Apr 24, 2011

A simple theory of efficient markets

Our starting point is an observation by a Professor Helbing  (not Helsing) in an interview with the FAZ. Here is the English translation:

On the question by the FAZ regarding the "intelligence of crowds" in bees and ants, Professor Helbing replies:

"In humans there is both, 'crowd intelligence,' and 'crowd madness.' We did some experiments recently where we asked subjects to provide estimates of "facts" of which they may not know much themselves (eg. the number of robberies in a given city). If subjects provide estimates independently of each other, the spread is large, but the mean is typically spot on. However, if subjects are being informed about the estimates of other subjects, they begin to rely on each others estimates. In the end, they agree, but the agreed-up value is often completely off the mark." (This phenomenon is know as groupthink in the Anglosaxon literature.)

Back to markets now. If we can generalize Helbing's results, markets participants will, on average, come up with prices that do reflect the underlying values efficiently, provided they are not relying on each other.

However, in reality, market participants do rely on each other to varying degrees---more in stress situations, less in calmer markets. This is crucial. In a crisis, market participants lose their bearings, group think takes over, and the resulting prices go off the charts. Market efficiency is lost. Eventually, the panic subsides, and relative efficiency is restored.

Our approach (others may have already said this more clearly, I don't know) eats the cake, and has it, too: In the long run, on average, markets are relatively efficient, but in crises they need not be. We use the semantics of "panic" to explain a behavior, which, via group think, and Helbing's results (I guess there are similar results out there from other workers) directly leads to inefficient prices.

George Soros, with his notion of reflexivity, may actually mean the same thing.

No comments: