Good morning, and welcome to the Annaly Capital Management Third Quarter 2020 Earnings Conference Call. Please note, this event is being recorded.

Good morning, and welcome to the Third Quarter 2020 Earnings Call for Annaly Capital Management, Inc. Any forward-looking statements made during today's call are subject to certain risks and uncertainties, including with respect to COVID-19 impact, which are outlined in the Risk Factors section in our most recent annual and quarterly SEC filings. Actual events and results may differ materially from these forward-looking statements. We encourage you to read the disclaimer in our earnings release in addition to our quarterly and annual filings. Additionally, the content of this conference call may contain time-sensitive information that is accurate only as of the date hereof. We do not undertake and specifically disclaim any obligation to update or revise this information. During this call, we may present both GAAP and non-GAAP financial measures. A reconciliation of GAAP to non-GAAP measures is included in our earnings release. As a reminder, Annaly routinely posts important information for investors on the company's website at www.annaly.com. Content referenced in today's call can be found in our third quarter 2020 investor presentation and third quarter 2020 financial supplement, both found under the presentation section of our website. Annaly intends to use their webpage as a means of disclosing material, nonpublic information for complying with the company's disclosure obligations under Regulation FD and to post and update investor presentations and similar materials on a regular basis. Annaly encourages investors, analysts, the media and other interested parties to monitor their company's website in addition to following Annaly's press releases, SEC filings, public conference calls, presentations, webcasts and other information it posts from time to time on its website. Please also note this event is being recorded. Participants on this morning's call include David Finkelstein, Chief Executive Officer and Chief Investment Officer; Serena Wolfe, Chief Financial Officer; Tim Coffey, Chief Credit Officer; Ilker Ertas, Head of Securitized Products; Mike Fania, Head of Residential Credit; and Michael Quinn, Head of Commercial Investments. And with that, I'll turn the call over to David.

