise coming through from the TDRs, and then you see the further improvement in the fourth quarter. On recovery rates, we talked about high-single-digit recovery rates Q1; mid-teens, Q2. It should drop -- go up to close to 20% Q3 and could be north of that in Q4. So that's part of the reason we're comfortable with our loss rate guidance as it could put a little pressure on your gross principal loss rates, but the better recovery rates mitigate it and keep us on track for that 6% approximate loss rate for the year.

Okay. And then second question is on leverage. You mentioned that it fell to 2.5x, and I think it should drop to the low 2s by the end of the year. I guess, when we think about the go-forward expectations on leverage, how much -- or how -- what's the level that you get to where you maintain all the optionality you need for a variety of different reasons, but also can be more aggressive on capital management? Maybe you could just talk about the go-forward expectations on capital management, and especially next year.

Yes, it's Ed. I think that, as you mentioned, we will continue to participate on the buyback front, also supporting capital in the Card Services area. So even doing that, yes, we're going to get close to about a 2% leverage -- 2x leverage by the end of the year. That's pretty conservative for us. We tend to fluctuate between 2 and 3. What we want to do is we want to keep, as you called it, the optionality going forward for whether there may be a significant file out there that requires additional capital or any other items that we want to deal with. So I think, Sanjay, I don't see us really heading back to 3. I think the 2 to 2.5 certainly seems very reasonable and gives us the type of dry powder that we could put to work very quickly, either on capital management, as you call it, or if there's a couple of big files out there.

And I'm sorry, are there a couple of big files out there as you're looking at the pipeline?

Yes.

Your next question comes from the line of Jeff Meuler with Baird.

Just trying to -- I would love to get a number just to better clarify the hurricane impact [ versus ] the non-hurricane impact. Outside of the FEMA zones, roughly what are the delinquencies up on a year-over-year basis in either Q2 or in June, if you have it?

So that's what we tried to provide you on that one slide, which is 30 basis points of the 40 basis points is attributable to the FEMA zones, which went into the TDR. So we would tell you that the natural migration or a run rate is that it would [ cost ] about 10 basis points up year-over-year. You can really see it in a couple of ways, as Ed talked about, the improvements in the early-stage delinquencies; if you look at some of the new account information in the trust, you'll see the new accounts for '18 are aging to loss -- or moving to loss lower or slower than what they were in '17 and '16. The natural trends are beneficial in terms of delinquency. You'll get a little bit of noise from the hardship program, as we talked about. That's the overall. The profile is meeting up to the expectations we had coming into the year. So about 10 basis points would be what we said it would be at June.

And I guess, the 25 basis point early stage, I get that that's improving. But on a 6%-ish rate, 25 basis points still sound somewhat material to me. Is there -- I wouldn't think there would be a TDR impact in there. Is there? Or just any color on what's going on that it's still up 25 basis points?

Well, I'd put it this way: There's substantial improvements from where it was. I wouldn't attribute very much of that to TDR. It just takes a certain period of time for things to work through, but we'd tell you it's heading in the right direction.

Okay. And then just finally, can you just clarify what tax rate you're assuming in the full year guidance since there was a Q2 benefit?

So for the first half of the year on core, it was a 24.2% rate. We have guided initially starting the year around 25%. So it could come in favorable, which would definitely push us toward the higher end of our core EPS estimate. Did I say 25%? I meant 24%, my bad.

Yes. We guided to 25%.

We guided to 25%. It could come in around as low as 24% for the year.

Your next question comes from the line of Ashish Sabadra with Deutsche Bank.

My question was about the recovery rate. The consumer is a bit more levered today than he was -- he or she was 2 years back. Does that affect any of your assumptions around the recovery rate even with ramping up in-house recovery?

No. The recovery rates themselves, much like how we spool up the various vintages, all has to do with -- the way the curve works, Ashish, is once we brought it in-house, that took time, then you needed to get your hundreds of collectors ready to go, train them up and then obviously they become more and more productive as time unfolds. And that's why we're in the, I guess, high single digits in the first quarter. We ended the second quarter around 15%. And the way the curve looks compared to what we did right after the Great Recession, it's almost bang on, so that as we head into Q3, Q4, it's really less to do with the consumer and more to do with having the internal group reaching full productivity and run rate. And so our recoveries should continue to creep up, even without factoring in the consumer.

