In a recent post I pointed out that the amREITs (agency mortgage real estate investment trusts) were acting as if they were going to trade at a discount to their book values.
A word on book values first. In its cleanest form, the book value of an amREIT is simply the liquidation value of its portfolio, less repayment of its debt, most of which is margin financing (repo) of the portfolio itself.
There are some other pieces, including value placed on intangibles like goodwill and customer relations (at NLY) and non-agency MBS they hold (MFA).
These non-Agency assets are generally minimal (a nickel or a dime of book value) compared to the book value derived from easily determined Agency securities prices. Working with the book values comes very close to telling you whether these companies’ portfolios are actually trading at a discount.
Adjusting the book values for market moves is another matter entirely. The research analysts that follow these firms have recently shown how hard it is to track the value of the portfolios of these beasts. In a quarter when the Fed aggressively bought Agency MBS for its own portfolio, the prices on MBS marched upward without too much volatility.
What didn’t march upward so steadily was the net effect of both the MBS the amREITs holdings of MBS and the hedges they used. Most of their hedges, by the way, are not to protect the value of the MBS. Rather, the amREITs hedge to protect their ongoing positive financing spread.
Put another way, the managers of the amREITs view their main operating risk to be the possibility that the short-term rate they pay to finance their inventory rises to the point that their business is not profitable. A secondary risk is that the value of what they hold will drop, leading to margin calls that force them to sell at a bad time. The third major risk is that the borrowers will repay their loans faster or slower than expected.
The cash Agency MBS market is pretty efficient and liquid. Various hedges can imperfectly protect the financing spread and the principal value, and even less perfectly offset the option cost of mortgages (the option being the borrower’s optional right to pay off the mortgage early).
Unfortunately, no combination of hedges perfectly matches all these risks, and any attempt to actively cover all the risks would cost more than the yield spread you would be trying to protect.
So none of the amREITs can actually afford to be fully hedged, anyway.
Some choose to take more risk in financing, or more risk in principal value, or more risk from the prepayment option, but they all have to pick their poison in order to deliver a good return to their investors.
As any borrower knows, most times you can get an ARM (Adjustable Rate Mortgage) at a lower rate than a 30-year fixed rate. A borrower that takes on the risk that rates will be higher when the reset date comes gets rewarded by a lower cost for the first few years.
From an investor’s standpoint, it’s just the other side of that same coin. A fixed rate MBS pays a higher yield, but has more price risk and more financing risk. Today’s 4% ARM MBS will at least reset to a higher rate in five years when financing it may well cost 6%, or 7%, or even more. Most 4% ARMs can reset over time to as high as 9%.
By comparison, today’s 5% fixed rate MBS will always pay only 5%. Imagine the pain if repo rates are up at 7% in the future, and looking like they will stay there.
The price risk is also much larger for fixed-rate MBS. From a practical market standpoint, fixed-rate MBS have historically traded between 80% of their face value and 110%. ARMs traded during the same period between 90% and 105%. Even at today’s 6:1 leverage levels, the amREITs are vulnerable to sharp declines in equity even if the MBS market only moves a few points.
It would not be right to let the existential risk go unmentioned: They all depend on margin lending with relatively short periods of financing. Even if they get 3 month’s financing at a time, that still leaves the lender the option of declining to renew the available credit as repo lines mature.
A year ago, the Fed stepped into the breach and provided the repo financing to the Bank and Dealer communities for MBS, corporate bonds, Agency debentures and even bank loans. Without that, we were on a one-way trip to the financial stone age and a world economy with as much as half or three quarters of its GDP simply impossible to produce.
Last March, the “echo” of that earlier disaster had most banks more worried about their survival (the dealers were essentially gone by then), so the market naturally assumed that even their amREIT customers holding newly-federalized MBS might not get the financing they needed.
That’s what took the Gang of Six (plus one) down to historical lows versus their book values, even though the rates environment was as friendly toward their business as it ever will be. Annaly (NLY) traded down to $11, Capstead Mortgage (CMO) under $8, Anworth (ANH) almost down to $4, Hatteras Financial (HTS) under $20, American Capital Agency (AGNC) to $14, and MFA Financial (MFA) to $4.25. Cypress Sharpridge (CYS) just had its IPO this summer, so it trades at or under book value today.
Among the names most often heard, NLY, HTS and AGNC tend to have more fixed rate exposure and/or rely on security sales to capture extra profits, with the attendant extra risk. ANH, CMO and MFA stay more in the ARMs, and hedge their exposure to repo rate reset risk in roughly that order (ANH the most robustly hedged). For my money, ANH and CMO could be called the most conservative.
Tomorrow we find out what ANH’s book value moved to during the third quarter. The others had nice increases. I took today’s sudden decline as an opportunity to add a little of the ANH I hope to buy if the entire group goes down with a major market correction.
I’m still mostly in cash, and plan to quadruple my amREIT holdings if that happens. The single-digit amREITs will have to decline a buck or two to get me that excited about them, and the higher-priced ones will all need to decline 20% to 30% to get me to commit that cash reserve.