Wednesday, September 24, 2008

Hand me down quote, and hand me the explanation

Some time ago, I blogged: From Kamal Jain to Christos Papadimitriou and via Moni Naor comes this modern take on the Turing test for the crowd:
If your laptop can't find it, neither can the market.
What did Kamal mean? He responds: A "rational" entity in theoretical economics is nothing but a Turing machine. So this quote basically means that at least one of the following must be false: Entities are rational; Economics system reach equilibrium efficiently; Equilibrium is not hard to compute by a Turing machine; In practice generic cases do not appear.If computer scientists have some evidence that computing Nash equilibrium is hard, then it puts the ball in economists court. So in theory economists should convey at least one of the following message:
  1. Entities are not fully rational, and therefore it is ian interesting question to study their irrationality. Though modeling would be hard, because a modeled irrationality may again look like a Turing machine (bounded rationality).
  2. Economics system may not reach equilibrium. If economics system converge towards theoretical equilibrium, then that again makes the market dynamics a computational model.
  3. Provide novel computational insights to computer scientists! Because in principle, one can simulate a set of rational entities by a bunch of laptops.
  4. Claim that the realistic situations are not generic. Well, in that case it is intersting to model the realistic situation. That again brings us to a computational model.
All in all, computational models are not forever ignorable to study economics systems. Or in some sense, algorithmic game theory is more realistic than algorithm-less game theory.


Post a Comment

<< Home