||The approval mechanism experiment: A solution to prisoner’s dilemma
||Tatsuyoshi Saijo, Yoshitaka Okano and Takafumi Yamakawa
||Players can approve or reject the other choice of the strategy after announcing the choices in a prisoner's dilemma game. If both approve the other choice, the outcome is what they choose, and if either one rejects the other choice, it is the outcome when both defect, which is called the mate choice mechanism. Theoretically, this mechanism implements cooperation in backward elimination of weakly dominated strategies (BEWDS) assuming that players are payoff maximizers, reciprocaters, inequality averters or the mixture of them, but it does not implement cooperation in Nash equilibria (NE) or subgame perfect equilibria (SPE). Although the coverage of behaviors shrinks, it also implements cooperation in neutrally stable strategies (NSS). Experimentally, we observe that the cooperation rate with the mechanism is 90% in round 1 and it is 93.2% through 19 rounds, and that subjects' behavior is consistent with BEWDS rather than NE, SPE or NSS behavior using questionnaire analysis. Utilizing off equilibrium path data, we find that payoff maximizers or reciprocators are 88-90%, inequality averters are 10-11%, and utilitarians are 0-1%.
|Revised version published in