Fairness through the Lens of Cooperative Game Theory: An Experimental Approach

This paper experimentally investigates cooperative game theory from a normative perspective. Subjects designated as Decision Makers express their view on what is fair for others, by recommending a payoff allocation for three subjects (Recipients) whose substitutabilities and complementarities are captured by a characteristic function. We show that axioms and solution concepts from cooperative game theory provide valuable insights into the data. Axiomatic and regression analysis suggest that Decision Makers' choices can be (noisily) described as a convex combination of the Shapley value and equal split solution. A mixture model analysis, examining the distribution of Just Deserts indices describing how far one goes in the direction of the Shapley value, reveals heterogeneity across characteristic functions. Aggregating opinions by averaging, however, shows that the societal view of what is fair remains remarkably consistent across problems.