That's helpful. And maybe a question on Epsilon. So the headwinds in agency, what -- that affected the first half of the year. But as you look at the second half, are there anything else to watch out for? So you have -- the Auto and CRM is definitely areas of strength. But are there any potential unexpected things that could crop up which could affect the growth profile in the back half or going forward?

Well, that's certainly a good question. I mean, we've been surprised in the past, as everyone knows. So we're being fairly cautious in our commentary around Epsilon's back half. What we're looking at is really the book that's built right now. And right now, it looks like the softness on the CPG side in agency is passing -- or sorry, it's going to hit its anniversary. And as a result, just based on that, that should pop us back to that mid-single-digit growth rate. Other than that, we're not going to try to figure out, hey, if everything goes great, we could be higher than that. We're just saying, just based strictly on what hit their anniversary levels, Epi should pop back to sort of mid-single digit by Q3.

Your next question comes from the line of Jason Deleeuw with Piper Jaffray.

The receivables growth for the file, it looks like it's now going to be about 12%. So I think it was originally mid-teens for the year. Can you just walk through kind of the details for why the growth was coming in a little bit slower?

Yes, sure. I think that, again, we tried to lay out the difference between sort of the true business itself, the active clients. What you're looking at is almost 20% growth in the active client portfolio itself. The difference between the 19% and the 12% is nothing more than the liquidations that have taken place. As we said, Bon-Ton is a big one. Bon-Ton was $800 million. And initially, based on what most retailers do, we assumed a prepackaged bankruptcy. They'd come out, and it would be a very slow attrition in the file. They went into liquidation, and -- as did the other two. And as a result, those files are going from $1.5 billion in receivables to $800 million by the end of the year, and that causes the reported number to be different from the active number.

And then I just want to touch on the strategic shift to off-mall and just seeing the credit sales and the receivables growth for the new clients, the 2015, 2018 signings, and then just seeing kind of the active clients’ category and the strong growth there. What should we think about the receivables growth for the portfolio after this strategic shift is complete? And then is there any help you can give us on the addressable market opportunity? Is it bigger or smaller than what you had before? And just any commentary on how ADS competes there? Do you feel like you compete there as well or better or worse? Just any color on that.

Sure. No, it's a good question. And the shift in the strategy is, again, your -- we have targeted a 15% growth rate in the file for the next several years. We believe the addressable market, which would be the traditional market, plus, frankly, a lot of new verticals that are shifting dollars out of traditional TV spend and radio spend and everything else and into this data-driven, personalized marketing, essentially trying to get a handle on the individual customer. We found that the addressable market, in our opinion, has actually grown. So if you have the traditional sort of sandbox that we play in, and then you add into that certain verticals that didn't even exist before, the big pure e-com players that are growing so fast and want us to take all that SKU-level information and slice it and dice it and run a million marketing programs, the wayfair.coms, the Build.coms, you'll see a number of those announce more in the fall. That's a whole new market for us, which I think is good. Additional verticals such as the IKEAs of the world and the Viking Cruises of the world, which, frankly, we never had before, it wasn't really something that they had a huge interest in, all this personalized marketing. And so those are new verticals as well that are coming in. And then finally, with the so-called -- I don't know if I'd call it the death of the department store, but certainly the troubles at the department store, is you're having a number of these very prominent brands that use the department store as their base, and therefore, were captured in the department store cards, which, as you know, those department store card portfolios are way too big for us. And so these brands are now not needing that base anymore and they're establishing boutiques with a big e-com play, and so there's probably 12 or 15 clients there. And then finally, another sector would be the beauty sector such as Ulta Beauty, which, frankly, they weren't even around before when most people were going into the department stores and to the beauty counters. And those things are growing like crazy. So I think the overall market, which we initially thought was somewhere around $35 billion, is probably more in the $45 billion to $50 billion at this point, I think, which sounds surprising. But the fact of the matter is, dollars -- everyone needs to replicate a certain business model that exists out there at a very, very, very large player, which is, I need to know who you are and what you bought down on the SKU level. And unless you have that ecosystem, you can't do it. And so that's what we're doing for all these retailers.

