Monday 29 June 2009

Physics envy, History envy

Physics is in some ways the geekiest science. It's fundamental, it has hard maths in it, and it has had enormous success at explaining the phenomena it tries to study. What other subject can successfully predict something to twelve decimal places?

As a result, some practioners in other fields have physics envy. This is a notable problem for finance quants, many of whom didn't make it as academic physicists (or did make but didn't like the salaries). Indeed in retrospect one can make a case that one of the causes of the Credit Crunch was the collapse of the Soviet Union - the argument would go that the collapse freed up lots of highly trained mathematians and physicists, some of whom came to work for investment banks - no bulge bracket firm was without its Academy of Sciences prize winner; the geeks used used the maths that they knew, which was mostly stochastic calculus, to model things; these models were dangerous but not easy to falsify (because they were only really wrong in a crisis); so the industry used them and was subsequently screwed. In one way at least communism brought capitalism down with it.

Anyway, the desire to build highly mathematical models has in practice lead finance down a dangerous path. Perhaps the aspiration was good, but the implementation has been deeply flawed.

Let me instead propose a different aspiration. History envy. History is a lovely subject. There are lots of facts, but most historians ignore many of the relevant ones. They are interested in motivations, in causes, in the evolution of ideas. They want to understand the why as well as the what. A good history text is carefully argued and insightful. It provokes discussion, and casts fresh light on the present. It's not clearly wrong, given the evidence, but it can never be said to be right, either.

How much better would finance be if it took these desiderata? Abandon the spurious and misleading quest for quantification. Just try to make an interesting argument about why things happen.

Labels:

3 Comments:

Blogger Dave said...

This is an interesting post, but I don't think you have it quite right.

Speaking as a trained mathematician, I believe that all that the quants are seeking is to do interesting, high-powered maths (and, as you say, also get well paid for it). They are not motivated by envy, simply by the satisfaction in creating things of beauty.

The fault lies with their corporate bosses (who I would doubt have any yearning to be physicists) who are seduced by the spurious accuracy of the models (managers always love spurious accuracy). Conversely, they would be put off by the vagueness and ambivalence of a historical approach (managers always hate ambivalence).

So, the underlying problem lies not with mathematics or physics envy but with a dysfunctional managerial culture that always favours "results" over complex conjecture.

11:41 pm  
Blogger Cognitive Overload said...

Surely the mathematicians are themselves limited by their education. It's taken a while for the maths toolbox to catch up with theories that take account of the complex effects of human involvement in the things that they are modelling (or as someone in a slightly different field recently put it "yes, you're right - this doesn't work if you include subterfuge"), or to look at such a small part of the picture that the whole system becomes vulnerable.

At least when you've seen the hordes come over the border a few times, you know to look beyond the actual border.

And managers *can* be educated into understanding conjecture and longer-term/ wider-ranging results. Sure it makes them nervous at first, but put carefully in terms of what they could lose and how, they tend to get it eventually.

Me, I'm suffering from psychophysics envy...

7:30 pm  
Blogger David Murphy said...

The real category error that both parties - quants and managers - made was to think that a price is somehow an objective thing. It's a platonic fallacy. A derivatives price is not like the fine structure constant: you can't determine it arbitrarily accurately if you do enough work. Firms that are comfortable with uncertainty, and who can act despite not knowing all the answers (and knowing that they don't) always outperform the confident but wrong in the long run.

7:39 am  

Post a Comment

<< Home