Hi, guys --
AHRPRD, the preferred stock of leveraged risky-commercial-real-esatate owning mortgage REIT Anthracite Capital Realty, yields more than 20%, and could therefore obviously double in price if things work out. If things don't work out, there's (squishy) downside protection so it'll probably outperform most long ideas.
I plan to touch very briefly on a bunch of stuff. I like to keep things short and informal, thus, "briefly"; this is complicated, thus "a bunch". Questions and objections are very welcome.
Just so everybody understands my limitations when it comes to the complicated parts: I work in structured finance and sit on the floor, but I don't deal directly with CMBS or commercial real estate for a living, and for a variety of reasons I haven't wandered over to the CMBS desk to pester them with questions. If any VIC members know some real CMBS guys, please send this to them because I'd love the feedback.
In other words, I know enough to be dangerous.
Here's my table of contents:
-- read 'em in any order you want.
- Commercial Real Estate Fundamentals
are thought to be reasonably healthy, on the supply-side anyway. Here, courtesy of the BEA via Calculated Risk, is a chart of non-residential construction as a per centage of GDP:
This is a really long time series, and obviously things can change over time so I don't read that much into it, but I'd describe it like this:
CRE construction fluctuated between 3.5% and 4.0% of GDP between 1960 and 1978. It then went on a huge tear, rising to almost 6% in 1982. It's been at or under 3.5% (the old bottom) since 1990. Industry insiders of course say that there's been no overbuilding this time, a claim made credible by eye-balling this chart. In addition, I'd suspect that some of the recent upturn is due to construction costs rising faster than general inflation.
Looking forward, commercial construction is of course headed down. Again courtesy of CalculatedRisk, here's the American Institute of Architects' Architecture Billings Index, a leading indicator of non-residential construction:
- Commercial Real Estate Mortgage
s are currently performing well, with CMBS 30 day delinquency levels at about 0.5% vs. the recent peak of ~1.5% in 2003. Delinquencies and losses are going to increase; the question is how far and how fast.
- Commercial Real Estate Securities
are in the crapper. Super-senior CMBS has recently been trading at LIBOR+400. This is insane. Let's step through a simplified example to see why.
SS CMBS has a 30% subordination level -- the deal has to realize 30% losses (NB, not 30% defaults) before the SS tranche takes a hit. At LIBOR+400, it's pretty much impossible to not receive a least LIBOR.
- $100 deal.
- You own the top $70.
- It goes for 10 years.
- Do whatever you want with the LIBOR component and stick the 400 bps spread under the matress. You wind up with $28.
So the deal needs to have realized losses of ($28 spread + $30 subordination) = $58 before you receive less than LIBOR. That's 100% default with 42% recovery. It's hard to see anyway for super-senior CMBS to underperform Treasuries at these levels because if defaults get that high the Revolutionary Council is going to repudiate the debts of the old regime anyway.
And this is where the safest and most liquid stuff is trading. Value investors might suspect that the riskier and less liquid securities are also trading cheap.
38% of par with a 22% yield is all the upside I need; what are the risks? Let's start with the income statement -- in particular, "Adjusted Net Interest Income". (God, I love this game!)
The most recent 10Q suggests (and I'd agree) removing the "commercial mortgage pools" from the calculation of net interest income. I'm going to go further (for calculating ratios only!) and treat the payments to the TRuPS as preferred dividends and chuck in $4MM from an equity investment in a real estate debt fund. We see:
$75MM Interest Income -
$37MM Interest Expense =
$42MM Net Interest Income
- 2.14X coverage, which is a lot. 50% of interest income has to disappear for AHR to be unable to meet interest payments. Preferred and TRuPS dividends are running about $8MM/quarter, so the coverage ratio there is ((37+8)=45)/75 = 1.67x, which is pretty healthy. The market is saying that income is going to evaporate at a rapid rate.
Let's look at the assets. AHR owns equity and some higher rated tranches from the following securitizations (HY stands for "High Yield"):
AHR I -- 2002
AHR II -- 2002
AHR III -- 2004
AHR HY IV -- 2004
AHR HY V -- 2005
AHR HY VI -- 2006
AHR Euro -- 2006
Outside the CDOs are a bunch of CMBS, $890MM purchase price/$540MM fair value, and $600MM of real estate loans. These are financed by ~$600MM of credit lines. In addition, the company has $315MM of term debt (closest maturity in 9 years) and $180MM of TRuPS. The debt and the assets outside CDOs more or less net out. There are some other assets kicking around, but let's throw the whole mess out and take a closer look at the CDOs.
