1
COMPLIANCE WITH INTERNATIONAL AGREEMENTS

Beth A. Simmons

A central theme in much recent international relations scholarship is the growing role of formal international agreements and supranational authority in the ordering of relations among sovereign states. Evidence of the growing range of authoritative commitments is evidenced by the movement since World War II to codify customary practices into explicit international legal instruments. The range of international agreements has grown rapidly over the past forty years with the development of rules that regulate economic, social, communications, environmental and human rights behavior. Evidence of the growth in supranational authority is provided not merely by the number of international organizations that have mushroomed in the postwar years, but also, most strikingly, in the growth and development of legally binding forms of third party dispute settlement: the evolution of the GATT's dispute settlement procedures into the more formal and less discretionary structure of the World Trade Organization, the 1996 inauguration of the International Maritime Court in Hamburg to handle disputes arising from the United Nations' Law of the Seas, the growing activism of the European Court of Justice, and the recent flurry of contentious case activity at the International Court of Justice1 are all examples of states agreeing voluntarily to give up a portion of their most basic aspect of sovereignty - the authority to act as the final judge of one's own actions - to authoritative international institutions. These developments are something of a puzzle for the study of international relations, the traditional assumption of which has been that national governments generally desire to preserve their legal sovereignty, particularly the sole authority to judge the acceptability of their policies in the international sphere. In much main line theorizing, states are viewed as making commitments - especially formal legal commitments - either cautiously or cynically, and are

During the Cold War, the Court decided only one contentious case on average per year; in 1995, however, the Court had a record number of 13 cases before it.

2
typically reluctant to delegate decision making to supranational bodies of any kind. Over the past two decades, a good deal of theoretical and empirical work has been devoted to explaining why states have entered into this vast web of agreements voluntarily. Explanations have focused on the functional need for agreements, given the growing level of interdependence (Keohane 1984), the desire for greater regularity and predictability in actors' mutual relations (Brierly 1963) and growing state responsibility in the economic and social realm generally (Röling 1960, Friedmann 1964). Dominant international relations paradigms would agree in general that governments usually agree to sacrifice a degree of their legal freedom of action in order to secure policy changes from others, or influence over their policies (Keohane 1993). Until recently, far less attention has been devoted to understanding why governments actually comply with such agreements, given that they can be costly in the short term and are not very likely to be centrally enforced. Four broad approaches to this question are reviewed here: realist theory, rational functionalism, domestic regime-based explanations, and normative approaches. These are not posed as mutually exclusive perspectives, and the less one is willing to straw-man the arguments of the major proponents the clearer become the numerous points of overlap. For example, while realist theory has rarely been articulated in such as way as to take international legal constraints seriously, some of its major proponents would admit that international law compliance is fairly pervasive (Morgenthau 1985). Similarly, scholars that focus on normative convergence as a source of compliance hardly rule out coercive processes to encourage such convergence (Bull 1977). Certainly, approaches that link domestic regime type with international rule compliance often tap into a deeper set of factors relating to the role of liberal principles and beliefs in securing international behavior consistent with the rule of law (Dixon 1993). And functionalist arguments can and have been made that point to domestic regime characteristics as a source of "market failure" that make international agreements all the more necessary. Nonetheless, these four broad approaches diverge in important respects and provide a useful way to arrange the growing literature on compliance with international agreements.

3
This review reveals that despite the recent interest in issues surrounding compliance and the effects of rules on international politics more generally, the effort to link theory with evidence is still in its infancy. Part of the difficulty flows from conceptual difficulties identifying compliance itself. Another problem has been methodological: difficulties in demonstrating causation, problems of selection bias in the use of cases and the analysis of data have been pervasive and difficult to remedy. Since the endeavor to understand compliance has been interdisciplinary, involving legal scholars and sociologists as well as political scientists, differing methods of analysis, reasoning, and standards of proof pervade the literature. While these differences have been enriching and have narrowed through scholarly cooperation, they do help account for the disparate nature of much of the relevant literature. This essay proceeds as follows. Section I discusses the concept of compliance, and presents some strategies for its measurement. Section II reviews four strands of international relations theory (inserting legal scholarship where arguments are compatible, even if they are not self-consciously writing within the tradition under examination) and culls from them a range of explanations and empirical findings regarding international legal commitments and compliance. Section III draws some conclusions about our state of knowledge regarding compliance with international agreements and suggests some directions for future research.

THE MEANING AND MEASUREMENT OF COMPLIANCE In his ground breaking study on compliance with international public authority, Oran Young (1979) suggested that "Compliance can be said to occur when the actual behavior of a given subject conforms to prescribed behavior, and non-compliance or violation occurs when actual behavior departs significantly from prescribed behavior." This definition distinguishes compliance behavior from treaty implementation (the adoption of domestic rules or regulations meant to facilitate but which themselves do not constitute compliance with international agreements). It also distinguishes compliance from effectiveness, since it is entirely possible that a poorly designed agreement could achieve high levels of compliance without much impact on

