“Markets don’t seek efficiency because investors aren’t rational.” Yeah, well, gas molecules aren’t rational either, but they obey very simple regularities in large numbers. Rational-expectations theory doesn’t actually require individual investors to be rational, it merely predicts that en masse they will behave as if they are.
I’ve made this analogy before and I think its good. Psychologists and economists, like physicists and chemists, study things that exist at much different levels of abstraction. An important implication of this is that behavior at the lower level doesn’t easily translate into behavior at the higher level. For example, the fact that individual molecules move in patterns that are called Brownian motion has little baring on the fact that when you put a few billion molecules of oxygen and a few billion molecules of hydrogen together you get a few billion molecules of water. Brownian motion is way neat, but if in some other universe it turned out that molecules just sat there laying still, it’s possible to imagine that it wouldn’t change the chemistry.
In fact, the chemistry was discovered before Brownian motion. In other words, we knew about the higher level behavior before we knew about the underlying mechanisms and this didn’t hamper our ability to do analysis at the higher level.
If this analogy works, then all this talk about animal spirits and psychology’s supposed obvious implications for economics is mis-guided. Its perfectly possible for individual humans to be irrational, idiotic twits and for the economic system that emerges from their individual actions to appear as though it is rational. If fact, the experiments of Vernon Smith show that this possibility is actually probable. In many experiments ((Here’s a counter-example; a paper on asset bubbles that shows price theory doesn’t seem to hold… asset price bubbles are common in the lab. Smith says his experiments show that the problem with theory is that it assumes common knowledge of rationality. It assumes people believe other people will do what’s in their best interest. Asset bubbles don’t arise in the lab after the same set of subjects repeat the experiment a few times and thus know that the others are rational.)), Smith shows the behavior of idiots, when aggregated, quickly converges to market outcomes that look like those outcomes predicted by neoclassical theory. In other words, behavior at the lower level of abstraction doesn’t necessarily impact behavior at the higher level of abstraction.
Here’s why, however, this might not be such a great analogy. In the case of physics and chemistry, the higher level abstraction, chemistry, is much more intuitive than the lower level abstraction, physics. We see chemistry every day; its very easy to create chemical reactions in your kitchen. The behavior and structure of atoms, on the other hand, is really strange and inaccessible. The solar system analogy helps, but atoms are still very strange places to think about. In the case of psychology and economics, on the other hand, the lower level abstraction is more accessible and the higher level abstraction is, well, more abstract. We all experience individual psychology and irrationality. Its so obvious. Economics — as I can attest to now having taught it for four years — is less obvious.