I was reading an article in the June 2007 issue of Scientific American. The author, Kaushik Basu is a professor of economics at Cornell. The article is about game theory, which has always been one of my favorite topics.

Since I love the area, I was excited to see an article about it. I was disappointed with Basu’s analysis. He starts by describing a game called the Traveler’s Dilemma (TD), which is a variant (or perhaps a generalization) of the better known Prisoner’s Dilemma. Using some examples from experimental game theory to show how people are irrational, he comes to a rather strong conclusion. As he sets forth, his analysis “undermines both the libertarian idea that unrestrained selfishness is good for the economy and the game theoretic tenet that people will be selfish and rational.”

In one shockingly bad part of his article, he discusses an experiment where members of The Game Theory Society played a version of TD. To explain why he’s wrong, I first have to briefly explain the game. There are two players. Each chooses a number from 2 to 100. If each chooses the same number, then that’s what they get. But if they choose different numbers, then the person who chose the lower number (n) gets that amount plus 2 (n+2) while the other person gets n-2. So if each chooses 100, then each gets 100. If A chooses 100 and B chooses 99, then B gets 101 and A gets 97. In game theoretic analysis of this game, the “Nash Equilibrium” has both players choosing 2. In other words, game theory predicts that rational people playing the game would choose 2. The rationale is as follows: If A thinks B will choose 100, then A chooses 99 and gets 101 instead of 100. If B thinks A will choose 99, then B chooses 98. Yada yada yada, they end up at 2.

So, 51 members of The Game Theory Society play the game. Each chooses a number, and that number is played against the numbers of each of the other 50 players (I oversimplified this part a bit, but in a way that doesn’t matter). The person who gets the most points gets $20 times their average payoff. Only 3 chose 2, while most others chose a number from 95 to 100. Aha! They’re irrational. Even these sophisticated experts in game theory didn’t make the right choices.

Basu offers some explanations but misses the obvious one — A rational person playing in this experiment would not choose the number 2. It is reasonable to assume that at least one other player will choose a number higher than 2, and indeed that several players will choose numbers close to 100. If you choose 2, your average score will be somewhere between 2 and 4, and your payoff if you win will be between $40 and $80. If you choose 100, and you get it only three times out of 50, your average score will be 6 and if you win you’ll get $120. In the actual game the winner had an average score of 85, and earned $1700.

What really bothers me about the article is how Basu leaps from people not behaving “rationally” for a very unusual game to the conclusion that free markets don’t work. He ignores so many aspects of what economics understands as rational — he even asserts that altruism itself is irrational. Economics is broad enough to allow rational behavior to include altruism. Markets work even for people who aren’t motivated purely by greed. There has also been plenty of other work showing people do not always make the most rational decisions – Kahneman & Tversky come to mind.

I certainly agree that markets are imperfect. Some of my favorite bits in economics concern the concept of market failures, especially externalities (like pollution) and asymmetric information (such as Akerlof’s classic “Market for Lemons” paper). But the fact that markets are imperfect is well known in the field. Some economists have proposed solutions to some of the market imperfections, such as Pigouvian taxes for pollution (e.g. drivers do not bear the full cost of the pollution from their tailpipes, so gas taxes can be used to increase the cost of using gas to account for that).

Unfortunately, Basu fails to provide an alternative system that is better than markets. Socialists, communists, and others who want to intervene in markets can and probably will use analysis like Basu’s to attack free markets and free trade. They have an alternate approach in mind, and it is not something that will be good for the world – only for those who have power through the government. Yes, free markets may be imperfect, but they are far, far better than any alternative. As someone who purportedly understands all of this, Basu should be standing up and fighting for free markets, not providing ammunition to those who would replace them with government control.