MFA Financial, Inc. (MFA)
Morgan Stanley Financials Conference
June 12, 2012 12:05 PM EST
Craig L. Knutson – Executive Vice President
Goodmunder Christiansen – Executive Vice President
Hi, good afternoon. I’m Cheryl Pate, the Specialty Finance Analyst here at Morgan Stanley and here to introduce MFA Financial. MFA Financial is a hybrid Mortgage REIT with a long track record. It was formed back in 1998 and is one of the few mortgage REITs operating today that has been in business through an entire interest rate cycle. The company has a strong track record in evaluating non-agency MBS and we look for the company to experience some expansion as some of the high cost flats roll off the portfolio over the next couple of years. As an internally managed REIT, MFA has among the lowest operating costs in our coverage group. And with us here today to present is Craig Knutson, Executive Vice President responsible for MFA’s non-agency portfolio and Goodmunder Christiansen, EVP that worked on the Agency side of the portfolio. So we have both sides of the portfolio coverage here which should make for interesting discussion.
So I’ll turn it over to the company and I’ll kick off the Q&A session once the formal presentation has concluded.
Craig L. Knutson
Thanks Cheryl. So as Cheryl pointed out, MFA is an internally managed mortgage REIT and we do invest in both agency and non-agency RMBS. Investment opportunities continue to exist in both the Agency and non-agency sector. The Agency MBS investments that we make continue to benefit from the steep yield curve, and despite repeated government actions and policies, we are not seeing prepayment rates spike as one might expect. And we reiterate this; our goal is really to generate low double-digit returns on equity with the appropriate amount of leverage. So have we done that over time?
So since January of 2000 we’ve put up a 15.2% annual return, that’s assuming reinvestment of dividends. So we are pretty proud of that.
So, this is what we call our asset allocation table and this is how we look at the company. So yes, we are invested in Agency and Non-Agency RMBS and so you could look across our entire company and look at our balance sheet and say it looks like you are about three-and-a half times levered. But the reality is we think it makes more sense to split out the two businesses solely for the purposes of equity allocation and leverage. So you can see the first column here is Agency MBS, and then you can see that debt to equity ratio is 6.87. So again, probably fairly typically on our space, six to seven or six to eight times leverage is what you typically see on Agencies. On the Non-Agency side, you can see that our debt to equity ratio is a little less than two times. So again, three-and-a half maybe is the whole company, but – and so that would be people to say, you are lower leveraged than most other REITS. We really think that you have to split out the two different components.
But if you look across the bottom at the yield on average earning assets, and this is for the first quarter of this year, you can see that in Agencies, our average yield on interest earning assets was 3.15, which is actually pretty high and is extraordinarily high if you consider the fact that we don’t own any 30-year of fixed rate mortgages.
The primary reason for that is that many of those are older securities that we bought some time ago that have generated higher yields. However, the next line that average cost of funds on our Agency you see is 1.71, that’s actually very high and the primary reason for that is again, we have a legacy book and we have a legacy book of interest rate flops and many of those interest rate flops that are at very high cost rates will be rolling off.
So somewhat uniquely we think that our cost of funds on our agency portfolio will decrease over the next twelve months or so, which is somewhat unusual on our space. Now, obviously at the same time, to the expense that we replace run off on the Agencies, we are not buying new securities that yield 3.15. So we think that while the assets that we add will obviously come on at lower yield, the funding cost somewhat uniquely will decrease for us.
On the Non-Agency side, you can see that the yield in the first quarter on the Non-Agency portfolio was 6.92%, so close to 7% yield. Our average cost of funds there a little over 2% for a net interest rate spread of about 4.75%.
So on the Agency side, we have a balanced portfolio of hybrid ARMs, it were approximately 75% hybrid ARMs and about 25%, 15 year fixed. Our overall premium exposure is quite low, our average amortized cost of our agency portfolio is 102.8%, and we believe that we have limited exposure to HARP 2.0, or HARP 3.0 or 4.0 or whatever we see there. Most of the hybrid securities that we own that are HARP eligible, so that’s 3 June of 2009 are interest only. So the underlying mortgages are not amortizing principal. So the purpose really of HARP is to push these forward into a 30-year fully amortizing mortgages, and because of those 30-year mortgages are fully amortizing, in most cases the payment is actually -- even though the rate is lower, the payment is actually higher than the payment that these borrowers are making now on these interest only loans. So again, that is our observation that the payment that the borrower makes is typically more important than the interest rate associated with the mortgage.
