An experiment studying the selection effect when assigning distributive power

In our experiment subjects first answer a personality questionnaire. They then take part in a 3-person game. In the game one player chooses between two possibilities. The first is an outside option which assigns a positive amount to that player, but leaves the two others empty-handed. The second option consists in the assignment of power over the distribution of a pie among the three players to one of the other two players. Our main treatments differ with respect to the selection procedure by which distributive power is assigned: to a randomly determined or to a knowingly selected partner. When making her decisions the choosing player could consult the answers to the personality questionnaire of the other two players. Our results show that, as a first reaction, knowingly selected players keep less for themselves than randomly selected ones. We also find that, in both treatments, they allocate more to the players that have given them distributive power than to third players. In our data competition has motivational and not only allocative consequences. *Institut d’Anàlisi Econòmica (CSIC), Barcelona. ** Max Planck Institut, Jena. *** Humboldt-Universität zu Berlin. Financial support by the European Union through the TMR research network ENDEAR (FMRX-CT980238) is gratefully acknowledged.