email icon Email this citation

Adversarial and Facilitative Approaches to On Site Inspection in Arms Control and Environmental Regimes

Ronald B. Mitchell

International Studies Association

March 18-21, 1998

Abstract

Do states account for likely sources of noncompliance in designing the compliance mechanisms of international institutions? What are the likely sources of noncompliance with a regime? States may violate international agreements because of intentionality, incapacity, inadvertence, or ignorance. The mix of these causes of violation varies across different issue areas: violations of arms control agreements are likely to be dominated by intentional violations while violations of environmental agreements are likely to be dominated by the other sources of noncompliance. What are the potential responses to noncompliance? Regimes can either adopt an "adversarial" approach of attempting to deter noncompliance or a "facilitative" approach of attempting to help noncompliers come into compliance. To the extent that the success of these different approaches depends on the types of noncompliance at play, do states choose different responses to respond to different noncompliance settings? Evidence from onsite inspection procedures in regimes regulating chemical weapons and wetlands demonstrates that the two regimes exhibit quite different causes of noncompliance and corresponding differences in the responses by the regime to noncompliance.

Introduction

Do international institutions make "rational" choices about institutional design? That is, do they adopt policies to accomplish desired goals based on evaluation of both the strategic context they face and the alternative policies they have available? For example, many international institutions seek to induce member states to comply with the proscriptions and prescriptions of a treaty. If states always violate regime rules for essentially similar reasons, we might expect little variation in the mechanisms international institutions use to induce compliance. Even if the causes of noncompliance do vary, we might expect little institutional variation if, of the alternative compliance mechanisms available, one is always more effective than the others. However, if noncompliance can occur for a variety of reasons, if the effectiveness of a compliance mechanism depends upon the causes of noncompliance, and if international institutions are responsive to these facts, then we might expect compliance mechanisms to vary across regimes in explicable ways. In short, are international institutions rational, and does their form follow their function?

Many scholars, extrapolating from realist and neorealist theory as well as from empirical experience in the areas of arms control and trade, have concluded that states violate international law intentionally whenever it suits their short term interests and that other states can induce compliance by credibly threatening potent sanctions (Downs, Rocke, and Barsoom, 1996). Even those scholars willing to accept recent neoliberal claims that "regimes can influence behavior" have argued "yes, but they do so only rarely and only through the use of enforcement" (Downs, Rocke and Barsoom, 1996). Sanctions are the most appropriate and effective response that international regimes can, do, and should adopt to enhance behavioral conformance with international agreements. Other scholars, extrapolating from neoliberal theory and, particularly, experience in international environmental affairs, contend that noncompliance can arise from incapacity, inadvertence, and ignorance as well as from intention. Therefore, sanctions are inappropriate, probably ineffective, and, in any event, unlikely responses to noncompliance. Regimes reduce noncompliance by enhancing states' technical or financial capacities, providing states with information, and shaping norms and ideas (Finnemore, 1996; Goldstein and Keohane, 1993; Haas, Keohane, and Levy, 1993; Jacobson and Brown Weiss, 1995; Klotz, 1995). Both sides in this debate frequently have argued the questions raised here in noncontingent terms; i.e., have argued for a single or "major" reason why states violate international agreements and a single or "best" mechanisms for getting states to comply.

This article argues that improving our understanding of the behavior of states, the structure and behavior of international regimes, and the interaction between the two, requires assessing whether both sides in this debate might be correct. Regarding the causes of noncompliance, we can better ask, not "are all treaty violations intentional" but, "under what conditions do treaty violations reflect intentional decisions and under what conditions do treaty violations reflect factors outside the state's control?" Regarding mechanisms to induce compliance, we can better ask, not "is compliance always induced by sanctions or their threat" but, "under what conditions do threats induce compliance and under what conditions do more facilitative approaches induce compliance?" These questions, framed in this way, raise the third question of international institutional design. If states recognize that the causes of noncompliance in one realm differ from those in another realm, and recognize that what constitutes the most effective response to a case of noncompliance depends on what the cause of that noncompliance is, then we can ask whether states act "rationally" and construct international regimes and institutions with compliance mechanisms attuned to the strategic context and to the mix of noncompliance sources they face. 1

To develop this argument, I first delineate the reasons why states violate agreements and how the causes of noncompliance are likely to vary depending on the nature of the strategic setting they face. Specifically, systematic differences in the strategic setting of arms control and environmental affairs is likely to produce different causes of noncompliance and corresponding variation in the "noncompliance demographic" that the international institution faces. Assumptions regarding the causes of noncompliance in these settings can be conceptualized as involving either an "intentional violator" assumption or a "good faith nonconformer" assumption (Mitchell, 1998, 117). I go on to identify two general approaches for responding to nonconformance, namely, an "adversarial" approach that seeks to deter intentional violators through sanctions, and a "facilitative" approach that seeks to encourage compliance through rewards, education, and capacity enhancement. Given variation in both the causes of noncompliance and the potential responses to it, I argue that we can expect states rationally to develop international institutions that are responsive to the most predictable and common sources of noncompliance, but that these institutions are likely to prove less capable of identifying and responding appropriately to the particular cause of noncompliance in an individual case. I then use experience with onsite inspection from the Chemical Weapons Convention and the Convention on Wetlands of International Importance to provide empirical support that even nominally similar institutional procedures are carried out in quite different manners due to the different noncompliance settings involved.

Sources Of, And Responses To, Noncompliance

Downs, Rocke, and Barsoom distinguish between two schools of thought on compliance, the "managerial" school and the "enforcement" or political economy school (Downs, Rocke and Barsoom, 1996). Although they distinguish and label these schools based on their responses to noncompliance, they also contend that these schools differ in their assumptions about the causes of noncompliance. Those in the enforcement school, according to Downs, Rocke, and Barsoom, consider noncompliance as almost always reflecting intentional efforts to deceive and defect, and therefore see it as empirically true and prescriptively appropriate that regimes, to be effective, must rely on sanctions. Managerialists, by contrast, view noncompliance as more often unintentional and see it as empirically true and prescriptively appropriate that regimes rarely induce compliance through sanctions, but instead rely on more positive responses including rewards and dialogue.

Downs, Rocke, and Barsoom's specific claims regarding the enforcement school rely on three distinct but related arguments. First, compliance problems derive from "the calculation of states bent on exploiting other states," rather than "the ambiguity of treaties, the capacity limitations of states, and uncontrollable social and economic changes" (Downs, Rocke and Barsoom, 1996, 392). Second, in a mixed-motive or PD type setting, compliance with "deeply cooperative" agreements can only be achieved through enforcement strategies, which involve credible and potent punishment and sanctions (Downs, Rocke and Barsoom, 1996, 387). Third, "critical regulatory arenas" usually involve situations properly characterized as mixed-motive games and prisoners' dilemmas (Downs, Rocke and Barsoom, 1996, 387). Thus, they make a nonconditional argument about the causes of noncompliance, a conditional argument about effective responses to noncompliance, and an empirical argument about the frequency of the conditions in the second argument.

As a whole, Downs, Rocke, and Barsoom argue that enforcement is necessary to compliance, at least in any cases "that matter." But the conditional clause can easily be used to convert the argument into a tautology (with the absence of enforcement indicating that the case does not "matter") and averts our gaze from much that may be of interest in the world of international relations. Even if Downs, Rocke, and Barsoom prove correct in their contention that enforcement is necessary to induce "deep" compliance in PD type settings, this leaves unaddressed the question of how regimes induce "deep" compliance in non-PD settings, or how they induce "shallow" compliance in any setting. If understanding how states induce deep cooperation holds theoretical interest because of how difficult and rare it is, understanding how states induce shallow cooperation holds theoretical interest because of how common it is. To directly engage their argument would require analyzing cases within their specified domain, to wit, regimes that are both deeply cooperative and involve mixed motive games. The argument that follows explicitly does not attempt to do that. Rather, it suggests that even if Downs, Rocke, and Barsoom's argument holds true within their specified domain, considerable variation of interest in the sources of noncompliance and the approaches to compliance management exists across domains. The sections that follow develop these arguments more fully while at the same time arguing that theories and assumptions about the causes of noncompliance do not inevitably and necessarily require corresponding assumptions about the best responses to such noncompliance.

Theories of the causes of noncompliance

We can develop useful insights into the range of potential causes of noncompliance based on insights from the managerial and enforcement schools, even while rejecting their claims about the empirical dominance of one of those causes. Taken together, these schools suggest that those actors regulated by treaties can fail to comply for a variety of reasons. States or nonstate actors may violate an agreement intentionally because they simply do not see it as in their interest to comply. States may fail to comply because they lack adequate or appropriate capabilities to ensure their own behavior compliance or that of the individuals or entities directly responsible for compliance. In some cases, states and nonstate actors may make good faith efforts to fulfill their commitments but nonetheless fail to do so. Finally, states or nonstate actors may be unaware of certain treaty requirements or the available policy and technology options for implementing them cheaply, efficiently and quickly, or the requirements may be ambiguous and vague (Mitchell, 1993).

From this list of intentionality, incapacity, inadvertence, and ignorance, we can develop two conflicting assumptions about the causes of noncompliance. The enforcement school, as delineated by Downs, Rocke, and Barsoom, argues that an "intentional violator" assumption is the most empirically accurate, while the managerialists argue that a "good faith noncomplier" assumption of incapacity, inadvertence, and ignorance more accurately captures the bulk of treaty noncompliance. These two assumption sets are delineated below.

The "intentional violator" assumption

Theoretical and empirical bases exist for the contention that state violations of international accords often, if not always, reflect intentional attempts to benefit by defecting from the agreed-upon cooperative equilibrium. On the theoretical level, the long history of realist theory and the compelling evidence supporting it gives primacy to a view that states are motivated by survival, short-term self-interest, and desires to achieve gains relative to other states. These state motivations provide predictions that states will decide between compliance and noncompliance based on careful and rational assessments of the relative costs and benefits of all available alternatives, and will choose the latter whenever doing so furthers their self-interest. These assumptions are reinforced by the conviction - frequently, though not always, correct - that international treaties and regimes reflect attempts by states to overcome PD type dynamics. If treaties are, in fact, efforts to overcome PD dynamics, then states will have strong incentives to defect, absent countervailing pressures from the institution. Realists accept that states may establish regimes to achieve other cooperative goals in other settings, as in treaties dealing with ship and airline navigation, satellite slot allocation, or regulatory harmonization (Krasner, 1991). However, these coordination games are discounted as fundamentally uninteresting with respect to compliance, since states have few incentives to defect from the cooperative equilibrium once it is established (Stein, 1990). 2

These largely theoretical arguments find reinforcement in the nature of certain treaties and issue areas. In many cases, the very nature of the behavior being regulated and the rules established to regulate it make "unintentional" noncompliance impossible, or at least improbable. Often, the behavior defined by the treaty as noncompliance requires intentional action, and could not happen by inadvertence. Consider, for

Theoreticaexample, agreements banning construction of particular weapons systems, capping tariff levels, forbidding torture, or banning commercial whaling. These actions cannot occur by accident, require intentional action to occur, and, thus, make an assumption of intentionality seem appropriate.

Equally important, there are reasons for believing that incapacity, inadvertence, and ignorance should be relatively rare forms of noncompliance. If states are as concerned with relative gains and wary of being taken advantage of as realists suggest they are, then existing members of a regime should not invite or accept the membership of any state that is not clearly capable of fulfilling its commitments under an accord. States should be unlikely to allow a state to become a member of a regime unless that state clearly has sufficient resources and control over outcomes to "deliver" the agreed upon level of compliance. With respect to ignorance, this approach tends to interpret a state's ignorance of either the rules or the strategies for compliance as post-hoc excuses and rationales for their noncompliance, rather than as causes. States negotiate vague rules precisely so that they do not create unnecessary political costs of following one's interests. In short, states are assumed to have adequate capacity, control, and knowledge to comply with a regime's rules, and therefore noncompliance must arise strictly from the lack of incentives to do so.

The "good faith noncomplier" assumption

As these last comments make clear, the claim that states intentionally violate international agreements depends on acceptance of underlying assumptions regarding capacity, control, and knowledge. First, for violations to constitute a "deliberate choice" requires that the state had the capacity - and, hence, could have chosen - to comply. Yet, regimes often seek to increase membership without regard to whether prospective member states are capable of fulfilling the corresponding commitments. This, at a minimum, creates the possibility for noncompliance due to incapacity. Treaties also may be designed as hortatory efforts that establish aspirational goals openly acknowledged as currently out of the reach of all or most member states. The financial targets established for membership in the European monetary union provide such examples. Especially in environmental treaties, states assume responsibilities with which they, knowingly or unknowingly, may not be able to comply. Actors who perceive compliance as beneficial may lack the necessary financial, administrative or technological resources to comply (Haas, Keohane and Levy, 1993; Jacobson and Brown Weiss, 1995; Young, 1992, 183-185). Non-compliance by corporate and individual citizens can arise because the government lacks informational or regulatory infrastructures adequate to elicit compliance. This can be especially common in realms experiencing high levels of uncertainty, since the resources required to fulfill the requirements may be unclear at the time that a state undertakes the commitment.

Second, for a violation to reflect a "deliberate choice" requires a direct linkage between state choice and state practice. However, if one abandons a "unitary rational actor" view of the state and recognizes the corresponding potential for principle-agent dynamics, one can imagine considerable opportunities for contradiction of the "intentional violation" assumption. For example, violations of arms control and human rights agreements could reflect "unauthorized" actions by government bureaucrats or contractors. Indeed, sales of nuclear-related equipment to Iraq by European and American suppliers constitute just such "unauthorized" violations of nonproliferation agreements. This imperfect relationship between action and outcome is also evident in noncompliance due to inadvertence. States may take actions sincerely intended and expected to achieve compliance but nonetheless fail to meet treaty standards. Indeed, the increasingly frequent replacement of "command and control" regulatory approaches with incentive-based approaches may well increase the frequency of inadvertent noncompliance. Inherent uncertainty in the impacts of incentive-based policy strategies, like taxes or subsidies, mean that even developed states' efforts to alter their citizens' behaviors may fail to achieve intended results (Epstein and Gupta, 1990).

Third, intentional violation assumes that the state was aware of and properly understood what was required of it, and understood available opportunities for achieving it. The traditional vagueness of treaties opens up the possibility not only for intentional "advantage taking" but also for a sincere misunderstanding of the requirements. One does not need to contend that the capacity, control, and comprehension assumptions of an "intentional violations" approach are always inaccurate to contend that, at least in certain circumstances, these conditions are not met. It is precisely these three assumptions that have been increasingly brought into question, particularly in the realm of international environmental affairs.

The "noncompliance demographics" of a regime

The foregoing description of potential causes of noncompliance suggests the likelihood of different reasons for noncompliance with a particular treaty, but also suggests the possibility that different treaties exhibit different "noncompliance demographics" or contexts. Different issues are likely to exhibit different "mixes" of noncompliance types, with intentional violations being dominant in some while inadvertence or incapacity is dominant in others. Thus, the accuracy of these competing assumptions regarding the causes of noncompliance will vary by issue area. An "intentional violator" assumption may provide a relatively accurate characterization of many security treaties while a "good faith nonconformer" assumption may provide a more accurate characterization of environmental issue areas. That said, however, even standards. Indeed, the increasingly frequent replwithin one of these broad issue areas, we should expect variation across different regimes and we should not expect either set of assumptions to fit any particular regime perfectly. Within any particular regime, we are likely to be able to find examples of all four types of noncompliance. Accurately characterizing a regime requires getting a demographic picture of the relative empirical frequency of these different types of noncompliance.

Theories of responses to noncompliance

The mix of types of noncompliance poses member states and treaty managers with the challenge of developing effective responses to noncompliance. Downs, Rocke, and Barsoom may be correct that regimes that seek "deep cooperation" within a PD-type mixed motive setting, must rely on enforcement, rather than alternative "managerial," strategies to induce compliance (Downs, Rocke and Barsoom, 1996). They also contend that compliance with "shallow" agreements is attributable to their shallowness. This implies that empirical variation in the compliance strategies of regimes will vanish if one examines that subset of regimes that seek and achieve "real" cooperation. Without affirming their argument that "deep" cooperation requires enforcement, we can build on the implicit and undeveloped portion of their argument that both the level of cooperation sought by a regime and the strategic setting it faces determine the form of compliance strategy it adopts.

If we accept the increasingly-supported contention that regimes can influence state behavior, then the question arises does success at this task always require the same functional form. Downs, Rocke, and Barsoom claim that regimes can induce compliance by only one means, but their caveats leave the question unaddressed. Even if enforcement is shown to be necessary to inducing deeply cooperative compliance in PD settings, other strategies may work to induce compliance with shallow agreements or in other strategic settings. "Shallow cooperation" may be less useful than cases of "deep cooperation" in compellingly demonstrating that regimes matter, but can prove quite useful as indicators of the mechanisms different institutions use to induce compliance in different circumstances. Of course, we must demand that the regimes we examine involve an attempt to induce true cooperation, i.e., that the regime sought to induce states to engage in (or refrain from) behaviors that they would not have otherwise engaged in (or refrained from). However, that does not require exclusive focus on deep agreements. As Downs, Rocke, and Barsoom note, the distinction between deep and shallow agreements is defined by " the extent  to which it requires states to depart from what they would have done in its absence" rather than the requirement for such change. (Downs, Rocke and Barsoom, 1996, 383 emphasis added). Thus, it remains of interest how regimes can induce compliance even with "shallow" agreements, not merely because they are common but because even shallow agreements require states to do things that they would not otherwise do.

Just as assumptions regarding the causes of noncompliance can be categorized into two schools, so can assumptions regarding the possible responses to noncompliance. By examining the possible responses to noncompliance, we can begin examining variation in international institutional design. If both the enforcement and managerial schools of compliance have some claim to empirical accuracy (even if both are victims of their own overgeneralization), then we can ask why some international regulatory regimes "enforce" compliance while others "manage" it?

An "adversarial" response model

The core of an adversarial response model is the notion that violations can and should be deterred. It frames the question as "how can we deter actors from violating?" Given that violations are intentional, those who violate could have chosen to comply, and violators will conceal evidence of the violation, states and nonstate actors can be induced to comply by establishing credible and potent sanctions. If sanction processes can be made sufficiently likely and sufficiently costly, those with incentives to violate will be deterred from doing so. Even when sanctions fail at the goal of "specific" deterrence, i.e., the sanctions fail to induce the violating actor to re-establish compliant behavior, they may achieve the goal of "general" deterrence with "onlookers" becoming less likely to violate after being made aware by example of the costs that defection and noncompliance will entail.

Sanctions appeal to us as a response for several reasons. First, both the logic and considerable evidence suggest that sanctions can effectively deter violations in a considerable number of cases. 3 Second, at both the domestic and international levels, sanctions have political appeal as "appropriate" punishment even when they are not "effective" at encouraging higher levels of compliance. Sanctions can almost always be made to punish the violator, even when they are inadequate to alter the violator's behavior. Third, sanctions have the advantage of only needing to be engaged in after violators have revealed themselves. By only needing to respond in a limited number of cases, the costs of the response system can be kept low.

Recent evidence has suggested that sanctions are both difficult to implement at the international level and are frequently ineffective even when they are implemented. Sanctions face traditional collective action problems - each state that contributes to sanctions incurs costs in doing so, while the benefits of doing so accrue to all relevant members of the international community. Thus, sanctions themselves are plagued by PD-type dynamics that make cooperation problematic (Axelrod and Keohane, 1986; Martin, 1992). Nor do they appear to be as effective as one might hope (Hufbauer, Schott, and Elliott, 1990). 4

Even when sanctions can be implemented and will alter the behavior of violators, this does not mean that they are the only effective means of achieving it. Although violations of the nuclear nonproliferation treaty are almost the quintessential evidence of the "intentional violation" assumptions of noncompliance, the nonproliferation regime appears to have spent as much effort developing technology denial procedures as ensuring that sanctions are credible or potent. Given the huge incentives some states feel to acquire nuclear weapons, it seems unlikely that they could be deterred and that denial rather than sanctions will be more effective.

Indeed, sanctions appear to have some perverse effects. First, they may make intentional violators even more obstinate. Domestic sociology literature as well as international empirical evidence suggests that states exhibit a "reactive resistance" to sanctions and a "rally round the flag" effect in which publics increase their support for a government's choice of violation instead of pressuring the government to alter that policy (Ayres and Braithwaite, 1992; Barber, 1979). At the domestic level, tax amnesty and gun buy-back programs demonstrate an increasing willingness, given these dynamics, to abjure traditional sanction-based responses in favor of alternative, more effective approaches, even if that means unambiguously intentional violations go unpunished. 5 Sanction-based approaches also reinforce incentives for those violating the regime rules to do so clandestinely, making it increasingly difficult to gather accurate information about the trends in the underlying problem the regime is seeking to resolve.

A "facilitative" response model

In contrast to an adversarial response model, the core of a facilitative response model is the notion that compliance can be encouraged or "facilitated." In this view, actors are assumed to generally be "good faith noncompliers" who, for various reasons, have less than perfect control over whether their behavior is in compliance or not. Actors are assumed to have values that are consistent with, even when not strongly supportive of, compliance. They are not driven by relative gains concerns, devious desires to garner the "temptation" payoff in a PD game, or general malicious intent, but rather are driven by absolute gains, a belief that cooperation will yield benefits, and a general (if not strong) respect for the rules of international law. Noncompliance occurs because states lack the capacity, control, or information necessary to bring their own behavior or that of substate actors within their state into compliance.

The general model of a facilitative response takes the form of "how can we help you comply?" A regime adopting such an approach tries to identify the obstacles to compliance, assuming that they are not motivational in nature, and tries to help or persuade the noncomplier to become a complier. Alternatives to the adversarial response model have become increasingly popular of late. Baldwin's seminal article on "positive sanctions" (Baldwin, 1971) as well as considerable economic theory suggesting the power of incentive-based approaches have provided the theoretical foundation for such efforts. Financial mechanisms have become frequent elements in environmental regimes, beginning with the Montreal Protocol, but more recently adopted in the FCCC and the Biodiversity Convention. Debt-for-nature swaps build on a notion that developing countries are willing and even eager to preserve their environmental resources but need assistance or incentives to do so. Other treaties have relied largely, if not exclusively, on increasing the flow of information about a problem or its solutions to alter behavior, particularly in areas where the belief is that, if we can provide states with information about the environmental damage they are causing to themselves, let alone others (thus convincing states that they are not in a PD situation), they will alter their policies appropriately (Levy, 1993; Mitchell, 1996). Several regimes include training programs to ensure that knowledge available to developed states is also available to governmental, corporate, and individual actors in developing states.

Facilitative approaches can induce different but equally perverse dynamics as those of an adversarial approach. On the one hand, facilitative approaches tend to be more costly because they place the regime at risk of "moral hazard" dynamics. By offering assistance to potential noncompliers, facilitative approaches provide incentives to those who might otherwise comply to threaten to violate the treaty to gain access to the offered assistance. This can make such approaches more costly than corresponding adversarial approaches which create incentives not to violate and need only follow through with a response with those actors who have demonstrated that the threat has failed. Facilitative approaches also assume that the actors being induced to comply do not have incentives to take advantage of the regime managers, a costly assumption if wrong.

On the other hand, facilitative approaches may work even when their assumptions about actors are wrong. If adversarial approaches can create a negative self-fulfilling prophesy in which actors become more committed to violating an agreement because they are being punished for doing so, facilitative approaches can induce a positive self-fulfilling prophesy in which the "naive" assumption that a state "wants to comply but can't" leads to an interaction which induces the state to adopt such a position. 6

The tendency to overgeneralize

Not atypically, the infancy of this debate over why states violate agreements and how to reduce this tendency has led to efforts on both sides to overgeneralize across issue areas. Scholars drawing on security and trade experience and hewing closely to a realist paradigm of state behavior (as well as many policy practitioners) tend to predict that environmental and human rights agreements, like arms control and trade agreements, will be plagued by intentional violations unless effective deterrent sanctions are put in place (Downs, Rocke and Barsoom, 1996). To reduce such behavior, the most effective response regardless of issue area is to create a system in which potential violators are convinced of both the likelihood that, and magnitude of, sanctions imposed for noncompliance. Indeed, in this vein, even those concerned with environmental compliance often critique environmental regimes for lacking "teeth," in the form of coercive enforcement (Barratt-Brown, 1991; Howton, 1994, 176; Jamieson, 1986, 218-219; Taplin, 1994, 277). More recent scholarship sympathetic to a neoliberalist paradigm of state behavior have argued that, even in arms control, "noncompliance is not necessarily, perhaps not even usually, the result of deliberate defiance of the legal standard" (Chayes and Chayes, 1991). 7 Compliance is less often ensured by threatened sanctions than processes of ongoing diplomatic management and persuasion.

Yet, both of these views seem to overgeneralize from one paradigm or issue area to the other. It seems unlikely that either of these "one size fits all" arguments is correct. Intuitively, one might well assume that the best means of inducing compliance would require adopting strategies that are responsive to the causes of noncompliance. Indeed, both sides in the debate rely on exactly such a notion that responses should reflect the causes of noncompliance. They err, however, in assuming that the causes of noncompliance are consistent across issue areas. There is no reason to believe that the "mix" of causes of noncompliance in arms control should look similar to those in environmental affairs.

"Rational" Choices Of Responses To Noncompliance

There are at least two reasons to expect that states will establish compliance procedures and institutions that are responsive to the noncompliance demographic context that they face. At first blush, we might expect a one-to-one relationship between the causes of and responses to noncompliance. As the foregoing discussion of responses suggests, any particular cause of noncompliance usually suggests a dominant or salient response that reflects beliefs about the "appropriate" response to the noncompliance, given its causes, and about the "effectiveness" of that response. Thus, for example, the enforcement school makes this linkage: sanctions are seen as appropriate responses to intentional regime violations and also are assumed, often implicitly and despite contradicting empirical evidence, to be effective at deterring such violations in the future. Managerialists also link the causes of noncompliance with the types of regime responses prescribed. On the one hand, they argue empirically that states don't violate treaties intentionally and so sanctions are neither appropriate nor likely to be effective. On the other hand, they make a prescriptive argument that, even if states do violate treaties intentionally, sanctions may be less effective than more reward-based or facilitative techniques.

This latter point suggests that the cause of noncompliance need not deterministically dictate one, and only one, "best" response. Considerable evidence suggests that sanctions often fail to achieve their specific or general deterrence goals, even when the violations being sanctioned are clearly intentional. Recent adoption of economic theory has led to sanctions increasingly being replaced by financial and other incentives in both domestic and international policy (Baldwin, 1971). Although they are undertheorized in international relations, alternative strategies such as the nonproliferation "technology denial" or domestic gun buy-back and tax amnesty programs mentioned above suggest that policy makers have a range of potential responses to a given noncompliance incident. 8

"Getting the response right on average"

If states create international institutions in a "rational" manner, seeking to create institutions that most effectively achieve the desired goals, we should expect states to create response strategies that are effective at inducing compliance. If the effectiveness of a strategy at inducing compliance is dependent on the causes of noncompliance than we can develop two hypotheses about the shape of compliance institutions. First, we should expect international institutions to create compliance systems that involve a dominant or "default" response that is effective in inducing future compliance given the mix or noncompliance setting it faces. This is a "get it right on average" criteria: states will create institutions that at least get the right response on average. Institutions will have strong incentives to take an actuarial approach to compliance in which the goal becomes to choose a single strategy of response that provides the best response to the most common cause of noncompliance. Thus, if adversarial approaches such as sanctions are deemed effective responses to intentional violators and facilitative approaches such as capacity enhancements are deemed effective response to good faith noncompliers, then we should expect states to evaluate the most likely types of noncompliance and develop institutional mechanisms accordingly.

"Getting the response right in individual cases"

A more refined notion of institutional rationality would suggest that states create discriminating institutions capable of getting the response correct more often than just "on average." To provide an example, faced with a noncompliance setting in which a majority (say, 70%) of the violations are due to incapacity problems, a facilitative approach would be better than an adversarial approach if a single response is to be used in response to all  violations. However, a more effective strategy would entail identifying mechanisms that could discriminate between different causes of noncompliance, and could respond to incapacity noncompliers with a facilitative approach while threatening sanctions against the remaining 30% of intentional violators. Thus, a more demanding standard of institutional rationality would lead us to expect institutions which develop both a repertoire of potential responses appropriate to the set of potential causes of noncompliance as well as the ability to discriminate the different causes of noncompliance. This is a "get it right in individual cases" criteria.

There are, however, several reasons why even rational states might not create institutions that meet this more demanding standard even while creating institutions that create the earlier, less demanding, one. First, states might make a rational institutional decision only to "satisfice," recognizing that they could get it "more right" with a more discriminating identification and response system but that the effort to design and maintain such a highly capable institution would not be offset by the increased levels of cooperation that it would produce. Second, states might recognize that international institutions are inherently hobbled in terms of access to information and flexibility that make creating such a discriminating and responsive institution impossible, even if worthwhile.

A Tale Of Two Cases

Case selection

Much recent "regime compliance" research has addressed the realist claim that state behavior is driven exclusively by power and interests (Chayes and Chayes, 1993; Chayes and Chayes, 1995; Downs, Rocke and Barsoom, 1996; Jacobson and Brown Weiss, 1995; Mitchell, 1993; Mitchell, 1994). In engaging this claim, many neoliberalist scholars have sought to use "hard cases" to demonstrate that international regimes can shape state behavior, producing outcomes different than would have occurred in their absence (Young, 1992). Well-selected "hard cases" -- in which regimes would have been highly unlikely to influence state behavior -- provide the analyst with compelling evidence that a state did not have any other reasons to behave as it did, thereby refuting subsequent, and tautological, claims in which a state's behavior is used to indicate the state's interest in engaging in it. Methodologically, hard cases allow more confident generalization: if regimes can be shown to influence state behavior in hard cases, then, a fortiori , they may influence state behavior in "easier" cases. Thus, both the nature of realist claims and explicit calls from realist scholars have demanded that neoliberal scholars demonstrate that regimes influence behavior in hard cases. 9

But, to understand international regimes and institutions more generally, we may be as interested in variation within the universe of cases of international cooperation as in the characteristics of the "hard" cases. We may be interested in the full repertoire that regimes use to induce compliance and not just how they operate when it is difficult to do so. We may be interested in the more common mechanisms for inducing compliance as well as the extreme cases. Rather than seeking generalizations that hold true across contexts, we should seek out the factors that cause variation across contexts. Rather than arguments that punishment strategies underlie deep cooperation in all issue areas or that managerial approaches will work in all contexts, we should look for factors that explain variation in the type of strategies states use in regulatory regimes.

The experience with onsite inspection in two multilateral treaties provide an initial basis for evaluating whether the dominant sources of noncompliance vary across issue areas and, more importantly, whether international regimes exhibit any degree of responsiveness to this variation in how they evaluate and respond to noncompliance. Onsite inspection (OSI) holds particular theoretical interest within the larger debate on the role of transparency in international regime maintenance. Transparency is considered to play an important role in enhancing the compliance and effectiveness of international regimes (Mitchell, 1998). As Downs, Rocke, and Barsoom note, transparency's importance "almost inevitably lies in the reaction that it provokes. This reaction is, of course, the substance of enforcement" (Downs, Rocke and Barsoom, 1996, 393). During the Cold War and in most recent arms control conventions, OSI has become a core mechanism for increasing transparency, and hence for increasing compliance and effectiveness. More recently, several scholars have urged the adoption of OSI or similar techniques in environmental regimes (Ausubel and Victor, 1992; di Primio and Stein, 1992; Fischer, 1991a; Fischer, 1991b; Greene, 1994; Kellman, 1995; Victor, Lanchbery, and Greene, 1994; Weidner, Zieschank, and Knoepfel, 1992). 10 Thus, onsite inspection represents, at least in some ways, a hard case for the argument here since most of this literature implicitly or explicitly suggests that OSI will be used in an adversarial mode in response to intentional violator contexts. Finding both an arms control and an environmental case in which OSI was used in a multilateral treaty setting but within different noncompliance contexts allows us to hold constant several factors that would influence response strategy (e.g., nominally similar procedures, number of actors) while allowing the issue area to vary.

