Good day, ladies and gentlemen, and welcome to the Bank OZK Third Quarter 2018 Earnings Conference Call. [Operator Instructions]. As a reminder, today's conference is being recorded. I would like to turn conference call to Mr. Tim Hicks. You may begin, sir.

Good morning. I am Tim Hicks, Chief Administrative Officer and Executive Director of Investor Relations for Bank OZK. Thank you for joining our call this morning and participating in our question-and-answer session. In today's Q&A discussion, we will make forward-looking statements about our expectations, estimates and outlook for the future. Please refer to our earnings release, management comments and other public filings for more information on the various factors and risks that may cause actual results or outcomes to vary from those projected in or implied by such forward-looking statements. Joining me on the call to take your questions are George Gleason, Chairman and CEO; Greg McKinney, Chief Financial Officer and Chief Accounting Officer; and Tyler Vance, Chief Operating Officer. Now let me ask our operator, Kevin, to remind our listeners how to queue in for questions. Kevin?

[Operator Instructions]. Our first question comes from Catherine Mealor with KBW.

Just wanted to start with a couple of questions on the 2 credits that had deterioration this quarter. And my first question is just, can you give us a little color about when the credits went substandard? And then how -- and then when the $19 million reserve was originally built into these credits?

Yes. I'd be happy to do so. The South Carolina credit went substandard in the second quarter of 2017. The North Carolina credit went substandard in the fourth quarter of 2017. And of course, our substandard and watch assets and other writings on our assets are reported in the notes in our Qs and Ks. So those were disclosed in those numbers, not specifically referenced, but disclosed in those numbers. The reserves for them, the $19 million, were accumulated over that period of time. Each of those credit was a watch-rated credit before it became substandard. So when it was a watch-rated credit, they added 2.5% allowance allocation for them. And then as they became watch-rated credit, those allowance allocations were increased based on the appraisals on the properties at the time. The 2 properties have both been appraised in the last 2 years, one, I think, about 2 years ago and one about a year ago. And then, of course, as the performance of those deteriorated this year, we obtained new appraisals on both properties in the quarter just ended. Obviously, the property performance had deteriorated significantly on both. That poor recent performance was factored into the assumptions in the appraisal, but led to a significant divergence in the most recent appraisals from the 2 appraisals done just 1 or 2 years ago on those properties. And hence, that's why the additional provision was needed over and above the $19.1 million that was previously reserved for these.

Got it, okay. And then, I guess, the big question with these credits in trying to figure out is, are these kind of one-offs? Or is there more risk in the rest of the portfolio? I mean, that's the big question here today. And I think, your chart on Page 4 of the management comments is super helpful in showing that you don't have a lot else that's in this vintage or these types of credit. So that's super helpful. But I think the question is it still was a very sudden and very large decline in the value, and then the lower appraisal. And so what can you tell us about the rest of your portfolio that can give us comfort that there isn't risk in the accuracy of appraisals for the rest of your RESG book. And particularly, maybe they're newer vintage credit, but they are significantly larger credits? That'd be helpful.

Yes, yes. I can give you quite a bit of color on that and probably some additional color that would be helpful as well. When we return to the 2 properties that we wrote down, the two credits we wrote down in the quarter and give you a little color on those, so South Carolina property is an enclosed shopping mall, and one of -- a tiny number of enclosed shopping malls that we have in the portfolio. And I'll discuss the others on those. The prior price along that had assumed that the mall would continue to gain tenants and fill vacant space with tenants and reach a stabilized condition. As it turned out, over the course of a year, several more small tenants exited the space. There was, until just a few months ago, a very active negotiation going on to refill some of that space and replace that lost rent with a new tenant. That negotiation collapsed. And when we ordered the appraisal based on the lack of leasing velocity, the new appraisal, essentially seeing -- assumes no leasing of vacant space and a continued erosion of the tenant base as leases expiring and so forth. So the appraisal's focus changed significantly from an operating property that will continue to replace tenants and go forward to a property that would just essentially melt down. And that's reflected in the extreme change and the valuation on it. The North Carolina property, which is a multi-phase lot and land and residential home facility, those guys have done a really good job improving the club operations and the sentiment around the property. And as a result of those improvements, there's been a significant number of homeowners, who have previously bought lots, begin to construct custom homes in the subdivision. So…

Yes. But what's the size of the watch -- the full watchlist credit?

The credit that is shown on the RESG bubble chart here?

Yes.

That credit is a revolver, and it's a revolver for lot development. It's a revolver for vertical home construction. The total outstanding balance on that today is about $49 million and change, and the total committed balance on that is $57 million -- $57.5 million total committed. And I saw one of the commentators who suggested that was a $100 million credit. That's not accurate. It's -- if fully funded, it would be $57.5 million. That is a very high-end project. Sales velocity and pricing trends are stable to positive on that. In the last 12 months, they've sold 16 lots and have four under contract to close between now and the end of the year. They've sold four townhomes out of the vertical inventory and have 6 more under contract expected to close before the end of the year. And those sales velocity numbers are consistent to improve with the prior years. That credit we expect to continue to operate as it is, and continue to generate accretive paydowns. And we expect that, that project liquidates out through normal sales of inventory that either have or will be creating, that will have more than enough proceeds from those sales to repay our principal and interest in full without any sort of delinquencies and so forth. Our current loan-to-value on that project is right at 97%. It is a watch credit, so we have 2.5% reserve allocation for it, which we think is quite adequate, given the sales trends and pricing trends on the project.

And one more on that credit, and I'll pop out of the queue. What market is that credit in?

That is California.

Our next question comes from Stephen Scouten with Sandler O'Neill.

I was curious about the frequency of appraisals. I know you spoke to the fact that they were relatively frequent. But can you give an idea about how many of those appraisals, if you know this number, might be over 2 years old at this point in time? And then assuming they're from what I hear 25k a piece on average, I mean, could there be an incremental cost that has to come through if we need to reappraise all these things?

Well, number one, we don't need to reappraise them. And the cost varied quite a bit from project-to-project. Depending on the complexity of the project, of course, that drives the complexity of the appraisal. Obviously, what we do is get a loan at origination on these credits. If our loan is more than 6 months old at the time we do a modification renewal, extension of the credit, we get an appraisal buyer. Most of our loans, as you know, are around 3 years before the maturity. Some are 24 months or 18 months. Some are 42 months. But the kind of the median and main probably on that is 36 months. If a project is experiencing any sort of unusual circumstances, difficulty, delays, problems, we will get a new appraisal for it. And our RESG loan documents, I think, almost universally, if not universally, require that if we have to or deem it appropriate to get a new appraisal, our customer pays for that. So the appraisals are as current as they need to be to give us an accurate view of property values to make sure that we're operating off relevant current market information. And if we find ourselves in the situation where we need to get another appraisal, that expense is almost always borne by the customer under the loan structure.

Okay, fair enough. And what would happen, I guess, if you were to stress the discount rate on your book as a whole and on those appraisals? I mean, have you guys done kind of math around that to see what downside would be from a stressed discount rate if that was part of the issue here?

Well, every loan is stress-tested if it's an income-producing property. 4 ways, and one's a series of economic stress based on vacancy and rental rate. One is an interest rate sensitivity stress test as to how much our loan could increase and still have amortizing debt service coverage at stabilization. One is a sale exit stress test based on the where secondary market pricing is for pertinent loans, how much could rates go up. And our sponsor still likes it with enough proceeds, pay our loan off. And then the fourth stress test is a sale exit stress test that is based on current cap rates and projected NOI and how much could that go up. So these loans are stress-tested in a lot of ways, and we feel extremely good about the quality of this portfolio. We acknowledge that -- the fact that we had 2 charge-offs in the RESG portfolio. This quarter we'll unnerve some people, and some will think the wheels are falling off. But in reality, we've had five losses in that portfolio and 15 years on there, at any point in time, in recent years, 300 to 400 credits in the portfolio. And we've had five cumulative losses on the life of the portfolio. We feel extremely confident in the quality of the portfolio, and that there's no significant risk beyond the two credits that we've written down. The watch credit we've discussed in detail. It has an elevated risk profile, which is why it carries a watch status. But based on our analysis of that project and so forth, watch, as opposed to substandard, is the appropriate writing. So we think it's relatively low, albeit elevated, risk associated with that. All the other credits, we believe, are clearly past credits with significant margins for errors on all sorts of stress metrics regarding the performance of those credits.

Okay, that's helpful. And one last one for me, I guess. I mean, you sound still pretty confident about the whole of the book. And obviously, seeing this bubble chart, I can see why based on where the LTVs are today. But I guess, given the kind of sentiment around the business, do you guys reconsider your risk appetite at all? Seeing the largest bubble here, I assume that's that $556 million loan I saw out in the press here recently. So can you take to the -- maybe that risk appetite for those larger-sized loans? And if -- just the overall sentiment market conditions to make you rethink how you approach this business moving forward.

We're a commercial real estate and construction and development lender. We're very conservative in the way we do that. When this year is over, I'm confident that, once again, this year for the 22nd or however many years, that's more than 22 years in a row, our losses will be below the industry's loss ratio. So we're going to continue to execute our business strategy. And it's not going to change our underwriting, our analytics or our structure or pricing on this. We will have an occasional loss in the RESG portfolio. We're in the lending business. Anybody that's in the lending business is going to incur, occasionally, some losses in your portfolio. But just as we've had 5 losses in the last 15 years in the portfolio, my expectation is that even though the portfolio is a lot bigger, we'll probably have a similar volume of losses over the next 15 years. We will occasionally, in a quarter here and there, in a year here and there, have a loss on the RESG portfolio. But we're not changing our business model at all because we believe it is very, very sound and very, very conservative. So it's business as usual for us, and we still are originating loans. The credit you mentioned that we originated right at the end of the third quarter, it is the largest credit we've ever originated. And it is a condominium loan credit in Miami. It is a great project, and it is a great sponsorship. And we're very excited and pleased to be the senior lender on that project, in fact, the sole lender on that project. The project is one we've been working on for over a year. And buildings 1 and two at this project have long since been built and…

Our next question comes from Arren Cyganovich with Citi.

A question on the portfolio management. I was surprised to see that you still had credits that were 10-years-plus old. To me, I would think that those credits would just naturally be somewhat problematic, and maybe you need to take reserves for something given that these projects typically don't last that long. So maybe just talk a little bit about why they stayed in the portfolio, why they weren't sold, why they weren't reserved for in a larger manner prior to these recent credit issues.

Yes. I'd be happy to do so. And I don't think there's any negativity associated with the fact that the older loans are in portfolio. I'm sure most banks have multiyear-old loans in their portfolios. So I don't know that there's any correlation there between any quality issues. In the pre great recession era of RESG, we tended to originate more loans that were sort of transitional or semi-stabilized sort of properties or stabilized properties for that matter. And that also meant we were doing loans at higher leverage. So this South Carolina credit on the shopping center, that really is the only income-producing property there in the sense of having tenants from paying rent. That is the only vestige from that era of the RESG portfolio when we were doing more of those sorts of loans and doing them at higher leverage than we do credits in the RESG portfolio today. As you'll notice from the color of dots on the little bubble chart there on Page 4 the management comments, the other five are either land and lot and vertical construction revolvers, which were intended to be long-term projects, and that would be the North Carolina credit and the watch credit. Or the other 3 loans that are down in the 30% loan-to-value range. Those are land or land development loans. And those, in some cases, a couple of those particularly were intended to be long-term land-held plays that were done at 50% or less leverage, and as they amortize every year. And they were expected to be longer-term credits. So given very low leverage of those and the quality of our sponsorship on those, we're quite comfortable with that book of business. So I think to really kind of drill in, I think, on where you're going, even among the 6 older credits in the portfolio, the two substandard and one watch credits are appropriately rated. The others are excellent.

Okay. In your prepared commentary or your commentary yesterday, that they -- you mentioned the paydown being accelerated and likely to accelerate or remain high in 2019. So it sounds as though that your prior years, the very high growth is now going to be much lower than it had been, if not flat, going forward. And the question is, if your credit does continue to perform as you expect, you should be generating pretty nice capital. And would you consider now, particularly with your stock down so much, instituting a relatively large stock buyback plan?

Well, that's a great question. And that question, of course, was asked on our conference call 3 months ago. Our board took up that issue in the August board meeting. And I would tell you that senior management and the board have a more constructive view of our growth projects -- our growth prospects over the next 3 to 5 years that, I think, is probably typical among the listeners on our call today. We've had a tough year of growth because we've had just such large paydowns as detailed in Figure 20 on Page 15 of our management comments, which you alluded to. And we do think that, that growth paydown continues. As we wrote in the management comments, we're expecting a record level of loan repayment in this current fourth quarter of 2018, so much so that our growth in non-purchased loans is probably going to be 0 or negative. And it could be quite a bit negative if all those paydowns materialize. We do expect to continue to have another high level of paydowns next year based on projects that we expect to see out in the sales on projects that will be reaching certificate of occupancy status next year. So I think we'll be dealing with a high level of paydown again in 2019. We are encouraged, as shown in Figure 21, by the fact that without giving any ground or diluting the quality of what we're originating or the return on equity on what we're originating that our originations in Q3 will were $1.47 billion. I think we've got a real good shot of having the best origination quarter of the year in Q4 if we can get everything closed that's in the path. So I think if you look at Q4, we're going to…

Our next question comes from Ken Zerbe with Morgan Stanley.

Great. I guess, maybe let's talk about your margin or your core spread just a little bit. Obviously, it was a pretty sizable decline, down 19 basis points in core spread and margin, I guess. Can you just help us understand -- I hear the stat about the 77% of the non-purchased loans are variable rate. It feels -- it seems they -- has been under pressure for a couple of quarters. It was down, obviously, this quarter because LIBOR didn't go up very much. But it was also down last quarter when LIBOR did go up quite a bit. I'm just trying to help us understand the dynamics of why we've seen sort of a couple quarters a week on margin and spread.

Well, that's a great question, Ken. And what I was most surprised about and most disappointed about in the quarter on the spread front was the fact that our yield on non-purchased loans actually went down 1 basis point. That -- we had to really dig into the minutia to come to the conclusion that, that was correct. A lot of factors played into that. And one was that we had what I think was a lower than average number of credits generated by loan prepayments this quarter. Now we had a record level of home prepayments, so you would think that would generate a lot of credits. But the loans that paid off for the most part were passed any sort of yield protection that we had on them. They -- we're at a point where most of the -- near maturity or at maturity. So all or substantially all of the deferred credits from the origination fees had already been earned in income on those as they approached maturity. And we just didn't get the extra bumps from minimum interest, prepayment, exit fees and so forth that we've gotten in some prior quarters. So it was a lagged quarter there. We had the charge off with the accrued interest on the 2 loans that we've put on nonaccrual that was about $570 -- Tim, it was 5...

$572,000.

$572,000. I mean, that's not a huge number, but it was in there. We had more loans pay off in our community bank and indirect portfolios. But those loans, a lot of those have net deferred debits instead of credits. So we sort of got dinked on the debit. You mentioned the LIBOR issue and the fact that on a period end to period end or an average basis, that was up 14 or 17 basis points when we have a 25 basis point move in the Fed funds rate that you would have expected to filter through there. And you can just go through all of the things. And normally, in a typical quarter, some of these things can lean to the positive side and some lean to the negative. Everything just leaned negative on that NIM this quarter. So we think based on the expectation of another Fed rate increase at the end of this quarter and what we think will be a normalization of some of those kind of unusual little things that go to and fro and push your yield up or down, we think we'll have a much more normal result this quarter. We believe that, that quarter, where our yield on non-purchased loans went down, is an anomaly. We think we'll see a return to a more normal pattern this quarter and in future quarters as long as the Federal Reserve continues to raise the Fed fund's target rate. Now in regard to the second quarter of this year in the core spread, yes, we were disappointed with that. To some extent, LIBOR didn't give us as much boost in Q2 as it did in Q4 of last year, Q1 of this year. So there was a bit of a LIBOR impact there. And of course, our deposit beta was about 21 basis points in Q2. Is that right, Tim?

The deposit beta was higher, but the average cost went up 21 basis points.

Yes, average cost went up 21 basis points. I'm sorry. I said that wrong. And we really started trying to work on the deposit costs in the quarter just ended. We've got a little bit of improvement there, the 18 basis points, but still not satisfactory. But it's going the right direction. And we've got a fairly detailed and comprehensive effort that a lot of folks are working on to really improve the data, the analytics, the pricing, the strategies and structures around deposit pricing and deposit price optimization. I think we'll see continued improvement in that, particularly as we get more of the technology and more sophisticated parts of that rolled out in 2019. So yes, we were disappointed, as you were and as everyone was, in the negative movement core spread over the last 2 quarters. We're working really hard to flip that back to a positive. I do think the yield on non-purchased loans in the quarter just ended was an anomaly, and we'll get to a more normal pattern of behavior there going forward.

