Nash Equilibrium and International Law

Abstract

A new wave of realists argues that international law loses its normative force because states that ‘follow’ international law are merely participants in a Prisoner’s Dilemma seeking to achieve self-interested outcomes. Such claims are not just vastly exaggerated; they misunderstand the significance of game theory. Properly conceived, international law is a Nash Equilibrium – a focal point for states as they make rational decisions regarding strategy. In domains where international law has the greatest purchase, the preferred strategy is reciprocal compliance with international norms. This strategy is consistent with the normativity of law and morality, both of which are characterized by self-interested actors who accept reciprocal constraints to generate Nash Equilibria and, ultimately, a stable social contract. These agents – ‘constrained maximizers’, as the philosopher David Gauthier calls them – accept constraints in order to achieve cooperative benefits. This article concludes that it is also rational for states to comply with these constraints: agents evaluate competing plans and strategies, select the best course of action, and then stick to their decision, rather than obsessively re-evaluating their chosen strategy. A state that defects when the opportunity arises may reduce its overall payoff as compared to a state that selects and adheres to a strategy of constrained maximization.

1 Introduction

For at least several decades, game theory has played a central role in the international relations literature. Only recently has it emerged as a powerful force in the international law literature as well. Political scientists learned as long ago as the 1960s – with the work of Thomas Schelling – that game theory offered a sophisticated matrix for modelling state relations.¹ The econometrics of game theory came with the promise of predicting behaviour: social scientists could not only explain why some states had acted the way they did, but might also predict future behaviour under certain conditions.² The Prisoner’s Dilemma provided an answer for problems regarding coordination and cooperation that had concerned the international relations literature for years.³

The central puzzle of the Prisoner’s Dilemma literature was the uncertain and uneasy relationship between a state’s selfish behaviour in international relations and a state’s commitment to international legal norms when those norms proved inconvenient or downright inconsistent with a state’s self-interest. One school of thought concluded that states generally act in their self-interest and seek to ignore the prescriptive power of international legal norms when the norms are sufficiently inconvenient.⁴ A second school of thought concluded that states are generally more receptive to international norms for a variety of reasons. For many scholars, receptivity to international legal norms could be explained by future costs associated with non-compliance (i.e., loss of reputation that might frustrate a state’s ability to negotiate future agreements), thus collapsing international law compliance into low-discount rate, self-interested behaviour.⁵ Or, in the alternative, some scholars concluded that compliance with international legal norms was internalized as a value that formed one part of a nation’s self-interest.⁶ In other words, fidelity to national values included, inter alia, compliance with international law, because some countries view participation in the global legal order (or fidelity to its underlying norms) as an essential part of their identity and constitutive commitments.⁷ Therefore, compliance with international law was a national interest to be included with other more egoistic national values. This novel move was simultaneously edifying and de-edifying in the sense that it elevated fidelity to international law to a high national interest (a good thing), yet simultaneously deflated international law compliance by turning it into just another interest in a field of interests, as opposed to a universal norm that demands compliance in the face of contrary self-interest. What each school rejected was what one might call a naïve account of international law: that states comply with international law simply because it is law.

While game theory offered theorists of international relations a model for explaining state relations, the methodology has had a far more explosive effect among international lawyers. Recent accounts have harnessed alleged lessons learned from game theory in the service of a new brand of realism about international law.⁸ These sceptical accounts conclude that international law loses its normative force because states that ‘follow’ international law are simply participants in a Prisoner’s Dilemma seeking to achieve self-interested outcomes.⁹ In short, these arguments can be distilled to the following elements. Effective multilateral agreements are rarely achieved, either in treaty or customary form.¹⁰ Most states consent to international legal norms through a process of bilateral agreements with specific partners who in turn have their own set of overlapping bilateral agreements.¹¹ Compliance with these agreements, whether via treaty or customary law, is usually based on considerations specific to a particular partner rather than general considerations regarding the content of the legal norm.¹² In other words, states comply with international norms in specific interactions with a particular state when there are good reasons to believe that the other state will reciprocate such compliance.¹³ This explains why a state might adhere to a particular legal norm with one partner but not with another. According to this school of thought, the vast majority of the content of international law fits this paradigm as opposed to one that posits general legal obligations to the entire world community.¹⁴ Reducing international law to a series of overlapping bilateral arrangements facilitates the use of the Prisoner’s Dilemma as a convincing model, though of course it is not necessary to limit the analysis to bilateral interactions. It is, after all, possible to have a multiple-player Prisoner’s Dilemma, though cooperation becomes more challenging as the number of players increases.¹⁵ In any event, the rhetorical advantage to the bilateral claim is clear: it makes the Prisoner’s Dilemma that much more intuitive as a model for international law.

The new realists proceed to argue that compliance in a Prisoner’s Dilemma is based on reciprocity that is hard to come by. A state will prefer to violate the treaty or customary rule while its competitor adheres to it, though this state of affairs is hard to achieve as all competitors share the exact same preference.¹⁶ Thus, in order to avoid the opposite result (mutual defection), states cooperate in the form of international agreements to produce the next-best preference: mutual adherence to the norm. Now comes the theoretical payoff, in the form of multiple claims: First, cooperation in the form of international agreements only shows up in the very limited situations when participants in the game have equal or near-equal bargaining power.¹⁷ In contrast, most cases of international relations involve unequal bargaining relationships, where a weak state is forced to adhere to the wishes of the stronger state or face unfavourable consequences.¹⁸ This reduces the scope of international law. Secondly, even in cases of comparable bargaining power, the application of the norm is based entirely on reciprocal compliance.¹⁹ States generally follow the norm only if their bilateral competitor also follows the norm. Unfortunately, international law has a relative paucity of enforcement mechanisms compared with domestic law, making assured reciprocal compliance through coercion rare and difficult to achieve. This further reduces the scope of international law. Thirdly, even when both states in a Prisoner’s Dilemma follow the norm, they are doing so out of state self-interest.²⁰ In other words, it is within a state’s self-interest to follow an international legal norm if and only if the other player is also following that same norm. Consequently, international law is really just a matter of self-interested behaviour on the part of states, not a robust system of law that demands compliance even when it conflicts with a participant’s self-interest.

Now comes the normative payoff of the argument, in the form of a fourth claim. Because international law is reducible to self-interested behaviour, states have no independent obligation to follow international law when it conflicts with their self-interest.²¹ International law is based entirely on the Prisoner’s Dilemma structure of self-interested behaviour; thus it has no independent normative force. If states wish to comply with international law, they may do so when it suits them. They may also structure international law obligations to their own benefit, but ought not to be concerned with how these norms affect humanity as a whole or the global community.²² Indeed, the claim is not just that states are not required to follow international law when it conflicts with their self-interest, but in fact that they should not. A government that follows international law when such law conflicts with the self-interest of the state is breaching its fiduciary duty to its citizens and placing the welfare of foreigners above the welfare of its citizens.²³ Partiality is not just permitted, but required.²⁴ This article takes aim at the validity of the third claim and its normative payoff. Since the third claim is based on a conceptual error, the supposed normative payoff is illusory.

Predictably, the new realism about international law sparked a serious counter-attack from both the professoriate and the international bar,²⁵ though such realism already had its adherents in some corners of the US Department of State (in previous administrations).²⁶ Most law school professors writing about international law are deeply committed to in the claim that international law has normative force and that states ought to follow it.²⁷ Consequently, scholars have mounted numerous defences of international law, cataloguing the effectiveness of human rights treaties and identifying the complex compliance and enforcement mechanisms that currently exist under international law.²⁸ Although most of these arguments are undoubtedly correct, they miss something fundamental and foundational about the new realism: the use of game theory as a mechanism for making claims regarding international law’s normativity – a claim that was largely absent from the international relations literature on game theory.²⁹ The use of game theory as an underlying methodology for understanding international law presents unique issues regarding the degree to which a descriptive methodology can yield normative conclusions regarding international law.

I argue here that the new realism about international law suffers from a profound misunderstanding about the significance of game theory. In short, the new realism misuses the methodology by concluding that self-interested behaviour and normativity are mutually exclusive.³⁰ Indeed, that is the conclusion that the new realists draw from the Prisoner’s Dilemma. This conclusion is false.

In order to defend this claim, we must engage in some preliminaries. First, section 2 of this article offers a more nuanced understanding of the Prisoner’s Dilemma in international law and explains how the international legal order promotes the creation of Nash Equilibria among its participants. Section 3 then explains the compatibility between rational self-interest and the normativity of international law, invoking the concept of constrained maximization. By invoking the rationality of plans, section 3 also explains why it would be rational for a state to follow international law even when it might defect with impunity. Finally, section 4 considers several objections, including the naturalistic fallacy, the unequal bargaining power of states, and the alleged inability of nation-states to bear moral obligations.

2 The Prisoner’s Dilemma and Nash Equilibrium

The outline of the Prisoner’s Dilemma story as told by the new realists is essentially correct, though at times it borders on unsophisticated and draws the wrong conclusions from the methodology. We shall start with the unsophisticated nature of the model and then proceed to the second question of the false conclusions drawn from it. As to the model, Goldsmith and Posner view international cooperation as a bilateral repeated Prisoner’s Dilemma.³¹ While this view is true, the model can be revised and tweaked. Properly conceived, the best way to understand international law is as a Nash Equilibrium – a focal point that states gravitate towards as they make rational decisions regarding strategy in the light of strategies selected by other states.³² In game theory, a Nash Equilibrium is defined as a solution in which each player evaluates the strategies of their competitors and decides that they gain no advantage by unilaterally changing strategy when all other players keep their own strategies unchanged.³³ A Nash Equilibrium functions as a kind of focal point, where participants in the game gravitate towards a particular legal norm and choose ‘compliance’ as their strategy if and only if the other players in the game are also choosing compliance as their strategy.³⁴ When a bilateral international agreement works, one state realizes that unilaterally choosing ‘breach’ as its strategy would confer no benefit because the costs associated with that shift in strategy are too high. So, the player sticks with compliance. If one player decides that a shift in strategy (i.e., breach) is indeed in his or her best interest, then the players fall out of Nash Equilibrium.³⁵

A Bilateral Agreements

In domains where international law has the greatest purchase, the strategy that results in the equilibrium is reciprocal compliance with international norms.³⁶ Consider a bilateral treaty negotiation regarding extraditions between two countries: state A and state B sign a treaty promising mutual extradition between their countries and establishing a legal framework governing these extraditions. Suppose that state A has custody of a suspect and must decide whether to comply with its obligations under the treaty regime. State A realizes that failure to comply with the regime will not only risk retaliation from state B in future extradition matters, but will also have numerous collateral effects – including possible retaliation in other bilateral contexts with state B as well as a loss of reputation in treaty negotiations with other states, which may now be less willing to sign agreements with state A.³⁷ Consequently, state A decides that compliance with the legal norm is in its self-interest and that it has no reason unilaterally to change its strategy. The cost of shifting strategies is just too high. Consequently, the states in this bilateral treaty regime are in Nash Equilibrium with each other because neither party has reason unilaterally to change its strategy. In this case, their compliance with an international treaty norm can be understood through game theory’s lens of self-interested behaviour.³⁸

Of course, one might point out that it may be beneficial for a state to defy the treaty commitment when it proves inconvenient, thus effectively transforming the state into a free rider that receives the benefits of the legal regulation but ignores the costs when they prove inconvenient.³⁹ This is certainly true, but the whole point of the structure of international law is that this outcome (free ridership) is more difficult to achieve ceteris paribus. Because states are linked together through mutual ongoing interactions that are explicitly legal in nature, a state cannot benefit by changing its strategy away from compliance. If it does so, it incurs costs associated with non-compliance that overwhelm any putative benefits from its defection against the norm. The whole point of international law is to create a structure whereby the cost of shifting strategy away from compliance becomes higher than it would be without legal regulation in that particular area. As a result, each state in the Nash Equilibrium decides to comply with the legal norm in question.

It is important to remember that the equilibrium need not be the most optimal or efficient legal regulation possible.⁴⁰ It might be the case that a different legal regime creates cooperation that produces greater benefits for every state.⁴¹ But this kind of Pareto optimality may be difficult to achieve. For example, it might be more efficient for the states to set up a bilateral international court to decide all cases of extradition between the two countries, though each state gravitates towards a Nash Equilibrium that is far below the Pareto optimal outcome for these two players. There is nothing in international law that promises that a stable set of legal regulations between competitors will be the most efficient regulations possible.⁴² Indeed, over time one hopes that the legal regime may evolve closer to Pareto optimality as initial cooperation yields greater cooperation. But in some cases the particular toolbox of compliance mechanisms in international law might limit the amount of optimality one can achieve in this context.⁴³ Although international law yields stable Nash Equilibria, it will never yield the kind of Pareto optimality that one finds in a domestic legal system.

B Multilateral Agreements

The same analysis would apply in a multilateral context. Consider, for example, the most important area of international legal regulation: the use of force.⁴⁴ This is also the most contentious area of international legal regulation, one that the new realists often use as a poster child for their contention that legal norms will give way to self-interest when the cost of compliance becomes inconvenient.⁴⁵ However, the Nash Equilibrium here is clear. The norm in question is the legal prohibition on the use of force, in both the UN Charter and customary law, unless such use of force is authorized by the Security Council – the central clearing house for decisions regarding international peace and security.⁴⁶ Some scholars trace the norm back to the Kellogg–Briand Pact, before which aggressive war was simply recognized as inevitable (and therefore not presumptively illegal).⁴⁷ This is too simplistic, since it was at the very least implicit in the notion of Westphalian sovereignty that states were free not just from outside interference in the widest sense, but also from outside attack in the narrowest sense.⁴⁸ In the current scheme, the prohibition against the use of force is now coupled with the Security Council’s authority to author ize use of force to restore international peace and security.⁴⁹

Unfortunately, Security Council authorizations for the use of force are rare, and, since the threat of a veto is always present, states cannot predict with any reasonable certainly when the Security Council will authorize such use of force.⁵⁰ Thus, state A complies with the norm and eschews the use of force. This strategy of compliance is made with the hope that the other players in the game will also favour compliance. However, no state can assume that competitors will adopt the same strategy; the competitors may choose violation as their strategy and in so doing reserve the right to use force at their discretion. Why would the second state choose this strategy? Perhaps because the costs associated with non-compliance are relatively mild. Although the state might be sued before the International Court of Justice (ICJ) and lose international standing (e.g., reputation), these costs pale in comparison to foregoing the use of force when your competitors refuse to do the same. This is why the international legal community has not navigated towards a Nash Equilibrium that grants the Security Council the exclusive authority to authorize military force. The stakes are too high and the legal prohibitions insufficient to incentivize reciprocal compliance. Simply put, each participant has an incentive to change its strategy away from compliance regardless of the strategy chosen by its competitors.

It is precisely for this reason that, at its earliest incarnation, international law gravitated towards a norm regarding the use of force that allowed unilateral exceptions to the prohibition against the use of force in cases of self-defence. Nineteenth-century treatises regarding public international law, in discussing the use of force, made clear that military force was legal in cases of self-defence or self-preservation.⁵¹ This exception to the norm prohibiting the use of force is as old as the prohibition itself. Although states were unwilling to adopt a strategy of compliance with a blanket prohibition on military force, they have been willing to adopt a strategy of compliance with a more nuanced legal norm that always allows military force in self-defence.⁵² A state can comply with this norm because even if a competitor in the game changes strategy, defects from the norm, and engages in aggressive warfare the first state can still use force in self-defence to protect itself, consistent with the legal norm. In other words, the cost of compliance with the norm does not require that a state risk its national security.⁵³

Consequently, states have a reason to stick with the strategy of compliance, even given the uncertainty regarding the strategy of their competitors in the game. That is why a Nash Equilibrium has developed round a prohibition regarding the use of force unless authorized by the Security Council or in self-defence. Each state benefits from the legal norm – a stable world order without aggressive force and constant warfare – and therefore complies with the legal norm because compliance with the norm is also consistent with purely defensive force when competitors in the game change their strategy.⁵⁴ So, no state has reason unilaterally to change its strategy in the game.

C Law and Self-interest

It is clearly correct, then, that international fidelity to the legal prohibition regarding the use of force can be described, using game theory, as self-interested behaviour on the part of states. However, this much was already clear in the previous wave of international relations scholarship 25 years ago.⁵⁵ Although advancements in the game theory models have only added sophistication to the analysis, they are hardly new. However, the new realists operating in the international law scholarship take all of this as evidence for a much more explosive normative claim: since compliance with international law is based on self-interest, international law has no normative pull.⁵⁶ The status of international law as law is seriously called into doubt.

There are many different ways of making this claim. One might conclude that international law is not law at all, or one might simply claim that international law is far less important than international lawyers think.⁵⁷ Or one might say that states comply with international law only when doing so furthers their self-interest and reject it whenever it does not, making international law different not in degree but in kind from domestic law.⁵⁸ All of these claims add up to an assault on international law’s normativity.

Of course, I am not the first to object to the new realism, and there is now a wide array of literature providing renewed justifications for international law in the face of the new realist attack.⁵⁹ However, none of the defences have, to my mind, adequately emphasized the specific methodological mistakes made by the new realists. Although game theory allows us to model international law as a game of self-interest, this picture is entirely consistent, pace Goldsmith and Posner, with international law’s normativity. Simply put, the Prisoner’s Dilemma also provides a model to explain morality itself (i.e., that of self-interested actors who accept reciprocal moral constraints on action as a social contract), and this dual nature of the Prisoner’s Dilemma cannot be taken as a reason to deny morality’s normativity, on pain of a reductio ad absurdum to complete moral scepticism.

