23 February 2004 The Customary International Law Supergame Order and law George Norman Joel P. trachtman Abstract Customary international law is an enigma. It is produced by the decentralized actions of states, and it generally lacks centralized enforcement mechanisms. Political science realists and some rationalist legal scholars argue that customary international law cannot affect state behavior: that it is"epiphenomenal. This article develops a model of an n-player prisoner's dilemma in the customary international law context that shows that it is plausible that states would comply with customary international law under certain circumstances. These circumstances relate to: (i)the relative value of cooperation versus defection, (ii) the number of states effectively involved, (iii) the extent to which increasing the number of states involved increases the value of cooperation or the detriments of defection, including whether the particular issue has characteristics of a commons problem, a public good, or a network, (iv) the information available to the states involved regarding compliance and defection, (v)the relative patience of states in valuing the benefits of long-term cooperation compared to short-term defection, (vi) the expected duration of interaction, (vii)the frequency of interaction, and(viii)whether there are also bilateral relationships or other multilateral relationships between the involved states One implication of this model is to lend credence to customary international law From a research standpoint, this model identifies a number of parameters for which data may be developed in order to test the model. From a policy standpoint, this model shows what types of contexts, including malleable institutional features, may affect the ability of states to reach stable and efficient equilibria in their customary international law William and Joyce Cummings Professor of Entrepreneurship and Business Economics Tufts University Professor of International Law, The Fletcher School of Law and Diplomacy, Tufts o George Norman and Joel Trachtman. All rights reserved. Preliminary draft. Please do not cite or further circulate without the authors permission
23 February 2004 George Norman * Joel P. Trachtman ** Abstract Customary international law is an enigma. It is produced by the decentralized actions of states, and it generally lacks centralized enforcement mechanisms. Political science realists and some rationalist legal scholars argue that customary international law cannot affect state behavior: that it is “epiphenomenal.” This article develops a model of an n-player prisoner’s dilemma in the customary international law context that shows that it is plausible that states would comply with customary international law under certain circumstances. These circumstances relate to: (i) the relative value of cooperation versus defection, (ii) the number of states effectively involved, (iii) the extent to which increasing the number of states involved increases the value of cooperation or the detriments of defection, including whether the particular issue has characteristics of a commons problem, a public good, or a network, (iv) the information available to the states involved regarding compliance and defection, (v) the relative patience of states in valuing the benefits of long-term cooperation compared to short-term defection, (vi) the expected duration of interaction, (vii) the frequency of interaction, and (viii) whether there are also bilateral relationships or other multilateral relationships between the involved states. One implication of this model is to lend credence to customary international law. From a research standpoint, this model identifies a number of parameters for which data may be developed in order to test the model. From a policy standpoint, this model shows what types of contexts, including malleable institutional features, may affect the ability of states to reach stable and efficient equilibria in their customary international law relations. * William and Joyce Cummings Professor of Entrepreneurship and Business Economics, Tufts University. ** Professor of International Law, The Fletcher School of Law and Diplomacy, Tufts University. © George Norman and Joel Trachtman. All rights reserved. Preliminary draft. Please do not cite or further circulate without the authors’ permission. The Customary International Law Supergame: Order and Law
The Customary International Law Supergame February 23, 2004 1. Introduction Customary international law( CIL) forms the infrastructure of international lay and is also part of international laws superstructure. It thus serves as its own foundation but also as the foundation for treaty law, and therefore for essentially all international law. How firm is this foundation CIL Seems to have fallen on hard times, as it has been overtaken by treaty as the principal source of new rules in the international community, as it often seems to be used by idealists to make boot-strapping arguments about the content of international law, and as its binding force---its social effect--is often not readily apparent. This at a time when the domestic analog of CIL-social norms---is the darling of legal scholarship CiL is under attack as behaviorally epiphenomenal and doctrinally incoherent By contrast, the central argument of this article is that CIL, while something of a trick of levitation, is theoretically plausible. Our argument makes one central claim: that while there are limits on and variations in the effectiveness of cil. there are circumstances where it may independently affect the behavior of states. There is no reason in theory, or in data adduced by others, to believe CiL generally epiphenomenal. In addition, as a by- product of the analysis supporting our central claim, we find that the cil doctrine that has developed in order to discriminate between valid and invalid assertions of the existence of rules of CiL is internally coherent, and consistent with our models rationalist perspective on state behavior See generally, J. Patrick Kelly, The Twilight of Customary International Law, 40VA J. INTLL.449(2000)
