4 ms·

Matt Tiabbi does write with a certain muckraker-type style, and because it's only an article, it can only go so deep in the story. Having said that, as far as I

Matt Tiabbi does write with a certain muckraker-type style, and because it's only an article, it can only go so deep in the story. Having said that, as far as I can tell what he says is pretty much spot on and very much factually correct. One example: He spends a few sentences explaining how banks created CDOs: take some good loans, take some bad loans, put them together and using some math convince regulators and investors that these things are AAA rated. Wired wrote an article (itself not that detailed) explaining the formula that allowed them to do this: http://www.wired.com/techbiz/it/magazine/17-03/wp_quant http://www.wired.com/techbiz/it/magazine/17-03/wp_quant I don't think there can be an article that explains this stuff. Some years from now, someone who is both smart and an excellent writer will be able to look back and write a book that explains everything, but until then this stuff will be about as good as it gets. Details on "how to prevent" are even harder, and you'll probably have to look at the upcoming G20 summit in April where the Europeans will try to at least start talking about some kind of regulatory framework, while the Americans will just want everyone to throw more money at the same people that fucked things up.

anamax 18y ago
> One example: He spends a few sentences explaining how banks created CDOs: take some good loans, take some bad loans, put them together and using some math convince regulators and investors that these things are AAA rated. You write that like it's wrong or in error. As Google has demonstrated, you can create something that it is more reliable than its parts. A bundle of somewhat independent things has less variance than the pieces.
- petewarden 18y ago
  "somewhat independent" is the crux of the problem. In a crash the only things going up are the correlations. The modelling behind mortgage securities assumed that defaults were uncorrelated based on the short time period they based them on. Now we know better.
  - anamax 18y ago
    Misunderstanding correlations among defaults was part of it but the other big surprise was that different asset classes which less independent than thought. For example, no one expected an increase in mortage defaults to be accomplied by problems in stocks. In fact, its likely that these two things fed each other, that the unexpected correlation in defaults is at least partly due to the unforseen correlation across asset classes and the unforseen correlation across asset classes is at least partly due to the unexpected correlation in defaults.

anamax 18y ago

> One example: He spends a few sentences explaining how banks created CDOs: take some good loans, take some bad loans, put them together and using some math convince regulators and investors that these things are AAA rated. You write that like it's wrong or in error. As Google has demonstrated, you can create something that it is more reliable than its parts. A bundle of somewhat independent things has less variance than the pieces.

petewarden 18y ago
"somewhat independent" is the crux of the problem. In a crash the only things going up are the correlations. The modelling behind mortgage securities assumed that defaults were uncorrelated based on the short time period they based them on. Now we know better.
- anamax 18y ago
  Misunderstanding correlations among defaults was part of it but the other big surprise was that different asset classes which less independent than thought. For example, no one expected an increase in mortage defaults to be accomplied by problems in stocks. In fact, its likely that these two things fed each other, that the unexpected correlation in defaults is at least partly due to the unforseen correlation across asset classes and the unforseen correlation across asset classes is at least partly due to the unexpected correlation in defaults.