3 Self-Interest and Normativity

In this section, I unpack the observation that game theory provides a model not only to depict international law as a game of self-interest but also to explain morality itself. In 1986, the moral philosopher David Gauthier published Morals By Agreement, a novel interpretation of social contract theory that harnessed the power of game theory to explain why rational actors would agree to a system that constrained their behaviour.⁶⁰Morals By Agreement provided, for the first time, a fully realized model of rational self-interested individuals agreeing to a social contract of morality.⁶¹ The relationship between reason and morality has a long pedigree, going back as far as Plato’s The Republic and, more explicitly, Kant’s work on the categorical imperative and the wave of contractarian theories following Rawls.⁶² But for Gauthier only game theory provided the necessary tools to explain how individual rationality and moral constraints might be consistent with each other.⁶³ Indeed, for Gauthier, the claim was even stronger: the latter could be derived from the former in the sense that one could demonstrate that rational agents ought to accept moral constraints.⁶⁴ In pursuing this account, Gauthier did not even resort to a universalized rational account of morality, i.e., he did not shift the focus from individual-level rationality to group-level rationality, arguing that a third-person point of view required the individual to recognize, on pain of contradiction, that accepting moral constraints was best for everyone.⁶⁵ Gauthier was unimpressed by such sleight-of-hand.⁶⁶ His vision of morality required that we face the hard question: is it rational for individuals, considering their self-interest, to accept the normative constraints of morality?⁶⁷

A Morality and the Prisoner’s Dilemma

The answer – almost a revelation for Gauthier – lay in the Prisoner’s Dilemma.⁶⁸ Rational agents must make decisions based on the expected moves of their competitors. Although the best possible outcome for a given player is defection in the face of compliance by all other competitors in the game, this outcome is also the outcome preferred by one’s competitors. If all competitors defect, the resulting payoff is extremely low, effectively throwing the game back into a state of nature where no one complies with any moral constraints, thus producing the worst possible outcome. The ration al solution to the game therefore requires acceptance of the objectively second-best (but, rationally, only possible) outcome: acceptance of reciprocal moral constraints on behaviour.⁶⁹ The purchase one gets from game theory is that this acceptance is itself demanded by self-interested behaviour. Rational agents seeking to maximize their own outcomes will choose moral outcomes as long as morality is a group endeavour.

Of course, this still leaves unresolved the cleavage between the rational agent at the social bargaining table – who is rationally compelled to accept reciprocal moral constraints – and the rational agent who must decide whether or not to comply with the social contract. It is one thing to demonstrate the rationality of bargaining for moral constraints and quite another to demonstrate the rationality of ex post compliance with the results of the social contract.⁷⁰ For Gauthier, such a rational agent must be considered a constrained maximizer, or an agent who ‘enjoy[s] opportunities for co-operation which others lack’, as Gauthier puts it, as opposed to a straightforward maximizer.⁷¹ The question is whether the constrained maximizer receives cooperative benefits that outweigh the risks associated with the strategy of constrained maximization – i.e., the risk that competitors in the game will defect and reject compliance as their strategy.⁷²

How can this be demonstrated? For Hobbes, the answer was simple: the sovereign itself ensures compliance, a fact that provided its own rationale for Hobbes’s specific rendering of The Leviathan.⁷³ Once one steps outside the scope of a total sovereign, though, the picture becomes more complicated. Various social institutions, both informal and formal, exist to promote cooperation among constrained maximizers: increased trust between cooperators, reputational gains, and community structures only open to cooperators, all of which have instrumental value for further cooperation.⁷⁴ Defectors, though they achieve some benefits from their straightforward maximization, lose all of the benefits of cooperation and suffer the community penalties for defection.⁷⁵ Consequently, constrained maximization is rational just so long as the community has the correct ratio of constrained maximizers to straightforward maximizers.⁷⁶ In a world filled with straightforward maximizers, the gains from (putative) cooperation would not outweigh the risks associated with the compliance strategy. However, in a world with a significant proportion of constrained maximizers, the strategy has a clear salience. Presumably, there is an empirical tipping point at which point the strategy of constrained maximization becomes rational.⁷⁷ The strategy becomes a Nash Equilibrium.

One might argue that the concept of constrained maximization is nothing more complicated than the concept of a long-term interest. Agents are typically concerned with maximizing their gains in the present and thus ignore strategies that will produce a maximum gain over a longer period. Whether one should maximize benefits now or later depends on what discount rate the agent applies to future benefits. If the discount rate is low (or zero), the agent will consider future benefits at full value when engaging in decision-making. If the discount rate is high, the agent will discount the future benefits and treat them as less valuable in deciding on a course of action today. Constrained maximizers certainly recognize that both the present and future benefits of cooperation will far outweigh the constraints of their behaviour. But the strategy of constrained maximization is about far more than simply long-term interests. The benefits of cooperation may be far in the future or immediate; similarly, the demands of constraint may impose themselves today or tomorrow. The real distinguishing factor of constrained maximization is a matter of pure strategy: go it alone and reap the benefits and consequences of such breach, or accept reciprocal constraints and receive the cooperative benefits that go along with them.

B Constrained Maximizers and International Law

One can see how the strategy of constrained maximization is directly applicable to international legal relations.⁷⁸ When one state decides on a strategy for diplomatic relations, it can choose to be a straightforward maximizer or a constrained maximizer. However, deciding to be a straightforward maximizer – although initially an attractive option – carries severe costs. A state that pursues this strategy will be branded a rogue nation and deprive itself of the benefits associated with cooperative constraints. Operating outside the community of nations carries enormous costs, as North Korea, Iran, and other isolationist states can no doubt confirm.⁷⁹ Those who adopt a strategy of reciprocal commitments to international law live in not only a world of relative security – fewer military interventions and aggressive acts – but also a world of bilateral treaty arrangements that would otherwise be unavailable to them. The rub of the argument is that the alleged dichotomy between fidelity to international law and self-interested behaviour turns out to be illusory.⁸⁰ The fact that states are self-interested in no way undermines the normativity of international law.

In the end, states cooperate by complying with international legal norms, and this commitment is necessarily grounded by their self-interest. The new realists claim that acting out of self-interest undermines the normativity of the subsequent constraints, especially that of customary norms built upon opinio juris. However, if this is the lesson that game theory teaches us for international law, then this must also be the lesson that game theory provides for morality itself. If the Prisoner’s Dilemma provides a reason to reject the normativity of international law, then it must also provide a reason to reject the normativity of morality. Consequently, the position of the new realists implies a reductio ad absurdum to complete moral scepticism unless the new realists can provide a salient asymmetry between international law and morality itself.

In searching for this alleged asymmetry, the new realists frequently resort to their old bailiwick: the lack of international enforcement mechanisms to punish defectors or straightforward maximizers.⁸¹ While it is no doubt true that international enforcement mechanisms are feeble when compared with their domestic analogues, the fact is often repeated to the point of exaggeration.⁸² It is certainly not the case that there are no viable mechanisms of enforcement; this point has already been exhaustively detailed in the literature and I shall not rehash the evidence here.⁸³ Further, even if we accept the proposition that international compliance mechanisms are compar ably feeble, the rest of the argument does not follow. The relative lack of enforcement mechanisms might be a relevant asymmetry between international law and domestic law, but it is hardly a relevant asymmetry between international law and morality. Indeed, moral norms – especially those that do not find their expression codified in the criminal law – rely on exactly the same kind of inchoate and allegedly nebulous mechanisms that punish defectors and provide most individuals with a rational reason to choose constrained maximization as their strategy. If this were not the case, then most individuals would reject as illusory all moral norms entirely. Both morality and international law create robust systems that reward cooperative behaviour.

Why do the new realists resist these arguments? Although the concept of constrained maximization is nowhere considered in The Limits of International Law, some clues are offered in Eric Posner’s work on social norms.⁸⁴ In Law and Social Norms, Posner concedes that rational agents will engage in ‘principled’ behaviour and will reap the reputational rewards associated with using the rhetoric of principle.⁸⁵ The language of principle has a signalling effect to potential associates: this agent can be trusted because he will never betray you.⁸⁶ But under Posner’s view of rational choice, such a blind commitment to principle is either illusory or insincere.⁸⁷ At some point, the costs of adhering to the demands of principle will become too high, and any rational agent (according to Posner) will defect and violate the demands of principle.⁸⁸ But saying ‘I will follow principle but only if the costs aren’t too high’ will not help one attract collaborators who are rightly scared away by such conditional language.⁸⁹ Thus, the result is that people cling to the rhetoric of absolute commitment to prin ciple, and when self-interest demands defection from the principle ‘they cheat and try to conceal their opportunism behind casuistry’.⁹⁰ The unprincipled attempt to have their cake and eat it too by attempting to blend in among a crowd of principled agents.⁹¹ This strategy works because it is very difficult for the public to distinguish between the principled and the unprincipled.⁹²

The same view apparently underlies Posner’s attitude about national compliance with international legal norms.⁹³ States will adopt the language of principled adherence to international law, but when self-interest demands defection, they can – and should – act out of self-interest.⁹⁴ Such a state may very well attempt to conceal its behaviour and develop obfuscations in an attempt to explain away the defection.⁹⁵ The state will attempt to defect and still reap the rewards of constrained maximization.⁹⁶

The question is whether a state can successfully adopt the insincere rhetoric of constrained maximization (i.e., fidelity to international law) while at the same time defecting and ignoring international legal norms. But in this respect, there is a relevant asymmetry between individuals and nations. While the individual can hide his decision-making process from potential collaborators, most modern nation-states conduct their decision-making through various internal actors. These debates are often – though not always – public or semipublic. Disputes with domestic constituencies are laid bare for the entire world to see.⁹⁷ If a domestic constituency presses the government to ignore international law out of self-interest, this plea will be heard not just by its own government but by the world. The ability to act insincerely is comparatively more difficult in the case of nation-states than it is with individuals. To the extent that some states, such as North Korea, conduct deliberations in secret, these states appear to be the least likely insincerely to claim adherence to international legal norms. Such rogue nations are often the least likely to tout publicly their adherence to, and participation in, international legal and regulatory regimes.

C Compliance and the Rationality of Plans

However, this still leaves a theoretical tension between the demands of rationality (occasional defection) and the demands of morality that counsels adherence to principle even in the face of rational opportunism. For Goldsmith and Posner, there is no moral basis to tell a state to follow international law when rational self-interest counsels in favour of defection.⁹⁸ And if indeed there arises a situation where the gains of defection outweigh the loss of cooperative opportunities at any given moment, rational choice would appear to demand defection. And since our account of morality is closely linked with rational choice, there would appear to be no basis to tell a nation to forego self-interest in favour of principle.

Gauthier’s initial answer to this conundrum was to frame his account in terms of dispositions to cooperate – dispositions that were themselves rational (and moral) insofar as one found oneself in a community with a sufficient number of agents who were similarly disposed.⁹⁹ In later work, Gauthier pushed beyond the concept of dispositions to cooperate in favour of an account of agency that linked intentions with plans and strategies that operate over time.¹⁰⁰ In other words, although rational choice theory – including Posner’s version – considers an agent’s all-things-considered judgement at each cardinal point in time, rational human agency operates in a far more subtle way. Were rational agents to recalculate rational choice at every cardinal time point, they would be exhausted and weighed down by the process of deliberation to the point of total collapse¹⁰¹ – literally, paralysis by analysis.

Instead, rational agency should be understood in terms of strategies selected after moments of deliberation, after which the chosen strategies only come up for re-evaluation at certain moments in time.¹⁰² What is missing from Posner’s account, in other words, is the concept of plans.¹⁰³ And plans are sticky in the sense that rational agents form an intention to follow a plan and do not give up the plan at the drop of a hat.¹⁰⁴ Living life as a rational agent requires the use of plans; rational agency would be unimaginable without them.¹⁰⁵

One might object that the stickiness of plans is irrational.¹⁰⁶ In other words, a truly rational agent should be constantly re-evaluating the rational benefits associated with their plan and all alternative plans.¹⁰⁷ The rational agent should be playing chess like Deep Blue (re-evaluating the benefits of every possible move at each move) and not like Garry Kasparov (pursuing and committing to a long-term strategy to win the game). To the extent that human agents are incapable of calculating like Deep Blue, perhaps one should count this limitation as a failure of rationality.

Gauthier, drawing partially on the work of Bratman and others, points out that the answer is not so simple.¹⁰⁸ Even if defection at any given moment is rationally beneficial, this is not the right comparison. Pursuing the strategy of rational choice at each cardinal time point may turn out to be less effective than choosing an overall strategy or plan that is rationally justified and then sticking to it.¹⁰⁹ Indeed, consist ently pursuing rational choice at each time point may end up being self-defeating in the long run.¹¹⁰ Plans provide stability for agents to pursue long-term interests and should be abandoned only in favour of new plans, not in favour of momentary and isolated desires.¹¹¹ An agent that is too easily lured from a stable plan by opportunistic defection is a myopic chooser.¹¹² Another way of stating the point is that the rationality of compliance with the reciprocal constraint – following the rules and resisting the temptation to defect – is conditional on the constraint’s place within the larger, rationally justified plan.¹¹³

The structure of this argument is well known to moral theorists who debate the relative merits of act utilitarianism and rule utilitarianism.¹¹⁴ Act utilitarians evaluate the consequences of each individual act and identify the moral thing to do based on this calculation.¹¹⁵ By contrast, rule utilitarians evaluate general moral rules based on their guidelines and then identify certain rules as amoral regardless of their consequences at any individual decisional time point.¹¹⁶ One reason for supporting rule utilitarianism is that, in the end, it may produce superior consequences globally, as compliance is better achieved in a world with sticky moral norms rather than constantly shifting moral evaluations. Ironically, constant re-evaluation of the consequences at each moment in time may end up being self-defeating.

None of this is new in the moral or political theory literature. Within the recent debate in the international law literature, the basic assumptions regarding rational choice among the new realists have gone relatively unchallenged. Some have questioned the wisdom of applying rational choice to international law; others have accepted the methodology but simply claimed that it yields different results.¹¹⁷ But what is badly needed is critical re-evaluation of the version of rational choice theory used by the new realists.¹¹⁸

There is strong reason to believe that states are rationally justified in pursuing a strategy of constrained maximization and sticking to it even when faced with the temptation of opportunistic defection. Even if states could masquerade as principled – a doubtful proposition – constant defection from international legal norms may produce negative outcomes over time. It might be more rational for states to pick the strategy that is rationally justified and stick to it: either try one’s best to engage with international institutions or ignore them. Although it is unclear whether this thesis could be empirically tested, it is very suggestive that the most successful nations in the world participate in international legal institutions whereas rogue nations on the periphery often are beset with hunger, famine, and war.

4 Objections to the Moral Obligation of States

At this point, several other objections to my account must be considered. The most worrying objection, addressed in section 4A, is that Gauthier’s theory of morality, and our extrapolation of that theory to the domain of international law, has fallen victim to the naturalistic fallacy. A second objection, outlined in section 4B, concerns the unequal bargaining strength of states – one alleged reason for stronger states to refrain from a strategy of constrained maximization. The third objection, addressed in section 4C, is that states are collective entities that are unable to bear a moral obligation, and that only individuals are directly subject to the demands of morality. If this view is correct, it would be nonsensical to say that a state has a moral duty to follow international law.

A Rationality: Normative, Not Descriptive

If the entire project is designed to derive morality from reason, then it would indeed appear as if we have attempted to jump over the is–ought gap. In his later work, Rawls famously distanced himself from any attempt to derive morality from reason,¹¹⁹ though in his earliest work, like John Harsanyi, he described his social contract theory as one piece of a general theory of rational choice.¹²⁰ Indeed, the whole project of deriving morality from reason stems from Kant and his obsession with practical reason and finds its apex in contemporary moral philosophers such as Rawls and Alan Gewirth.¹²¹ In his later work, Rawls took great pains to emphasize the role of reflective equilibrium in his methodology: not a top-down derivation but rather a theoretical device for navigating toward a coherent vision of justice as fairness both in the original position and in defensible principles of justice.¹²²

Have we fallen victim to the naturalistic fallacy here? The answer requires an im portant clarification. In deriving morality from rationality, we are not deriving morality from the fact of rationality. Rather, we are deriving moral value from ration ality as a value.¹²³ Simply put, constrained maximizers ought to pursue compliance as their strategy if they are committed to rationality as a value. If they aspire to be rational, then this is what rationality demands, though there is nothing that requires them to be rational. Moral value turns out to be somewhat parasitic on normative rationality – precisely the lesson that game theory has taught both moral philosophers and international lawyers.

Are individuals committed to rationality as a norm? They certainly are, and undeniably so, insofar as they hope to exercise rational agency.¹²⁴ Indeed, even the most elementary forms of agency require a commitment to rationality in the form of means–end reasoning, the transitive ordering of preferences, and the law of non-contradiction. It is very difficult – perhaps impossible – to imagine inter-human relations, including language, without this commitment to basic principles of rationality. And the fact that individuals may be imperfectly rational is entirely irrelevant to the point here. One’s normative commitments may fall well short of perfection, or even large-scale success, but that is not evidence that one is not committed to the value in question. No one achieves perfect rationality, just as no one achieves perfect morality. But this is trivial; the point is that if individuals are committed to rationality, then they ought to be committed to a strategy of constrained maximizing in the form of accepting reciprocal moral constraints. And, as it happens, all individuals are committed to rationality as a norm because this value commitment is constitutive of rational agency itself. Committing to rationality is part of what it means to be a rational agent.¹²⁵

Can the same thing be said about states? Are they committed to rationality as a value? The question is best pursued from the opposite direction: how could we deny that states are committed to rationality as a norm? States have interests and pursue collective projects on the international stage in order to maximize those interests.¹²⁶ Those projects involve rationality over time and necessarily require basic principles of rationality such as the transitive ordering of preferences and fidelity to the principle of non-contradiction.¹²⁷ The only relevant difference between states and individuals is the lack of phenomenological unity among the former.¹²⁸ While each individual typically enjoys a unified phenomenological point of view, states are composed of many individuals, each of whom represents her own unified phenomenological point of view.¹²⁹ But the lack of phenomenological unity of the state does not prevent it from exercising rational agency. Although the phenomenological unity of the individual certainly facilitates rational integration (viz., self-knowledge and direct epistemic access to one’s own thoughts), none of this implies that there cannot exist external means of displaying a shared commitment to rationality. This is precisely what a state achieves through government, a system of representation and deliberation, and diplomatic representation on the world stage.¹³⁰ To deny the rational agency of states would be to deny the foundations of international relations.

