05 June 2005

Attention smarter people.

Interesting little piece in this week's Sunday Times Mag, by Stephen J. Dubner & Steven D. Levitt, about some capuchin monkeys that have been taught the concept of money. Oddly, the article never once uses the word "abstract," even though what's exciting about this experiment is that money is an abstraction and the ability to grasp and utilize abstractions is precisely what separates human beings from all other life forms on this planet. But most of all I'm befuddled by this here passage:

Chen next introduced a pair of gambling games and set out to determine which one the monkeys preferred. In the first game, the capuchin was given one grape and, dependent on a coin flip, either retained the original grape or won a bonus grape. In the second game, the capuchin started out owning the bonus grape and, once again dependent on a coin flip, either kept the two grapes or lost one. These two games are in fact the same gamble, with identical odds, but one is framed as a potential win and the other as a potential loss.

How did the capuchins react? They far preferred to take a gamble on the potential gain than the potential loss. This is not what an economics textbook would predict. The laws of economics state that these two gambles, because they represent such small stakes, should be treated equally.

So, does Chen's gambling experiment simply reveal the cognitive limitations of his small-brained subjects? Perhaps not. In similar experiments, it turns out that humans tend to make the same type of irrational decision at a nearly identical rate. Documenting this phenomenon, known as loss aversion, is what helped the psychologist Daniel Kahneman win a Nobel Prize in economics.

Am I missing something here? I spent like 20 minutes poring over these three paragraphs over dinner tonight, convinced that I must be overlooking some sneaky mathematical legerdemain of the kind exemplified by the Monty Hall Problem. But the subtlety, should it in fact exist, continues to elude me. So tell me, wonks: What's irrational about the decision arrived at by both monkey and man? In the first scenario, it makes perfect sense to "gamble," since you can only win—you're guaranteed one grape and have a 50% chance of getting another. In the second scenario, you already have the two grapes, and "gambling" can only halve your supply. There is no reason not to gamble in Scenario #1, and zero incentive to gamble in Scenario #2. And yet apparently "economics textbook[s]" maintain that the rational approach would be to treat them as equivalent. I'm open to correction, but my sense right now is that a poker table full of economists would offer a whole lot of dead money.

Speaking of which: Fuck. I was there at the time, found out how long you can keep your hands on your head before your arms start to get really tired. (About 25 minutes.) I've found another, much smaller club a bit further south, and am still NOT making a bundle; it's a blow all the same, though, as I'd developed a real camaraderie with many of the dealers and the other regulars. Let's get the game legalized, please. It is a contest of skill, involving merely an element of chance. (See also: every other competitive endeavor you can think of.)

No comments: