Heartland Financial USA Inc (HTLF) 2017 Q2 法說會逐字稿

Greetings, and welcome to the Heartland Financial USA, Inc. Second Quarter 2017 Conference Call. This afternoon, Heartland distributed its second quarter press release, and hopefully, you've had a chance to review the results. If there is anyone on this call who did not receive a copy, you may access it at Heartland's website at www.htlf.com.

With us today from management are Lynn Fuller, Chairman and Chief Executive Officer; Bruce Lee, President; Bryan McKeag, Executive Vice President and Chief Financial Officer; and Andrew Townsend, Executive Vice President and Chief Credit Officer.

Management will provide a brief summary of the quarter, and then we will open up the call to questions from analysts.

Before we begin the presentation, I would like to remind everyone that some of the information management will be providing today falls under the guidelines of forward-looking statements as defined by the Securities and Exchange Commission. As part of these guidelines, I must point out that any statements made during this presentation concerning the company's hopes, beliefs, expectations and predictions of the future are forward-looking statements, and actual results could differ materially from those projected.

Additional information on these factors is included from time to time in the company's 10-K and 10-Q filings, which may be obtained on the company's website or the SEC's website.

(Operator Instructions) As a reminder, this conference is being recorded.

At this time, I'd like to turn the call over to Mr. Lynn Fuller at Heartland. Please go ahead, sir.

Thank you, Tim, and good afternoon. We appreciate everyone joining us today as we discuss Heartland's performance for the second quarter of 2017.

For the next few minutes, I'll touch on the highlights for the quarter. I'll then turn the call over to Heartland's President, Bruce Lee, who will cover progress on our key operating strategies. Then, Bryan McKeag, our EVP and CFO, will provide additional color on Heartland's quarterly results, followed by Drew Townsend, our EVP and Chief Credit Officer, who will offer insights on credit-related topics.

Well, I'm very pleased to begin with news that Heartland reported an excellent second quarter with favorable results in every key performance measure. The company experienced solid credit quality, continued growth in demand deposits and positive improvement in our efficiency ratio. And just after quarter-end, we announced the completion of the most significant acquisition in our company's history.

And now on to the specifics.

For the quarter, Heartland earned net income available to common shareholders of $21.9 million, a 5% increase over the second quarter of 2016.

Now on a per-share basis, Heartland earned $0.81 per diluted common share for the quarter.

And year-to-date, net income available to common shareholders was $39.9 million or $1.49 per diluted common share, in line with earnings of $40.8 million in the first half of 2016.

Return on average assets for the quarter increased to 1.06%, while return on average common equity was 11.13%, and return on average tangible common equity was 14.07%. Our tangible common equity ratio improved to nearly 8% for the quarter, reaching the top of our target range for this ratio.

Likewise, book value and tangible book value per common share continued to increase, ending the quarter at $30.15 and $24, respectively.

Now looking at the balance sheet. Total assets moved down slightly to $8.2 billion for the quarter as we've been implementing our strategy of shedding lower-margin business and exiting shared national credits, with the intent to ensure we remain beneath the $10 billion asset threshold in 2017.

With the recent closing of Citywide Banks' acquisition, which added $1.4 billion in assets, we are now at $9.6 billion in total assets. This leaves us with sufficient room for loan growth to new and existing clients, while remaining below $10 billion through the end of 2017 and delaying the negative impact of Durbin until the mid-2019.

Heartland's security portfolio currently represents 25% of assets. With our target at 20%, we still have room to convert cash flow from our securities portfolio into loans. Currently, our available-for-sale portfolio duration is 4.5 years, with the portfolio yield now at 3.1%.

Overall credit quality remains sound, with nonperformers consistent with previous quarters and net charge-offs decreasing from the first quarter. And in a few minutes, Drew Townsend will provide more detail on credit-related topics.