Your next question is from the line of Andrew Jeffrey with SunTrust.

Just as a follow-up on Jason's question, it's helpful to see the active portfolio performance. Can you compare that to what you have seen in the past -- in past vintages as far as how those -- how newer customers spool up? Is this a faster ramp? And does mobile and do other initiatives contribute to that? Or is this about right for, say, prior 3 years' vintage performance?

I guess the way we would answer it, Andrew, is you're correct in your assessment that the new programs we're adding are much quicker in ramp. If you go back 8, 9 years ago, Ed and I would tell you that a new Private Label program would go from $0 to $50 million in receivables over a 3-year period. We now are adding Private Label programs that are going from $0 to $200 million plus in receivables over a 3-year window. So we're definitely bringing in a different group of retailer. It doesn't have the same soft goods bias. It's ramping much quicker, so the growth profile is much nicer than what we historically would have said for a new Private Label start-up.

Yes. I would also say, Andrew, and I think it's a great question, this cuts across all the verticals. So it's not just the clients being somewhat different. There's a sense of urgency out there like you wouldn't believe, right? It's the fear of, I've got to catch up, I've got to get this ecosystem of my own where I can recognize who it is down to the individual level, what they purchased online, offline, and find that price point that would trigger one extra sale per year than otherwise. And so the sense of urgency which probably comes with the shrinking life or sort of career life of the various execs, probably has something to do with it as well.

Okay. And if -- just as a follow-up, if an investor asks me or asserts, "Hey, it looks like Alliance is seeing greater impact from the hurricanes from some of the forbearance in the TDRs than other lenders and what's going on there," how do you address your -- the impact you're seeing versus maybe what other card-based businesses would be seeing in those FEMA zones?

Yes. I think, frankly, I think it's a fair question, and it's -- a lot has to do with how we approach our clients. We take everyone who is in that FEMA zone, and the big question is why are we still talking about hurricanes this late in the game. And what we did is we took everyone in the FEMA zone, we froze those accounts October, November last year. We started to unfreeze them December and January. And then once we could reach out to these folks, those who needed it were put into their hardship programs, which was first quarter. We reserved against it, and now they're going to ping-pong around between curing or writing off. To us, it's a tail that exists out there that needs to bleed off. Does it change our P&L? No. Does it change what we are going to expense? No, it's been already taken care of. So maybe that we're all Private Label, and Private Label got hit a little bit harder as opposed to co-brands or general purpose cards, but -- or it could be the way that we do a blanket freeze on these FEMA-related, if you want to call it, a bit more consumer-friendly. But at this point, it's noise that's in the delinquency numbers. Already been reserved for, so it's really a nonissue at this point.

Your next question comes from the line of David Togut with Evercore ISI.

Could you quantify same-store credit sales in Q2?

Well, you'd have in there, David, a lot of the ramp-ups of the newer vintages as well. So as you know, when we're starting a program, the retailer could be doing 2% growth, but we're ramping up a card program, so it could be 20% growth at those retailers. And then some of the more mature retailers, they're probably doing -- we're probably doing, I would say, 2% or 3% comps on that. You throw in the new retailers that are ramping up very, very quickly and you're moving into the -- probably into the high single digits, something like that.

Got it. And then just as a follow-up, with the all-important holiday retail season around the corner, are there any specific online marketing initiatives you can implement at your -- let's say, your more traditional mall-based retailers to help them get through Q4 successfully?

Yes. That's a great question, yes. We are -- I can tell you, there is a huge amount of interest obviously from that client base to make sure that the holiday season is a good one. And so we are loading up, and we're getting in front of it now, a lot of these programs that are, if you want to call them, follow-up digital-type campaigns, so that if you went into the store or you bought a significant item for a couple hundred dollars, we will use the SKU-level information and behavioral information and we will find a way to you, come back to you, most likely in an online capacity, with an offer for an accessory at the right price, reach you on the right channel. And right now, our retailers run about 20% of their sales are still -- are online. We're running over 40% of our sales are online, meaning that the vast bulk of what we do are these follow-up type campaigns that's driving the incremental sale to the retailer. So expect -- if you haven't been flooded yet during last holiday, you can assume that we'll be flooding you this year.

