Good day, and thank you for standing by. Welcome to the M&T Bank third quarter 2021 earnings conference call. Currently, all phone participants have been placed in a listen-only mode. Following management's prepared remarks, we will open the call for your questions. If you have a question at that time, please press star one on your telephone keypad. If you wish to remove yourself from the queue, press the pound key. We do ask that you please pick up your handset to allow optimal sound quality. Please be advised that today's conference is being recorded. Lastly, if you should need operator assistance at any time, please press star zero. I would now like to hand the call over to Don MacLeod. Please go ahead.
Thank you, Emma, and good morning. I'd like to thank everyone for participating in M&T's third quarter 2021 earnings conference call, both by telephone and through the webcast. If you have not read the earnings release we issued this morning, you may access it along with the financial tables and schedules from our website, www.mtb.com, and by clicking on the investor relations link and then on the events and presentations link. Before we start, I'd like to mention that today's presentation may contain forward-looking information. Cautionary statements about this information, as well as reconciliations of non-GAAP financial measures, are included in today's earnings release materials, as well as our SEC filings and other investor materials. Those materials are all available on our investor relations webpage, and we encourage participants to refer to them for a complete discussion of forward-looking statements and risk factors.
These statements speak only as of the date made, and M&T undertakes no obligation to update them. Now I'll turn the call over to our Chief Financial Officer, Darren King.
Thanks, Don, and good morning, everyone. As we noted in this morning's press release, the favorable results we reported for the quarter highlight the strength and diversity of M&T's business model in the current challenging environment. Revenue in our fee-generating businesses was particularly strong, including mortgage banking, trust and brokerage, and payments. Credit trends are stable to improving, illustrated by net charge-offs about half our long-term average, a modest reserve release, and little change in the level of nonaccrual loans. In alignment with the strong revenue trends and the improved profitability over the last year, incentive compensation is rising as well. We'll offer some details on that in a moment. Lastly, our capital levels continue to rise. The CET1 ratio is near a record high as we await the closing of the People's United merger. Now let's review our results for the quarter.
Diluted GAAP earnings per common share were $3.69 for the third quarter of 2021, improved from $3.41 in the second quarter of 2021 and $2.75 in the third quarter of 2020. Net income for the quarter was $495 million, compared with $458 million in the linked quarter and $372 million in the year-ago quarter. On a GAAP basis, M&T's third quarter results produced an annualized rate of return on average assets of 1.28% and an annualized return on average common equity of 12.16%. This compares with rates of 1.22% and 11.55%, respectively, in the previous quarter. Included in GAAP results in the recent quarter were after-tax expenses from the amortization of intangible assets amounting to $2 million, or $0.02 per common share. Little change from the prior quarter. Also included in the quarter's results were merger-related charges of $9 million related to M&T's proposed acquisition of People's United Financial.
This amounted to $7 million after tax, or $0.05 per common share. Results for this year's second quarter included $4 million of such charges, amounting to $3 million after-tax effect, or $0.02 per common share. Consistent with our long-term practice, M&T provides supplemental reporting of its results on a net operating or tangible basis, from which we have only ever excluded the after-tax effect of amortization of intangible assets, as well as any gains or expenses associated with mergers and acquisitions. M&T's net operating income for the third quarter, which excludes intangible amortization and the merger-related expenses, was $504 million. Compare that with $463 million in the linked quarter and $375 million in last year's third quarter. Diluted net operating earnings per common share were $3.76 for the recent quarter, improved from $3.45 in 2021's second quarter, and up from $2.77 in the third quarter of 2020.
Net operating income yielded annualized rates of return on average tangible assets and average tangible common shareholders' equity of 1.34% and 17.54% for the recent quarter. The comparable returns were 1.27% and 16.68% in the second quarter of 2021. In accordance with the SEC's guidelines, this morning's press release contains a tabular reconciliation of GAAP and non-GAAP results, including tangible assets and equity. Let's take a look at some of the details that drove our results. Taxable equivalent net interest income was $971 million in the third quarter of 2021, compared with $946 million in the linked quarter. Higher income from PPP loans accounted for the majority of the $25 million quarter-over-quarter increase in net interest income. As the second round of PPP loans began to receive forgiveness from the Small Business Administration.
The net interest margin for the past quarter was, excuse me, 2.74%, down just 3 basis points from 2.77% in the linked quarter. We estimate that the higher balance of cash on deposit at the Federal Reserve contributed about 13 basis points of pressure to the margin. Largely offsetting that was the higher income from PPP loans, both scheduled amortization and accelerated recognition of fees from forgiven loans, which added an estimated 10 basis points to the margin. All other factors, including lower income from hedges, a slightly lower cost of deposits, and an additional accrual day, netted to zero impact. Compared with the second quarter of 2021, average interest-earning assets increased by 3%, reflecting a 22% increase in money market placements, primarily cash on deposit with the Fed, and a 3% decline in investment securities. Average loans outstanding declined about 3% compared with the previous quarter.
