Wait, Wait...Just Say Gaussian Copula Base Correlation

I finally got around to reading this paper on models & the Gaussian copula that I’d seen discussed on Dealbreaker and elsewhere. At some point while reading I realized there was basically only one fundamental observation I was looking for in it, a subconscious litmus test I had for whether the paper was insightful or rubbish, and finally around p. 44 I saw it:

…the fact that the Gaussian copula base correlation model was a market standard provided a considerable incentive to keep using it, because it avoided having to persuade accountants and auditors within the bank and auditors outside it of the virtues of a different approach.

[...]

…If a trader started to use a model substantially different from the Gaussian copula, then a position that was properly hedged on that model would not look properly hedged to risk controllers, unless the trader could succeed in the difficult task of persuading them to stop using the market-standard model.

In short: Sometimes, models are a way of keeping people off your back.

‘Everyone’ decides and ‘knows’ that such and such model is the ‘market standard’, so as a result to fight the consensus would be such an uphill campaign and so disruptive to daily business that it’s just not worth it. Just go with the consensus ‘right answer’.

I am reminded of the Simpsons episode “Much Apu About Nothing” in which Apu is getting his US citizenship:

Proctor: All right, here's your last question. What was the cause of
the Civil War?
Apu: Actually, there were numerous causes. Aside from the obvious
schism between the abolitionists and the anti-abolitionists,
there were economic factors, both domestic and inter--
Proctor: Wait, wait... just say slavery.
Apu: Slavery it is, sir.

Good paper.

About these ads