Other Regarding Preferences: Outcomes, Intentions, or Interdependence

The Ultimatum Game seems to be the ideal experiment to test for the structure of preferences or the sequential rationality assumptions underlying subgame perfection. We study the theoretical implications of introducing the possibility of misconceptions - that actions may potentially affect continuation payoffs - and show that the set of Perfect Bayesian Nash Equilibria does not converge to the subgame perfect equilibrium when the possibility of misconception approaches zero. The perfect equilibria studied corresponds qualitatively to the experimental findings of offers made and unfair offers rejected.