In the big Friday lead-up to America’s most politically incorrect semi-holiday (Columbus Day), I received a link to one of Jody Shenn’s articles on Bloomberg with a reasonable question:
“Does the increase in early prepayment have any role to play in the recent amREIT decline?”
So now that I’ve survived the heat of a weekend that should have been a brisk leaf-peeping delight here in New England, let me reflect on some of the data in Jody’s article.
First, we need to note that any econometric data point is judged vs. expectations, not for any absolute level. Second, we need to get inside the numbers to see how our amREIT holdings might have been affected.
In Jody’s article, we see that the recent plunge in mortgage rates attributed to Operation Twist seems to have spurred a mini-burst of refi activity. But that activity is limited to the borrowers who recently passed through the post-crisis underwriting gauntlet (the people paying 4.5% on their mortgages that are part of today’s 4% MBS). For those recent borrowers, their home appraisals reflect the typical pendulum swing to the conservative side that follows any local optimistic price peak. Also, mortgage loans originated in the past two years went though a documentation of family income and assets that might make a colonoscopy seem less invasive.
The people paying 5.5%, 6% or 6.5% since 2005 and 2006 are still locked out.
That means the amREITs that own seasoned 5.5% MBS are still enjoying incredibly low prepayments and many basis points of excess yield.
Those who bought 4% MBS earlier this year might be losing the 2% or so premium they paid on the one and a half percent (17.7% CPR) of those MBS that paid off last month. Certainly not the end of the world, since even a portfolio consisting entirely of that very specific MBS was probably already expected to have at least a half a percent to one percent monthly prepayments.
So let’s say the amREIT holding a chunk of recently originated FNMA 4% MBS had an extra 1% prepayment last month. That was an extra loss of premium of just 1% times maybe 2.5% of premium, or two and half basis points. Multiply by 8 (for 7x leverage), and the “hit” to asset value is whopping 20 basis points. Far less, in other words, than the daily price-spread volatility every MBS holder experiences. Put in terms of earnings, the maximum hit to any of our universe of amREITs is still less than 10% of the ongoing spread income, and that “hit” is limited to that fraction of the portfolio that is recently originated fixed-rate MBS.
There’s no way this uptick in prepayments can explain the hit to stock prices we see in the sector.
Before I let this go, I need to comment on the season. Prepayments for October also include mortgage loans paid off as the borrower sold their homes. Those real estate closings and loan prepayments actually happened in August and early September, and rippled through the MBS payments system to become a decline in principal in the MBS in October. That coincides with the traditional selling season for families with children. Historically, except for resort, retirement or vacation homes, we see a peak each year in prepayments in September, and especially in October.
I’m quite sure the amREIT management teams take seasonality into account in their internal projections of portfolio income, though they all do boil the prepayment expectations down to a single number for investors and analysts to worry about.
Which brings us back to Columbus Day, which was the peak for foliage in Connecticut for most of my youth and early adult years.
A few years back, I took a picture of the huge (about fifty foot tall) Japanese maple near my driveway when it was showing its peak color. I feel very lucky that whoever owned my house in the 1920’s planted that tree, which has spawned several dozen offspring as an understory throughout the woods behind my house. It is simply the most amazing red color I’ve ever seen on a tree, though I still prefer the variegated splendor of sugar maples when it comes to New England fall color.
FWIW, temps are in the 80’s and we’re still a few weeks away from peak fall color. I checked for some friends, and most B&B’s in the area are booked solid for the weekends, but midweek reservations seem to be available. My best guess is that two or three weeks from today might be near the peak for Litchfield county.