Thank you, Purvi. Good morning, everyone, and thanks for joining us for our third quarter earnings call. I'll briefly provide an update on the broader market and how we manage our portfolios during the quarter. And before Serena reviews the financials, I'll discuss the factors that underscore our business principles and risk posture and have contributed to our performance. The U.S. economy rebounded at a faster pace than many had anticipated in the third quarter, best exemplified by the decline in the unemployment rate to 7.9% in September. Household spending rose sharply in certain sectors of the economy, arguably led by housing, and have made a substantial improvement with current activity well above pre-COVID levels. However, momentum is slowing as the service sector is unable to fully recover while the virus persists. The extended unemployment benefits supported under the CARES Act, which meaningfully boosted incomes, expired in July, although only half of all jobs lost during the pandemic have been restored. Monetary policy accommodation continues to be critical, especially given challenges on the fiscal front, which is further complicated by the uncertainty around the course of the virus as well as the outcome of next week's presidential election. Similar to the second quarter, the Federal Reserve continues to use all available tools to smooth market functioning while signaling they stand ready to provide more support if needed. Suppressed interest rate volatility remains a positive backdrop for our agency MBS and credit businesses, contributing to our ability to generate a 6.3% economic return during the quarter and core earnings in our dividend by $0.10. Additionally, we achieved these results while reducing our leverage to 6.2x. It is important to note that historic volatility we faced in March and a pending risk event in the election are reasons for maintaining a cautious approach. And while we would characterize the third quarter core earnings as a near-term peak, we do expect to outearn the current dividend again in the fourth quarter, all else equal. Now shifting to portfolio activity and beginning with our agency business, large-scale Fed purchases, now totaling over $700 billion, net of pay down since March, and strong nominal carry have offset high levels of supply and elevated prepayment speeds. While our agency portfolio was largely unchanged in notional terms, we further shifted out of higher coupon pools in favor of lower coupon TBAs. The current attractiveness of TBAs is driven by an improved deliverable resulting from the Fed taking out the most negatively convex pools and implied financing rates that are well below repo rates. And while roll specialness will not last in perpetuity, it contributes to excess returns and serves to mitigate episodes of spread widening, such that we anticipate maintaining our overweight in TBAs while the Fed sustains a heavy presence in the market. Our specified pool holdings do continue to provide considerable value as they offer meaningful call protection, have more accurate model durations and exhibit better supply and demand dynamics than generic pools. Prepayments are currently at their cycle highs, and we expect speeds will remain elevated given historically low mortgage rates, a strong housing market and improved uptake in technology, facilitating refinancing efficiencies. The investments we have made in specified pools over the last number of years are paying off in this environment. Our portfolio prepaid at just under 23 CPR in the quarter, more than 10 CPR slower than the GSE universe of 30-year fixed rate MBS, despite a weighted average loan rate on our portfolio that is 17 basis points higher than the universe. And with 96% of our Pool holdings characterized by some form of prepayment protection, we are well-positioned to withstand the current environment, and we anticipate our speeds will continue to outperform the broader market. Our outlook for the agency sector remains constructive as recent spread compression is supported by continued Fed support, attractive financing, strong nominal carry and subdued volatility. Our hedging activity was focused on protecting the portfolio from tail risk as we opportunistically added out-of-the-money swaptions at attractive pricing and also added modestly to our swaps and futures positions. Funding markets are healthy and liquid as capacity remains ample with abundant reserves in the system. Also, financing spreads for credit products, which were elevated through much of the summer, have recently contracted and we expect further improvement to come. Notwithstanding this liquidity in financing markets, we are cognizant of risks on the horizon and are, therefore, operating with conservative leverage and nearly $9 billion of unencumbered assets, the strongest liquidity position for the firm in years. Now shifting to residential credit, we maintain our positive outlook on the sector, considering strong fundamentals and technicals such as stabilizing delinquencies and favorable housing supply dynamics. Less than 6% of the overall housing market is now in forbearance, which is down from 9% in late May. The pace of improvement in housing has helped return markets to a more normal state and led to firmer asset spreads and a rebound in securitization volume. We completed 2 securitizations in the third quarter, totaling over $1 billion, which did reduce our overall capital allocation of the sector. However, this decline was partially offset by retained bonds on our new securitizations as well as opportunistic investments in NPL and RPL securities. The non-QM loan market also exhibited meaningful spread contraction since the end of the first quarter. Consequently, credit and convexity evaluation are as critical as ever in this environment, given less cushion in underlying yields. We believe the limited layer risk and delinquencies and convexity profiles of our loans demonstrate our proactive asset selection and focus on borrowers with significant equity. Our belief in this strategy is evidenced by our sizable retained interest in our securitizations, well beyond the mandated 5% risk retention in all securitizations at the time of issuance. Now although non-QM loan originations have been more limited since the first quarter, we are working to rebuild pipelines and source additional supply as non-QM originators reopen operations. In commercial real estate, while our portfolio was roughly flat quarter-over-quarter, we have seen a notable uptick in activity with loan and new issue security channels and are selectively evaluating new opportunities. In