B Bargaining Power

We must now redeem a promissory note and account for the fact that states bargaining for international legal norms do not stand in a position of equal bargaining strength.¹³¹ Up to this point, we have assumed that participants in the Prisoner’s Dilemma are bare self-interested agents, without further consideration of their particular strengths and weaknesses that might affect their ability or willingness to defect.¹³² Indeed, the new realists make much of the unequal bargaining power of states and conclude that stronger states will ignore international law simply because they can.¹³³ Given that the unequal bargaining power of states is undeniable (even though it stands in tension with the formal equality of all states under international law), it would seem that our account is impoverished at best and irrelevant at worst.¹³⁴

This anxiety is misplaced. The unequal bargaining power of states is relevant under our model because it affects the costs of non-compliance and the benefits associated with cooperation. As to the former, stronger states will face less retaliation for their non-compliance because weaker states may feel that they need to sign agreements with the stronger state even if previous defections alert the weaker state to the risk that the strong state will again defect. The unequal bargaining power might bring the weaker state to the table in spite of this prediction. Secondly, the benefits associated with cooperation are less significant for stronger states. Their stronger status might open up avenues of cooperation simply because they are stronger and because other states therefore need their cooperation – cooperation that is entirely independent of their strategy of constrained maximization.¹³⁵

Three points are in order here. First, the difference in bargaining power will be most salient when strong and weak states bargain against each other, but will be irrelevant when strong states bargain against each other and weak states do likewise. Secondly, the difference in bargaining power does not prevent strong and weak states from signing agreements; it simply increases the likelihood that the stronger state might be tempted to defect. In cases where the stronger state is strong enough to eschew constrained maximization entirely in favour of a strategy of straightforward maximization, the state may indeed defect. This is most likely in contexts where international law has weak enforcement mechanisms. In contexts where the enforcement mechanisms – however diffuse and informal – are working properly, constrained maximization will continue to be viable even for strong states.

Thirdly and most importantly, the fact that some states will defect in favour of straightforward maximization in some contexts is completely irrelevant for situations where constrained maximization continues to be most valuable. Indeed, this is the logical error made by the new realists. They point out the situations where unequal bargaining power and lack of enforcement allow stronger states to ignore international law with impunity, i.e., to act in their own self-interest.¹³⁶ In such cases, it is indeed correct to suggest that international law is ineffective. But the new realists then use this fact as a pivot to say something about situations when international law is effective, i.e., when states agree to follow international law because the value of constrained maximization is high. Since this situation is also governed by self-interest, the new realists implicitly conclude that even these domains of international cooperation have little or no normative pull because there is no sense of legal obligation.¹³⁷ This is an error. One should see immediately that the underlying current of self-interest works differently in each case. In the latter, self-interest entails constrained maximization and fidelity to international norms; in the former, self-interest entails defection. The fact that some states will violate international law when reason counsels defection does not mean that they are not following international law when reason demands respect for it.

It should come as no surprise both that there are areas where international law remains ineffective due to insufficient enforcement mechanisms, and that the absence of enforcement is less of a constraint for the most powerful nations. Though this is a pedestrian point, it does point to one aspect of the new realist critique that demands further study: the degree to which power imbalances change the tipping point at which a state has reason to shift from straightforward maximization to constrained maximization. This question is largely empirical and ought to be studied more systematically by scholars with training in empirical legal studies. However, the goal of such research would not be to undermine the normativity of international law, but rather to determine with empirical rigour those areas where international law is least effective and where systems of enforcement ought to be strengthened.¹³⁸

C The Moral Obligation of Groups

This article now concludes by briefly rejecting another alleged reason why states need not follow international law: the supposed inability of collective entities to bear moral or legal responsibilities. According to the new realists, corporate bodies (including states) are incapable of bearing such obligations.¹³⁹ Although corporations enjoy legal rights and bear legal responsibilities, they do so because their constituent parts – officers, directors, employees, and shareholders – all benefit from, and consent to, corporate obligations.¹⁴⁰ Shareholders accept the risk of paying for corporate obligations (including unforeseen liabilities) because they also accept the promise of future dividends based on their equity stake.¹⁴¹ Although states do not demonstrate the same kind of internal organization, they nonetheless do organize themselves so that they can act on the world stage, form alliances and agreements with other states, and enjoy all of the cooperative benefits of constrained maximizers. Although the citizen does not receive dividends like a shareholder, the citizen certainly enjoys the cooperative benefits of living in a state that engages in international relations: everything from economic opportunities fuelled by trade to the peace dividends that flow from the prohibition on the international use of force. Citizens do not consent in the same way as do shareholders who purchase stock, but their acceptance of the benefits of citizenship certainly functions as tacit consent.¹⁴²

The new realists also claim that states cannot be morally bound by international law because they are incapable of consenting to their obligations, a fundamental precondition of international treaties and customary law.¹⁴³ Under this view, corporations have the power to make binding commitments only because doing so increases the autonomy of its individual members: thus, the corporate power to consent to obligations only has instrumental value.¹⁴⁴ When the corporate commitment is too burdensome to the individuals, they demand the corporation change the commitment, whereas citizens of a state allegedly have no such authority. Once a state accepts a legal obligation, it remains operative for future generations even after the original citizens are dead.¹⁴⁵ Although international legal obligations are surely dynamic in nature as states abrogate, amend, supplement, and revoke treaties constantly, the persistent objector doctrine and jus cogens may limit a state’s opportunities for revising customary international law.¹⁴⁶

Furthermore, the new realists reject the possibility that the autonomy of states has intrinsic value.¹⁴⁷ The warrant for this conclusion is that states, unlike individuals, have no life plans, and therefore are not valid subjects of the principles of autonomy that are required for an agent to realize a life plan.¹⁴⁸ This conclusion bears scrutiny.¹⁴⁹ If a state lacks the agency necessary to realize a life plan, it is unclear how a state has enough agency to exercise supposedly self-interested behaviour on the world stage. Implicit in the notion of self-interested behaviour, consistent with the Prisoner’s Dilemma, is the notion of a rational agent with enough foresight to have long-term interests (through subsequent iterations of the game). If the possibility of a state’s life plan is rejected, then so is the entire applicability of the game theory methodology to international law and international relations; one would effectively have to throw out the baby with the bathwater. A state’s normative agency inevitably entails the construction of long-term interests, which renders the state capable of consenting to (and bearing) legal obligations.

One might respond that there is a difference between a state’s capacity for agency and whether this autonomy is intrinsically valuable. On this view, states are capable of exercising collective agency, though this agency has instrumental value only insofar as it facilitates or maximizes the autonomy of individual citizens whose life plans may require organization into collective units (states) that can operate on the world stage. Something along these lines might be implicit in Kymlicka’s account of individual human flourishing within a community.¹⁵⁰ That is, the life plan of the state has no independent intrinsic value.

I reject this view of the collective entity as having no independent moral value, though I cannot defend fully the claim in this limited forum.¹⁵¹ Nations, both in the cultural abstract and in their particular organization as nation-states, contribute to the rich tapestry of human existence.¹⁵² However, merely assuming arguendo that states have no independent autonomy does not by itself require the conclusion that states are incapable of bearing moral obligations. There is a missing proposition in the argument, namely that moral obligations at the collective level evaporate if they fail to maximize the autonomy of individuals.

This need not be the case. One might coherently argue that once properly formed from the material of rational individuals, states become distinct entities whose interrelations are governed by an autonomous sphere of legal relations that are independent of the domestic laws governing their citizens internally. Just as one might call a corp oration a legal or metaphysical fiction (though I do not subscribe to this view),¹⁵³ one might just as well dismiss a state with the same epithet. But the fiction might be sufficiently robust that its own collective agency generates corresponding moral duties even if, at the end of the day, its moral significance originally emerged from its constituent parts. A state without citizens would not have any value; however, once a state is composed of individuals and begins to exercise collective rationality in its engagement with other states, it becomes capable of bearing moral obligations. Indeed, I have tried to demonstrate in this article that a state’s collective rationality (in the form of constrained maximization) requires that it follow international legal obligation.

5 Conclusion

Although the new realists offer an academic argument, it is important to remember that their game-theory-fuelled scepticism codifies a view that extends far beyond the academy – it pervaded American foreign policy for much of the last decade.¹⁵⁴ Although many commentators have exposed the flaws in such reasoning, few of the criticisms have – as we have done here – explicitly focused on the link between national self-interest and fidelity to legal norms as being essentially the same dynamic underlying normative rationality and normative morality. This is an ambitious claim; those who reject this account of morality might also reject its relevance for a theory of international law. For some, the notion of constrained maximization may leave little room for our folk concept of altruism or for doing what’s right when it requires significant sacrifice. But this account of morality leaves open the possibility that in any one scenario, fidelity to norms may require significant sacrifice; the account simply insists that, as an overall and long-term strategy, constrained maximization is rationally justified. For the very same reason, states might still comply with international law with opinio juris – a sense of obligation – knowing that in any one context it might involve a sacrifice but with full knowledge that in the long term constrained maximization is in the nation’s self-interest. In a way, this is the lesson that was lost in our foreign policy over the last decade.

1

See T.C. Schelling, The Strategy of Conflict (2nd edn, 1980), at 3–20 (discussing the ‘retarded science of international strategy’).

2

Cf. J.P. Trachtman, The Economic Structure of International Law (2008), at 4–5.

