Use access key #2 to skip to page content.

A Quality CDO Analogy From Matt Taibbi



May 13, 2011 – Comments (7)

I think this is right on.


A Quality CDO Analogy From Matt Taibbi
 Posted by kid dynamite on May 13th, 2011

Matt Taibbi’s latest tome on Goldman Sachs is harsh enough that it actually played a roll in Dick Bove’s downgrade of $GS stock today.  I wanted to focus on a specific analogy Taibbi used to explain CDOs, because I think it’s an excellent one, although it could use some slight tweaks to make it even better.  Taibbi writes:

“A normal CDO is a giant pool of loans that are chopped up and layered into different “tranches”: the prime or AAA level, the BBB or “mezzanine” level, and finally the equity or “toxic waste” level. Banks had no trouble finding investors for the AAA pieces, which involve betting on the safest borrowers in the pool. And there were usually investors willing to make higher-odds bets on the crack addicts and no-documentation immigrants at the potentially lucrative bottom of the pool. But the unsexy BBB parts of the pool were hard to sell, and the banks didn’t want to be stuck holding all of these risky pieces. So what did they do? They took all the extra unsold pieces, threw them in a big box, and repeated the original “tranching” process all over again. What originally were all BBB pieces were diced up and divided anew — and, presto, you suddenly had new AAA securities and new toxic-waste securities.

A CDO, to begin with, is already a highly dubious tool for magically converting risky subprime mortgages into AAA investments. A CDO-squared doubles down on that lunacy, taking the waste products of the original process and converting them into AAA investments. This is kind of like taking all the kids who were picked last to play volleyball in every gym class of every public school in the state, throwing them in a new gym, and pretending that the first 10 kids picked are varsity-level players. Then you take all the unpicked kids left over from that process, throw them in a gym with similar kids from all 50 states, and call the first 10 kids picked All-Americans.”

This is almost perfect, although in reality we need a few minor adjustments:  it’s not really like taking the top 10 kids from the rejected unpicked batch from all 50 states and calling them “All-Americans,” – they don’t get even better, they’re just rated AAA too -  it’s like calling them varsity as well.  And the (flawed/faulty/crazy) logic behind it, with the mortgage CDO analogy, went something like “they’re all from different parts of the country, so they won’t all suck at the same time.”  Or, “They all suck in a different way, so they won’t all suck at the same time – that’s what our model says.”

In any case, I think Taibbi’s gym-class pick analogy is one of the best I’ve seen to describe CDO-squared to the layman.


7 Comments – Post Your Own

#1) On May 13, 2011 at 9:56 AM, kstarich (28.75) wrote:

Finally these guys are getting exposed.  Since the S&L crises they have gotten a pass. Will be a very interesting Summer and Fall.


Report this comment
#2) On May 13, 2011 at 10:33 AM, leohaas (30.08) wrote:

Excellent analogy.

It is immediately obvious that the label "All-American" is not appropriate. However, it was not immediately obvious that the "AAA" label stuck on may CDO tranches (especially the "squared" ones) was wrong. The investing world had a lable sticker for this purpose: the ratings agencies. Those ratings agencies were bought by the CDO issuers: the CDO issuers would pencil in an expected rating when seeking ratings on their CDO tranches, and not pay the ratings agency if it did not confirm the pencilled-in ratings. We all know what happened...

What bothers me the most is that these ratings agencies are still in business. I understand that there is no legal case against them (they will win a "Free Speech" defense any day in court), but why are we still listening to folks who got it all so incredibly wrong?

Report this comment
#3) On May 13, 2011 at 11:41 AM, binve (< 20) wrote:

kstarich ,




Report this comment
#4) On May 13, 2011 at 12:59 PM, MegaEurope (< 20) wrote:

In my opinion, the analogy fails because turning mortgages into multiple tranche CDOs involves chopping them up, not using them whole.  The volleyball team would be composed of random body parts, stitched together and reanimated by Coach Frankenstein.

Report this comment
#5) On May 13, 2011 at 3:27 PM, rfaramir (28.64) wrote:

The first failure was government involvement. Only agencies you are forced to use don’t care about their accuracy. The private economy actually cares about results and has a feedback mechanism for enforcing it, called profits and losses.

Second, when you don’t care about accuracy, you will choose models which are easy, especially in cases like these which are naturally hard to examine. They doubtless assumed normal distribution of failures.

Third, when you are a Keynesian, your model will use aggregates, no matter that you cannot just add people together, no matter that their behavior will not conform to ‘normal’ curves.

Lastly, the main cause of non-normal distribution is central planning (socialism). If the states were significantly different, you could reduce risk by diversifying across them, each its own independent experimental laboratory. But a centrally managed money supply blows bubbles across the nation; you cannot get away from it by moving to a different state nor by investing across states. The effects of increasing the money supply were not exactly uniform, but the Fed made it a universal phenomenon (well, national plus some exporting of it).

Report this comment
#6) On May 13, 2011 at 5:15 PM, leohaas (30.08) wrote:

#5: what were you on when you wrote your comment? I'd like to get some of that stuff...

CDOs (the topic of the post) have nothing to do with socialism. And nothing with Keynes.

The only valid points you make in the discussion is that ratings being required is part of the problem, and that these agencies assumend normal distributions. The analogy points out that in this assumption lies the problem.

Report this comment
#7) On May 13, 2011 at 6:24 PM, MegaEurope (< 20) wrote:

Aggregating economic data is what all economists do, it's not specific to Keynesians.

The few who refuse to add together data should really be called economic philosophers rather than economists.

Report this comment

Featured Broker Partners