addition, underlying property sales and leasing activity is beginning to pick up as well. We remain focused on ensuring the health of our current portfolio through an active dialogue with sponsors to closely monitor underlying performance trends and we are comfortable with our current portfolio positioning. It is likely that parts of the commercial real estate landscape will be systemically changed by the pandemic, but it's still too early to judge the full extent. Despite declining rents and lower net operating margins, broadly characterizing the near-term CRE environment, cap rates should continue near current levels given record low interest rates, ample liquidity in the commercial real estate sector and the significant amount of capital raised in the sector pre-COVID, all of which could prove to be longer-term tailwinds. Turning to our middle market lending business. As we have spoken about extensively, we have taken a countercyclical approach over the past few years and the economic downturn we're experiencing validates this investment strategy. Our view is that our portfolio is in a sound position, given our financing of nondiscretionary companies, the vast majority of which have been deemed essential operators. Over diversification has been the downfall of many in late cycles, and we have opted instead for concentration among our top sponsors in certain niche industries, as evidenced by the fact that over 50% of our portfolio is predicated on government spending, with the majority of individual assets in excess of $60 million. Regarding portfolio fundamentals, underlying cash flow trends have been encouraging year-over-year as we've seen EBITDA and revenue growth that has consistently delevered the portfolio companies. Also to note, last quarter, we highlighted our conservative approach to reserves and CECL adjustments. To illustrate, our middle market lending reserves of $33 million this quarter are down from $51 million, which is roughly 1.5% of our $2.1 billion portfolio, as Serena will discuss in further detail. Now with respect to our outlook for capital allocation, I noted on our last earnings call that our allocation to agency would increase this quarter. However, with tighter agency spreads and our credit exposure at the lower end of the range, we are focused on responsibly increasing our credit portfolios. But as we are still early into this recovery, attractive opportunities are episodic. Now a further note on use of capital decisions about when to raise or buy back common stock are evaluated as part of our capital allocation framework, and similar to how we analyze investment opportunities across our businesses. It is a dynamic process, consider it of liquidity, time horizon and relative returns. We've repurchased over $200 million in stock throughout the past 6 months at times when our evaluation deemed at the most attractive use of capital. And we will consider using the buyback as a tool to generate shareholder return when appropriate to do so. Now an additional topic I want to cover today is related to how we manage our overall business and risk profile. And specifically, the 3 different forms of leverage we have available to enhance returns: First, at the corporate level is capital structure leverage, which includes preferred equity and unsecured corporate debt; second is structural leverage at the asset level, including subordinate securities retained through securitization; and third, traditional balance sheet leverage, primarily in the form of repurchase agreements. Now beginning with capital structure leverage, as of the end of the quarter, our preferred equity represents just under 15% of our long-term capital, which is in line with our average over the past few years. We've been very intentional with the construct of our capital structure by analyzing the relative attractiveness of the complementary financing and capital options. While capital structure leverage can enhance returns during periods of benign volatility, that benefit can quickly evaporate when not prudently managed. As the REIT sector witnessed earlier this year, capital structures can become overburdened which is exceedingly difficult to manage when access to common equity is constrained by depressed valuations. Now at the asset level, when we consider structural leverage relative to balance sheet leverage, we focus on synchronizing our financing with the liquidity of our investments. An example of this is our securitization platform, which matches term financing with less liquid residential whole loans. We have been very conservative with respect to repo leverage on top of structural leverage, as evidenced by the relatively low balance sheet leverage applied to our credit businesses, and importantly, certain assets within those businesses. Collectional leverage is a relative value analysis, and we have been very careful about not over-layering multiple forms of leverage, which is not always apparent in reported figures. And we believe this discipline was very evident in our relative performance this year as markets experienced unprecedented volatility. And finally, I want to touch on transparency and governance as more mortgage-related companies have publicly announced or completed transactions. We welcome a renewed focus on the sector since it will come with it an expanded investor base increased capital and greater disclosure. Mortgage investing is the core of our business, and we have a differentiated platform that has proven across all market environments. Importantly, we're a long-term model with an ownership mentality and accountability embedded in how we operate the business. Earlier this month, on the 23rd anniversary of our IPO, we published our inaugural Corporate Responsibility Report, which details our significant corporate governance enhancements, responsible investment, corporate philanthropy initiatives and human capital, diversity and inclusion commitments. It is available on our website, and I encourage everyone to read it. We published the report during a time of considerable global challenges. The COVID-19 pandemic, resulting economic and market upheaval and systemic social justice issues continue to have far-reaching impacts. Especially in times like these, we are guided by strong corporate values that enable us to successfully manage risks and take advantage of new business opportunities. Like others around us, we are using this unparalleled moment in history to lead with increased purpose and impact, both societal and economic, and we hope others will do the same. And with that, I'll hand it over to Serena.

