« Am I A Bigot? | Main | I'll Drink to That »

CDO's Explained

05 Dec 2007 05:44 pm

This here from Portfolio is probably the best brief explanation I've seen yet of what a "collateralized debt obligation" is and why it matters to you. It's also just a nifty application of web technology.

Share This

Comments (8)

That's actually one of the dumber descriptions of CDOs I've seen

CDOs are seldom made up of underlying AAA securities and the business of "making more buckets because there is so much money the buckets overflow" (I paraphrase) is inane

Great visuals, but it was designed by people who don't understand what they are looking at to be consumed by people who still don't understand it after reading/viewing. Like most of the rest of the financial press...

Actually it is right on. Makes sense and is visual for people that are not numbers orientated. Pretty innovative. Bravo Matt.

Sorry Jozef your arrogance blinds you to the truth and the truth shall set you free go forth and be humble for you look like a fool...

This article actually does a better job of explaining CDOs for grown-ups, and includes some helpful graphics: "The Panic of 2007" Chart 43 (it comes from a finance textbook -- there aren't that many charts in the article) explains the key point Jozef alluded to: how BBB-rated mortgage-backed securities were turned into AAA-rated CDO tranches.

For what it's worth, I didn't know exactly what CDO's were going in beyond a vague idea that they were "financial instruments tied to real estate", and that's my level of knowledge now. Don't know if this reflects more poorly on me or the link, though.

I'll try Fred's link and see if that helps.

Was it helpful for you, gz?

Yes. It also had the effect of making me appreciate the flow diagram in the graphic Matt linked to. Though, if I understand it, the real argument in your link is that it's not the dollar value of bad loans, but the bogus ratings which "contanimated" AAA rated bonds.


At least as of August, which seems like ancient history now.

gz,

The number of bad loans and the rating system that carved AAA-rated CDO tranches out of pools of BBB-rated mortgage-backed securities are related. The rationale for giving these tranches AAA ratings was that, since any defaults would be first absorbed by the lower-rated tranches, there would probably be no impact to the AAA-rated tranches, hence they deserved the highest credit quality. It appears that the defaults may be higher than expected, and there's a chance that there could be enough defaults to impact the AAA-rated tranches, so there is a lot less confidence in these ratings now. As of last week, the ABX index which tracks CDOs priced the 2006-vintage CDOs' BBB-rated tranches at 17.5 cents on the dollar and the AAA-rated tranches at 68.92 cents on the dollar.

It may turn out that this ABX sentiment was overly pessimistic -- that the defaults on mortgages won't be so high as to warrant such declines in the value of these CDOs. Ordinarily, the market would give you a good estimate of the extent of the damage. The problem is that the market for CDOs has pretty much shut down, since no one trusts the credit ratings and there is so much uncertainty about what anything is worth.

The way this impacts the availability of mortgage credit is that most mortgage lenders don't hold the loans they issue -- those loans are sold and securitized and become part of mortgage-backed securities, which are in turn bought by CDOs. Since the market for mortgage-backed securities has essentially dried up, a lot of smaller mortgage lenders have gone out of business; survivors are forced to tighten their lending standards to have any chance of unloading their new loans, and even then, are sometimes losing money on the securitizations.

Far from being "ancient history" (I don't think the structure of CDOs has changed since August), the article was prescient -- particularly in one of its prescription of "Hedge Funds to the Rescue". That may have started to happen, with Citadel swooping in to buy E-trade's mortgage-backed securities for something like 27 cents on the dollar last week.

The graphics are quite misleading, but the text in the animation is much better.

I'll see if I can give a simple description here (disclaimer - I'm a journalist covering mainly European securitisation).

The first graphic for RMBS is quite accurate (if generic). Mortgage backed securities pool thousands of mortgages. As repayments are made, they are diverted through what is known as a cashflow waterfall, which basically ensures that AAA tranches are repaid before lower rated tranches. Any defaults in the underlying mortgages are taken from the lower rated tranches first (after they've eaten through excess spread and the reserve fund).

Contra the graphic, CDOs don't get what's left over from the mortgage payments after the RMBS has been paid. Rather the cashflows come from the tranches of the RMBS themselves.

A CDO (in its simplest form), does exactly the same thing as an RMBS but with a pool of mortgage backed securities. The cashflow from the securities goes through the waterfall, paying off the senior tranches of the CDO first. Any defaults in the underlying RMBS are taken from the lower tranches first. Depending on the rating of the underlying securities, the AAA notes are supported by varying amounts of mezzanine tranches - typically 15% to 30%. So if you have a CDO backed by 50 securties, and five default (with no recoveries, for the sake of illustration), then the AAA, AA and maybe even A tranches should be fine.

The problems in CDO land stem mainly from the fact that correlation between (subprime) RMBS tranches has been much higher than anticipated. CDOs of RMBS would be fine with very high levels of losses in a small number of RMBS, but that's not what's happening. What's happening is a low (in absolute rather than historical terms) level of losses in a very wide range of RMBS. So even if few AAA RMBS tranches are going to be hit, many BBB and single-A tranches are. If you have a CDO backed entirely by BBB or worse tranches, then it turns out the credit enhancement doesn't do much good. It's a binary scenario - either nothing defaults or almost everything does, wiping out the whole CDO.

It's important to distinguish between CDOs of "high grade" RMBS, for which the credit enhancement does seem to be working because they are backed by AAA to single-A rated RMBS, and CDOs of mezzanine ABS, which are the worst affected. For instance, Citi's recently announced expected write downs for CDOs of ABS were roughly 60 cents on the dollar for CDOs of mezzanine but only 30 for high grade (and then only for high grade CDOs issued after 2005).

One thing to watch out for next is commercial real estate CDOs, which tend to be backed by BBB and BB CMBS and subordinated loans. Consequently in a widespread commercial property downturn they are likely to suffer the same problems as CDOs of mezzanine RMBS. That said, it is unlikely that a commercial property downturn will be as severe or as widespread as the US subprime crisis.


Comments closed December 19, 2007.

Copyright © 2008 by The Atlantic Monthly Group. All rights reserved.