Testing the Waters

Abstract

Individuals often fail to cooperate because they are not sufficiently reassured that others involved share their desire for reciprocal cooperation. Such situations may be seen as Assurance games. The existing literature fails to examine the problems posed by lack of information about others' preferences, either assuming that information is perfect, or that it may be made perfect by mutual verbal reassurances. This article shows that in sequential public goods supergames, players with Assurance preferences may gather information about others' preferences from their game moves, and that it may pay them to take risks of short-term losses in order to do so. The most efficient information-gathering strategy for such a player is cooperation. The model helps us understand why players of Assurance often appear to take risks in order to “test the waters” with cooperative moves, and why the problem posed for collective action is sometimes successfully resolved.