Looking at the loans by category on an average basis compared with the linked quarter. Overall, commercial and industrial loans declined by $3.3 billion, or 12%. The primary driver was a $2.4 billion decline in PPP loans. Dealer floor plan loans declined by $803 million, reflecting the ongoing impact from vehicle production and inventory issues seen across the industry. All other C&I loans were essentially little changed from the prior quarter. Commercial real estate loans were also little changed from the second quarter. Residential real estate loans declined by just under 4%. There are a few moving parts underlying that figure that are worth highlighting. Balance decreases due to normal prepayments and principal amortization, including the Hudson City portfolio, drove some of the decrease, as well as re-pooling of loans previously purchased from Ginnie Mae servicing pools.
Those were offset by retention of new home production, which will be a bigger factor in the fourth quarter. Consumer loans were up 3%, consistent with the recent quarters and continuing to be led by growth in indirect auto and recreational finance loans. On an end-of-period basis, total loans were down 4%, reflecting most of the same factors I just mentioned. The 11% decline in C&I loans include a decline of PPP loans outstanding to $2.2 billion at September 30th. Average core customer deposits, which exclude CDs over $250,000, increased 2% or $2.8 billion compared with the second quarter. That figure includes $3.8 billion of non-interest-bearing deposits, partially offset by lower interest checking deposits. Turning to non-interest income. Non-interest income totaled $569 million in the third quarter, compared with $514 million in the linked quarter.
The recent quarter included an insignificant valuation gain on equity securities, largely on our remaining holdings of GSE preferred stock, while the prior quarter included $11 million of valuation losses. Mortgage banking revenues were $160 million in the recent quarter, compared with $133 million in the linked quarter. Revenues for our residential mortgage business, including both origination and servicing activities, were $110 million in the third quarter, compared with $98 million in the prior quarter. Residential mortgage loans originated for sale were down about 7% to $1.1 billion when compared with the second quarter. The lower volume was more than offset by higher gain on sale margins. Commercial mortgage banking revenues were $50 million in the third quarter, compared with $35 million in the linked quarter. Those results reflect a strong originations quarter, combined with prepayment fees on loans previously securitized.
Trust income was $157 million in the recent quarter, compared with $163 million in the previous quarter. Recall that the second quarter's results included $4 million of seasonal fees arising from tax preparation work we undertake for clients, which did not recur in the third quarter. In conjunction with the transfer of M&T's retail brokerage and advisory business to the platform of LPL Financial in mid-June of this year, about $10 million in revenues associated with managed investment accounts previously classified as trust income are now included in brokerage services income. Service charges on deposit accounts were $105 million, compared with $99 million in the second quarter. The primary driver of the increase was customer payments related activity. Turning to expenses. Operating expenses for the third quarter, which exclude the amortization of intangible assets and merger-related expenses previously mentioned, were $888 million.
The comparable figure was $859 million in the linked quarter. Salaries and benefits were $510 million for the quarter, compared with $479 million in the prior quarter. As was the case in previous quarters this year, the higher salaries and benefits reflect revenue generation in certain business lines and incentive compensation associated with that revenue, notably commercial mortgage banking and trust. Also, at the enterprise level, we are accruing the corporate incentive at a higher rate in 2021, reflecting our expectation that M&T's full year earnings and profits will be higher than they were in 2020. Other costs of operations in the recent quarter including $5 million from the accelerated amortization of capitalized mortgage servicing rights as a result of the prepayments of previously securitized commercial mortgage loans that we referenced earlier.
Lastly, year-over-year growth in trust income and assets under management in our retirement business has carried with it an increase in the share of those fees paid to sub-advisors, which are included in other cost of operations. Also recall that the other cost of operations for the second quarter included an $8 million addition to the valuation allowance for our capitalized residential mortgage servicing rights. There were no adjustments to the valuation allowance in the third quarter. The efficiency ratio, which excludes intangible amortization and merger-related costs from the numerator and securities gains or losses from the denominator, was 57.7% in the recent quarter, compared with 58.4% in 2021's second quarter. Next, let's turn to credit.
Credit trends continue to stabilize, but as has been the case for the past little while, some industries are improving more rapidly than others, reflecting challenges such as the supply chain, pressure on materials costs, and the costs and availability of labor. The allowance for credit losses declined by $60 million to stand at $1.5 billion at the end of the third quarter. This reflects the $20 million recapture of previous provisions for credit losses, combined with $40 million of net charge-offs in the quarter. At September 30th, the allowance for credit losses as a percentage of loans outstanding was unchanged from June 30th at 1.62%. Annualized net charge-offs as a percentage of total loans were 17 basis points for the third quarter, 18 basis points in the second quarter.