Thank you, David, and good morning, everyone. I will provide brief financial highlights for the quarter ended September 30, 2020. And while our earnings release discloses both GAAP and non-GAAP core results, I will be focusing this morning primarily on our core results and related metrics all excluding PAA. As David mentioned earlier, the suppressed interest rate environment, continued Fed support, and our active portfolio management has enabled us to do more with less, less leverage that is, as we've generated the best core return per share we've had in 5 years with 6.2x leverage. Our book value per share was $8.70 for the third quarter, a 4% increase from the second quarter, and we generated core earnings per share, excluding PAA, of $0.32, a 19% increase from the prior quarter. Book value increased on GAAP net income of approximately $1 billion or $0.70 per share, which includes a $0.02 benefit related to reversals of CECL and specific reserves that will be addressed in more detail later in my comments, partially offset by common and preferred dividends and lower other comprehensive income of $253 million or $0.18 per share. GAAP net income increased in comparison to the second quarter due to gains on the swap portfolio of $107 million as well as higher GAAP net interest income of $447 million, up from $399 million in the prior quarter, primarily due to lower interest expense from lower average repo rates and balances, partially offset by lower net unrealized gains on instruments measured at fair value through earnings of $121 million, down from $255 million in the prior quarter. Since the disruption in March, we have recovered a large portion of the decline in the fair value of our portfolio. We have included a slide in our investor presentation that provides additional color and detail on the assumptions utilized in the evaluation of our CECL reserves. In the current quarter, we have seen a general improvement in market sentiment and the economic models we use in this process. The key macroeconomic assumptions used in determining our reserves being GDP, unemployment and commercial real estate values are still projecting unfavorable assumptions comparative to prior years. However, they have improved in comparison to our Q2 assumptions used. We recorded a modest reversal of reserves associated with our credit businesses of $22 million on funded commitments during the third quarter. Driven by portfolio activity, primarily loan payoffs and the aforementioned improvement in macro assumptions. And now total reserves net of charge-offs comprised 4.56% of our ACREG and MML portfolios as of September 30, 2020, versus 5.32% as of the prior quarter end. As always, we do not simply book what the model tells us we should reserve. And therefore, in conjunction with our investment teams, we spend time analyzing the results of the reserve calculation and ensuring our expectations align with the quality of the portfolio and the performance of the borrowers, which supports the release of reserves. Consistent with prior quarters, we continue to think it critical in the current environment to err on the side of conservatism, given the significant uncertainty around economic and market value scenarios. We remain comfortable with our existing credit portfolios and the associated CECL reserves, and we'll continue to monitor specific asset performance and economic projections as we determine future reserves. Turning to earnings. The largest factors quarter-over-quarter to core earnings were a lower interest expense of $115 million compared to $186 million in the prior quarter, primarily due to lower average repo rates, lower compensation and G&A expenses as a result of cost savings from the company's internalization, as well as record high TBA dollar roll income of $113 million in comparison to $96 million in the prior quarter, due to higher average balances as we took advantage of specialness in the TBA market and increased concentration in dollar rolls compared to pools. These increases are partially offset by lower coupon income on agency MBS due to lower average balances. Recorded benefit from a reduction in financing costs with lower average repo rates down to 44 basis points from 79 basis points in the prior quarter. Combined with lower average balances, down to $67.4 billion from $68.5 billion in the prior quarter, and weighted average days to maturity relatively consistent with the prior quarter at 72 days versus 74. Our all-in average cost of funds this quarter was 93 basis points versus 1.29% in the prior quarter. While we continue to benefit from the 0 interest rate policy, our treasury department has a keen eye on the market and has renegotiated all debt agreements possible to take advantage of the optimal financing environment. We have opportunistically turned out repo where it has made sense. However, with ample reserves in the system and deliberate forward guidance on the part of the Fed, we're of the opinion that repo rates will be lower for longer. Therefore are striking the appropriate balance between shorter-term funding and modestly more expensive longer-term repo agreements. We reduced our leverage, and our economic leverage declined to 6.2x from 6.4x quarter-over-quarter, primarily due to lower repo balances of $2.5 billion, higher equity balances of $397 million, due to favorable mark-to-market changes for swaps and instruments measured at fair value through earnings, realized gains on investments and derivatives, and lower net payables for investments purchased of $254 million, partially offset by higher TBAs of $1.9 billion. In addition, we reduced the balances on our FHLB facility in preparation for the securitizations we closed during the quarter and the upcoming sunset of our FHLB membership. As David has mentioned before, you cannot unsee what we went through in March. And therefore, we are proud of the strong results for the quarter, achieved with prudent management of our leverage given lower risk tolerance levels. The portfolio generated 205 basis points of NIM, up from the second quarter of 188 basis points, driven primarily by the decrease in cost of funds that I mentioned a moment ago. Annualized core return on average equity, excluding PAA, was 13.79% for the quarter in comparison to 12.82% for the second quarter. Our core return on average equity, excluding PAA, per unit of leverage improved to 2.22% in Q3 versus 2% in the second quarter. Again, illustrating the notion of doing more with less. We have started to see the benefits of our internalization with improvements to our efficiency metrics in the third quarter, being 1.32% of equity for the third quarter in comparison to 2.01% in the second quarter of 2020. And 1.73% year-to-date, which is within the range of expected cost savings disclosed in Q1 with our internalization transaction announcement, even with the reduced equity base as a result of the March disruption and our share buyback program. Annaly ended the quarter with an excellent liquidity profile with $8.8 billion of unencumbered assets, an increase of almost $1 billion from the prior quarter, including cash and unencumbered agency MBS of $6.9 billion. After Labor Day, we began a hybrid working arrangement with employees voluntarily working from our Manhattan office periodically. While we recognize the benefits to communication and collaboration from being physically together a few days a week, we continue to work seamlessly outside of the office, and we'll continue to monitor state and city data in determining future return to office plans. And that concludes our prepared remarks, and I'll turn it over to the operator for questions.

