MFA Financial Inc (MFA) 2017 Q3 法說會逐字稿

Ladies and gentlemen, thank you for standing by. Welcome to the MFA Financial Inc. Third Quarter Earnings Conference Call. (Operator Instructions) As a reminder, this conference is being recorded.

I would now like to turn the conference over to your host, Harold Schwartz. Please go ahead.

Thank you, operator. Good morning, everyone. The information discussed on this conference call today may contain or refer to forward-looking statements regarding MFA Financial Inc., which reflect management's beliefs, expectations and assumptions as to MFA's future performance and operations. When used, statements that are not historical in nature, including those containing words such as will, believe, expect, anticipate, estimate, should, could, would or similar expressions are intended to identify forward-looking statements. All forward-looking statements speak only as of the date on which they are made.

These types of statements are subject to various known and unknown risks, uncertainties, assumptions and other factors, including those described in MFA's annual report on Form 10-K for the year ended December 31, 2016, and other reports that it may file from time-to-time with the Securities and Exchange Commission. These risks, uncertainties and other factors could cause MFA's actual results to differ materially from those projected, expressed or implied in any forward-looking statements it makes.

For additional information regarding MFA's use of forward-looking statements, please see the relevant disclosure in the press release announcing MFA's third quarter 2017 financial results.

Thank you for your time. I would now like to turn this call over to MFA's Chief Executive Officer, Craig Knutson.

Thank you, Harold. Good morning, everyone. I'd like to welcome you to MFA's third quarter 2017 financial results webcast.

With me today are Steve Yarad, CFO; Gudmundur Kristjansson, Senior Vice President; Brian Wolfson, Senior Vice Resident; Hal Schwartz, Senior Vice President and General Counsel; and other members of Senior Management.

Before we begin, I'd like to acknowledge the comfort and kindness extended to us by investors, shareholders, equity analysts and other mortgage REIT management teams after the sad passing of Bill Gorin in August. Bill's absence is a big loss to MFA and to all of us personally, and we're grateful for the support you have provided to us.

We can now move to the earnings presentation. MFA had an active third quarter of 2017 as our team successfully transacted on multiple fronts within the residential mortgage universe. With continued strong pricing in mortgage and credit assets, it is challenging to find attractively priced assets. Our investment team has worked tirelessly to find opportunities while maintaining pricing discipline.

Our third quarter earnings was $0.15 per share, down from $0.20 in the second quarter due primarily to 2 factors. The first is CRT prices, which were negatively impacted in August and again in September, largely due to concerns about possible property damage caused by Hurricanes Harvey and Irma. This price decline was modest, approximately $5 million. But since we account for these assets under the fair value option, this $5 million price decline is reflected as a loss in our income statement. Furthermore, this compares to a significant price increase of nearly $14 million for these same assets in the second quarter, which was reflected as a gain in our income statement in the second quarter.

The second factor that negatively impacted third quarter earnings was a onetime expense incurred pursuant to contractual obligations of the company in connection with the death of Mr. Gorin. Book value declined slightly to $7.70 from $7.76 in the second quarter due primarily to the declaration of a $0.20 dividend, which exceeds earnings by $0.05. This was the 16th consecutive quarter in which we have paid $0.20 dividend.

MFA's book value has been remarkably stable. For the last 5 quarters, book value has ranged from a low of $7.62 to a high of $7.76. This range is less than 2% from low to high. This is due to our asset selection process coupled with very modest leverage levels.

We remained very active in the investment market in the third quarter, investing or committing to purchase nearly $600 million of assets. Our portfolio runoff was again high in the third quarter, approximately $1 billion, but was down from the unprecedented level of $1.75 billion in the second quarter. And we expect that our portfolio runoff will continue to trend lower in the fourth quarter.

Finally, our estimated undistributed taxable income as of September 30, 2017, was $0.15 per share.

If you could turn to Page 4, please. MFA began operations nearly 20 years ago, and the company has generated strong and consistent long-term returns to investors through volatile markets and through various interest rate and credit cycles. Since January of 2000, we've generated annualized shareholder returns of approximately 15% as we have over the last 5- and 10-year periods. And over the last 12 months, our total shareholder return has exceeded 25%.

If you could turn to page 5, please. On Page 5, we summarized our current investment strategy. In 2017, we focused primarily on credit-sensitive residential mortgage assets. The credit assets we've acquired continued to perform well, tend to be short term in nature and, therefore, have less interest rate sensitivity. Many of our assets were purchased at a discount, so they actually benefit from increases in prepayment rates. Investor expectations of improved economic growth have positively impacted credit-sensitive assets, as have continued home price appreciation and repaired borrower credit profiles. While these trends are positive for the assets that we own, the resulting higher market pricing makes new investing more challenging by altering risk-return profiles. We have maintained pricing discipline, which means sometimes we don't win bids while continuously reevaluating our existing asset classes as well as exploring new opportunities. One such opportunity has been MRS-related investments, which we will talk about shortly.

Our strategy does require staying power, which gives us the ability to invest in and hold long-term, distressed, less-liquid assets and positions us to take advantage quickly in a meaningful way should market disruptions create attractive investment opportunities. We have permanent equity capital. Our debt-to-equity ratio is low enough to accommodate potential declines in marks. And our liquidity combined with portfolio runoff from short assets gives us the ability to invest significant amounts as we identify attractive investment opportunities. And we invest with a focus on the long-term performance.

Turning to Page 6. The supply of credit-sensitive residential whole loans is sporadic. That is, we can only buy them when they are for sale as opposed to, for instance, Agency MBS, which can be bought or sold every day. Despite a challenging investment environment, we purchased or committed to purchase nearly $600 million of assets in the third quarter.

Portfolio runoff, while still high, slowed from $1.75 billion in the second quarter to $1 billion in the third quarter. We did see some spread widening in CRT securities due to concerns about the 2 hurricanes that made landfall in August and September. But as more data becomes available, it appears that much of this concern was likely an overreaction and spreads have largely recovered.

Now I'm going to turn the call over to Steve Yarad, who will provide further details on the results for the third quarter.

Thanks, Craig. As Craig has outlined earlier in this presentation, MFA's financial results for the third quarter and for the year-to-date were highlighted by book value and dividend stability. However, our Q3 earnings were lower than prior quarters.

Please turn to Page 7, where we present a summary analysis of the key items impacting net income this quarter. In reviewing our results, you will note we've continued to see an increasing contribution of other income items to overall net income. Recall that we discussed this in our second quarter presentation. This trend continues to be primarily attributable to the following 3 factors. One, runoff in assets and associated liabilities that generate net interest income, including Agency MBS, Legacy Non-Agency MBS and RPL/NPL MBS. Two, acquisition of assets that we account for using the fair value option, such as NPL whole loans and CRT securities. When we elect the fair value option on an asset, we present the changes in the fair value of the asset each period directly in other income. For NPL whole loans, this includes any cash coupon payments received during the period and gains that typically occur when loans are liquidated or otherwise paid off. And three, higher funding costs, primarily reflecting recent Federal Reserve interest rate increases, which has impacted net interest spreads.

Further, as just noted, while income from NPL whole loans is reported in other income, we are required to present the funding costs for this portfolio as interest expense, which reduces net interest income due to the mismatch in the reporting of income and funding costs for this asset. Consequently, as that NPL whole loan portfolio continues to grow, we would expect that the contribution of other income to overall net income would increase, all else remaining equal.

During the third quarter of 2017, we continued to see the impact of this shift, as net interest income declined compared to the second quarter. In addition, as Craig has discussed, other income also declined principally due to decreases in unrealized gains from CRT securities, which exhibited price volatility this quarter, primarily due to market concerns related to hurricanes in August and late September.

Overall, for the quarter, we recorded unrealized losses on our CRT securities accounted for using the fair value option of $5.2 million. These compare to unrealized gains on this portfolio of $13.8 million in the second quarter. In EPS terms, this represents a sequential quarter decline in earnings of approximately $0.05 per common share.

On our second quarters earnings call, we noted that the increased contribution of other income to MFA's overall earnings, including some items accounted for using the fair value option, may result in fluctuations in the overall level of MFA's net income in future quarters. Having said that, it should be noted that at September 30, 2017, our CRT portfolio is in an overall unrealized gain position of approximately $41 million. In addition, since quarter-end, CRT prices have exhibited less volatility, have again tightened and, in fact, are approaching previous tight levels.

In addition to the impact of lower other income, earnings were also impacted by higher-than-usual G&A expenses this quarter due to the aforementioned nonrecurring impact of the company's contractual obligation to the estate of Bill Gorin. The net impact of the additional expenses recorded reduced EPS by approximately $0.01 per common share.

And now I'd like to turn the call over to Gudmundur Kristjansson, who will provide more details of our investment activity and portfolio performance for the quarter.

Thank you, Steve. Turning to Page 8, where we show third quarter investment flows. We invested approximately $600 million in the quarter and found opportunities in nearly all of our asset classes, with most activity in RPL/NPL whole loans, RPL/NPL MBS and MSR-related assets. Assets runoff declined from the second quarter, but remained elevated at approximately $1 billion in the third quarter.

We expect runoff to continue to decline in the fourth quarter. We've added MSR-related asset to our asset table as these reinvestments have grown to an excess of $400 million. These investments are securities and loans backed by mortgage servicing rights. Their performance is dependent on the performance of the underlying servicing rights. But importantly, due to structural features and overcollateralization, the price volatility of these assets is substantially lower than the underlying servicing rights.

Turning to Page 9. MFA's yields on interest-earning assets and net interest rate spread remain attractive and relatively stable despite recent Fed's fund increases and the persistent overall low interest rate environment.

Our leverage was little changed from last quarter, but was lower than a year ago due to large pay-downs in the second quarter and the fact that we have been replacing higher-levered assets, such as Agency MBS, with assets such as whole loans that utilize less leverage.

Turning to Page 10. Here we show the yields, cost of funds and spreads for more significant holdings. Given the current yields of our assets and the yields we're seeing in the marketplace, we believe that with the appropriate amount of leverage with each of our asset classes, we will continue to generate attractive returns for our shareholders.

Turning to Page 11. Here we will review MFA's interest rate sensitivity. Our asset duration changed little in the third quarter, declining 2 basis points to 131 basis points, as our asset allocation was relatively unchanged. The size of our interest rate swap portfolio was unchanged in the quarter at $2.55 billion, while the hedge duration declined approximately 20 basis points to minus 230 basis points as our swap shortened naturally over time. In addition to market value protection, our interest rate swaps currently hedged about 1/3 of our repurchase agreement. MFA's net duration was unchanged in the third quarter at 76 basis points. In addition to low net duration, it is important to emphasize that due to the company's emphasis on credit-sensitive assets and the seasonal nature of our portfolio, we have very limited exposure to long-term interest rates.

Turning to Page 12. MFA's strategy of limiting interest rate risk has consistently delivered book value stability, limiting the quarter-over-quarter changes in book value we have experienced. In fact, in the last few years, the largest quarter-over-quarter change in book value was 4%, with the average change being less than 2%.

In addition to book value stability, our portfolio continues to exhibit low sensitivity to changes in prepayment rates, as the discount accretion in our Legacy Non-Agency MBS portfolio continues to dramatically outpace the premium amortization in our Agency MBS portfolio. And the impact of Agency MBS premium amortization on MFA's earnings has diminished over time. This is primarily due to the fact that the 29-point average purchase discount on our Legacy Non-Agency MBS is much larger than the 4-point average purchase premium on our Agency MBS portfolio.

In addition, strong home price appreciation and better access to credit for Legacy Non-Agency borrowers continue to support prepayments on our Legacy Non-Agency MBS, while seasoning of our Agency MBS portfolio and lower loan balance for our 15-year fixed Agency MBS continues to limit prepayment volatility on our Agency MBS holdings.

Due to these factors, we believe MFA's earnings will continue to be relatively insensitive to changes in prepayment rates.

With that, I will turn the call over to Bryan Wulfsohn, who will discuss our credit-sensitive assets in more detail.

Thank you, Gudmundur. Turning to Page 13. The residential mortgage credit market continues to enjoy both fundamental and technical support. Interest rates and mortgage rates remain low by historical standards.

The unemployment rate is down to 4.2% from 4.9% a year ago. And the number of homes with negative equity continues to fall, estimated at 5.4% in September versus 7.1% a year ago. When people have equity in their homes and are employed, they tend to make their mortgage payments.

According to the CoreLogic Loan Performance Insights report released in October, the latest reported month-end delinquencies dropped 0.9% to 4.6% versus a year ago.

Turning to Page 14. We continued to have success acquiring packages of nonperforming and reperforming residential assets, adding over $200 million in the third quarter. Returns continued to be in line with our expectations of 5% to 7%. We believe our oversight of servicing decisions and active management of portfolio produces better economic outcomes.

Again, as a reminder, our credit-sensitive whole loans appear on our balance sheet on 2 lines: loans held at carrying value, $639 million; and loans held at fair value, $1.1 billion. This election is permanent and is made at the time of acquisition. Typically, we elect carrying value for reperforming loans and fair value for nonperforming loans.

Turning to Page 15. This slide shows the outcomes for loans that we purchased prior to September month-end of 2016, therefore, owned for more than 1 year. 1/3 of loans that were delinquent at purchase are now either performing or paid in full. Another 1/3 have either liquidated or REO to be liquidated. And another 1/3 are still in nonperforming status. We are pleased with our performance since modification, as over 80% of our modifications are either performing or paid in full. These results have outperformed our initial expectations for reperformance.

Turning to Page 16. We scaled back our purchases of RPL/NPL MBS in the third quarter. The current market yields for A1s are approximately 3% and around 5% for A2s. This asset class has seen strong demand, leading to existing deals getting reissued once callable at lower yields and tighter spreads. Significant appetite from unlevered buyers continued to be the driver for the spread tightening.

Turning to Page 17. The credit metrics on our loans underlying our Legacy Non-Agency portfolio continue to improve. 100% of the loans underlying our Legacy Non-Agency portfolio are now amortizing. This principal amortization, together with home price appreciation, continues to reduce LTVs.

Delinquencies are cheering. 60-plus-day delinquencies as of September 30 for the portfolio were under 12%. On this page, we illustrate the LTV distribution of current loans in the portfolio. The red bars on the right-hand side represent the at-risk loans, where the homeowner owes more on the mortgage than the property is worth. These are the loans we worry most about transitioning to delinquent and defaulting in the future because the borrowers are underwater. As you can see, these red bars are disappearing and almost gone entirely. Please also note the increasingly large black bars on the left side. Loans with LTVs below 80% are attractive refinancing candidates and 80% -- 86% of the current loans have LTVs at 80% or lower. A combination of low rates available today and a 30-year amortization term versus a 20-year remaining term on these loans today can offer homeowners substantially lower monthly payments.

Of course, given our deeply discounted purchase price of these assets, we are very happy when the underlying loans prepay.

And now I'd like to turn the call back over to Craig.

Thank you, Bryan. So in summary, we've remained active in the investment market. We have maintained our disciplined pricing approach, which sometimes means we don't win bids. We're investing for the long term, so we are keenly aware that reaching too much for investments can lock in years of suboptimal returns. Our investment team is working harder than ever to reevaluate various asset classes and diligence new opportunities. We cannot always predict what the next attractive investment opportunity will be, but we are quite confident that we will have a seat at the table, the expertise to understand and structure the transaction and ample capital to be able to invest in meaningful size.

This concludes our presentation. Operator, would you please open up the call for questions?

(Operator Instructions) Your first question comes from the line of Douglas Harter from Crédit Suisse.

This is actually Sam Choe filling in for Doug Harter. So you guys mentioned that there was a reduction in credit reserves. I believe it was on the Legacy Non-Agency assets. I was wondering if you can walk through that decision and kind of elaborate more on how you see credit quality trending in the next quarter and into 2018.

