Let me begin by thanking the editors of International Theory for the opportunity to discuss my book, and the respondents who have kindly agreed to comment. My task is to start the conversation with a primer on the contents of the book and some observations about what the book seeks to achieve. Needless to say, it is impossible to provide anything like a comprehensive summary of the book in the space available and so I instead hope to give a flavor of the book and perhaps even persuade people to read the book if they have not yet done so.
How International Law Works offers a theory to explain why states comply with their international legal obligations. It then applies that theory to the field of international law, testing the conventional wisdom on how international law operates and why it looks the way it does.
The writing of How International Law Works began with a commitment to speak to at least two different strands of literature that exist within the field of international law. The first of these is what might be called a traditional legal approach to international law. In response to the question of whether and why international law works, this school of thought has long relied on the notion that states have a propensity to comply with international legal rules.Footnote 1 This can be interpreted as a preference for compliance or an internalization of norms of compliance and, in that sense, can be reconciled with some constructivist approaches.
A second strand of writing is deeply skeptical of international law and institutions. This work argues that international law does not exist, is ineffective, or is sufficiently weak as to be dismissed (Goldsmith and Posner, Reference Goldsmith and Posner2005). The analogous argument in the international relations literature is that international law is epiphenomenal. This is, of course, a significant debate in international relations, and in significant ways that debate has driven the discussion within international law. Ultimately the question of whether international law affects the behavior of states is an empirical one, and no theoretical treatment can settle the question; though my own view is that there is strong evidence – whether anecdotal, qualitative, or quantitative – that it does (Guzman, Reference Guzman2008: 10, 11).
It is not the task of this book to speak to the empirical side of this debate, however. Its goal is to build a stronger and more useful theoretical apparatus with which to examine the institution of international law. To do so, the book adopts standard rational choice assumptions and sets out to explore how much of international law can be explained and understood through that lens.
Before proceeding, a final note as to methodology may be useful. In addition to its goal of speaking to major debates within international legal scholarship, the book aspires to contribute to discussions of international law within the international relations literature. Interdisciplinary work of this sort is challenging because the two fields have very different norms and practices. That which is obvious in one field is often novel in the other. At times, therefore, the book engages in a discussion that is familiar in one of the fields, but requires development for the other. Nowhere is this more true than in the development of the book’s most basic theoretical approach, to which I now turn.Footnote 2
The Three Rs of Compliance
The book starts by engaging the most commonly cited complaint about international law – that it lacks coercive enforcement mechanisms. The absence of coercive enforcement is a signature feature of international law, and a central element in the theory developed in my book. The book reminds us that in a repeated game, cooperation can emerge even in the absence of enforcement. This simple point opens the door for much of what follows in the book. If international law is to work, I argue, it must rely on this sort of cooperation. It is the repeated interaction of states that leads to what I term the ‘Three Rs of Compliance’: reciprocity, retaliation, and reputation (Guzman, Reference Guzman2008: 33–48).
Though the book emphasizes that all three of the Three Rs are important contributors, the workings of reciprocity and retaliation are fairly straightforward and so the discussion of reputation is much more extensive. Even within international relations, where reputation has attracted more attention than in international legal scholarship, theories of reputation have tended to be underdeveloped and,Footnote 3 when treated more thoroughly, are often heavily focused on matters of security.Footnote 4
My book seeks to address this weakness in the literature by developing a more complete model of reputation, including how and why a state’s reputation can change, and the circumstances under which it matters more or less. The theory reveals, for example, that international law can influence states even if reputations are compartmentalized by subject area (Downs and Jones, Reference Downs and Jones2002). Compartmentalization tends to make states more compliant in some areas of international law, and less compliance in others (Guzman, Reference Guzman2008: 100–111).
Application of the theory
Once the basic analytical structure is in place the field of international law looks quite different from the way it has traditionally been viewed, and much of the book is devoted to revisiting the entire subject area through the lens of the Three Rs of Compliance.
The book takes a contractarian approach, as discussed (and at times critiqued) by the commentators. Rational states exchange promises in order to improve their position in much the same way that private parties do when entering into contracts, and we can turn to theories of contract to understand the behavior of states. The first lesson from contracts is that states entering into an agreement seek to maximize their joint gains, subject to transaction costs.