The first question today comes from Steve Delaney of JMP Securities.

A lot of good numbers in this report, but I'd like to just comment that I think from my view, possibly the best number is the 70 basis point drop in expenses to average equity. So we know where that came from. And congratulations on the internalization. Your dividend coverage, obviously, was excellent, strong quarter and 145%. David, we heard your comments that this 3Q may be a little peaky in terms of the near-term cycle. But Serena, I'm wondering if you have available the information to comment on any undistributed taxable income that Annaly may have had as of September 30.

So there are significant differences regarding core versus taxable income. One of the largest differences is the amortization of swap losses, which we spread over the term of the original hedging instrument, while for GAAP, we account for it all at once and exclude it for core. As of the third quarter, we have no concerns about meeting our redistribution requirements for the full year.

Understood. It seems there won't be any tax-driven requirement to increase the dividend or pay a special dividend; that decision will depend on other factors rather than being mandated by taxes.

That's the correct assumption, Steve.

Sure, Steve, it's great to hear from you. Regarding the loans and securitizations, these originated from our expanded prime program and are mostly composed of loans that are difficult to decipher. Looking at the overall picture of the securitizations we've completed, which total more than $5 billion, we have retained approximately 18% of our total securitization.

Excellent. And those are nice bonds and I think have the possibility to see ratings upgrades, right, as those loans season? And you'd be...

Yes, definitely. To my earlier point, Steve, regarding balance sheet leverage, we apply very minimal leverage to these assets. The return on our last deal was approximately 12.5% on the balance sheet.

The next question comes from Rick Shane of JPMorgan.

Look, in a normal environment, you guys faced binary risk in terms of rate direction. But we're in a period now where the distribution of outcomes is more skewed than it's probably ever been. The other risk you normally face is determining the timing of those rate moves. And the Fed is creating an environment through their communications that deliberately diminishes that uncertainty. I think that this creates a setup where there's less uncertainty on a theoretical basis than you guys have ever faced. I'm curious how you balance being able to maximize profitability in that environment with also managing risk. And there was an interesting comment from storing that related to one place where you're enhancing profitability by maintaining exposure to the shorter end of the repo financing curve. Can you talk about how you look at this going forward?

Sure. So to your point, the rate risk is skewed. It's asymmetric. We don't expect the Fed to go to negative rates, and we're at the 0 lower bound for the next number of years, and they've been very deliberate with their forward guidance, which has been encouraging. But that being said, as I mentioned in my prepared remarks, there is always that risk that we could see rates out the curve increase, and there's a lot of uncertainty around that. So our hedge profile is set up as it has been for the last number of quarters with a modest steepening profile because there could be events that do lead to higher rates out the curve. We're very comfortable at the front end of what the Fed has conveyed to us. But you could see volatility at some point out the curve. And so that's why we position hedge profile to reflect that. And that also is why we've added more swaptions to the hedge portfolio to help us mitigate against that tail risk. Now with respect to financing and our average days, which I believe was 72 days at the end of the quarter. We do look at our financing profile relative to the cost of hedging short rates with other instruments. And we are very comfortable with the level of financing rates out the near-term horizon. The curve is relatively flat from one month to a year, we understand that. But there is an incremental cost of turning everything out to a year when we don't think it's necessary to do so and we can manage those risks and hedge those risks more cheaply with other interest rate instruments. And we do have a fair amount of short-dated swaps that do protect for that.

Got it. So does that strategy make sense because the cost of swaps is low in a mobile environment? Does that allow you to take a bit more risk on the financing? I don't want to overemphasize the risk because it makes sense, but is that how you view it?

