A Simple Approach for Organizing Behavior and Explaining Cooperation in Repeated Games

We introduce a novel approach for organizing behavior and explaining cooperation in repeated games. Our approach is based on the idea that players differ according to an inherent propensity to cooperate that systematically affects behavior and cooperation levels. We formulate the empirical implications of this idea and test them in the lab. Our data support our approach. Our main conclusions are: (i) players' strategies in a repeated game can be ranked along a single dimension, (ii) this ranking remains stable across repeated games, and (iii) the composition of a group, in terms of its players' propensities, strongly affects cooperation levels.