To illustrate how this would apply to states, consider the sort of agreement among states that would be predicted by a simple contract model. Suppose two or more states face a common problem – this can be virtually any sort of problem, but is most easily thought of as some form of prisoner’s dilemma, such as an agreement to deal with the local environmental impacts of business activity on a shared lake. As mentioned, the states can be expected to maximize the joint gains from the agreement. This means that they will identify the value-maximizing substantive terms (e.g. how much waste each country can dump into the lake), price the agreement to ensure that both sides are better off with the agreement than without it, and enter into a fully binding agreement. Because the international system does not provide a convenient set of enforcement tools, states must develop ways to encourage compliance. The agreement might be supplemented, for example, with a system for monitoring compliance and for penalizing violations with sanctions that, in an ideal world, would approximate expectation damages so as to allow efficient breach. None of the available options approximates the strength of domestic legal systems, so one might expect states to use the strongest available tools to make their promises credible.
When we look at the practice of states, however, we again see all sorts of behavior that is at odds with this description. States frequently enter into ‘soft law’ agreements that fall short of a formal treaty, for example. This is particularly surprising because the reality of weak enforcement makes even a formal treaty less of a commitment than what private parties accept when they enter into a binding contract. Furthermore, monitoring systems are often less extensive than one would expect, dispute resolution clauses are the exception rather than the rule, and so on.
This behavior presents a puzzle for a rational choice model – why would states not want to enter into agreements that are more, rather than less, binding. People working in the contracts area might suggest that states may be trying to avoid the problem of over-deterrence. In international law, however, this seems implausible. Given the weak enforcement systems that exist, and the weakness of even the most stringent mechanisms that states can build into an agreement, it is difficult to believe that there is a worry about excessive compliance.
The explanation for this puzzle becomes clear only when one focuses on the way in which compliance comes about in international law, and in particular the way in which reputation matters. Notice first that the key function of any effort to increase the credibility of a state’s commitment is not to trigger or enforce some explicit penalty or punishment. Dispute resolution, monitoring, reporting, transparency requirements, escape clauses, exit clauses, and so on all serve primarily to define and make clear when an obligation has been violated and perhaps to call on the offending party to come into compliance.
In this sense, credibility-enhancing devices serve to provide information to the system. This raises the question of what role information serves and why it might promote compliance. Most directly, it serves to make clear to all states – including those with no direct involvement in the dispute – what has happened and who is legally at fault. This matters to states because a losing defendant suffers a reputational loss if it is found to have violated international law. This threat of a reputational loss explains one way in which international law can generate ‘compliance-pull’, but it also explains why states may be reluctant to incorporate credibility-enhancing devices into their agreements. The loss of reputation suffered by the losing defendant is costly to that state. There is an offsetting gain because observing states get better information about the defendant and its behavior, but that gain is enjoyed by all states in the system, whether or not they are party to the agreement. When entering into a bilateral (or small multilateral) agreement, therefore, states view the potential reputational loss as a net loss for the parties to the agreement – the losing defendant loses reputation and the other parties to the agreement enjoy only a fraction of the total gain. The sanction, in other words, is costly. It increases the probability of compliance, which the parties want, but in the event of a violation it imposes a loss on the parties as a group. The aversion to credibility-enhancing devices, therefore, can be attributed at least in part to the reluctance of states to adopt an information and enforcement system that has a negative-sum character.
This explanation of state reluctance to increase the credibility of commitments is different from the commonly advanced explanation that states often prefer soft law because it preserves their flexibility (Lipson, 1991). The latter argument relies in one way or another on the risk aversion of states to explain why they prefer more flexibility in their commitments.Footnote 5 As I discuss later in this comment, it seems more appropriate to view states as risk neutral in their international legal obligations.
Notice that each of the credibility-enhancing elements mentioned above (among others) serves to adjust the de facto commitment that states are entering into. Changing either the set of behaviors that are classified as violations – as is done by escape clauses and reservations, for example – or the likelihood of a violation being observed – as is done by monitoring and dispute resolution, for example – alters the burden imposed by the agreement. If states want to increase the level of commitment they can, for instance, move from soft law to hard law, setup a monitoring mechanism, build a dispute settlement system, reduce the availability of escape clauses, forbid or restrict reservations, or limit the ability to exit.
Returning to where we started, rational states will make choices among the available design elements in order to maximize the joint value of the agreement. The fact that the agreements actually used by states vary widely in how these design elements are used suggests that the relative value of the various elements changes from one context to the other.