Total assets eased a bit in the second quarter, though the key non-interest demand deposit category increased, now representing 34% of total deposits. And in a few minutes, Bruce Lee will comment on loans and deposits in more detail.

Net interest margin continued strong for the quarter at 4.14%, reflecting reductions in funding costs and improved yields on our securities portfolio.

Heartland continues to carefully manage non-interest expense, and we are pleased to see positive results as reflected in our efficiency ratio, which dropped into the 65% range for the quarter. And in a few minutes, Bryan McKeag, will address non-interest expense in more detail.

Now on the M&A front. Early this month, we announced the completion of our acquisition of Citywide Banks of Colorado with $1.4 billion in assets. This transaction represents a major milestone for our company in terms of asset size, additions of new talent and sizable market share in one of America's best growth markets, Denver, Colorado.

With several opportunities still on the table, we remain committed to reaching $1 billion or more in assets in each state where Heartland operates. However, we will pursue only those deals that will be accretive to our shareholders' earnings per share, producing minimum internal rate of return of 15%, and a maximum earnback of 3 years.

In concluding my comments today, I am pleased to report that at its July meeting, the Heartland Board of Directors authorized a dividend of $0.11 per common share, payable on September 1, 2017.

I'll now turn the call over to Bruce Lee, Heartland's President, who will provide an overview on the company's strategic initiatives. Bruce?

Thank you, Lynn. Good afternoon. I'm pleased to comment today on the results achieved by Heartland's 10 community banks for the second quarter of 2017. Turning first to lending, total loans decreased slightly in the quarter as organic loan growth remains a challenge.

On a more granular level, however, we see several improving trends that provide good reason for optimism. First is new money advanced in our commercial and ag portfolios, which totaled $214 million in the second quarter. This represents an increase of $91 million over last year's quarter, and $48 million over the first quarter this year.

A second indicator is the increase in brand-new borrowing relationships for the banks, with $34 million of new advances in the second quarter compared to $23 million in the first quarter.

We are also encouraged to see net increases in loan portfolios at 5 of our banks in the second quarter, a significant turnaround from the first quarter when none of the banks reported organic loan growth.

Additionally, pipelines are better now than we have seen in recent quarters. And as a group, our bank presidents are optimistic about net loan growth for the next 2 quarters.

To summarize, we are seeing encouraging results at the banks, with loan growth heading in the right direction and new originations closer to offsetting runoff. Our lending teams are hitting on more cylinders, and we are receiving optimistic reports from the field.

One more point on loan growth deserves mention, and that's our pricing strategy, which is based on the value we bring to the client relationship. That value includes the credit experience and skills of our bankers, our knowledge of local markets and a strong emphasis on relationships rather than transactions. Price competition for quality loans remains intense, and as we have stated previously, we continue to practice our disciplined approach to credit and have not stretched in pursuit of quantity.

Looking now at deposits. Total deposits eased in the second quarter, consistent with some previous years' second quarters, when depositors pay tax obligations. We are very pleased to see continued growth in non-interest-bearing demand deposits, which increased in total dollars and grew 34% of total deposits -- grew to 34% of total deposits. We expect to see this proportion increase further in the third quarter with the addition of the Citywide Banks deposits, which includes 44% in non-interest-bearing demand.

Heartland is well positioned relative to funding, with a strong base of core deposits and very little in brokered or FHLB advances, providing ample room to secure additional funding for loans when needed. Related to upward pressure on deposit costs, we are seeing some, but not to the extent we see on loan pricing.

Moving to our mortgage business. Residential real estate originations increased from the previous quarter, driven by seasonality. In comparison with the previous second quarters, however, originations have decreased, reflecting both a softer mortgage market and our scaled-down mortgage unit that now operates within the Heartland banking footprint.

We continue to fine-tune our mortgage operation to respond quickly to changes in demand. Additionally, we are implementing a plan to offer an online mortgage product in the coming months.

Heartland's loan -- Heartland's mortgage loan servicing portfolio exceeds $4.3 billion on June 30, an increase of over $100 million during the past year. At June 30, we have $31.6 million of MSRs on our books with a fair value of approximately $45.9 million or $14.3 million more than book value.

On previous calls, I've highlighted the progress of Heartland's commercial card payment solutions unit. The unit stepped higher in 2016, rising 11 spots to rank #54 among all U.S. Visa and MasterCard commercial issuers according to data released last month by the Nielsen report.

Our unit recorded purchase volume of $264 million, an increase of 89% over 2015. Similarly, our units purchasing card moved up to rank 28th, among purchasing card issuers with an impressive increase in purchase volume of 199%.

Finally, I want to provide an update related to our recently completed acquisition of Citywide Banks in Colorado. As Lynn mentioned, this was our most significant acquisition, providing scale and visibility at one of America's best growth markets. We intend to capitalize on Citywide Banks' 11% trend in loan growth and the bank's superior deposit mix, with approximately 44% in demand and 90% of total deposits in non-time.

We're also excited to unite with a highly-respected community bank in the Denver market that enjoys a solid 53-year reputation for its community support in sound banking practices.

The process of integrating our 2 companies is well underway, with the Citywide systems conversion planned for October. With the merger now complete, senior bank leadership is in place and has hit the ground running led by longtime Denver banker, Kevin Quinn. Notably, the credit cultures of both organizations are very similar, and the management teams are working closely together. We anticipate a smooth transition in the coming months.

With all the positive activity taking place at the Heartland member banks, I am really looking forward to the second half of this year. With that, I'll now turn the call over to Bryan McKeag for more detail on our quarterly financial results.

Thanks, Bruce, and good afternoon. I'll begin my comments today with the tangible common equity ratio, which showed good improvement again this quarter, increasing 47 basis points to 7.97%. The increase in the market value of our investment portfolio and derivatives added 5 basis points to the ratio, with the remaining 42 basis points increase coming from retained earnings.

Lynn and Bruce have already commented on loans and deposits, so the only other balance sheet item I'll comment on is borrowings, which declined a total of $17 million this quarter and have declined $175 million since the end of last year. This is largely the result of our ongoing efforts to delever the balance sheet by shedding low margin business.

Moving to the income statement. Net interest income totaled $74.6 million this quarter, up $1.6 million from the prior quarter. The drivers of the increase were an increase in average earning assets of $84 million and one additional calendar day, partially offset by a slight decline in net interest margin.

The net interest margin on a tax-equivalent basis remains strong at 4.14%, which is a 2 basis point decline from last quarter.

Yields on loans increased 2 basis points, while investment yields decreased 8 basis points, and interest costs on deposits and borrowings declined 1 basis point compared to last quarter.

This quarter, the net interest margin includes 12 basis points from the amortization of purchase accounting discounts, which compares to 14 basis points in the prior quarter.

Non-interest income totaled $25.6 million for the quarter, down $300,000 from last quarter. When compared to last quarter, gain on sale of securities was down $1.1 million, and the gain on sale of loans for the quarter was up slightly or $700,000, reflecting modestly higher loan application activity. That was up almost 25% from last quarter.

Service charges and fees increased again this quarter, up almost $250,000 over last quarter, primarily from continued growth in commercial card volumes, as Bruce mentioned in his comments.

Shifting to non-interest expense. Total non-interest expense was $69.3 million this quarter, a decrease of $2.4 million from the prior quarter, with most categories showing improved or flat changes from Q1.

M&A-related costs were approximately $1 million this quarter compared to $1.6 million last quarter.

Our largest expense category, salary and benefits, decreased slightly or $600,000 as compared to last quarter.

Professional fees were down $700,000 from last quarter as consulting and other costs related to M&A activities were down $600,000 compared to the prior quarter.

Advertising costs were down $1.1 million this quarter, reflecting a more normalized level, as the prior 2 quarters included elevated costs related to deposit campaigns.

