First Horizon Corporation (FHN) Earnings Call Transcript & Summary

Good morning, and welcome to the First Horizon IBERIABANK Investor Community Information Session. [Operator Instructions] Please note that this event is being recorded. I'd now like to turn the conference over to Ms. Ellen Taylor, Head of Investor Relations. Please go ahead.

Thanks so much, Nick, and good morning, everyone. We really appreciate you joining us today. I'm so excited to be a part of the First Horizon team. Needless to say, this has been quite a challenging year, and certainly, CECL implementation has been an additional factor that's added to those challenges. So given that the First Horizon IBERIABANK MOE is one of the first transactions under CECL, we thought it might be helpful to have some experts provide some additional clarity. So as a result, we're really pleased to have Greg Norwood and Jonathan Prejean from Deloitte with us today. They are planning to provide some prepared remarks, and then they will open it up for questions. You can find the presentation they'll be referencing on ir.fhnc.com. And then I, of course, need to remind you that this presentation is not intended to cover anything related to our MOE or any historic or pending transactions, and Deloitte will not address any questions specific to any past or pending transaction. So with that, I'm going to hand it over to Greg Norwood.

Thanks, Ellen. And on behalf of Deloitte and especially, Jonathan and myself, thanks to First Horizon and IBERIABANK for hosting this session. We are happy to share some thoughts about a key aspect of CECL. So let's turn to Slide 2 of the deck we provided. Today, we will look at PCD accounting or purchase credit deteriorated by reviewing, first, the definition and ways of identifying such loans. We will also look at the accounting at closing. We will look at what the expected future impact could be based on consummation assumptions. And finally, we will look at how measurement could change based on changes in the future. As Ellen said, we will go through the slides with prepared remarks, but we'll leave ample time for questions at the end. Before we get into the discussion, let me be clear. Our comments today are not reflective of any specific transaction, and examples are for illustrative purposes only. Also, the amounts in the example are purposely simple and not intended to portray any real-life examples of measurement. Rather, they are simple examples to highlight the basics of the accounting measurement and how they can change in the future and the impact to key financial statement reported amounts. I will walk you through the definition of PCD loans and some possible selection criteria. Then Jonathan will take you through the accounting construct and highlight the impact of gross up and double count, terms you certainly heard in the marketplace. Then I will walk you through how the accounting can affect the reported allowance, provision and NII, given how the future may play out. With that, let's move to Slide 3. Before I start, much has been said about why didn't FASB just treat all acquired loans as PCD. Given where we are today, I will not try to speculate, so I will move directly to what the rule does say. First, we should remind ourselves that CECL is a principle-based standard, not prescriptive. FASB developed CECL using broad concepts and no detail "how to" requirements, and that is certainly true in their guidance for PCD accounting. So while FASB was clear in the accounting and the financial statement geography for PCD, they did not give rules on how to identify PCD versus non-PCD. So let's focus on the definition, and importantly, what's not in the definition. The definition is more than insignificant deterioration in credit quality since origination. Let's start with credit quality, which is the measurement criteria. Credit quality is a relative concept. PCD is not trying to identify bad loans, which is why FASB did not define the identification concept as expected loss or risk of loss or some other concept of loss measurement. Importantly, FASB went out of their way to say PCD was not the same as legacy PCI definition, which for shorthand was probable default. A simple way to think of credit quality is on the consumer side. For example, 2 borrowers with FICO scores of 790 and 690 may both meet the definition of a PCD loan if their credit quality declines since origination. So when we think about PCD loans, they could be both high-quality credit or lower-quality credit, and the classification is determined by the degree of change. So now let's look at the degree of change required to be PCD defined as greater than insignificant. This is highly judgmental. FASB provided no guidance as to how to define the level of change. And like other areas of CECL, banks may come up with different thresholds. But it's clear, FASB said that PCD should be a much bigger population than PCI. Now let's talk about the accounting geography. Let's start at the end. All loans, whether PCD or non-PCD, will have an allowance based on CECL principles after consummation and day 1 accounting. And going forward, all changes in expected loss estimates for both types of loans will be reflected in the provision line on a quarterly basis just like originated loans. So if credit improves on the acquired book, the benefit will be reflected in a lower provision expense in future periods. Also charge-offs for PCD and non-PCD loans will go through the allowance just like originated loans. In the end, FASB wanted all loans whether originated or purchased to be treated the same. Now let's talk about the charge for the consummation and day 1 allowance as reflected in the financial statements as well as interest accretion. For PCD loans, the allowance amount is grossed up. And what this means is there is no charge in the income statement. Rather a CECL allowance is recorded at consummation and the difference between the total fair value mark at consummation and the CECL allowance is accreted into interest income in the future. For non-PCD loans, the amount of the CECL allowance is charged through the buyer's income statement in the period of closing, which is referred to as the double count. And the entire fair value mark, both interest and credit components, is accreted into interest income in the future. Jonathan will emphasize the accounting also when he goes through the examples, which will help put a picture to my words. So now let's move to Slide 4. The PCD selection process has 2 discrete steps. First is to identify a bank's view or processes for assessing credit quality. This would likely vary by bank and could include many different processes. Slide 4 lists some possible credit quality triggers or buckets. For reference, FASB included an example of criteria and other perspectives in standard. For those so inclined, it is example 11. FASB's obvious criteria included nonaccrual loans and delinquent loans, but also included loans that have been downgraded or whose credit spreads have widened since origination. And the second step is for each criteria identified, the bank must determine the threshold for measuring greater than insignificant. This threshold can differ by category and by relative origination credit quality. And what I mean by that is that each one of these buckets may have a different measurement and a high quality, like in our example of FICO, there might be different measurements for changes relative to individual borrowers. Before I turn it over to Jonathan, let me highlight one last perspective in determining PCD loans. Like any credit quality assessment, both idiosyncratic factors and the general economic environment can cause borrowers to suffer credit quality and both should be considered in this process. Now I will turn it over to Jonathan, who will begin on Slide 5.

