[FoRK] "Common" delusions? was re: How Complex Systems Fail

Jeff Bone jbone at place.org
Sat Nov 7 20:39:38 PST 2009


Bill wrote:

> Perhaps (aforementioned delusion) as common as belief in 'efficient  
> market theory' & pixie dust.. :-)

And Ken responded:

> Believe it. :-) ...ken...

I, um, "know some folks" who work in applied economics and various  
financial markets.  ;-)  I don't actually know *anyone* who "believes"  
in "efficient market theory."  (More correctly, efficient market  
hypothesis... and in fact there is no such singular hypothesis, there  
are several;  they are expressions of the behaviors, limits and  
constraints of idealized markets under certain assumptions, which  
nobody actually claims are true of any particular actual markets.)

I think the common or at least popular delusion here is that anyone  
with a clue actually regards any efficient market hypothesis as being  
some realistic model of real markets.  Of *course* real markets aren't  
efficient;  they are incomplete, fragmented, non-transparent,  
asymmetric, have inconsistent internal and external rules, have lossy  
and laggy information paths, are often driven by outright incorrect  
information, and are very much "dispersed" in the very sense  
previously discussed. But the fact that they aren't efficient doesn't  
mean that the abstract model of market efficiency isn't useful;  it's  
just a limited model, a lossy compression of reality.  As such it's no  
more or less useful than, say, Newtonian mechanics;  it's a reasonable  
approximation of certain things under certain conditions, but with  
*known* limits.

Thinking that obvious inefficiencies in real markets somehow damningly  
"disprove" efficient market hypotheses is similar to the Kirk Cameron  
and friends' "eyeball" arguments against evolution;  i.e., it's only  
an argument that somebody who doesn't understand what others are  
really claiming would make, perhaps one that only somebody would make  
who doesn't understand what a hypothesis is in the first place would  
make.  (Actually, this isn't a very good analogy.  It's undoubtedly  
the case that early EMH-ers literally believed their hypotheses were  
higher-fidelity models of the phenomenon of interest than they  
actually are.  But in part, this was because they didn't have good  
theories of complex systems:  chaotic equilibria, information-, game-,  
and measure-theoretic views of things, networking queueing theory, and  
so on.  It is *precisely* through such things --- without invoking the  
spectre of "behavioral" economics or anything similar --- that we can  
understand the recent failures --- and potentially avoid future ones.   
If only we're reasonable enough, determined enough, and diligent  
enough to do so...  which seems unlikely in the extreme.)

Here's your homework:  consider why special relativity *proves* that,  
in the limit, efficient markets (in the EMH sense) aren't possible.   
Then consider whether this means that various EMH variants are  
therefore without useful application...


IMHO,


jb



More information about the FoRK mailing list