And finally, other non-interest expenses were up $800,000 over last quarter, as this quarter included over $800,000 of additional write-downs on partnership investments and tax-related projects, while there were none in the prior quarter.

For the quarter, the efficiency ratio was 65.61%, down considerably from the 69.95% last quarter, as core operating revenues increased $2.3 million, primarily due to higher net interest income. And core operating expenses decreased $2.8 million, primarily due to lower advertising and M&A-related costs.

The effective tax rate for the quarter was 26.85%, which included the benefit from several tax credits. We believe the normalized tax rate of 30% to 31% is reasonable on a full go-forward basis.

Next, I'll share some comments regarding the impact of the Citywide acquisition, which closed on July 7, 2017. First, the addition of Citywide will increase loans by $980 million, net of a loan mark in the 2% range. Investment should increase $230 million. Total deposits will increase $1.2 billion, of which $530 million or 44% are non-interest-bearing demand accounts.

Second, at closing we transferred $58.6 million of cash and issued approximately 3.2 million shares of Heartland common stock to Citywide shareholders. This will increase Heartland's common equity by about $153 million. We are currently estimating goodwill at $94 million and the core deposit intangible at $16 million. Obviously, these estimates could change slightly once all the fair value marks are finalized.

So with all that in mind, our expectations for Heartland's results for the rest of 2017 would be the following:

First, net interest margin on a tax-equivalent basis will remain strong in the 4.15% range, and this is inclusive of Citywide's core margin, which has been slightly above 4%, and its purchase accounting impact.

Mortgage production in Q3 is expected to be somewhat better than Q2 and then show the normal seasonal decline in Q4.

Core fee income, exuding mortgage and security gains, is expected to show continued improvement, and then will increase with the addition of Citywide, which the run rate is about $1 million per quarter.

Core expenses, excluding M&A-related costs, should remain well controlled at $68 million to $68.5 million per quarter, and then will increase with the addition of Citywide with a run rate of just over $9.7 million per quarter.

Further, we anticipate cost saves to be about $2.5 million per quarter, starting in December, and should be fully phased in by January 2018.

In addition to the core expenses, we will also see higher costs for M&A and system conversion activities for Citywide, which we estimate to be between $3 million and $3.5 million. A large portion of those costs will be in the third quarter with a much smaller portion in the fourth quarter.

And with that, I'll turn the call over to Drew Townsend, our Executive Vice President and Chief Credit Officer.

Thank you, Brian. This afternoon, I'll begin my credit-related remarks by discussing the changes in Heartland's nonperforming loans during the first quarter -- second quarter.

Total nonperforming loans increased by $2.4 million in the second quarter or from 1.19% to 1.24% of total loans. During the quarter, $7.4 million of nonperforming loans in all loan categories were resolved. New nonperforming loans identified during the quarter equaled $13.7 million, of which $12.7 million were originated by the Heartland member banks, and $1 million came from Citizens Finance, Heartland's consumer finance company.

The new bank nonperforming loans by loan type included $5.8 million or 46% attributed to the commercial loan portfolio, $1.2 million or 9% in the agricultural portfolio and $5.7 million or 45% from the retail portfolios.

Heartland-wide, there are only 6 nonperforming borrowers with loans outstanding exceeding $1 million. In aggregate, these 6 borrowers total $30.1 million or 46% of total nonperforming loans.

In the retail portfolios, $16.3 million or 25% of total nonperforming loans are repurchased residential real estate loans from our service loans portfolio. This represents a $2.1 million increase in total nonperforming loans. The majority of these loans are government-guaranteed, and our estimated loss exposure is considered to be minimal.

For strategic reasons, a change was recently implemented in which all new loan originations in the government-guaranteed residential real estate portfolios are now being originated and sold on a service-released basis.