3

See Schelling, supra note 1, at 7 (noting that ‘[w]hat is impressive is . . . how vague the concepts still are . . . and how inelegant the current theory of deterrence is’); ibid., at 213–214 (explaining the Prisoner’s Dilemma); ibid., at 225–226 (using the Prisoner’s Dilemma to explain coordination and cooperation regarding warning systems).

4

See, e.g., Abbott, ‘Modern International Relations Theory: A Prospectus for International Lawyers’, 14 Yale J Int’l L (1989) 335, at 337–338; Setear, ‘An Iterative Perspective on Treaties: A Synthesis of International Relations Theory and International Law’, 37 Harvard Int’lLJ (1996) 139.

5

See, e.g., Setear, ‘Responses to Breach of a Treaty and Rationalist International Relations Theory: The Rules of Release and Remediation in the Law of Treaties and the Law of State Responsibility’, 83 Virginia L Rev (1997) 1, at 8 and 74–75.

6

See, e.g., Katzenstein, ‘Introduction: Alternative Perspectives on National Security’, in P.J. Katzenstein (ed.), The Culture of National Security: Norms and Identity in World Politics (1996), at 13, 17–26; see also A. Wendt, Social Theory of International Politics (1999), at 198.

7

See, e.g., T.M. Franck, Fairness in International Law and Institutions (1995), at 42–45; T.M. Franck, The Power of Legitimacy Among Nations (1990), at 25; L. Henkin, How Nations Behave: Law and Foreign Policy (2nd edn, 1979), at 46–48; Koh, ‘Why Do Nations Obey International Law?’, 106 Yale LJ (1997) 2599, at 2602–2603.

8

See, e.g., J.L. Goldsmith and E.A. Posner, The Limits of International Law (2005); Posner, ‘Do States Have a Moral Obligation to Obey International Law?’, 55 Stanford L Rev (2003) 1901. The current wave is ‘new’ because it harkens back to a first wave of prominent sceptics of international law. See generally H.J. Morgenthau, Politics Among Nations: The Struggle for Power and Peace (2nd edn, 1954).

9

Goldsmith and Posner, supra note 8, at 184 (concluding that ‘[w]hen states cooperate in their self-interest, they naturally use the moralistic language of obligation rather than the strategic language of interest. But saying that the former is evidence of moral motivation is like saying that when states talk of friendship or brotherhood they use these terms, which are meant to reflect aspirations for closer relations, in a literal sense’). Goldsmith and Posner thereby presume that the language of morality and the language of interest are mutually exclusive categories – a proposition they never explicitly defend. See also ibid., at 100.

10

Ibid., at 36–37 (arguing that, in treaty contexts, states may achieve ‘shallow multistate cooperation’ and that, in the context of customary international law, ‘genuine multistate cooperation is unlikely to emerge’); see also ibid., at 87 (asserting scepticism that ‘genuine multinational collective action problems can be solved by treaty’).

11

Ibid., at 87 (describing how cooperation in pairs creates a multilateral regime).

12

Ibid., at 88 (describing the ‘strong pattern in international law’ whereby threats of retaliation are nearly always the responsibility of the victims of violations and concluding that the ‘enforcement of multilateral treaty regimes is usually bilateral’).

13

See ibid., at 87–88.

14

Ibid., at 66 (arguing that theorists inflate context-specific and temporally-limited behavioural patterns, coincidences of interest, and situations of coercion into exogenous rules of customary law).

15

Ibid., at 36 (discussing the costs associated with the multilateral model, including increased costs of monitoring and the risk of undetected free-riding).

16

Ibid., at 32–35 (describing coordination problems).

17

Ibid., at 60.

18

For a full resolution of this point see infra sect. 4B.

19

See, e.g., Goldsmith and Posner, supra note 8, at 150–151 (discussing reciprocal compliance in the context of GATT).

20

See ibid., at 100.

21

Ibid., at 185 (arguing that a moral obligation to comply with international law is illusory).

22

Ibid., at 205–206.

23

Ibid., at 209–215.

24

See also Goldsmith, ‘Liberal Democracy and Cosmopolitan Duty’, 55 Stanford L Rev (2003) 1667, at 1675–1682.

25

For a particularly trenchant example see Hockett, ‘The Limits of Their World’, 90Minnesota L Rev (2006) 1720 (reviewing Goldsmith and Posner, supra note 8).

26

See, e.g., Franck, ‘The Power of Legitimacy and the Legitimacy of Power: International Law in an Age of Power Disequilibrium’, 100 AJIL (2006) 88, at 90 (‘[n]ot surprisingly, however, the claim [of law’s fecklessness] resonates strongly in the halls of American governance’).

27

For a classic example see M.E. O’Connell, The Power and Purpose of International Law: Insights from the Theory and Practice of Enforcement (2008).

28

See, e.g., A.T. Guzman, How International Law Works: A Rational Choice Theory (2008), at 13 (providing an explanation of international law’s effectiveness from a rational choice perspective).

29

Although Guzman uses game theory models expertly to demonstrate the effectiveness of international law, ibid., he does not directly dwell on the issue that I have raised here, i.e., whether the assumption of self-interest implicit in the Prisoner’s Dilemma undermines international law’s essential normativity. Guzman has pursued his analysis in a number of important essays: see, e.g., Guzman, ‘A Compliance-Based Theory of International Law’, 90 California L Rev (2002) 1823 (hereinafter Guzman, ‘A Compliance-Based Theory’) (presenting a theory of international law in which compliance occurs in a model of rational, self-interested states); Guzman, ‘Reputation and International Law’, 34 Georgia J Int’l & Comp L (2006) 379 (hereinafter Guzman, ‘Reputation and International Law’) (describing expected loss of reputation as one mechanism of ensuring compliance); Guzman, ‘Saving Customary International Law’, 27 Michigan J Int’l L (2005) 115 (hereinafter Guzman, ‘Saving Customary International Law’) (mapping out a theory of customary international law based on a model of rational choice); see also Dunoff and Trachtman, ‘Economic Analysis of International Law’, 24 Yale J Int’l L (1999) 1.

30

See, e.g., Goldsmith and Posner, supra note 8, at 100. Other commentators have noted the lack of support for this assumption: see, e.g., Norman and Trachtman, ‘The Customary International Law Game’, 99 AJIL (2005) 541, at 541–542. The argument presented by Goldsmith and Posner relies on the proposition that customary international law is based on opinio juris and that acting in self-interest precludes acting out of a sense of legal obligation: see Goldsmith and Posner, supra note 8, at 14–15. The answer to this sceptical challenge lies in properly understanding opinio juris as ‘the intent of states to propose or accept a rule of law that will serve as the focal point of behavior, implicate an important set of default rules applicable to law but not to other types of social order, and bring into play an important set of linkages among legal rules’: Norman and Trachtman, supra, at 542; see also Alvarez, ‘A BIT on Custom’, 42 NYU J Int’l L & Politics (2009) 17, at 43 (‘[t]hat states have or may have had “economic” reasons to conclude a treaty does not exclude other normative effects produced by these treaties’ entry into force, subsequent practice under them, or efforts to enforce them’).

31

Goldsmith and Posner, supra note 8, at 87–88.

32

Cf. Benvenisti, ‘Customary International Law as a Judicial Tool for Promoting Efficiency’, in E. Benvenisti and M. Hirsch (eds), The Impact of International Law on International Cooperation: Theoretical Perspectives (2004), at 85, 92–94 (describing the use of customary international law for shifting to a new, more efficient equilibrium); Lipson, ‘Why Are Some International Agreements Informal?’, in B.A. Simmons and R.H. Steinberg (eds), International Law and International Relations (2006), at 293, 297–298 (discussing how tacit understandings and implicit rules may create a Nash equilibrium).

33

See D.G. Baird, R.H. Gertner, and R.C. Picker, Game Theory and the Law (1994), at 39–41.

34

See K. Basu, Prelude to Political Economy: A Study of the Social and Political Foundations of Economics (2000), at 114–116 (describing the problem of choosing between multiple Nash Equilibria); see also Baird et al., supra note 33, at 39–40 (discussing a classical example of a focal point); Schelling, supra note 1, at 110–112 (discussing focal points).

35

Cf. Katz, ‘The Strategic Structure of Offer and Acceptance: Game Theory and the Law of Contract Formation’, 89 Michigan L Rev (1990) 215.

36

See Norman and Trachtman, supra note 30, at 542, 571.

37

See, e.g., Guzman, ‘A Compliance-Based Theory’, supra note 29 (presenting a theory of international law in which compliance occurs in a model of rational, self-interested states); Setear, supra note 5, at 1 (examining the international legal rules that govern responses to treaty breaches from the perspective of rationalist theories of international relations).

38

However, pace Goldsmith and Posner, the parties’ self-interested compliance does not preclude their acting out of opinio juris: see Norman and Trachtman, supra note 30, at 541–542; see also Alvarez, supra note 30, at 44.

39

See Goldsmith and Posner, supra note 8, at 87 (arguing that the free rider problem is worse when an agreement involves large numbers of states).

40

See Basu, supra note 34, at 114 (discussing problem of multiple Nash Equilibria).