Your next question is from the line of Darrin Peller with Wolfe Research.

Just the first question is on -- just to be clear, there's no new noncore or problematic portfolios you see right now that will become noncore or nonactive as far as you can tell right now? I just want to verify that, first of all.

That's correct. I think Bon-Ton is, by far, the big one in the tent.

All right. And then the timing that we should expect, the -- everything to be really clean with TDRs passed through at this point, in your opinion, would be late third quarter into fourth quarter that we get clean results. I just want to make sure we all know what to expect exactly.

Yes. Darrin, history would say that by the end of the third quarter, it should pretty well have cleared through.

Okay. All right. And then just let me shift to more of a strategic question. On the Epsilon side of the business, I think there's some assets there that aren't performing as well as you would've expected going back a couple of years ago. And so when you start with Epsilon, are there any strategic areas that you could potentially prune or you think might need restructuring further that could either be sold off? Just if you could touch on that and then maybe follow up with just revisiting the time line in your mind and the board's mind around how long you'd be giving yourselves and the stock and -- before you decide to actually execute on other strategic initiatives for the whole company.

Yes. Obviously, I need to be a bit careful in how I position it, but needless to say, I think that Epsilon's overall performance has been softer than anticipated over the last several years, frankly. And I think that there have been discussions at the board level that started a couple of years ago. I can tell you right now, they have picked up fairly dramatically over the last year. And so from that perspective, we are not sitting around with hope as our strategy. And we are looking at all the assets of the company on now a more intensive basis. And I think that by -- as we move into the back half and we look at the performance, if the performance is there, that's great. If the performance is only modest, my guess is there'll be more pressure to do something. That's probably about as far as I can go.

Your next question is from the line of Vincent Caintic with Stephens.

And I also want to -- thanks for the details on the card sales for the new and total active clients, that's very helpful. And I notice you've been getting some account wins, and then conversely, that there's been some issues with peers in terms of retailer accounts and lots of switching of retailers with new card partnerships. I'm just wondering if you could discuss the competitive environment. And then also, when you win accounts or when you renew accounts, how are the margins upon renewal and for accounts that you might win going forward?

Yes, it's a fair question. I think that there's really sort of 2 markets that are out there. One is the store cards that are co-brands, and that means the Visa, MasterCard, big programs. There's probably, what, 2 dozen very large portfolios that are multibillion-dollar portfolios that are out there, and they're kind of up for grabs every time there's renewal talk. And frankly, those are the ones that sort of the big players, the big banks are vying for on an ongoing basis, and my suspicion would be that the competitive level is quite high. What we have done is we have sort of confined ourselves to almost exclusively Private Label, which is less attractive to the big banks because the balances are so small. If it's a $500 balance versus a $3,000 balance, that's of less interest to a big bank to move the needle. And that has proved to be very effective over the years. Also, 90%-plus of our new clients are starting from scratch, which means that you need to have had a number of vintages that you've signed every single year to keep that ramp going, and that requires a lot of patience and it requires bringing a client from nothing to significance over a longer period of time. Again, that doesn't really move the needle at the big banks, and that's our sweet spot. If it is a co-brand, most likely, I would say, 85% of our co-brands are accommodations to clients who already have a Private Label card, so that maybe there's an extra 5% or 10% of the customer base that would like the co-brand option as opposed to just Private Label. But since that's the case, we already have the Private Label. We're going to get the co-brand. We don't run into the tough competition that I think is out there in the general purpose card arena. So long story short, as long as we can stay in our sandbox and focus on what we do, then I think we're in good shape. As soon as we start venturing into these multibillion-dollar co-brand files, frankly, I just don't think we're going to be competitive. It's just our uniqueness is more about nurturing an account from 0, getting it going, getting the data going, getting the campaigns going, layering on the Conversant product, which allows us to do prospecting on behalf of the client, et cetera, et cetera. So it's just a different world right now. I don't know whether it's better or worse. It's just different.

Okay, helpful. And just to put a finer point on that. So the economics that you're getting off of the renewals, on the new relationships, those are relatively unchanged, would you say?

Yes. I mean, I would say -- obviously, I can't get into specifics. But what are we running, Charles, on our ROEs, 35%?

Yes.

About 35% return on equity. That would be a good figure to use.

Perfect. And just maybe switching gears for a quick one. So this conversation about the TDRs related to the hurricanes, could you actually size -- actually, how big of an impact -- or how big of your portfolio is this actually because if it's all of the TDRs for these hurricanes, but it's only, I don't know, 1% or 2% of your portfolio, it might not be a big deal. So if there's -- if you could size up just how much of an impact this is to the actual portfolio, that would be great.

Yes. So what you can do, Vincent, you can go to our first quarter 10-Q. There's -- I think it's Page 17, where we discussed the modified receivables. It delineates the number of accounts in restructuring. It delineates the amount. We've not put it up for Q2, but in the first quarter it was less than 3% of our overall AR. That's where you can see the dollars quantified, and you can see the reserve rate at the end of Q1 was 24%. We have disclosed that it's now 28% at the end of Q2. But that would give you the best information surrounding that, Vincent.

At this time, we have reached the allotted time for questions. Your last question will be from Dan Perlin with RBC Capital Markets.

So my first question, Ed, is just what are the reconciling items for you to only do mid-teens in the third quarter? Because when I look at it and hear everything you just said and you just put up 31%, and I appreciate some of that's tax, but a lot of it was actually outperformance on a couple of units. And all you're saying is charge-offs are going to get better. Loyalty sounds better. Epsilon sounds better. Tax likely to stay similar to better. So I'm just trying to understand what pressures this in the third quarter to make sure I understand the cadence of how this year should play out.

Yes. I think that -- and I'll let Charles jump in. But directionally, you're right in the sense that you started with Q1. You had the acceleration in Q2 as losses came down and the LoyaltyOne segment hit the ground running. As we move into Q3, you're going to have losses come down again and you're going to have LoyaltyOne continue to contribute, and now you're going to have Epsilon. So it would seem that Q3 should even be faster. The fact is, from a dollar perspective, you did about $4.50 in Q1. You did $5 in Q2. You'll do a $6 handle in Q3. So from a dollar perspective, you're certainly getting that type of acceleration. The only difference is that loss rates last year in Q3 were lower than Q4. Whereas this year, Q3 will be higher than Q4. I think in Q3, Charles jump in, I think we had a number of sales that we did that helped lower the Q3 rate last year.

Correct. So last year, the loss rate was 5.5 in Q3. It went up to 6 in Q4. So it really comes down to reserve methodology. This year, we're saying high 5s in Q3, mid-5s in Q4. So you just kind of flipped the reserve pattern from the prior year.

Okay. So that cadence is a little bit off from what we saw last year. So if we model that up, we have to make sure that the third quarter is trued up for that provision expense. Is there optionality for you guys to have a reserve release later in the year, given maybe over-provisioning, given the results from the hurricane stuff? Or that's just locked in and that doesn't change?

I wouldn't necessarily call it a reserve release because you're still growing your receivables and you're putting up an allowance for it. You could see though the allowance rate fall. So we were 6.8% in Q1, 6.8% reserve in Q2. Based upon the trends we see in loss rates, I would expect the reserve rate to drop Q3 and to drop a little bit further in Q4. So you're still putting up a build, it just would be a lesser build that what you would have had otherwise.

Right. And then just one last one for the sake of time. The operating expense leverage on your ARs in the fourth quarter, my suspicion is that could be pretty significant, but I just wanted to quantify that a little bit. If you could help us, that would be great.

Yes. That's one we probably need to talk about offline, Dan. We can just work with your model at that point.

Okay. I think that's it. So thank you, everyone, and we'll catch up to you next quarter.

Thank you. This concludes today's conference. You may now disconnect.