Copulas and Cults

Felix Salmon has written an interesting piece on the copula model. Naturally the model does not come out well. He quotes me, accurately, as disapproving, to put it mildly, of models that depend on correlations between companies.
But that’s only part of the problem. Far more serious, because it extends to all of finance not just to a single model, is the poor education that people get in university financial engineering programs and also the blind-following-the-blind behaviour that is so common throughout the industry.

The copula model is not robust to changes in model assumptions. Black-Scholes is. Did you know that? Or maybe I’m wrong. Would you like to know the truth?

Yes, I could tell you. I could spoonfeed you. You’ve got used to being spoonfed, haven’t you? But you’re passing the buck there, putting an awful lot of responsibility on my shoulders. I can cope, as I’m sure David Li can cope. But you’re a big boy/girl now, you should be able to think for yourself. Isn’t that part of your job description?

It’s getting quite tedious me telling people to get off their backsides and test the models for themselves. Don’t believe anything I say, don’t believe anything Nassim, also quoted in the Salmon article, says. Question everything. Switch your brains back on.

In the late 1970s I had the dubious pleasure of attending a Billy Graham evangelical event for the followers of the christian cult, as a guest of some born-again nutters. Over the last decade I have had the equally dubious pleasure of attending many conferences on credit, listening to various academics, let’s say “Professor X,” for example, preaching about the copula cult. I use the word ‘cult’ in this context because of the similarities between the unthinking adoration I witnessed at both types of event. I found the Billy Graham event hilarious, I found the credit events disturbing. In both cases the audiences were intelligent people, in both cases there was only one non-sheep among them: me.

Paul Volcker recently spoke about financial engineering and it being, er, not quite what it’s made out to be. Setting aside the obvious discussion of what took him, and so many others, so long to realise this, he did make one point that matches my experience. And that is the use of the “Nuremberg Defence” by financial engineers, “Don’t blame me, I was only following orders.” I agree with Volcker that this is pathetic. And I’ve heard the same excuse from hundreds of people over the years, well before the recent crisis. I will be teaching people about the boundless possibilities of mathematical modelling, far beyond anything in most textbooks and certainly beyond anything in university programs, or I’ll be explaining the dangers of Value at Risk, etc. and the audience will almost invariably say that they’d like to try out the new ideas when they get back to their office…but…but…but they won’t be allowed to because the bank’s policy is to use ‘here insert name of stupid model that they’ve never properly tested.’

Gentlemen, and ladies, look in the bottom drawer, or wherever it is that you put them, and get out and dust off your cojones. Stand up, be counted, and stop bleating.