And then finally I alluded to this a minute ago, our agency funding costs are expected to decline. We have a little close to $900 million of interest rate swap, with an average fixed pay rate of 3.79% that will roll off in the next year or so. So we think that will help to decrease that funding cost on Agencies.
We’ve also been increasing our Non-Agency portfolio. So we own approximately $4.5 billion market value of Non-Agency securities with an average amortized cost of 73% of par and again, as you saw in the prior slide in the first quarter these assets generated a loss adjusted yield, and I’ll get into what we mean by loss adjusted, a loss adjusted yield of close to 7%. And we believe that these assets are much less sensitive to changes in the yield curves and interest rates because the return on these may very well increase if prepayments increase because the purchases is such a large discount. But in addition, on an asset that trades at a 7% yield in a world where the 10 year yield is 1.5%, we think that the vast majority of that yield is really due to credit and not the interest rate sensitivity. So what happens if we see higher rates? We’ll if you see higher rates we think that likely the economy is probably doing better and that’s why we are seeing higher rates and if that’s the case, then the credit component of these securities will probably dominate their performance. And we have some historical context for this, not in Residential Mortgage back Securities that traded at $0.77 on the dollar because that never happened before a few years ago. But if you look at junk bonds over time, you’ll see that the junk bonds tend to be more correlated with the stock market and inversely to interest rates and it’s again because of this credit component that a large portion of that yield is really due to credit rather than interest rate sensitivity.
So we think in the overall portfolio that the two asset classes are very complementary. Obviously we have interest and prepayment risk on the Agency side and while we have credit risk on the Non-Agency side, we think we mitigate much of that interest rate risk.
So this is a busy slide, but I will point to couple of the red circle numbers at the bottom. These are 20 largest Non-Agency positions. So couple of things that I would point out. One, it’s pretty diversified, there’s only one position that’s more than 2% of the whole portfolio. But if you look down at the bottom it’s a cycle score so this is the weighted average cycle score when these pools were originated. The weighted average cycle there is 733. So that tells you something about the nature of these mortgages. These are typically prime type of borrowers. If this were subprime that – that average cycle score would be more than 100 points lower. So 733 indicates primarily prime and near prime assets.
The next column, which is not circled, is also quite important. The WALA, that’s the Weighted Average Loan Age. So the Weighted Average Loan Age here is 66 months. So we have 5.5 months of seasoning on these loans. So when we go – when we do our credit work, the challenge in the credit work is to estimate how many of the loans are current today will default in the future. And the second we have five to six years of pay history is very revealing. So several years ago people would say this particular bond has a lot of no dock or low dock loans. And surely that was worrisome in the early stages of our mortgage. But if you have five and six years of pay history, we think that, that’s a lot more significant than how sick the file was six years ago, right? It’s really the fact that the borrowers, they are making payments and in some cases, may have never missed the payments in the last six years. So we think that that’s pretty powerful and that pay history is a very good indicator.
On this portfolio, that’s five and a half years seasoned, we have approximately 19% of the underlying mortgages are 60 plus days delinquent. And I’ll get into our last subject in a minute, but the 6.92% loss adjusted yield across these particular 20 bonds is assuming that 39% of the underlying mortgage is default. So not only the 19% that are delinquent but again, an equal amount after that of loans that are current – again, on which we have 5.5 years of pay history. So people look at that and say that’s pretty bad, you think that 40% of the loans are going to default and I guess their response is that well, I guess that’s pretty bad but those are the numbers that are baked into our yield assumptions. So if you believe that maybe it’s not that bad, the implication is that the actual yields that were recognized will be higher.
So this is to try to illustrate what we mean by a loss adjusted yield. So if you go back to 2005 let’s say and you look at a non-agency security and granted they didn’t trade a dollar price of $72 then, then they all traded at par. But if they had traded at dollar price of $72, with a 4.5% coupon and of course since these all were borne as triple As in 2005, everybody knows that you will never lose any money on a triple A rated security. So the default in some very assumptions are zero and typically non-agency is prepaid pretty quickly. So you probably would have used the 15% CPR assumption. That being the case at a dollar price of $72 you would have expected a yield of 14%.