Sure, Sam. So I think we probably -- it is on the legacy book to answer your question. But I think we've probably had a similar credit reserve reduction in every quarter for the last, I don't know, 3, 4, maybe 5 years or so. So our evaluation process is basically to take a subset of all those legacy bonds. So every bond gets looked at approximately at least once per year. And so that analysis will include such things as updating the mark-to-market LTV, looking at delinquency trends that we've seen in the past, look at prepayment rates and, again, we're just -- we're evaluating those initial assumptions or the most recent assumptions on each of those bonds, and the $15 million basically comes out of that process of bottom to top review of those assets. And our expectation of future cash flows has improved. Now the way that impacts us, it increases the yield perspectively on those bonds. So it has no effect on book value, has no effect on earnings. But if we were accreting to, let's say, 91%, and we change the credit reserve and now begin accreting to 92% that will increase the yield on that bond.

Got it. So look at it from your perspective instead of earnings, I guess.

So just to clarify just one other thing that Craig has said. As it increases the yield, it will result in increased future interest income on that. It doesn't impact the period in which we make the credit reserve change.

Got it. The current -- okay. Got it. And then from a modeling perspective, I mean, there are a lot of one time items on the operating expenses. How should we be thinking about the normalized expense level?

This is Steve Yarad. If you look at our G&A over the past 7 or 8 quarters and as the business has changed and grown a little bit, we've sort of been in G&A to equity in the sort of 1.5% to 1.6% range, which equates to $1 value of that G&A, including compensation of somewhere around $11 million to $12 million a quarter. And I think that's kind of the way you should be thinking about it as we move forward.

Your next question comes from the line of Jessica Levi-Ribner from B.Riley FBR.

The first one on CRTs. With spreads out this quarter, I guess we would've expected you to be more active in the market, especially if the hurricane fears were overblown. So how did you think about that? And how do we think about kind of the marks in the fourth quarter, given maybe some spread normalization?

So as far as opportunistically buying these as spreads widened. Unfortunately, there were very few trades. The spreads widened, but most of the holders were reluctant to sell at those levels. So the marks were lower but, truthfully, there was very little volume that transacted. They probably widened by 30 basis points from the tights. And much of that was just -- was an overreaction. What's happened since then is you've seen a little bit of analysis. So for instance, if you look at delinquencies after Katrina, delinquencies spiked immediately, but within a few months had come down significantly. In those areas, delinquencies could have been as high as 30% initially. But ultimately losses, I think, were less than 1.5%, so less than 1.5% of those borrowers ultimately defaulted. You've had estimates that have come out from the GSEs about how affected the various areas were. And look, even those initial estimates changed a lot as well. So initially, I think the Fannie Mae, and they use FEMA data, projected 2.3% exposure to Harvey. They then revised it to 0.4%. And initially they expected 5.3% exposure to Irma, which they revised down to 1.8%. So it's just a lot of excitement and sort of irrational and very little trading. What's happened since then is, as you've seen some of these estimates come out of Street firms and the GSEs, the prices have largely recovered. So very rough numbers, and we don't have final numbers for October yet. But I would say that the $5 million that our CRT prices were down in the third quarter, we've probably made that back in October.

Okay. So almost a full reversal.

Correct.

Just shifting over to the whole loan acquisition. Can you talk a little bit about the competition in that market? And how you think about NPL versus RPL? How many -- are they -- are you looking at earlier clean pays like [707]? Or how do you think about that market in the competition?

We have seen over time the competition continues to grow, especially on the reperforming space, where we've seen a little bit less competition. And just sort of a static pool of buyers is in the -- on the nonperforming side just because of all the sort of the work and the infrastructure that needs to be in place to be able to handle those types of assets. So I think you can assume where we have been more active is on the nonperforming side. And it's because that's where we tend to see a little bit better returns and better perspective returns in these spreads.

Your next question comes from the line of Steve Delaney from JMP Securities.

If I could, can we start with the undistributed taxable income on Slide 3. I'm just curious, Steve, if there's -- what the timing requirements might be with the IRS? Or are you going to be able to carry that into 2018?

Thanks, Steve. Yes, so the way it works is that you have until approximately 9 months into 2018 just to fully distribute your taxable income. So we would expect that, that balance would be sustained through the -- we may reduce a little bit and we would carry it in 2018.

Okay. So this largely stems from maybe 2017 tax year items that the distribution can be delayed as late as September of '18? Is that the way to think about it?

Pretty much, Steve, yes.