Thanks, Greg. So if we move to Slide 5, we're not going to -- this is not an accounting class, so we don't want to get too accounting technical with you. But we think it's helpful for you to see how these transactions will be recorded. So that you fully understand what Greg was talking about when it comes to the day 1 accounting and then the impact from the subsequent accounting. So the first example on Slide 5 shows the accounting for PCD assets. So in this example, we have a $10 million balances outstanding. There's a 2.5% rate mark. So that's referred to as the noncredit discount. And then we have an allowance, which is determined under CECL, and we'll approximate the credit mark associated on -- with the fair value calculation. As you can see here, they talk about a gross up. The standard talks about a gross up. But really, what you're doing is your -- if you paid $9.25 million for this portfolio of loans, a component of that discount is the rate component or the noncredit component, and the other component is the credit component. So as you record it, you'll split it out. It will actually be recorded on your balance sheet. For those who remember the PCI calculation, the old 0 3 3 calculation, that credit mark or the part you don't expect to collect would not be recorded in the financial statements. And thus, any charge-offs associated with that $500,000, in our example, would not be -- would not flow through as a charge-off in the bank's credit metrics going forward. Now with the introduction of PCD accounting, that $500,000 will be recorded as an allowance and will act as allowance going forward. Thereby, any improvements in credit will be a reduction in the allowance. Any deterioration in credit will be an increase in the allowance through the provision. And of course, any charge-offs will flow through the credit metrics of the organization. So if we move to the next slide, this is where the accounting for the non-PCD assets. And as Greg mentioned, there's a uniqueness to this one because there is a so-called double count associated with this one. And I'll explain what that means. So in our example, we have the loans with a balance of $5 million, unpaid principal balance. And the fair value is $4.9 million. That fair value is comprised of $25,000 of rate mark or noncredit discount and a $75,000 credit mark. Now as you'll see on this one, it is not identified as an allowance. Because on day 1, when you acquire the loans and do purchase accounting, you would record that the combination of the $25,000 and the $75,000 or $100,000 would be your discount that gets accreted up over the life of the loan. On day 1, we'll refer to it as day 1.5. So upon acquisition, you would also establish a CECL reserve. In our example, the CECL reserve in line 4 and 5 approximates the fair value credit mark. And so that establishment of that allowance will go through the provision expense. So it will be a P&L impact on day 1 or day 1.5 of the acquisition. And therefore, you can see the double count. You have the CECL allowance of $75,000, along with the credit mark of $75,000. This is different than what was historically done prior to CECL, where the credit mark was considered -- was booked similarly, but you didn't have to book the additional $75,000 CECL allowance. And you wouldn't book an additional allowance until your estimated losses actually exceeded the existing -- the purchase discount. So the other component is the 4 $100,000 will be accreted through income. So that $75,000 will come through over the life of the loan and offset -- although not at the same time, will offset the $75,000 initial impact to retained earnings through the provision expense. So if we move to Slide 8. Very basic, the allowance that has discounts associated with the portfolio will be removed. The loans, both the PCD and non-PCD, will be initially recorded at fair value. Where the difference will be is that on the PCD assets, you will have the fair value broken down between a credit and noncredit component. And for those following along, this is on Slide 7. You'll have that broken down between a credit and noncredit component with no further impact to tangible book value. On the PCD -- on the non-PCD side, you initially recorded fair value. The combination of your credit and noncredit mark will be your loan discount, which will be accreted in the future. And then you will record an additional allowance associated with that portfolio. That will be another impact to tangible book value. So that's how it's set up on day 1. Now Greg is going to walk through what happens both from an expected -- so your expected performance of the portfolio and then when your expectations change so that you can see how that happens. The one thing I do want to point out is that after this initial recording of the loans, whether they're PCD or non-PCD, they will be accounted for in the same manner going forward. So there will not be a distinction between acquired PCD, acquired non-PCD and originated loans. All will look the same going forward. So Greg, do you want to go through the P&L impact?