The allowance for credit losses at the end of the quarter reflects our assessment of credit losses in the portfolio under the CECL loss measurement methodology, which includes our macroeconomic forecast. As we've previously indicated, our macroeconomic forecast uses a number of economic variables, with the largest drivers being the unemployment rate and GDP. Our forecast assumes the national unemployment rate continues to be elevated compared to pre-pandemic levels, averaging 5.5% over 2021, followed by a gradual improvement reaching 3.5% by mid-2023. The forecast also assumes that GDP grows at a 6.8% annual rate over 2021 and 2.7% annual rate during 2022. Non-accrual loans were essentially flat at $2.2 billion compared with June 30th, but increased as a percentage of loans to 2.4%, compared with 2.31% of loans at the end of June.
We also expect to disclose that our level of criticized loans is little changed from the second quarter when we file our third quarter 10-Q in a few weeks. Loans past due on which we continue to accrue interest were $1 billion at the end of the recent quarter. Turning to capital. M&T's Common Equity Tier 1 ratio was an estimated 11.1% at quarter end, compared with 10.7% at the end of the second quarter. This reflects continued strong organic capital generation combined with lower risk-weighted assets. As previously noted, while the People's merger is pending, we don't plan to engage in any stock repurchase activity. Now let's turn to the outlook. As we enter the final quarter of the year, we see little need to change our outlook for the remainder of 2021.
We expect year-over-year loan growth to be flat to out slightly on a reported basis and flat to down slightly excluding the impact from PPP loans, which reflects the decline in dealer floor plan loans. We continue to expect net interest income to be down a low single-digit % from full year 2020. We noted on the July conference call that we expected net interest income in the third and fourth quarters to, on average, be in line with the $946 million in the second quarter. We still expect that to be true, but the faster than expected forgiveness of PPP loans did pull some of that income forward into the third quarter from the fourth quarter.
We've slightly exceeded our outlook for low single-digit growth in non-interest income. As we begin to retain the majority of residential mortgage loans we originate on the balance sheet, residential gains on sale will be primarily driven by Ginnie Mae re-pooling gains. Momentum in the trust and payments-related businesses remains strong. Expenses have grown faster than we forecast in January, but with most of that growth directly connected to better-than-expected revenue and better-than-expected net income trends. The credit environment continues to improve along with the overall economy, but some segments are recovering more slowly than others. We believe criticized assets are at or near their peak, but there still remains some risk of downgrades within criticized loans from accruing to non-accruing. Our preparations for completion of the merger with People's United continue while we wait for regulatory approval.
Our projections as to the financial impact remain largely in line with what we offered at the time of announcement this past February. Following our usual practice, we'll offer our thoughts for 2022 on the January conference call. Of course, as you are all aware, our projections are subject to a number of uncertainties and various assumptions regarding national and regional economic growth, changes in interest rates, political events, and other macroeconomic factors, which may differ materially from what actually unfolds in the future. Let's open up the call to questions, before which Emma will briefly review the instructions.
If you would like to ask a question, please press star one on your telephone keypad. If you wish to remove yourself from the queue, press the pound key. We do ask that you please pick up your handset to allow optimal sound quality. We will take our first question from Gerard Cassidy with RBC.
Thank you. Hi, Darren.
Morning, Gerard.
A couple questions for you regarding your swap book. Can you share with us where it stands today and how that bleeds into net interest income over the next 6 or 12 months? Simultaneously, you've been very conservative in the positioning of your cash, particularly obviously at the central bank, the Federal Reserve. It's up dramatically year-over-year. You've been talking about it all year. When do you start to lengthen out the duration in that line item on the balance sheet?
Sure thing. Well, I've started writing your questions down, so now I won't miss them. On the swap book. The swap book has been fairly steady in terms of notional amount, right around $19 billion. It's been that way since the fourth quarter of last year. The received fixed yield we're getting on that book has been declining as new hedges roll on and older hedges at higher rates roll off. This quarter it added around $65 million to net income and impacted the margin by slightly less than 20 basis points. We'll see the notional amount come down next quarter. Will start to come down over the course of the following four quarters into the end of 2022.
The received fixed coupon will come down a little bit as well, but the bigger impact will be just the decline in the size of the outstanding notional. The impact of swaps and the hedging, just to remind everyone, for full year 2021 is really close to what it was in 2020. In 2022, the impact will be about half. It will be highest in the first quarter of 2022 and lowest in the second quarter, or in the fourth quarter, sorry, of 2022. To go along with that and to then answer or address your second question about cash, we've talked a lot about the cash position over the course of the year. If you look at the uptick in the latest quarter, it has been largely, actually, trust demand deposit growth in the quarter that drove a lot of the increase.
However, we are seeing some of the PPP balances as they decline end up in cash. What you saw this quarter, I think if you look at the securities balances at the end of the quarter, is they were actually up by about $300 million. We replaced runoff + $300 million. We've been in generally replacing the mortgage-backed securities that are running off. We're kind of targeting a duration there or thinking about 20-year MBS on average. Sometimes we're buying 20, sometimes we're manufacturing 20s by a mix of 15-year and 30-year. We were in a little bit, although not in a big way, in some of the five-year spaces as rates have moved around. That was more back-end loaded this quarter as rates have started to move up.
