Customary international law (CIL) is widely recognized as a fundamental source of international law. (1) While its continued significance in the age of treaties was once contested, it is now generally accepted that CIL remains a vital element of the international legal order. (2) Yet CIL is also plagued with conceptual and practical difficulties, which have led critics to challenge its coherence and legitimacy. (3) In particular, critics of CIL have argued that it does not meaningfully affect state behavior. (4) Traditional CIL scholarship is ill equipped to answer such criticism because its objectives are doctrinal or normative--namely, to identify, interpret, and apply CIL rules, or to argue for desirable changes in CIL. For the most part, that scholarship does not propose an explanatory theory in the social scientific sense, which would articulate how CIL works, why states comply, and why and how rules change. (5)
In recent years, scholars have begun to fill this gap by proposing explanatory theories based on the rational choice paradigm. (6) These theories generally describe CIL rules as cooperative equilibria in repeated prisoner's dilemmas, sustained by mechanisms such as reciprocity, retaliation, and reputation. (7) However, these theories raise their own difficulties. In essence, they are general theories of social norms or international cooperation applied to CIL, without fully considering CIL's distinctive legal characteristics and institutional setting--which set it apart from social norms or treaties. (8) They focus primarily on the mechanisms that support compliance in equilibrium, and generally neglect changes in the CIL rules themselves, which they see as driven by exogenous factors. (9) Most importantly, their conclusions often contradict traditional CIL doctrine and the understanding of lawyers, judges, and officials as to how CIL works, (10) which has obstructed dialogue between explanatory and normative or doctrinal debates about CIL.
In this article, we propose a new explanatory theory of CIL, which highlights two ways in which CIL differs from social norms and treaties. First, unlike treaties, CIL offers few legal and institutional features--such as detailed obligations, flexibility clauses, and structured countermeasures--to support reciprocity or retaliation. Second, common knowledge that a rule is CIL comes with shared legal understandings about how CIL works. These shared understandings run against "tit-for-tat" equilibria based on direct reciprocity. CIL rules are commonly understood to be universal in nature: they create rights and obligations that apply to all states, regardless of whether they have themselves accepted them. This universality complicates the systematic use of reciprocity as a stabilizing mechanism, while contributing little to reinforce retaliation or reputational sanctions.
While these characteristics undermine the decentralized punishment strategies emphasized by previous theories, they also lead to an alternative rationale for compliance: a state may comply because it knows its decision to defect creates a precedent that may undermine a cooperative norm it values. Indeed, traditional CIL doctrine recognizes that violations of a CIL rule weaken that rule and may lead to its demise or the emergence of a new rule. (11) This precedential effect arises through at least two channels. First, the state's defection weakens the incentives of other states to continue upholding the rule. Second, courts and other legal actors that apply CIL sincerely--that is, by assessing the prevailing practice and opinio juris--will see the defection as evidence that the rule has disappeared or changed. Thus, if a state values the continued existence of the cooperative norm and believes its decision to defect will create a precedent that undermines the norm, that state may refrain from defecting despite short-term incentives to do so. But if the state does not value the rule or believes that its defection will not much affect the behavior of others, it may choose to defect and free ride on continued compliance by others.
Our theory yields several insights. First, it has important implications for the understanding of how CIL works and changes. It suggests that there may be a "tipping point" beyond which defections accelerate and lead to the rapid demise of a CIL rule and to the formation of a new one--a point we illustrate by reference to several major CIL changes. It also suggests, contrary to existing theories, that in some circumstances powerful states may have a greater stake in complying with CIL rules that they value, because the precedential impact of their actions is greater. Second, our theory is consistent with the available empirical evidence on CIL formation and change. Indeed, we initially developed it while conducting the first systematic empirical study of a change in CIL, which examined the transition from absolute to restrictive sovereign immunity in 121 states. (12) Finally, our theory takes traditional CIL doctrine seriously, while also clarifying and refining it. As such, it bridges the gap between traditional scholarship and the "new" explanatory work on CIL.
Before moving on to articulate our theory, we note some important qualifications and limitations of the scope of our argument. First, while we recognize the growing role of nonstate actors, our theory follows most doctrinal and explanatory accounts in emphasizing the behavior of states and their organs. However, we also explore the role of international courts and tribunals in shaping CIL. Second, while we argue that the legal and institutional characteristics of CIL complicate the use of direct reciprocity, our theory is still based on reciprocity in a more diffuse sense. (13) Indeed, it makes sense for states to be concerned with the precedential impact of their actions on a CIL rule only if they stand to benefit from continued compliance by others, and vice versa. Third, we do not address broader normative questions as to whether specific CIL rules---or, indeed, the CIL formation process itself--is efficient from a welfare standpoint. (14) While our theory may contribute to such analyses, we cannot explore this complex issue within the constraints of the present article.
Part I of the article develops our theory of CIL. After reviewing existing rational choice theories, it explains how institutional limitations and shared legal understandings of CIL undermine the view that cooperative CIL norms are sustained by decentralized threats of punishment. It then articulates an alternative rationale based on the precedential effects of defection and explains its main theoretical implications, including the "tipping point" phenomenon and the special incentives of powerful states. Part II explores our theory's implications for salient areas of traditional CIL doctrine, including several topics that come within the ongoing program of work by the International Law Commission (ILC) on the formation and evidence of CIL. It shows that, while our theory is broadly consistent with mainstream doctrinal views on CIL, it also sheds new light on several debates and areas of ambiguity. Part III discusses the scope and limits of our theory. In particular, it addresses the relationship between precedential concerns and the decentralized punishment mechanisms emphasized by existing theories of CIL. It also explores the potential application of our theory to international cooperation problems that do not clearly constitute prisoner's dilemmas, such as international human rights.
A PRECEDENT-BASED THEORY
Reputation, Reciprocity, and Retaliation
Rational choice accounts hold that many CIL rules--like other international law rules or cooperative social norms--can be understood as equilibria in repeated prisoner's dilemmas. While states can derive mutual benefits from cooperation, they also face short-term incentives to defect and free ride on compliance by others, thus making it difficult to sustain the cooperative equilibrium. (15) The CIL rule of foreign state immunity provides an example: each state benefits when the courts of other states refrain from exercising jurisdiction over it. A state could free ride, however, by allowing foreign states to be sued in its courts--thus allowing local plaintiffs to recover--while continuing to enjoy protection abroad as long as other states continue to observe the rule.
Rational choice theories explain stable cooperation despite such short-term incentives through three mechanisms: reciprocity, retaliation, and reputation. (16) Reciprocity refers to a state's withdrawing its own cooperation in response to defection by another. (17) Because this withdrawal leaves the defecting state worse off than under mutual cooperation, carefully calibrated threats of reciprocity can deter defection. (18) For example, in the case of immunity, other states--either the state affected by the defection or third states--could deprive the defecting state of immunity in their own courts as long as it does not return to compliance. This tit-for-tat strategy is widely considered the most efficient way to achieve stable cooperation in the context of a repeated multilateral prisoner's dilemma. (19)
Retaliation refers to actions that states take to punish a defecting state. For example, states can terminate diplomatic relations, cut off foreign aid, impose trade sanctions, or take military action. (20) In theory, credible threats of retaliation can provide incentives for cooperation. Retaliation has disadvantages, however, compared to direct reciprocity. First, it is typically difficult to produce effective retaliatory sanctions. Unilateral sanctions are often insufficient to deter defection, which means that the providers of retaliation must overcome a collective action problem. (21) As a result, retaliation is often weak, and its threat may not be credible, especially for relatively minor offenses...