Assorted links


Re the Tierney piece... A lot of the behavioral effects wash out when you look at big groups over long periods of time. That's why the rational hypothesis works. A lot of confustion (such as the efficient market hypothesis) arises from misinterpreting a hypothesis tested only against group behavior as a hypothesis of psychology. Psychology and economics should eventually produce consistent predictions. But psychologists are speculating when they say groups are irrational the same way that Fama and French are speculating when they say no individual can beat the market over time.

Why do people insist on claiming that The Ultimatum Game (et al.) tell us anything about whether humans are rational? ALL DELIBERATE BEHAVIOR IS RATIONAL. If I reject the other players offer in The Impunity Game, it does not mean I am irrational. It means I place some expressive value on rejecting the offer, or derive value in some other way from refusing my share of the cash. Those invisible, impossible to measure values are every bit as real as the value of money. Journalists and economists both must absolutely cease to state that choosing non-material value over material value is irrational.

Those invisible, impossible to measure values are every bit as real as the value of money.

Right, but those 'invisible values' are not impossible to measure, but doing so might be costly to the experimenter. So it's not really surprising that a subject would reject an offer of $1 out of $10 because, presumably, the self-respect value of rejecting mistreatment is greater than a paltry $1. But now let's change the experiment, and the offer is a $100,000 share of $1,000,000. Now -- how many of the same people would reject *that* offer? Far, far fewer I would predict--especially in the 'Impunity' version of the game.

Raise the stakes high enough so that the amounts overwhelm the subjective valuation of unfairness and I would be people would start looking a *lot* more rational in accepting unfair offers.

Noah, I couldn't agree more. When I was learning economics in college (at GMU), I was told that economics makes no claims on what people want to maximize, just that they do want to maximize. That's why we have a word like "utility" that doesn't refer to anything specific. Also, I was taught that "rational" simply means: If someone prefers A over B, then, all else being equal, that person will choose A over B.

Also, I'd like to know how much money is involved in these impunity games. In the ultimatum games, rejection of the offer quickly decreased as relative values increased.

"This is a humpty-dumpty style argument - you want 'rational' to mean exactly what you say, common usage be damned."

I don't think there is a common usage. If there is, it's probably something like: "If I understand his choice, he must be rational. If I don't understand his choice, he must be irrational."

Re: #5, what if the experiment was set up slightly different. Instead of choosing between shrimp, mullet, etc. You have an option of choosing between two different kinds of shrimp or even oysters. Certainly there are going to be some foodies that prefer oysters from X waters instead of Y waters, but I can't imagine that many. They alter the price of the two different kinds of oysters and see the response.

Of course people have a preference between two completely different foods. Some people don't like peanut butter and chocolate, but prefer caramel. If I have a choice between two caramel chocolates and one costs more than the other, I am more likely to think it is premium and, therefore, choose that one. This is one reason why luxury/premium brands exist (or any brand for that matter)...people think they are better even if they are not.


In the ultimatum game, rejection of unfair offers has been observed even when the stakes were equivalent to several months' wages.

Although the Ars Technica article doesn't mention it, the "impunity game" researchers and participants (see original paper) were Japanese. Cultural factors may have come into play (conformity vs. individuality, group loyalty vs. selfishness, "the nail that sticks out gets hammered", etc).

Also, age may have been a factor, since the game participants were undergraduate university students. As street gang and hip hop culture shows, young men in particular have an aversion to being "disrespected" and sometimes react in impulsive or even violent ways, even if this may have profoundly negative or self-destructive consequences. Unfortunately, the researchers didn't provide any breakdown by sex of whether young men were more likely to reject the offers than young women.

Also, the amount of money being divided was only 1000 yen (a bit more than $10), and in two of the three studies involved, participants received 300 yen and 700 yen respectively just for showing up. So a purely symbolic rejection didn't involve any real sacrifice.

Here are the instructions given to responders in the "private impunity game", copied and pasted from the original paper (presumably, this is a translation from the original Japanese):

Part One: "A decides how to divide the 1,000 yen that has been given to the two of you. A gets the amount of money he/she has allocated him/herself, and the experiment is over. B receives the money allocated by A."

Part Two: "Although we have not told you in the previous instructions, you will also make a decision. You have to decide whether to ‘accept’ or ‘reject’ the decision made by your partner. If you choose to ‘accept’ your partner’s decision, you will receive the amount of money allocated by him/her. If you choose to ‘reject’ your partner’s decision, you will not receive the money allocated by him/her. (That is, you will earn nothing.) Your partner does not know that you have a choice between ‘accept’ and ‘reject.’ He/she decides how to allocate the 1,000 yen without knowing whether you will ‘accept’ or ‘reject’ his/her decision."

This is unnecessarily convoluted and indirect. Part Two partly overrides and contradicts Part One (the experiment is not over, and B does not necessarily receive the money allocated by A), so a less than careful reading (focusing on the final sentence in particular) might construe this to be an ultimatum game, with "reject" interpreted as rejecting the entirety of A's allocation. Based on its opening sentence, Part Two is very likely on a separate sheet of paper, but nowhere mentions that A gets the money regardless of rejection (this is implied only by omission).

Aside from a few genes, I don't know that I possess anything from Sweden.

Tony Coen:

It was an editorial, not a research paper. It was notable because it comes from a left-of-center magazine that adores Swedish social democracy.

The interesting thing I see, is while people in the United States think "capitalism is in crisis", people in Europe are lamenting how "socialism is in crisis". Stranger still, when the differences between the United States and Europe are largely stereotypes (both are mixed economies with fairly similar amounts of regulation and social welfare spending).

What seems to be having a crisis is the institution of government itself. The crisis in the United States was largely that the government could no longer maintain its state-sponsored real estate bubble (the real estate bubble was, in effect, a massive state-organized welfare/jobs program in the United States). In Europe, where citizens are already more comfortable with the welfare state operating more openly, people directly see the success and failure of the state, and so they see their social welfare apparatus in crisis.

In the United States, people think that "Yes We Can" solve the problem of a bankrupt and failing imperialist state by even more massive statism, because Americans associate the failing of our state with some vague notion of capitalism. In Europe, people have a more difficult time fantasizing about economic messiah to save them... They can't blame everything on some non-existent laissez faire capitalist system.

Yeah, what could a guy named Ruben Andersson possibly know about Sweden anyway.

Also, the slant of the article actually seems leftwing (and published in The Guardian, to boot), so dismissing him as some sort of righty trying to smear the wonder that is Sweden doesn't seem appropriate either.

I've participated in one of these "impunity games" at UPenn, when I was in high school (four years ago), and I think I might have some insight into why the participants might reject the raw deal: because theyfucking knew each other!! It wasn't some random stranger that was offering me $3 out of $10 instead of $5, it was somebody that I knew and I would see again. (Furthermore, it was a small sum of money, so it was worth it to screw over your friend who was trying to screw you.) I didn't read the methodology on this one, so maybe the people were random, but I can tell you for sure that the people at Penn sure as hell didn't think about this when they were designing their game.

On Ultimatum Games and the like, people should really read Binmore's book Does Game Theory Work?: The Bargaining Challenge. The book makes a pretty strong case for looking at evolutionarily stable strategies instead of thinking about "fairness" preferences. It turns out that the interpretations of these experiments is more tricky than most commentators think, particularly in the presence of mixed strategies.

Comments for this post are closed