Two Harbors Investment Corp (TWO) 2019 Q2 法說會逐字稿

Good morning. My name is Gerod, and I will be your conference facilitator. At this time, I would like to welcome everyone to Two Harbors Second Quarter 2019 Financial Results Conference Call. (Operator Instructions)

I would now like to turn the conference over to Maggie Field with Investor Relations for Two Harbors. Please go ahead.

Thank you, and good morning, everyone. Thank you for joining our call to discuss Two Harbors' Second Quarter 2019 Financial Results. With me on the call this morning are Tom Siering, our President and CEO; Mary Riskey, our CFO; and Bill Roth, our CIO.

After my introductory comments, Tom will provide an overview of our quarterly results and long-term strategy; Mary will highlight key items from our financials; and Bill will review our portfolio and investment opportunities. The press release and financial tables associated with today's call were filed yesterday with the SEC.

If you do not have a copy, you may find them on our website or on the SEC's website at sec.gov. In our earnings release and slides, which are posted in the Investor Relations section of our website, we've provided a reconciliation of GAAP to non-GAAP financial measures. We urge you to review this information in conjunction with today's call.

I would also like to mention that this call is being webcast and may be accessed on our website in the same location.

Before I turn the call over to Tom, I would like to remind you that remarks made by management during this conference call and the supporting slides may include forward-looking statements. Forward-looking statements are based on the current beliefs and expectations of management, and actual results may be materially different because of a variety of risks and other factors. Such statements are typically associated with the words such as anticipate, expect, estimate and believe or other such words. We caution investors not to rely unduly on forward-looking statements. Two Harbors describes these risks and uncertainties in its annual report on Form 10-K for the fiscal year ended December 31, 2018, and in other filings it makes or may make with the SEC from time to time, which are available in the Investor Relations section of Two Harbors website and on the SEC's website at sec.gov. Except as maybe required by law, Two Harbors does not update forward-looking statements and expressly disclaims any obligation to do so.

I will now turn the call over to Tom.

Thank you, Maggie, and good morning, everyone.

We hope that you had a chance to review our earnings press release and presentation that we issued last night.

Please turn to Slide 3 to review our results. This quarter was highlighted by our strong book value growth. We delivered a 5.4% quarterly return on book value and a 14.7% return on book value for the first half of 2019.

Our book value performance was a result of both portfolio positioning and active hedging. We are also pleased to have completed our first MSR securitization in the quarter, which has competitive terms compared to our bilateral facilities and gives us more financing flexibility for MSR.

Please turn to Slide 4. In the second quarter, our sector traded under stock price pressure, which we believe was driven by dividend cut concerns, lower interest rates and a flatter yield curve as well as an emergence of faster money buyers and sellers, which created more intraday volume and volatility.

As you know, our goal is to deliver book value stability through a variety of market environments. We construct our portfolio with this long-term focus in mind, while endeavoring to minimize the impact of short-term changes in rates.

In particular, we believe our strategy of pairing MSR with Agency RMBS generates a more stable risk-adjusted return throughout market cycles compared to hedging agencies with swaps only.

Additionally, our unique portfolio of legacy non-Agency securities is one that can't be easily replicated, and we believe this strategy will continue to generate attractive returns given the strong tailwinds in housing.

On this slide, you can see how these differentiating factors have resulted in total stockholder outperformance of 45% since our inception.

Recently, we have received some questions on possible GSE reform. We think it is unlikely that anything of substance happens prior to the 2020 election. Obviously, that outcome could greatly influence any reform efforts.

Finally, I would like to congratulate Bill on behalf of Two Harbors on his retirement at the end of the year. Under his leadership, we have built a strong team of investment professionals. We'd also like to congratulate Bill Greenberg and Matt Koeppen on their appointment as Co-Chief Investment Officers, beginning in 2020. We have a great team at Two Harbors, and we are very excited about the opportunities ahead.

I will now turn the call over to Mary to review our financial results.

Thank you, Tom. Turning to Slide 5, let's review our financial results for the second quarter. We generated comprehensive income of $201 million or $0.74 per share, and our book value was $14.17 per share compared to $13.83 at March 31. The increase in book value was primarily driven by the underperformance of higher coupon agencies in specified pools with MSR performing as expected and offsetting current coupon spread widening. Credit spreads also improved and contributed to positive book value.

Finally, I'd note that through the end of July, our third quarter book value is up between 2% to 3% after accruing for dividends.

Moving to Slide 6, let's review our core earnings results. Core earnings, including dollar roll income, was $0.39 per share in the second quarter, representing a return on average common equity of 11.1%.

Core earnings was impacted primarily by the higher prepayment environment, however, we also realized a minor impact from the compression of the LIBOR repo spread. We avoided much of the effects from spread widening that occurred in the first half of this year by maintaining longer repo maturities.

We do expect to see a small impact from this in the third quarter. However, we are actively managing our repo laddering to minimize both costs and exposures.

I'd like to take a brief moment to discuss a change we made this quarter to core earnings. The change involves the modifications approach we used to calculate MSR amortization. Specifically, an adjustment for any gain or loss on the capital used to purchase the MSR.

We recognize that core earnings is an important metric to the investment community. We believe that our new approach allows core earnings to better reflect how the carry earned on MSR varies as a function of prepayment rates.

As such, we believe it is a much improved and more accurate reflection of the economic returns our portfolio can generate. Further details are provided in the Appendix Slide 19.

Finally, I'd note that our other operating expense ratio, excluding noncash LTIP amortization, was 1%, down slightly from 1.2% in the first quarter.

As a reminder, we anticipate that our expenses will remain stable in the low 1s in 2019.

Turning to Slide 7. Our portfolio yield declined in the quarter to 3.93%. This was primarily driven by lower agency yields, reflective of the lower rate and higher prepayment environment. We also purchased discounted legacy non-Agencies at lower base yields. However, as we discussed in the past, we anticipate that over time, these discounted bonds will generate double-digit total returns through a combination of both strong yields and price appreciation.

Now let's review our financing profile, as shown on Slide 8. Our economic debt-to-equity ratio, which includes the implied debt on our TBA positions, was 7.8x at June 30. Our average economic debt to equity was up slightly quarter-over-quarter at 7.2x from 7x.

Our diverse financing profile includes a mix of traditional repo, convertible debt, revolving credit facilities and MSR secured term notes. We have 25 active agency repo counterparties and the market continues to function efficiently for us.

As I mentioned earlier, we are focused on managing the laddering of our repo maturities and minimize costs and exposure to repo LIBOR spread changes.

The improvement in financing for both MSR and non-Agencies presents an ongoing opportunity for our business. On the MSR front, we closed our first financing securitization this quarter. This $400 million securitization of 5-year secured term notes has 12-month prepayment protection and a spread of LIBOR plus 280.

The benefits of this structure includes scalability, competitive advance rates and pricing, and the length of the term is greater than bilateral facilities.

Across all of our bilateral MSR facilities, we had $300 million outstanding with the total capacity of $790 million as of June 30.

With respect to non-Agencies, our custom spreads have continued to be favorable and were consistent quarter-over-quarter.

With that, I will now turn the call over to Bill for our portfolio update.

Thank you, Mary, and good morning, everyone. Please turn to Slide 9. In the second quarter, global growth concerns, especially trade issues with China and Mexico, fueled the rally in interest rates and impacted economic growth expectation. This resulted in a flatter yield curve and 3-month LIBOR above most longer-term rates. This curve shape reflects the market's expectation that the fed will lower rate, which ended up occurring in July and the anticipation is that there will be further rate cuts in the latter half of 2019.

In the mortgage market, the agency basis widened with current coupons widening by about 1 point, while higher coupon mortgages outperformed widening only by about 0.375 points. Specified pools performed quite well in the rally, offsetting much of the base's widening. MSR performed as expected, declining in value, in line with its duration and current coupon spread. Residential credit assets continued to perform well. Discounted legacy non-Agencies benefited from lower expected forward LIBOR, which would have the effect of increasing the excess spread available to cover future losses.