Other real estate owned decreased from $11.2 million in the first quarter to $9.3 million in the second quarter. The reduction was due to several commercial and retail property sales that occurred in the second quarter.

In total, nonperforming assets as a percent of total assets, increased slightly from 0.90% as of March 31 to 0.93% as of June 30.

Based on the current information available on the legacy Heartland loan portfolios as well as our understanding of the credit metrics of the acquired Citywide Banks, it is anticipated that the ratio of nonperforming assets to total assets will show improvement in the third quarter of 2017.

When reviewing Heartland's overall loan quality metrics during the second quarter, the company continued to demonstrate a favorable level of total sub-rated loans, those risk-rated watch or substandard. At 6.92%, the non-pass credits as a percent of total loans are at a level which compares very well to most quarters over the past several years.

With respect to delinquency totals, the 30 to 89 days delinquency ratio reduced from 0.44% to 0.38% from the prior quarter to the current quarter. This level of delinquency remains consistent and within the range of those percentages reported in the last several quarters.

As we review the allowance for loans, it is noted that provision expense was $889,000 during the second quarter. The lack of organic loan growth, the lower volume of acquired loans, which moved out of the purchased accounting pool, generally positive changes in overall credit quality in the performing portfolio and the variability of other factors considered in the allowance calculation, all were contributing factors to Heartland's second quarter provision expense being reduced by $2.8 million from the $3.6 million reported during the first quarter.

It remains noteworthy that $853 million of loans from our most recent acquisitions still reside in the purchased accounting pool and are covered by the valuation PCI reserves.

As credit decisions are made on these loans in future quarters, a provision expense will be necessary to establish the associated allowance for these acquired loans.

As shown in the earnings release, our coverage ratio of allowance for loan losses as a percent of nonperforming loans was 81.78% in the second quarter, down slightly from 86.29% in the first quarter. The allowance for loan losses as a percent of total loans remained relatively unchanged, decreasing slightly from 1.03% to 1.02% this quarter. The valuation reserves totaling $22.8 million are recorded for the aforementioned loans obtained from acquisitions.

Excluding those loans covered by the valuation reserves would result in allowance to loans ratio of 1.20% as of June 30, down slightly from the March 31 ratio of 1.22%.

In summary, nonperforming assets as a percent of total assets increased slightly during the second quarter due to a modest increase in nonperforming loans as previously outlined. Conversely, total sub-rated loan levels as of quarter end continued to show improvement, and trends for various other asset quality metrics, including delinquency levels, net charge-offs and total other real estate owned, continued to demonstrate favorable results.

Overall, the outlook for improvement in credit metrics in the third quarter is considered to be positive.

That concludes my remarks. I will turn the call back to Lynn and remain available for questions.

Thanks, Drew. Now, Tim, we'd like to open the phone lines for questions from our analysts.

(Operator Instructions) Our first question comes from the line of Jeff Rulis of D. A. Davidson & Co.

This is Matt on for Jeff. My question is about loan growth. What are your expectations for loan growth in the second half of 2017? And also, are you seeing any particular strength in any geographical areas?

Matt, as I mentioned, we had 5 of our 10 banks that did show positive loan growth during the quarter, so I would start with those banks. And we do expect to see positive organic loan growth in the last half of the year. I'm not sure I'm quite ready to talk about what -- the percentage growth, but we do expect it to be positive. And we think you can see the momentum that we've had starting to grow in that direction. And we think the addition of Citywide also will help that with their $1 billion of new loans and their historically positive growth trends.

Our next question comes from the line of Damon DelMonte of KBW.

So my first question was I think, Lynn, in your opening comments, you referenced that, in an effort to remain under the $10 billion asset threshold, you guys have been kind of working some credits off the books. I think you mentioned maybe some shared national credits and maybe some loan categories that you don't view as core. Could you just kind of refresh those comments for us? I didn't get everything that you said.

