Power, hierarchy and social preferences

I ran an experiment in order to evaluate the relationship, if any, between power, or the search for power, and the degree of altruism. In particular I experimentally tested whether an organization structured in a strictly hierarchical way was able to reduce the degree of altruism of a group of experimental subjects. The subjects were divided into groups and played a series of dictator and ultimatum games with the members of other groups; for each experimental euro that they earned, the experimenter assigned half of it to the group. Two different settings were analyzed according to how this group surplus was distributed among group members. In the control setting (treatment A) the group surplus was distributed equally among group members, while in the power setting (treatment B) there was a ranking of the earnings in the group, and the subject who earned the higher sum was given the power to decide the distribution scheme of the group different from her own. It was found that the introduction of a hierarchical structure generated a significant decrease in the rate of altruism, measured in terms of the allocation given to the receiver in the dictator game. In this case the tournament among group members for leadership and the competition for power was a very strong means to induce behaviour more in line with the classical assumption of economics. A remarkable gender effect emerges, suggesting that women seem less attracted and trapped by competition for power. 1 Introduction A world in which agents are heterogeneous and selfish individuals cohabit with other-regarding ones and in which agents display reciprocative behaviour raises completely new problems in the design of an optimal economic policy (Bowles, 2006). Already analyzed has been the problem of motivational crowding-out (and crowding in) which emerges when external intervention via monetary incentives or punishments may undermine (and under different conditions strengthen) intrinsic motivation (Frey and Jegen, 2001). As a consequence, agents' behaviour is affected by the rules used to interact with other subjects, and the institutional framework in which they are bounded to play significantly shapes their behaviour. The institutional setting and the social values systems play a crucial role in supporting and spreading certain behaviours among agents. Even in the absence of crowding out/in effects, the institutional design would play an important role in shaping individual behaviour by harnessing selfish actions and fostering pro-social actions. To date, very little attention has …