Abstract
This paper experimentally investigates how economic principles affect communication. In a simple sender–receiver game with common interests over payoffs, the sender can send a signal without a pre-given meaning in an infrequent or frequent state of the world. When the signal is costly, several theories (focal point theory, the intuitive criterion, evolutionary game theory) predict an efficient separating equilibrium, where the signal is sent in the infrequent state of the world (also referred to as Horn?s rule). To analyze whether Horn?s rule applies, and if so, which theory best explains it, we develop and test variants of the sender–receiver game where the theories generate discriminatory hypotheses. In costly signaling variants, our participants follow Horn?s rule most of the time, in a manner that is best explained by focal point theory. In costless signaling variants, evolutionary game theory best explains our results. Here participants coordinate significantly more (less) often on a separating equilibrium where the signal is sent in the frequent state if they are primed to associate the absence of a signal with the infrequent (frequent) state of the world. We also find indications that a similar priming effect applies to costly signals. Thus, while the frequency with which participants follow Horn?s rule in costly signaling variants is best explained by Horn?s rule, the priming effect shows that some of our participants? behavior is best explained by evolutionary game theory even when signals are costly
