Responses to the EMH

This is a belated response to Gabriel's question,

What does the EHM has to do with Normal distributions?

The first thing to say is that I was retelling Cooper's argument, not my own. And secondly, it's hard to take Gabriel seriously. Are you arguing that EMH has nothing to do with Normal distributions, or that it doesn't necessarily have to? Do you want me to provide examples of investment banks that have used statistical models that combine (i) an assumption that asset prices are "correctly" priced; (ii) an underlying probability distribution that is Normal? Do you want me to point out that asset return distributions have fat tails?

Gabriel is better qualified to discuss the ins and outs of EMH than I am, and can thus allude to finer distinctions between necessary and sufficient conditions. But any answer to his question must be either blindingly obvious or incredibly deep and profound. You might find EMH models built on Tukey distributions, but that's not the point here. The issue is how has this financial crisis occurred, and a misguided faith in equilibrium theorising combines with Gaussian distributions surely played a major role.

For what it's worth, as the RBC crowd remain silent, we're left with a choice between two theories of bubble activity. The Minsky approach blames the inherent volatility of capitalist systems, and places faith in the omnisicience and benevolence of ignorant and self-interested politicians to solve them. The Hayekian approach explains how errors occur, and why there needn't be 'blame' per se, rather an explanation for how people are fooled when economic signals are disrupted. I find the latter more convincing.

Comments

[Gabriel]

Actually, my question was "for real" although the wording was unfortunate. My knowledge of finance is very much not there :-)

I know there are a lot of models used by financial institutions that use log-normal shocks (geometric Brownian motion) or the like.

On the other hand, if I am to make a point, it would be that this has little to do with the models and proofs, the theory of EMH. Although I might be wrong.

[aje]

I would love to see evidence of finance companies retaining the meat/principle of EMH but modelling under non-Normal conditions. Maybe they do, but my admittedly crude understanding is that the two come hand in hand. But I buy the counter argument that they needn't.