The Convention on Wetlands of International Importance Especially as Waterfowl Habitat (1971), known as the Ramsar Convention for its place of signature, and the Convention on the Prohibition of the Development, Production, Stockpiling, and Use of Chemical Weapons and on their Destruction or Chemical Weapons Convention (CWC) provide useful cases in which to examine the issues delineated in the foregoing paper. Both have incorporated onsite inspection procedures as mechanisms for enhancing compliance. Both are active, multilateral agreements. The availability of at least one environmental case allows us to examine whether the recommendations to adopt compliance mechanisms tested in the field of arms control for use in the field of the environment have validity.

These two cases present an empirical puzzle. In the arms control context, OSI looks quite adversarial while, in the environmental context, OSI has a much more facilitative feel. Should we expect similarity? If so, how do we explain the difference? Does a comparison of the two cases support the argument made above that states are constructing compliance mechanisms responsive to their noncompliance context? The evidence presented here suggests that OSI has the potential to play an important role in environmental regimes, but that it will operate by rather different causal pathways than in arms control context.

Comparing the likely causes of noncompliance: strategic setting and "noncompliance demographic"

The Ramsar Convention and the CWC face different strategic settings that produce different noncompliance demographics, or distributions of causes of noncompliance. In short, the CWC more closely fits the "intentional violator" assumption, while the Ramsar Convention more closely fits the "good faith nonconformer" assumption. In both cases, the shape of the noncompliance demographic results from an interaction between the factors inherent to the strategic setting and to institutional choices about the definitions of compliance and noncompliance. Noncompliance in the CWC seems likely to be dominated by intentional violations by states seeking to develop chemical weapons to provide for their security, while avoiding other states finding out about these developments. Threats to wetlands stem from inadvertence, ignorance, or the incapacity to properly protect their wetlands. which to e

PD dynamics and relative gains concerns

The contrast in the strategic setting of regulating chemical weapons production and wetlands degradation provide core reasons for expecting different noncompliance demographics in the two arenas. Consider first the applicability of the PD paradigm to the two issue areas. The CWC represents a quintessential arms race dynamic that can be readily compared to a PD dynamic. It seems appropriate and accurate to assume that most states rank their preferences as 1) get other states to forego chemical weapons while acquiring them yourself, 2) forego chemical weapons and get other states to also forego them, 3) acquire chemical weapons and have other states also acquire them, and 4) forego chemical weapons while other states acquire them. This standard PD preference ranking suggests that states will have ongoing incentives to defect from any agreement that is made. In contrast, the PD assumptions map less well onto the wetlands preservation case. Especially with growing knowledge and concern about the local benefits of wetlands, from sewage treatment and water recharge to bird habitat and ecotourism, the goal has been to induce compliance with wetland preservation regulations by increasingly convincing states that wetland preservation is a dominant strategy. Indeed, as I argue elsewhere, the Ramsar Convention's major task of inducing wetlands preservation may well involve education more than PD type compliance management. In addition, if we accept that states are not always and exclusively concerned with relative gains (Powell, 1991), it seems reasonable to suppose that states tend to be more relative-gains-driven in security realms while being more absolute-gains-driven in environmental realms. Both sets of considerations suggest that states in the CWC will have incentives to use any cooperative arrangements that are made to gain advantage of other regime members by violation while these incentives for intentional violation will be far less in the wetlands case.

Compliance definitions, intentionality, and capacity

The nature of the primary obligation in the two regimes also differ in ways which differentiate how intentionality, incapacity, and governmental control relate to noncompliance. In the CWC case, states assume a primary obligation to refrain from developing or acquiring chemical weapons. Violations of the CWC require intentionality. Chemical weapons programs do not happen by accident. They require sufficient resources and engage sufficiently high-level security concerns that they can be safely assumed to be undertaken only with the explicit knowledge and support of the highest levels of a government. Wetlands degradation represents almost the polar opposite. Wetlands destruction is rarely undertaken as the intentional goal in itself. Even in historical periods, wetland drainage has been undertaken as a means to a higher priority goal, such as land reclamation or mosquito eradication, rather than as an end in itself. Thus, whereas governments are quite likely to be committed to chemical weapons programs as the only means to achieving their security goals, governments seem less likely to hold the same level of commitment to destroying wetlands, so long as they can be provided with alternative means for achieving the economic or policy goals that motivate them. Even when the failure to fulfill the Ramsar Convention's requirement of "wise use" results from government action, for many states this is likely to reflect undesirable (from that state's point of view) and often unintentional (even if usually foreseeable) results of actions undertaken for other reasons, while failure to fulfill the CWC's requirement to refrain from chemical weapons development reflects an intentional, deliberate, and high level policy choice. That said, states generally committed to Ramsar may nonetheless decide quite consciously to place other priorities, such as economic growth and development or national security, ahead of wetland protection with respect to a given wetland. Thus, intentional degradation of a wetland will remain a possible outcome.

Given the nature of the behaviors and definitions of compliance and noncompliance, financial, technical, or administrative incapacities will lead states to comply with the CWC's requirements, not violate them. By contrast, these same incapacities in the Ramsar case will lead states to violate the convention's requirements. Thus, for example, the administrative incapacity to prevent drainage of a wetland by local farmers or to provide alternative sources of water to people using the wetland for drinking water can lead to serious wetland degradation. Notably, although incapacity seems unlikely to produce CWC violations, that does not mean such causes of noncompliance are limited to the environmental sphere. Incapacity can be, and has been, a problem in other arms control conventions, most notably the START accords in which requirements to dismantle nuclear weapons have been violated by former Soviet states claiming they lack the necessary financial or technical resources to comply.

Inadvertence concerns

Inadvertence would seem unlikely to play any role in violations of the CWC while being a regular feature of violations of the Ramsar Convention. In the CWC case, one can imagine the possibility of a substate actor either within or outside the government undertaking a chemical weapons program. However, if the actor were outside the government, it seems unlikely that the CWC would even consider this to be a violation of the convention. 11 In contrast, if the actor were within the government, the nature of the decision process and assumptions about the necessary involvement of high level actors would lead to deep skepticism about a claim of inadvertence and lack of control over an allegedly "rogue" element of the government. However, in the Ramsar case, inadvertence is likely to be a common, and accepted, reason and/or excuse for noncompliance. Developments planned to take wetland preservation into account may nonetheless fail to do so either because of poor understanding of local hydrology and wetland dynamics, or because those building housing, a hotel, or a road fail to take the necessary actions to accomplish that primary goal while also properly preserving the wetland ecosystem.

Ignorance and its implications for violation or compliance

Finally, consider ignorance as the cause of noncompliance. The requirement to refrain from developing a chemical weapons capability is sufficiently clear and salient that noncompliance due to ignorance of the obligation and what it entails can be precluded on its face. And, as with incapacity, ignorance in the CWC case is likely to be ignorance of how to violate the agreement, and hence is likely to reinforce compliance. Nor is knowledge necessary to successfully comply, since, for most states, simply doing nothing keeps it in compliance. In contrast, ignorance both of what is required and how to achieve it play important roles in the Ramsar context. The vagueness of the requirement that states make "wise use" of their wetlands has been a consistent problem for states. The Contracting Parties have made numerous attempts to clarify what is meant by wise use which has produced an extensive exegesis of what is meant by the term, but has by no means eliminated a sincere post hoc claim that a government did not understand that a particular behavior was forbidden (or required) by the "wise use" criteria. Indeed, knowing whether a state is complying in a strict sense with the "wise use" requirements of the Ramsar Convention would be quite difficult. Even if a state knew what was required, it could well claim ignorance about how to achieve it. Understanding how to prevent certain changes in the ecological character of a wetland requires scientific knowledge that is often beyond the expertise of those tasked with protecting a given wetland, particularly when those tasked with such are overburdened, underfinanced, and often powerless bureaucrats.

The CWC is, thus, dominated by intentional violations with very few convincing reasons to expect noncompliance due to incapacity, inadvertence, or ignorance. The Ramsar Convention, on the other hand, faces a much more varied context with noncompliance due to all four sources possible but with intentional violations likely to be a far less dominant theme.

Comparing the noncompliance approaches

The divergence in the likely sources of noncompliance appears to have produced a corresponding divergence in the two regime's approaches to OSI. The CWC's approach to OSI appears to be fundamentally adversarial while the Ramsar Convention's approach to OSI appears fundamentally facilitative. 12 Rather than attempt to make any claims regarding exactly how facilitative or adversarial the approach of either convention is, this section addresses this evaluation by comparing the regimes to each other with respect to the goals of inspection, control over the terms of inspection, the incidence of inspection costs, how the regime responds after an inspection, and the perceptions of inspection by those being inspected.

Goal of the inspection

Although the two OSI procedures operate quite differently, they both seek to induce the inspected actor to come into compliance with the terms of the international treaty. Common notions regarding OSI have been influenced by its birth in the arms control context. Within that context, OSI was considered a necessary and appropriate response because of deep-seated beliefs regarding 1) the ubiquity of incentives for intentional violation and deception; 2) the inability of establishing transparency by other means because of the resistance to providing regular, accurate, and voluntary reports; and 3) the effectiveness and, indeed, necessity of an adversarial approach in inspections of, and responses to, noncompliance. This "arms control" variant of OSI is best exemplified by the current IAEA inspections of Iraq. An alternative variant of OSI seems to be at work in the Ramsar Convention. In this regime, the operational assumption appears to be that noncompliance arises more often from incapacity, inadvertence, and ignorance. Indeed, given this assumption, it might seem that OSI is, at best, unnecessary and, at worst, inappropriate and counterproductive. However, as the Ramsar Convention experience illustrates, that is to retain assumptions of an adversarial response to noncompliance, which appear not to be valid.

From a different point of view, the two OSI procedures can be seen as diverging in terms of the informational goals of the inspection. In the CWC case, assumptions regarding the nature of violations correspond to a notion that the target of the inspection already knows the "truth" about the target's behavior. The inspection serves to provide other regime members with information to reassure them that they also know the "truth" about the target's behavior. In a Ramsar Convention inspection, corresponding to the assumption that violations are not intentional, the target state itself may be unaware of the degree to which the wetland has been degraded. The inspection serves not to provide other members with a "truth" about target behavior that the target already has, but to provide both the target and the inspectors with the "truth" about a wetland that neither side has prior to the inspection.

Control over the terms of inspection

Comparing who defines the terms of reference under which inspections occur provides the most telling evidence of a difference in approaches in the two regimes. In the CWC, the "target," i.e., the actor to be inspected, has little say about the timing, content, scope, or depth of an OSI or about the inspection team membership. By contrast, the Ramsar Convention's Management Guidance Procedure (MGP) for onsite inspection is initiated by the country to be inspected, who also dictates the terms under which the inspection will be undertaken.

The CWC provides for three types of inspections: initial, routine, and challenge. Initial inspections of a "declared facility" subject to the CWC provide the basis for negotiating a formal "facility agreement" between the Organization for the Prohibition of Chemical Weapons (OPCW) and the contracting party. This facility agreement sets the limits on the "amount and type of inspector access" during routine inspections of declared facilities. In the event that sufficient suspicious evidence becomes available to warrant a challenge inspection, the OPCW can authorize an "inspection mandate" to identify "any evidence that might suggest that a facility is producing, storing, diverting, or exporting proscribed or limited CW agents, equipment or munitions." 13 Although these inspection mandates are usually limited in scope, authorizing inspection for particular chemicals or facilities, these limits are established by the OPCW, not by the target of the inspection.

Even the regular routine inspections under the CWC reflect basic suspicions of malicious intent on both sides. On one side, the multinational group of inspectors selected by OPCW from a standing pool of inspectors, conduct the inspection with an initial perspective that the government or facility being inspected has incentives to violate the terms of the CWC, and may have acted on those incentives already. On the other side, governments and facilities being inspected assume that inspectors have incentives to collect valuable military, economic, or industrial information that is unrelated to the task of treaty verification. 14 Although the OPCW provides a suspected violator little leeway to dictate the legal terms of reference for a challenge inspection, it is clear that both governmental and private targets have strong incentives to minimize the practical access that inspectors are given during any inspection.

By contrast, the processes of the MGP established in 1988 by the Ramsar Convention's Standing Committee reflect assumptions of mutual assistance rather than mutual mistrust. 15 The Convention requires Contracting Parties to notify the Bureau when "the ecological character of a listed wetland is changing or is likely to change as a result of technological development, pollution or other human interference." 16 The Ramsar Bureau (the regime's secretariat) can be notified of a threat to a listed wetland by the Contracting Party, but can also have its interest triggered by another Contracting Party, an NGO, or an individual. Either the Contracting Party or the Ramsar Bureau can make the initial recommendation to add the site to the Montreux Record of endangered sites, with listing on the record dependent upon acceptance by the relevant Contracting Party. Once a Contracting Party lists a site, a Management Guidance Procedure is initiated by the relevant Contracting Party extending an invitation to the Ramsar Bureau to undertake an inspection. Unlike in the CWC case, the Bureau creates an inspection team composed of Bureau staff members and technical and scientific experts who visit the site under terms of reference dictated by the Contracting Party, including the dates for the inspection, the people who should be contacted, the information to collect, and other information desired by the Contracting Party. For example, in a Trinidad and Tobago case, the terms of reference requested an analysis of the condition of the wetlands, identification of the impacts of resource exploitation, and recommendation of potential strategies for multiple sustained use.

The activities undertaken during an MGP demonstrate its facilitative nature. An MGP involves "one or more site visits by Bureau staff and its consultants, specialists who are experts in the particular problems involved. They may be representatives of other Contracting Parties or nongovernmental partner organizations. The subsequent report includes a detailed analysis of the situation and recommendations for future action in order to arrive at an acceptable solution to the problem" (Davis, 1994, 55). Inspections usually entail equal time spent investigating the environmental status of the wetland and spent meeting and presenting results and recommendations with government officials. The purpose of the Montreux Record is to "identify priority sites for positive national and international conservation attention" (Ramsar Convention Bureau, 1996). The purpose of the MGP "is to bring about the steps necessary for the removal of the site from the Montreux Record" (Ramsar Convention Bureau, 1996). During an MGP, the Contracting Party being inspected usually facilitates access to not only the areas already identified as necessary, but may provide greater access than provided for in the terms of reference, as it becomes clearer how access to additional areas would help the inspectors better understand the causes of the wetland degradation and the possible remedies for it.

Incidence of inspection costs

The two regimes also diverge with respect to the costs that an inspection involves and who bears those costs. In both cases, the costs of the inspection itself are covered by the secretariat, either the OPCW or the Ramsar Bureau. Notably, however, NGOs foot the majority of the bill for the MGP inspections whereas the costs of OPCW inspections are covered almost exclusively by state contributions.

More important than the direct costs of the inspection are the extent to which the target being inspected views there as being indirect costs to an inspection. A target of a CWC inspection is likely to view an inspection as involving certain unavoidable collateral costs as well as the risk of even greater costs. The indirect but unavoidable costs of a CWC inspection involve the fact that normal operations of the chemical facility are either disrupted or completely halted by the inspection. The far greater risk, however, is that inspectors will gain militarily-sensitive or economically-proprietary information unrelated to the CWC convention but nonetheless valuable to the inspectors. As SIPRI notes in its analysis of the CWC, "in view of the scope and intrusiveness of the verification regime, great emphasis has been put on the obligation of the OPCW to carry out verification in an unobtrusive manner and to safeguard confidential information received from the states parties" (SIPRI, 1997). Targets, informed by the basically competitive atmosphere of arms control and security dynamics, understandably concern themselves with the risk that inspectors from particular rival countries, even if "cleared" by the OPCW, will use the inspection to gain a military or economic advantage that will be used against them in some other realm and at some later date.

The nature of the Ramsar Convention make it highly unlikely that any inspection will uncover information that is militarily or economically valuable to the inspectors. Conforming to the notion that relative gains concerns are more at play in the security arena than the environmental arena, targets do not assume that inspectors will use the MGP to get information that will advantage them in some other realm of competition. Indeed, to the extent that the MGP inspectors exceed their mandate, they are likely to turn up information that the target itself was not aware of and will help it either do a better job of managing their wetland or some other environmental amenity. Thus, inspections that exceed their mandates are likely to provide benefits to the target in the Ramsar case while imposing costs on the target in the CWC case.

One way to conceptualize this difference is in terms of the value of secrecy and private information vs. the value of transparency. The nature of arms control creates a significant need for "collateral secrecy" in which revealing any unnecessary information is considered costly. Even the most "good faith" member of the CWC has strong incentives if inspected to minimize the information revealed to inspectors. Besides the incentives those interested in cheating on the regime have to limit inspection depth and breadth, even those supportive of the regime will want to minimize transparency because of the value of secrecy to other, closely linked, arenas. Equally important, "private information" holds considerable value in the arms control context, whereas the less competitive nature of the environmental arena makes private information far less valuable, and indeed, may make aggregate, collective information particularly valuable. Although the aversion to revealing information is inherent in the issue area, this aversion is reinforced by an adversarial approach to inspection. Even without the threat of sanctions, a state developing chemical weapons would have incentives to do so secretly, at least until it was strategically useful to reveal their acquisition, to maximize the security benefits of the acquisition. However, these benefits are reinforced by the adversarial approach taken by the regime, since secrecy then provides a way of reducing the likelihood of being sanctioned.

Additional indirect evidence of the difference in these concerns can be seen in the governmental level at which the regime is addressed. Contact with the Ramsar Bureau occurs at a relatively low bureaucratic level. Ramsar Contracting Parties must designate an agency to be the "Administrative Authority" responsible for treaty implementation. Almost every state has designated a division of their agriculture, environment, or natural resource ministry. For example, the administrative authority for the United States is in the Office of International Affairs, Fish and Wildlife Service, (Ramsar Convention Bureau, 1997). Leaving Ramsar implementation to these relatively low levels of government in general, also tends to give them free play in arranging OSIs. 17 By contrast, in the CWC, OSI visits involve not only high level decision makers within the ministry of defense but also consistently involve personnel of the ministry of foreign affairs and top level executive branch decision makers. 18

After inspection, what?

Additional evidence of the facilitative nature of the inspection are the subsequent reports. In the Ramsar case, the recommendations of the MGP report appear to consist largely of identifying and prioritizing the causes of the wetlands degradation and urging the Contracting Party government to undertake actions that address those causes. In general, the recommendations call on the government to halt actions by government agencies or private actors that are harming the wetland, to enforce existing regulations that would slow wetland degradation, or to commence actions that could reverse trends in wetland degradation. Often, even the request that an MGP be undertaken may reflect efforts by the government to identify the sources of the problem. A request can also reflect efforts by a relatively weak element in a government to raise the political salience of a known source of wetland degradation, thereby increasing their ability to accomplish their wetland preservation task against other bases of power within the government. 19 Recommendations do not provide any sense that the degradation reflects intentional government efforts to destroy the wetland. Rather, the recommendations take a tone that wetlands degradation has occurred because of government ignorance of the impacts of given policies on a wetland or of the ecological state of the wetland; government failure to give wetland preservation priority, by failing to adopt or to enforce appropriate policies; or government incapacity to undertake actions that could preserve the wetlands. 20

Notably, the final recommendation of an MGP often entails urging the Contracting Party to "remain in close contact with the Convention Bureau with a view to seeking technical assistance funds" (Davis, 1994, 57). For example, as a result of the 1990 MGP to Jordan, the World Bank and the United Nations Development Program allocated $3.3 million through the Global Environment Facility for wetland protection projects (Davis, 1994, 57). The Fourth Conference of the Parties at Montreux in 1990 also established a Wetland Conservation Fund (WCF) specifically targeted at assisting the wetland conservation efforts of developing states. The WCF creates an incentive for developing countries to join the regime and implement the corresponding commitments. Although the funds directly provided through the WCF are quite limited (and provided in large part by NGOs), the Ramsar Bureau has administered the WCF to enhance the funds available to developing countries for wetlands protection by helping countries gain access to grants from the World Bank, other multilateral and national development banks, or other nongovernmental sources (Garcia, 1992). Even when states are not complying or not even members of the Convention, the Ramsar Bureau seeks to use MGPs as educational tools. In some ways this can produce a process parallel to the "general" deterrence function of sanctions. Thus, for example, the Ramsar Bureau "stressed the importance of the Monitoring Procedure at Nariva Swamp as a case study, from which lessons could be learned and applied in wetlands with similar problems elsewhere in the Neotropics, but particularly in the insular Caribbean region. Especially important are the sharing of coastal management and planning needs, community co-management, national administration and wise use of wetlands , by those other countries with similar population, social and economic realities" (Ramsar Convention Bureau, 1996).

By contrast, a CWC report is far more likely to identify that the target was engaged in activities banned by the convention and that the behavior was intentional. The target is likely to be told what it needs to do to bring itself back into compliance. Although, as with any international treaty, sanctions are likely to be difficult to get implemented by member states, they are explicitly delineated in the CWC's Article 12 as the sole response to noncompliance. At the same time that it seems unlikely that states would actually undertake collective sanctions against a state revealed as undercutting the CWC, it does seem likely that publication of any information within a CWC inspection report that reflects even the possibility of clandestine violation is likely to produce considerable negative diplomatic response to the target state. Precisely because the suspect state would be, most likely accurately, assumed to have violated the convention intentionally, the notion of other states identifying ways to provide resources to facilitate compliance seems unlikely. That said, were a CWC OSI to turn up a violation that appeared to have been caused by an incapacity to control certain actors or behaviors and were it clear that the target state had no prior knowledge of this and sought assistance in avoiding such outcomes in the future, other regime members might well provide the requested assistance.

Almost as indicative of a facilitative approach to compliance is the response to those who are complying. The two cases suggest that regimes may be distinguishable based on whether compliance is expected and noncompliance is to be frowned upon or noncompliance is expected and compliance is to be praised. In the Ramsar case, MGP reports frequently commend the inspected state for efforts made to comply, even when those acts have fallen short of producing the desired result. Given that compliance requires acts of commission, that these acts are costly, and that these acts require foregoing other normatively-appropriate goals, it is not surprising that the Ramsar Convention praises states, particularly though not exclusively developing ones, for taking such steps. By contrast, compliance with the CWC requires acts of omission, essentially requiring states simply to refrain from actions to acquire chemical weapons. Since such restraint involves only the costs of foregoing certain security options which the regime itself seeks to characterize as normatively-inappropriate, compliance is considered to be expected rather than deserving of praise, and noncompliance is to be frowned upon.

Response of targets to inspections

Finally, we can consider how targets perceive inspections as evidence of whether an OSI system is facilitative or adversarial. In the CWC case, considerable evidence exists that states are reluctant to be inspected and understandably concerned about the possibility of an inspection even when they have no intention of violating the agreement. Considerable political concern has been expressed in the US, for example, about the intrusiveness of inspections, the requirements and costs imposed on chemical plants and facilities that are subject to inspection, and the economic and military risks of being willing to submit to inspections, even while recognizing that doing so is a necessary (though not sufficient) quid pro quo to gain access to inspections in other states. Targets consistently see OSI as a costly and risky burden, only worth accepting if the benefits of cooperative arms control can be shown to outweigh it.

In quite stark contrast to this perception of OSI in the CWC is the common response to Ramsar's MGP. Receipt of the 1995 MGP inspection report on a wetland led the Ramsar delegate from Trinidad and Tobago to recommend "more frequent use of the [MGP] to other Contracting Parties, based upon the technical analysis and advice that resulted and the momentum that was created" (Ramsar Convention Bureau, 1996). This view of the MGP as providing positive benefits rather than involving costs poses a striking contrast to the CWC case.

Analysis And Conclusion

The foregoing suggests that states have established a facilitative approach in the Ramsar case and an adversarial approach in the CWC case. Without comparing the effectiveness of the two approaches at inducing compliance, we can inquire into the institutional design question of why they have chosen the different approaches to OSI and compliance management that they have. Such an analysis suggests that the differences in approach reflect differences in the noncompliance demographics and strategic setting that they face, different levels of fault tolerance, and different levels of concern over the problem.

In both the CWC and the Ramsar cases, the approaches to OSI appear to clearly be responsive to the most common causes of noncompliance. The CWC faces a noncompliance environment in which most noncompliance is likely to reflect intentional decisions to violate and has adopted an approach to OSI and to responses to revealed violations that relies almost exclusively on an adversarial model of compliance. The Ramsar Convention faces a noncompliance environment in which most noncompliance is unlikely to be intentional, being due either to lack of concern, lack of ability, lack of information, or lack of control. In response, the MGP has developed a far more facilitative approach to compliance management, seeking to help nations do better at preserving their wetlands rather than sanctioning them for failing to do so.

