The Rise and Fall of the Gaussian Copula Function

Felix Salmon has a terrific cover story in this month’s Wired on the all-too-seductive Gaussian Copula function, which made it easy to calculate risk on Wall Street. Take a big basket of financial instruments, each with their own risks, through them into the Copula function, and presto: you get the net risk. It was so beautiful and apparently bulletproof that investors began treating it like a black box. Then it all went wrong.

This story profiles David X. Li, the inventor of the formula (now living in China), and how its adoption led to the chronic mispricing of risk that toppled some of the largest banks.

It’s a hell of a story, for students of Wall Street and mathematics alike. Here’s the function, annotated (yes, we’re aware that it seems to be missing a parenthesis, but that’s in the original paper, too):

copula

Go to the feed source of this article
Go to publisher’s website

This article appeared on the website mentioned above and the author and/or the publisher are to be credited explicitly for the content. Alphaverse.com is not affiliated in any professional way with the publisher of this article and uses its content purely for education purposes.

Comments

  1. KrisBelucci says:

    Hi, good post. I have been wondering about this issue,so thanks for posting.

  2. cheers! In that case you also might want to check the article below and its technical appendix where NNT explains it a bit more eloquently, in his own way…:-) The appendix is the mathematical debunking of the Gaussian.

  3. THE FOURTH QUADRANT: A MAP OF THE LIMITS OF STATISTIC
  4. Technical Appendix and empirical data on the shortcomings of the Gaussian

Speak Your Mind

*