CDOs

[koganbot]

Good piece by Steven Pearlstein from last December that's helping me understand Collateral Debt Obligations better. ("Better" here is a relative term, however. Still wouldn't say that I understand them.)

It's Not 1929, But It's The Biggest Mess SinceNow, to truly understand I would have to know what was going on in this paragraph (Pearlstein has

Comments

[dubdobdee]

did you ever look at that econ diagram i linked to the one with the water and the buckets: i thought that got across how tranching works (and fails) quite vividly --- not sure if i can recall where i found it but it is linked by me in one of your earlier econ threads on lj

the point here is not that tranching actually changes the intrinsic value of the asset, rather that it changes its rating

the assumption made in the financial industry is (was) that by bundling low and high value assets, the overall value is pushed up ( low value asset is just as likely to fail -- ie fall on its own to a trigger-value -- but the combination is less likely to fall to ITS trigger value, bcz the high-value asset is thought NEVER to fall below a certain value)

a more cynical view was modelled by tom: it's like buying a bundle of singles, sight unseen except for the one on top (ie they're bagged up and you only get to see the to one) -- well in the second-hand record business, the top one is ALWAYS a high value one, the rest ALWAYS rubbish, but the finance milieu was such until recently that it was in everyone's interest to assume that, "statistically", a lot of the other singles are likely not to be terrible, because no one's looked, so plenty of them, by random process, must be pretty good

(it had not occurred to one and all -- or possibly it had but they busily repressed the worry - that an unscrutinised industry will attract more than the usual number of dodgy deals)

[koganbot]

Here are the buckets (though I think your verbal explanation is easier to understand).