The Customary International Law Supergame February 23, 2004 1 1. Introduction Customary international law (CIL) forms the infrastructure of international law, and is also part of international law’s superstructure. It thus serves as its own foundation, but also as the foundation for treaty law, and therefore for essentially all international law. How firm is this foundation? CIL seems to have fallen on hard times, as it has been overtaken by treaty as the principal source of new rules in the international community, as it often seems to be used by idealists to make boot-strapping arguments about the content of international law, and as its binding force—its social effect—is often not readily apparent. 1 This at a time when the domestic analog of CIL—social norms—is the darling of legal scholarship. CIL is under attack as behaviorally epiphenomenal and doctrinally incoherent. By contrast, the central argument of this article is that CIL, while something of a trick of levitation, is theoretically plausible. Our argument makes one central claim: that while there are limits on and variations in the effectiveness of CIL, there are circumstances where it may independently affect the behavior of states. There is no reason in theory, or in data adduced by others, to believe CIL generally epiphenomenal. In addition, as a byproduct of the analysis supporting our central claim, we find that the CIL doctrine that has developed in order to discriminate between valid and invalid assertions of the existence of rules of CIL is internally coherent, and consistent with our model’s rationalist perspective on state behavior. 1 See generally, J. Patrick Kelly, The Twilight of Customary International Law, 40 VA. J. INT'L L. 449 (2000)
The Customary International Law Supergame February 23, 2004 This article refines and extends an emerging rationalist understanding of CIL Pioneering work in this field, notably that of Jack Goldsmith and Eric Posner, has begun to articulate a rationalist theory of CIL. Goldsmith and posner have argued that CIL(a) does not exogenously affect state behavior and (b) is internally incoherent, as states not motivated by opinio juris to comply with CiL.' We show why the first assertion is either tautological or incorrect as a matter of theory, and, to the extent that it purports to rely on factual observation, is not supported by the data presented. The tautology arises from a false dichotomy between motivation by self-interest and motivation by law. In a ationalist model, behavior is assumed to be motivated by self-interest. If law is artificially separated from self-interest, then of course a rationalist model would assume that law has no motivating force. Yet we show how CiL rules may modify the payoffs of behavior and therefore affect behavior through self-interest. While cil is endogenous to states as a group, it is an exogenous influence on the behavior of each individual state CIL may affect behavior even if it only does so at the margins. As to the second assertion, our analysis provides a plausible basis to assign a discriminating role to opinio juris, and therefore we find cil doctrine internally coherent in at least its core dimension This article develops an iterated multi-player prisoners dilemma model of CIL. Of course, game theory can only provide a crude representation of highly nuanced decision-making. Our analysis focuses on parameters of the multi-player prisoners dilemma in the CIL context: (i)the relative value of cooperation versus defection, (ii) the number of states effectively involved, (iii) the extent to which increasing the number of See, e.g., Mark A Chinen, Afterword, 23 MICH. J INT'L L 201(2001); Mark A Chinen, Game Theory and Customary International Law: A Response to Professors Goldsmith and Posner, 23 MICH. J INT'L L 143 (2001): Vincy Fon francesco Parisi, International Customary Law and Articulation Theories: An Economic Analysis, George Mason Law and Economics Research Paper No. 02-24, available at http:papersssrn.com/sol3/papers.cfm?abstractid=335220;JackL.Goldsmith&ericA Posner, Further Thoughts on Customary International Law, 23 MICH. J INTL L. 191 (2001); Jack L. Goldsmith& Eric A Posner, A Theory of Customary International Law, 66 U. CHI. L REv. 1113(1999) hereinafter Goldsmith Posner 1999]; Andrew T. Guzman, A Compliance-Based Theory of international Law, 90 CAL L REV. 1823 (2002); Francesco Parisi, The Formation of Customary Law, George Mason Law and Economics Research Paper No. O1-06, http:/papers.ssrn.com/sol3/papers.cfm?abstractid=262032;edWardt.SwaineRational Custom, 52: 3 DUKE L.J. 559(2002); Pierre-Hugues Verdier, Cooperative States International Relations, State Responsibility and the Problem of custom, 42 VA L REV 839(2002) he leading article here is goldsmith posner 1999, supra note 2. pcc "self-interest "we mean merely to refer to maximization of preferences--these rences could be other-regarding or altruistic See the cautions expressed in Goldsmith Posner 1999, supra note 2, and the broader treatment in Duncan Nidal, The Game Theory of International Politics, in Kenneth A Oye, ed. COOPERATION UNDER ANARCHY (1986)
