Jan 28, 2020
Operator
Good day, and welcome to the Flagstar Bank's Fourth Quarter 2019 Earnings Call. Today's conference is being recorded.
At this time, I would like to turn the conference over to Mr. Kenneth Schellenberg.
Please go ahead, sir.
Kenneth Schellenberg
Thank you, Anna, and good morning. Welcome to Flagstar Fourth Quarter 2019 Earnings Call.
Before we begin, I would like to mention that our fourth quarter earnings release and presentation are available on our website at flagstar.com. I would also like to remind you that any forward-looking statements made during today's call are subject to risks and uncertainty.
Factors that could materially change our current forward-looking assumptions are described on Slide 2 of today's presentation, in our press release and in our 2018 Form 10-K and subsequent reports on file with the SEC. We are also discussing GAAP and non-GAAP financial measures, which are described in our earnings release and in the presentation we made available for this earnings call.
You should refer to these documents as part of this call. With that, I'd like to now turn the call over to Sandro DiNello, our President and Chief Executive Officer.
Alessandro DiNello
Thank you, Ken, and good morning to everyone listening in. I appreciate you taking the time to join us today.
In addition to Ken, I'm joined this morning by Jim Ciroli, our Chief Financial Officer; Lee Smith, our Chief Operating Officer; and Kristy Fercho, our President of Mortgage. As usual, I'm going to start the call by providing a high-level view of our performance for the quarter.
Then I'll turn the call over to Jim for details on our financial results. Lee will follow with a review of our business segments and strategic initiatives, and I'll conclude with guidance for the first quarter before opening up the line for questions.
As reported earlier today, we produced strong fourth quarter earnings, capping off what I consider perhaps the most successful year in our company's history. For the quarter, we posted net income of $58 million or $1 per diluted share, up 38% from the adjusted results in the same quarter last year.
For the full year, we achieved adjusted net income of $199 million or $3.46 per diluted share, topping our adjusted results for the full year 2018 by 15%. As our collective prepared comments will highlight, our fourth quarter results build upon the momentum we have seen in all of our business lines over the past year.
As I've said in prior quarters, we built this company to have diverse revenue streams and a flexible balance sheet, which provide us optionality. In 2018, we were able to grow earnings despite a challenging mortgage market that was characterized by narrowing margins and lower volumes.
We did this by growing our community bank and servicing businesses and by aggressively managing costs in our mortgage business. In 2019, we faced pressure on our net interest margin due to declining rates, but we were able to temper this by capitalizing on favorable conditions in the mortgage market.
The result, we grew diluted earnings per share by $0.44 for the full year compared to 2018 and grew our tangible book value per share by $4.67. We continue to operate in a beneficial mortgage market with our mortgage teams taking advantage of that in a disciplined way.
The investments we made in people and technology, not to mention the expansion into the bulk correspondent business and more broadly into the through acquisitions and lift-outs continue to pay off. As always, we are managing our volume to maintain industry-leading service levels while producing margin and revenue we haven't seen in a very long time.
Our performance in the fourth quarter and full year 2019 was strong in all 3 of our primary business lines. These results further validate the power of our business model.
Our servicing business continues to grow and perform well, while our banking team focuses on adding high-quality relationships, the kind of relationships that produce spread income, even in uncertain interest rate environments, tapping it off is our mortgage segment, which we built to leverage scale, operate strategically in all economic scenarios. All in all, I think we are positioned well heading into 2020.
I also think we've done a great job of keeping credit quality in check. It's front and center with us all the time.
We keep doing that, and our business model provides a natural hedge against volatility in earnings. When interest rates are falling and net interest margin is narrowing, our mortgage business becomes more profitable, offsetting the narrower NIM.
When interest rates are rising, our NIM improves and offsets the reduced profitability for the mortgage business. Meanwhile, our servicing business continues to grow, adding fee income that is quite resilient to interest rate fluctuations.
As a result, our ROAs and ROEs, quarter-by-quarter, have been pretty consistent and strong over the last two years, a period in which we experienced quite a bit of volatility in interest rates. That concludes my comments.
Now I'll turn the meeting over to Jim for more detail around the numbers.
James Ciroli
Thanks, Sandro. Turning to Slide 6.
Our net income this quarter was $58 million or $1 per share. This performance compares to the $63 million or $1.11 per share last quarter and adjusted net income of $42 million or $0.72 per share in the same quarter last year.
For the full year, we had adjusted net income of $199 million, $3.46 per share, a 15% improvement in earnings per share over the adjusted $3.02 we earned in 2018. Diving deeper into this quarter's performance.
Net interest income was up $6 million or 4% over the prior quarter. Average earning assets grew $1.7 billion, mostly in our loans held for sale and warehouse loan portfolios.
The net interest margin decreased by 14 basis points, primarily due to the impact of the September and October rate cuts. Additionally, we took advantage of the favorable mortgage environment to grow our loans held for sale portfolio, which brings more net interest income dollars, albeit at a lower NIM.
We'll review these numbers further in a couple of slides. Mortgage revenues were $98 million, a decrease of $10 million or 9% compared to the prior quarter.
During the quarter, we optimized volume and margin and continue to capitalize on the favorable mortgage market, resulting in a gain on sale margin of 123 basis points, an increase of 3 basis points compared to the prior quarter. Asset quality remained strong as net charge-offs were only 10 basis points, and our nonperforming loan and allowance coverage ratios remained flat at 21 basis points and 0.9%, respectively.
We'll provide more details when we get to the asset quality slide. Capital also remained solid.
