Timing is Everything

This week marks the annual conference in New York for small and medium-sized financial companies sponsored by FBR (Friedman, Billings and Ramsey, the high-flying specialty broker that made itself a subsidiary of its mortgage REIT a few years back).

I used to make it a point to attend this conference, both to renew old friendships, and to hear from CEO’s of companies I invested in.

I’m sure this year’s confab was much, much smaller.  Still, from the ashes rise the Phoenix, and my friend was intrigued by the presentation by Eric Billings, the ‘B’ in FBR.

I got this e-mail from an investing friend today, and halfway through answering his question, I realized it would probably be a decent blog post.  So here goes:

Good morning, Howard,

I listened to the presentation from Arlington Asset ( NYSE:AI ) at the FBR Capital Markets Conference yesterday and heard Eric Billings and Rock Tonkel lay out their strategy for turning $165M in invested capital into $300M or more in 12-18 months using non-Agency RMBS without leverage. The prime and alt A assets have been substantially already purchased. They believe they will get this return through the reflation of these securities back toward par. This will benefit from their NOLs, making it mostly tax free.

If you were a handicapper, how would you rate the likelihood of this return? I’m a rube in this game but on the lookout for a double like everybody else.

Here is the link if you have 40 minutes to listen in and view these slides.

Arlington Asset slide show (painless registration required)

Thanks for the help,


My reply:

Hi, XXX….

I will check that out later… thanks.

The run to par may not go all the way to par. if they bought the bonds at the absolute bottom (jan – mar this year), they paid in the 30’s or very low 40’s, and the double is in the cards, assuming they bought 2005 or earlier vintage

The problem with the alt-As is that the AAA bonds only had 2% to 6% subordination under them, and foreclosures are nearing 40% on a lot of the pools, even 2004 and 2005… on the good news side, those pools often had half the pool pay off before the storm hit, so subordination was able to nearly double.

Still, losses are extraordinary as some people take being underwater as an excuse to stop paying and live rent free for a year or longer… that puts the servicer on the spot to advance the money for the missed payments until they actually foreclose….

Why does this matter?

Because the bondholders are getting those advanced dollars, the servicers stand in the absolute front of the line when a foreclosed property is sold…. if they advanced taxes, insurance and payments equal to 10% or even 15% of the loan amount, there is not much left over to pay bondholders…

Look at it this way:

If the house had an 80% LTV ($400k) mortgage based on a $500K appraisal, then the advances from the servicer might add up to $40K to $60K. Add in the legal costs to foreclose, the maintenance (some borrowers are taking more than their furniture when they finally leave), and the cost of selling the house, and you might be looking at another $40 to $60K in sale proceeds that need to be reimbursed. Take the middle of this range – $100K as an estimate.

Now assume the house sells for $250K (could be lower in bubble states). Even so, that leaves only $150K to cover the bondholders for the $400K of mortgage debt, making the loss severity a whopping 62.5%.

What this means for Arlington Capital’s prime and alt-A bonds is that we are still looking at allocating a substantial amount of loss to those AAA bonds if the house prices don’t recover before the wave of foreclosures finishes.

In my example of an alt-A deal that now has 40% of its loans heading for, or already in, foreclosure, 0.625 * 0.4 equals total loss of 25% of the current balance. Even if the subordination has grown to 5% or 10% of the deal, that leaves 20% to 25% of the AAA bonds turning into credit losses. If you bought them for 35, you only need to recover 70 to double your money, so if the slide stops here, you’re fine. If you paid 45 or 50, you will never double your money, especially if another third or half of the pool slides into foreclosure.

Same is true for prime deals, though today’s foreclosure pipeline is more like 10% to 12% of those deals. On the other hand, subordination is bupkus in those deals, often around 1% or maybe 2%. Also, those prime bonds only got down to the high 40’s at their bottom, so if Arlington paid an average price near 50 (quite likely), they can’t double their money.

The last qualifier is that the money only doubles if the remaining bond can be sold at par in 12 to 18 months. No guarantees here, but I think private-label MBS aren’t likely to trade at tight spreads vs government MBS any time before I finally retire.

Having said that, it does seem like 70% to 80% total return is in the bag. The question is whether they can realize that return in 1.5 years, or whether it takes 4 years… that will affect the annual return.

