Fairness under uncertainty

Abstract

With the help of a repeated matching pennies game, we model a situation in which one of two players may accidentally get into an adverse situation that the other player can take advantage of. In doing so, we vary the degree of uncertainty about future rounds and especially about the frequency and role distribution of situations in which fairness and positive reciprocity are possible. It turns out that uncertainty can greatly increase the number of fair moves when unfair behavior leads to a loss of efficiency. If there is no loss of efficiency and if unfairness only leads to a redistribution between players, we observe almost exclusively selfish behavior. We also vary the social distance and find that small distance leads to a significant increase in the proportion of fair moves even in games where there is no loss of efficiency.