Endogenous vs. Exogenous Transmission of Information: An Experiment

Based on Gossner, Hernández and Neyman’s (2006) 3-player game (hereafter GHN) we analyze communication efficiency in the lab. In that game, player 1 represents random nature an i.i.d. procedure, player 2 is a fully informed player (wiser), and player 3 is the less informed player (agent). The game is repeated and players 2 and 3 get 1 if both actions match nature’s actions and 0 otherwise. We propose an experiment following this game. We implement two treatments: one without chat (NC) and one with chat (C). In the treatment with chat, players may first send messages to each other through an online chat application, and then play the game. After the chat time, only the wiser player has perfect information on the realized (random) sequence played by nature. The players then play the finitely repeated binary game. In treatment NC, subjects just play the game. In the experiment we observed endogenous communication treatment NC as well as exogenous in treatment C, both of which result in higher payoffs. Furthermore, when explicit communication is possible we observe a chat effect which can be interpreted as a higher level of efficiency in communication. Strategies used by subjects are in line with GHN strategies.