Comments on: Journalism’s welcome longevity

By: djseattle

djseattle — Wed, 15 Feb 2012 22:27:45 +0000

Yglesias appears to have fallen for the lump-of-paper fallacy.

By: realist50

realist50 — Wed, 15 Feb 2012 21:05:08 +0000

I thought Felix’s Wired article was one of the most interesting pieces of financial journalism that I’ve read over the past several years, though I’d agree that the line quoted by dedalus is over the top. To me the broader point is that common sense and the wisdom gained from experience always have a place in assessing if a course of action is reasonable, especially so when one is relying on the output of a model that’s ultimately dependent on human behavior.

By: Curmudgeon

Curmudgeon — Wed, 15 Feb 2012 11:31:53 +0000

The fact of the matter is that we don’t know what content will stick around. A dozen years ago I wrote a series of articles that were only peripherally related to the technical publication in which they appeared. I wouldn’t call them works of art by any means, but one of them is still one of Google’s highest ranking search items related to me. I’ve had requests for copies and reprints from around the world, and it has been translated into at least four languages. Go figure.

By: Th.M

Th.M — Wed, 15 Feb 2012 10:27:51 +0000

I think this policy is a boon for authors as well, as this copula paper of yours for instance is how I became aware of your writing.

By: Danny_Black

Danny_Black — Wed, 15 Feb 2012 09:29:15 +0000

DCWright, unless you are liquid enough to outwait the predators.

By: dedalus

dedalus — Wed, 15 Feb 2012 06:50:50 +0000

I liked your essay but thought your Great Man Theory of history went embarrassingly overboard:

“Li’s Gaussian copula formula will go down in history as instrumental in causing the unfathomable losses that brought the world financial system to its knees.”

Hmmm, well, it’s pretty to think so.

By: spiffy76

spiffy76 — Wed, 15 Feb 2012 04:31:04 +0000

There has always been this tension in journalism. They used to say that yesterday’s news was only good for wrapping fish. That’s why they always charge so much to look up old articles in their archives.

By: DCWright

DCWright — Wed, 15 Feb 2012 00:55:11 +0000

I hope you didn’t pick that ridiculous title. (1) The underlying problem was the assumption of stable correlations, not the particular copula that was used to model them. (2) The fact that these models were not good for tail risk was very well known. People used different models and historical re-runs to try to quantify extreme events. (3) Different models might give different probabilities (all low) and different losses (all high) for extreme events, but in the end it doesn’t matter much. In a crisis, all correlations go to 1. If you are in a market when it collapses, you will loose everything, regardless of how good your risk models were.