The other thing which I mentioned in the prepared remarks, we're starting to retain some retail mortgage production. As we think about duration, and we think about what is in our eyes the best choice or maybe the least bad choice of how to invest the dollars right now, we do like the mortgage space. If we have a choice between the MBS or retaining our own production, we like the yield better on our own production. We retained about $400 million of production. That started through this quarter but had the biggest impact in September. We'll continue that practice through the fourth quarter and into the early part of next year. Obviously, we'll be paying close attention to where rates are and how we feel about holding duration either in the mortgage line or in the securities line.
That's something that we'll be watching to put that cash to work. I guess as we look forward at that plan and how it all unfolds, taking into account PPP and the hedges that are rolling off, we see net interest income kind of dropping first quarter and not declining from there or in fact, increasing, depending on what you see as the forecast for rates as the year goes on. When I look at the storm we've weathered and the patience we've exhibited on putting that cash back to work, I see us turning the corner in the next two quarters.
Very good. Just to clarify one of your comments on the notional amount, $19 billion. You mentioned $65 million improvement to net income, but you meant to say net interest income, is that correct?
Yes. Net interest income. That's correct. Yep. Thanks, Gerard.
Okay. Yep. I just wanted to check on that. The follow-up question, you were very good at giving us the details on the commercial and industrial loan portfolio, what happened this quarter with PPP and also with the floor plan loans. Can you share with us what you're seeing outside those areas, your other customers? Second, Darren, what's the total dollar amount of the floor plan portfolio?
All right. I got to write these down again, Gerard. Other industries, what's going on, the floor plan. Floor plan balances are roughly, call it a billion and a half. What's interesting, Gerard, is when you look at floor plan balances, really it's much about line utilization and the commitments you have with those floor plan dealers. Line utilizations are running around 25%, which this is the part of the year when those are typically lowest, but never that low. When you think about where they are on average in normal times, they're about 70%. When I look at where those floor plan balances are compared to what I would consider normal, it's probably $2 billion-$3 billion lower than what might normally be the case.
When you look outside of the PPP loans, you're seeing some movement up in C&I business in the areas that you might expect, in some construction, in manufacturing. Interestingly, a little bit in resources and ag, and less in some of the other areas. When I look at new originations and what the trends are in new C&I originations, they've ticked up every quarter this year. It would be nice for all of us if they were screaming up, but that's just not the case. They are consistently going up, and this quarter's C&I originations were the highest in the last seven quarters. You all are well aware the fourth quarter tends to have a little bit of seasonality as folks in industry, both in C&I and in commercial real estate look to get deals done before tax years end.
Holding the fourth quarter to the side, it was another constructive quarter of improvement and it's slow and steady moving up in a direction that we like. The margins are holding fairly solid as well. It's encouraging trends. We'd obviously welcome a little bit faster, but the trend is our friend here.
Thank you so much.
One point of clarification I'd add is on the outlook for the hedge benefit next year. That assumes that short-term rates, whether the Fed funds target or LIBOR or SOFR, do not start to ramp up rapidly.
We'll go next to Matthew O'Connor, Deutsche Bank.
Good morning. I was hoping you could update us on the latest estimate on pro forma capital once the Invesco deal closes and the marks are done.
Sure, Matt. Just to make sure I've got the question right. Is the question specifically will the deal still be capital accretive, or is the question, once the deal's done, how will we move down the CET1 ratio?
The initial estimate at closing. You're at 11.1 right now. If you had closed the deal, your latest thoughts if you closed the deal right now, where does the 11.1 go to?
Got it. Okay, perfect. It's going to be right around there. It's probably within 10 or 20 basis points of 11.1. The reason why I haven't given you an exact number is when we do the final marks, where interest rates are will have a big impact on the interest rate marks, which was one of the pieces of helping us be capital accretive. The other thing, obviously, is the credit marks. With the credit improving. That lowers the amount of loans that are classified as Purchased Credit Deteriorated and increases the non-PCD loans, which increases the day two famous double count. That will perhaps bring down the impact on capital or reduce capital at close. As you recall, that will just increase the accretion to EPS as we go forward. We should be right around that 11 spot.
Just north of it would be my belief based on what I see today.
Okay. What are the thoughts on kind of what a near or medium-term target is for that CET1 ratio? Obviously, where I'm getting at is how aggressive will you be on buybacks to get there? Thank you.
I think it's fair to say that harboring capital has never been a characteristic of M&T. As we mentioned, at 11.1, we're near an all-time high. We'll be looking to bring that down. Ideally, we'll bring it down because we'll have asset growth, and we'll see the CET1 ratio come down as we continue to build the balance sheet and support our customers, both the existing M&T and the new People's customers. Based on the trends, that will probably be slow. We'll look to bring that ratio down. When you're generating 30 basis points to 40 basis points of capital every quarter, obviously, you've got to be fairly bold in your buybacks to bring that down. To run at a pace where it comes down much faster than net 20 basis points to 30 basis points a quarter would be pretty tough.
That would be, I think, a reasonable glide path to use as you think about what the rate of decrease might be. As a place to look, if you go back to how CET1 came down post Hudson City and that path, that rate of decrease, is probably a reasonable guide of how it might unfold post completion of the People's merger.
Okay. There's a lot of numbers you threw out there. Just to summarize, you generate 30 basis points, 40 basis points of capital per quarter. You suggested maybe bringing down the CET1 by 20 basis points to 30 basis points, just to summarize that.
That's right. Yep.
Yep. Okay. Thank you very much.
We'll take our next question from John Pancari with Evercore.
Morning.
Morning, John.
On the loan front, I appreciate all the color you gave around the loan growth dynamics and what you're seeing there. I want to see if you could talk a little bit about your plans for the commercial real estate portfolio longer term. I know right now you're a bit outsized there at 30% of loans and your peers are closer to 15%, and I believe you had expressed the interest in shrinking that portfolio. Could you maybe talk about what the long-term target is for the size of that book as a percentage of loans and how we should think about the cadence of that decline? Thanks.
Yeah. I'm happy to talk about that. I think the future plans for M&T and CRE have been highly dramatized in the last little while, notably by a favorite reporter in New York City. If you look at the impact of CRE in the stress test, in the December 2020 stress test, it suggested that there might be more capital-friendly ways to participate in the CRE industry. When we look at the relationships we have with clients and how we approach the market, our job is to be advisors and financial intermediaries for those customers. Historically, we've fulfilled that role by holding the loans on our balance sheet. We will still do that in the future, but we'll think more broadly and include other sources of capital and act as an intermediary on behalf of those clients.
As we shift to that kind of a thought process, it will be a very gradual and controlled shift. As you think about CRE balances and how they might change over time, there's some natural decrease that was going to happen no matter what, and that comes in the form of the construction portfolio. Right? There's a number of projects that are ongoing. If you look at CRE balances over the year, they've been relatively flat. That's some of those construction projects going through their natural cycle with those lines drawing that are building those balances. As those projects come to their natural conclusion and turn into permanent mortgages, you'll start to see those balances come off our balance sheet, and you'll see a natural decline in CRE, not unlike what we saw in the 2017, 2018 time frame with a similar period.
As we look at new originations, we'll still think about obviously supporting our clients from a construction perspective. As we look at permanent mortgages and other forms of real estate lending, we'll look not just on our balance sheet but outside. If you think about in real estate, we use our M&T Realty Capital Corp for agency loans with Fannie and Freddie, and we'll continue to do that. We also use and place some loans with insurance companies. So we'll look to broaden those into non-agency types of relationships as well. It's going to take some time. So, I guess the long story short here is the path we're on is to be able to actually provide a broader level of service and access for our clients to be capital efficient.
We'll see some of those net interest income dollars become fee income dollars as we make that transition. The decrease that you will see in balances will be gradual as we kind of work our way from where we sit today through the next few years. We don't necessarily have a hard target of what we're trying to get to. We're just trying to get to be a little bit better balanced than we might have been coming into the crisis.
Got it. Okay. That's helpful. Separately, on the expense side, could you maybe just talk about any observation of wage inflation impacts that you're seeing in the business and could that impact at all your expectations around the People's deal, either the cost save expectations or the merger related costs?
Yeah. I think the pressures we see on wages both at the bank and in our customers are not much different than is being seen around the country. The number of folks exiting the workforce is really putting a strain on those that are looking for folks who are still in the job market. We're seeing our people being coveted by others, which isn't necessarily new, but seems to be exacerbated right now. I think it's particularly acute in the IT space, which again was always a place where there's been a lot of competition, given what's happening in the world with the prevalence of technology in business. We're seeing it spread to other parts of the bank as well.
When you look at our expense growth that's related to comp, really what I would describe as regular salaries have been impacted by it, but the full effect won't be seen until next year because it's been going up throughout the year. You won't have a full one year run rate until you get to 2022. In the meantime, what's been driving our increase has been increase in revenue related or profit related compensation. For our corporate pool, we target a percentage of net income for that pool, and as net income has gone up, so has that accrual. For some of our fee businesses, as those fee revenues increase, so do the commissions earned by the employees. I think it's interesting.
I spent a bunch of time with the team going through our expenses. If you actually looked at our expense growth in 2020 over 2019, our expenses actually went down. Against our peer group, we were one of three banks that actually decreased expenses in 2020. What we're seeing this year is a little bit more of a return to normal with a little bit of pressure that we talked about from the outside world, which is making the print this year higher than everyone else just because of math. If you look over the course of the last couple of years, our expense growth is pretty much right in line with the peers, if not a couple % basis points below. As it relates to PUB, that's our internal acronym for People's United Bank.
We believe that the expectations for accretion are still in line with what we thought. Obviously they're subject to the same pressures that everyone else is. By and large, the cost takeouts are right where we thought they would be. The revenue synergies are what we thought they would be. From a capital perspective, as we mentioned a few minutes ago, we don't see any change there. The rate of return on the deal is still very attractive versus alternative uses of capital. No real change there. It's a challenging operating environment, that's for sure.
Got it. All right. Thanks, Darren.
We'll go next to Frank Schiraldi of Piper Sandler.
Morning. Just, Darren, follow-up on the dealer floor plan utilization comments. As you plan for budget out 2022 and think about balances in that year and growth in that year, what are your thoughts in terms of could we see, is it expected that we'll see sort of a bounce back to those older utilization rates as supply chain issues subside? Just what are your expectations for 2022 there?
Well, at the risk of being an over M&T optimist, we believe that those balances will start to come up. Knowing our conservative nature, we don't see a bounce back in those utilization rates back to normal right away. We do see an increase year-over-year. From what we can see in the industry, what we hear from the dealers, what we believe how the manufacturers will be looking and thinking about things, we think it's probably into 2023, early 2024 before you're back to the utilization rates that you were at pre-pandemic. Given that, and without any better intelligence, I would think about kind of straight-lining that over the course of the next couple of years. As we get better information, obviously, we'll share that with you.
Just watching how you're seeing manufacturing come back online, we haven't seen signs of a material uptick yet in terms of production, and therefore, we don't think we'll see the inventories bounce as quickly. We should see some steady improvement over the course of the next four to six quarters.
Okay. Just a follow-up on the securities book, adding the securities in the quarter. Could you just share what the pickup in yield is there? I know there's a lot of unknowns in terms of interest rates and moving parts, but is that kind of a reasonable expectation of moving into securities purchases on a quarterly basis, maybe 5% increase in those balances? Is this more about just pre-funding the People's United book?
I guess on the securities book and on the yield pickup, the yields that we're putting on now are lower than the roll-off yields. That's the downside. The upside is we're putting things on at call it 150 basis points, 160 basis points, and we're taking it out at 15 basis points. The pickup there is pretty good. When you look at the rate and the pace of us investing in securities, I think what you saw this quarter is a good indication that we, at a minimum, don't want to decrease the securities portfolio. We'll work to keep it flat. Realistically, we probably think about $200 million, $300 million increase a quarter in addition to replacing the runoff is a good target. At least that's kind of as we look at it and think about it, that's where our heads are today.
Again, I'll remind you that the other part of that thought process is retaining retail production, right? So we're getting mortgage exposure both in the securities portfolio as well as in the mortgage line on the balance sheet. So you'll see us deploying that cash and kind of as our treasurer would describe to me, legging into the trade as we go through the coming quarters, doing a bit of dollar cost averaging into that and rather than trying to solve it all at once.
Okay. All right, great. Thank you.
Yeah.
We'll go next to Ken Usdin, Jefferies.
Thanks. Hey, good morning. Darren, just have one follow-up on NII, understanding the PPP pull forward. You said that NII would be down in the fourth quarter. I'm just wondering what type of PPP are you expecting? Is it most of the remainder, or what type of level do we step back down to given that pull forward from third?
Well, it's a bit of a guess just because the timing of when our customers might seek forgiveness is a little bit unknown. Broad strokes, we're thinking it's kind of half the rate of accelerated capture of those origination fees in the fourth quarter versus what it was in the third. Just for perspective, the third quarter was the highest in terms of the dollar volume of those origination fees that we pulled forward. When I look at where we stand from the whole program, we're the better part of 75% of the way through, approaching 80% of the way through. There might be another $75 million of those origination fees to come. We think we see the bulk of that, I should say probably 2/3 of that in the fourth and first quarters. Then it falls off from there and trickles down.
Think in the range of $30 million in the fourth quarter compared to the $60-ish million we were in those fees this past quarter, and then continuing to drop from there.
Yeah. Okay. Got it. Thank you. If we just take that out of the third and fourth, what's happening with underlying NII then? Is that more stable, or what's happening-
It is.
underneath the surface when you kind of talk to that? Because given that you kind of still talked about your down low single digit full year guide.
Yeah. No, you've got it exactly right, Ken. The down is also a function of the swaps and the hedge portfolio that we had talked about. When we look underneath at the core, I've got to be careful when I use the word core because that means different things to different people. When we look just underneath at the balance sheet and we look at the loans and the deposits and the costs, the yields on the loans, they've been relatively stable for the last I would say four quarters. We had a slide that outlined that in the document that we posted earlier this quarter. LIBOR's been fairly stable, and most of the loans are priced off LIBOR, and so you've seen those yields fairly stable for the last little while, and the deposit costs have been fairly stable for a while.