The good news from buying all these MBS so cheap is that even today’s low ARM coupons (3.5% or so) are really 7% to 10% yield on those low prices, so the pain of waiting for the principal payoff is not as bad as it might seem.  Fixed rate MBS are even better, paying 5%- 6% on the face amount, which translates into current yield as high as 15% or 20% given the discount purchase price.  That certainly takes the sting out of waiting for the recovery.

As they described the trade (and when they bought the kinds of bonds described), I would put the unlevered IRR as between 40% + (very fast recovery) and 20% + (if it takes a long time).


8 Responses to Timing is Everything

  1. George R. says:

    Eric Billings is now Chairman of Arlington Asset Investment Corp. and no longer affiliated with FBR Capital Markets Corp, or is AI connected to FBCM. In fact, most of the $165M in capital AI has invested in non-Agency RMBS came from Arlington’s total liquidation of its stock in FBCM. Arlington sports a book value a little north of $17.84 and closed today at $13.50. There are only 7.7M shares outstanding.

    • hhill51 says:

      Gets more interesting as we go. Thanks, George. Without knowing the precise timing of the purchases, I’m pretty comfortable estimating yield on invested capital at 15% or higher, with the potential of going twice that high. (Leaving the question of what they do for an encore if the first round of investments matures on the fast side.)

  2. Patrick says:

    Does it not seem like a bit of schadenfreude that Arlington’s overview slide in the presentation touts the company’s “straightforward, C-Corp structure”? Tell that to executives at HomeBanc, New Century, or American Home Mortgage.

    • hhill51 says:

      Good point. The opposite of the great Groucho Marx ditty: “Whatever it is, I’m against it.”
      In this case, whatever we are, that’s the best. Especially odd considering their prior life extolling the virtues of the REIT structure. Are you the same Patrick that got into the REITs and started putting those good independent comments on the web about five years back? If so, glad to have you reading. (And even if not.)

      • Patrick says:

        Yes I am one and the same. I’m flattered that you remember me. I’m very glad you’ve started up this blog – I’m a daily reader. So hard to find good information on mortgage REITs out in the blogosphere.

  3. Tom says:

    Good tutorial. Thank you. Your comment on subordination caught my eye. If half the loan pays off, then subord effectively doubles. Is that always true? My understanding was that different deals seemed to be different. Some times the equity tranch got paid off first in repays. Some times it was equally distributed to different tranches. Some times the AAA tranches got paid off first. Is that correct? Is one way more prevalent than others?? Thanks.


    • hhill51 says:

      Hi, Tom —

      You are correct that there are lots of variations (even including the arbitrary reversal of priority in some Goldman synthetic CDO deals.
      Among the various RMBS types, I suppose the subprimes were most uniform, conforming to the model I described, where early principal went exclusively to senior bonds, effectively increasing subordination. That form of deal, called “shifting interest” did that until month 37, at which point the mezz and junior tranches got all the principal, as long as delinquencies and losses were below key thresholds and subordination had doubled since the deal was initially priced.

      Alt-A deals used both shifting interest and parallel pay structures, but by 2006-2007, the shifting interest was more common. Those deals also tended to divide the “natural” AAA into senior/junior or even senior/mezz/jr structures. The problem with those turned out to be the fact that AAA payments are first determined for each month (principal and interest), and then divided among the tiers of AAA bonds. In effect, that allowed the mezz and junior AAA’s to be paid principal in parallel with the senior-most tranches during the first few years (some called this “leaking out” to the subordinated classes), which then set up even the super-senior tranches to take losses at the back end of the deals.

      Prime deals tended to have such small subordination below AAA that it hardly matters what was going on with the mezz and junior classes once the tidal wave got going. For them, the division of the original AAA like the alt-A AAA’s has come into focus, making investors shy away from even bonds that had initial subordination many times the expected losses.

      That’s why even pretty good 2005 vintage alt-A or prime RMBS that still have AAA ratings trade at pretty hefty discounts (20 to as many as 50 points), while a lot of subprime AAA bonds are trading surprisingly close to par (96 or 97 is not uncommon) if they are currently receiving principal in the “waterfall” of principal payments.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: