IL/IR Symposium: The Engines of Compliance

[Jana von Stein is a Senior Lecturer in Political Science and International Relations at Victoria University of Wellington (New Zealand) and a Faculty Associate at the Center for Political Studies (University of Michigan)]

When do – and don’t – states comply with international rules? For instrumentalists (adopting Keohane’s –admittedly simplified– categorization of the literature as ‘instrumentalist’ and ‘normative’, the puzzle starts with the observation that no overarching power exists to enforce international law. If there is no ‘highest power’ to enforce rules, why follow them? The ‘engines of compliance’ are typically more diffuse than in domestic systems, but they are nonetheless real:

• International inducements. Sometimes a state benefits enough from having others follow the rules that it pays the ‘cost’ of ensuring compliance itself, whether in the form of ‘carrots’ (e.g., trade concessions) or ‘sticks’ (e.g., economic sanctions). Inducements are typically decentralized and based on self-help, so their application can be uneven. Inducements also face typical collective action problems, and so often work best when a powerful state is doing the heavy lifting.
• Reciprocity. Axelrod demonstrated long ago that reciprocity can be an engine of cooperation if the involved parties are sufficiently sure that they will interact into the future. The same logic holds for compliance, under certain conditions. Reciprocal noncompliance must harm the party that is tempted to renege: this is why (direct) reciprocity is rarely useful in international human rights law, but can work in the realms of trade and war conduct. Reciprocity is also problematic if the ‘punishment’ can’t be limited to the violator, as is often the case in international environmental affairs.
• Reputation. For instrumentalists, reputation is a means to an end: a reputation for keeping promises can make it easier to secure cooperation more broadly or in the future. Reputation is important for predicting future behavior, not for punishing past actions. Scholars debate just how much reputation carries over from one issue-area to another, or from one government administration to another. What is more, concerns about reputation can sometimes push governments not to comply, for instance if they want to foster a reputation for protecting their interests or their friends. 
• Domestic politics and institutions. Scholars have typically focused on three mechanisms that are typical of liberal democracies (see the groundwork laid in Slaughter). Relatively independent courts provide an element of ‘enforcement’ that the international system, by and large, lacks. Elections force leaders to be vigilant about breaking promises. Finally, the fractious bargaining process of negotiating treaties when legislative constraints are present results in agreements that are more likely to stick. The empirical record supports these ideas in many instances, but it is more mixed than one might have anticipated. In part, this is because scholars have relied on blunt measures. This is improving. But with further scrutiny, it is also clear that there are many contingencies and even countervailing pressures.

I have grouped another set of studies under the rubric of ‘normative approaches,’ which is perhaps a misnomer. These approaches are diverse, but united in their view that compliance is not chiefly about concerns about the ex post costs of noncompliance.

• Compliance as a problem of management. For scholars like Chayes and Chayes, the fundamental principle of state consent, and the care that states take when negotiating and entering into treaties, suggest that governments generally intend to keep their international legal promises. Noncompliance typically results from treaty ambiguities, capacity problems, and significant changes over time. Punitive sanctions are hard to mobilize and are rarely work when noncompliance stems from capacity problems. The managerial perspective seems to give us purchase in areas where scientific and/or technological expertise are crucial, such as environmental affairs. But it gives us less traction in the not-hard-to-find cases where governments willingly and knowingly violate international law. Another criticism comes from Downs et al., who argue that it has serious inference problems. I discuss this below.
• Norms and identity. People don’t operate in a social vacuum. They develop norms, which eventually become so deeply ingrained that the latter assume a ‘taken for granted’ character. Koh argues that repeated transnational interactions lead to a reshaping of identities, which then make their way into domestic policy, law, and social practice. At this point, governments don’t just comply; they obey. Johnston emphasizes the roles of persuasion and social influence in promoting compliance. Persuasion is about changing minds and attitudes; framing and argumentation are important. Social influence, in contrast, often results in compliance without internalization: actors comply because they care about perceptions, seek social rewards, and/or want to enhance esteem. Goodman and Jinks emphasize acculturation, which entails a similar process, but with greater emphasis on the disconnect that can emerge between practices and beliefs. Validating these propositions empirically is very challenging. Another criticism is that this literature sometimes maps attributes of individuals onto governments and even states, which is an uncomfortable fit.
• Legitimacy and fairness. For Franck, a rule exerts a ‘compliance pull’ when people perceive it as legitimate and fair. Legitimacy is about right process: a rule’s clarity, its authority as expressed through symbolic rituals, similar treatment of like cases, and procedures that an organized community accepts. Critics charge that this leads to a preference for rules that tick procedural boxes, even if they are unjust. Franck’s subsequent work on fairness is in part an attempt to address that question. Ultimately, the back-and-forth involved in discussing and debating policies and legal interpretations results in fairer rules. This, in turn, elicits better compliance. Critics point to conceptual problems: legitimacy is hard to gauge independently of the compliance it is meant to explain. Moreover, Franck’s arguments beg the question: why do governments ever sign on to agreements that are not ‘legitimate’ and/or ‘fair?’

The chapter discusses three additional current questions of interest to IL and IR scholars studying compliance. First is the potential problem of ‘endogeneity’: as Downs et al. point out, high compliance rates and limited ‘punishment’ may simply reflect that states create and commit to agreements that require minor departures from what they would have done anyhow. These concerns have led quantitative researchers (see von Stein) to use increasingly sophisticated methods to control for endogeneity. This is an important improvement, but there is still little agreement on how best to tackle this problem statistically. To say that endogeneity is a problem, of course, is not to say that it tells the whole story. States are typically motivated by various factors when designing and/or deciding to join international agreements. What’s more, it is impossible for governments to fully anticipate the range of challenges that can arise and undermine compliance. A critical question is this: when circumstances change, do governments continue to comply (setting aside the limited cases in which clausula rebus sic stantibus applies)?

A second question involves definitions. I define compliance as adherence to rules. This is appropriate if we view law as rules intended to regulate behavior. But, if we have a process-based theory, attitudinal alignment is what we should care about. What’s more, because rules are constitutive and not just regulative, it is impossible to separate them from behavior or to make causal statements about the two (C.f. Brunnée and Toope; and Kratochwil and Ruggie). A second criticism by Martin suggests that a focus on compliance is itself misguided; instead, we should be looking at effectiveness and implementation. This has at least two implications: we need to think more carefully about the counterfactuals, and quantitative-oriented scholars at very least need to rethink their dependent variable. These are important points. But in my view, compliance, far from being orthogonal to understanding causal effects, is an important component of how we tackle the question of treaty effects.

Empirically, we know a lot more about compliance than we did just a decade ago. Scholars have made choices about which questions and research tools to privilege. One example of this is the ‘quantification’ of the study of IL. The use of statistics has made it possible to move beyond the idiosyncrasies of a particular time period or country; it has allowed comparisons across agreements and issue areas; and it has made it possible to grapple with difficult questions of endogeneity. At the same time, there are perils: we lose the richness of understanding that comes with honing in on a few cases; existing measures are often unable to distinguish between causal mechanisms; and the wealth of findings we now have has not necessarily generated consensus.

IR and IL scholars are talking to each other with much greater frequency now than they were even a decade ago. They are asking many of the same questions, sharing theoretical approaches, and using many of the same empirical methodologies. This is all for the good.