So what changed? Well, the first thing that changed is these securities aren’t paying at anywhere near close to 15% CPR with a cost basis of 73%, we only wish they were paying at 15% CPR. So if you slow the speed down to 6% all of equal that 14% yield drops to 10% and then finally as we’ve said, we have significant default and loss severity assumptions that are also built into these yields. So if we default 44% of the underlying loans with a 50% loss severity, that 10% yield now goes down to 7%. So we say its 7% loss adjusted yield. That’s what we mean by that.
And while Housing Fundamentals certainly are still weak, we believe that we have appropriately factored this into our credit reserve estimate. Most of these securities still have credit enhancements underlying them. So when they first came out, they were all top of the stack, triple A rated, they had subordinate pieces beneath them. We still have across the portfolio about 4.5% of structural credit enhances. So these are tranches junior to these most senior tranches that will absorb losses before they hit the senior bonds.
The second thing that protects us is a discounted purchase price. So on average because our amortized cost is 73% of par, we could lose 27% of the face amount of these and still get back what we paid for them. And finally, we have a significant credit reserve. So we have – I have another slide on this coming up but we have about a $6 billion face amount portfolio and close to a $1.5 billion credit reserve.
So that -- you can see on this slide here. So over on the left, the current phase of the Non-Agency portfolio is $6 billion or purchase price of $4.4 billion or about $0.73 on the dollar. So the question is what happens to the discount, the purchase discount – the difference between par and 73% which is about 27 points. Well, if you look over on the right, you’ll see that the vast majority of that purchase discount, we’ve set aside as a credit reserve. So we are not accruing that into income. The income on non-agencies is the coupon income and then it’s at a very small blue portion at the bottom, only about 5% of our discount we are amortizing into income.
So the other way to look at that is that our average purchase price is about 73% of par and if we have a 22% credit reserve, then essentially we are saying we’ll paid 73% for these bonds and we expect to receive back approximately $0.78 on the dollar.
Then finally there’s the liability side of the house and we’ve actually made a lot of progress on this front in the last six to eight months or so and we can say that less than one third of our non agency mortgage back securities are funded with short term repo. So if you look down here you can see that the market value is $4.5 billion in that first column there. Repo less than 12 months is only $1.2 billion or 27% of the total funding. We have repo $200 million or so that’s 12 months and longer. We have securitized debt. These are re-securitizations. We’ve done four re-securitizations and when we do a re-securitization we create new triple A securities which we sell to third party investors. So that’s really the best type of financing that we get because it’s permanent financing. There’s no mark-to-market. There’s no margin call. It’s non recourse to the company. So that’s about $1 billion and then in addition we’ve done $500 million, a little bit more than that actually of multi year, this is three year collateralized financing.
So we think we’ve really shored up the capital structure around the non agency securities and you can see obviously the longer term financing is more expensive than the short term financing. So this isn’t really an interest rate play the reason that we’re putting these longer term facilities in place. It’s to insure against rollover risk. So we know this portfolio was financed for a long time. These multiyear facilities also give us the ability to substitute securities. So for instance if we’re accumulating securities to maybe do a re-securitization, we could put those securities on that line when we accumulate sufficient securities to do it we can substitute other collaterals. There are certain bonds that may not re-securitize well for a variety of reasons. So this gives us another alternative for longer term financing.
So again just to review MFA. We’re internally managed. So and I think that’s increasingly among the minority, but as an internally managed REIT, we have no incentive to issue equity other than it makes sense to issue equity. So I think our interests are aligned with shareholders. So we don’t get paid more if we issue more equity. We get paid in ROE so I think we care about the same thing that you care about. Our goal is to generate double digit returns on equity and so far, at least since the year 2000 I think we’ve done a pretty good job at doing that. While housing fundamentals – we’re certainly not waving a checkered flag on housing fundamentals. We think we’ve appropriately factored those projections into our cash flow projections in our current reserve estimates and I will say versus standing here a year ago that a year ago there really were no bright spots anywhere in terms of the housing market and housing data and again we’re not waving a checkered flag.
