Reciprocal climate negotiators

Do reciprocal preferences matter for participation in international environmental agreements? Combining the two canonical models of Barrett (2003) and Rabin (1993), I find that 1) the empty coalition is always stable; 2) the grand coalition is stable if reciprocity preferences are sufficiently strong; 3) if a minority of countries are not reciprocal, a majority coalition can still be stable, if reciprocity is sufficiently strong and widespread; and 4) there is also a stable minority coalition, if costs are moderate and the number of countries is not too small. The latter coalition is weakly larger than the maximum stable coalition with standard preferences, but barely improves on welfare, and is characterized by negative sentiments. The analysis illustrates that although models of reciprocity tend to be complex, the context of a specific applied game may reduce this complexity.