Okay. All right. Very good. I have to apologize, I had not focused last quarter on the MSR-related investments that appeared to be related to some -- the MSR, obviously, being the collateral for some term loans. And I'm just curious how you found that opportunity. Obviously, if there's something proprietary, I don't want -- I'm not asking you to disclose that, but that -- is there a market there that, that could be a repeat opportunity, because, obviously, the coupons at 6% and the leverage would suggest a very high ROE for you on that asset class. So we'd just like to hear a little bit more about that opportunity.

Steve, I'll talk a little bit generally and then if you want to get into more details, we can have Gudmundur address that. By the way, you didn't miss anything last quarter, because it would have been exceedingly difficult to figure out that we had any MSR-related investments.

You buried it. It was kind of buried last quarter.

So we broke it out for the first time this quarter. But these -- they were -- actually, our first such investment was almost a year ago. We didn't have a lot of size. We didn't, frankly, from a competitive standpoint, want to publicize this initially. This is the market that we've actually been involved in for a little bit over a year. Over the years, we've been asked about MSRs as a possible asset class. And while we've looked pretty hard at these, we never bought MSRs. Because we don't own 30-year fixed-rate agencies, these wouldn't have grown to be as complementary as they might have had we owned 30-year fixed rate agencies. In addition, we would've encountered some operational complications because we're not licensed to own MSRs. And then finally, we also discovered in the process that it's very difficult to obtain leverage on MSRs. But ultimately, our work on MSRs paid off, because we had a seat at the table as a somewhat innovative structure was developed to obtain leverage on MSRs. And as Gudmundur said before, the instrument that we own is structured such that we can also borrow against it in the repo market. So our investment vehicle is backed by MSR cash flows. It's overcollateralized. It's also backed by the issuer. But our cash flows are floating rate LIBOR based. So the value of our investment is much more stable than the value of the underlying MSRs. And maybe I'll let Gudmunder talk a little bit about future opportunities and the size of that market.

Yes, Steve. So it's an interesting opportunity. And as Craig pointed out, I mean, we've spent time around the MSR space. And then we kind of learned that it was complicated and somewhat illiquid to finance MSR assets at the end of last year. And so we got involved in some creative transactions there. And the reason why this is a bit complicated is because it's -- you have the underlying MSR assets, as you know, which is volatile in nature, and you have to actually analyze a few payments and understand the underlying assets well. But on top of it here, you have some structure and creativity. And then you really have to look to the partner, which, in this case, is a servicer, and figure out how they're doing from a company point of view. So there's definitely some creative work being done there and there's numerous transactions that have happened where the structure is ideally loan or a note where the MSR is pledged in some nature as collateral to the asset. And there's probably -- going forward, as you think about that, maybe there's 1 or 2 transactions per quarter. It's definitely not a large space at the moment, but it definitely is interesting and provides attractive opportunities, as you noted, based upon the high coupon rates that we see. Interest -- also importantly, just to understand and so we're clear, the interest rate sensitivity of our instrument is low because it has a floating rate coupon, and it has a fixed maturity opposed to the underlying MSR, which is obviously, as you understand, quite sensitive to changes in interest rates.

That's very helpful background, and maybe we'll follow up offline on that just to get a little more better understanding about the opportunity. I think we actually saw, I believe, earlier this year securitization transaction backed by MSRs. I think that was Ginnie Mae MSRs. But glad to hear that MFA is plugged in. Not surprised but pleased to hear that you've identified that opportunity. And just one last thing, just looking at your capital structure, we're all kind of wondering where rates are going to go on the longer end. And given the tightness in the corporate debt market that we've -- we are seeing still, just wondering when you look at your balance sheet and you look at your 8% baby bonds and your 7.5% preferred, when you look at those coupons versus investment returns, do you see any opportunity for partial redemption there to -- maybe to lower your blended cost of capital? That's my last question?

Sure, Steve. So yes, we do look at that. Understand that whether it's preferred or the baby bond structure, those are structured as retail securities. So typically, they're pretty expensive to issue. They cost over 3% to issue. So it's not quite as simple as just comparing the coupon, but we're very much aware of where corporate bonds and even junk bonds for like companies have been done as well as the preferred space.

Your next question comes from the line of Bose George from KBW.

It's Eric on for Bose. Can you guys just give a little more detail about how you envision deploying some of your excess capital and runoff going forward? And then how big would you actually like to see the MSR portfolio grow over time?

Sure. So in terms of growth, I think our growth asset class is still credit-sensitive loans. As Bryan said, at least for the moment, I think it's probably more skewed to nonperforming than reperforming loans. But obviously that can change. But as I also said in my prepared remarks, as market spreads change across all of our asset classes, I think we try to continually sort of reexamine all the asset classes. So as an example, we really haven't been active in the agency market for some period of time, probably 4 years or so. And spreads in agencies are very tight at the moment. But relative to some of the tightening that we've seen in say legacy or in RPL/NPL securities, agencies probably, on a relative basis, look better now than they did 3 or 6 months ago. So again, I'm not suggesting that we're going to run out and buy agencies. But we do undertake that process, so we look across all our asset classes. There are some possibly new loan-related opportunities that we can look at as well. So we're pretty active. It is a tough market environment to find cheap assets. But I think we're -- we've shown in the past that we have been somewhat innovative in our ability to analyze different and creative structures and commit significant capital. So the good news is, when something new does come up, we typically get the first call -- one of the first 3 calls. And as far as the MSR-related assets, I think we'd love to be able to grow that asset class. As Gudmundur said, it's maybe 1 or 2 deals a quarter. So it's not the type of assets that we could see having billions invested in, but certainly hundreds of millions. It's over $400 million now. We actually had a deal, that first deal that we bought, actually, was called a year later. So the number was actually over $500 million, now down to $400 million and change. But that could grow by maybe $100 million a quarter, $150 million a quarter, if things go really well.

Yes. That's a helpful answer. And then I feel like this is always a question that we return back to every couple of quarters. But just monetizing more of your unrealized gains and just the pace that we should think about that going forward, particularly from a modeling perspective, of course.

So yes, unfortunately, I think your pain was the modeling perspective. But I'll actually give you some details. Typically, we don't really discuss specifics of our Legacy Non-Agency portfolio management. But I'll share some details of these third quarter sales. And I characterize these as very opportunistic, but also part of a consistent strategy of managing a mature portfolio. So we sold about $45 million market value of bonds this quarter, which was comprised of 12 different positions. 8 of those 12 positions were position sizes below $3 million. So I'm sure you know odd-lot positions often trade at a discount to round lots. So if we can achieve round lot pricing on odd-lot positions, we think these have prudent sales.

7 of the 12 bonds -- obviously, some of these hit multiple categories. 7 of the 12 bonds are low-loan kind of bonds. So that's fewer than 100 loans in the pool. And what happens with low-loan kind of bonds is that as they factor down and have fewer loans, they also trade at a concession because your monthly prepayment or the fall percentages can be very idiosyncratic. And again, to the extent that we can execute a sale without a price concession, we think those are prudent sales. 5 of the bonds that we sold traded over par and 1 at a price of nearly $102. So these are bonds on which we have, obviously, realized significant price appreciation, and in most cases, we see very limited further upside. And then finally, the largest position we sold this quarter, which was a $17 million current base position, was a bond that we sold slightly over $101.5, and this bond is currently callable at par So again, these are very opportunistic-type sales. And it's part of just managing a mature portfolio. But hopefully, that will give you a little bit of color as to sort of what goes into the thought process as we do monetize those bonds, as you said.

Your next question comes from the line of Rick Shane from JP Morgan.

My questions have been asked and answered. We're all going to miss Bill a great deal. He was a lovely man, and we're sorry to see him pass away.

Thanks, Rick.

(Operator Instructions) And at this time, there are no further questions.

All right. I'd like to thank everyone for joining us today. And we look forward to speaking with you again next quarter.

Ladies and gentlemen, this conference will be available for replay after 1:00 Eastern Time today through February 2. You may access the AT&T Teleconference replay system at any time by dialing 1 (800) 475-6701 and entering the access code 432695. International participants dial (320) 365-3844. That does conclude your conference for today. Thank you for your participation and for using AT&T Executive Teleconference. You may now disconnect.