Sure. There were even periods last quarter where we paid fixed at negative rates. And so it's attractive to do so as opposed to terming out repo from that standpoint. So it's pure relative value.

Okay. And I do have one last question. I apologize to my peers who are waiting in queue. Just given the volatility we've seen in the equity markets headed into the election, can you comment on the volatility that you've seen during that same period in your core asset classes as well?

Sure, sure. And look, we do have this risk event next week. I think heading into the latter part of last week and this week, there were expectations of stimulus and the potential for a democratic sweep. So we saw an increase in rates, predicated on the possibility for stimulus. That seems to have subsided, given the lack of progress on the stimulus front. But I will say there are a variety of outcomes that could occur with respect to the election and fiscal policy that does have us relatively cautious. That's why we're hedged well, and that's why our leverage is at the low end of the range over the last number of years. And so we think this conservative approach is appropriate. And it also does happen to more than cover the dividend. And so we're pretty comfortable with our operating environment. We have a lot of latitude, particularly with low leverage right now.

The next question comes from Kenneth Lee of RBC Capital Markets.

Appreciate the discussion of the different types of leverage. Wondering if you could just further expand upon it, perhaps share your appetite for potentially increasing leverage within each of the buckets just given the current attractive environment for investing?

Sure. I'll go through the various asset classes and what we find appealing in terms of leverage. Starting at the corporate level, I mentioned that it's not beneficial to increase capital structure leverage due to the costs associated with preferred equity, so that's not something we're considering. Regarding balance sheet leverage versus structural leverage, balance sheet leverage is currently quite affordable, thanks to the Federal Reserve's policies, and this primarily influences our agency business. Thus, our focus in the agency space is on balance sheet leverage. In contrast, structural leverage is more appealing on the credit side, as evidenced by our securitization returns. We apply more structural leverage in our credit portfolios, where our balance sheet leverage is approximately 1.5 times. We believe we've found the right balance by not combining structural leverage with balance sheet leverage, and we're satisfied with that. When it comes to increasing leverage, we would need assets to become cheaper to consider that. We're quite comfortable with our current position. A leverage ratio of 6.2 times gives us significant flexibility to adapt as the market changes. If asset prices decline, we might increase leverage naturally through equity degradation. However, we have the capacity to acquire assets at lower prices if we believe the decline is temporary. While we don't anticipate asset prices dropping given the Fed's situation, if that were to happen, we could take the opportunity to buy assets. On the other hand, if there’s ongoing spread tightening that decreases our leverage due to appreciation in book value, we wouldn't necessarily acquire more assets to keep leverage stable. We might even consider the possibility that assets are overvalued and reduce leverage further by selling assets. Overall, we have considerable flexibility to make decisions based on market conditions, asset prices, and our liquidity and leverage status.

Got it. That was very helpful information. I have a quick follow-up regarding equity capital allocation. It seems that part of the increase towards agencies was a result of the securitization transaction. I would like to know your updated thoughts on how that equity capital allocation might trend in the near term.

Yes. Sure. And as I said in the prepared comments, we would like to tick up our allocation of credit. But there's still a lot of uncertainty. We don't know the exact course of the virus. We don't know how policy is going to shape up on the fiscal front. We are still learning about how this economy is evolving. So we have to be very methodical about how we add credit assets. Our pipelines are filling up. We're seeing very good opportunities. But again, it is episodic. And so we'll see how it plays out. And we'd like to get a lot more certainty with how the economy shapes up before we make a stronger commitment. But at the margin, I would say we would like to get our allocation and credit marginally higher.

The next question comes from Doug Harter of Crédit Suisse.

Can you talk about kind of how book value has performed in October so far?

Sure, sure. We're up a few pennies quarter to date. Nothing to write home about, and we have a long way to go left in the quarter.

Understood. Can you discuss how the events of next week might affect your actions compared to the other uncertainties present in the market?

Yes. Well, unfortunately, it's likely the case that the passage of next week may not lead to a resolution in the outcome of the election. And so that's something we're certainly concerned about. But there's different states of the world based on how things evolve. If you get a blue wave, for example, then you could see a continuation of that slight sell-off we saw last week, and before, with the optimism around stimulus. If you get a split Congress, if the President retains the White House and you have a Democratic Congress, that could lead to a moderate stimulus. And if it goes the other way, you potentially, and this is something that concerns us, could see a repeat of what we saw in 2010, where you had austerity coming from Congress from the Senate and the White House in need of stimulus. So there's a variety of circumstances that could prevail. And again, it does come back to taking a cautious approach as we do figure it all out.

The next question comes from Bose George of KBW.

Can you talk about the level of specialness in October versus the last quarter? And then just how do you think about the level of your net TBA position as a percentage of your total MBS exposure? Could we see that increase if returns remain elevated? Just curious how you're thinking about sizing that?

Sure. No, that's a very good question, Bose. So specialness in October is still quite prevalent. It has moved down in coupon into 1.5s and 2s. But generally speaking, the financing rates, implied financing rates on those coupons are negative 60 to negative 80 basis points depending on underlying assumptions. And that compares almost on par to what we experienced in the third quarter. We generally, on average, rolled our portfolio at south of negative 50 basis points on the quarter. As the market does repopulate with new collateral, you should see specialness dissipate somewhat as time passes and the Fed doesn't take all of the lower coupons out of the market, but it should be here for a little while. Now in terms of how we think about it, this is a carry-driven market. In the third quarter, we had about $4.8 billion in portfolio runoff in the agency portfolio. About half of that was allocated to TBAs to grow the TBA position, with the other half, more early on allocated to pools. Now we're at a point where specified pools have gotten to a point where they're pretty fully valued. And so the way we're thinking about it currently is runoff in the fourth quarter will more than, on average, go into TBAs. But we are certainly considerate of what the portfolio looks like, and we have to manage the portfolio, not just for near-term carry, but we have to think about a variety of scenarios as to how markets can evolve. And our specified pool portfolio is the core of our business. And those cash flows, long-term cash flows that we put on the balance sheet are going to remain the core. So to make a long story short, we'll probably grow the TBA portfolio modestly in the fourth quarter. But we do remain focused on our pool portfolio, and we think it's performed quite well, as I talked about in the prepared comments with respect to portfolio speeds.

Can you provide the incremental difference in the TBA returns compared to the spec pools?

Sure. I would say pools are in the high single digits and the very low double digits in TBAs.

Okay. Great. And actually, just one clarification. The cost of funds that you guys reported, the 93 basis points, does that include the implied funding benefit of the TBAs? Is that's just a balancing number?

No. No, that's a subjective assessment because you'd have to incorporate a prepayment assumption. So when we give you our cost of funds, it's entirely objective data.

The next question comes from Brock Vandervliet of UBS.

David, on the CPR speeds, you remarked, we are kind of in an unusual environment here with low rates for looking like a very long time. But seasonally, there should be a bit of a decel in activity. Help us kind of get our head around the forward look on CPR. And also with respect to the available pool of mortgages that pencil for refinancing? And where that is now versus a couple of months ago?

Sure. Let's start with the pool of refinanceable mortgages. Currently, about 80% of the total is 50 basis points or more in the money. This figure has slightly decreased due to heavy market issuance and lower loan rates. However, the majority remains advantageous for refinancing. Regarding speed predictions for the near term, we anticipate a modest increase in speeds during the fourth quarter, estimating it will be around 5% to 7% faster. Concerning seasonality, it's important to recognize that originators are operating at full capacity and expanding. Even if we experience lower housing turnover in the winter, which contributes to seasonality, we expect resources to shift more towards refinancing. Thus, we don't anticipate seasonality will lead to a significant decrease in prepayments in the first quarter, which is generally when the decline is most observed. If anything, we expect it to remain flat or increase.

Okay. Understood. Regarding funding costs, there was a significant decrease this quarter. How much more improvement do you think is possible, or is this the extent of it?

Well, look, I think when we look at repo rates on the agency side from overnight out to a year, we're talking 13 basis points so thereabouts out to an upwards of 30. And so as the portfolio does roll, we will be resetting the agency portfolio at lower rate. As of today, our average rate is 36 basis points, so where our current repo rates are today. So that's a decline to some extent, and we'll see is more repo rolls.

The next question comes from Mark DeVries of Barclays.

You mentioned looking to start to reallocate back towards credit investments, but the opportunities you're seeing are episodic here. Could you just talk about where you're seeing the most attractive relative value and where we should expect you to kind of reallocate capital?