And are you able to quantify -- I think points number 1 through 3 that you mentioned seemed like they're just -- they have not -- almost nothing to do with sort of the loan yields or sort of the interest rates. Like if fourth quarter went back to normal, and you had better credits on the pre-phase, and things like in the community banks, how much -- what are we -- I mean, can you size that for us? Like are we talking that your core spread actually jumps up by x basis points because of a normalization? I'm just trying to get a size there.

Tim, go ahead.

Yes. Ken, this is Tim. Really, look at Page 17 and 18 of our management comments. We enumerated, and George explained a lot of those points. I think we've got 5 different bullet points there. Just that $572,000 write-off was worth 2 basis points on the margin. And you can kind of go through all of those different points that we have there and they're worth a couple of basis points each. And so I'm not going to try to size what that would be. But you could then look at our Figure 24, which shows all of our increases on our non-purchased loan yield, the trend over a several quarter period. And if we start with 4Q '16 and go through 2Q '18, the average is 16 basis points of increase per quarter. And even if you throw in this past third quarter, which we think was an anomaly, that's still 13 basis points of average increase every quarter. So I do think, to George's point that this is an anomaly. We feel like we'll get back to a normal range in the fourth quarter on non-purchased yields. And then as George mentioned, obviously, we're showing improvement on the deposit cost side. So we hope to continue to show that improvement. That sets us up for a much better result for Q4 core spread. So again, we've got some things that -- all of which moved kind of against us during the quarter. And we've got those kind of outlined in our management comments.

Got you. That's helpful. And then just one last question in terms of rates or specifically the deposit beta in your deposit gathering strategies. Given the expectation for no loan growth next quarter, and 2019, obviously, is a little bit weaker than probably what we have expected, are you able to roll back some of the spin-up markets? Or -- because, I guess, I'm trying to think of like, can you stop paying up for deposits since you, I guess, technically don't need them as much to fund the business?

Yes. I'm going to let Tyler answer that. But I'm going to into Tyler's answer to that by saying spin-up is not so much a term we're using. And if you think about how we've described spin-up in the past, it's been a very geographic-focused market segmentation strategy. And in reality, it's George's old strategy. And it's worked well for us for a number of decades, but that strategy has not worked as well for us this year and this cycle, as it has in the past, because the competitive world has changed with a lot of online competitors and consolidations that's occurred. And the banking industry means that we're in -- in some of the markets we're in now means we're up against much more regional competitors, some of whom are very aggressive on pricing. So our strategies needed to change. So what we've tried to do is start a process of evolving to a much more sophisticated strategy that's focused on customer product and geography. So we're really not doing what under our old rules of engagement strategy is for spin-up. We're moving to a much more nuanced and sophisticated strategy. And this is the Tyler Vance, Cindy Wolfe new world of dealing with this issue for our company, and I think it's very timely. And Tyler, you might want to add some comments on that.

Yes. The only comments I would add is that certainly, we are focused on these additional analytical capabilities. And those in the third quarter started to give us much more visibility in terms of our deposit activity in various markets, as George mentioned, market segments from a geography product and also a competitor standpoint. So there's analytical data and modeling enhancements that are underway or planned, we think, will continue to contribute to better performance in terms of beta. Deposit specials are a continuing part of our deposit acquisition strategy, as George mentioned. And we do expect to continue to offer various CD or other specials from time-to-time in certain of our pricing regions. But our deposit team's prime focus is on increasing our volume of core deposit customers. And the management commentary saw some of those results in terms of net checking, which has, once again, been excellent at 22,099 net new accounts through the first 9 months of this year. And of course, that number's already in excess of last year's full year number of 22,013. And we expect it to grow further in the fourth quarter. Now the original question, I didn't think some of the slower growth that we're describing obviously should translate through to a lower beta in Q4. Or that's certainly our hope. It gives us an opportunity to continue to take out some of our highest cost deposits as we move through the quarter. So we're hopeful about that.

Our next question comes from Michael Rose with Raymond James.

I just wanted to go back -- I just wanted to go to that Figure 21 that you referenced earlier. And the originations clearly are down. Now George, in past calls, you've talked about a lot more of the markets and submarkets being balanced from a supply/demand dynamic. And now not only are paydowns seemingly structurally elevated, but you're seeing more competition from traditional bank lenders that have come into the states that you guys operate in with RESG. I guess, my question is -- I mean, have we peaked in terms of originations? Or do you think this is the right time to continue to try and gain market share, perhaps at the expense of cost and rate?

Well, what I would repeat, Michael, is what we've said for my 39.5 years as Chairman, CEO, and that is that quality is nonnegotiable. We're not going to do things that don't meet our very stringent asset quality standards. And we're, secondly, only going to do things that give us a satisfactory return on equity for the capital invested. So we're not going to try to gain market share in a highly competitive market by cutting our credit standards or getting too aggressive on pricing. That discipline is the singular reason that our volume in -- of originations in Q1 and Q2 of this year were half the volume of originations in Q1 and Q2 of 2016 or 2017 in round numbers, about half the volume. We maintained, and I spent a lot of time talking with our lenders over the course of Q1 and Q2, that if we would hold to our standards and continue to work hard and call on customers and turn over rocks for opportunities that we would be able to get our volume back to a better level without giving up credit quality or giving up pricing. And I'm very gratified. It's a big win for us even though it's lost in all the noise of the quarter, that our volume came back up to $1.47 billion of originations in Q3 without us giving up our credit or our pricing standards. Again, as I mentioned, we think we're well positioned to potentially have our largest origination volume for the year in Q4. We've certainly got the deals lined up to close. We've got to get them closed, but we've got good volume there, and we seem to have good pipelines going into next year. So I think that we've held through our principles. We've held…

Okay. So I guess, the follow-up to that is you built out a couple of these other businesses. It's really contributed to the overall growth of the company. Now I would suspect that those verticals, just from a holistic point of view and what they could contribute to growth relative to your national RESG business, are smaller. So I guess, what I'm struggling with is the headwinds in RESG accelerate. And you peaked in terms of -- or at least it's going to be really difficult to get back to the same origination volume in RESG. I mean, are we looking at a period where, especially with the run-off portfolio, where you could actually see little to no loan growth for the next few years?