The discussion so far has ignored the actual substance of international agreements. It turns out that we can easily add the substantive terms of an agreement to the discussion, and we can understand those terms in much the same way. Just as the de facto obligation of a state is influenced by the way in which the various design elements already discussed are used, it is obviously also affected by the substance of the agreement. There is, in other words, a tradeoff between the substance of an agreement and its formal provisions (Raustiala, Reference Raustiala2005). The burden of an agreement can be affected by changing the substance, but the same can be achieved through changes to the form of the agreement. In a human rights treaty, for example, one can reduce the practical impact of the treaty on states by watering down the formal requirements, by eliminating monitoring, by permitting reservations, etc.
Risk and risk preferences
Many of the arguments advanced in my book are best described as complementary to at least some existing views. As mentioned above, for example, there are explanations for the choice between hard and soft law that I find persuasive and that are quite different from what I propose. My claims, then, represent additional reasons why soft law instruments are used.
At other points, however, I find myself in disagreement with widely held views. One such instance, of particular relevance to international relations scholars, concerns the way in which states react to risk. It is commonly assumed in both international law and international relations writings that states are risk-averse and this is said to help to explain a variety of behaviors in which states prefer to accept less rather than more constraining commitments (Koremenos et al., Reference Koremenos, Lipson and Snidal2001). As mentioned above, they might, for example, prefer a soft law agreement to a hard law agreement because they prefer to preserve the flexibility to depart from their commitment, even if this means the agreement has a lower expected value.
It seems to me, however, that it is more appropriate to model states as risk-neutral with respect to international legal commitments. At any given moment states have a large number of legal commitments – perhaps thousands – and these commitments are in a wide range of subject areas. In this sense they are ‘diversified’ in their legal commitments. Because states (and their leaders) are concerned with the overall risk of the ‘portfolio’ of legal commitments, much of the risk associated with individual obligations can be ignored. Just as a diversified financial portfolio allows risk-averse investors to behave in a risk-neutral fashion, a diversified portfolio of international legal commitments allows states to behave in a risk-neutral fashion, even if one thinks they are risk-averse (Guzman, Reference Guzman2008: 122–126). To be sure, diversification requires that the risks involved are not too closely correlated with one another, but this is surely true with respect to international agreements. It is difficult to imagine that compliance with or violation of an extradition agreement is closely correlated with compliance with or violation of an environmental agreement or a trade agreement.
One might think that obligations implicating sufficiently important issues – national security, for example – are so critical to a state that they cannot be diversified away. One must remember, however, that while states do indeed enter into international commitments that implicate issues of great import to the state, the legal commitment itself is never essential. The worst that can happen as a result of a legal commitment is that the state may violate international law. Doing so may impose costs, but this brings us back to the diversification claim. Because a state is never forced to comply with its obligation it can choose to breach and suffer the consequences – and this risk can be diversified.
Conclusion
How International Law Works is an attempt to contribute to a more robust theory of international law. The strategy is to make clear and standard assumptions (in this case, rational choice assumptions) and see how far we can go toward understanding how international law works and explaining what we observe in the world. In the end, I think it is fair to say that we are able to explain a fair amount of what we see. Certainly there remain puzzles. Some of these can probably be explained within a rational choice model, while others may require alternative assumptions.
Because the assumptions drive the theory, there are at times contradictions between traditional views of international law and what emerges in the book. I argue, for example, that a more sensible view of customary international law would dispense with the requirement of widespread state practice and instead rely only on a particular understanding of opinio juris.
The most important divergence between a traditional international law perspective and that advocated by the book concerns the very definition of international law. The conventional definition includes formal treaties and customary international law, but excludes ‘soft law’ and ‘mere norms.’ The book points out that if international law is to be coherent it must include a broader category of legal commitments, including soft law and norms. This is so because soft law is best explained as a less robust version of hard law, and a norm is best viewed as a weak version of customary international law.Footnote 6
Furthermore, customary international law must work through the same mechanisms of retaliation, reciprocity, and reputation, as do agreements. We should, therefore, think of the entire field of international law, including soft law and mere norms, as operating through a common set of principles relating to how the costs and benefits of compliance are changed by legal obligations. This represents a significant departure from the way in which we have traditionally thought about international law – at least in the legal academy – but strikes me as a necessary step in bringing the tools and insights of social science to bear on this important institution.