As a result, prices were modestly higher by about 1 point.

Let's move to Slide 10 to review our portfolio, which, at June 30, was comprised of $32 billion of assets and about $9 billion of net long TBA.

From a capital-allocation perspective, 76% of capital was allocated to our rate strategy and 24% to credit.

In terms of portfolio activity, as the market rallied in the quarter, we added Agency RMBS in part to manage our duration exposure but also to take advantage of the wider spreads available in the market. Specifically, we added approximately $5.7 billion of 3.5% and 4% coupon lower pay-up specified pools. We think that these bonds have excellent prepayment protection and convexity characteristics at reasonable prices.

We also reduced our 4.5% and 5% coupon higher pay-up specified pool position by approximately $950 million. As we felt these pools, we're priced at levels that had more risk than reward.

Despite the higher agency balances, however, our exposure to mortgage spread was slightly lower since our mortgage spread risk declines as interest rates fall.

One question we've been getting recently is our outlook for leverage as it has been drifting higher the past few quarters. We think about leverage in our portfolio from a risk exposure standpoint, similar to the way we think about interest rate exposure. In short, our focus is more on the sensitivity of book value relative to movements in mortgage spreads than it is on just the nominal leverage number. Although this quarter we added agencies, which increased our leverage, our mortgage spread risk actually declined.

As we've discussed in the past and as you will see on the next slide, we continue to believe that pairing agencies with MSR allows us to have higher agency leverage but with lower overall risk. We are comfortable with our current level of leverage and do not expect it to change materially from here.

On the MSR front, we did not purchase any bulk servicing in the quarter as we didn't see any packages that met our criteria, either collateral-wise or price-wise. We do expect to see a pickup in transaction volume in the second half of 2019 as there will likely be an increase in supply driven by higher refinance volume.

Finally, in our credit strategy, we added approximately $370 million in market value of discounted legacy non-Agency, which, at an average price of $58, we believe have the potential for upside price appreciation.

Turning to Slide 11. Let's discuss our risk positioning. As seen on the top of this slide, our exposure to changes in rates remains small. Consistent with our typical practice in the second quarter, we utilized an extensive amount of swaptions and mortgage options to protect our portfolio against big rate moves, either higher or lower. This hedging practice contributed to our strong performance. The bottom table of this slide shows our spread exposures to rates, up and down 25 and 50 basis points.

As we've discussed in the past, MSR has negative duration and hedges both interest rate and mortgage spread risk, which you can see in this table.

Please turn to Slide 12. The second quarter presented an interesting rate environment with a flat yield curve, LIBOR elevated relative to longer-term rates and the potential for increased refinance activity.

Nonetheless, we delivered strong returns because of our portfolio positioning, principally holding higher coupons, specified pools and MSR and also because of our dynamic use of options in hedging. Given the current market backdrop, we believe that we've positioned both our rates and credit strategies to deliver strong long-term, risk-adjusted returns. In our rate strategy, we view levered returns on agencies as currently in the low double-digit range.

We continue to believe our holdings of specified pools and higher coupons will generate attractive long-term returns. Returns on agency paired with MSR are also in the low double digits, but as we discussed on the last slide, this comes with a lower risk profile.

I'd like to spend a moment highlighting the total return opportunity that exists in our credit strategy. We believe that our portfolio of legacy non-Agencies will continue to benefit from the strengthening housing market.

Additionally, the lower rate environment is beneficial to our discounted portfolio. Residential credit tailwinds have been strong. And as you can see in the charts on this slide, sub-prime non-Agency prepayments have increased while loss metrics have come down.

Continued re-equivocation over time has resulted in and can continue to result in increased prepayments and lower delinquencies, defaults and severity. We believe these favorable dynamics will persist, driving bond prices higher and generating strong total returns going forward.

