Agent-based simulation of trust games for communication and information

A trust game is a two-player game in extended form. The sub game perfect equilibrium of the game is that a player called an investor does not invest the wealth. Because, this behavior is the best response of the investor according to the best response of the another player, called an investor, "not invest". Based on the experimental results of Blacht and Feltovitch(2009), many human subjects choose the equilibrium strategies. However, who are allowed to communicate with the opponent by cheap talk or to observe past activities of the opponent sometimes choose cooperative strategies not strategies. In this study, an agent-based simulation experiments are conducted to analyze the effect of the communication and information effects in the trust games. The experimental result indicates that the effect of information of the past behavior of the opponent to the cooperative behavior is larger than the effect of communication.