But that data is at least a little bit mixed right now. You do here about some pockets where – Phoenix for instance where housing prices they’re up and you read about bidding wars. So we think again we still have a long way to go, but for the first time in the last three years we’re starting to see at least some positives on the housing front. And finally we’ve diversified our funding sources, particularly for non agencies which includes securitized debt, structured financing and equity in addition to repo.
And with that I’ll open it up to questions.
So maybe I’ll just kick it off. Maybe we can talk about the best case view you have in terms of the macroeconomic, the likelihood of QE3 and how that impacts your thinking on both the agency and the non agency portfolios in that base case view.
Craig L. Knutson
Sure. So in terms of probability and likelihood, if you look at market and the prices on the side and the research from the Wall Street firms are saying that if people are anticipating probably 50% to 80% chance of QE something, whether that is additional purchases or growth of the balance sheet or if it’s just extension of operation twist, which could take any form of just finding longer dated treasures and selling shorter dated treasures, or selling shorter dated treasures and buying into more securities. Our own thought process on this is we start by looking at how do we want to construct our portfolio from an interest rate point of view and interest rate risk and the overall strategy for the company as a whole has been to try to minimize interest rate risk. That’s where the angel portfolio is and why we emphasize high arms and 15 year fixed. But it also comes into play with the non agencies because they have very limited interest risk as well. So the company as a whole has limited interest risk .
But in terms of QE and kind of big picture, the fed already owns about between 15% and 20% of the agency mortgage market and on the treasury side they own about 15% to 20%. So the bigger question is also how big do they want to be in the marketplace and how big can they be until they disrupt market flows and liquidity? Our view has been that we don’t want to be in that position where we’re trying to front run the fed, just try to pick up those securities that they will hopefully buy and then we’ll gain from that. We want to take more of a big picture approach where we think about interest rate risk over the long term, not over the next quarter or two.
Craig, can you talk a bit about what’s going on in non agency where non pricing supply where you’re seeing the access come from.
Craig L. Knutson
Sure. So a year ago we had some pretty good price weakness in non agencies, the second quarter and certainly the third and fourth quarter and it was all precipitated by several things. But one of them was certainly Europe. Another was fear of impending supply. So the feeling was a year ago that the European banks were going to have to sell large holdings of non agencies. You had the early maiden lane transactions that went well at first and then really did not go well. And there was also the thinking that we were going to see poor supply from domestic banks that would be forced to sell these securities because of risk weighting on assets they were getting downgraded.
So we still have Europe. We didn’t see the large selling that we got out of Europe, but certainly they still own that paper and the domestic banks obviously still own this paper. So to show what’s different this year because we really haven’t seen prices weaken over the last month or so when you probably would have expected that and to some extent I think there’s a little bit of a buyer strike. I’ll tell you that we looked at bonds this week and last week that we looked at a month ago and it feels like we ought to be able to buy them cheaper given what’s happened in the stock market and general risk off. But it’s not really the case. So the prices have held in a lot better than they did last year. So why is that? Well, there’s actually a lot of reasons.
The other larger factor last year was the Wall Street dealer inventories and if you look at trace data the Wall Street dealers sold $20 billion of non agency RMBS in the second half of last year. Fast forward to this year, the dealer inventories are very light. So there’s really not that selling pressure there. European banks we’ve seen very little selling out of them. Domestic banks they just came out with the new risk based capital guidelines this week and while they’re incredibly difficult to understand and get your arms around, I think the early conclusion is that these guidelines are somewhat enlightened and will not likely cause them to have to sell these securities. And finally the main lane liquidations that you saw early this year went exceedingly well. It was a food fight for bonds. So prices have actually hung in very well over the last month or so. Obviously they were up a lot in the first quarter and again it feels like bonds should be trading cheaper, but we’re not really able to pick up bonds cheaper. The sellers are not motivated sellers or not forced to sell. So sellers are willing to sell if prices come to them, but there is quite a bit of an equilibrium in the marketplace.
Anything new on the SEC review of residential mortgage RITs and the leverage they take? I haven’t heard much about that recently.
Craig L. Knutson
You’re referring to the concept release? Around 3C 5C. So the SEC put out this concept release. They solicited comments and letters and they pick all that back and they take it into review. So there’s been no official communication or word as to where that stands and unfortunately it’s highly possible that we may never hear anything about it. So in other words there’s no requirement that they publish some conclusion about this. That oh, we’ve decided not to do anything or we haven’t decided to do something. So it’s really, there is no new official information. I think the one thing that market participants look at is the fact that there have been certainly several billion dollars of equity capital raises that have been done since that concept release and so certainly I’ve heard this argument that gee, if the results of this were going to be really bad, will those capital raises have been permitted to proceed. But there’s no official word.
Most mortgage REITS are trading around book value. So that’s probably price into it to a certain degree. But if they came back and said that six or seven times leverage is inappropriate and they mandated something less which would impair your business model, what I guess would be this reaction to solution to something like that which could obviously roil the market short term?
Craig L. Knutson
Well, it certainly could. But I don’t think that it’s even within their purview to come back and mandate leverage. So the debate really is are these investment companies or not and if they’re investment companies and they’re regulated under the 40 act and then they’re limited to I believe one and half times leverage. So it’s not a question of figuring out what the magic leverage is. It’s are they investment companies or not because we all rely on this greasy 5C exemption to be exempt from the 4EX. The conclusion is we’re not investment company. But even that being said and I got a call from one of our largest shareholders when this concept at least first came out, I don’t know what was it, a year ago? Probably close to that and so what’s the disaster scenario? So if I have this right you guys wouldn’t have to own agencies. You could just own non agencies and so at least yes you’d have to run with less leverage but you’d have assets that yield 7% and he’s right. That is because the 55% whole pool test gets passed with whole pool agency securities and non agency securities that we own are not qualifying real estate assets for purposes of the 3C 5C exemption. Probably more than anybody ever wanted to know about 3C 5C, right?
Maybe one last from my perspective. Can you talk a little bit about the appetite for re-securitization market on the non agency side?
Craig L. Knutson
Sure. So right now this has been the case for a year and a half or so. The only rating agency that’s’ still active in the re-securitization market is DVRS. And so the last three deals that we’ve done have all been with DVRS and while we’re very happy with those deals, the appetite for DVRS rated triple A securities is not the same, is not as deep as it would be for instance SNT or movies or fixed rated triple As. So it’s worth finding for. It’s actually we saw levels, we got more triple As in the last deal that we did. We got 43% triple As versus 37% in the prior deal which doesn’t sound like a lot but with the structural leverage and re-securitization it’s actually very profound the difference. So we continue to see it as a very viable market. We’ve heard rumblings out of SNT that they’re close to coming back in being able to rate re-securitization. So I think obviously that would help the market.
To some extent it’s a double edged sword because if SNT were to come back and make re-securitizations more viable then more people would do them and then the asset prices go up and then we’re chasing yield again. But it still is a very viable and vibrant market. The one difference in the last deal that we did is we sold fixed rate bonds. In the prior deals we’ve sold floating rate bonds. When we last sold floating rate bonds it was probably June of 2011. So it was before the fed came out and said we expect to see low rates from 2014. Once that became the case buyers are not as anxious to pay up to get floating rate protection if they’re buying a two year security and don’t think rates are going to go up for two years. So the cost is a little bit higher on that last deal, but because we got a much better execution in terms of more triple As, the ROE if you will on that single trade was probably the best that we’ve ever had.
Continuing along the financing theme, on the non agency side, do you have a target level that you’re looking to reduce the short term repo to? Or should we expect to see more of these multiyear collaterals type longer term dated financing come through?
Craig L. Knutson
So I think we’re pretty comfortable with where we’re now. We’ve tried to mature the capital structure around these assets. If an opportunity comes up in a particular pocket for 18 month money or 123 month money obviously we’ll look at those. But I think we’re very happy about how we’ve reduced our reliance on this 30, 60, 90 day repo. And again, there is some cost associated with that, but the asset yield is sufficient that we think it’s proven and we can afford to do that. Well, thanks everyone for coming and thanks for your interest.
Right. Thanks Craig and Goodmunder.