Yes. It was pretty simple, Damon. Basically, we had some shared national credits that amounted to almost $25 million that we moved out. They were transactions, okay, so there wasn't a full relationship. And then looking across the member banks, at any type of business that really isn't core, the whole purpose here is that we want to keep our high-margin, well-rounded relationships and ensure that we can still add to those without crossing that $10 billion mark until 2018. And we should have plenty of room to do that. We bounce between $9.6 billion and $9.8 billion in total assets with the addition of Citywide Banks, and that gives us plenty of room. Plus we've still got cash flow coming off of our bond portfolio. We're running 25% to 26% of our assets in our bond portfolio, which more typically would be at 20%. So we can certainly take the cash flow off of that, put it into quality loans for both new and existing relationships. And then there's still probably about 6 to 8 different strategies on the liability side that we could employ if we start bumping up against that $10 billion in 2017. So we've had these strategies in mind throughout the year because we know that we'll cross that $10 billion mark in '18, but we clearly don't want to cross that in '17. Durbin -- so the ultimate effect of this is Durbin won't impact us until July of 2019. And we think that gives us plenty of time, that 18 to 24 months, to both grow organically and pick up additional acquisitions to get our total assets up close to that $12 billion mark, which will easily cover the loss in Durbin fee income and cover some of the cost of passing that $10 billion asset size.

Got you. Okay, that's helpful. I appreciate that color. I guess, Bryan, you had mentioned there's about $1 million of merger charges that were incurred this quarter. From a modeling standpoint, where would we find that broken out in the non-interest expense categories?

I'd say most of it is sitting in the professional. I think there's a little bit probably down in other. Some of our travel expenses are up, et cetera, as we've had to travel out to Denver to start getting ready for system conversions and things like that.

Got you, okay. And the just lastly, if we look at the core margin, could you just talk a little bit about some of the dynamics that you're seeing just given the couple of rate increases that we've seen over the last few months?

Yes. I think, there's lots of things going on in our margin with purchase accounting and everything. But the core margin really has been, every time you dial back to 12 to 15 basis points from purchase accounting, has been right in -- on a tax-adjusted basis, right in the 4% range. And as rates have moved up a little bit, we've seen a little less premium amortization so that's helped the investment books yield. And then the adjustment. So when we model our internal rate -- interest rate scenarios, a 25 basis point move to us is about 6 basis points or so if deposit rates don't move. And so this last time, deposit rates didn't move, but the last move was in June. So we didn't get a whole lot of that this quarter. We might get -- maybe we got 1 basis point this quarter. We might get 2 or 3 more basis points next quarter. All of that is flowing through. And then you've got the purchase accounting that will go down. And when Citywide comes in, they'll have a new piece of purchase accounting. So I think it's all going to blend out, and I think it's going to hang kind of where it is, maybe up a tick or 2. And then, we'll see what happens with purchase accounting and the Fed.

Okay. And your cost of interest-bearing deposits is hanging in there pretty good. Are you starting to see more pressure on those accounts at all?

Not really. Even though, again, interest rates have been moving up, we haven't seen that much upward pressure on deposit rates yet. It's much more competitive on loan pricing than it is on deposit pricing right now.

Our next question comes from the line of Steve Moss of FBR.

I was wondering with regard to M&A, if you could provide us any update with regard to discussions you're having in pricing?

Well, yes. Clearly, pricing has gone up. And I mean, we've been pretty disciplined, as I mentioned. Our transactions absolutely have to be accretive to our current shareholders' earnings per share. We certainly on, what I would consider to be conservative forecasts and conservative assumptions, we need at least a 15% IRR. And then finally, we like to have an earnback in 3 years. So thank God, we've got a very deep pipeline of acquisition opportunities, and we'll run our models. And we basically tell the sellers that we pushed it as far as we can to make sense out of it. And we've been fortunate that our model is quite attractive to community banks, and they understand that if we report an acquisition that is not seen as being favorable to the market, that hurts them as much as it hurts us because most of these banks are taking a high percentage of Heartland's stock with the acquisition, in the range of anywhere from 70% stock to as much as 90% stock, the balance in cash. So to the extent we can do deals that make sense, our stock price continues to go up, and we would hope that the sellers would see that as attractive as we see it. So we won't get every deal, Steve, but we've got a deep pipeline so we can move to the next deal.