When you look at that book, things have been relatively constant. Those yields, the spread, the margin has been very steady. The portfolio, absent the PPP, has been reasonably steady as well. There's a consistent stream of net interest income. More typical M&T low volatility. It's these other factors that have caused some of that. Those are starting to run their course, right? As I mentioned, PPP, we had a very strong quarter. There's a little bit left to go. The hedging is running down, and it's starting to run its course.
The more normal, what I would put in air quotes, M&T balance sheet and lack of volatility will start to show itself, and that's why we think we start to hit a trough in net interest income in the first quarter and start to build back from there, especially as we take some of the actions that we just talked about with rebuilding the securities portfolio and holding some of the mortgage loans on our balance sheet as we start to inflect and go back in the other direction.
Got it. Okay. One quick follow-up. Do you know just the amount of Ginnie Maes that you sold and how much that helped the mortgage banking line as far as that in and out of keeping 400 of resi but selling some of the EBO stuff?
I'll be close, Ken, but I won't be exact. I think we were within $350 million-$400 million of Ginnies that went out this quarter, and the gain, I think, was about $9 million. Just kind of round numbers of where we are. There's still a decent balance of Ginnies on the books. The rules there were they changed all the time, which is why they're still on our books. The rules were you had to be six months of consistent payments before you could re-pool it. Sometimes that got moved around a little bit and got extended to seven months or eight months . We're at the point where those should start to re-pool and come off the balance sheet. You'll see some of that in the fourth quarter and into the first quarter.
Again, those will put a little bit of pressure on the overall balances. But will show up in fee income, and is reflected in our comments about growth over the rest of the year.
Perfect. Okay. Thank you, Darren.
We'll go next to Bill Carcache with Wolfe Research.
Thank you. Hi, Darren.
Morning, Bill.
Just wanted to follow up on your comments on pursuing a less balance sheet-intensive strategy to reduce your exposure in CRE. Can you elaborate a bit on how that looks and what kind of revenue impact you'd expect that to have? Would we lose the credit exposure, which would be beneficial given the harsh treatment that regulators have been imposing on CRE as an asset class? But curious how to think about the top-line effect.
Yeah. Sure, Bill. I guess the short answer is I don't have a great answer for you. The reason is we're still early in the process. When you look at what we do today, you see a good analog is what we do in our realty capital business. It produces a steady stream of fee income for us every year. A nice margin, obviously, it's very capital friendly. The pace at which we switch and build the fee income from the net interest income, it's still nascent in terms of our development. If you look at what we're getting in that line of business today, it's probably in the $30 million-$40 million range outside of our Realty Capital Corp business.
If you look at where some of our competitors are and how big of an impact it has on their fee income, we're bullish that it can be a good source of income, capital-friendly income, that will offset what might be a decline in net interest income. I guess I would start to look at the fees that you might see at some of our peer banks as a way to start thinking about what it could become. At this point, I don't have big plans. We don't have big plans for that in terms of hitting the income statement in 2022, and we'll talk more about 2022 when we get there. We're still doing our work.
For now, I would think about it as if there's a decrease in CRE net interest income, it will be offset by fees as we do gain on sale in that line of business. Then we'll come back to you with a little bit more detail on what that shift might look like as we finalize our plan.
Got it. That's really helpful, Darren. Thinking about CCAR as a follow-up to that as your binding constraint. Any thoughts on how long before, I guess, these actions get reflected in your stress capital buffer? It sounds like it's going to take some time. From that standpoint, would it take some time to show up in the lower SCBs such that we could see it migrate back down your required level of capital somewhere back towards the 9% range?
Yeah. Bill, you're thinking about it exactly the right way. Obviously, we're a Category IV bank. As a Category IV bank, we go through the stress test every other year. This year that's upcoming, 2022, will be a CCAR year for M&T. We won't have made these changes to be reflected in that stress test, which means you're at least two years away from that. What will happen is the impact will work its way through the stress test in a couple of places. The first place you'll see it, obviously, is in the risk-weighted assets and in the losses that are assumed under stress. The other place that you see it is you see it in the PPNR, right?
The PPNR is a percentage of risk-weighted assets should go up, which will be better producing of income and capital to absorb losses under stress. Over time, that will work its way into the stress test as well. Not much you will see in the 2022 version. We'll start to see it in the 2024 version, but I would really think you'd see it in 2026 is when it would start to come in. As I said, there's a number of reasons why we're thinking about this, not just because of the capital efficiency, but we think it actually broadens the level of service we can provide to our customers. Just as a reminder, our current SCB is 2.5%. The implied one from December would've brought it up higher, but that's not our actual SCB.
We'll see where the scenario ends up this year, meaning in 2022. If it's like it was in the stress test just passed, we would not expect to see an SCB at the rate that it was calculated in the December 2020 scenario that was run, which was a pretty severe scenario.