The Customary International Law Supergame February 23, 2004 2 This article refines and extends an emerging rationalist understanding of CIL. 2 Pioneering work in this field, notably that of Jack Goldsmith and Eric Posner, has begun to articulate a rationalist theory of CIL. Goldsmith and Posner have argued that CIL (a) does not exogenously affect state behavior and (b) is internally incoherent, as states are not motivated by opinio juris to comply with CIL. 3 We show why the first assertion is either tautological or incorrect as a matter of theory, and, to the extent that it purports to rely on factual observation, is not supported by the data presented. The tautology arises from a false dichotomy between motivation by self-interest 4 and motivation by law. In a rationalist model, behavior is assumed to be motivated by self-interest. If law is artificially separated from self-interest, then of course a rationalist model would assume that law has no motivating force. Yet we show how CIL rules may modify the payoffs of behavior and therefore affect behavior through self-interest. While CIL is endogenous to states as a group, it is an exogenous influence on the behavior of each individual state. CIL may affect behavior even if it only does so at the margins. As to the second assertion, our analysis provides a plausible basis to assign a discriminating role to opinio juris, and therefore we find CIL doctrine internally coherent in at least its core dimension. This article develops an iterated multi-player prisoner’s dilemma model of CIL. Of course, game theory can only provide a crude representation of highly nuanced decision-making. 5 Our analysis focuses on parameters of the multi-player prisoner’s dilemma in the CIL context: (i) the relative value of cooperation versus defection, (ii) the number of states effectively involved, (iii) the extent to which increasing the number of 2 See, e.g., Mark A. Chinen, Afterword, 23 MICH. J. INT'L L. 201 (2001); Mark A. Chinen, Game Theory and Customary International Law: A Response to Professors Goldsmith and Posner, 23 MICH. J. INT'L L. 143 (2001); Vincy Fon & Francesco Parisi, International Customary Law and Articulation Theories: An Economic Analysis, George Mason Law and Economics Research Paper No. 02-24, available at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=335220; Jack L. Goldsmith & Eric A. Posner, Further Thoughts on Customary International Law, 23 MICH. J. INT'L L. 191 (2001); Jack L. Goldsmith & Eric A. Posner, A Theory of Customary International Law, 66 U. CHI. L. REV. 1113 (1999) [hereinafter Goldsmith & Posner 1999]; Andrew T. Guzman, A Compliance-Based Theory of International Law, 90 CAL. L. REV. 1823 (2002); Francesco Parisi, The Formation of Customary Law, George Mason Law and Economics Research Paper No. 01-06, http://papers.ssrn.com/sol3/papers.cfm?abstract_id=262032; Edward T. Swaine, Rational Custom, 52:3 DUKE L.J. 559 (2002); Pierre-Hugues Verdier, Cooperative States: International Relations, State Responsibility and the Problem of Custom, 42 VA. L. REV. 839 (2002). 3 The leading article here is Goldsmith & Posner 1999, supra note 2. 4 By “self-interest” we mean merely to refer to maximization of preferences—these preferences could be other-regarding or altruistic. 5 See the cautions expressed in Goldsmith & Posner 1999, supra note 2, and the broader treatment in Duncan Snidal, The Game Theory of International Politics, in Kenneth A. Oye, ed., COOPERATION UNDER ANARCHY (1986)
The Customary International Law Supergame February 23, 2004 states involved increases the value of cooperation or the detriments of defection, including whether the particular issue has characteristics of a commons problem, a public good, or a network, (iv) the information available to the states involved regarding compliance and defection, (v) the relative patience of states in valuing the benefits of long-term cooperation compared to short-term defection, (vi) the expected duration of interaction, (vii)the frequency of interaction, and(viii) whether there are also bilateral relationships or other multilateral relationships between the involved states The parameters we identify are incorporated in our model as independent variables, but from a normative standpoint it is possible for policy initiatives to select or manipulate these parameters. That is, by developing a positive theory suggesting the parameters for determining whether Cil will affect state behavior, this article opens the way to normative institutional design. States may determine to restructure certain nstitutions in order to facilitate the formation and operation of CiL. There may be circumstances under which it will be normatively attractive to facilitate the development of CIL, rather than to engage in more self-conscious and static treaty-making. The institutional dynamism and social immanence that attracts some scholars to social norms n the domestic context may also be attractive in at least some international contexts This article is organized as follows. The remainder of this introduction provides a short definition of CIL and briefly locates this article in relation to four literatures: law and economics, social norms, international organization and industrial organization. Part 2 explains our choice of the iterated n-person prisoner's dilemma as the basis for our model, as well as our choice of an assumed strategy for players within this model. Part 3 explains the other assumptions and parameters of our model. Part 4 sets forth the formal model. Part 5 presents some implications in terms of (i) the plausibility of CIL, (ii) policy, (iii)international law doctrine, and (iv) research. Part 6 briefly concludes a. Customary International Law and its Doctrine As an introductory matter, it is useful to review the fundamental definition of CIL. Article 102 of the American Law Institutes Restatement (Third) of Foreign Relations Law states that"Customary international law results from a general and consistent practice of states followed by them from a sense of legal obligation. The sense of obligation is referred to in Latin as"opinio juris sive necessitatis. " Articl 38(1(b) of the Statute of the International Court of Justice, listing the sources of international law applicable by the Court, includes"international custom, as evidence of a general practice accepted as law. CiL has two core doctrinal problems relating to opinio juris. First, can CIL ever come into existence if it requires opinio juris-a sense of legal obligation--before it can exist? The second relates to the motivation of states. Are states ever motivated by opinio Daniel A. Farber, Positive Theory as Normative Critique, 68 S CAL L REV. 1565 (1995)
The Customary International Law Supergame February 23, 2004 3 states involved increases the value of cooperation or the detriments of defection, including whether the particular issue has characteristics of a commons problem, a public good, or a network, (iv) the information available to the states involved regarding compliance and defection, (v) the relative patience of states in valuing the benefits of long-term cooperation compared to short-term defection, (vi) the expected duration of interaction, (vii) the frequency of interaction, and (viii) whether there are also bilateral relationships or other multilateral relationships between the involved states. The parameters we identify are incorporated in our model as independent variables, but from a normative standpoint it is possible for policy initiatives to select or manipulate these parameters. That is, by developing a positive theory suggesting the parameters for determining whether CIL will affect state behavior, this article opens the way to normative institutional design. 6 States may determine to restructure certain institutions in order to facilitate the formation and operation of CIL. There may be circumstances under which it will be normatively attractive to facilitate the development of CIL, rather than to engage in more self-conscious and static treaty-making. The institutional dynamism and social immanence that attracts some scholars to social norms in the domestic context may also be attractive in at least some international contexts. This article is organized as follows. The remainder of this introduction provides a short definition of CIL and briefly locates this article in relation to four literatures: law and economics, social norms, international organization and industrial organization. Part 2 explains our choice of the iterated n-person prisoner’s dilemma as the basis for our model, as well as our choice of an assumed strategy for players within this model. Part 3 explains the other assumptions and parameters of our model. Part 4 sets forth the formal model. Part 5 presents some implications in terms of (i) the plausibility of CIL, (ii) policy, (iii) international law doctrine, and (iv) research. Part 6 briefly concludes. a. Customary International Law and its Doctrine As an introductory matter, it is useful to review the fundamental definition of CIL. Article 102 of the American Law Institute’s Restatement (Third) of Foreign Relations Law states that “[c]ustomary international law results from a general and consistent practice of states followed by them from a sense of legal obligation.” The sense of obligation is referred to in Latin as “opinio juris sive necessitatis.” Article 38(1)(b) of the Statute of the International Court of Justice, listing the sources of international law applicable by the Court, includes “international custom, as evidence of a general practice accepted as law.” CIL has two core doctrinal problems relating to opinio juris. First, can CIL ever come into existence if it requires opinio juris—a sense of legal obligation—before it can exist? The second relates to the motivation of states. Are states ever motivated by opinio 6 Daniel A. Farber, Positive Theory as Normative Critique, 68 S. CAL. L. REV. 1565 (1995)
The Customary International Law Supergame February 23, 200 juris? The rationalist assumption of states motivated by their own preferences seems prima facie inconsistent with the Cil requirement of motivation by opinio juris, and this is the criticism that Goldsmith and Posner level at CIL doctrine. They argue that what appears to be cil is actually motivated by coincidence of interest, coercion, cooperation predicated on self-interest, or coordination predicated on self-interest. We respond to thi argument in part 5. For the social scientist studying law, the critical descriptive question relates to the effects of legal rules on behavior. For CIL, this descriptive question is also a doctrinal question, as Cil doctrine requires some level of generality and consistency of practice some quantum of state behavior. And again, at least under the restatement formulation, this behavior must be motivated by opinio juris. Under the Statute of the International Court of Justice, the custom itself may serve as evidence of acceptance as law, of opinio juris. Some commentators have suggested that opinio juris in a formal sense may not be necessary at all, but that the requirement should be understood in terms of state consent or acceptance As we address the theory of CIL, it is useful to have in mind some examples CIL. CIL exists in a wide variety of fields, and coexists in many areas with treaty law As already noted, CIL is its own foundation. Thus, there is a CiL of CIL. In other words, the cil doctrine discussed above is understood as law pursuant to CIL Furthermore, the law of treaties came into existence pursuant to CIL, although this law is today codified in treaty. But moving beyond these foundational areas, CIL addresses, for example, issues of territorial integrity, jurisdiction to apply law, diplomatic immunity, the rights of states with respect to coastal areas, human rights, cross-border pollution and the use of force to settle international disputes These are diverse fields. each with its own characteristics. Our theoretical approach calls for discrimination among these fields. The parameters for discrimination are indicated by the model set forth below. Simply put, we would expect greater possibility for formation of and compliance with Cil in some fields than in others. Of course, this is an area in which empiricism would require analysis of areas in which CiL has not developed: the dog that did not bark. Thus, we might ask, why is there no Cl rule that prohibits over-fishing in international waters? Why is there no Cil rule that requires extradition of criminals? Our theoretical approach also accepts the possibility of linkage among diverse fields, integrating and therefore homogenizing the behavioral effect of each rule of law b. Four literatures 7 Goldsmith Posner 1999 .supra note 2 Maurice H Mendelson, The Formation of Customary International Law, 272 RECUEIL DES COURS 155, 268-293(1998)
The Customary International Law Supergame February 23, 2004 4 juris? The rationalist assumption of states motivated by their own preferences seems prima facie inconsistent with the CIL requirement of motivation by opinio juris, and this is the criticism that Goldsmith and Posner level at CIL doctrine. 7 They argue that what appears to be CIL is actually motivated by coincidence of interest, coercion, cooperation predicated on self-interest, or coordination predicated on self-interest. We respond to this argument in part 5. For the social scientist studying law, the critical descriptive question relates to the effects of legal rules on behavior. For CIL, this descriptive question is also a doctrinal question, as CIL doctrine requires some level of generality and consistency of practice— some quantum of state behavior. And again, at least under the Restatement formulation, this behavior must be motivated by opinio juris. Under the Statute of the International Court of Justice, the custom itself may serve as evidence of acceptance as law, of opinio juris. Some commentators have suggested that opinio juris in a formal sense may not be necessary at all, but that the requirement should be understood in terms of state consent or acceptance. 8 As we address the theory of CIL, it is useful to have in mind some examples of CIL. CIL exists in a wide variety of fields, and coexists in many areas with treaty law. As already noted, CIL is its own foundation. Thus, there is a CIL of CIL. In other words, the CIL doctrine discussed above is understood as law pursuant to CIL. Furthermore, the law of treaties came into existence pursuant to CIL, although this law is today codified in treaty. But moving beyond these foundational areas, CIL addresses, for example, issues of territorial integrity, jurisdiction to apply law, diplomatic immunity, the rights of states with respect to coastal areas, human rights, cross-border pollution and the use of force to settle international disputes. These are diverse fields, each with its own characteristics. Our theoretical approach calls for discrimination among these fields. The parameters for discrimination are indicated by the model set forth below. Simply put, we would expect greater possibility for formation of and compliance with CIL in some fields than in others. Of course, this is an area in which empiricism would require analysis of areas in which CIL has not developed: the dog that did not bark. Thus, we might ask, why is there no CIL rule that prohibits over-fishing in international waters? Why is there no CIL rule that requires extradition of criminals? Our theoretical approach also accepts the possibility of linkage among diverse fields, integrating and therefore homogenizing the behavioral effect of each rule of law. b. Four Literatures 7 Goldsmith & Posner 1999, supra note 2. 8 Maurice H. Mendelson, The Formation of Customary International Law, 272 RECUEIL DES COURS 155, 268-293 (1998)