I don't think so, Michael. I mean, it may be, in terms percentage terms, less loan growth than we've experienced over the last several years, but I think we will generate a very healthy positive growth right in non-purchased loans going forward. And the diversification is a big part of that as well as the fact that our confidence in RESG has never been higher than it is today.

Yes. Michael, this is Tim. I mean, our purchase portfolio is $2.3 billion right now. Obviously, it's a lot less of a headwind than it has been in previous years. I think it ran off 38% year-over-year. So even if you took that 38% or somewhere around that, obviously, that headwind is a lot less. We said in the comments that we would have -- they would expect 2019 growth to be better than 2018's growth. So from our standards, it's not as high growth as we've had, but still pretty good growth. And it's got a good mix. So I'd just -- I'd point you to those data points.

I think we will see -- as Tim referenced in the management comments, we will see volatility and payoffs and originations from quarter-to-quarter. But we're pretty constructive on our opportunities to grow in 2019. I will also -- I want to acknowledge because I don't want anybody to be surprised by it, we're going to have a lot of payoffs and paydowns in the portfolio in this current fourth quarter of 2018. And it is very possible that we'll have 0 to negative growth in the portfolio this quarter. So no one should be surprised by that. But I think when we get past this quarter, I think we've got pretty a decent loan growth trajectory going forward.

Our next question comes from Matt Olney with Stephens.

George, I appreciate the details on underwriting construction loans. And one of the items that you mentioned was...

Yes, I don't think. I didn't mention the use of interest reserves. But I'm happy to address that question. In most complex or large construction loans, you have reserves built into the loan structure for interest, for taxes, for insurance, for everything that's going -- every cost that's going to be incurred during the construction and development period to a stabilization. So that's just an RESG deal or Bank OZK deal. It's a standard practice throughout the industry. And we adhere to that practice. So you'd say, "Wow. You're including in your loan a reserve for the interest on the loan." But in reality, we're calculating all that into the required equity of the borrower. So we want the borrower to put their equity in before we fund our loan. So the borrower's money, if they're putting in 50% of the cost of the project, and we're putting in 50%, that is our average in the portfolio. Then they're going to pay for the land and pay for the closing cost and pay for the loan fees. And they're going to pay for construction and development and other costs until their 50% of the money is in the project. And then when that's in, if the budget's in balance for the project, we will start funding our loan. So that typically means that we are funding the interest out of our loan proceeds, but that was just factored in as a total cost of the development. And just the sequencing, the interest is do later in the project. And that's when we're funding, after all the equity is in the project. So I think we have a very standard and very conservative approach to using interest reserves. If we have a loan that has matured, part of the renewal or extension of that loan as the sponsors got to rebalance the reserves. So for example, we were talking about some of these land loans that go all the way back to 2006 and 2007 and 2008 that are 30-some percent loan to value. Our typical approach on those loans is that they're renewed every year or every two years when we renew them. We'll get a new appraisal, make sure that our loan to value is still going down. We'll get amortizing annual payments on those. And if the modification, the borrower will have to deposit with us an interest -- an operating cost reserve that will cover the renewal period. So the sponsor on renewals, we just don't keep putting interest up on loans. The sponsors after the original period have to rebalance and replenish the reserves in a manner to our satisfaction. So I think our use of interest reserves is very prudent, very conservative, and it is consistent with the highest standards that the industry follows.

And then going back to RESG. I think historically, you said that many of your borrowers have paid RESG a premium price due to the higher quality of service and the faster execution. But now you've got increased competition. So I'm curious if you still believe that you get paid; a premium price for RESG.

That's a great question, Matt. And if you -- you've listened to our calls for a number of years. So you'll recall in early '16 -- in 2016 and so forth, when we said because a lot of competitors have pulled back from the space, we were getting more premium pricing than we had in, say, 2014 or 2015. And certainly, as competitors have come back in, in large numbers, both bank and nonbank competitors, into the space, our ability to get a premium price has diminished. But I think there are still a large number of our customers who value the quality of execution, the speed of execution, the quality of our asset management capabilities. And that helps us in 2 ways. In some cases, I think we get a premium price, albeit probably not as much as we were getting 2 or 3 years ago premium, from those customers. And two, a lot of times, when you're in a competitive situation, you may have a competitor who's sitting around on top of your price and deal structure. And yes, we get that nod in many cases because of our track record and proven capability to execute and service the assets and work with customer in a very intelligent and understanding way. So I think, yes, it's still an important part of our story. It's still an important part of our business model. It's not getting us paid this higher premium as it would in a less competitive environment.

Our next question comes from Jennifer Demba with SunTrust.

George, just following up on the question from before about share repurchase appetite. I'm assuming you anticipated that question today. But I'm just surprised you wouldn't look at share repurchases now given the stock is at such a cheap valuation. I don't know if you expect it to go down this much -- or the sector to go down this much in the last several weeks.