Our total risk-based capital was 11.5% at year-end, relatively flat to the prior quarter, despite nearly $1.2 billion of period-end asset growth. More than 90% of this period-end asset growth was in loans held for sale, a category that has very low-risk content.
We'll go through a more extensive analysis of our capital later. So let's turn to Slide 7 and dive deeper into the income statement.
Net interest income grew $6 million to $152 million for the fourth quarter, which was 4% higher than last quarter. The results reflect a 9% increase in average earning assets, led by loans held for sale and commercial loans.
The commercial loan growth was split between warehouse loans and commercial real estate, largely, homebuilder loans. With the increase in loans held for sale, we were able to book more net interest income, albeit at a lower spread, enhancing return on equity and reducing NIM by 3 basis points.
On the deposit side, the impact of the rate cuts in late September and October, reduced our cost of deposits by only 6 basis points, as 38% of these deposits are technically noninterest-bearing. However, if you consider the rates we credit to our sub-servicing customers, the cost of deposits came down 18 basis points, a beta of 36%.
We'll discuss earning asset growth more on the next slide. Noninterest income decreased $9 million or 5% to $162 million due to seasonally lower mortgage revenues, strengthened by continued robust refinance activity.
Our gain on sale revenue of $101 million represented a decrease of $9 million or 8% from the prior quarter. The decrease primarily reflected $1 billion or 11% in lower fallout-adjusted locks.
Our retail mix was roughly the same as last quarter, and refinance activity comprised 56% of our lock volume. We remained disciplined in our pricing and continue to focus on execution to ensure that we protect our gain on sale margin.
This quarter saw us return to the RMBS market, with two issuances totaling $725 million. As a result of all these activities, gain on sale margin expanded by 3 basis points to 123 basis points.
As you'd expect, the increased refinance activity accelerated prepayments, creating a net loss on our mortgage servicing rights of $3 million compared to a net loss of $2 million last quarter. Noninterest expense was $245 million, up $7 million from the prior quarter.
The increase was primarily related to $5 million of balance sheet cleanup and discretionary expenses that took place in the fourth quarter, including fixed asset write downs, costs related to the secondary offering by our largest shareholder and a donation to the Flagstar Foundation. Mortgage expenses were relatively flat to last quarter as our closing volume was essentially flat.
We primarily recognize expense when we close loans in contrast to revenue recognition, which is a rate lock. This quarter, mortgage closings of $9.3 billion exceeded lock volume of $8.2 billion.
The result of which is that mortgage revenues declined from last quarter, yet mortgage expenses were flat. We expect that in Q1, we'll see a reversal of this, that locks will exceed the level of mortgage closings.
It's interesting to note that mortgage expenses were 112 basis points of closing volume, virtually the same as last quarter. The remaining expense categories were relatively flat with the prior quarter, demonstrating the scalability of our business and cost discipline of our employees.
When you look at expenses year-over-year, and if we adjust for the expenses related to the branch acquisitions and the $5 million of expenses we called out earlier, what you find is that mortgage expenses are $79 million higher, driven by an increase in our retail mix. Normalizing for the run rate expenses of the acquired branches, all other non-mortgage expenses are only $30 million higher, the result of growth in the servicing business and in community banking.
It's important to note that this expense growth is coming in at a marginal efficiency ratio under 40%. Let's now turn to Slide 8, which highlights our average balance sheet this quarter.
Average loans held for investment grew $0.4 billion, driven by an increase in average warehouse loans and CRE. Most of the CRE growth was driven by our home builder finance team as this sector continues to be strong and deal terms remain lender favorable.
Average deposits increased $0.1 billion, driven by higher custodial balances, which rose $0.2 billion or 5%, due to an increase in the number of accounts serviced. Finally, we continue to grow our tangible book value per share, which ended the year at $28.57, up $0.95 from last quarter end.
So let's now turn to asset quality on Slide 9. Credit quality in the loan portfolio remained strong.
Early-stage delinquencies continue to be negligible, only $14 million of total loans were over 30 days delinquent, and still accruing at December 31. Our allowance coverage was 0.9% of total HFI loans.
I'd point out that this coverage reflects 23% of these loans being warehouse loans. So excluding warehouse loans for the denominator, given their relatively clean credit loss history, our coverage ratio withstand at 1.1%, among the strongest in the industry.
Turning to Slide 10. I'd also like to give an update on CECL.
We now expect our reserves will increase approximately 25% to 35%. This includes the impact on our reserve for unfunded commitments.
As you know, the impact of adoption does not affect income at implementation. The CECL impacts each of our portfolios quite differently, depending on the life of the loan and the collateral type, including whether the loan is secured or unsecured.
Considering our current economic outlook and the mix and credit characteristics of our loan portfolio, the day 1 change breaks down by portfolio as follows: we expect our residential mortgage portfolio, excluding loans with government guarantees, to be reserved at 0.8 to 1.0%; our consumer loan portfolio to be between 1.9% and 2.2%; our commercial loans, excluding warehouse loans, to be between 0.6% and 0.8%; and our warehouse loan portfolio to be minimally impacted by CECL. Additionally, the reserve for unfunded commitments is expected to increase by approximately 3x, largely due to existing lines with our nonwarehouse commercial customers.
We also continue to maintain reserves for our loans with government guarantees and for specifically measured loans whose amounts are not impacted by the CECL methodology change. Given our forecasts for growth in each of these portfolios, we'd expect our first quarter loan loss provision to be between $3 million and $5 million.
This contemplates a replacement of net charge-offs, our expectations for mix and other changes in our loan portfolio and no changes to our economic outlook. Turning to Slide 11.
Despite balance sheet growth, our capital ratios remained relatively unchanged. Total risk-based capital was 11.5% at December 31, down only 2 basis points, while our CET1 ratio of 9.3% was up 7 basis points.
Our Tier 1 leverage ratio of 7.6% was down as we allowed the balance sheet to flex up this quarter. As you're aware, these ratios are based on old capital rules under new capital simplification rules, which we're now subject to.
Our Tier 1 leverage ratio at December 31 would have been 8.0%, 43 basis points higher, and our total risk-based capital ratio would be 11.7%, 22 basis points higher. It might be helpful to look more closely at the quarter-to-quarter change in our total risk-based capital ratio, which is still our tightest capital ratio.
Earnings retention increased the ratio by 31 basis points. Balance sheet growth in the loans held for sale portfolio alone, reduced this ratio by 25 basis points.
All other activity accounted for a net decrease of only 8 basis points. We're quite comfortable with the ratio compression that is coming from growing loans held for sale.
Loans held for sale are carried at fair value and that portfolio turns over on a regular basis. Between loans held for sale in our warehouse loan portfolios, we have approximately 450 basis points of total risk-based capital dedicated to 2 asset categories that have very little risk content.
In warehouse lending, which has a 100% risk weight, we've had under $5 million of losses cumulatively over the last 12 years, a period that includes the Great Recession, higher or lower balances. And these 2 categories do not meaningfully change the outputs from our capital stress testing models.
So this is why we remain comfortable with our capital levels at December 31. I'll now turn to Lee for more insight into each of our businesses.
Lee Smith
Thanks, Jim, and good morning, everyone. We're very pleased with our net income of $1 per diluted share for the fourth quarter and adjusted net income of $3.46 per diluted share for fiscal 2019.
These earnings resulted in a return on equity of 12.7% for the quarter, an adjusted return on equity of 11.7% for the full year. The last 12 months have seen declining interest rates put pressure on net interest margin for many banks.
But our unique business model, which is well balanced between interest income and noninterest fee income revenues, has delivered outstanding results. To further illustrate this point, during the year, 49% of our revenues came from net interest income, while 51% came from noninterest or fee income.
As a result, we believe in our ability to deliver predictable and sustainable earnings in any interest rate environment and continue to generate strong returns for our shareholders. There were several other notable developments during the year and quarter, which included average interest-earning assets, increased an impressive $4.3 billion or 26% in 2019, with $1.7 billion or 9% occurring in the fourth quarter.
We saw good growth across all commercial and consumer loan categories, while our net interest margin increased from 2.89% in 2018 to 3.05% in 2019, despite there being 3 rate cuts during fiscal 2019. Average retail deposits increased $1.5 billion year-over-year, in large part, a result of the 52 branches we acquired at the end of 2018.
Average custodial deposits also increased $2 billion in 2019 as we continue to successfully grow our subservicing business and add more loans to our servicing platform. Mortgage banking revenues increased $105 million or 44% to $341 million in 2019 versus $236 million in 2018, as our mortgage origination business was able to stay disciplined and take advantage of the lower interest rate environment during the year.
Our subservicing business grew 28% in 2019. And we ended the year servicing or subservicing approximately 1.1 million loans, which generates consistent noninterest fee income for the bank.
The integration of the default service operation we acquired at the beginning of the fourth quarter has also gone very smoothly. During the year, we closed out the last remaining legacy regulatory issue with the expiration of the CFPB consent order on September 30.
Finally, capital remains solid, and we've just announced an increase in our quarterly dividend to $0.05 per share, after we restarted our dividend program at the beginning of 2019, which has enabled our shareholders to further benefit from our success. It's been another exciting year for Flagstar, and we're proud of what we've accomplished.
Furthermore, given our business model, we believe we can continue to grow and reward our shareholders with even higher returns in the future. I will now outline some of the key operating metrics from each of our major business segments during the fourth quarter.
Please turn to Slide 13. Quarterly operating highlights for the community banking segment include average commercial and industrial, and commercial real estate loans increased $128 million or 3%, with the growth being driven predominantly by homebuilder finance loans within our commercial real estate division.
Average consumer loans held for investment increased $58 million or 1% in the quarter as we continue to add high-quality non-auto indirect loans and HELOCs to the portfolio. We've added almost $900 million of consumer loans to the balance sheet over the last 12 months, with most of that growth coming from non-auto indirect loans and HELOCs.
Average warehouse lending loans increased $239 million or 10% to $2.7 billion in the quarter as a lower interest rate environment increased refinance activity in the mortgage industry. Overall, this means average loans held for investment increased $425 million during the quarter, which drove the $6 million or 4% increase in net interest income quarter-over-quarter.
Given our expanded footprint and relationship-based approach to lending, we will continue to originate high-quality consumer, commercial and warehouse loan balances going forward. Average demand and savings deposits increased $133 million, while average custodial and escrow deposits increased $222 million as a result of the increased number of loans subserviced and the lower rate environment, increasing refinance activity.
As we look ahead, we will continue to grow the community bank given it produces predictable and stable earnings for the bank. Please turn to Slide 14.
Quarterly operating highlights for the mortgage origination business include: fallout-adjusted lock volume decreased 11% to $8.2 billion quarter-over-quarter, while the net gain on loan sale margin increased 3 basis points to 123 basis points. As a result, gain on sale revenues decreased to $101 million in the fourth quarter.
The slight decrease in fallout-adjusted lock volume was due to the usual seasonal slowdown in the mortgage industry, but the continuation of low interest rates, driving increased refinance activity offset some of this decline. Refinance activity accounted for 56% of our lot volume during the quarter.
Mortgage closings of $9.3 billion in the fourth quarter were consistent with loan closings in the third quarter. And this is testament to our mortgage operations team and how they optimize our available capacity without compromising service quality to our customers.
We continue to maintain our disciplined pricing approach to ensure we focused on originating the right loans and maximizing revenues with the capacity we have available. It's this approach that enabled us to increase gain on sale margin 3 basis points to 123 basis points in the quarter.
We're very pleased with how we've been able to take advantage of the lower interest rate environment in the mortgage business. It has generated significant noninterest fee income during the quarter and year, and is a natural hedge to some of our other businesses in a declining interest rate environment as we have seen throughout 2019.
Moving to servicing. Quarterly operating highlights for the mortgage servicing segment on Slide 15 include: we ended the quarter servicing or subservicing approximately 1.1 million loans of which over 918,000 or 84% are subserviced for others.
We increased the number of loans serviced or subserviced by 97,000 or 10% in the quarter and 240,000 or 28% in the last 12 months, and have been one of the fastest-growing subservices in the industry during that time. Today, we have the capacity to service or subservice 2 million loans as well as provide ancillary offerings such as recapture services and financing solutions to MSR owners.
If you look at Slide 38, you will see that we are achieving the high end of our $4 million to $6 million of operating profit before tax guidance for every 100,000 loans we add to the platform. And in fiscal 2019, our servicing business generated $49 million of operating profit before indirect allocations and tax.
During the quarter, we sold $4.9 billion of MSRs, albeit flow sales and 88% of the loans sold was servicing retained, which gives you an indication of the strong relationships we have with MSR buyers and the confidence they have in the underlying quality of our originations and seamless transfer to our subservicing model. The integration of the default servicing operation in Jacksonville, that we acquired at the beginning of the quarter, has gone very smoothly indeed and gives us full control over the life cycle of any loan.
Custodial deposits averaged $4.8 billion in the fourth quarter, an increase of $222 million or 5% from last quarter. This is an illustration of one of the additional benefits we get from our subservicing business as it provides liquidity that helps fund our balance sheet.
We couldn't be more pleased with the growth and results of our subservicing business and believe we have the right team, platform and offering to continue its success in the future. Moving on to expenses on Slide 16.
Our total noninterest expenses increased 3% or $7 million to $245 million quarter-over-quarter, while total revenues decreased slightly by $3 million to $314 million during the same period. The main drivers of the increase in expenses were as follows: there were $5 million of balance sheet cleaner and discretionary expenses in the fourth quarter, and investment in our mortgage business of $2 million during the quarter as we continue to take advantage of the lower rate environment.
If you look at Slide 17, our core nonmortgage, noninterest expenses were $136 million in Q4, after excluding the previously mentioned $5 million, which is flat when compared to Q3. Expenses tied directly to the mortgage origination business were $104 million, an increase of $2 million versus Q3, given mortgage closings which drive mortgage expenses, were flat quarter-over-quarter, and we continue to invest in this business.
Approximately 75% of expenses tied to the mortgage origination business are variable in nature and paid on a per unit basis or semi-variable, which means they are just quickly and easily if mortgage volumes decline. Mortgage revenues, which are a function of mortgage locks, decreased $10 million quarter-over-quarter.
And we would expect this to translate to lower mortgage closings, and therefore, costs associated with the mortgage business in the first quarter. Our efficiency ratio was 78% for the fourth quarter, which was a worsening of 3% from the prior quarter for the reasons I've just outlined.
Going forward, we will continue to include this granularity to our expense reporting, and would focus on optimizing our core noninterest expenses in 2020. We estimate noninterest expense will be between $228 million and $233 million during the first quarter.
The decrease in expenses being the result of less mortgage activity, given the seasonality of the business. It's been another strong quarter and an outstanding year.
All three of our major businesses: the community bank, mortgage originations and servicing, have made significant contributions to the bottom line. We believe we have a diversified business model where the different businesses complement each other, but also act as a natural hedge, meaning we can be successful and generate consistent and predictable earnings in any interest rate environment.
Given this unique business model and seasoned management team, we believe we will continue to create value for our shareholders going forward. With that, I'll hand it back to Sandro.
Alessandro DiNello
Thank you, Lee. I'm going to close our prepared remarks with some guidance for Q1, and then open the call for questions and answers.
Please turn to Slide 19. We expect net interest income to decline 5% to 10% due to a seasonal decline in interest-earning assets, together with a narrower net interest margin.
As Jim noted, we expect a provision expense of $3 million to $5 million. We anticipate mortgage revenue, including gain on sale and net return on MSR, will decrease 10% to 15% due to the seasonality of the mortgage market.
Our other noninterest income is expected to decline slightly. As we noted, we anticipate noninterest expenses to be $228 million to $233 million, reflecting lower mortgage-related expenses as well as the elimination of the discretionary expenses that he noted.
And finally, we expect the effective tax rate to be consistent with Q4 2019. Finally, I'd like to expand on expenses a bit.
When the mortgage environment is favorable, we have the ability to quickly scale up to take advantage of revenue growth opportunities. And because of the way we have built this business, we can just as quickly ratchet down costs when the demand for originations falls off.
We did this not only in 2018 but consistently in the past with the ebb and flow of the mortgage market. That's not to say that we can't become more efficient.
Improving efficiency is a major focus for our team this year, and we are determined to make progress. As a start, this quarter, as you've seen, we introduce more granularity in our expense numbers to help you model more accurately.
This concludes our prepared remarks, and we'll now open the call to questions from our listeners. Emma?
Operator
[Operator Instructions]. Our first question comes from George Bose with KBW.
Bose George
Okay. I wanted to ask just about the net interest margin, first.
Just going forward, you noted in 1 quarter, your guidance is the 5 to 10 basis point decline, driven just by the December rate cut, can you just talk about any impact from the changing asset mix, I guess, loans held for sale will decline? How that plays into it as well?
Alessandro DiNello
Sure, Bose. So let me first say that, I'm really not too concerned with the NIM compression that we've seen and what we're projecting.
And there's a couple of factors that come into play here. The $2 billion of deposits we purchased through the branch acquisitions are very low cost, as I'm sure you recall, and really, they can't get any cheaper.
And Q1 will be the first quarter where our commercial book feels the full impact of the last fed move. Second, the average balance of our loans held for sale grew $1.1 billion in Q4.
And -- at an average yield of about, I think, it's 3.92%. While -- and while that's a great arbitrage, its spread is relatively narrow.
So while we see loans held for sale shrinking a bit in Q1, the average yield will still be -- will probably decline a little bit actually. And while it's still a good arb, still with negative impact on overall NIM.
Plus, we expect warehouse will probably shrink a bit in Q1. And as you know, that's a very high-yielding asset.
So when you really look deep, I think you can see that the NIM compression is a function of the makeup of the balance sheet, which is different than most other mid-sized banks, but something we're very comfortable with. And I think when you see that this happening to NIM, the dynamics that are causing that to happen are actually the same dynamics that are causing our mortgage business to see better revenues.
Bose George
Okay. That makes sense.
But -- and just I know you don't sort of give guidance sort of more further out, but if you look out past the first quarter, the impact of the rate cuts are over -- is this sort of -- should we think about more stable margins after that?
Alessandro DiNello
Well, it all depends on what happens to the interest rate market, right? So it's very difficult for me to answer that question with a high degree of certainty.
But you would think for everybody if the Fed doesn't change the position that stability in margin is more likely, but we're not -- I can't be specific about that.
Bose George
Okay. Makes sense.
And then actually, just switching over to the mortgage side, just wanted to ask a regulatory question. Just with the proposed changes that the CFPB is suggesting to the QM patch, potentially reduced -- taking -- removing the 43 DTI, just wanted to get your thoughts about how that plays out?
Whether you feel like that's going to have -- that's going to be good or bad, just your thoughts on that would be great?
Alessandro DiNello
Yes. I'll let Kristy comment on it.
But I think generally, we view that as positive.
Kristy Fercho
Yes. We do.
I think one of the first things that would happen -- and sorry, Bose, this is Kristy. One of the first things that would happen is they would look to potentially do an extension of that.
But I think the proposals on the table, looking at kind of the APR at 1.50 [ph] is something that we feel very comfortable with, and we think we'll actually be pretty easy to manage to. So I think the one thing everybody is committed to is ensuring a smooth transition and no disruption to the market.
And so I think whatever ultimate decision gets made, and certainly, all the proposals that are on the table today. We feel very comfortable with, and we'll be able to adjust quite easily.
Operator
Our next question comes from Scott Siefers with Piper Sandler.
Robert Siefers
Let's say, I guess, Jim, maybe first question's for you, just regarding the overall size of the balance sheet side, sounds like we will get some relief in the first quarter. But just curious for your thoughts on appetite to thereafter, grow the balance sheet at the same kind of pace you've been doing, just given where you are with existing capital levels, even under the simplified method?
Alessandro DiNello
Let me give to you some thoughts first on that, Scott, then I'll let Jim chime in here. We've always been able to optimize the growth opportunities, and we'll continue to do that.
I think you've been following us a long time, you know that we have pretty good success doing that in the past. We know the return on equity of every loan that we put on our books.
And so if it doesn't hurdle our return expectations, we just don't make it. So we'll do as much as we can, given that requirement while not compromising credit at all.
Now depending on what those opportunities are, whether they might be in C&I, CRE, builder, warehouse or even consumer nonmortgage, that will determine just how much held for sale we carry in the mortgage business. So all of that in mind, we still have to make sure that we keep our capital levels where they need to be.
So I think that this is the real beauty of our balance sheet that it gives us the flexibility to adjust our held for sale to offset either greater or less growth in our commercial lines of business. So just overall, that's how I view the growth of the company going forward, which isn't any different than the way we've managed it in the past.
Jim can get into the more accounting-related issues here, and I'll let him do that.
James Ciroli
Yes, yes. So I was going to get right there, and I'm glad that Sandro did.
Because he made those comments in his prepared remarks that we have somewhat, and we've said this for a few years now, unparalleled flexibility in our balance sheet. So what you saw this past quarter was that we increased our loans held for sale because we had that opportunity.
And we took -- we did that, albeit it hurt NIM a little bit, but improved the returns that we had on equity because we were able to flex that up and fully invest the balance sheet. As we move through the quarter, I think you're reading our guide right, which is that we'll expect that portfolio and likely warehouse loans, depending on where the mortgage market goes in the first quarter.
Those will flex down a bit. So I think that you'll see that spin down, but I wouldn't think anything of that.
And to the extent that we have the capacity, just know that we always have the ability to flex up the loans held for sale. If we don't find the growth opportunities that we like in commercial lending or in consumer lending.
We always have that capacity that reflects that one category up, like we did this past quarter, to maximize our ROE. And we've been saying over the past 5 years, and I know we have a slide in the deck that talks about our CAGR over a 4-year time period of growing earning assets, I wouldn't expect the future pace to be as brisk as it has in the past.
Because 4 years ago, we had a situation where we are underinvested from a capital standpoint, and now we're no longer in that position. So I would imagine, just like this past quarter, where we were able to add $1.2 billion of assets, but keep our capital ratios flat.
That our capital retention is going to really influence the amount of balance sheet growth that we have.
