Article #7129 in support of the "It's all bullshit" hypothesis

Rationality of markets my ass. From what I can see, the stock market is very loosely based on reality and ration decisions based on all available information. Mostly short term trends are dictated by the whim of a few large market movers and the self-fulfilling prophecies of their economic models.

Here's the thing: all of classical economics is based on the assumption that agents in an economic system are rational, meaning they make logical decisions based on all available data. There are two problems with this model. First, people aren't rational. If you don't believe this, look at the shit people buy on ebay. Large groups of people behave slightly more rationally, but they are still prone to herd mentality, and they're not completely rational besides. Bucking the conventional wisdom has a higher cost than making "irrational" decisions. Yes, there have been studies to this effect. And while "more" rational is in some sense better, it nonetheless fucks up game theoretic models nicely. Being just slightly willing to dick over someone else at the expense of your own payoff can completely change the outcome of a game (think about almost any game of Monopoly you've ever played...assuming you've ever actually finished a game). Also, some people (and even groups of people) are just plain stupid. For example, think about the game of rock, paper, scissors. It is very easy to see that the optimal "rational" strategy for this game is to pick your choice completely randomly with equal probability for each choice. This is provably so assuming rational opponents. However, when a tournament of computer programs was held, the winner was not a bot that used the random strategy. The strategy it used is a bit more complicated than I'd like to go into here, but basically it calculated the top few most likely strategies an opponent might take to try to game the system and generated counter-moves for them. Such a strategy is provably sub-optimal, but it's nonetheless brilliant. And it won the tournament. The optimal strategy wasn't optimal.

The other problem is the whole "all available information" requirement. There are always unknowns. In any real problem, there are always factors of the system that aren't known, or at the very least it is prohibitively expensive (in terms of money, time, people, etc.) to find out. It is, in fact, a stated desirable quality at many companies to be able to make quick decisions based on incomplete information. Bullshitting is coveted. Moreover, people evolved to be good at that kind of thing! The human brain is generally better than computers at games because it can make educated guesses based on context that a computer simply can't. So why is it surprising, then, that in any economic system the agents will take actions based on intuition that cannot be justified given the amount of information in their possession? Acting on intuition is what people do! Any macroscopic model of human behavior is useless if it does not incorporate this tendency into its model!

Oh, and then there's always the issue that anything you might want to study in economics is too terminally tangled in inter-dependencies that trying to isolate variables is a laughably doomed endeavor. Any statistical study of an economic system will likely say more about the presenter's ability to manipulate a set of numbers to show what he or she wants to than about the actual subject at hand, as far as I'm concerned.

Can you tell economists annoy me?

I've been wondering lately if it's possible to derive a model based on irrational and/or stupid and/or herd-following agents. Perhaps such a model could incorporate the fact that said agents use a rational agent model. Seems to me it should be possible to construct such a thing and game the system. I've heard there is an emerging field of economics called "behavioral economics," but I don't know much about it. I believe it deals with some of these issues, but I'm not sure. And I'm too lazy and too tired to google it.