Sure. Thanks, Jonathan. So moving to Slide 8. Again, we tried to get this very simple so that we could focus on the geography of the accounting and how that can change rather than the actual numbers themselves. But as you can see on the top of Page 8, the total discount of $350,000 is the sum of the numbers from the prior page. The charge-offs, again, are the sum of the expected credit mark, assuming that there is no change. And as you see on Page 8, the accretion is fixed over the life of the loan. And as we'll see in other examples, only changes relative to prepayment, and this is different than the PCI accounting. As you can see, the "double count" is reflected in the $75 mark, that's on day 1 and in year 1. But this basically shows that the accretion of the total $350,000 will be recognized over the life of the loan. And that the allowance in this example, assuming perfect foresight, is reduced each year by our pro rata share of the charge-offs. So that is the basic construct of what the portfolio would perform, if all of management's assumptions prove out perfect and there's no changes in overall credit quality. Let me now turn to Page 9, where we have 2 examples to try to very simply identify what can happen in different forms and different fashions and different timing. The top half of the page is faster-than-expected prepayments. So the same kind of activity in years 1 and 2. But in year 3, at the end, you've experienced the proportionate charge-offs for that period, but all the loans prepay. So what happens then is you accelerate the accretion from years 4 and 5 at the time the loans prepay. And because there's no future losses, the remaining allowance for loan loss under CECL is reversed. So you can see that, that can have significant impacts. And that prepayment, importantly, will impact both the accretion timing as well as the provision/benefit. If we go to the other type of example is where credit deterioration happens post the balance sheet date of recording the acquisition. So here, you see in year 3 the company has concluded that this portfolio needs an additional provision of $25. So as you can see, the accretion remains the same in years 4 and 5. But because of the higher expected credit losses, you'll see charge-offs increase to absorb what is now a $600 estimated loss against the portfolio. So again, 2 examples to really try to highlight that the prepayment assumptions and the reality of how loans prepay or charge-off will impact both the accretion and the allowance balance and changes in credit quality of the portfolio or importantly, in CECL, changes in credit quality of the environment or the reasonable and supported forecast can both impact future provision and expense. So with that, Ellen, let's turn it back over to you, and happy to answer any questions relative to the concept of PCD accounting.

Sure. Operator, can you give the Q&A?

[Operator Instructions] Our first question comes from John Helfst of Voya.

This is a really stupid question. But why -- what's the advantage of having non-PCD and PCD? Like it seems like -- couldn't it be easier just to have one classification? Like it's like -- it's almost like we're driving with metric and miles per hour. I don't understand what's -- I feel like moronic asking this question, but like what's the added advantage to having both technology or both classifications?

That's a great question. I assume you're meaning from a FASB's perspective as to why they wanted 2 different classifications.

I don't want to point fingers, but like from a user standpoint, like it is more complicated, right? So like I'm actually following you along, but then if I try to turn around and explain this to someone I know I couldn't do it very well. So that means I don't really understand it as well as I would like to. So what's the added benefit to the user of having the 2 buckets, in your mind? Again, I'm not trying to blow up FASB out of the water here, right. What is the advantage?

So that's a fair question. And I would say -- I wouldn't say there's an advantage to having 2 separate classifications other than, historically, we've had 2 separate classifications. Although admittedly, what we're looking at today is a lot different, right? Because historically, you were purchasing credit impaired loans. Now we have credit deteriorated loans. I would say the benefit of what happened was the day 2 accounting is the same. So whether -- we could debate whether or not we need the 2 separate classifications, and we have what we have. I think the benefit, however, is that going forward, there -- the accounting is the same. Whereas in the past, you had to understand what -- on the PCI side, what wasn't expected to be collected on day 1. If that improves, how does that impact interest income, et cetera, et cetera. Now after the acquisition is booked, the accounting for any portfolio, whether it's originated, acquired PCD or acquired non-PCD, that accounting will be the same. And so you're really looking at what's the impact on credit that's going through the allowance/provision. What's the impact on interest income relative to speed of prepayments, et cetera? So I think the benefit is really the day 2 that they've aligned on a date -- on a go-forward day 2 accounting as opposed to the initial classification, which I understand your point, but they -- I guess, they wanted to keep something of old, but still want that to align on a day 2 accounting.

I think, again, one person's opinion, when you think about it is when I talked about idiosyncratic events causing PCD classification, I think if you look at a portfolio that would be non-PCD, a lot of those would be very recent originations and recent relative to the environment. But also, if you think about it, if it's a relatively new -- newly originated loan, in which case, there's just not enough time for deteriorate, those would be reflected as a debit in the income statement just as though the buyer had originated. So there's an element of if it's something that's newly originated or just very recently originated with no change, then they wanted that to mirror originated account. Again, I said there was a lot of debates about this back and forth, and FASB just continued to believe this was the right measurement standard. But one of the things that we think is important is to try to demystify this as to just really understand how the numbers flow through the provision line, how they flow through capital and how they flow through the NII line.