4
the phenomenon of concern (pollution levels for example). While compliance may be necessary for effectiveness, there is no reason to view it as sufficient. The literature reviewed here is almost exclusively concerned with the conformity of behavior to rules, rather than the ultimate outcome of such conformity (Jacobson & Brown Weiss 1995, 1997). Furthermore, most of the literature reviewed here is concerned with compliance with explicit rules or agreements, often of a legal character or of normative import, and not with "compliance" with the demands of an adversary or requests of an ally. The concern has typically been with obligations that flow from authoritative agreements, widely held normative prescriptions, or authoritative interpretations of proper behavior, rather than acquiescence to unilateral political demands based on the exercise of power alone.2 An important distinction also needs to be made between what Fisher (1981) refers to as "first order" versus "second order" compliance. The former refers to compliance with standing, substantive rules often embodied in treaty arrangements (Downs & Rocke 1995, Chayes & Chayes, 1995). Second order compliance refers to compliance with an authoritative decision of a third party such as a panel of the World Trade Organization, the United Nations Human Right Committee, or the International Court of Justice (Bulterman & Kuijer 1996, Fisher 1981). The study of first order compliance raises difficulties of establishing an underlying "rate" of compliance, since it is far from clear how to conceptualize a denominator for such a rate. As a result, researchers looking at the same set of behaviors can disagree vehemently over whether "most" foreign policy actions are effectively governed by law, rules, and agreements, or whether such considerations have little effect on state behavior (Henkin 1979). Studies of second order compliance can often more convincingly establish such a rate, as well as narrow the range of behavior that would constitute compliance by focusing on a particular, often precisely rendered, decision. (Fisher 1981). Unfortunately cases of rulings represent only the tip of the iceberg of the larger compliance problem, and are likely to represent a biased set of observations, especially
2

In practice, of course, agreements among asymmetrically endowed actors are rarely perfectly voluntary, and the decision to "conform to prescribed behavior" might rest on an amalgam of obligation and felt coercion.

5
since only governments willing to make concessions are likely to submit to authoritative decision-making processes (Coplin 1968). Finally, most researchers will readily admit it is very difficult to judge whether a particular policy constitutes compliance at all (Jacobson & Brown Weiss 1997). Often international agreements are written so as to permit a range of interpretations regarding the parties' obligations. Furthermore, compliance is rarely a transparent, binary choice. Often actors will behave in ways that are intentionally ambiguous, dilatory, or confusing, frequently under conditions in which verification is difficult (Young 1979). In other contexts, actors may make good faith efforts to comply that nonetheless fall short of an agreement's prescribed behavior. Some researchers have dealt with these difficulties by making such assessments in the context of generally prevailing expectations (Chayes & Chayes 1995). Going further, constructivist approaches assume that standards of compliance are socially constructed, and must not be imposed by the analyst, making each assessment highly context specific.

INTERNATIONAL RELATIONS THEORY AND THE PROBLEM OF COMPLIANCE Compliance and the Realist Tradition For realists, power, rather than law, has traditionally been the primary determinate of the course of interstate relations. Most realists - theoreticians and practitioners - tend to be highly skeptical that treaties or formal agreements influence state actions in any important way (Boyle 1980, Bork 1989/90). While Hans Morgenthau (1985) was ready to admit that "during the four hundred years of its existence international law ha[d] in most instances been scrupulously observed" he thought that this could be attributed either to convergent interests or prevailing power relations. Governments make legal commitments cynically, and "are always anxious to shake off the restraining influence that international law might have upon their foreign policies, to use international law instead for the promotion of their national interests...." (Morgenthau 1985).3

See also Aron, 1981, 110: "...juridical interpretation, even when it is concretely improbable...is utilized as a means of diplomatic pressure."

6
Similarly, Stanley Hoffmann (1956) described the nation state as "a legally sovereign unit in a tenuous net of breakable obligations." In this formulation, there is little relationship between what governments may be legally bound to do or refrain from doing, and their actual behavior, except insofar as this results from a coincidence of law and national interest. Raymond Aron (1981) put it succinctly: "International law can merely ratify the fate of arms and the arbitration of force." The decentralized nature of the international legal system is viewed by realists as its prime defect. International agreements lack restraining power, especially since governments generally retain the right to interpret and apply the provisions of international agreements selectively (Morgenthau 1985). Major powers and the pursuit of important interests are viewed as highly unlikely to be constrained by legal authority or prior agreement. In short, realists have typically assumed that international law is merely an epiphenomenon of interests or only made effective through the balance of power (Oppenheim 1912). Aron and other realists were prepared to admit that "the domain of legalized interstate relations is increasingly large" but argued that "one does not judge international law by peaceful periods and secondary problems." (Aron 1981). In the realm of high politics, realists have been especially skeptical about the rule of law and legal processes in international relations (see for example Diehl 1996, Fisher 1981, Fischer, 1982, Bulterman & Kuijer 1996). For the most part, realist perspectives have focused on the fundamental variables of power and interest, rarely feeling compelled to inquiry further into states' compliance with international agreements.

Compliance and Rational Functionalism A different set of expectations is suggested by a more functional approach to the study of international institutions. Most basically, functionalist approaches view international agreements as a way to address a perceived need: international legal agreements are made because states want to solve common problems that they have difficulties solving any other way, e.g., unilaterally or through political means alone (Bilder, 1989: 492). Rational functionalism shares realists' concern

7
with the incentives states faced to comply with or disregard international agreements, but has been far less likely to denigrate the cooperative problems that pervaded the realm of so-called "low politics." Certainly by the mid-1980s it was increasingly difficult to relegate economic, social, and environmental issues to the status of "secondary problems," given the robustness among developed countries of the postwar peace. Unlike realist theorists, those taking a functional approach did not assume that these issues posed trivial problems for compliance. On the other hand, rational functionalism starts not from the assumption that states cynically manipulate their legal environment, but that they "engineer" it in what are taken to be generally sincere efforts to address an otherwise suboptimal outcome. Both realism and rational functionalism are interest-driven approaches in which incentives play a crucial role.4 The latter has, however, been far more willing to view the particular agreements and even the international legal system in toto as a collective good, from which states collectively can benefit, but to which none wants to contribute disproportionately and by which none wants to consistently be disadvantaged. The focus of analysis in this approach has been on the perceived benefits of a system of rule based behavior, and the individual incentives states face to contribute to, or detract from, such a system. Because they are often crucial to getting solutions, agreements are taken seriously. In the absence of severe unresolved collective action problems or overwhelming incentives to defect that have not been addressed, obligations are therefore likely to be carried out. Though functionalist theories have concentrated largely on why states obligate themselves rather than why they comply with their obligations, theorists in this vein have

There are important differences between and within these approaches regarding the role of sanctions versus the use of incentives to "manage" the process of compliance. Chayes and Chayes (1995) emphasize that international law essentially has an important persuasive function, and to enhance compliance scholars and practitioners should move away from an enforcement model which focuses on sanctions and punishments to a management model that focuses on positive incentives and negotiation to achieve compliance. Critics respond that such a "managerial approach" to compliance will only go so far; that deep cooperation - agreements that proscribe behavior that is truly difficult to forswear or prescribe behavior that is costly in the short term - will require some form of enforcement (Downs et al 1996). The distinction between enforcement and management approaches is often made in the context of domestic law enforcement (Hawkins & Thomas 1984, Snavely 1990).

8
suggested a number of mechanisms that potentially influence compliance behavior. The central mechanism for securing compliance is related to reputation: states anticipate that they will pay a higher cost in the long run for agreements they break in an effort to achieve immediate gain (Keohane 1984, Schachter 1991). Indeed, one function of international agreements is to enhance the reputational consequences of noncompliant behavior, by providing mechanisms that increase transparency and therefore improve information regarding other states' behavior (Keohane 1984, Milgrom et al 1990, Mitchell 1994). Some authors have argued that reputational explanations for compliance are especially relevant for new and developing countries, who have an interest in developing a reputation as "rule of law" countries (Shihata 1965). Greater transparency and

opportunities for reciprocity also enhance compliance where there is repeated play amongst a small group, for example, in the EU or among the large countries in the WTO. Functionalist accounts often emphasize that international institutions play a crucial role in providing a clear focal point for acceptable behavior (Garrett & Weingast 1993). A such, they facilitate the convergence of expectations and reduce uncertainty about other states' future behavior. In common with the more normative research agenda discussed below, some scholars point out that the focal points created by international agreements or institutions can actually gain a high degree of legitimacy both internationally and domestically (Franck 1990, Peck 1996, Tacsan 1992). This legitimacy, in turn, can have important political consequences (Claude 1966). In this view, the search for a legitimate rule is a rational response to the need to find a stable solution to an otherwise costly, intractable problem or dispute. The starting point for functionalist explanations of international agreements and compliance relates to the inability of states to solve a problem without the institutional device. Like realists, most functionalist approaches to international politics begin with the premise that states delegate sovereignty begrudgingly. Certainly, there is a strong preference for solving international controversies through political means, even unilaterally if necessary. However, functional theories recognize that for a number of reasons, this may not be possible. Most

9
functional theorizing has been systemic, focusing largely on international market failure and problems of collective action (Keohane 1984). Relatively little attention has been given to the domestic political reasons why international agreement may not be possible in the absence of an international institution, but in principle, the source of the "suboptimality" in functional theory could be domestic or systemic in nature. One important exception to the systemic focus of most functional accounts of compliance is a theoretical study of the GATT rules by Downs and Rocke (1995). The central contribution of this work is that uncertainty regarding domestic politics and interest group demands have a tremendous influence on the nature of international agreements undertaken, the severity of sanctions required, and hence the degree of compliance to be expected from participants in the GATT regime. Specifically, the authors argue that GATT's weak enforcement norm is a result of uncertainty about the future demands of interest groups: most states do not want aggressive enforcement of the GATT because there may be conditions in the future under which they themselves will be compelled for domestic reasons to violate GATT obligations (Downs & Rock, 1995). This uncertainty tilts preferences toward shorter and less stringent punishments, which in turn reduces the cooperative demands of the treaty agreement (as the costs of defection rise, a highly cooperative treaty becomes too risky to be practical). In this model, arrived at deductively through a game theoretical representation, domestic uncertainty makes it more difficult to punish rule violation, which makes it more difficult to secure compliance, the expectation of which contributes to a watering down of the agreement in the first place. Other studies that locate the source of suboptimality at the domestic level have focused less on the implications for sanctioning and more on the role that international agreements or authoritative interpretations of obligations themselves play in creating constraints that resonate in domestic politics. This approach begins with the observation that domestic institutions can at time be a barrier to the realization of benefits for society as a whole: preference outliers can capture domestic institutions and thus hold policies hostage to their demands; well-organized interests can exert particularistic influences on policy, decreasing overall welfare; decisionmakers

