Friday, July 06, 2007

Dangerous to your hedge fund, Commander, not to this portfolio

Are AAA CDOs trading at 30 cents on the dollar? Are AAA-rated CDOs much riskier than other AAA assets? Are AAA CDOs getting downgraded? Is the CDO market falling apart?

No, maybe but not by much, no, and no.

There has been a lot of misreporting about the CDO market, much of which involves lumping all types of CDOs together, when the problems are really concentrated in a single area. The following is a graph of AAA CDO spreads for the last 6-months. Without looking at the legend, you should see that a specific area of the CDO market is in trouble...

For background, please read this and this. I want to first point out that in order for the most senior tranche in most CDO deals to default, a hell of a lot has to happen. High-yield Collateralized Loan Obligations (or CLOs, which is the green line above) normally have 25% or so of subordination. So in order for the AAA senior tranche to fail, at least 25% of the collateral has to default.

But wait, there's more!

You assume you'll recover something in bankruptcy, so defaults have to be even higher.

But wait, there's more!

The deal has triggers, which divert cash away from junior tranches for the benefit of senior tranches. So when defaults start rising, interest that would have gone to a junior tranche is actually used to pay down the AAA tranche.

The fact is that AAA rated CDOs with no sub-prime exposure are showing no sign of distress.

Now, let's turn to ABS deals, which is where the sub-prime problem is. There are two basic types: High-grade (red and yellow lines above) which invest in investment grade pieces of ABS. Most of these deals I've seen have average ratings of A. The other type is Mezzanine (or Mezz, blue line above). These invest in BBB and BB ABS pieces.

Remember that a CDO of ABS is structure on top of structure. By this I mean, ABS are often tranched, where there are junior and senior bonds. So let's assume you build a portfolio of ABS, all of which are junior tranches rated BBB. Then you make a CDO out of that portfolio. The junior tranche of your CDO is backed by junior tranches of a ABS deal. Structure on top of structure.

Both the CDO structure and the ABS structure cause defaults to hit your junior tranches first and your senior tranches only after the junior tranches have been wiped out. So let's say that 10% defaults wipes out a BBB-rated ABS tranche. If that's held in a CDO portfolio, that's a defaulted bond in the CDO. What if the same thing happens to all the BBB-rated CDOs within the CDO? Suddenly 10% sub-prime defaults nationwide turns into 100% defaults in the CDO!

Now, that's not going to happen to many CDO deals, because few are actually 100% sub-prime RMBS, and you'd never see exactly the same defaults in every ABS piece you hold. So no one is going to suffer 100% defaults. But this does serve as an example of how defaults get amplified when you have structure on top of structure.

Consider how different the situation is if the CDO owns all A-rated bonds. Maybe those need 20% underlying defaults before they fail. While 10% might seem pretty reasonable, 20% is far less likely. This is why the senior AAA tranche in high-grade structured finance CDOs are performing just fine, while the junior AAA tranche isn't.

Presently few (if any) AAA CDOs are getting downgraded in the wake of this sub-prime mess, but the market is telling you that many will be. Whether defaults actually turn out to be bad enough to sink any AAA tranches remains to be seen.


Anonymous said...

Excellent blog and well written!

I think your key phrase is
Maybe those need 20% underlying defaults before they fail. While 10% might seem pretty reasonable, 20% is far less likely

I think the market shares your conclusions for the time being.

However, simulations that I've done show that if sentiment shifts to '20% is pretty reasonable and 30% is far less likely' -- you will see signficant price erosion on the senior tranches of CDOs.

The price relationship to expected losses is decidely non-linear.

It is the same with any short option trade. You can get out while a little ahead and leave money on the table or lose everything in the wink of an eye.

David Merkel said...

Great blog. I've seen in my lifetime AAA CDOs, Franchise loan ABS and Manufactured Hosing ABS go into default. Recoveries were pretty good as defaults go, but the market prices of those instruments declined tremendously because they could only really be held by investors with a strong capital base that didn't care about short-term reputational risk. That's a small segment of the total fixed income buyer universe.

You're correct that losses to AAA CDOs are unlikely, but many investors will have to live with mark-to-market implications and additional capital charges if downgrades take place.


Accrued Interest said...

Certainly there are many holders of AAA ABS CDOs who have seen their bonds widen 50-75bps and aren't too happy. And I'm sure some deals will see the AAA come close to losing money. I just think most of them will hold together.

There are really two interesting questions. One is about sub-prime performance generally. But as someone who doesn't hold any such paper, I'm more interested in the contagion. That's why I'm watching CLO spreads so closely.

Anonymous said...
This comment has been removed by a blog administrator.
Anonymous said...

all i hear constantly is that this stuff is small potatoes.........and the bears always agree, but counter with - "these funds are levered up 10 to 20x1, thats the problem"

that may be the case, but not everybody is piggybacking like they were to LTCM in 98 - which is what everyone is waiting for. there are many who are in the same trade, for sure, but i believe the contagion will be mild. sure its sobering up BX and KKR, maybe even some banks are cooling it a bit. if they are, and i sure hope so, this will be a fantastic buying opportunity.

this blog rocks plain and simple.

Accrued Interest said...

I really think the contagion will be temporary, but if it is larger, its a great chance to get long risk.

Latin in London said...

Excellent blog! Congratulations.

I would like to ask you a question.

You wrote:

"..Both the CDO structure and the ABS structure cause defaults to hit your junior tranches first and your senior tranches only after the junior tranches have been wiped out.."

If a tranch in a CDO is defaulted, let us say the junior one, but still got to pay the more seniors, can you say that the CDO defaulted??

Because otherwise I do not understand why do you say (in you e.g) that a 10% default in the ABS Junior Tranch, that clearly defaults the CDO Junior Tranch, implies a default of the CDO overall.

I would appreciate your answer.

Thank you for your time

Accrued Interest said...

Thanks for the kind words. Sorry I haven't posted yet this week.

I honestly don't know if its generally said that a CDO deal has defaulted if just one of the tranches is in default. When I was writing I was using the term "default" on a tranche specific basis.

Anyway, here is what can happen.

Let's say you own 100 different Baa2-rated sub-prime RMBS deals. Let's say any of your RMBS pools can only withstand 8% defaults before its out of subordination. 10% therefore puts any one of your deals in default.

Now, let's say that nationwide, sub-prime loans default at a 10% rate. Well there is a pretty good chance that most of your RMBS tranches will default. Now your whole CDO, from top to bottom, is in trouble.

Anonymous said...

I like the news, even know I have just read the news now, I became interested to know what is a Hedge fund and what it really means.