by Jurgen Brauer, March 2006
Copyright: J. Brauer. No reproduction without permission.

We know from our consumption behavior that we do not always rationally calculate benefits and costs before engaging in a purchasing decision, nor do we make rationally calculated decisions on which job to take, whom to marry, or how to reward or punish our children. Instead, often we go on hear-say, gut feelings, habits and tradition, or impulsive actions, especially for complex decisions.

Economists are intensely interested in people's decision-making, as are other social scientists. But until recently economists did not possess laboratory tools to observe and interpret individual decision-making. By default, we therefore observed average behavior, and on average it turned out that people acted "as if" they were rational. For example, when prices fall, consumers on average increase the quantity demanded on the item in question. But just because the price of a gallon of milk falls does not imply that every consumer now buys more milk than before. Yet so long as the average consumer bought more, that was good enough for economic theory and for certain predictions the theory makes.

We used to observe average behavior, and on average it turned out that people acted "as if" they were rational. Now we can study decision-making in the laboratory.

Recent advances, however, now permit scientists to observe individual decision-making behavior, usually by means of clever laboratory experiments with human subjects who sometimes are even hooked up to functional magnetic resonance imaging (fMRI) equipment that monitors brain activity in "real time" as decisions are made.

The advance in the laboratory now also permits economists to study a long-standing question: can it pay to behave irrationally in the market place? The fascinating answer is "yes, it can." In a slew of recent articles, experimental economists report that a minority of irrational participants in a laboratory experiment can sway a majority of rational decision-makers.

If I expect others to push the market irrationally upward, then it becomes rational for me to invest more as well even as I know that the market will eventually collapse.

Consider a stock market game in which stock prices have increased drastically over time. The parameters of the game are such that fairly complex calculations are necessary to determine whether or not to keep pumping one's (play) money into the market. Only a few of the randomly chosen experimental subjects will per chance have the ability to make the required sorts of rational calculations. Most subjects will make their decisions irrationally, on the basis of criteria such as a "rising tide lifts all boats." They pump money into the game market, average stock values rise, and an artificial stock market "bubble" develops. In contrast, the rational players recognize that the numbers don't work out. Rationally speaking, they have an incentive not to put more money in the market. And yet, the experiments show, they also put more money in, on the expectation of what the other (irrational) players are going to do. So if the market stands at an index of 5,000, and I expect others to push it irrationally to 6,000 then it becomes rational for me to invest more as well even as I know that the market will eventually collapse because I calculated that the good fortunes cannot last. All I need to do is to jump out of the market just in time. This is an example of strategic complementarity. Whereas my own rational calculation would tell me to sell stocks, strategically it may be better to complement the irrational players' actions (i.e., to do what they do), at least for awhile.

Note that I switched the question. From "can it pay to be irrational?" to "can it pay to behave irrationally?" It almost never pays to be irrational, but it can pay to behave irrationally.

Experiments have also shown the opposite result, that of strategic substitution, the case of doing the opposite of what others are doing. In this case, a minority of rational players can compel a majority of irrational players to switch their decisions. Consider betting on the outcome of a sports game and suppose that there are a few well-informed and many poorly-informed traders in the market. If the market as a whole under or overpredicts the point-spread in the game, the well-informed traders can make money as they bet in the opposite direction of the poorly-informed traders. Thus, the influence of the poorly-informed participants is counteracted, instead of being amplified as in the case of strategic complementarity.

This sort of evidence is exciting because for the first time economists and other scientists are getting an experimental handle on certain aspects of the notions of rationality and irrationality. This permits us to explain a broader class of empirically observed phenomena under a unified theory - always the goal of theory.