Wednesday, September 7, 2011

Econophysics: an introduction

I have criticized a number of times Econophysics as a rather naive venture of physicists into Economics, where there is too much focus on "automatic" data exploration and too little use of theory and understanding of what the data measure. But may it is just my prejudice against and my ignorance of Econophysics.

B. G. Sharma, Sadhana Agrawal, Malti Sharma, D. P. Bisen and Ravi Sharma offer in six pages an account of what Econophysics is, what its goals are, what it can contribute and where it is headed. The basic idea is that economic agents are like particles in that they are in large numbers and interact in complex ways. The dynamics of such complex processes are studied with powerful statistical tools in Physics, and physicists think that this should also apply to Economics. The focus is very much on the stock market, probably because physicists have realized where money can be made. There is no sense that there would be an attempt to improve welfare. They are also much more likely to completely discard a model in one set of observations does not corroborate it. Physicists are especially critical of how economists stick to rejected dogmas and of their inability to explain how small shocks can pan out into large crises.

The focus is really on the description of data process and documenting there statistical properties. In particular, econophysicists want to find ways to exploit even the smallest opportunities for arbitrage by finding, often through obscure and complex black box processes, the right price of an asset at any moment in time. However, there is no attempt at understanding why these arbitrage opportunities arise, say because of some form of irrationality, asymmetric information or perverse interactions in the price mechanism. From this I conclude that Econophysics can be interesting to make money on the stock market, but at least at this point, does not help us in any way in understanding why the world is like it is. Which I find rather ironic for Physics.

4 comments:

Kansan said...

Why does it not surprise me that a paper written by physicists is only six pages long, does not develop any argument and needed five co-authors?

ivansml said...

I'm also somewhat skeptical about econophysics - although I can't say to have studied it in detail, the few papers I've seen were pretty lame. And the fixation on financial markets can be annoying, especially when accompanied by strong, yet unsupported claims about economics and economists in general.

Nevertheless, there are surely better reviews of the field than this one, which is basically just a rehash of general information from Wikipedia plus authors' uninformed commentary. For example, (although I haven't read the whole thing,) recent paper by Chakraborti et al. (arXiv:0909.1974) seems to provide much better review of actual research done by econophysicists.

Anonymous said...

Apparently both the invisible hand and market efficiency are axioms of economics. Ugh. I've read enough.

V.L Rendoumis said...

Being a Physicist and having read a lot of papers in the field of Econophysics, I do not agree with many of my colleagues who are looking into finance with a "the sky is the limit" attitude. Maybe theories in Economics are not that consistent and accurate but none can make new theories out of scratch. There is at least some truth. Another strange thing, though, is that most of our fellow Physicists, manage to publish only in some Physics journals and not in serious economic or finance journals. Well...there is a reason for that, relative to the structure of a "physics" paper and an "Economic" paper. I have talked much with researchers in Economics and most of them agree that for a paper to be published in an Economic journal there must be a good story supporting it. And by story I don't mean the actual idea for the modeling and the derivation of results. Sometimes, I fear, that this story is considered to have more impact than actual data. In contrast, a paper in Physics looks first at the data and than tries to make an appropriate story-theory. I agree, though, that a lot of papers in Econophysics do not provide any story. Maybe, in this context, I can answer to Kansan, why papers written by physicists are only six pages long. But, their effort is to analyzing the data an not linking them to some broadly known theory or story.