Here are some cashflow projections for CDOs I-V:
Date Principal Interest Cash Flow
384,945,598 240,929,012 625,874,609
09/19/2008 0 0 0
10/15/2008 0 1,978,631 1,978,631
10/15/2009 0 31,123,497 31,123,497
10/15/2010 0 30,096,982 30,096,982
10/15/2011 0 27,554,734 27,554,734
10/15/2012 93 23,853,025 23,853,118
10/15/2013 67,740,482 21,898,332 89,638,814
10/15/2014 29,677,609 22,716,427 52,394,035
10/15/2015 21,732,651 15,077,406 36,810,057
10/15/2016 77,459,295 11,762,007 89,221,302
10/15/2017 21,664,810 10,783,712 32,448,523
10/15/2018 31,258,344 11,597,250 42,855,594
10/15/2019 68,902,253 8,708,967 77,611,220
10/15/2020 8,218,583 8,151,911 16,370,493
10/15/2021 9,935,325 7,464,368 17,399,693
10/15/2022 15,818,211 4,569,583 20,387,794
10/15/2023 27,204,023 1,965,363 29,169,386
10/15/2024 1,060,987 1,066,305 2,127,292
10/15/2025 893,242 223,062 1,116,304
10/15/2026 947,088 168,974 1,116,062
10/15/2027 1,004,629 111,853 1,116,482
10/15/2028 827,088 53,929 881,017
10/15/2029 600,886 2,695 603,581
-- That's about $619MM over the next 15 years. You need to haircut this by $50MM because there's one big loan from '98 ("The Senior Living Portfolio") that's been in bankruptcy since 2002 and is about to keel over. Call it $570MM, or on average $37MM/year.
The assumptions here are 0.9% default -- which translates into 1.8% 60+ delinquency -- with 60% recovery. The intent is to simulate the "Moderate Stress" scenario from Fitch's "Stress Testing: Expected Losses in US CMBS - 2006 & 2007", which can be found here:
There are tons of knobs and levers involved in this, and I might not be pressing and twiddling the right ones. Factors which tend towards conservatism are:
- Fitch is looking at '06 & '07, which are poor vintages. These CDOs contain a smattering of '05 stuff, but are mostly earlier with a heavy concentration of '98 deals.
- I'm cranking us up to 1.8% 60+ delinquency immediately. If we get there, it'll probably take a year or so.
- I'm assuming immediate liquidation on the loans. In reality, they'll linger in special servicing for -- ? -- 6-12 months.
I left out the Euro CDO because it's got a bunch B-piece loans which makes it unamenable to this kind of analysis, and the AHR VI because I'm not able to run it any more. For what it's worth, under slightly different assumptions, AHR VI PVs to $114MM at a 5% discount rate, and the Euro CDO is something like $50MM at 50% of par.
The main point is that there is substantial value in AHR's older vintage CDOs.
Here are some scenarios:
1: AHR goes belly up tomorrow, and you lose your entire investment. Oh, well!
2: AHR is forced to eliminate the preferred dividend. This might happen, but I think losses will be minimal because:
A: It'll take at least a couple of years, so you get to bank 40%, and
B: Distressed preferreds tend to bottom out around $5 purely due to option value. Say assets=$3B, liabilities=$3B, equity=$0. In theory the preferreds are worthless. But if you can purchase them at 20% of par -- $40MM market value in this case -- and the value of the assets increase by 2%/year for 5 years, you've got a 6-7 bagger on your hand.
3: AHR decides to eliminate the preferred dividend. They'll do this if it's not in the common's interest to receive dividends. I don't think it's in the common's interest to receive dividends right now, so this is a real possibility. The preffered might trade down, but I'd welcome this scenario, because given the opportunities available in the market, it would render scenarios 1 & 2 highly unlikely. At these prices, I'm more than willing to defer dividends for a couple of years to ensure ultimate success.
4: AHR staggers along with weak results. You get the coupon and some price improvement.
5: Things are more or less stable at current levels, and the D's go back to $20+.
Whoops, got to go to work, where I can't post; I'll be back with "OTHER THOUGHTS" later in the day.
Bailout may return some rationality to CMBS pricing.
|Entry||10/02/2008 04:58 AM|
|Can you talk about scenarios under which AHR would have to sell assets. You mentioned 600 mil of debt... Is that CDO debt? are there covenants.
I agree that super senior CMBS CDOs really ought to be worth par... but hey, didn't Merril just sell super seniors fo $.22?|
|Subject||Writeup Part II|
|Entry||10/05/2008 08:17 PM|
|Hi, guys --
I guess Monday wasn't the best day to recommend investing in leveraged risky commercial real estate mortgage companies. Sorry about that, and also for submitting this in two parts; these days, I do this when I can, not when I want to. The idea hasn't attracted much interest, probably because nobody could figure out what I was talking about. Maybe this addendum will help.
Here's some more BACKGROUND:
CMBS Classic -- pre-2005 -- was underwritten something like prime residential mortgages, where you lose money on a loan because, hey, stuff happens, like you build a convenience store and gas station and laundromat across the highway from the mall and then the mall burns down. Just like everywhere else in CreditLand, standards have declined over the last few years. The combination of looser underwriting, a weakening economy and loans that just shouldn't have been made, like the Riverton loan --
, make it unlikely that '05-'07 vintage CMBS will perform as well as earlier cohorts. But the earlier stuff is (fingers crossed here) pretty solid. Realized losses to date are less than you'd expect from investement grade credits. What that means is that CMBS-style commercial mortgage loans have been better than BBB corporates for the last 10 years.
AHR owns the most subordinated part of the CMBS -- the "B-piece", which goes from the unrated residual tranche up to the BB+ level. The B-piece ranges from 2-6% of the total capital in the deal. B-piece investing is harder and riskier than buying subordinated tranches of, say, RMBS, because CMBS loans sizes are heterogeneous -- the three biggest loans could be 30% of the deal. If one of those sourers the B-piece could be wiped out. Special exprertise in real estate and securitization are requirements for success. This niche _seems_ like it should produce superior returns over the cycle, but there's no evidence of that from the stock prices of publicly traded participants.
The B-piece is sold on a discounted basis. The buyer gets a high coupon to start with and if things work out a capital gain at the end. The B-piece is also the "controlling class", the beneficial owner if you will, able to direct the course of work-outs, etc. This is the opposite of CDOs, where the most senior tranche is the controlling class.
APPROPRIATE LEVERAGE for Commercial Real Estate Mortgages:
If some dude wants to buy an apartment building and talks the bank into letting him borrow at 75% LTV, and the current rent roll is big enough to cover the expenses and pay off the loan and maybe deliver a decent cash return, nobody's going to call that nuts. Maybe it's a good deal, maybe it isn't, but 3:1 leverage doesn't sound inappropiate for commercial RE equity. If 3:1's OK for the equity, then, jeeze, you've got to at least add one turn for the mortgagor. So 4:1 is an appropriate amount of leverage to apply to a portfolio of first-interest CRE loans.
The assets can be put into THREE HOUSES:
- Brick. These are the equity and a couple of other tranches from CDOs I, II and III. AHR went public in 1998 and so owns a bunch on '98 subordinated CMBS. CMBS Classic loans had a 10 year maturity. 1998 -- 2008. It's 10 years later. Those 10 year loans have (at least so far) been able to refinance or pay off or get extended. I think I, II and III are in aggregate worth around par -- of course they couldn't be sold for that, or anywhere close, but that's what I think they're worth. It'd take more than a huff and a puff to blow these down.
- Wood. CDO IV and V are (mostly) the subordinated tranches backing up I, II and III. Because of the seasoning, it would take at least a huff and a puff to blow these down.
- Straw. This would be everything else. I could blow it down, and I smoke. But there's an awful lot of it.
The cashflow projection in my writeup is brick mixed with a declining amount of wood. There's no straw there.
LIMITATIONS of the CASHFLOW PROJECTIONS:
I was trying to present a scenario where 9% of outstanding CMBS loans default over the next ten years. You might wonder, "What's to prevent 9% defaulting over the next two years?" Nothing.
I haven't looked at AHR's older loans one-by-one. I scoped out the biggest ones, and for the rest am relying on using loss numbers twice what management assumes.
GAAP is USELESS
Just throw it away. GAAP is broken for CDO issuers to begin with, and they're using fair value for all the assets and liabilities. A decline in the market value of their debt means a profit. This is kind of good news, because it's practically impossible for them to be insolvent under GAAP, and really hard for them to show a loss.
OPERATING INCOME ACCOUNTING
Operating income relies heavily, surprise! on management estimates. My impression is that for the sub-prime-bought-at-a-discount CMBS part, they take a guess at realized losses and then recognize income based on an IRR calculation -- roughly, they bought some '07 stuff at 50% of par for a 10% coupon, expect to get the principal back, and thus will report 7.5% or par as income.
But, whoops, now it's bed-time -- I _will_ finish this write-up, even if it kills me!
|Entry||10/05/2008 08:27 PM|
|Hi, Doob --
Sorry, I can't blame you for being confused; this is a work in progress.
> Can you talk about scenarios under which AHR would have to sell assets.
They've got some market-value based credit. Market value is going down fast. In addition, these lines are being down-sized. They might be forced to toss some stuff overboard.
>I agree that super senior CMBS CDOs really ought to be worth par... but hey, didn`t Merril just sell super seniors fo $.22?
I don't know if recent senior CMBS should be trading at par ... is it _really_ AAA? But at LIBOR+400, it doesn't matter if it's really AAA or not. The stuff that Merrill sold/gave away/hid/whatever is completely different. These were (I'd imagine) mostly super-senior mezzanine ABS CDOs, or in other words, repackaged BBB tranches of sub-prime mortgages.
|Subject||RE: RE: Confused|
|Entry||10/05/2008 08:44 PM|
|So your theory is that Merril pooled the Mezz pieces of various pools of subprime mortgages into a structured ABS CDO trust, in which they retained the Senior most floaters, which they then repo financed with the expectation to hold to maturity, but when the haircuts escalated they decided to seller finance them to a more stable counter-party (Lone Star), rather than default on their repurchase obligation and lose their equity in the senior piece of the ABS-CDO pool of mezz pieces of the subprime whole-loan pools?
Wow, it makes so much sense now!|
|Entry||03/04/2009 11:56 PM|
I'd love to hear your current take on this idea, Bowd. Buying opportunity or avoid? TIA