Okay. I mean, I get the day 2 elegance, but I would just think like you could solve, just reclassify everything day 1 and just wash it up. I don't know. Honestly, like I've gone back in time and read FASB bulletins like ASC 001, 002, 003. And frankly, the old accounting is confusing to me as well. So I'm not hating on CECL as -- I'm just hating on the whole accounting construct that's unnecessary complicated, I feel like.

Our next question comes from Will Nance, Goldman Sachs.

Could you maybe talk a little bit about the PCD versus non-PCD selection process? And what kind of delineates the 2? You touched on it a little bit, but I guess, particularly in an environment like we're all in where you would think every loan in the industry has deteriorated somewhat, how do you kind of set that demarcation about what qualifies as PCD versus non-PCD? How much flexibility is there to do more classification as PCD versus non-PCD? It would seem like accounting incentives given the double count would kind of incentivize you to do more PCD versus non-PCD, if kind of given the option. So I'm just kind of curious how much flexibility there is to kind of toggle between the 2 when there has been some kind of exogenous event like we're all living through?

Yes, great question. This is Greg. And when I think about it, maybe a way to kind of layer through that is there's a lot of flexibility to address one of your premises, and banks will do it different. And it certainly does look at what are the processes that the bank uses either in benign times or in times that we're facing today relative to monitoring credit quality. So as a bank looks at this, we would expect that they would really focus on what are the management routines within the organization? How are they looking at it? And if management routines are heightened because of an environmental concern across the macroeconomics, then those factors would be something that would make sense, perhaps, to consider. So even on Slide 4, where we talk about modified loans or loans in forbearance, watch list loans, those type -- the heightened monitoring, those types of things might be ways that a bank could look at whether or not credit quality has declined. So the -- what we've seen is really looking at the routines that management uses and then both trying to identify idiosyncratic, which often banks are very good at, right? That's what they do. They look at their credits. But also looking at the more macro environment and looking at other factors that might say credit quality has deteriorated. And I think an important part that we tried to hit is this is credit quality, right? It doesn't mean that a loan has to be perceived as bad. It's just a relative measure of where it was at origination versus where it is at consummation date. So I guess, to summarize, a lot of flexibility looking at the way management looks at credit quality and monitors it. And if there are events at a macro level or a geographic level or in a particular asset class, all of those would be factors that I think a management team would look at to consider the buckets for PCD determination.

Got it. I appreciate that. That makes sense. And then just maybe a separate kind of related answer. I'm guessing the credit market is determined on the day of the close. So can you just talk about how you would look at determining that credit mark on the day of the close? And then how does the flow-through of a larger credit mark in a deteriorated credit environment when a deal closes? How does that roll through to impact tangible book value dilution and accretion? So if the mark ends up being larger than you may have expected, I would assume that's more book value dilution and more accretion down the line. But can you just kind of help flesh out a couple of the moving pieces there?

So just for clear, you're talking about -- in general, are you talking about the PCD determination?

I guess, across either asset classes, if a deal is announced and there's some sort of expectation of credit quality and then upon deal closing credit quality generally deteriorate across the industry, how does that flow through to metrics such as tangible book value dilution and earnings accretion? And what's the trade-off between the 2?

Yes. So I'll start on the PCD side. So well, actually on both sides, if my -- if I announce, and I have a fair value estimate of my loans and then I close, call it, 6 months later, and that fair value estimate changes, obviously, that impacts your tangible book value accretion, et cetera. On the PCD side, when you have the fair value and you have that credit mark, you have the day 1 establishment of your PCD portfolio, which includes your allowance. The size of the credit mark relative to the total discount will impact future accretion because basically, what you're doing is you're splitting out that discount on your fair value between the CECL allowance and the noncredit discount. So the higher -- the more credit mark associated with that discount will reduce the future accretion. On the -- so -- but it would -- but the tangible book value impact would be whatever -- based on the fair value of the assets. So you won't have a day 1 impact beyond the fact that -- of your fair value. On the non-PCD, you're going to have an impact from the day 1 provision, right? So in addition to booking the fair value of the loans on day 1 or, as I mentioned, 1.5, and I mentioned 1.5 because it goes through the P&L and it wouldn't go to goodwill. So that would be an additional impact to tangible book value because you're now establishing that CECL allowance on day 1.5 for the -- in our example, the credit mark and CECL allowance are the same. So you would have that similar impact, an additional impact to tangible book value from the credit mark. And that's really -- I mean, any changes in your estimations along the process prior to close are going to impact it. But at the close, the fair value at the close, you're going to have additional impact because of that CECL allowance on the non-PCD assets. That would -- and the impact -- go ahead, Greg.

I was just saying, another perspective relative to this is kind of in the CECL environment, right? So the allowance and the fair value mark of a transaction that was closed in 2018 or 2019 would have one construct under CECL environment. But obviously, under CECL now, you could have the exact same number of loans in a non-PCD category, right? So they haven't suffered any deterioration. But if you looked at CECL and what CECL might project to happen a year from now, you could see where that day 1 allowance that would go through P&L as a double count could get larger not because there are more loans, but just because there's a worse outlook under CECL. And that day 1 dilution, again, that it creeps back over in our example, 5 years, at the end of 5 years, you're back to 0, but on day 1, a worst CECL outlook could cause that number to increase. And then relative to what Jonathan is saying, from a fair value mark, the same would be true for your PCD classification as well.

Got it. That's helpful. I'm going to attempt to go for 3 here and see if I don't embarrass myself. But I think under the old accounting policy, when you had, I believe, it's acquired noncredit impaired loans that had a credit mark against them. You -- when those loans would either refinance or pay off, you would have a lot of accelerated accretion. But in a refinance scenario, you would have to actually reestablish an allowance. And so you have the scenario where banks that would announce an acquisition, they would have some ongoing level of accretion and a relatively depressed level of provisioning. And as loans would refinance, you would have accelerated accretion. So net interest income ends up being higher than expected, but that reestablishment of the reserve upon refinance would end up elevating the provision. It seems in your scenario as you're going through in your accelerated prepayment scenario, there's an acceleration of the accretion and there's a releasing of the reserve. It seems -- it doesn't seem like the drivers of the provisioning are the same under this. Can you talk to that? And am I getting those moving pieces right?

So let me see if I can break it down. So there's a couple of things going on that caused that prior to CECL. One was when I bought loans and at a discount, right, when I looked at my provisioning and my allowance determination, I was able to take full credit for that discount. So for example, bought $100 loan at a $10 discount, so I paid $90 for it. So I had that $10 discount. If I calculated my allowance and through my normal process, and I say, well, my allowance should be $2. I wouldn't have to record any allowance because I was still covered with the discount. So first, CECL takes that away. So you can't look -- you can't take full credit for that discount. So you have a little high -- and that -- in my example, I would still book an allowance. I'd book the relative 2% of $90, call it. So that's one component. The whole refi if it's determined to be a new loan, and we don't want to get into accounting, but if it's a new loan and you refinance it, you would have the same phenomenon that you had under prior accounting. Meaning if I refinance the loan into a "new loan," I would release whatever discount I have, which -- and then -- and whatever provision I have on that loan and establish a new provision. Now whether it's -- you would expect it to be very similar. So those may net out effectively. But you would get that income statement, that accelerated accretion and the provision establishing the new allowance under the loan. So you would have the same phenomenon. I think part of the depressed provisioning was because of the discount associated with the acquisition. And so therefore, the allowance, we call it accrete to impair, right? So you would accrete up until you don't have enough discount to cover your allowance, then you start booking incremental allowance. So I think that had -- with some of the phenomenon of why you had lower provisioning, but -- and that's gone under CECL because you don't -- you're not allowed to "accrete to impair" under CECL. So you may not see that as much. But you would have the acceleration of accretion if it's become -- if it's refinanced into a new loan. Does that make sense?

Got it. I think so. So the accretion still accelerates, but because you already have an allowance against that loan when you do the refi, the kind of rebooking of the provision is not as pronounced under the new system?

Yes. I mean, obviously, that's fact to circumstances dependent, but generally speaking, yes.

[Operator Instructions] Our next question comes from Steve Covington of Stieven Capital.

Just have you heard any will or desire from the FASB to fix the double counting piece of the standard? It seems like an obvious and simple thing to fix. Have you heard anything from them?

This is Greg. I mean, I guess, what I would do is recap history. I mean letters were written prior to the standard and comments were made that did not result in a change. And letters were written and comments were made subsequent to the standard, and there was no change. So I don't hear folks talking about it, frankly, at all. Jonathan, I don't know, from your vantage point, what are you hearing?

I agree. Last fall, the -- there was a -- it was put on the list of potential agenda items to remove it, and the FASB didn't take it up. So unless something significant changes going forward, I don't see them revisiting. It's been asked numerous times.

Okay. And one -- appreciate that. And one quick follow-up. So I've asked this of a number of CEOs that I've been meeting with. In a normal environment, assuming we ever get back to a normal environment, I asked them, could they sell an at-the-market rate loan at par? And most of them say, "Yes, we could sell new production at the market at par easily and sometimes at a premium." And so if that's the case for a non-PCD loan, could a management team argue that there really isn't any necessary credit mark because you could sell it at par, especially in an environment where you have to put up a day kind of this add CECL double count at the same time?

Well, let me take a stab. So one way to think about this is the allowance is a CECL expected loss measurement. And for a lot of assets in particular or certain economic times, that may line up with a fair value measurement as well. But there are some differences to the techniques used in commercial loans is the example people look to because a commercial loan for CECL only goes out to the contractual life, whereas a market participant might look at the renewal probability and the value of that renewal. So the other is again, an expected perspective from a management team may or may not be a fair value number, and you could look at that in today's environment, right? There's a lot of different nuances that are pretty significant in how to look at valuing assets, whether it be under market participant exchange or under CECL. So I also think the -- consistent with what your premise of your question is, FASB expected the non-PCD category to be rather small. And if you haven't had any deterioration in credit quality, not only would the population be small, but it would be logical that it wouldn't have a significant provision, perhaps. But a lot of that comes into the facts and circumstances of the portfolio as well as the environment for which the measurement assessment has to be done.

