Drunken Ultimatums

   The so-called ultimatum game contains a world of psychological and economic mysteries. In a laboratory setting, one person is given an allotment of money (say, $100) and instructed to offer a second person a portion. If the second player says yes to the offer, both keep the cash. If the second player says no, both walk away with nothing.

The rational move in any single game is for the second person to take whatever is offered. (It's more than he came in with.) But in fact, most people reject offers of less than 30 percent of the total, punishing offers they perceive as unfair. Why?

The academic debate boils down to two competing explanations. On one hand, players might be strategically suppressing their self-interest, turning down cash now in the hope that if there are future games, the "proposer" will make better offers. On the other hand, players might simply be lashing out in anger.

The researchers Carey Morewedge and Tamar Krishnamurti, of Carnegie Mellon University, and Dan Ariely, of Duke, recently tested the competing explanations — by exploring how drunken people played the game.

As described in a working paper now under peer review, Morewedge and Krishnamurti took a "data truck" to a strip of bars on the South Side of Pittsburgh (where participants were "often at a level of intoxication that is greater than is ethical to induce") and also did controlled testing, in labs, of people randomly selected to get drunk.

The scholars were interested in drunkenness because intoxication, as other social-science experiments have shown, doesn't fuzz up judgment so much as cause the drinker to overly focus on the most prominent cue in his environment. If the long-term-strategy hypothesis were true, drunken players would be more inclined to accept any amount of cash. (Money on the table generates more-visceral responses than long-term goals do.) If the anger/revenge theory were true, however, drunken players would become less likely to accept low offers: raw anger would trump money-lust.

In both setups, drunken players were less likely than their sober peers to accept offers of less than 50 percent of the total. The finding suggests, the authors said, that the principal impulse driving subjects was a wish for revenge. CHRISTOPHER SHEA

Great short piece from the New York Times showing how logical economic behavior is far from real economic behavior. Economics is a long way from being a science, but it is the closest we have.