Despite the fact that the legacy non-Agency sector is shrinking, we have a substantial and unique portfolio of deeply discounted bonds, and we continue to find pockets of value and opportunities to add to our holding.

In conclusion, we believe that we have positioned our portfolio through both security selection and active hedging to drive strong returns over the long term. We are very excited about the opportunities in both our rates and credit strategies in the latter half of 2019 and beyond.

I will now turn the call back to the operator for Q&A.

(Operator Instructions) We will take our first question from Doug Harter.

Obviously the kind of the start to August has been quite volatile. Just -- if you could just talk about kind of how your rate positioning is for kind of this kind of magnitude of rate declines we've seen. And kind of how you see the prepay outlook changing with this move?

First of all, I'd like to apologize for the technical glitch we had this morning. With that, I'll turn your question to Bill.

Doug, thanks for joining us. Yes, so there were couple of questions in there. I think Slide 11 is sort of a relatively good depiction of where we were at the end of June. We continue to maintain a slightly long bias in duration, but we have a substantial amount -- I talked on the call, and one of the things that benefited us in the second quarter is the extensive use of options. Certainly, as markets become more volatile and whip around, protecting the mortgage -- the portfolio by use of options is very important and effective. In terms of what we have seen in August, it's a little hard to make sense of what's going on, but we continue to stay close to home. If the primary rate comes down, and it's been pretty sticky so far, but I guess it depends on what rate level we settle at. If the primary rate comes down, then I think you'll see a reasonable pickup in generic prepays -- or prepays on generic collateral. One of the reasons that we've been active -- continue to be active in adding specified pools, especially off of lower coupons, as I discussed on the call, was because that creates more stability in the portfolio, it's easier to hedge, and there's less prepay volatility there. So we continue to move in that direction.

Yes. The team has done a really good job managing this, Doug, and that continues into August as well.

Great. Just I guess on that, the relative -- where do you see relative attractiveness of kind of specified pools today? Obviously, given the commentary that generic speed is going to likely continue to increase but, obviously, pay-ups have increased substantially.

Yes. Great question. Yes, I think the positioning that we talked about on the call is representative of what -- where we think value is. We've been trying to buy lower pay-up off of, let's say, 3.5. We haven't seen as much off of 3s yet. But we sold pools that had performed very well and pay-ups went up quite a bit off the higher coupon. And so by recycling into lower pay-ups off the lower coupons, we come up with better convexity with less premium at risk. And we think those represent good value, and we continue to pursue those.

We will now take our next question from Mark DeVries.

Let me be the first one to congratulate you, Bill, on your retirement. I know we may have you for another quarter, but it's been a pleasure working with you over the years. First question, as you noted in the prepared remarks, leverage went up almost more than 0.5 turn in the quarter. Although, it sounds like part of that was efforts to manage both spread and duration risk. Are there any implications we should think about, though, from a higher leverage in terms of higher earnings power at the end of the quarter versus the average for the quarter?

Mark, well, thanks for the good wishes. I appreciate it and enjoyed working with you as well. The average economic leverage for the quarter went from 7 to 7.2. And keep in mind that we think of leverage in terms of including net TBAs because that is economic leverage. And I think as we discussed on the call, we frankly think about really more our risk metrics, which we talked about on Slide 11, which is, because we have MSR, MSR helps dramatically hedge mortgage spread risk. So even though we have more agencies, we actually have less risk to spread moves. That being said, we do not expect to see a leverage change materially from the levels that we've been running recently.

Yes. Leverage and [inflation] is sort of a hollow metric, Mark. Really, we think in terms of risk over kind of the headline leverage metrics. Obviously, that's an important number but in isolation, it doesn't mean all that much.

Okay. Fair enough. So I think you also noted that through some of the leveling -- or I'm sorry, laddering of your maturities on repo, you've at least mitigated some of the impact of the bit of a dislocation we've seen in that kind of repo LIBOR spread. But how much of a drag is that on earnings? And if that normalizes, how much of a lift might we see from kind of core earnings?

Yes. That's a great question. Mary's is going to answer that one.

So yes, we realized this is a small impact to Q2, about $0.01 per share due to our maintaining the longer repo maturities. I think repo costs, obviously, haven't come down like the rest of short-term rates. And it's not really the availability but really the pricing. We are keeping our eyes on it and expect it to normalize as time passes.

We will now take our next question from Bose George.

Just in terms of incremental returns, you gave us some numbers on that. But I'm just curious how the incremental kind of spreads that you're getting compares to the level you were in the quarter?

Could you rephrase that, Bose? I wasn't sure I...

Yes, sure. Just the spreads on new investments and how that compares to, I guess, the 138 basis points spread you guys had at quarter-end?

Yes. Well, I'd say a couple of things. The first thing is we talked on the call about our expected total returns that we think we can realize, generally being in the low double digits. And then on the non-Agency, clearly, the -- we've talked about this in the past sort of the baseline is lower than what we think the total return will be. If you think about it, right, we talked about the NIM. The prepaid, the repo spreads, and then obviously some purchases at lower yields of assets. But as you know, we don't really manage our portfolio based on a NIM. We really are thinking about it in terms of where we can generate the best total returns while keeping book value to stable. So I think it's kind of hard to talk about the spread, vis-à-vis, today versus the historical accounting spread rather we think about that the returns that we discussed on the call, low double digits and then potentially strong total returns on non-Agency being better than that.

The -- so then maybe just in terms of the low double -- the returns from an ROE standpoint. I mean just give us a bit of a range, so is that similar to what you have in the portfolio? Is that a little lower? Or just how would you characterize it versus the ROE you generated this quarter?

Well, I think it's probably pretty similar to what you saw this quarter. I mean obviously, one of the things we've been positioned in has been higher coupons in the specified, which did well last quarter. They continue to do well this quarter. Mary talked about our performance through July and so far this quarter. And so I think that's relatively consistent with what we saw.

Yes. Bose, we're kind of people that when we stink, we have our hats firmly in our hands. For instance, in the fourth quarter, we were very honest about our disappointment in our portfolio performance. So this was a great quarter for us. We generated a lot of alpha, and the opportunities for us going forward are abundant. So we're excited.

And then actually just on the MSR securitization that you guys did, should we think about it more as just a better structure because of the term funding? Or is there a -- any funding benefit as well from that?

Bose, so I think we think of it -- a number of benefits from it. As a term, it's definitely beneficial. The scalability, the ability to add additional term notes in the future, and it's non recore.

Okay. Great. And then actually one just on the regulatory side. You noted that GSE reform is unlikely, but just curious, what your take is on the how the QM patch issue gets resolved?

Yes. That's a good one. I think what we saw was sort of an initial shot across the bow, if you will. I think it sounds like the FHFA, they would like to get a number of things accomplished. And yet, at the same time, it seems like Mr. Calabria has also aware that there's a lot of it that might need legislation. And so I think by putting something out there on the QM patch expiring, the idea is not necessarily to have it expire but maybe have a discussion and dialogue as to what should it look like and how can the GSE continue to support housing but with the modified footprint. So I think that -- I think this is just the beginning of the discussion of how that patch and how the QM -- how that is going to play out. And I think it'll probably take a decent amount of time so...

Okay. And let me add my congratulations to you, Bill, on your retirement and to Bill Greenberg and Matt on their promotions as well.

Thanks, Bose.

Thank you.

We will now take our next question from Trevor Cranston.

I'll let my congratulations on your retirement, Bill. Question on the updated methodology you're using for your core earnings calculation on the MSR amortization. I guess I wanted to make sure I understand the change better. Is the new method that you're using, I guess, similar to what we would think of in the Agency MBS market is using like a lifetime assumed yield on the MSR that you're multiplying by the amortized cost? Or maybe you can just expand a little bit on exactly what the change was there and how we should think about modeling that going forward.

Trevor, this is Mary. So as I discussed in my [part] we did modify the amortization to include an adjustment for the gain or loss on the capital used to purchase the MSRs. So this is consistent with the rest of our portfolio where the hedging or the gains and losses on our RMBS portfolio are not included in core earnings. Appendix Slide 19 provides some more details, but I think you can think about MSR for income as the amortized cost that -- at the beginning of the period times the original pricing yield. And we really believe this is a more accurate reflection of the economic return and the carry-on MSR.

Okay. So I guess so when we think about, for example, this like third quarter to-date where rates have dropped significantly, does the new methodology imply that there wouldn't be a significant uptick in amortization because the original pricing yield hasn't changed? Or am I not thinking about that correctly?

Yes. That's correct.

Okay. Great. And then second question, I guess, more on the macro side. With the volatility we've seen this quarter and expectations for the fed to likely continue dropping rates pretty significantly. Just curious if you guys had any thoughts on the likelihood if we do continue to see things move lower in the rates market? If you think the fed would be the likely to potentially reinstitute balance sheet growth or the QE programs? And whether or not you think that might be focused on the MBS market?

That's a great question. Matt Koeppen is here, he's going to answer that one for you.

That's an interesting question. Maybe some other people in the room want to weigh in. It's a -- we're definitely in the midst of a transitioning market here. Back to the world where we are in a low rate environment and the fed is cutting, we definitely -- I think you would -- you could expect to see a higher volatility environment, mortgages continuing to be under pressure. I think in terms of looking forward to QE again, I don't think anybody is really talking about that or discussing that. That seems -- that's probably a ways down the road. But we are continuing to -- like we said earlier, we are continuing to manage the portfolio consistent with how we have in the past. We are certainly glad in an environment like this to be maintaining low-risk exposures, low duration and low spread risk. This is the environment that we talk about and prepare for lots of time. So we feel like we're holding in fairly well.

We will now take our next question from Rick Shane.

Bill, those looks that everybody's giving you at the table are jealousy, congratulations.

Thank you, Rick. Appreciated.

That's a good one, Rick.

I have my moments. Tom, you talked about the alpha that you guys created in the quarter, and given the challenging quarter, it really stands out. And I think we understand a lot of it. I think the call protected or the prepayment protected Agency MBS did very well. Obviously some challenges on the MSR side. But when we look through the details, there are some significant gains -- these are realized gains on swaps that really makes sense to us. There's $97 million of realized gains on other derivatives. What is that? And just help us understand when you think about the alpha and the outperformance this quarter, what the puts and takes were?

Yes. Thanks for that question, Rick. Bill Greenberg is going to take that one.

Yes. Rick, good to talk to you. You mentioned some causes for our outperformance this quarter. And I would say, they're really threefold. One -- and besides the fact, one is the asset selection that we have are focused on being exposed to hard coupons is one. The existence of MSR in our portfolio, which reduces our overall mortgage spread risk, is another. And then lastly, as was alluded to on the call, our active hedging strategy and experience. And all 3 of those things really should contribute a lot. I think without going -- without diving too much into detail, so exactly what I was talking about, I would expect that those realized gains on those derivative result from realizing gains on options that we had. Typically, when we buy, say, out-of-the-money options in order to hedge for tail risks and the market rallies, those options become in the money or at the money. And rather than keeping them there with reduced convexity benefit, we will roll those options into ones that are further out of the money, and I expect that those gains result in that.

Got it. And is that why the notional on the swaps and the swaptions didn't seem to change much? It was that you took the swaptions and rolled them?

Yes. That's right. So things -- for revised swaptions at some -- again, forget about the portfolio effect, at some price market rallies, they go up in value. Their convexity increases for a little while and then it starts to decrease. And then in order to maintain the convexity benefit, the protection that we want or continued rate rallies, we basically will sell those and buy other ones oftentimes at equal notionals, and we'll reset the value of the convexity and the portfolio protection characteristics.

We will now take our next question from Matthew Howlett.

Bill, congrats. Just 2 questions. First, on the MSR. Bill, did I hear you correct that you said the UPBs sort of would start to pick up? The portfolio on UPBs would start to grow again in third quarter with the full arrangements? And then what would it take -- you said you're out of the bulk market. What would it take to get more active? Is that more just where prices come down too? Or is it just more of an outlook where you think speeds are going to ultimately end up?

Matt, thank you. Appreciate the good wishes. Yes, so a couple of things. First, we have seen slow volumes pick up, which is not surprising given that we start to see higher refinancings come in. Just to be clear, we are not out of the bulk market. We have bought a number of packages first quarter, as you know. And we did not buy any -- close any in the second quarter. But that was more a function of what was available and the prices and the collateral. So it was more sort of like saying why we didn't see any bonds we like at the prices that they were trading at. We do expect to see an increase in our bulk volumes through the remainder of the year. Obviously, there's a lot of volatility lately. So there might be a little bit of a pause from potential sellers until the market settles down a bit. But given that refi volumes are clearly picking up, we would expect the second half of the year eventually to see a pickup in bulk volumes, and we anticipate participating if the prices make sense to us.

Got it. Okay. And what explains sort of the lack of packages coming out in the second quarter? Was it just the sellers didn't like the prices or gains on some margins have improved that they don't need to sell MSRs?

Well, we don't originate loans and look at whether we should keep servicing or not. But I can tell you, first, volumes earlier in the year were quite substantial. One thing that we did see as the market rallied was we did see people who wanted to get a certain price, hold off because they thought, well, they wanted yesterday's price effectively. And so there were a lot of packages that didn't trade that might've traded. But if somebody makes them and maybe they don't hedge and keep the duration in line with what the market's doing, they might just say, "Well, I'll wait." So I think it's a combination of factors. As you know, once the market rates get to a certain point and stabilize and there is more visibility and clarity, what I'll call the bid offer, if you will, will tighten up, and you'll see a lot more transaction volume.

Got it. Okay. And by the way just one quick question, the FHLB advances. I know you took that down. I'm just curious on what -- why you decided to take that down.

Sure. So we had -- advances that were rolling off. We still maintained $50 million position with a much longer term. We value that relationship but we -- with the agency book, the advances and rates are not better than what we can get in the repo market.

Yes. We're always going to -- most economically is the short answer.

Right. The wholesale markets, they're better financed. And then last thing just on the -- I know you guys said low -- sort of low double-digit on the agency -- just the agency just putting aside the MSRs for a second. Is there sort of -- when we look at the spread, it was realized probably as 1.1% When you look at sort of 3Q, you have this dynamic where the fed cut and swap rates are going down again, the prepays look like they were up about 30% in July and the yields were lower. I mean is that -- can we expect that spread to be stable sort of everything taken together and going on. Any sort of color on where that spread could head in 3Q?

Sure. Yes, I mean some of the things you say about well, obviously prepays were off, et cetera, and there's a question of one of the yield on the cheapest to deliver, if you will. When we look at deploying capital, whether it's in credit side or on the agency side, we're basically serving across all coupons, all collateral types, all specified pool types, et cetera, and accessing what we think is the most attractive, that's sort of what we view our jobs ask to drive long-term returns. So as a result, when we cite those kind of returns, we're really looking at what we think is the best of what we can purchase, which is definitely not necessarily the cheapest to deliver on the TBA, which is arguably not going to be nearly as attractive as certain specified pools.

Right. Got it. So it's very just really asset selected -- selection literally at what it really comes down to as you look at the portfolio.

That's exactly correct.

That's the name of the game.

As there are no further questions, I will turn the call back to Ms. Field for concluding comments.

Thank you, and thank you for joining our conference call today. We have planned to participate in the Barclays Global Financial Services Conference on September 10, and our presentation will be webcast live on our website under the Events and Presentations link. We will look forward to speaking with you then. Have a wonderful day.

Ladies and gentlemen, this concludes today's conference call. Thank you for your participation. You may now disconnect.