International law provides an ideal context for studying the effects of freedom from coercion on cooperative behavior. Framers of international agreements, no less than the authors of private contracts, can choose between self enforcement and coercive third-party mechanisms to induce compliance with the commitments they make. Studies of individual contracting provide some evidence that coercive sanctions may crowd out self enforcement, implying that too great a propensity by external actors to intervene in the contractual relationship may produce welfare losses. We explore the possibility that too much coercive third-party enforcement similarly can reduce the value of international agreements. We argue that, in spite of the obvious differences between state and individual decisionmaking, enough similarities exist to make the inquiry worthwhile. Using analytic moves worked out in the context of private contracts, we make two general claims about international agreements, one conventional and one controversial. First, we maintain that one usefully can evaluate efforts to frame and implement international agreements in terms of optimal enforcement structure. Choosing from a broad range of normative criteria, one still can distinguish between better and worse enforcement strategies. Second, we argue that the optimal enforcement structure for any particular international agreement will depend on both the goals of the agreement and the context in which it designed and implemented. Because these goals and contexts are diverse, the set of optimal enforcement structures is heterogenous. Some optimal enforcement structures will depend largely on self enforcement, while others will not. Central to our claim is an appreciation of the interaction of self enforcement and third-party coercion including binding arbitration, use of international courts, and enforcement by domestic actors. We maintain that in a far from trivial number of instances subject to international agreement, self enforcement and coercive enforcement may be rivalrous and the optimal enforcement structure would preclude or limit coercive enforcement. In particular, we argue that good theoretical arguments buttress the general tendency of domestic courts not to extend their coercive powers to implement an international agreement without a clear signal from the framers of the agreement that this coercion is desired.