41

See H. Gintis, Game Theory Evolving: A Problem-Centered Introduction to Modeling Strategic Interaction (2nd edn, 2009), at 109.

42

See, e.g., Guzman, ‘Public Choice and International Regulatory Competition’, 90Georgetown LJ (2002) 971, at 975.

43

See ibid., at 984 (noting that there are ‘problems with international cooperation that make it inferior to well-functioning domestic systems’).

44

See M.J. Glennon, Limits of Law, Prerogatives of Power: Interventionism After Kosovo (2001), at 3.

45

See ibid.

46

See generally ibid., at 17–19.

47

See Y. Dinstein, War,Aggression and Self-Defence (4th edn, 2005), at 83.

48

See M.E. O’Connell, International Law and the Use of Force: Cases and Materials (2005), at 114–117.

49

Glennon,supra note 44, at 17–19.

50

Cf. Lobel and Ratner, ‘Bypassing the Security Council: Ambiguous Authorizations to Use Force, Cease-Fires and the Iraqi Inspection Regime’, 93 AJIL (1999) 124, at 154 (discussing the Council’s impotence and failure to act in this area).

51

See, e.g.,J. Westlake, Chapters on the Principles of International Law (1894), at 115; H. Wheaton, Elements of International Law (8th edn, 1866), at 90; T.D. Woolsey, Introduction to the Study of International Law (5th edn, 1879), at 184.

52

See O’Connell, supra note 48, at 240 (discussing the fact that the UN Charter prohibits force generally while leaving a limited exception for self-defence).

53

But see Murphy, ‘The Doctrine of Preemptive Self-Defense’, 50 Villanova L Rev (2005) 699, at 702.

54

It is certainly true that not all states comply with the prohibition regarding the use of force. However, Henkin must surely be right that ‘almost all nations observe almost all principles of international law and almost all of their obligations almost all of the time’: seeHenkin, supra note 7, at 47 (emphasis omitted). If there is any doubt regarding the veracity of the maxim, one need only ask what the world would look like today if the prohibition regarding the use of force was not followed most of the time.

55

Cf. Schelling, supra note 1, at 119 (suggesting that game theory can be more extensively used to analyse non-zero-sum games of strategy).

56

See, e.g., Goldsmith and Posner, supra note 8, at 184 (arguing that states use moralistic and legalistic rhetoric merely to disguise purely self-interested motives).

57

Ibid., at 225 (‘[i]nternational law is a real phenomenon, but international law scholars exaggerate its power and significance’).

58

See, e.g., Glennon, supra note 44; Lobel and Ratner, supra note 50.

59

See, e.g., Hockett, supra note 25; Norman and Trachtman, supra note 30, at 541–542.

60

D. Gauthier, Morals By Agreement (1986).

61

Many moral philosophers have pursued similar themes, but without explicitly invoking game theory as a methodological tool: see, e.g., T. Nagel, The Possibility of Altruism (1970); T.M. Scanlon, What We Owe to Each Other (1998).

62

See generally J. Rawls, A Theory of Justice (1971) (describing and elaborating upon the conception of justice that is implicit in the contract tradition).

63

Gauthier, supra note 60, at 9 (‘[m]orality does not emerge as the rabbit from the empty hat . . . . [I]t emerges quite simply from the application of the maximizing conception of rationality to certain structures of interaction’); see also Kraus and Coleman, ‘Morality and the Theory of Rational Choice’, in P. Vallentyne (ed.), Contractarianism and Rational Choice: Essays on David Gauthier’s Morals By Agreement (1991), at 254, 255 (arguing that rationality cannot provide the substantive content of morality).

64

Gauthier, supra note 60, at 9 (‘[r]eason overrides the presumption against morality’).

65

See ibid., at 10 (emphasizing that his theoretical focus is on why it is rational for individuals to agree to constraining principles ex ante as well as to comply with such agreed constraints ex post).

66

This is one way of understanding Hobbes’s theory: collective rationality and mutual benefit demands a social contract, but individuals might prefer free riding, thus requiring the Leviathan to enforce individual compliance. For further discussion of the relationship of collective rationality to the social contract in Hobbes’s theory see P. Pettit, Made with Words: Hobbes on Language, Mind, and Politics (2008). See generally D.P. Gauthier, The Logic of Leviathan: The Moral and Political Theory of Thomas Hobbes (1969).

67

For more recent projects pursuing the same line see generally K. Binmore, Game Theory and the Social Contract II: Just Playing (1998).

68

Gauthier, supra note 60, at p. v (‘[t]e present enquiry began . . . when, fumbling for words in which to express the peculiar relationship between morality and advantage, I was shown the Prisoner’s Dilemma’).

69

Ibid., at 177 (arguing that cooperation and constraint by all would ‘yield nearly optimal and fair outcomes’).

70

Ibid., at 14–15.

71

Ibid., at 15.

72

Ibid., at 175–176.

73

See Pettit, supra note 66, at 108 (‘Hobbes’s picture is that as [people] each contract to create a commonwealth, people know that should they later defect, then the sovereign, drawing on the strength of the rest, will be there to punish them’).

74

The value of reputational gains and the costs associated with reputational losses will depend on the degree to which reputation is carried over from one legal context to another: see, e.g., Guzman, supra note 28, at 100–111 (discussing the compartmentalizing of reputation); Downs and Jones, ‘Reputation, Compliance, and International Law’, 31 J Legal Studies (2002) S95 (outlining empirical and theoretical reasons for believing that the actual effects of reputation are both weaker and more complicated than the standard view of reputation suggests); see also Swaine, ‘Rational Custom’, 52 Duke LJ (2002) 559, at 618 (noting that ‘states do not, in fact, interact solely with respect to one rule or the other, and it is also possible to understand their interaction with respect both to an individual rule and to the system of customary international law’).

75

Gauthier, supra note 60, at 176–177.

76

Ibid., at 176.

77

See ibid., at 174–175.

78

Gauthier himself published work regarding Hobbes’s theory of international relations: see, e.g., Gauthier, supra note 66, at 207–212 (discussing Hobbes’s views on the state of nations, and observing that the development of nuclear weapons is ‘bringing the state of nations nearer to the true Hobbesian state of nature’); see also Gauthier, ‘Deterrence, Maximization, and Rationality’, in D. MacLean (ed.), The Security Gamble: Deterrence Dilemmas in the Nuclear Age (1984), at 100, 107 (drawing on Hobbes’s theories of nature to defend the rationality of deterrent policies).

79

See, e.g., A. Chayes and A.H. Chayes, The New Sovereignty: Compliance with International Regulatory Agreements (1995), at 70.

80

For a similar argument see Spiro, ‘A Negative Proof of International Law’, 34 Georgia J Int’l & Comp L (2006) 445, at 447.

81

Cf. Glennon, supra note 44, at 60–64 (discussing desuetude).

82

See also C.R. Beitz, Political Theory and International Relations (1979), at 47–49 (noting that, in the face of considerable empirical evidence to the contrary, people continue to suggest that international relations resembles a Hobbesian state of nature).

83

See, e.g., Chayes and Chayes, supra note 79, at 22–28 (proposing different methods of ensuring compliance); see also D. Cortright and G.A. Lopez, The Sanctions Decade: Assessing UN Strategies in the 1990s (2000) (analysing the effectiveness of 12 cases of UN sanctions).

84

See E.A. Posner, Law and Social Norms (2000).

85

Ibid., at 187 (asserting that people rationally use the rhetoric of principle ‘in order to obtain strategic advantages in their interactions with others’).

86

Posner further elucidates the idea of an absolute principle as a claim regarding incommensurability: see ibid., at 192–198. In other words, if someone says that no amount of money will convince them to give up a much-needed vacation with his or her family, that person is implicitly saying that the value of money and the value of time with his or her family are incommensurable and cannot be compared. If they could be compared, according to Posner, there would be a price at which the amount of money would outweigh the value of the time with the family: ibid., at 193–194. Posner concludes that incommensurability claims signal to others that one will not cheat: ibid., at 197.

87

Ibid., at 195–197.

88

Ibid., at 190 (asserting, as an example, that ‘a rational person will sacrifice his reputation when the gains are sufficiently high’).

89

Ibid., at 197.

90

Ibid., at 195.

91

Ibid., at 197–198.

92

Ibid., at 197.

93

See, e.g., Goldsmith and Posner, supra note 8, at 167–184 (discussing their ‘theory of international rhetoric’).

94

Ibid., at 185.

95

Ibid., at 169 (‘states provide legal or moral justifications for their actions, no matter how transparently self-interested their actions are’).

96

Ibid., at 172.

97

Cf. ibid., at 178–179 (discussing how leaders will address their speech to foreign leaders but intend their talk for domestic audiences).

98

Ibid., at 185.

99

See Gauthier, supra note 60, at 182–184.

100

See, e.g., Gauthier, ‘Commitment and Choice: An Essay on the Rationality of Plans’, in F. Farina et al. (eds), Ethics, Rationality, and Economic Behaviour (1996), at 217, 217–219 (arguing that plans serve as a rational guide for subsequent actions); see also Gauthier, ‘Intention and Deliberation’, in P.A. Danielson (ed.), Modeling Rationality, Morality, and Evolution (1998), at 41, 53 (discussing why an agent would rationally deliberate about plans).

