New Experimental Evidence On What Makes Cooperation Possible

03 Oct 2005

Freedom to choose your associates makes sustained cooperation between individuals possible, according to new experimental research by Professors Talbot Page, Louis Putterman and Bulent Unel, published in the October
2005 issue of the Economic Journal.

Cooperation is critical to the functioning of both economic and political organisations, but standard economic theory has long predicted that rational people will fail to cooperate when they interact in groups for any specific
number of periods. This article reports that when experiment subjects get to choose who they associate with, cooperative individuals band together to defend against opportunists, and the incentive to acquire a good reputation leads even some opportunists to act more cooperatively.

The researchers placed subjects in small groups in which each make a series of decisions about whether to retain an endowment of money for themselves or contribute it toward a group project. Group members were better off when all contributed than when none did. But individuals could be better off still by letting others contribute while holding on to their own money.

Without freedom of association, experimenters have found that contributions decline over time as more cooperative individuals tire of being taken advantage of by more selfish ones. But with freedom to rank prospective
group members and periodic formation of new groups on the basis of the rankings, the new experiment found that cooperation was both initially greater and more sustained.

By itself, the authors explain, standard theory predicts that voluntarily based regrouping should make no difference to cooperation levels if people are strictly self-interested and believe this to be true of others.

The authors show why the presence of cooperation in their experiment demonstrates that their subjects generally believed that some of them could be genuinely inclined to cooperate, even at some sacrifice of their material payoffs. They show how the number of subjects who cooperate in the final period implies that at least 59% of their subject pool had a genuine willingness to cooperate even at some cost to personal earnings, a substantially higher
estimate than found in previous experiments.

The experiment is related to experiments of researchers at the University of Zurich, led by economist Ernst Fehr, and a group that includes Elinor Ostrom, at the University of Indiana, who have provided evidence that many people have a taste for punishing selfish behaviour. These researchers found that the opportunity to impose costly punishments (sanctions) on others is a way to sustain cooperation in the laboratory and, by extension, in some real group situations.

The authors of the present article compare the effects of voluntary regrouping with the effects of punishment. They find that both the opportunity to punish and the opportunity to regroup were similarly effective in inducing cooperation, but the group formation treatment generated more benefits because punishing carries costs both to the punishers and to those punished. The combination of punishment opportunities with control over group composition also gave less benefit than the latter alone.

The new research qualitatively replicates many aspects of Fehr''s and Ostrom''s work. All three sets of researchers see themselves as helping to demonstrate the presence of cooperative propensities that can be harnessed
for social benefit by careful institutional design. All three research groups also conclude that their findings call for a rethinking of some aspects of economic theory.

''Voluntary Association in Public Goods Experiments: Reciprocity, Mimicry and Efficiency'' by Talbot Page, Louis Putterman and Bulent Unel is published in the October 2005 issue of the Economic Journal. Talbot Page and Louis Putterman are in the Department of Economics at Brown University. Bulent Unel is at the University of Texas at Austin.

Louis Putterman

+1-401-863-3837 | louis_putterman@brown.edu