A competitive test of the core solution in side-payment games

Abstract

This study investigates multiperson decision making in group systems. An experiment was conducted to test the core solution against other theories of payoff allocation in n-person games. Five hundred seventy-six subjects participated in three-person cooperative, superadditive, side-payment games with nonempty core. The manipulated experimental variables included core size, location of the equality payoff vector with respect to the core, and strength of the 23 coalition. Results indicate that the core solution has greater predictive accuracy in games having small rather than large cores and in games with the equality vector located inside rather than outside the core. More important, the results show that the core solution is less accurate than the Shapley, nucleolus, and disruption nucleolus solutions. These findings cast doubt regarding the empirical accuracy of the core solution in side-payment games.