Sure. So we can go around the room here. We do have the full room of all our business heads. And again, opportunities are episodic. We have started to rebuild our residential whole loan pipeline, and we're seeing our other businesses pick up in activity in terms of at least looking at a lot of opportunities. But again, we do have to take the backdrop of what the uncertainty entails with respect to the economy and how credit could evolve. So we're certainly cautious, but let's just start with Mike Fania on the resi side.

Sure. Thanks, David. In the residential credit sector, we believe non-QM securitizations are yielding leveraged IRRs in the low to mid-double digits. A significant factor is the limited availability of supply, which has been relatively new. Agency pipelines are full and have high gross margins, and there was also a reduction in programs starting in April. We are optimistic about this sector and are working to strengthen our originator network and strategic partnerships. Additionally, we find the unrated NPL or RPL security sector appealing. Currently, senior securities yield unlevered rates of 3.125% to 3.25%, and senior support securities yield 5.25% to 5.50%. We expect leveraged IRRs in this area to be in the high single digits for senior securities and between low to mid-double digits for senior support.

Mark, this is Mike Quinn. So on the commercial side, I would say, the most attractive things that we're seeing today are in the transitional whole loan space. And I think that's for a couple of reasons: One, obviously, we're being cautious just given the uncertainty going forward. But assets have repriced. And I think we're seeing opportunities where we're able to lend to borrowers at lower levels today than we saw, call it, 12 months ago. In addition, spreads are quite a bit wider today than they were. I think all-in coupons are probably pretty similar, maybe a little bit higher, but the credit quality of those loans is quite a bit better. So we are seeing opportunities to make new investments on the whole loan side.

Mark, it's Tim Coffey. I think on the middle market side, we will just continue the pattern of what we've been able to accomplish to date. The portfolio remains very clean. So we expect to see continued tack on acquisition activity on our existing book. You'll see it in the number of issuers that we've been able to maintain, if it continues to be at 50 on the head. And so I think the benefit of having a clean portfolio is that those names have the ability to go out and execute with tack-on acquisition activity, thus resulting in financing needs to which we can obviously leverage our past success. So the bulk will be coming from what we've done today. And continue to increase our exposures on our existing stable of names.

And Mark, we do also have Ilker Ertas here on the agency side, but it looks as though I've already stolen all of this thunder in prior comments, I think, where we stand on that front.

Okay. Great. That's very helpful. Could you just provide us an update on kind of the efforts to manage weather retroactively on existing portfolio perspective and the new underwriting kind of COVID exposure in the commercial and middle market lending?

Sure. Mike can start...

Yes. So Mark, this is Mike Quinn again. I think the COVID exposure in the commercial portfolio hits us most in two places: One is the hotel space and two is in retail. So first, on the hotel side, we've really got 5 positions in the hotel space. And in every one of those positions, sponsors are contributing capital to those deals to support those assets. And 4 out of our 5 hotel positions are currently paying. In the retail side, I think we've been hit as well. I think similar themes in that space in that the positions we have, sponsors are supporting those deals and infusing capital. I think in some instances, those loans will need to be extended in order to get through this kind of low liquidity period. But we like our basis in those assets, and we feel comfortable with the outcomes going forward.

Mark, on the MML side, it's quite evident that we've ironically benefitted significantly from the impact of COVID in our portfolio. While we don't like to emphasize that too much, our portfolio has effectively taken advantage of the past six or seven months. In contrast to many of our peers, our portfolio has experienced a fundamental improvement across individual holdings. Additionally, while some select names have faced challenges over these months, I've observed that in the last couple of months, those initially affected by COVID have made considerable progress. They may have faced early challenges, but have seen a positive turnaround recently. Overall, we're optimistic about the current state of our portfolio following the recent market volatility.

This concludes our question-and-answer session. I would like to turn the conference back over to David Finkelstein for any closing remarks.

Thank you, Elissa, and thanks to everybody for joining us. We hope everybody is staying healthy in these times. And in line with my earlier commentary around purpose and impact in this environment, irrespective of your political affiliation, just want to conclude by reminding everyone to get out and vote next week, if you haven't already. And we'll talk to you soon. Thanks.

The conference has now concluded. Thank you for attending today's presentation. You may now disconnect.