10
can have time-inconsistent preferences that cause them to pursue short term interests at the expense of longer term gains; political polarization can lead to suboptimal outcomes or decisional paralysis at the domestic level. Under these circumstances, actors may not only have incentives to make international agreements (to freer trade, fixed exchange rates, convergent macroeconomic policies), but also to comply with them as a way to solve an intractable domestic problem. In this view, international agreements place a desired constraint on policy, where domestic politics alone have proved socially suboptimal. Furthermore, authoritative external decisions may reduce the domestic political costs of particular courses of action - an argument regularly invoked to explain compliance with the EMU and the IMF, for example. Those who have examined bargaining in the context of legal dispute settlement have argued that concessions tend to be easier to make, from a domestic political point of view, when legally required by an authoritative third party (Fischer, 1982; Merrills, 1969). The value of compliance, in this view, flows as much or more from the domestic political benefits as from the value of securing changes in behavior on the part of other states in the system. Finally, a large and growing literature has focused on one of the most tangible sources of suboptimal social behavior: domestic administrative or technical incapacities. A host of studies, many of them relating to compliance with environmental accords, point to the inability, as distinct from the willingness, of governments to comply with their international obligations. An ongoing study by a consortium of scholars headed by Harold K. Jacobson and Edith Brown Weiss (1995, 1997) comes to the conclusion that a crucial factor contributing to variations among the compliance performance of nine countries across five environmental accords over the past ten years has been their degree of administrative capacity, including knowledge and training of personnel responsible for environmental policy, adequate financial resources, the appropriate domestic legal mandate/authority to accomplish program implementation, and access to relevant information. Lacking such administrative or technical capacities, rule-consistent behavior may simply not be within a signatories' choice set. Outside agencies can help countries develop such capacities; the function of international agreements, in this case, is to not only specify obligations,

11
but to facilitate their attainment for certain classes of signatories deemed unable, without external resources, to meet particular standards of behavior (Haas et al 1993).

C. Compliance and the Nature of the Domestic Regime Another approach that has recently received some attention in the study of interstate disputes and more recently in legal circles may be termed "democratic legalism."5 In this formulation, regime type is crucial to understanding the role of law in interstate relations (Slaughter, 1995). While the domestically-formulated functionalist literature focuses on a range of domestic conditions that can contribute to or detract from compliance, this line of research looks at the distinctive features of democratic regimes that tend to bind them into a "zone of law" in the conduct of their foreign relations. The thrust of much of this literature is that democracies are more likely to comply with international legal obligations, and two kinds of reasoning are advanced to sustain this argument. One line of reasoning suggests that because liberal democratic regimes share an affinity with prevalent international legal processes and institutions, they tend to be more willing to depend on the rule of law for their external affairs as well. The argument depends on the notion that norms regarding limited government, respect for judicial processes, and regard for constitutional constraints carry over into the realm of international politics (Dixon 1993). Thus, countries with independent judiciaries are more likely to trust and respect international judicial processes than those which have no domestic experience with such institutions. Political leaders accustomed to respect constitutional constraints on their power in a domestic context are more likely to accept principled legal limits on their international behavior. This should lead us to expect that governments with strong constitutional traditions, particularly those in which intra-governmental relations are rule governed, are more likely to accept rules-based constraints on their international behavior. These arguments dovetail nicely with a growing agenda in political science that argues

This is not a term used by proponents of this approach, but it is a convenient appellation for purposes of this article.

12
that liberal democracies are more likely than are other regime types to revere law, promote compromise, and respect processes of adjudication (Doyle 1986, Dixon 1993, Raymond 1994). Specific mechanisms that might tighten the link between the domestic rule of law and international behavior include the absorption of the latter into the corpus of domestic regulation itself. Assuming a close correlation between regime type and meaningful domestic legal restraints on the public exercise of authority, "...[O]ne of the best ways of causing respect for international law is to make it indistinguishable from domestic law" (Fisher 1981).6 One

example of such absorption can be found in military manuals with respect to the laws of war: both Britain and the US have imported international law concerning warfighting into their military handbooks. The idea is to make the two sets of law incentive compatible, such that "[p]atriotism and national loyalty will be aligned on the side of compliance" (Fisher, 1981, 147). One can easily imagine the limits to such a strategy, however. Replication of international rules at the domestic level hardly guarantees their potency. Where international law is easily absorbed into the domestic system of rules one can wonder if behavior would have been much different in its absence; where international rules do not comport well with indigenous legal culture, expectations for compliance should not be be high. A second, distinct mechanism has also been used to back expectations of the importance of democracy for law compliance. This rests on the observation that leaders in liberal democracies may be constrained by the influence that international legal obligations have on domestic groups, who are likely to cite such rules or rulings to influence their own government's policy. In one version of this argument, the mechanism through which the compliance pull operates is domestic interest groups, who may have an interest in or preference for compliant behavior (Young 1979, Schachter 1991). In studies of compliance with environmental accords, for example, democracies were found to provide much more freedom for non-governmental organizations to operate, supplying the opportunity for the formation and strengthening of

This parallels Robert Keohane's notion of the "enmeshment" of international comitments into domestic politics and political institutions (Keohane 1992).

13
transnational coalitions to influence government compliance efforts (Jacobson & Brown Weiss 1997). NGOs have similarly been crucial in the human rights area, and countries that have embarked at least upon a transition to democracy have been influenced by their presence (Sikkink, 1993). The weight of an international obligation or authoritative legal forum may be crucial in convincing various domestic audiences actively to oppose a government's policy, raising the political costs of non-compliance. As Fisher (1981) has argued that these costs are likely to be especially heavy in the case of second order compliance involving violation of a specific decision against a government by an international authority, rather than a standing rule about which the government is likely to argue. There is an affinity with this line of argument and various strands of functional reasoning that view international institutions as crucial in influencing the domestic political debate surrounding a controversial foreign policy choice. The distinctive contribution of democratic legalism is its expectation of systematic differences between liberal democracies and nondemocracies in this regard: domestic political constraints encouraging law-abiding behavior are assumed to be much stronger in the former than the latter. For these reasons, more democratic countries are expected to be more willing to use legal institutions to regulate international behavior and to settle disputes, and comply more readily with these agreements once they are made.

Normative approaches to compliance Normative considerations are present to some degree in the observation that democratic norms relating to the rule of law may be operative in governments' attitudes toward international law compliance. But there is a growing school of research that places normative considerations at the center of their analysis of state behavior. This approach has been more willing to accept that normative concerns are capable of driving perceptions of interest, and that the most appropriate way to understand normative influences is through a subjective - rather than an analytically

14
imposed - framework of meaning. In this view, normative standards of appropriate conduct are socially constructed references points against which state behavior can be gauged. Normative influences have a long tradition in the study of international law compliance. At the turn of the century, Eliahu Root cited "moral force" as a reason for compliance with the decisions of arbitration panels, for instance (Root 1908). The predominance of realist theory in the study of international relations after World War II largely edged aside such arguments as naive, until a more subtle understanding of the relationship between international power and international society was articulated. Hedley Bull (1977) provided an early antecedent to what loosely can be refered to a somewhat more "constructivist" approach to the problem of international law compliance. While a firm believer in the ultimate importance of the balance of power in international politics, his work emphasized the importance of international society (shared norms and beliefs) as crucial to the effective functioning of international law. Parting with realist skeptics such as Aron, Bull thought international law existed by virtue of the fact that activities of actors in international relations are carried out under the assumption that the rules they are dealing with are legal rules. This did not however, justify "our treating them as a substantial factor at work in international politics..." Bull cited notbehavioral evidence, but drew on the discourse that diplomats have used to describe, justify, and excuse their actions: "What is a clearer sign of the inefficacy of a set of rules is the case where there is not merely a lack of conformity as between actual and prescribed behavior, but a failure to accept the validity or binding quality of the obligations themselves - as indicated by a reasoned appeal to different and conflicting principles, or by an unreasoned disregard of the rules." Bull's work opened the way for a more interpretive, contextual approach to the understanding of compliance than had his realist predecessors, with whom he shared a healthy respect for the balance of power. He viewed the primary function of international law as helping to mobilize compliance with the rules of what he termed international society, but thought it erroneous to view the principal contribution of international law as lying in its imposition of restraints on international behavior. Law could perform such a function only under a narrow set

15
of conditions: "International law can mobilize the factors making for compliance with rules and agreements in international society only if these factors [a social system marked by shared norms and beliefs] are present" (Bull 1977). Bull's central insight was explicated by international regimes7 theorists working in the constructivist mode over the course of the 1980s. That insight, that actor behavior alone failed adequately to convey inter-subjective meaning, and hence was an insufficient indicator of the role of rules, norms and agreements in international politics, was explicated most brilliantly by Friedrich Kratochwil and John Ruggie (1986). Rather than look to the violation of norms as the entire rule compliance story, these scholars argued that analysts needed to understand state behavior as interpreted by other states and intended by the actors themselves. Agreeing with Bull, they noted it was important to research how states justified their actions, and whether the international community responded to proffered rationales. "Indeed," they wrote, "such communicative dynamics may tell us far more about how robust a regime is than overt behavior alone." In this view, even divergent practices of actors could express principled reasoning and shared understanding. What constitutes a breach of obligation is therefore not simply an "objective description of a fact" but an inter-subjective appraisal. These observations comport well with the distinguishing features of social constructivism: its concern with the nature, origins, and functioning of social facts, the understanding of which is viewed as a major limiting feature of more utilitarian approaches. The implications for theory and research were profound. For one, these scholars suggested that the emphasis on "rational institutional design" - a prime interest of the more functionalist approaches discussed above - was fundamentally misplaced. Whereas scholars such as Downs and Rocke argued that relative incentives, rather than concepts such as "fairness" drove the compliance decision, the more subjective, normative approach suggested that "rational institutions" can be undermined if their legitimacy is in question (Ruggie & Kratochwil 1986).
7

"International regimes" were conceived as not isomorphic with international law, but the overlap is significant. International regimes were defined by Stephen Krasner (1983) as "principles, norms, rules, and decisionmaking procedures around which actor expectations converge in a given issue area."

16
This assertion dove-tailed with a large and growing legal and sociological literature that attempted to understand voluntary law compliance for individuals as well as groups and organizations as a function of perceived legitimacy of the law itself and legal processes themselves.8 The thrust of that literature is that perceived legitimacy of a legal rule or authority heightens the sense of obligation to bring behavior into compliance with the rule. But if legitimacy was central to voluntary compliance, how could one explore this relationship in a non-tautological way? One approach has been to locate the compliance pull of international norms in the nature of the norm itself. Thomas Franck has argued that the legitimacy of a rule - its ability to exert a "pull" toward voluntary compliance - should be examined in light of its ability to communicate, and this in turn is influenced by the rule's determinacy and its degree of coherence (Franck 1990). Similarly, Jeffrey Legrow has proposed rule attributes such as specificity, durability, and concordance as one way to think about the effect of norms on outcomes. In his view, the clearer, more durable, and more widely endorsed a prescription, the greater will be its impact on compliance behavior (Legrow, 1997). Others have focused on the substance of the rules as underpinning its moral force and contributing to legitimacy. For example, Fisher (1981) asserts that rules will be better complied with when they follow commonly held notions of fairness and morality; for example, against killing rather than informing on friends, or prescribing reciprocal rather than uni-obligational behavior. He holds that the more "elemental" the rule - the more it reflects malum in se rather than malum in prohibitum - the more first order compliance should be expected. Arguments with a similar flavored have been advanced by political scientists working in the area of human rights compliance. Margaret Keck and Kathryn Sikkink (1997) have argued that among the wide array of human rights standards embodied in international agreements, two kinds of prohibitions are most likely to be accepted as legitimate transnationally and cross-culturally: norms involving bodily integrity and the prevention of bodily harm for vulnerable or "innocent" groups, and norms

At the individual level see Tyler (1990). With reference to collective organizations, see Knoke and Wood (1982).

17
involving the legal equality of opportunity. These norms, they argue, transcend culturally specific belief systems and resonate with basic ideas of human dignity common to most cultures, enhancing their legitimacy as behavioral prescriptions. International, transnational, and nongovernmental organizations play a significant role in the normative processes cited in much of this literature. While the rational functionalist literature acknowledges that these kinds of entities can provide "focal points" that function to narrow the range of equilibrium outcomes, scholars working in the normative vein would go much further: international institutions and organizations provide legitimacy to particular rules, thus enhancing their effectiveness via a heightened sense of obligation rather than through their mere instrumental value as a convenient point of convergence (Claude 1966, Peck 1996). Joaquin

Tacsan, for example, writes of the International Court of Justice as a producer and disseminator of new and consensual knowledge, and therefore able establish a point where normative expectations can converge through an interactive bargaining process. He argues tha it was the ICJ's redefinition of the notions of self-determination, non-intervention, collective self-defense, and use of force in a regional context which determined the normative expectations which ultimately prevailed in Central America's peace settlement (Tacsan 1992). Authoritative institutions that review states' policies for consistency with their international agreements are valued in this view not because they have the formal power of sanction, but because such procedures provide legitimacy to the attempt to find a gap between governments commitments (formal stances) and their actions. It is for this reasons, some have argued, that NGOs have been very much in favor of creating a Sustainable Development Commission at the 1992 UNCED in Rio: not because it would have the formal power of sanction, which realist and some functionalist perspectives would point out as a major weakness, but because it is a forum which legitimates the very process of holding governments accountable for their behavior and its relationship to their stated positions. In short, normative approaches to the problem of compliance have focused on the force of ideas, beliefs, and standards of appropriate behavior as major influences on governments'

18
willingness to comply with international agreements. The hall mark of much of this research is its departure from the radically individualized methodology of some variants of realist and functionalist approaches, and its embrace of an understanding of international obligations as social constructs that must be understood and analyzed in a highly inter-subjective framework of meaning. Despite some attempts to marshal positive methodologies in the study of normative influences (Kacowicz, 1994; Kegley and Raymond, 1981)9 the inter-subjective formulation of the problem has largely resisted such approaches. When pressed, scholars who have devoted attention to the social ideational influences on international politics submit that factors such as aspirations, legitimacy, and the notion of rights, fall into the category of reasons for actions, which are not the same as causes of actions (Ruggie, unpublished observation). Moreover, its explicitly inductive and high contextual methodology distinguishes this approach from much of the (neo)realist and functionalist literature.

CONCLUSIONS Studies of compliance with international agreements encompass a wide range of approaches to the study of international relations and cross over into the disciplines of law and sociology. Yet despite what may appear to be a sprawling literature, the empirical work that might link theory with observed behavior (or subjective understandings of such behavior, as constructivists would have it) has only begun to accumulate in the past few years. Part of the difficulty in the empirical study of compliance has been methodological. If the central analytical issue is to understand the conditions under which states behave in accordance with rules to which they have committed themselves or, more broadly, to prevailing norms of international behavior, then it is important to isolate the impact of those rules and norms. Several studies have tried to demonstrate a correlation between legal standards and state behavior, and have sometimes employed large databases and statistical techniques to do so, but
9

Kacowicz (1984) has looked at normative convergence as reflected in substantive treaty provisions and its effect on the propsects for peaceful territorial change. Kegley and Raymond (1981) have used "quasi-authoritative treatises" drawn correlations between legal norms prevailing at the time and states' use of force.

19
these have often been unconvincing in demonstrating causation, or even in providing a explanatory link between actions taken and the existence of agreements or normative considerations. Thus it has been possible to show that there has been much international behavior that is consistent with international law, even in the conduct of hostilities between states (see for example Kegley & Raymond 1981; Tillema & Van Wingen 1982); it has been far more difficult, however, to show and causal link between between the two . Establishing causation has been complicated by problems of selection bias and endogeneity that exists on many levels. As realists have noted, selection bias operates to proliferate rules in issue areas in which problems of strategic cooperation are in many cases minimal. From Aron's complaint that it is no test of international law merely to look at "secondary" problems, to Downs and Rocke's observation that compliance is not very interesting if the body of international agreements really just look like "flight control" agreements from which no one has much interest in defecting, there are good reasons to expect that the problem of compliance with treaties will be easier than the problem of international cooperation in general (Downs et al 1996). This is essentially due to the fact that treaty negotiation is endogenous: governments are more prone to make agreements that comport with the kinds of activities they were willing to engage in anyway, and from which they could foresee little incentive to defect. The endogeneity of treaty agreements also serves to weaken the standards of behavior within agreements, which might be expected to improve compliance, but only because the rules are quite lax. As a result, it has been very difficult to show that a rule, commitment, or norm per se influenced governments to take particular positions that represent compliance. For example, in Oran Young's (1979) account of compliance with the partial test ban treaty, he gives away the reasons for compliance in the absence of enforcement mechanisms by allowing that the states involved really had extremely weak incentives to go against the agreement anyway. It has been far more difficult to construct research around crucial case studies in which agreements can be