Jennifer, I don't have anything, really, to add on that comment other than what I've already said. I don't think the board will look at it in November. I think we're very anxious to get our Q4 numbers out there, which we expect will be really good numbers. I would expect, as part of our annual strategic planning and budgeting process at the end of the year and going into the February board meeting, that, that will be a subject that gets looked at. And the board will look at it, again, in February. And again, I'm not going to prejudge their conclusion. That will depend, in large part, on our expectations for future growth opportunities.

Okay. Appreciate that. One more thing. You've obviously -- in RESG, you had the biggest exposures in New York City and Miami or South Florida residential real estate. Could you remind us of -- I know you have a variety of properties, I'm sure, at different price points. So can you just remind us kind of some standard metrics, if you can, of what kind of projects you have within those 2 portfolios?

Well, Tim, what page are you on there? Page 11.

Figure 13 and 14.

Yes, Figure 13 and 14 really talks about the product diversity in our RESG portfolio and the product diversity by geography in our RESG portfolio. And you are correct, as shown in Figure 14, New York and Miami are our largest markets, followed by L.A., Chicago and Dallas and Denver. Condos are the largest. Multifamily is second, followed by office, hotel and mixed-use projects as far as product guide. And you can see the numbers and the leverage points on all of that. The New York metro area portfolio has a lot of condos in it. It has a lot of multifamily in it. It has a fair amount of office, a fair amount of hotel. And they had a fair amount of mixed-use, and then a smattering of other things from land to some miscellaneous sort of industrial commercial. The Miami portfolio is much more concentrated in condominium loans. We've got 13 or 14 projects that I don't know the exact number now, but I think we're about 80% sold on those projects. The typical contract for sale either has 30% or 50% nonrefundable earnest money up. So we've got, Tim, 6 or 7 of those projects that are CO-ing this quarter and first and second quarter of next year, that sales are already 80-something percent, and they're going to pay off. And as soon as they CO, we'll get paid off on those condos. So the Miami portfolio is very low leverage, as you can see, 42% loan-to-value, 37% -- I'm sorry, 42% loan to cost, 37% loan-to-value. And that just reflects the very low leverage we do those condo loans at.

Our next question comes from Matthew Keating with Barclays.

What I'm curious about, it does seem on the loan growth side, we've obviously talked in this call and in the prepared remarks how non-purchased loan growth, in your view, should improve next year. If you harken back to earlier this year, the back was forecasting record non-purchased loan growth this year and a record non-purchased loan next. However, on the July conference call, you kind of said there were too many variables at stake to really have a lot of visibility on 2019. So I'm just curious, like what variables have changed? Or do you have more clarity on how and what level of comfort? Obviously, that's your expectation. What level of comfort do you have with that expectation for non-purchased loan growth next year to be better, given that we are expecting record near-term repayments?

Yes. That's a fair question. And I will apologize to our listeners for the fact that on the call 90 days ago, I said I don't know or we don't know that several times in response to different questions. You'll notice we have delayed our call several days from the schedule that we previously operated on. And the reason that we delayed the call is we wanted to produce more informed management comments document and have the time to produce that, so we could give much more detail and written material to our listeners. And also, we wanted to make sure that we had time to really bet some of the questions that we have not had time to fully bet in preparation for the call last quarter. So I apologize for saying we don't know several times last quarter. There are always, of course, things we don't know. But we've done a fair bit of analysis on expected payments on loans next year, prepayments, payments from sales of condos and lots and land and projects stabilizing. We've got some visibility. I think it's pretty good into the path line early in the year next year. So we're feeling fairly well informed and fairly constructive when we say that we think we'll do better on non-purchased loan growth next year than we've done this year. Obviously, having missed in our guidance this year, I'm not going to say any more than that. And we feel good about that. We're optimistic about that. And I'll apologize that we didn't hit our growth numbers this year. We did not fully project or expect the level of prepayment velocity that we've had in the RESG portfolio.

Understood. Earlier, you talked about, I guess, one of the larger credits -- or the largest credit the bank has was over $500 million. But that was just recently, I guess, provided or taken out. What's the largest credit that's outstanding within the RESG portfolio at the moment?

Well, yes, it's not significantly advanced at this time. But that's the largest credit in the RESG portfolio. That was closed on the last week of September. So it's newly mended and just starting to fund. There was a -- it's a 2-tranche funding. We already had the land loan on the project at about 40% or 50% loan to cost, loan to value on the land loan. So we -- when we rolled from the land loan into the vertical construction loan, we did a tranche funding, first tranche funding that was the land part and rolled our loan balance into that. And then the sponsor puts in all their money, and we're the last guys to fund on the vertical construction. So it's a 2-tranche funded deal on that project.

Okay. No, that's helpful. And my final question would be on deposits. It's great to hear that Cynthia and Tyler are looking at new data analytics to make more informed decisions, et cetera. It is, though, our understanding that the company has been pretty data-intensive in the past. And so maybe if you could provide a practical example of some of the potential changes like -- so we can understand how that might improve deposit beta performance. I understand you're now not just targeting geography, but also product type and consumer. But maybe if you could provide at least one example of how that might ultimately improve the bank's deposit costs over time.

Well, I think you guys are just going to have to take our general comments on that. I mean, I'm not going to get into the details of strategy or the details of what we're doing and how we're doing. We are making significant changes. Those have started to be implemented in a simplified version. The degree of sophistication and analytics and algorithms that go along with that will steadily increase as we build up that model and capability, which we're already working on. And we've already significantly increased our data analytics capability. So we're just trying to be much more sophisticated and intelligent about what we do there than we have in the past. And that will be a multi-quarter evolution that probably won't be fully rolled in until late in 2019. But we hope that as we roll it in each quarter, we hope you'll see a few basis points or a couple of basis points quarter improvement in that deposit beta as we do a better job managing that.