Okay. And then with regard to the loan growth -- the loan pipelines here, is it more C&I-bent? Or is it commercial real estate that is driving the improvement on the pipelines?

Yes. It's much more operating companies and less commercial real estate. As we've talked in previous quarters, we've worked very hard both in Colorado and in California, to reduce our overall real estate exposure. When we made those acquisitions, they were a little heavier than we would've liked them to be. So we worked that down. Now we have plenty of room to grow real estate, but our focus over the last really, 3 quarters, has all been on strategic calling on operating companies.

Okay, that's helpful. And with regard to mortgage here, I was just wondering how much was volume down year-over-year and if you could give color around gain on sale margin versus last year.

Yes, this is Bruce. Volume is down about 30%, and margin is actually up. So we're pleased with our margin. It's been one of our initiatives in that business. And the volume, part of that decrease again is strategic as we're now feeling the full effects of exiting several of out-of-footprint offices. And the balance is really just a decrease in the overall mortgage market. Remember that our -- about 70% of our volume is purchase volume, so we expect to have strong summer months because of that focus as opposed to the refi business.

Our next question comes from the line of Andrew Liesch of Sandler O'Neill.

Just one question that I have that you guys haven't covered yet. Just going back to the commercial real estate in Colorado. I know Citiwide has got a pretty diversified loan book. But are there any types there that you may be looking to -- that may be higher-risk or portfolios that you don't want to retain that we might want to -- that we might see decline in the coming quarters?

No. We're very pleased with their portfolio and their diversified mix. We like their -- the asset quality. We like the asset classifications. And my comment earlier about Colorado was really as we were deleveraging the real estate book at Centennial. And so right now, we are very pleased with the combined portfolio from a mix standpoint.

Andrew, this is Bryan. I just wanted to add in here that when we look at our 300 and 100 levels, I think that our 300 level is right at about 155%, and we're like 50%-something on the 100 level for construction. So we're now in a spot where we've got that in a pretty good place. So we're comfortable with the level. Even when Citywide comes in, we're comfortable with them coming in with a little bit higher CRE. So that will all work good.

We have plenty of room to grow our real estate portfolio in Colorado. And again, as we've been talking, our major focus for the last 3 quarters has all been on operating companies, and that's consistent with Citywide's philosophy as well.

Andrew, I think there's something unique about the Citywide real estate relationships as well. And that is in many cases, where you see investors in non-owner occupied real estate, they generally don't have a lot of deposits. They don't have a lot of non-interest-bearing DDA. In the case of Citywide's clients in this area, they tend to be very good-sized depositors, and they tend to turn their properties. So we really like the mix of business that they have in that area. It's quite different from what you might see at other community banks.

Just I guess one other question on the mortgage banking now that I think of it. So with the new online offering plus expanding the -- being, I guess, larger in California and larger in Colorado, would you expect your gain on sale numbers to be better next year than this year?

We've been working very hard with margin, with the online offering and expanding. We would expect them to be slightly better as we build out both Colorado, California. And also, we're building out Arizona. So we would expect them to increase, and we expect them to be, as Bryan mentioned, better in the third quarter than they were in the second. Seasonally, the fourth quarter is tough, especially December. And we'll have to get a little closer to year end before we start forecasting out what '18 will be. But as we grow out those markets, those 3 that I mentioned, we would expect to see more volume.

Our next question comes from the line of Nathan Race of Piper Jaffray.

A question on the agricultural portfolio, first off. Saw some nice seasonal growth this quarter. I'm just curious how much of that is just seasonal increases. And if you're seeing any additional -- intentional runoff of some borrowers that may not be of the financial merits that you guys typically look to portfolio?

Nathan, this is Drew. I would say, probably the former would be my estimation of the growth. I think it had some seasonal increase there. I think there was a few million dollars of growth. I wouldn't say there's any particular segment or -- we've had, again, as we've alluded to maybe in past quarters, we've had a handful of downgrades in this portfolio. I mentioned one specifically. And the good news, that one actually got paid off today. That was one of the non-performers that was a growth piece for us in the second quarter. So that one actually went away. So no, overall, I mean, I think we're kind of steady as she goes. We haven't been real quick to step on the gas or pull back. I think we try to be pretty measured in our approach always. So the portfolio obviously is experiencing some strain with the different commodity prices where they're at. But overall, I think we're in pretty good shape.

I would suggest, Nathan, we don't see much in the way of losses in that portfolio because we've been very careful not to over-leverage those borrowers. We've kept loans to appraise values very, very low. Because we stress those credits to see what happens to them with the rising rate and low commodity prices. And when you do that, you can see that they just can't afford a lot of debt on the land.

And the only thing I would add, Citywide and I know you're probably already aware of this, but Citywide doesn't bring any ag to the portfolio per se. So we don't expect an increase due to the latest acquisition.

Right. Got it. And just one housekeeping item. Bryan, in terms of the Durbin impact in the back half of 2019, I have in my notes, just under $2 million per quarter. Is that accurate?

It would be pretax. So on an annual basis, pretax should be $5 million when we get to $10 billion. But by the time we get to Durbin, which would be June of '19, we may be bigger. So that number could be bigger at the time it actually impacts. But I would say it shouldn't probably be more than $6 million. And in fact, just thinking about it today, we have been modeling kind of our run rate out to $10 billion. I think when Citywide comes in, I think they even have less on a percentage basis, of debit income than we do. So we may even be lower than the $5 million I've been projecting because as you blend that in and add their assets, we might be running at a little bit lower rate. So hopefully that helps.

(Operator Instructions) Our next question comes from the line of Daniel Cardenas of Raymond James.

Just a couple of questions. On the new mortgage online product that you guys have rolled out, is that a national product, or is that just within your footprint?

No. It's really intended to be in our footprint. It's not a national product at all. It's really meant just to provide another channel as borrowers in our footprint would like the option to go through multiple channels. Some of the people like to at least start it online, and then it's usually finished with an MLO.

Okay. And how long has that program been in operation?

We haven't started it yet. We're going to launch it in the third quarter, we believe.

All right, okay. Fair enough. And then just as a quick reminder. How big is your [SNIC] portfolio at the end of the quarter? And do you anticipate any additional rundown in that portfolio in 3Q?

Daniel, this is Drew. The [SNIC] portfolio is at 0. So if we have additional rundown, something went wrong.

There are no further questions over the audio portion of the conference. I would now like to turn the conference back over to management for closing remarks.

Thanks, Tim. So in closing, we're very pleased with Heartland's excellent financial performance for the second quarter as well as the first half of 2017. And just to summarize, for the quarter, we maintained our margin at 4.14%. We continued improvement of our efficiency ratio, which is now in the 65% range. We maintained solid credit quality and actually improved our charge-off ratio. We improved our deposit mix with non-interest-bearing demand deposits now representing 34% of total deposits. And we strategically managed our balance sheet to remain under $10 billion in total assets for this year, while advancing numerous prospects for acquisition thereafter, which will help mitigate the adverse impact on both revenue and the cost of regulatory compliance.

I'd like to thank everybody for joining us this afternoon. I hope you can join us again for our next quarterly conference call, which will be October 30, 2017. Have a good evening, everyone.

This concludes today's conference. Thank you for your participation. You may disconnect your lines at this time. Have a wonderful rest of your day.