Ultimatum and the Kibbutz

I’ve been enjoying my new tenure as a research fellow, for the next six months, at the Centre for Co-operative and Community-Based Economy (formerly, the B.C. Institute of Co-operative Studies). Every Friday, at 10 am, the staff, my fellow fellows and anyone else interested in what goes on here gathers for tea and conversation. (If you’re on campus, you ought to come by!) On a recent Friday, talk turned to the intriguing results of the socio-economic experimental “game” called “Ultimatum”.

In Ultimatum, two players must decide how to divide a sum of money (say, $100). The first person makes an offer; the second person can either accept or reject that offer, but if he or she rejects it, both players get nothing. It’s a one-offer deal. There is no negotiating.

The rational response of Player #2 should be to take whatever is offered by Player #1—it’s better than squat. However, in practice, many players will reject offers of 30% or less. Why they would take nothing rather than a little is one question. Why Player #1 is such a Scrooge is another—although, at least in pure economic science, players are acting “rationally” (i.e., in their own best interests) if they low-ball their compatriot, who has little leverage beyond accepting or rejecting the offer. Further studies have shown variations based on culture and gender: women behave more cooperatively, apparently, and are more likely to propose an even split.

Later, I came across a brief article in the New York Times Magazine annual Ideas issue that described a “drunken Ultimatum” experiment, in which imbibers at a bar were asked to play. They were even less likely to take anything less than a 50:50 split, suggesting that short-term revenge (or a sense of injustice) rather than long-term strategizing was behind this seemingly irrational behaviour.

Of course, after learning about Ultimatum, I immediately wondered: W.W.K.D.? What would a kibbutznik do?

I wasn’t surprised to learn that someone had already answered this question. Because of its unique communal set-up and voluntary membership, the kibbutz is one of the most intensely studied communities in the world, generating thousands of academic studies over the years. (I can only imagine what it must be like to be kibbutz-born identical twins, one raised there and one raised elsewhere—you’d be the ideal choice for every social science, nature-vs.-nurture experiment on the planet!)

While they didn’t use Ultimatum, Bradley Ruffle and Richard Sosis deployed a similar experiment to compare the level of co-operation of kibbutzniks (taken from relatively well-off, still-traditional kibbutzim) between fellow (but anonymous) kibbutz members versus “outsiders” (i.e., likely townsfolk). To simplify, the two players were told that an envelope held 100 shekels and they could take as much or as little as they wanted, but so could the other mystery player. If what the two players took totalled more than 100, then they would receive nothing. If it totalled less than 100, then what remained was multiplied by 1.5 and divided equally between the two players, plus they could keep whatever they had removed originally. The idea is to distinguish motives for individual gain (i.e., taking a lot and hoping your opponent takes a little) from those of collective gain (i.e., taking nothing or a little, knowing that what remains will be redistributed fairly with dividends).

The 2006 paper documents some surprising results. When faced off against another (unknown) kibbutz member, kibbutz subjects showed a willingness to behave cooperatively and only take a little from the envelope. When matched with an “outsider”, they behaved exactly as other subjects in the control group of non-kibbutzniks and were tempted to take more from the envelope. (Sociologists describe this as an example of “in-group bias”.) Equally intriguing, the longer that someone had lived at a kibbutz—i.e., if they were a born and raised kibbutznik versus a new member—the more likely they were to behave less rather than more cooperatively. New members—at least those who choose to join the still-traditional kibbutzim (the experiment was done in 2000, before the acceleration of privatization)—tend to be more ideological in their communalism than existing kibbutzniks.

As the co-authors write in their conclusion:

Despite the promise of a universally cooperative group, kibbutz members cooperate more with members of their own kibbutz than with city residents. What is more, when paired with one another, kibbutz members and city residents exhibit identical levels of cooperation. In this sense, kibbutz members may be said to be conditionally cooperative individuals. Our findings attest to the strength of the psychological foundations of in-group-out-group biases, in spite of a society’s efforts to train its members otherwise. Even members of this once idyllic, voluntary, cooperative community do not treat all individuals alike. Instead, they appear to form expectations concerning others’ degree of cooperation and reciprocate in kind.