The quick follow-up to that is there anything that could change your mind as to perhaps unwinding the decision to kind of the asset light thought process around CRE, or is that a direction that you want to head in? I guess it's not a foregone conclusion that regulators-- or is it a foregone conclusion that regulators are going to be harsher, treating CRE more harshly?
I guess I wouldn't say it's a foregone conclusion, Bill. The way I think about it is if you look over time, there's been a stress test for a number of banks every year and for many banks every other year. Each time there's a stress test, there's a different scenario that goes through that process and different asset classes are stressed. Each one of those scenarios leads you to an outcome and gives you more insight into how portfolios perform under different economic scenarios. Sometimes they stress CRE, sometimes they stress mortgage, sometimes they stress credit card, sometimes they stress indirect auto. So you got to go through a series of these, and each of those are data points that help inform your thinking. What we got in December 2020 was a data point, and it's informing our thinking.
It's not telling us that we want to never do another CRE loan as long as we live, but it said to us that there are certain asset classes and certain types of loans and how long they might exist on your balance sheet that carry a different loss assumption, and therefore, a different level of capital that you need to support them. We'll look at the mix of assets that we have on our balance sheet and be thinking like we always do about how to optimize returns, and what's the best use of our shareholders' capital. We'll still have CRE on our balance sheet, but we think that moving down this path has a number of upsides that we described in terms of capital efficiency and balancing out our income streams between net interest income and fees.
Most importantly, it gives us another outlet to help support our clients. So for those reasons, we'll move down this path. I don't think there's any turning back from that thought process. I think the question really is just as we talked about, I think in answering your first question, what's the pace at which we get there? The market will, to some extent, dictate that pace based on clients and talent. Like anything we do, we'll be thoughtful and measured in how we execute.
That's super helpful, Darren. Thank you for taking my question.
No problem.
We'll take our last question from Christopher Spahr with Wells Fargo.
Hi, good afternoon. My question is related to the role of technology and the integration of People's United, and if you can contrast that with the Hudson City deal and the integration there. This is all assuming that the integration takes place in the first quarter of next year.
Okay. I don't think my answer will change on the impact of technology if it doesn't happen in the first quarter, but I'll start with that one. I guess if you look at technology, Chris, different conversions and different mergers bring with them different levels of complexity. To kind of compare and contrast before People, our most recent two mergers, Hudson City was probably one of the least complex system integrations we did. I mean, we were basically converting mortgages and time deposits. When you think about complexity of products, those are about the easiest ones there are. There was no trust business. There really was no commercial business to speak of, no cash management, treasury management. The usage of the web and mobile were pretty light. That was a very straightforward, less complex conversion.
If you went all the way to the other end of the spectrum, I would look at Wilmington Trust, right? Now, there was a merger where we had traditional bank products, both consumer and commercial, but we also inherited a number of new wealth and trust businesses that we weren't really in in a big way. That one was much more complex in terms of the planning and setup for that. I would characterize the People's as down the middle, and that it certainly looks much more like Wilmington Trust with a little less emphasis on trust, but we gained some outside businesses in mortgage warehouse lending as well as in some equipment finance that we wouldn't have. We'll be looking to maintain those systems and set them up and be able to run them.
The flip side is the core banking system that People's runs on is FIS. That's one de-conversion from FIS onto our platform. That's something we've done before. We've done it a number of times before. We know the game plan of how to execute that. When I look at the different things that we need to get done to prepare for that integration, we're progressing along. We know what the path forward is, and we're preparing to do what we need to do to minimize the disruption to customers, to make it easy for employees to be able to service those customers. The path is known. I would characterize this one down the middle of the two ends of the spectrum.
Not without its challenges, but in the grand scheme of things, not the most complex system conversion that we've had to execute.
Thanks. I guess I should be clear. Assuming that you do get approval of the fourth quarter, you can begin the conversion process in the first quarter of next year, correct?
Assuming we get approval, I guess I would just caveat the fourth quarter. As you know, if we get approval on December 1st or past, it's probably going to make the first quarter a challenge. Depending on the timing of the approval, there might be some pushback on when we actually complete the system conversion. We'll see where that happens. It's obviously our objective to complete that system integration and conversion as quickly as practical once we have our approval and we complete the legal close. As I know you're well aware, Christopher, we've never been folks to maintain multiple deposit systems and multiple loan systems any longer than we have to because it creates risk, and it's not cost effective.
For those reasons, we would look to find an ability and a path to complete the system conversion as quick as practical once we know for certain when our approval is and our legal close.
Thank you.
This does end our allotted time for question and answers. I'll turn the call back over to Don MacLeod for closing remarks.
Again, thank you all for participating today. As always, if clarification of any of the items on the call or news release is necessary, please contact our investor relations department at area code seven one six eight four two five one three eight . Thank you and goodbye.
This concludes today's conference call. Thank you for your participation. You may now disconnect.