MFA Financial, Inc., (NYSE:MFA)
Credit Suisse Financial Services Forum
February 8, 2012 2:15 pm ET
Stewart Zimmerman – Chairman of the Board, Chief Executive Officer
William S. Gorin – President, Director
Craig L. Knutson – Executive Vice President
Thanks. Our next company is MFA Financial. We’re happy that MFA is joining us here today. They haven't yet reported fourth quarter, but again we’re very happy to have them. MFA is a residential mortgage REIT who invests across both the agency and Non-Agency MBS sector. We like the flexibility this gives the company to find the most attractive risk adjusted returns.
With that now let me turn it over to CEO, Stewart Zimmerman; President, Bill Gorin; and Craig Knutson, who is responsible for Non-Agency investing. Thanks.
Thank you. Okay, good afternoon and thank you for your interest in our company. We hope you find this informative. We are an internally managed REIT positioned to benefit from an investment in both Agency and Non-Agency residential mortgage backed securities. So, let me say that one more time, we are self advised and self managed. Again very, very different than those companies that have an external advisor and again I think it's an important part of the fact that we are aligned with our shareholders in terms of your interest.
We have an experienced management team focused on residential mortgage backed securities. I've been doing this since 1967. So it's more than a couple of years and actually Bill and Craig. Craig and I worked together once upon a time, number of years ago. So, we've all been involved in the residential mortgage backed area for a number of years.
Investment opportunities certainly exist in residential mortgage backed securities particularly in the Non-Agency sector. Agency MBS investment returns continue to benefit from the steep yield curve and despite government [elections] refinancing rates are not spiked.
In this period of economic uncertainty our goal remains to generate double-digit returns on equity with an appropriate level of leverage. Again, I think you’ll find that we have been a very disciplined company and that we’ll continue to be so. Slide I’m very proud of, we’ve had as 14.5% annual returns since January of 2000 and you can see as we’ve depicted here MFA relative to the S&P 500 financials and the S&P 500.
Having said that, as introduction, I’m going to turn it over to Bill Gorin, who is the President of the company.
William S. Gorin
Thanks Stewart. So, while Stewart mentioned our focus is residential mortgage backed securities. We allocate our equity and assets over Agency and both Non-agency. So, the first column is an agency model, which some of you are familiar with, deleverage, debt to equity ratio is 6.7 times. So, not different than any companies you’ve seen. The yields in our agency assets as of the third quarter, in the third quarter was 3.37%, which again in line with other companies. Although, we achieved that without owning any 30-year fixed rates.
So, we’re very happy with the yield on our Agency portfolio. Our cost of funds in our Agency portfolio is 1.74% and we’re not very happy with that. That reflects the fact that we invested in prior time periods when interest rates were a lot higher. And the good news is we expect this cost of funds to go down because many of our high cost swaps will be running off over the next 12 months.
But the spread on agencies was 1.63% which is high for us on a historic basis and, spread we’re happy with.
The next column is the Non-Agency MBS portfolio. So this is a portion of the company that we've grown really, we timed it right, we got in the bottom at the end of ’08, I believe. The leverage is two to one and we've used different forms of leverage overtime. So when we reentered the Non-Agency space we’ve financed all the assets just with equity. And eventually as the market became available and as we became more comfortable, we added some repo financing.
Again we waited for the market to become more efficient and we added secularization financing, and what we’re going to tell you today in this presentation is, we’ve actually added three-year term structured financing of $500 million. So we continue to diversify our funding sources of longer-term funding for a Non-Agency MBS. And why is that interesting to us? Well, the unlevered loss-adjusted yield our Non-Agencies as of the third quarter was 7.29%.
And today we’re going to tell you what loss-adjusted means. Our funding cost was 1.61, so the spreads were 5.68 and with two times leverage, the ROE there is in the mid-teens. We own cash and cash doesn't earn much for you and lowers your returns somewhat, but we will always own cash, because of all the uncertainties in the world. So the spread across the whole company was 3.5 times leverage was 2.9.
On the agency side, we’re about 75% hybrids, in case you don't know a hybrid might be fixed for an initial term and then become a one-year adjustable. And 25% of the portfolio or so is 15-year fixed rate.
Our amortized cost basis is 102.6, I think that would compare very nicely to other [reach] you may be familiar with. We believe we have limited exposure to HARP 2.0. So, really, it seems the goal of HARP 2.0 is to get people to refi from high-cost mortgages into new lower cost 30-year, amortizing fixed rate mortgages.