20
shown to have created new and powerful incentives, or to have altered actors' perceptions of their interests. The problem of selection bias creates inferential difficulties for the study of second order compliance as well. Where significant national interests are involved, governments are arguably likely to resist or ignore international jurisdiction (Fischer 1982). As a result, cases in which third parties render authoritative rulings with which the parties are legally bound to comply are likely to be "easy" cases involving parties eager to settle their dispute and therefore willing to make concessions anyway (Coplin 1968, Schwartzenberger 1945), those countries that tend in any case to be law-abiding, or perhaps small countries whose weak negotiating position gives them little to lose by going through legal processes. However, the nature of the selection bias may differ across issue areas depending on the nature of the dispute settlement mechanism in place. In the human rights areas, the requirement of the "exhaustion of domestic remedies" might have the opposite effect on the pool of litigated cases: as a result, states with good domestic rules are not likely to produce a lot of internationally reviewed cases. With respect to trade disputes under the GATT and now the WTO, incentives exist to sue the largest traders, as the payoffs if successful are so much greater than going after small markets. An in some issue areas, third parties themselves may have the potential for acting strategically, altering the nature of cases brought before them; for example, finding ways to avoid taking jurisdiction for cases with which they anticipate a lack of compliance. The point is, for any given issue area, the "litigation pool" is hardly likely to be typical of the international community of states. The impact of these biases will influence our ability to draw inferences regarding compliance with authoritative third party decisions. In some cases, adjudication or arbitration may be likely to complement what states would have done anyway; in other issue areas, its impact may be expected to vary widely. Not much work has been done to determine whether assertions of selection bias are truly pervasive, or are rather more limited than some have implied. For example, it should not be too difficult to produce the evidence that might support a claim that countries or country pairs that agree voluntarily to third party arbitration or other judicial processes for settling their disputes are

21
systematically different from a random sample of countries. Endogeneity of substantive agreements poses more difficulty and may require another strategy. One is to select cases in which there is strong evidence of a shift in state interests over time, by which a previous commitment becomes inconvenient or even costly to maintain, at least in the short run, under some later circumstance. The endogeneity problem would be minimized to the extent that compliance with agreements made at an earlier point in time are delinked from obvious explanations of strong, narrow and immediate "interest-based" reasons to comply. There are other ways to address the endogeneity of treaties themselves. One is to view the negotiation phase as an integral aspect of the compliance process. Scholars who have emphasized the persuasive functions of legal agreements and discourse have also been interested in the contribution that participatory negotiation makes to eventual compliance. Highly compatible with a more constructivist approach to compliance, rather than worry about "controlling" for endogenous effects of treaty negotiation, one might incorporate its discursive elements into a fuller story of the compliance process. Governments persuade and become convinced of the value or appropriateness of particular standards of behavior over the months, years, and even decades they spend in their formulation. This research agenda might even call for an examination of the discourse used by participates as such negotiations unfold; attitudes toward compliance are necessarily shaped and reflected in this discourse. This strategy uses the negotiation process as data to reveal more regarding attitudes toward compliance, rather than viewing it as a source of "bias" in making inferences. Despite the conceptual and methodological difficulties, research into the question of compliance with international agreements represents a substantial advance in the study of international law and institutions as important influences on international politics. Realism's assumption that such effects were marginal discouraged empirical investigation and broader theoretical innovations that might address variations that patently did exist ascross issue areas, among nations, and over time. The current emphasis on the effects of rules on behavior has gone well beyond the functionalistist studies of the 1980s which was primarily concerned with the

22
phenonenon of rule creation and evolution rather than on their behavioral effects. Nonetheless, we are a long way theoretically and empirically from an understanding of the conditions under which governments comply with international agreements. Pockets of progress mark particular issue areas, such as the environment and human rights, but more effort is needed to subject these findings to the broader concerns of international politics.

Literature Cited:

Aron R. 1981. Peace and War: A Theory of International Relations. Malabar FL: Robert E. Krieger Publishing

Bilder RB. 1989. International third party dispute settlement. Denver Journal of International Law and Policy. 17(3): 471-503

Bork RH. 1989/90. The limits of 'international law'. The National Interest 18: 3-10 Boyle FA. 1980. The irrelevance of international law. California Western International Law Journal. 10

Brierly JL. 1963. The Law of Nations. London: Sir Humphrey Waldock. 6th ed

Bull H. 1977. The Anarchical Society: A Study of Order in World Politics. New York: Columbia University Press

Bulterman MK, Kuijer M. 1996. Compliance with Judgments of International Courts. The Hague: Martinus Nijhoff

Chayes A, Chayes AH. 1995. The New Sovereignty: Compliance with International Regulatory Agreements. Cambridge MA: Harvard University Press

Claude IL. 1966. Collective legitimization as a political function of the United Nations. International Organization, 20(3): 367-79

Coplin WD. 1968. The World Court in the international bargaining process. inThe United Nations and its Functions, ed. RW Gregg and M Barkin, 313-31. Princeton NJ: Princeton University Press