And just one last question on deposits. I think in previous quarters, you've given out the New York City deposit balances. I'm not sure if I saw that this quarter. So I think those deposits stood at around $1.8 billion at the end of the second quarter. Could you update us on where those were at the end of the third?

That's a pretty close number. I think it was down maybe ever so slightly from that. But obviously, that will be in our public filing.

Our next question comes from Brock Vandervliet with UBS.

What's the average loan size within RESG?

Go ahead.

Brock, it's Tim. Yes, average size $58 million assuming full funding. So total commitment, $58 million.

Okay. What's the top 10 loan exposure as a percentage of the total portfolio?

I don't have that number.

I don't have that either, Brock. What I can tell you is, every few quarters, or at least annually at the longest, the RESG staff comes and takes a lengthy time in our board meeting and goes through their 50 largest credits. When we did the 50 largest credits at the August board meeting, the 50th credit was $102.5 million.

Okay. I think it would be really reassuring to get some better sense of granularity on that, but...

We can provide some data on that in the -- or perhaps in the Investor Day. We're going to be updating the investor deck shortly.

Okay. And going back to these two credits, what, I guess, I don't understand is I can -- we see charge-offs quite frequently, where -- and there's a storm or some sort of economic change. And the business models just evaporates, and there's a charge-off. And I think everyone can understand that. But these credits you lived with for a long time. And could you walk through kind of -- how many -- were there any off-ramps to either offload risk and sell pieces of these loans or get -- claw more equity out of the borrower or the bartering syndicate? What was done short of this?

Well, good question. And yes, there probably were opportunities where we could have sold these whole loans in the past at higher prices. That was not a strategy we were pursuing. And obviously, despite the outcome on these 2 credits, we have a good working relationship with the sponsors in these transactions. And we think highly of them, and they've worked hard and made really good efforts to make these projects successful. So the shopping center credit, obviously, has suffered from the general malaise and negativity around shopping centers, and particularly shopping centers that are not the super prime -- super high-end shopping centers that are your kind of destination event sort of shopping center. And we continued to work with the sponsor on these. They were never TDRs. They were never restructured. We were always getting the market rate of interest on in these loans that had amortizations on the shopping center loan up until now. And the shopping center loan's not even past due now. It will go past due in October. But it's not -- it was not a delinquent credit at the end of the quarter. So it's been amortizing and been playing down. Yes, if our crystal ball had been better, and we've been able to see the future and so forth, we would have probably put these loans up for sale and sold them when it was a better environment to sell them. But we didn't do that. And in regard to the lot and home development loan, a decision, strategic decision was made to try to accelerate the sale of lots and the progress in that development by developing a significant amount of vertical home inventory to sell. That strategy seemed prudent to us. It seemed prudent to our sponsor. Everybody was on board. That was a good idea. No one foresaw that when that product came to market and the other improvements and enhancements were made to the club, and kind of just the general sentiment around that development, that it would create a wave of resells from existing homeowners that would undercut our sponsor's product and price. So that didn't turn out the way we expected. We make a lot of good decisions. Those were not good decisions.

Going forward, would you be more aggressive in selling pieces of larger loans? I don't think that's been a strategy in the past.

I don't think that the outcome on these two properties in any way is changing our strategy. I mean, we view these as very much isolated issues. And we feel very good about the strategy, with which we have built and operated that portfolio and our bank in general for decades. So no, I don't think this will change our thought in that regard.

Our next question comes from Brian Martin with FIG Partners.

George, just one comment. Just going back to the buyback for a minute. And that was -- I mean, if you -- even if you were to evaluate it at this point, I guess, do you have a sense where, since you've already done some work, it sounded like after last quarter, I mean, the capacity you'd have to do that based on kind of capital levels, just existing capital levels without potentially adding some to do a buyback. But just kind of -- if you could kind of frame that. Or is that not something you can give a little color on?

Brian, anything I would say there would just be speculative, and it's not an issue that's on the table today. So I'd prefer to not comment on that and just let the board continue to make its evaluation of that. And we'll get ready to do something. We'll clearly quantify that.

Yes, okay. I got you. and then just one thing. Going back to the margin for a minute, I guess. If -- I guess, I appreciate the commentary you guys have kind of given and the factors you laid out. But I mean, if you get some better alignment between LIBOR and Fed funds. And as to Tim's point, you get some normalization on those multiple items that you've kind of identified. And the deposit betas do moderate some, just given less growth and the recent initiatives you're talking about. I guess, are we understanding it right? At least your hope would be that the margin or the core spread, however you want to look at it, should be -- have a bit of an upward bias from where we are today, assuming the rate forecast, where we get a couple of more rate increases. And I mean, just kind of -- I know you don't give guidance, but just in general, is that a fair characterization of how you guys would be thinking about the margin today?

Brian, this is Tim. Yes. I mean, to your point on the spread between LIBOR and in either Fed funds or OIS, if you looked at where LIBOR was a year ago and where it is at the end of September, it's up about 100 basis points. So if you look at that spread and how that spread operated throughout that 1-year period, it did kind of -- it did expand in kind of that 4Q, 1Q time frame. And it has contracted down to a more normal range between that spread. So I think we're at a level that is a more normal level between the two and so we would expect that our LIBOR and Fed funds, and therefore our yield on non-purchased loans to be more highly correlated in future quarters.

Okay. So I guess -- what I said, I guess, at least -- so you're kind of -- are you, I guess -- is the belief that the margins should go higher if all that happens? Tim, I guess, I'm not saying it's going to happen, but, I guess, that would be the thought, I guess, if I'm understanding that correct?

And I think you've got to take each component of the margin by itself. You got the purchased loans that are running off. Obviously, those are higher yielding. But that's $654 million when it's yield. And non-purchased yield, even at $607 million, is coming to a closer point, where those could converge in the next several quarters. Deposit costs, we'll give comments around deposit costs around that, too. So I think you've just got to do your math around all those specific components and look at the margins there.

And Brian, I would add to that. We expect to continue to add liquidity. Our liquidity ratio went up several points in the quarter just ended. When we get the call report data, people who file -- people who calculate that will see that liquidity ratio went up quite a bit from June 30 to September 30, where you added $88 million in securities that was part of that liquidity build. Part of it was pre-nup pledged collateral and so forth. We expect to continue to build liquidity. There will be a cost of doing that to our margins. So if you're running the model on this, you ought to factor in a continued increase to some degree in that bond portfolio, $100 million or $200 million, whatever you want to put in, something per quarter of growth in that bond portfolio at whatever. That kind of short-agency paper is at a 3-year average life. That starts yielding about $320 million now, I think. So we're going to continue to put more liquidity on the balance sheet because we think that's a prudent thing to do. So that will weigh on margin. I would echo what Tim said. When we get a point if we do where our non-purchased yields rise to the point that they cross over the line, the yield line for the purchased loan, I think that is a positive point for our net interest margin. Because then you've got the higher-yielding part of the loan book growing, the lower-yielding part of the loan book shrinking. And obviously, the dynamic of that would -- all other things being equal, and everything else is not equal. But all other things being equal, that would have a positive implication for net interest margin.

Okay, that's helpful. I appreciate the color. And just last two things, and I'll step back, was just the expense base, George. If you can just comment on it. This quarter's a pretty good level absent the one-time items you guys called out. And just lastly was the origination volume that you talked about picking up, if there's any reason you can point to. I guess, you've seen that's taking that volume higher and giving you some optimism, the fourth quarter could be the best.

Yes. On the origination volume, what I would tell you is our guys have been working really hard, and they've been out working with a lot of customers on a lot of transactions. And we gained some new customers that we expect to close our first loans work that are really some marquee relationships we've never had before that we hope to close transactions with in Q4. So the origination volume has been garnered by hard work. Hopefully, we can get them all closed. Greg, do you want to comment on the net interest expense?

Yes, yes. Obviously, Brian, the name change and the rebranding cost had a heavy impact on the third quarter. Absent those, I think that the third quarter run rate from a noninterest expense standpoint is fairly clean. We continue to work to finalize the buildout of some of the infrastructure that we've been talking about for a number of quarters. We think we're getting -- starting to the late innings there. We currently still have a goal of continuing to reduce reliance on third-parties' consultants at all as we bring that expertise and have that expertise and bring that in-house. We're continuing to identify, recruit and hire those people. And you have processes. It's continuing, but it is getting towards the latter part of the deal. So I guess, I'd say that and just best circle back around and say, yes, I think the third quarter run rate is fairly clean, particularly if you look at it over the last couple of quarters. Although we are continuing our efforts to bring some of that expertise on board, in-house and reduce our reliance on third parties.

Our next question comes from Catherine Mealor with KBW.

It's a long call. I just had two follow-up questions. The first was just on the growth and in your confidence for growth for next year. And George, do you believe that your confidence for better growth next year is based more on what you see in your pipeline for funding the products that you've already committed or more of the pipeline of potentially new commitments that you're gathering right now?

Catherine, it's a mixture of both. I mean, obviously, if we don't originate any new loans, we will not have growth next year. But we do have a strong top line of loans already closed that are funding, and we're getting into a good funding cycle on a number of those projects next year. So it's a mixture of both. We've got to continue to originate new volume if we're going to have growth. So -- and that's nothing different from what it's always been.

Got it. But could you argue the funding cycle for this upcoming year is better than it was this past year?

I don't have comparable data to know that. I know what's in our forward projection. I haven't looked at and compared that to where we were a year ago. But I know what's in our projection that Tim does. And Tim, you might have some color on that.

Yes. Catherine. I mean, the funding is from a quarter-to-quarter, it obviously increases slightly based on just the projects we have and very slightly from quarter-to-quarter. But it's a fairly consistent funding per quarter as these projects are completing throughout their life cycles. So -- and I would expect 2019 fundings to be on average slightly better than they were in 2018. But -- and then you got originations that kind of come in there as well. And as we said earlier in the call, prepayments I think will vary from quarter-to-quarter. So you can't have different results from those prepayments on a quarter-to-quarter basis.

Okay, got it. That's helpful. And then my last question is just as you think about the higher charge-offs that we saw this quarter, how do you think that's going to change the regulator's comfort in your commercial real estate to capital ratio over 300%? And could that impact your future growth or your capacity for future growth or even your capital ratios?

I think our regulators are informed and intelligent enough to look at this and realize these were two assets we had marked as substandard. We knew there was exposure in them. We didn't know how much exposure until we got the new appraisals. But we had those things adequately reserved for based on the appraisals that were within a year old at the time or 2 years old at the time. So I think the regulators know we're doing the right thing. And hopefully, a couple of one-off credit issues that we had will not change their attitude. I would not expect it to do so. One thing I would point out in that regard is that even with all the negativity and the anomaly that we had regarding our yield on non-purchased loans and everything else, we still generated a 1.33% annualized ROA for the quarter. So when you have a really bad quarter and you generate a 1 33% return on assets for the quarter, that's not the end of the world sort of outcome. It -- the credit issues this quarter were a black eye. They're not a malignant melanoma or anything. So keep it in perspective as we go forward.

And I'm not showing any further questions at this time.

All right. Thank you, guys, for all joining our call today. We appreciate your questions and your attention, and look forward to talking with you in about 90 days. Thank you so much.

Ladies and gentlemen, this does conclude today's presentation. You may now disconnect, and have a wonderful day.