Now, the impacted vintages would be 2009 and prior and the assets we owned from that vintage are not fixed rates, they’re hybrids. The coupons were about to reset down, they are interest-only. So, there is not much incentive for people to refi out of these pre-existing hybrids into new fixed rates. And as I mentioned, our agency funding costs were expected to decline, about $740 million with an average cost of 3.66, we’re running off in the next 12 months.
So, as of the end of September we had 4.1 billion market value of Non-Agencies, with an amortized cost of 73% of par. The loss adjusted yields in the third quarter was 7.29% without any leverage. Now, when you have such a high yielding asset, it’s not clear that these assets are correlated with interest rates. In fact, what you would have seen last year was that the value of these assets was correlated to equity markets and now they’re all correlated to treasury markets for example.
We’re also a lot less sensitive to prepays, so if you buy an asset at a approximate 30% discount and prepays were to go up, you capture a lot of value there or put another way, if our agencies have a two point premium, and our Non-Agencies have a 30 point discount, if CPRs were to go up across the board, we would be better off by 15 times when non-agencies are capturing a 30 point discount versus [accruding] a 2% premium. Public knowledge of prices of Non-Agencies declined throughout 2011 and we were a buyer, not a seller of Non-Agency. It was an opportunity for us.
And with that I'm going to let Craig run you through a little mathematical example of what loss-adjusted yield means?
Craig L. Knutson
So, thanks Bill. So, yeah, people always ask us well, what exactly does loss-adjusted mean? So, we’ll lay that out here. If you look at the first line, the historic base cases, this is what I’ll call the old days, which is six, seven years ago. If you bought a Non-Agency security at a dollar price of 72, which is not really realistic as six or seven years ago, they all traded at par. But, if you had bought it at 72, you had a 4.5 coupon; you would assume no defaults and no loss severity because of course these are AAA securities. So, they always return par and you would have used a pre-payment speed of 15% or perhaps even faster. And so you'd get a 14% yield on that bond.
So now we change things one at a time. So, what we now know is that because most of these borrowers don't have equity in their homes and because refinancing is much more difficult today for homeowners especially with jumbos the speeds are a lot slower. So change nothing else except for speed; slow your speeds down to 6% and your yield goes down to 10%. Now we’ll add in our loss adjustments. And so this is assuming that 44% of the loans in the pool default with a 50% loss severity with the same 6% CPR, and now your yield goes to 7%.
So that's what we mean by a 7% loss-adjusted yield. People compare our product to jump-ons, and so jump-on yields maybe quoted at 7% or 8%. But jump-ons are typically not quoted at loss-adjusted yields. So, that’s before any losses.
So, housing fundamentals remained weak, but we believe that our credit reserves have appropriately factored into our assumptions these weak fundamentals. Some of the things to keep in mind, the securities that we owned in the portfolio, we still have credit enhancement underlying. So this is, credit enhancement that sits below our level of our bond. It was credit enhancement that was originally structured into the deal. That's a little bit less than 6%. So before we take dollar loss one there’s almost 6% of credit enhancement beneath us.
Second thing is a discounted purchase price. We certainly don't expect to get back par in these securities, but if our average purchase price or amortized cost is $0.73 on the dollar then not only do we not expect to get back, you don't need to get back par.
And then finally we have a substantial credit reserve. So when we book income and I have a couple of slides to detail this a little bit more. We have a credit reserve that's approximately $1.2 billion to cover losses and this again this is after the credit enhancement that exists already in a $5.6 billion phase portfolio.
So, this illustrates what we mean by this credit reserves. So over on the left, you have current phase of $5.6 billion, $5.7 billion, you can see the purchase price in the blue is about 73% of par. So we had a significant discount or $0.26, $0.27 in the dollar. So the question is what do we do with that? Well, of that discount, if you look over on the right, we've reserved most of that as a credit reserve. And only a small portion at the bottom as an accretable discount. So what does that mean? That credit reserve actually for us is very different than it would for a bank for instance. So a bank when they have loan loss reserve, that loan loss reserve goes through their income statement. If they were worse, it comes back through their income statement. This actually sits on our balance sheet. It doesn't come through our income statement and what we're saying is we don't expect to get back that principal amount on those bonds.
So, the portion that we do, the portion of the discount that we expect to get back is accretable discount at the bottom. So, our yield on those securities is the coupon obviously divided by the price and then we have a small amount of discount accretion because we’re only accreting part way back to par.
The other thing is that the credit reserve of 21%; in essence what we’re saying is that we expect to get back $0.79 on the dollar for those securities that we have an amortized cost of 73 for.
So, I won’t go through a lot of numbers on this page. This is our 20 largest Non-Agency positions and a couple of things I'll point out, you can see that there is no one position that’s even 2% of the whole portfolio. So, it’s pretty diversified. If you look down at the bottom, we've circled several numbers there. So the average cycle, this is that origination, is 734, which gives you some sense for sort of the overall quality of the portfolio.
But then people look and say, well you have 60 plus day’s delinquent to 22% of the portfolio is 60 plus days delinquent and yet you're projecting defaults of almost twice that, of 43%. And people say, well, that looks pretty bad. And I guess our response to that is, well, we hope it is pretty bad. But, that's the assumption that we’ve baked into these securities. So the yields that we book on these securities assumes that twice the level of delinquent loans defaults over time. And again we're using approximately a 50% loss severity.
The other side is with Non-Agencies, it’s not just about asset selection, it's also about the capital structure that we employ to hold those assets. So, we have a market value of Non-Agencies of approximately $4.1 billion, we had $1.8 billion of repo, we had $1.4 billion of equity and we had securitized debt of $1 billion.
Now, the securitized debt, these are these resecuritization that we've done, so that we bundled the Non-Agency securities, then we issued new securities and we get a certain percentage of those, couldn’t be anywhere from 25% to 40 something percent are rated AAA, because we subordinate the rest of the cash flows to those and those we’ve sold to third-party investors at LIBOR plus, 100 LIBOR plus 125 type rates.
So, as excited as we are about the secured financing from the resecuritizations, as Bill said before, we’ve now added a fourth layer of funding to our Non-Agencies. In the fourth quarter we entered into a three-year collateralized financing arrangement that essentially gives us $300 million and we then increased it by another to $200, so it’s a total of $500 million now of three-year committed financing for Non-Agencies.
So, we think that's very significant to the extent that we can reduce our reliance on 30-day repo. That's always a good thing. This three-year financing, of course as you would expect is more expensive than 30 day repo. And so we tell you here that it's approximately a 100 to 150 basis points more expensive than 30-day repo. But we think the certainty of three-year financing is very, very significant.
So just a sort of review the high points, MFA, we’re internally managed. Our goal is to continue to generate double-digit ROEs. We've generated a 14.5% annual returns since 2000. While housing fundamentals do remain weak, we believe that we've appropriately factored this into our cash flow projections and our credit reserve estimates. And finally, we’ve worked hard to diversify our funding sources for Non-Agency MBS, including the securitized debt and this three year financing.
And with that, I’ll open it up to questions.
Can you talk a little bit more about the three year structure now, what type of advance rates do you get on that and how does that compare to repo?
William S. Gorin
So the advance rates on that are pretty similar to repo. So, it’s a give or take approximately 30% type haircut. We haven’t disclosed exactly what the funding rate is for competitive reasons, but we've told you that it is more expensive in 30-day repo, which one would expect.
And what do you think that the depth of that market is? Is that something that could grow over time?
William S. Gorin
William S. Gorin
It's only could, its only could grow over time. Again we did $300 million in the fourth quarter and we added another $200 million. So, I think for us, it's about having multiple and diverse funding sources. So it's not about having, [exclusive] if we get finance our entire portfolio to resecuritization that would be the best way to do it, but you can't do that, because you still end up with a large, end up with a part that you don’t sell. So, yes I think, it could, it certainly could grow.
(Inaudible) lot of the other questions that are in the marketplace, and its an area which as Craig has said, we continue to make the structure, financing structure with much more efficient.
Unidentified Company Representative
The other advantage that give us is, as we look at assets and as we purchased assets we know, we have to think about how we're going to finance those assets, so it's not just about finding the right bond and getting it at the right price, because we don't own them simply for equity, right. So, we have to consider that the whole sort of package of ROE. So for instance, when we look at a bond, we know which bonds will resecuritize most efficiently. And so we might be willing to pay more for those bonds because we know that that we can resecuritize those and achieve efficient financing there. There may be other bonds that don't resecuritize very well, either because we don't get very, many AAAs or for a variety of other reasons. So having a committed line like this is great, because those assets might trade cheaper because of the fact that they’re more difficult to resecuritize but we have another means to finance those.
William S. Gorin
So, in terms of these programs and how they might effect Non-Agencies, I think the recent plan announced by the President for the first time addressed Non-Agency securities, obviously it’ll be a great thing if we had more prepayments than we expect right, because prepayments when you pay 70 for an asset are always a par and that's always a good thing, plus if they prepay, they can’t default. So, we’d love to see that. We certainly are not at a point where we would change any of our assumptions. And I think there are many political hurdles to actually achieving that. So I guess in short, it would be a great thing, but we're not drinking champagne yet, I’m sorry, the second part of your question? Okay, so for the securities that we own, which are primarily prime, prime jumbo site assets. The services are really the big four or five services. So you don't really see the same thing that you see in subprime paper where it makes a huge difference, [who] the service was, because they, it’s how much they advanced.
