This quarter was the most-covered I’ve seen in years. (Is that an early top indication?)
There were write-ups on mREITs from nearly a dozen firms. I’ve had the pleasure of reading coverage from Deutsche, BankAmerica Merrill, Wunderlich, Credit Suisse, Citi, Stern Agee, KBW, JMP, and Jeffries. No doubt I’ve left one or two out. A special thanks to readers who forwarded reports. Keep ’em coming.
Not since Friedman Billings and Ramsey shocked the equity world by being number one IPO underwriter (virtually all mREITs) has there been as much interest in the sector.
Before summarizing my thoughts on the several dozen equity research pieces I’ve read and handful of calls I’ve listened in on, I’d like to reiterate the basic economics of this business and the state of the financial union.
First, we need to recognize that there are rules for mREITs that force everyone in the business to implement variations on the same strategy. The two rules that overwhelm the others are the earnings distribution requirement and the “good REIT asset” test.
Everybody with even a passing interest in REITs knows the 90% distribution rule. In layman’s terms, all REITs have to distribute at least 90% of their taxable earnings during the calendar year they earn them. (The fiscal year is the calendar year for REITs because of this.) Slight variations from this rule-driven payout are the declaration/distribution calendar and the carryover tax (4%) on profits that are held for distribution before the end of the next calendar year.
The dividend of a REIT is considered to have been paid in the year it was earned if it is declared in that year, and if the record date is before December 30 of that year and if it is actually paid out by January 31st of the next year.
In practice, a number of mREITs pay out quarterly dividends with an uneven schedule because of this. For the first three quarters of the year, they wait until the accounting is finished for the quarter, and then declare and pay their dividends. For the fourth quarter, they don’t have this luxury. They have to declare the dividend even before the quarter is finished. For this reason, some mREITs pay in May, August, November and January. That can make for a happy holiday season, but the feeling of an income drought waiting for that first quarter dividend.
Other mREITs will make a practice of carrying over fairly substantial earnings, and paying the 4% excise tax. It gives them an extra eight or nine months to use the capital, and works out to be relatively cheap capital compared to their common stock or preferred. Some mREITs, most notably RWT, have made this carryover a regular practice for years. Until the meltdown, Redwood’s stated policy on dividends was to pay out a steady small dividend for three quarters (what they considered the absolute worst case) and then a giant dividend at the end of the year to bring their payments into compliance with the REIT distribution rules. That all changed in 2008, when they found out, along with everyone else, that their “worst case” was nowhere near how bad it could get.
The other, less known, but also very influential rule is the REIT-qualifying asset requirement that 75% of their assets must be “undivided” interests in real estate or real estate debt. From a practical standpoint, this means mREITs have just two choices — to invest in unsecuritized loans (also known, coincidentally, as “whole loans”) or to invest in “whole pools” of Agency-guaranteed securitized loans, or MBS.
Today the unsecuritized, unguaranteed mortgage loan market is basically dead, so even mREITs that focus on non-Agency paper still have three quarters of their portfolio invested in new Agency MBS pools. The exception is the investor that also can service distressed seasoned unsecuritized loans, like IVR, with its connection to Wilbur Ross and his vulture fund investment in the servicing operation of American Home Mortgage.
The reason the qualifying pools are almost all new issue is that they get auctioned by originators each month from their production flow, and a buyer can bid a couple of “tics” (32nds of a percent) higher than the generic price to own the whole pool.
Once a pool is in the generic secondary market, it gets broken up into pieces over time as investors buy and sell. This happens naturally, because trading happens in lots of $1, $5, $10, $20 or even $50 million. Since the sale of $10 million might include pools that have partially paid off when the trade happens, it wouldn’t be uncommon to have a $10 million sale consist of three or four pieces of seasoned MBS pools that had $15 or $20 million original face amounts. Over time, almost every pool that started out undivided gets broken into pieces if it trades in the secondary market.
Of course, that doesn’t apply to the whole pools that mREITs own, but it also prevents them from selling pools that have shifted in relative valuation unless they are comfortable that they can replace those paid down or sold MBS with new “whole pools.”
Once you realize that most mREITs have the bulk of their portfolio formed by pools bought when those particular MBS first hit the market, then you know why an mREIT can really only distinguish itself from the competition “around the edges.”
They do have the choice when buying these new pools to choose ARMs or fixed rate MBS. The managers can’t choose to buy something that isn’t being originated, so even though they may like the investment characteristics of FNMA 6.5% MBS today, there just won’t be any new whole pools of that coupon available, because that would mean some originator came to Fannie Mae last month with a big slug of conforming 7% fixed-rate mortgages. Won’t happen.
Same with the desire to buy discount bonds a few years ago. With no new 4.5% loans being made in mid-2005, you couldn’t buy a Freddie 4 in whole pool form at that time.