101

See Gauthier, ‘Commitment and Choice’, supra note 100, at 219 (discussing how agents who adopt a plan restrict subsequent deliberation for actions that are compatible with that plan).

102

Ibid., at 219–221 (describing conditions for rational reconsideration).

103

See generally M.E. Bratman, Intention, Plans, and Practical Reason (1987) (providing a more elaborate discussion on the role of plans in understanding the relationship between rationality and action).

104

Ibid., at 64–65 (discussing plan stability in order to explain the rationality of an agent’s reconsideration or non-reconsideration of a plan); see also Gauthier, ‘Commitment and Choice’, supra note 100, at 221 (‘a full appreciation of the role that plans play in deliberation requires revisions in the orthodox view of economic rationality’).

105

Bratman, supra note 103, at 2–3.

106

Cf. Gauthier, ‘Commitment and Choice’, supra note 100, at 221–222 (‘from the standpoint of the economist . . . [a]n agent’s reasons for an action are adequate just in case he prefers the expected outcome of that action no less than the expected outcome of any of its alternatives. The expected outcome of an action is the probability-weighted sum of the possible outcomes of the action’), with Gauthier, supra note 60, at 184–185.

107

Gauthier, ‘Commitment and Choice’, supra note 100, at 222.

108

See ibid., at 222–223 (discussing how the thesis that human beings are maximizing individuals can be applied to rational planning). For further discussion see generally E.F. McClennen, Rationality and Dynamic Choice: Foundational Explorations (1990) (analysing how normatively to justify principles of rationality).

109

Gauthier, ‘Commitment and Choice’, supra note 100, at 228 (challenging the view that ‘directly maximizing considerations are to be brought to bear on each particular choice’). Simply put, commitment to a plan ‘makes planning maximally efficacious in co-ordinating one’s own actions . . . with those of others, so that one may best realize one’s objectives’: ibid.

110

See ibid., at 242–243.

111

See Bratman, ‘Following Through with One’s Plans: Reply to David Gauthier’, in Danielson (ed.), supra note 100, at 55 (arguing that rational deliberation and plan stability are linked by the concept of planning). Along with Gauthier, Bratman believes that deliberation about future actions ‘is justified by appeal to its expected long-run impacts’: see ibid., at 59. Bratman concludes that reconsideration of a plan is rationally justified if the agent believes that a specific alternative will better achieve ‘the very same long-standing, stable and coherent desires and values’: ibid., at 61; see also Bratman, ‘Planning and Temptation’, in L. May et al. (eds),Mind and Morals: Essays on Cognitive Science and Ethics (1996), at 293, 294 (suggesting that coordination is impossible without stable intentions and plans).

112

For a discussion of myopic choosers see McClennen, ‘Rationality and Rules’, in Danielson (ed.), supra note 100, at 16. A formal model was first offered by Strotz in ‘Myopia and Inconsistency in Dynamic Utility Maximization’, 23 Rev Econ Studies (1955) 165, at 173.

113

Claire Finkelstein developed a version of this view in her essay, ‘Acting on an Intention’, in B. Verbeek (ed.), Reasons and Intentions (2008), at 67, 83.

114

See, e.g., Smart, ‘Extreme and Restricted Utilitarianism’, in J. Rachels (ed.), Ethical Theory (1998), at 286, 286–288.

115

Ibid., at 286–287.

116

Ibid., at 287.

117

See, e.g., Guzman, supra note 28, at 15–22.

118

Cf. Goldsmith and Posner, supra note 8, at 7–10 (addressing the methodology of rational choice theory, but mostly addressing constructivist challenges).

119

See J. Rawls, Political Liberalism (expanded edn, 1996), at 52 (‘[j]ustice as fairness . . . does not try to derive the reasonable from the rational’).

120

Cf. Rawls, supra note 62, at 16 (noting that contract terminology conveys the idea that principles of justice are principles that would be chosen by rational persons and that ‘[t]he theory of justice is a part, perhaps the most significant part, of the theory of rational choice’), with Harsanyi, ‘Morality and the Theory of Rational Behaviour’, in A. Sen and B. Williams (eds), Utilitarianism and Beyond (1982), at 39, 39 (noting that his ethical theory, while based on intellectual traditions in moral philosophy, makes essential use of the modern Bayesian theory of rational behaviour).

121

See, e.g., A. Gewirth, Reason and Morality (1978).

122

See J. Rawls, Justice as Fairness: A Restatement (ed. Erin Kelly, 2001), at 29.

123

For a discussion of the structure of the naturalistic problem regarding rationality see de Sousa, ‘Modeling Rationality: A Normative or Descriptive Task?’, in Danielson (ed.), supra note 100, at 120 (defining normativism as the claim that ‘within all attempts to model actual reasoning processes there must be an ineliminable element of normativity’).

124

See Korsgaard, ‘Personal Identity and the Unity of Agency: A Kantian Response to Parfit’, 18 Philosophy & Public Affairs (1989) 101, at 109–115 (discussing how rational, unified agents make life plans).

125

See I. Levi, The Covenant of Reason: Rationality and the Commitments of Thought (1997), at 1–19.

126

This view is arguably implicit in J. Rawls, The Law of Peoples: With ‘The Idea of Public Reason Revisited’ (2001), at 32. See also H. Kelsen, Principles of International Law (1952), at 154.

127

See also Pettit, ‘Collective Persons and Powers’, 8 Legal Theory (2002) 443.

128

For more on the irrelevance of this distinction for purposes of the commitment to rationality see C. Rovane, The Bounds of Agency: An Essay in Revisionary Metaphysics (1998), at 132.

129

Cf. Rovane, ‘What Is an Agent?’, 140 Synthese (2004) 181 (discussing group agents).

130

See I. Levi, Hard Choices: Decision Making Under Unresolved Conflict (1986), at 151 (‘even students of market economies attribute beliefs, desires, goals, values and choices to families, firms and, of course, government agencies.’).

131

See generally Franck, supra note 26 (discussing international law in an age of disparities of power).

132

See supra notes 21–22 and accompanying text.

133

See, e.g., Goldsmith and Posner, supra note 8, at 116 (discussing how weaker states can be coerced into compliance by more powerful states).

134

For a discussion of this problem see Beitz, supra note 82, at41–44, 47–48.

135

For a general discussion of how underlying geopolitical realities can preclude establishing effective rules of international law see Glennon, supra note 44.

136

See, e.g., Goldsmith and Posner, supra note 8, at 66–75 (discussing patterns of state compliance – and lack thereof – with regard to the customary international law exempting fishing vessels from right of capture during times of war); ibid., at 116–117 (discussing strategic coercion, ‘often in violation of international law’ by stronger states to make weaker states comply with human rights norms when such compliance is in the interest of the stronger states).

137

Ibid., at 90 (‘we have explained the logic of treaties without reference to notions of “legality” or pacta sunt servanda or related concepts. As was the case with customary international law, the cooperation and coordination models explain the behaviors associated with treaties without reliance on these factors, or on what international lawyers sometimes call “normative pull”’).

138

See, e.g., Frischmann, ‘A Dynamic Institutional Theory of International Law’, 51 Buffalo L Rev (2003) 679, at 681; Goodman and Jinks, ‘Measuring the Effects of Human Rights Treaties’, 14 EJIL (2003) 171; Scharf, ‘International Law in Crisis: A Qualitative Empirical Contribution to the Compliance Debate’, 31 Cardozo L Rev (2009) 45.

139

See Goldsmith and Posner, supra note 8, at 186.

140

Ibid., at 187–188.

141

Ibid., at 188 (citing C. Kutz, Complicity: Ethics and Law for a Collective Age (2000), at 253).

142

See J. Locke, Two Treatises of Government ss. 119–122, ii, at 347–349 (1690, ed. Peter Laslett, rev. edn, 1988).

143

See Goldsmith and Posner, supra note 8, at 189.

144

See ibid., at 187–188.

145

See ibid., at 190–191.

146

See ibid., at 189.

147

Ibid., at 191.

148

Ibid.

149

The conclusion that the autonomy of states (and nations) has no intrinsic value can and should be resisted, though a full account is impossible here. See, e.g., C. Taylor, Sources of the Self: The Making of the Modern Identity (1989) (presenting a history of the modern identity); see also W. Kymlicka, States, Nations and Cultures (1997) (arguing that group rights are derived from enlightenment commitment to individual flourishing); Buchanan, ‘Democracy and the Commitment to International Law’, 34 Georgia J Int’l & Comp L (2006) 305, at 320; Margalit and Raz, ‘National Self-Determination’, 87 J Philosophy (1990) 439, at 443.

150

See Kymlicka, supra note 149, at 35.

151

For a full defence of the value of collectivities see G.P. Fletcher and J.D. Ohlin, Defending Humanity: When Force Is Justified and Why (2008).

152

See ibid., at 136–147.

153

See F.K. von Savigny, System des Heutigen Römischen Rechts (1840), ii, at 236 (asserting that juridical persons are fictitious but are nevertheless entitled to rights by extension).

154

Cf. J. Goldsmith, The Terror Presidency: Law and Judgment Inside the Bush Administration (2007).