Diehl PF. 1996. The United Nations and peacekeeping. In Coping with Conflict After the Cold War., ed. E Kolodziej, R Kanet, 147-65. Baltimore: Johns Hopkins University Press

Dixon WJ. 1993. Democracy and the management of international conflict. Journal of Conflict Resolution. 37(1): 42-68

Downs GW, Rocke DM. 1995. Optimal Imperfection? Domestic Uncertainty and Institutions in International Relations. Princeton NJ: Princeton University Press

Downs GW, Rocke DM, Barsoom PN. 1996. Is the good news about compliance good news about cooperation? International Organization. 50(3): 379-406

Doyle MW. 1986. Liberalism and world politics. American Political Science Review 80: 115169

Fischer DD. 1982. Decisions to use the international court of justice: four recent cases. International Studies Quarterly. 26(2): 251-77

Fisher R. 1981. Improving Compliance with International Law. Charlottesville: University of Virginia Press

25
Franck TM. 1990. The Power of Legitimacy Among Nations. New York: Oxford University Press

Friedmann W. 1964. The Changing Structure of International Law. London: Stevens and Sons

Garrett G, Weingast BR. 1993. Ideas, interests, and institutions: constructing the EC's internal market. In Ideas and Foreign Policy, ed. J Goldstein, RO Keohane. Ithaca NY: Cornell University Press

Haas PM, Keohane RO, Levy, MA. 1993. Institutions for the Earth: Sources of International Environmental Protection. Cambridge MA: MIT Press

Hawkins K, Thomas JM. 1984. Enforcing Regulation. The Hague: Kluwer Nijhoff

Henkin L. 1979. How Nations Behave: Law and Foreign Policy. New York: Columbia University Press and Council on Foreign Relations. 2d ed.

Hoffmann S. 1956. The role of international organization: limits and possibilities. International Organization, 10(3): 357-72

Jacobson HK, Weiss EB. 1995. Compliance with international environmental accords. Global Governance 1

_____. 1997. Compliance with international environmental Aaccords: achievements and strategies. Harvard University Seminar in Law and International Relations. 5 February

26
Kacowicz AM. 1994. The problem of peaceful territorial change. International Studies Quarterly. 38: 219-54

Keck, Margaret and Kathryn Sikkink. 1997

Kegley CW Jr, Raymond GA. 1981. International legal norms and the preservation of peace, 1820-1964: some evidence and bivariate relationships. International Interactions. 8(3): 171-87 Keohane RO. 1984. After Hegemony: Cooperation and Discord in the World Political Economy Princeton NJ: Princeton University Press

______. 1992. Compliance with international commitments: politics within a framework of law. American Society of International Law Proceedings. 86: 176

______. 1993. Sovereignty, interdependence, and international institutions. In Ideas and Ideals: Essays on Politics in Honor of Stanley Hoffmann ed. LB Miller, MJ Smith. 91-107. Boulder Colo.: Westview Press

Krasner SD. 1983. Structural causes and regime consequences: regimes as intervening variables. In International Regimes. ed. SD Krasner. 1-21. Ithaca: Cornell University Press.

Kratochwil FV, Ruggie JG. 1986. international organization: a state of the art on an art of the state. International Organization. 40(4): 753-75

Legrow JW. 1997. Which norms matter? revisiting the "failure" of internationalism. International Organization. 51(1): 31-63

Merrills JG. 1969. The justiciability of international disputes. Canadian Bar Review. 47: 24169

Milgrom PR, North DC, Weingast BR. 1990. The role of institutions in the revival of trade: the law merchant, private judges, and the champagne fairs. Economics and Politics 2(1): 123

Mitchell RB. 1994. Regime design matters: intentional oil pollution and treaty compliance. International Organization 48(3): 425-58

Morgenthau HJ. 1985. Politics Among Nations: The Struggle for Power and Peace. New York: Knopf. 6th ed.

Oppenheim, L. 1912. International Law. London: Longmans, Green. 2d ed.

Peck C. 1996. The United Nations as a Dispute Settlement System. The Hague: Kluwer Law International

Raymond GA. 1994. Democracies, disputes, and third-party intermediaries Journal of Conflict Resolution, 38(1): 24-42

Röling BVA. 1960. International Law in an Expanded World. Amsterdam: Djambatan

Root E. 1908. The sanction of international law. Proceedings of the American Society for Itnernational Law. 2: 14-17

28
Ruggie JG. forthcoming. Constructing the World Polity: Essays in International Institutionalization

Shihata IFI. 1965. The attitude of new states toward the International Court of Justice. International Organization. 19(2): 203-22

Schachter O. 1991. International Law in Theory and Practice. Dordrecht: Martinus Nijhoff Publishers

Schwarzenberger G. 1945. International Law. London: Stevens & Sons Ltd.

Slaughter AM. 1995. International law in a world of liberal states. European Journal of International Law. 6: 503-38

Snavely K. 1990. Governmental policies to reduce tax evasion: coerced behavior versus services and values development. Policy Sciences. 23: 57-72

Tacsan J. 1992. The Dynamics of International Law in Conflict Resolution. Dordrecht, Boston, London: Martinus Nijhoff

Tillema HK, Van Wingen JR. 1982. Law and power in military intervention. International Studies Quarterly. 26(2): 220-250 Tyler T. 1990. Why People Obey the Law. New Haven: Yale University Press.

Young OR. 1979. Compliance and Public Authority. Baltimore: Johns Hopkins University Press