Alessandro DiNello
And if I could just emphasize this concept of flexing the balance sheet to take advantage of our ability to hold mortgages longer and therefore, get the arb on that, that's always going to have a negative impact on the NIM. Because while the arb is solid and has absolutely no risk, it's narrower than our -- than the NIM in our banking business.
So when you see that NIM compressed, not necessarily a bad thing. I mean, we showed an increase in net interest income.
An increase in net interest income produces more earnings, more earnings produces a better return on equity, that should turn into shareholder value. So I can sit here and I can -- we can decide not to grow mortgages held for sale by $1 billion, so we don't have mortgages that have a yield under 4% going on the books for 30 days.
And I can report a NIM improvement. But is that better for the company?
I don't think so.
Robert Siefers
Okay. I appreciate all that color.
And then if I could ask a follow-up, just want to move to the cost side, and I appreciate the extra detail you guys gave on the mortgage versus non mortgage. Maybe just a couple of things in there.
Is there anything that would prevent you guys from getting back down to the -- that sort of $55 million to $60 million per quarter in mortgage-related costs, i.e., changes you guys have made to the mix? And then is there any opportunity to reduce costs, absent more of a decline in originations?
And I guess the context in which I ask that is, we sort of know what the aggregate cost base will be in the first quarter given your guidance, but thereafter, presumably originations pop back up in the middle quarters of the year. So just wondering if there's an opportunity to pare back any of the cost base absent something more reductive, so to speak, in the origination side?
Lee Smith
Yes. Scott, it's Lee.
So a couple of things. On the mortgage side, and I mentioned this in the prepared remarks, approximately 75% of the mortgage expenses are variable or semi-variable in nature.
So they will adjust quickly if mortgage volumes were to decline and we feel very good about the variable cost model we've put in place on the mortgage side. We've given you the additional granularity.
So you can now see not the cost for the non-mortgage businesses. And both Sandro and I mentioned our focus in 2020 is to get more efficient as an organization.
So that $136 million in Q4, we will be focused on that particular number.
Kristy Fercho
The piece, if I could add to, Scott, when you talk about the efficiency and Lee talked about some of the technology investments that we made in mortgage, that kind of additional $2 million, some of those efficiency -- or some of those technology investments will actually help drive additional efficiency in some of our businesses. And so, so we'll continue to look for and as Sandro talked about the efficiency exercise that we're doing.
We're definitely looking at where we can get more efficient and continue to take cost out of the business.
Alessandro DiNello
But look, to get back to $55 million in mortgage expense, which is what it was in the fourth quarter of '18, that wouldn't be a good thing, right? That would mean that the market is small...
Lee Smith
Much smaller.
Alessandro DiNello
What we've tried to lay out here by giving you this granularity is to show you that while there's been this increase in mortgage expenses, the increase in the revenues have been more and at a really nice leverage. And so we can go back down to 55% as quickly as we need to if the revenue isn't there.
That's the point we're trying to make here, and that's why we're trying to show you the difference between the core expenses and the mortgage expenses. But if we're going to do the volume that we did this past quarter.
And the mix that we have this past quarter, which we think we're optimizing, we can't do that with half of the mortgage expenses that we have today. Yes.
I got to tell you, I mean, I couldn't be more pleased with the way this mortgage business is being run. We could have originated $40 billion to $50 billion of mortgages last year, but the margin would have been narrower, expenses would have been a lot higher, and our exposure to MSR risk would have grown exponentially.
No $20 billion bank in the country is averaging -- is originating over $30 billion of mortgages. And I think the way we manage this business keeping it profitable no matter what's going on with interest rates, managing MSR risk expertly, avoiding consumer compliance issues.
I don't know how we could do it much better.
Operator
Our next question comes from Daniel Tamayo with Raymond James.
Daniel Tamayo
Well, you just -- Sandro, you just touched on the question I was going to ask, which was on the -- if you're passing on mortgage origination opportunities, which just sounds like you are, so I guess, given that, is it -- in the gain of sale -- gain on sale margin is really, really high. You guys didn't expect it today up as high as it was last quarter and it rose.
So I guess the question again is, if you are passing on these opportunities, can you -- do you think you can continue to see gain on sale margins at an elevated rate for the near term, at least?
Alessandro DiNello
Well, I mean, execution's really important. And what the FHFA makes Freddie and Fannie do with pricing issues, I mean, there's a lot of things that, that come into play here.
So it's really difficult to predict with a tremendous amount of accuracy. We try to give you the best guess we can based on what we know today in the first quarter guidance.
I think that we've been pretty darn good at being able to manage the margins at levels that are -- that compare very nicely to what's going on in -- with our peers in terms of what's available publicly. It's really hard to compare us to others because mix is different in every company.
And we put a little bit more emphasis on the retail business, where we think we've been able to attract some very, very strong people. So I think in terms of total revenue, I think that we're going to be able to keep the profitability of the business where it's been, and maybe we can make it better.
We're certainly always trying to do that. So I'm pretty confident in our ability to continue to execute, but there is just so many factors that come into play here.
It's hard to be real to predict out for any real length of time.
James Ciroli
And what I would say, Danny, just to follow-up on that. The strength and scale we have at all of our channels is somewhat unparalleled.
And we try to message the optionality, the strength and scale in those channels gives us. So it's hard to predict where market forces might take that cost revenue, but I can assure you, we're doing what we need to do to optimize the revenue versus the contribution that, that revenue makes to the bottom line.
Lee Smith
Yes. And Danny, we're very aware of what our capacity is.
And we want to give our customers a great experience and certainty, and I think they appreciate that. We don't want to overshoot capacity and have our customers having a bad experience, that's not what Flagstar is about.
And if you look at Slide 17, which is a really more of a cost slide, but the third bullet, just from a mix point of view, year-over-year. So in Q4, this Q4, 22.7% of closings came from our retail channels versus 14.3% a year ago.
And obviously, the retail business is typically the higher-margin business.
Daniel Tamayo
That's all really helpful. I appreciate that color.
All right. I'm going to switch gears here.
The tax rate guidance you gave lower than what we've seen in the past, obviously, in line with the fourth quarter. But what are the investments you're making that are driving the tax rate lower?
And then, I know you called out the benefit from the state net operating losses, but how much of that is the lower tax rate guidance? And how long would you expect to benefit from that?
Alessandro DiNello
Can't give you all of those secrets here. But I mean just to add a little -- maybe he can give you a little more color.
James Ciroli
Yes. So we look at what the larger institutions are doing to do put in place tax planning and to satisfy their CRA requirements.
So of course, we're investing in things like low-income housing in other tax credit-related items. We are also, as we talked about in the release, we're now at this point, fully utilized on all of our net operating losses, both -- we're in a position where we're a taxpayer.
So we're utilized on our net operating losses at the federal level. We have some still that are limited.
And by being in that position now, these tax planning strategies start to make financial sense. And so as they've come into the horizon for us to implement those, we've done them.
But it's things that you'd see at larger, more regional type banks. But just typical things that those banks would do.
Daniel Tamayo
So fair to say that the benefit from the state net operating losses is minimal and most of the new guidance is based on the investments you're making?
Alessandro DiNello
In the new guidance, yes.
Operator
Our next question comes from Henry Coffey with Wedbush.
Henry Coffey
I know you don't like to talk about future quarters, but given the shocking level of provision based on CECL -- no, given the fact that you're actually going to start booking somewhat CECL-related provision expense, should we assume that something around $3 million to $5 million is the new normal? Or given that we've -- you've had no loan quality issues there 1 or 2 over the last 5 years, and we're just sort of plugging in a number, which is usually 0.
Alessandro DiNello
Yes, yes.
Henry Coffey
So it is. That new number instead of being 0 in my model be $3 million to $5 million?
Or what is the thought process going forward?
Alessandro DiNello
Well, it can't be 0 if you're growing the balance sheet, right? Because you've got to provide for anything new.
So that's all a function of how much the balance sheet grows, and then what categories it grows. And Jim included in his prepared remarks some pretty specific information on how it impacts the various different bucket.
So we're not -- we don't know where the growth is going to come. So I can't even answer that question.
I don't -- even if I wanted to, I couldn't because I don't know where the growth is going to come. But I can tell you this, in my -- in one of my comments, I think, and maybe I've answered in one of the questions, I referenced the fact that every loan has to hurdle from an ROE point of view.
And so we're going to take into consideration the CECL impact on the growth at any particular loan would cause. And if it doesn't have the return, including that CECL impact, then we're not booking the loans.
So at the end of the day, I don't think it matters. I think it -- ultimately, we're going to be just as profitable as it would have been without CECL, but it will impact a little bit in terms of how we make decisions on where we grow the balance sheet.
Henry Coffey
I mean, that's how you think about these issues because, frankly, you're one of damn good bank. Are your competitors also sort of thinking in the same terms that you have a new -- they have a new cost element to put into the ROE equation, and therefore, they need to be more disciplined on pricing?
Or are you still waiting for the market to kind of wake up?
Alessandro DiNello
Well, I was going to ask you.
Henry Coffey
Or have you've already analyzed...
Alessandro DiNello
Yes. I don't know.
I mean, obviously, we look at what everybody reports. And we want to make sure that we're not crazy in the way we're thinking about things.
But I don't know. I think we all are learning as we go along here on CECL.
And so we'll just quarter-by-quarter, get a little bit smarter about it and be able to be more precise.
James Ciroli
What I would say, Henry, is it's not a cost, it's not numerator, it's capital, it's denominator. Numerators can be the same.
We look at these loans from a profitability perspective the same way. Economically, that we do, but CECL definitely forces a higher capital cost.
I'm not sure how other people are going to look at it. That's what we're going to do.
We're going to stay true to our return objectives on those loans. And if those return -- if we can find good opportunities, we'll make those investments.
And if we can't, again, we have balance sheet flexibility where we can find net interest income without having to rely too much on growing loans held for investment.
Alessandro DiNello
We've always liked the warehouse business. Now we really like the warehouse business because it has very little CECL impact.
And so we'll keep looking for opportunities to grow that as much as we can within our concentration limits. But those are things you look at now, you know that you might have looked at it a little differently before.
Maybe that's what the whole point of the thing is, right?
Henry Coffey
We can talk about that off-line.
Alessandro DiNello
But trying to cover some rationalization.
Henry Coffey
Preferably over a really nice dinner, some place.
Alessandro DiNello
Yes.
Henry Coffey
So look, the business is absolutely proven itself. But you're -- and you do have a good level of capital.
But as you think about return dynamics, what's preventing you all from taking a more aggressive stance towards either dividend or buybacks? However, you prefer to bump up the payout, what are the restraints to doing that?
What are the opportunities that would encourage you to be more aggressive? Simply, what are your thoughts on where you are today and where you'd like to be with return on capital measures?
And with that, I'll just go back into the queue.
Alessandro DiNello
It's all about our ability to get return on our investments. If we can get the return on the investments as opposed to a buyback or a dividend.
If we can do it just through the normal business. We think that is a long-term and the better interest of the shareholder because it grows earnings, it grows the strength of that organization.
It opens up other opportunities for potentially acquisitions and such. So I think that growing the company and getting the return that way is certainly my preference.
Now if we don't think we can do that or we think we need to support the stock in some fashion, then a buyback and/or increased dividend is appropriate. And of course, we've done both, right?
We instituted a dividend a year ago, we just increased it a little bit. We did do some buyback last year.
So we are looking for those opportunities. And I think that the fine sight is of any value.
I think the way we did it last year, the combination of the dividend, a little bit of a buyback and then growing earnings almost $5 per share on tangible book value, created a 45% increase in our stock value. And I know that the market forces have a lot to do with that, but it was the best amongst all mid-sized banks in the country.
So I think that's evidence that the choices we've made, at least historically, have been pretty good, and we'll try to balance all of that in the same way going forward.
James Ciroli
Yes, keep in mind, Henry, just to reemphasize the growth preference. As I said in my prepared remarks, our marginal efficiency ratio and that growth in servicing and community banking is sub-40%.
That's a pretty good dynamic. So we'd like to continue growing.
But if -- I mean, the important thing of having that tool in the toolbox is that we don't have to over rely on growth to deliver earnings per share growth.
Henry Coffey
No. I agree.
It's just helpful to hear your perspective on it.
Operator
[Operator Instructions]. We'll take our next question from Chris Gamaitoni with Compass Point.
Edward Gamaitoni
A quick question. I just wanted to get your thoughts on loan growth for 2020.
And what areas you envision might be adding the most balances?
Alessandro DiNello
I don't know. I really don't.
It all depends on where the opportunity is. I've talked a couple of times here and Jim has as well, about the importance of getting the right return on the loans that we grow here.
And so we're going to keep doing that. And historically, I've answered this question the same way every quarter, somebody always ask it.
And I just don't know. We're prepared for whatever the opportunity is, CRE, C&I, homebuilder, warehouse, non-QM mortgage, HELOCs, indirect, non-auto, all those things.
They are all categories that we're comfortable with. We're keeping within our relatively conservative concentration limits, making sure that the ROE hurdles on every single loan we put on the books.
So we'll just see where the opportunities are.
Lee Smith
I think the only thing I'd add, and Sandro has gone, I mean, we -- you can -- we have a very diversified model, Chris. So I mean, it can come from commercial consumer, warehouse, multiple different sources, and we have all of those in-house.
So we'll be ready for the opportunities as they present themselves.
Operator
Our next question comes from Steve Moss with B. Riley FBR.
Stephen Moss
We talked a lot about mortgage expenses here, but just on the revenue side, I mean, we've had a pretty good move in rates in the last couple of days. And just kind of wondering what are the underlying assumptions to your mortgage revenue guidance for the quarter?
And maybe a little color, if you think about, if longer-term rates hold here for the remainder of the quarter, what maybe some of the upside could be?
Alessandro DiNello
Well, from a planning perspective, we've always said that we look at the average expectations of the MBA, Fannie and Freddie as a basic guideline, and then we try to follow that through all of our entire company, whether it's warehouse originations or MSR prepayments and all those things. So right now, we expect that the volume for 2020 is going to be a little less than 2019.
And that's just because that's what majority of the experts bank. As that view changes, we'll begin to adjust our business.
And again, that's how we've always done it. And as I said in my prepared remarks, there's a lot of ebb and flow in the mortgage business.
And we've shown our ability to stay profitable, no matter what. So we're very comfortable with that.
Stephen Moss
All right. That's helpful.
And then in terms of just the other consumer loans here, basically more than double year-over-year. Just kind of wondering the types of loans you're doing?
I see the local growth is in indirect lending. And any color you can give there would be helpful.
Alessandro DiNello
Yes. So in fairness, doubling is off of a small number.
So it's not like it grew really a lot.
Lee Smith
Yes. I think I mentioned in my prepared remarks, we grew the consumer loan base about $900 million year-over-year.
Most of that was indirect lending, as you say, in HELOCs.
Alessandro DiNello
All of it was indirect lending, in the other consumer portfolio.
Stephen Moss
Right. And so are those purchase loans from other institutions, I take it?
James Ciroli
No. Yes.
Steve, $400 million of the $300 million growth was in indirect. So that was the business we just started from scratch a couple of years ago, and it just continues to kind of mature into the -- its equilibrium point.
Alessandro DiNello
And these are RV and Marine.
Lee Smith
Non-order.
James Ciroli
Yes. That's like a -- there's more detail in the press release on that.
Operator
I would now like to turn the conference back over to Sandro DiNello for closing remarks.
Alessandro DiNello
Thanks, Emma. Overall, 2019 was a very good year for Flagstar.
In fact, Flagstar was the number one performing bank stock, as I mentioned earlier, in the mid-sized space in 2019, with our stock price increasing 45%. We're here to create shareholder value, and no mid-sized bank did it better than Flagstar in 2019, and that's a great tribute to the Flagstar family.
An important point that I think bears repeating is the consistency of our earnings quarter-over-quarter. For the year, our adjusted ROE was 11.7% in Q4, and even better at 12.7%.
The knock-on Flagstar by many has been concerned with how the volatility of the mortgage business will impact our earnings. In my humble opinion, the market lead different mortgage markets we've seen over the last 24 months.
And the consistency in the returns we've generated during that period, should lay that concern to rest. For the Flagstar folks listening, you made this possible.
My deepest thanks to you for your continued commitment to move our company from good to great. Our journey continues.
Thanks once more to everyone listening in. I look forward to speaking to you again in three months.
Operator
Thank you, ladies and gentlemen. This concludes today's teleconference.
You may now disconnect.