Yes. And I'll just add that if you're selling it at par, I guess, from a fair value perspective, you may argue that there is either no or minimal credit mark, right, because the interest rate is covering it. The challenge is you may be able to argue that. But from a CECL front, you wouldn't be able to. And the reason being is that, that argument was put forward that why should we record an allowance on day 1 under CECL. And I don't -- we don't need to rehash CECL, but that argument was made because I knew the originated loan. I price it to cover the risk of loss. It's covered and so forth and so on. FASB acknowledged that, but still said that they -- their view was -- their position was that they wanted to get full expected losses on the balance sheet on day 1 regardless of how it came in over time and the fact that you're getting paid to cover it. So I think even from that point -- even if you were to argue, there was no credit mark in my fair value, you would still have that day 1 CECL allowance for non-PCD assets.

Our next question comes from Catherine Mealor of KBW.

This is a really helpful call. My question is it kind of piggybacks earlier question, but I just want to kind of confirm that I'm thinking about this right. So when you're on slide, let's see, on Slide 9, that looks at what day 2 looks like for the provisions and the accretion. So in the scenario where credit is better or there are earlier payoff, it looks like both the accretion and the provision release, if you will, are both accelerated. So if we kind of over mark, then in periods after that, we're going to see a really -- we're going to see lower provisions, and we're going to see higher accretion kind of at the same time. So that's kind of question one. And then question two is then in a scenario where credit is worse, it appears that the accretion doesn't change. And so I just kind of want to confirm that and that accretion stays the same, but the only difference then is that we see a higher provision. Is that the correct way to think about it?

Yes, I think so. But go ahead, Jonathan.

Yes, I was going to say, and these are some of the 5 examples. But yes, I mean, the -- even though you're booking an allowance, as long as the loans are still performing and not on nonaccrual, you would still have the same accretion, right? If you -- if we wanted to dive down in here and say, well, that $25 is because we put loans on nonaccrual or stuff like that, you might see it, but the accretion would continue as long as it's an accruing loan. It's just we may have had a change for that.

I see. So basically nonperforming -- if they go nonperforming then in that credit deterioration scenario, you could actually see less discount accretion?

Yes. I mean, if your loans are nonaccrual or something like that. That's right.

That makes sense. But you should assume that in a scenario where there's credit deterioration, there probably will be nonperforming, especially if they'd have charge-offs?

I think that's a great nuance to the question because the CECL component. So if we just assume on Slide 9, the $25 in the bottom example, was all forecasts. So like today, not a lot of charge-offs, not a lot of change in nonaccruals for whatever reasons, and that's another discussion. But you could see where you're not seeing charge-offs, you're not seeing nonaccrual, so your accretion would go forward. But you believe unemployment is going to peak at a higher level in the second quarter than the first quarter, then that $25 would come in, and you wouldn't have "a lot of loans" going nonaccrual, so the accretion could continue.

That's right.

Got it. Okay. So I mean, in theory, these loans are still going to act very similarly, I guess, outside of the discount, but just from the kind of provision cadence as your legacy non-acquired book from a CECL perspective. As you think your losses are higher, you'll take a big catch-up kind of provision in that quarter as you think losses are lower, then you'll release some of your allowance kind of all at once in that quarter. Is that a fair way to think about it?

Yes. And that's really the big change with the PCD accounting is that historically when you had improvements in credit, you would take that in over time through yield, right? And now under CECL, you just take it in immediately through provision.

Got it.

So the yield stays.

Our next question Stephen Scouten of Piper Sandler.

I'm curious just if you could -- maybe going back to your illustration, if we could frame this up in a way, probably a lot of us are used to seeing it. And I know you didn't want to talk about specific company examples. But a lot of times, when we see these in presentations when a deal is announced, we'll see, "Okay, here's the credit mark. Here's the breakdown of PCD and the non-PCD, and then here's the CECL adjustment." And so if I think about your example that way, is it -- am I doing this right when I say there's a $575,000 credit mark, $500,000 of which is PCD, $75,000 of which is non-PCD? Is that correct on a top level?

That's correct.

Okay. And so when I think about that -- and then there's a CECL adjustment of this $75,000. And so when I think about that cumulatively, does that mean that entire $650,000 would go into the loan loss reserve on day 1?

That's what I was just about to clarify. So you're right, okay, that you've got a $575 credit mark for your fair value. But what happens is, when you look at Slide 6, and you've got the rate fair value mark, credit fair value mark, right, $25 and $75. That $75 that's called in line 2B, that actually gets accreted. The -- and so what we're looking at from an allowance perspective is the $500 on -- from Slide 5, plus the line 5 on Slide 6, $75, right? So you're going to accrete the $75,000 "credit mark" of the non-PCD assets. So you're not going to have $650,000 as an allowance. You'd have $575,000 as an allowance.

And that's what you're showing on Slide 8. You're showing that coming through in year 1. Why does it come through immediately in year 1 and not over the life of the portfolio?

Because on day 1, you have to establish an allowance for non-PCD assets and you establish that allowance through the provision. So that -- what we're showing on Slide 8, that day 1, where we say day 1 that $75,000, that's establishing that allowance for the non-PCD assets. And so in total, you would have $575,000, which is the combination of the non-PCD $500,000 and this $75,000 day 1 adjustment to establish the allowance for non-PCD.

And while this is extremely detailed, and maybe this one is, is if you look at Slide 6, what you're seeing in the example as the provision is line 4, right? So you're seeing that in the provision. The $75,000 on line 2b, the double count has to be eliminated, has to be put back into capital. So the -- by accreting the $100,000 over time, that line 2b is going in as increased NII. It's being recharacterized in the financial statements as NII versus a credit component. And by doing that, it basically de facto offsets over time the line 4 provision expense. I don't know if that helped clarify it, but it's basically, the...

Not really. The presentation just doesn't tie it together very well or reference the numbers, but that's neither here nor there, I suppose. But my main question is, so on day 1, the reserve would be $575,000 not the $650,000. And then, I guess, from a P&L impact, it would be $150,000 that would run through the P&L, right? The credit mark on the non-PCD and then the CECL adjustment?

Credit.

No.

No.

No. The $75,000 from line 2b is accreted into NII over the life of the portfolio.

Okay. But at...

I think you got -- the $150,000.

I can't track how it's going to fund.

Sorry. I apologize. But I think I see -- the $150,000 just happens to be the $75,000 and the $75,000 combined. It's just -- now that I look at it I probably could have put a different number. But you don't take the full 2b through income on day 1. That goes in over the life of the loan, like any other discount would. What the only thing you take the full amount to P&L on day 1 is the $75,000 of provision.

Yes. This is Ellen. We probably could have taken the discount accretion on Slide 8 and split it between the 2.

Yes.

Our next question is from Jack Bohlen of KBW.

It's Jackie. I have a question. And to a certain extent, I think maybe you've alluded to this when you said that if I understand you, it sounds like you said FASB intent was that the non-PCD portion was expected to be rather small, at least for the limited number of transactions that I've seen under CECL with PCD, that has not been the case. My question is, given the flexibility that a company has to mark alone as PCD, why wouldn't you try to mark everything you possibly can as PCD because it would avoid the day 1 provision and the double counting?

I guess, I'd just reiterate what FASB said relative to their belief of the outcome of the standard is that PCD would be a smaller portfolio enough to double count, would be less significant to the financial reporting. So I think as managements goes through and makes their determination, it is important to understand, particularly in a principle-base standard, what the objective was. And the PCD categorization was clearly expanded. I mean, the definition was changed during the process of developing the standard. So I do think as management looks at it, and it looks at how it monitors credit quality in any given particular environment, I think management teams will assess what does classify as PCD. And as that evolves, the focus of FASB saying it should be a larger PCD category and a larger -- a smaller non-PCD category, I think, we'll see that play out.

Okay. So am I correct in my interpretation that from an accounting standpoint -- I mean, obviously, it impacts your goodwill creation and everything. But there's an advantage to the PCD mark, if I understood earlier comments correctly, doesn't imply that those loans are necessarily bad, just that there's been even a small piece of deterioration since origination. When -- it avoids a large provision on day 1, and it limits some of that double counting and the accretion back into income. So it makes for cleaner financial statements going forward, right?

Well, I think you've hit 2 points there. One is that the categorization should be a larger number. And I think the struggle to one of the earlier questions is, why didn't they just have one category. I think as an investor and an analyst, the key here is, yes, the numbers for the double count may be smaller, but depending on how you look at your modeling, there would still be kind of a look-through to tangible book value to say, "Okay, how do I think about tangible book value dilution versus the day 1 entry and then the accretion back in, how do I want to assess that in my modeling?" So unfortunately, it might be smaller, but it still creates kind of a requirement to assess what goes through the income statement on day 1 and then how that's reflected back into the NII on day 2 through the life of the loan.

Next, we have a follow-up question of John Helfst of Voya.

Okay. So maybe one question, one statement. I feel like the extra provisioning that we have, like if we're a bank buying another bank that's got banged up credit, prior bad credit performance, you're going to have like a massive provision upon closing, which to me then jams up all your valuation metrics that sticks for the history of mankind. It's actually to me non -- I don't know how to think about, like noncore, but it's actually less user -- adds less user value. Actually, it actually obfuscates the core profitability of that company. Like I know probably everyone's been complaining about that. But anyway. And then my other question -- that was a statement. You can comment on that. Maybe I'm wrong. My question then is like with this whole COVID, we're in a period of forbearance. And then I think we're going to move to modification before year-end under the CARES Act. Can you walk us through what we should be expecting in terms of like we've seen at a banner, I know we're not supposed to name names, they had some for CECL outlook for 2Q and some of it was driven by loan downgrades. And that was like helpful for me, understanding the process of what to expect going forward. So I would think in a normal environment, we have that, but now we have CARES Act, which may actually limit the amount of -- probably a long downgrade should be the same, but then we may have less charge-off experience with the modification process and TDR accounting. Am I like crazier or am I on to something?