For ARMs, most borrowers choose hybrid loans that are fixed for 3, 5, 7 or 10 years, and then adjust every year or even more frequently. If you’re buying whole pools, that’s basically what you’ll get. That leaves a longer period of fixed rates than a typical repo financing (most repo’s are three months and less), so the portfolio manager has to manage that fixed-rate period until the loans begin adjusting on a regular basis. Even then, there are floors, caps on rate adjustments, and often a full year between adjustments, so fully seasoned ARMs still need to be hedged, or at least held with the understanding that there is both price risk and net income margin spread risk while holding them.
Having said that, the wide variance in performance last quarter tells us that different tactics and strategies can make a big difference, even when the “long” portfolios are mostly the same. First among the differences is that fixed-vs-floating choice made when selecting the “core” portfolio of REIT-qualifying assets. Second is the choice(s) made to hedge.
The crew at Two Harbors is choosing to make cash-flow hedges in the form of excess interest IOs and Inverse IOs from structured deals. When added to their more “normal” LIBOR swap hedges, that gives TWO about 75% hedging. That 75% figure is my educated guess given their slightly sub-1 year effective duration gap. Annaly chooses to go unhedged on about three quarters of their portfolio, as indicated by their effective duration north of three years. Anworth, Capstead and Hatteras try to keep their portfolios 100% hedged (as to funding cost risk), a fact we can infer from their duration gaps in the three month neighborhood.
MFA is mixing deep discount and high premium portfolio elements so they can withstand bull and bear moves in the bond market on a capital basis, though it leaves them exposed to big moves in funding costs. They are (my opinion) proceeding with the assumption that the values of their non-Agency MBS are likely to enjoy the benefit of increased collateral value if inflation rears its ugly head, and likely to enjoy increased yield spread on the whole-pool REIT-qualifying portfolio of premium MBS if ordinary rate increases come to bear and slow down prepayments. In this sense, the premium they paid for the Agency paper is just equal to owning par bonds plus IOs, an echo of the TWO cash flow hedges.
All the mortgage REITs are acutely aware of the importance of asset selection and hedging, though they only talk about the asset selection in their conference calls. I believe that’s because they lose their audience quickly when explaining the hedges, and because the hedges are traded quite actively to maintain the risk posture they desire. Remember, they have that 75% block of assets that really shouldn’t be sold.
The other place we see variance is in the funding strategies. Today the option of getting permanent financing by issuing CDO’s is gone, and unlikely to return. Still, an echo of that strategy is what we see when CIM and MFA do re-REMICs. Once the re-REMIC (or CDO) bonds are placed, the financing cannot be called, and has no margin calls. The downside of “permanent” financing is that the remaining asset is illiquid even in normal times, and virtually unsaleable in a crisis.
For those still with me, I’m now ready to make my basic market observation.
The banking industry still owns tons of future losses in its portfolios. In spite of two full years of zero interest rates, they are still largely dependent on government guarantees to stay in business. Our housing finance business has not recovered to a healthy mix of two thirds “government” and one-third “private label.” Instead, it’s 90/10 or worse. At the peak of the bubble it got up to 50/50, a clear indication that the private label business was over-reaching.
Our financial industry has been very successful crying poverty. They continue to get more help than any other financial industry, except for the basket cases in Ireland and Greece. Those “socialist” countries in continental Europe only briefly guaranteed their banks’ obligations, and only direct obligations. Those countries did not do as we did by stepping in and guaranteeing the assets the banks held. We have to wean our banks off the taxpayer guarantee heroin before they become permanently addicted.
As long as our banks can borrow at 0% to 0.25% and lend to the taxpayers at 3% to 4% yield, that’s all they will do. As we can all see in the new terms on our credit cards, when the banks in America lend to us on our own credit, they charge more than ever today and give out less credit, in spite of the low cost of funds they enjoy. I don’t count their loans on housing, since they take that risk for all of five minutes by getting government guarantees on the loans from Fannie Mae and Freddie Mac.
I was talking with a bond buddy today and he remarked that the securitization business has basically recovered in Europe, with both mortgage and asset-backed securitizations once again being launched across the continent, all without government guarantees on the resulting bonds. They may be socialists over there with their free medical care for all, education through the university level without taking on back-breaking debt, their years of unemployment coverage, and their strong unions for most ordinary workers, but it does hold down on the mortgage and credit card defaults. Except in markets like Spain that got tons of speculative investment and “hot money” foreign buyers fueled by easy credit, housing values in Europe were just as high or higher than our bubble peak, yet those values didn’t collapse anything like ours has. We might consider asking what they did right and we did wrong.
Having said all that, the amREITs are doing exactly what the banks are doing, borrowing at low rates and holding government-guaranteed paper that yields 3% or more. They don’t have nearly the same leverage as the banks, but they also don’t have sources of funding like deposits, a funding source that doesn’t look at the assets every night and issue margin calls the way a repo lender does.
