to the agreement to appropriate a particular amount (e.g., 5 tokens). If this is the case, coordination on “good” equilibria seems possible. Given these extensive communication opportunities and the fact that there is usually a substantial fraction of reciprocal subjects, it seems quite likely that communication raised efficiency because subjects could coordinate their choices on more efficient equilibria.
Social interactions frequently are associated with social approval or disapproval. The anticipation of such social rewards and punishments may have important economic consequences. For example, it may affect the efficiency of team production and the decisions in diverse areas such as tax evasion, the exploitation of the welfare state, criminal activities, union membership, and voting behavior. The behavioral role of social rewards and punishments is stressed in social exchange theory (Blau, 1964). In contrast to pure economic exchanges, social exchanges involve not only the exchange of economic rewards but also the exchange of social rewards. The admiration or the contempt that is sometimes expressed by parents, teachers, professional colleagues, and spectators are prime examples of a social reward. In general social rewards are not based on explicit contractual arrangements but are triggered by spontaneous positive or negative emotions that can be interpreted as approval and disapproval, respectively.
Of course, approval as well as disapproval can be communicated and can have an important impact on individual behavior in a common-pool resource game. People who talk to each other enter a social relationship. Within this relationship, exchange of approval and disapproval is possible. Two assumptions must be met in order to observe more cooperative behavior compared to SNE, however. First, there must be subjects who actually care about approval or disapproval and who change their behavior in the expectation of such approval or disapproval. Second, there must be subjects who actually express approval or disapproval. The first condition is obvious. The second condition is important because it is usually not costless to express approval and in particular disapproval. Our point is that reciprocally motivated subjects are willing to bear the cost and are willing to reciprocate the cooperative or noncooperative actions by others. Thus, preferences as assumed in our model may explain why communication in combination with the expression of approval and disapproval can have a positive impact on cooperative behavior.
Taken together, we have described two potential channels through which communication may elicit cooperative behavior in the presence of reciprocal preferences. Although the first rests only on the exchange of information, the second is built on the possibility of communication face to face. We would expect, therefore, that communication effects are particularly strong if face-to-face communication is possible (as it is the case in the treatment already discussed).