Let me separate those into the 2 -- what I think was 2 questions. So the first one, somebody buys, I think your words, was banged up credit. So if you buy a troubled bank, so to speak, let's just say for arguments, virtually all or all the loans would be PCD. And therefore, there would be no non-PCD and no debit through the day 1 income statement. So you would basically record it at fair value, book the allowance. And if your assessment of the banged up credit was perfect, you would not have any income statement impact in the future. So I think that example is -- leans towards what FASB was saying is that would be recorded on day 1. You wouldn't see it in the income statement because it was banged up credit. The dialogue around COVID and forbearance, I think what I would just say there is it really focuses on the difference between deterioration in credit quality and measuring the risk or the expected loss. And it's -- that's probably one of the most fundamental issues here is CECL is estimating expected losses. What PCD classification is trying to do is find those loans where you've had deterioration in credit quality. So a borrower that's asked for forbearance arguably has perceived his or her or the company's credit quality is changing and that might be a factor that a management team may consider. And to your point, a year from now, the stimulus may have kept that from ever being an actual charge-off, but that's a different measurement than whether or not the credit quality of the borrower has deteriorated since origination.

So as a purist philosophy major, I should be downgrading my loans, depending on the cash flows, my assessment and my quarterly cash flows and my borrowers and therefore, booking a larger CECL provision -- or just provision in 2Q, 3Q? And then another question would be cash flow accounting. Is -- do these loans charge-off in 2021 or not? Is that what you're saying?

Well, what I was trying to say is how you measure CECL is a risk of measuring expected losses. That is fundamentally different than how do you classify a loan that's had greater than insignificant deterioration in credit quality. So you could -- those are totally different measurement paradigms.

Well, I mean, I'd argue with you. I was a lender, and you rent -- you put loans on a risk rating of 1 to 9. Like 9 is highly doubtful, right? One is gilded edge. So bell shape curve, 3, 4, 5, cash flows are deteriorated. You're going to downgrade. And that's like a Moody's BBB to a BB. Like how is that not greater than a significant expected loss? Or are you saying really, you're looking at the bottom, like 7, 8, 9, is that what you're saying for the CECL?

Well, I think an example might be, even if you use where stimulus might help a particular industry, that industry may have a deterioration in credit quality. But because of the stimulus, a company may conclude that the risk of loss ultimately is not as great as you might think. But that doesn't mean that, that industry hasn't had a deterioration in credit quality. And again, I think in the slide we used, PCD is not a distinction of good loan or bad loan. So even in the next -- someone said, who knows when the normal comes back, but let's just say, 2 years from now, you have a transaction. And you could be looking at this and classification of PCD doesn't mean you think that's a bad loan. It's just -- I mean, you could have a Grade 1 commercial loan go to a Grade 4, still a pass grade and have a bank conclude that that's a greater than insignificant deterioration. It's still a pass grade loan.

Yes. Meaning no additional reserve?

Correct. Or very modest. But again, that's why we said even high-quality borrowers can have a greater than insignificant deterioration in credit quality.

And our final question comes from Dan Mazur of Citadel.

This relates a little bit to Catherine's, but just to confirm. With -- prior accounting provided a measurement period where you could adjust with credit deterioration that didn't flow through the income statement but was an adjustment to goodwill. The example in your presentation is kind of year 3. But is there a measurement period now? Or is it just -- it's kind of stuck at day 1? So any deterioration even in the first few months would be what would flow through the income statement?

So CECL doesn't change the purchase accounting guidance, which provides for the measurement period. CECL doesn't provide for a subsequent measurement period. And so you should have your day 1 established reserves, and then anything after that would be fall under your normal CECL process. And so if you have deterioration post the next month or 2 months later after close, that would just be part of your normal provisioning process.

Okay. That makes sense. And then just wondering if CECL accounting has a potential to impact, whether acquire books of purchase, a bargain purchase gain versus a negative goodwill or contract would be. So like in a hypothetical situation, of course, like an all-stock deal where the seller is below tangible book value, but the buyer is either -- is very close to tangible book value. What's the relevant measurement period for tangible book value for the buyer? I'm just wondering if CECL would have an impact on that on a lower tangible book value kind of post-close versus pre-close?

Yes. And when you get into that kind of discussion, it really gets very facts and circumstances relative to the different accounting measurements across all types of asset classes, be it securities, be it loans, be it unfunded commitments. So it's really hard to really walk through how that might play out.

This concludes our question-and-answer session. Now I'd like to turn the conference back over to management for closing remarks. Thank you.

Well, just wanted to say thank you to you all for joining us again today. Hopefully -- and thank you so much to Greg and Jonathan and the whole team at Deloitte. We really appreciate their finding time to help walk all of us through this. And certainly, we hope it was useful. And don't pretend that we've potentially answered every question. But certainly, we'll stand ready going forward to help provide additional assistance. And thank you so much.

Conference has now concluded. Thank you for attending today's presentation. You may now disconnect.