In my opinion, the banks have enough losses embedded in their portfolio to offset the spread income from this big giveaway for another couple of years. That’s how long I think Ben Bernanke and the Fed will continue 0% bailout rates, asset buying, emergency lending (like TALF), or whatever else.
Too many people forget that the Primary Mission of the Fed has nothing to do with unemployment, inflation, etc. Their true reason for existence is make sure the banking system survives. If that survival is threatened, the Fed will do whatever it takes, including actions that cause double-digit unemployment or inflation.
Now that we know the environment for owning levered portfolios of MBS is likely to stay great for another year or two, and we know that the signal it’s coming to an end will be banks looking to lend aggressively to consumers and small businesses (without taxpayer heroin), then we can invest in these mREITs.
Selecting which mREITs and how to allocate is the trick. Also recognizing how their portfolio strategies turn into taxable income and capital gains and losses will make an investor less nervous holding these stocks.
The fact that Annaly took the largest book value hit last quarter was no surprise at all, given their heavy exposure to fixed rate MBS and their very light hedging ratio. Never mind the fact that none of their hedges (undisclosed, so I’m only guessing) have anything to do with protecting capital. In my opinion, what hedging they do is geared only to protect part of the income stream from spikes in the repo rate.
Similarly, I wasn’t surprised that ANH took a very small hit to book value. Even that was higher than it should have been, were it not for the Agency delinquent loan buyout program.
The final push of Agency loan buyouts hurt ARMs disproportionately, but that’s now over. In general, the more ARMs you have, the less capital risk you have, and the more seasoned ARMs you have in your portfolio, the less capital risk vs. your competitors with ARMs still waiting for their first reset date.
The smaller REITs with portfolios in the $5 billion neighborhood could afford to focus on “niche” plays and buy meaningful pieces (for them) of those MBS.
One reader mentioned a refinance program being offered to very high FICO borrowers that allowed them to get a lower fixed interest rate in spite of losing their equity in their homes. Fully qualifying that kind of borrower as to assets, income and credit history can make for a very slow paying MBS that has virtually no default risk (unfortunately today’s primary source of prepayments). According to that reader, the boys at AGNC were quick to jump on these, and they managed to buy a good-sized slug of them before the market repriced them higher than conventional fixed-rate paper. It’s interesting also that a program of special refinancing like that could give the team under Kain (former Freddie Mac guy) a chance to buy them when the loans first got pooled, so they can be REIT-qualifying while giving a higher prepay-adjusted yield. Nice. Much harder to “move the needle” with a portfolio ten times that size.
Bear in mind that only Annaly has a portfolio that ranks up with major insurance companies, banks, and fund managers. The rest of the amREIT universe is still relative pipsqueaks in a $5 trillion market where hundreds of billions of trades settle every day.
As I mentioned, I wasn’t surprised by Annaly, nor by Anworth. The AGNC and TWO crews were too new to gauge last June, but they’ve both performed as they said they would in their presentations from Q1 and Q2 this year. I’ve gotten more confident in their ability as a result. I’ll still withhold judgement until I see them survive a one or two year bear market in bonds. That’s the real test. Hedging and smart ownership of premium bonds and IOs will be the way to survive when that happens.
I liked what I saw at CYS, as well, so I got in on the price decline associated with the recent secondary offering.
I continue to see CIM as fully priced, and I continue to think that many of its investors don’t recognize the potential for required dividend distribution that exceeds their positive cash flow. It can happen when too much of your earnings are accrual of discounts. Merrill and Credit Suisse both lowered price target on this stock, as did a couple of others. That’s not to say the 17% dividend isn’t a fine reason to hold it, even if there is no more capital gain to be had.
I hold my MFA without selling, but also without adding.
CMO disappointed. I had hoped to load up if the market became disappointed enough to sell more shares to me at $10.50. Didn’t happen. The shares I hold cost well below $10, so I’ll have to have something really good to do with the money before I sell in my taxable accounts. My main disappointment was the fact that they had allowed their cost basis to creep up high enough that the Agency buyouts hit book value. I would be happier if they kept their basis under 101.5% on the stuff they own, and then they kept their duration gap inside three months, as before. Then I’d be willing to own it up to 10% or maybe 15% above book value. Yes, it would mean lower dividend yield, but I can live with a well-managed low volatility 12% yield.
I think I’ll cut this off now and add comments as more reflections on the quarter that was come to mind.
Before I go, I should give you the good news for the sector. In the month of October, Agency MBS outperformed US Treasuries almost as much as they did any time in history. Most of the damage has been repaired, in other words. Check out this Bloomberg article that lays out the fact that generic MBS beat Treasuries by 92 basis points last month. Wow!