So at the low-end they might be advancing 30% or 40%, and at the high-end maybe 70%. I would say that the advance rates on the securities that we own are typically mid-to-high 90s. So it's a very rare on a jumbo loan that a service level stay, I don't think I'll be able to cover my advances. So it's one thing on a (inaudible) $150,000 that’s been vacant for three years, how you could see a 100% loss severity in the service we can’t recoup their interested advances. On a $700,000 to $800,000 property it’s hard to conceive that you have 100% loss severity on that. So, it's a good question, but it's doesn't really affect our cash flows all that much.
Unidentified Company Representative
Unidentified Company Representative
So as we mentioned, we invest in both Agencies and Non-Agencies. And in the Agency space 75% of the assets are hybrids where they are fixed for a period of time and then become one-year adjustable, so we don't have a lot of interest rate risk on the Agency hybrid space, where they’re become an one year adjustable. We do have about 25% of our Agencies in 15-year fixed rate, which amortizes over a 15-year life and we have fairly high coupons there, I think they’re 4 plus percent, so, again not high interest rate sensitivity on the agency fixed-rate.
The more interesting story is Non-Agencies. So, we told you in the third quarter that the Non-Agencies were yielding an excess of 7%. So, if you're been asked that yielding an excess of 7%, and you know the treasury rates at that time were probably got two-ish percent, so 5% of it is credit risk and two percentage interest rate risk. So, the majority of the value that asset and the cash flow is tied to credit.
So theoretically you would say it's not a high correlation between changes and interest rates and the change in the value of the asset, that's the theory, and while we have that theory you'd look at high-yield bonds over time. And once the spread over treasuries was an excess of 300 to 400, the asset did not trade with interest rates. In fact, it’s a negative correlation. When the economy improved and interest rates went up the value that high-yield bond actually went up.
So here we’re talking about assets that yield yet more, more than 300 or 400 over the 500, 600 over, that's the theory, what's the practice? So, last year these assets trades like stocks. They were very highly correlated to what the stock market was doing and negatively correlated to the treasury market they trade in the opposite direction. So, we believe that there is not interest-rate risk in the Non-Agencies we own in fact, is the opposite of the risk they probably move the opposite way of interest rate. If what causes interest rates to go up is an improving economy, the value of Non-Agencies will go up. The one exception it could be the stagflation people lose control of all interest rates then the correlation might be get somewhat tighter, but again Non-Agencies with its very high yield are not going to be tied to changes in interest rates.
So, that is how we look at our interest rates exposure. On the 15 years that we do own by the way, we have more swaps and we have 15 years. So we've hedge it out the risk on the 15 years.
Great, if you could just touch on how the performance on underlying credit performance of the Non-Agency bonds has been recently whatever you saw, kind of seen in the inflows of new delinquencies and kind of [rollover]?
William S. Gorin
So, again because we haven’t reported, we really have to go back to third quarter, I guess two years ago, we had fairly significant move, or we move to those about $100 million, I think from our credit reserve to our accretable discount. Since then each quarter, each year we've had small moves or smaller moves the original large move was primarily due to the fact that those were assets that we’ve purchased in the depths of crisis end of 2008 and 2009 and (inaudible) some of those we probably used assumptions which are little too harshly default 90% of the loans with 70% loss severity in the whole three years later that was probably a little bit too harsh.
So we had, like say we had a fairly large adjustment there, we continue to see those adjustments, but we’re slower to take those adjustments, because we're making assumptions and projections on these bonds that go out 25 years. So we’re little reluctant to base our vision on three months or six months worth of data. So we don't really want to move credit reserve in, and then move it out to next quarter, move it into the next quarter. So I think we look for a little bit more evidence that they really are performing better. Our voluntary speeds have generally always been faster than how we’ve model them. So that obviously is to good think. But I think we've been somewhat slow, but nevertheless we have made those adjustments.
And we have two people all they do is surveillance on our existing security. So I think we understand and the data that’s available on these securities, it's frightening how much data is available. The real challenge is thinking about what to do with the data, but, [probably the] example is pay history. So we can look at these loans and the average even as of the third quarter, I think the average seasoning was five years. So, we have five years of pay history on all these loans.
So, is there another question in the back?
William S. Gorin
Okay. All right, well thanks everyone. Appreciate your attendance.