How a tricked out Gaussian copula function brought about the financial crisis

But Mr. Li cut that corner by using prices for credit default swaps as a proxy for the actual data.

This is basically equivalent to an actuary using the price of life insurance to guess your life expectancy. It works fine, until someone closes the loop.

If you read "The Greatest Trade Ever," one of the common themes is that the prices don't correlate the way they should--defaults don't affect CDS prices, for a while, then CDS prices overshoot, then defaults catch up.

This is pretty much what Nassim Taleb alluded to in "The Black Swan": relying on Gaussian-based models for phenomena that fall in "extremistan" instead of "mediocristan" when Mandelbrotian-based (more fractal like) is dangerous.

Looks like this has already been extensively discussed on HN: http://news.ycombinator.com/item?id=811864

I also remember this article from Wired last year.

http://www.wired.com/techbiz/it/magazine/17-03/wp_quant?curr...