Both approaches seem particularly well-suited to the noncompliance environment they face. At an intuitive level, it seems unlikely that the facilitative techniques of the Ramsar Convention, even were it politically possible to adopt them, would be effective in the CWC context. Likewise, when faced with both the frequency with which wetlands degradation arises because of incapacity or ignorance and the low priority given to wetlands conservation relative to other concerns, it seems that the sanctions implied within the CWC context would be more likely to lead states to leave the regime rather than to conform to its dictates.

That said, however, some evidence suggests that either regime could also have adopted other approaches. The gun buyback and tax amnesty approaches occur in noncompliance contexts that are clearly dominated by intentional violations but have succeeded at inducing levels of compliance among at least some actors. Likewise, in the Ramsar case, sufficiently large and powerful sanctions might prove effective, especially if they were large and likely enough to lead states to dedicate resources currently dedicated to other uses to wetlands protection, thereby overcoming the incapacity or ignorance problems that were causing the noncompliance. Indeed, if the level of cooperation within the Ramsar regime deepens in the future and if Downs, Rocke, and Barsoom are correct in saying sanctions are essential to inducing compliance with deeply cooperative agreements, sanctions may become an increasingly common policy option.

Notably, neither regime appears to have adopted a particularly discriminating approach to compliance management. The CWC does not appear to have created any institutional mechanism for identifying or responding differently to those cases, however rare, in which noncompliance was due to inadvertence, ignorance, or incapacity. Likewise, the Ramsar Convention does not appear to have created institutional machinery for identifying or responding to intentional destruction of wetlands or at least those cases in which states clearly understand that their behavior will lead to the degradation of a listed wetland.

The two approaches to OSI differ with respect to how assumptions about noncompliance, even when erroneous, can influence behavior. An adversarial approach assumes project participants are trying to cheat, assumes that nonperformance is likely to occur and likely to be intentional, and assumes that deception or concealment is widespread. Targets are likely to reciprocate this adversarial approach, seeking to avoid revealing even minor or unintentional problems, particularly if any observed nonperformance is likely to bring sanctions. With such an approach, acquiring accurate and timely information can be difficult and costly, if not impossible. A facilitative approach assumes that participants are trying to implement successful projects, that nonperformance if it occurs is likely to be inadvertent, and that first-party reporting is likely to be largely honest. Participants are likely to reciprocate this approach as well, revealing project problems before they threaten performance, particularly if the response is likely to be useful assistance in solving problems and bringing nonperforming projects back into compliance. Thus, to the extent that participant behavior is not determined independently of the conduct of monitoring, both these approaches can set in motion reciprocal behavioral dynamics that make their embedded assumptions self-fulfilling. The facilitative approach encourages the good-faith implementation it assumes, at the cost of being slower to detect the intentional nonperformance or attempted deception that it assumes to be unlikely. The adversarial approach may be quicker to detect intentional cheating, at the cost of making implementors who are ambiguous in their intentions more likely to try it. There are no instances of strongly adversarial approaches in real international environmental regimes, though environmentalists often advocate them. Strongly facilitative approaches are more common, such as the wetlands convention (Ramsar Convention Bureau, 1990).

The two strategic contexts also appear to vary in the "fault tolerance" of the member states. In the CWC, a sanctions-based approach seems more sensitive to ensuring that certain basic behavioral proscriptions are not violated by any actor, because violation even by a single actor could be devastating. In contrast, the Ramsar Convention appears more concerned with maximizing the amount of behavioral change by member states, even if some do not change their behavior at all. This divergence in the goals of the two regimes suggest the need for more careful distinctions about the goals of different types of regulatory regimes in the future.

The foregoing evidence does not directly contradict Downs, Rocke, and Barsoom's argument that deep cooperation requires enforcement. Concern over chemical weapons is particularly strong among nations, and the eventual conclusion of the CWC demonstrates a major attempt to induce deep cooperation that will be costly and risky for member states to follow through on. The regime also clearly uses the enforcement type of strategy Downs, Rocke, and Barsoom advocate. Wetlands protection, even though negotiated into an international convention over 25 years ago, has far shallower support. The Ramsar Convention requires only that member states must list a single site, and must make "wise use" of their wetlands. To the extent that a state can even be aware of what is required of it with respect to "wise use," it is unlikely that this requires "deep" cooperation. Given that there is considerable evidence that wetland degradation continues at a rapid pace, it also seems that wise use, however defined, has not become a commonplace. That said, however, it also seems that the Ramsar Convention in general and the MGP, the Montreux Record, and the Wetland Conservation Fund have together led to some level of enhanced protection of some wetlands that would not occurred otherwise.

This suggests the need to begin analyzing more carefully how regimes induce cooperation. At a minimum, it seems that the compliance institutions states build into regimes need not be identical. These differences in compliance institutions do not reflect merely random variation but reflect apparently careful choices in response to divergent noncompliance settings. The Ramsar Convention appears to have been "effective" to some extent, producing some outcomes with respect to wetland protection that would not have occurred otherwise. Even if this cooperation does not meet appropriate criteria for "deep" cooperation, it does suggest that international institutions can influence state behavior through approaches other than sanction-based enforcement. The evidence here suggests that questions of regime effectiveness need to be asked more carefully, with attention to the contexts within which regimes exist, and to the type of approach the regime adopts. Increasingly, we need to ask not merely whether regimes are effective at influencing state behavior, but under what conditions and by what means can regimes prove effective? Can either facilitative or adversarial approaches be adopted under different conditions? Can the success of one approach be generalized to other contexts? Can a facilitative approach work in arms control arenas? Can adversarial approaches work in environmental contexts? These and other questions remain to be answered.

References


Note 1: My current thinking on this subject has benefited greatly from my involvement in the Rational International Institutions Project being directed by Barbara Koremenos, Charles Lipson, Brian Portnoy, and Duncan Snidal at the University of Chicago (Koremenos, Lipson, Portnoy, and Snidal, 1997). Back.

Note 2: Given these factors, noncompliance with a coordination game based regime can reflect either an effort to shift the equilibrium or that we have misidentified the type of game involved, but it may also indicate that other, non-intentional factors are causing noncompliance. Back.

Note 3: Reverting to an "enforcement" argument has theoretical value in that it minimizes the "lack of fit" between new evidence that regimes can influence behavior and the realist paradigm that power and interests dictate behavior. By claiming that regimes only influence behavior by resorting to enforcement, realism need only relax the assumption regarding states as the only actors relevant to international interaction but need not relax the assumption regarding the factors which influence state behavior. Essentially, an enforcement argument limits the damage to the realist paradigm by interpreting new evidence as requiring revisiting "who does what" in international relations without revisiting "what they do." Back.

Note 4: Insert Papes, 1997 citation here. Back.

Note 5: Notably, tax amnesty and gun buy-back programs have the virtue of overcoming the incentives of violators to conceal their violations. Back.

Note 6: Whether this is a theoretical possibility or an empirical reality, of course, remains to be seen. Back.

Note 7: Need page numbers. Back.

Note 8: Elsewhere, I develop the notion that policy responses can be divided into six categories based on whether consequences, opportunities, or perceptions are being manipulated (Mitchell, 1996; Mitchell, 1997). Back.

Note 9: For example, Downs, Rocke, and Barsoom's recent article critiques the "managerial" school of compliance for selection bias in the cases used to demonstrate that enforcement is unnecessary to international compliance. They claim that, to the extent that a regime relies on "managerial" rather than enforcement strategies, any observed compliance can be explained by the regime's failure to require substantial behavioral change, or "deep cooperation" (Downs, Rocke and Barsoom, 1996). Back.

Note 10: Need to get citations for urging of using OSI in environmental sphere. For example, see (Schwabach, 1989) article on Rhine where on p. 475 he discusses onsite inspection stuff. Also discusses sanctions favorably, and repeats call for them in conclusion. Back.

Note 11: Consider, for example, the sarin attack in the Tokyo subway which did not raise accusations of CWC violations against Japan. Back.

Note 12: Although the CWC's approach to OSI has primarily adversarial characteristics, the regime as a whole uses considerable facilitative mechanisms to attract states to become members. Indeed, both the OPCW and the Ramsar websites have pages dedicated to delineating the advantages of membership in the convention (see http://www.opcw.nl and http://www.iucn.org/themes/ramsar). Back.

Note 13: DTIRP, Chemical Weapons - The Impact (http://www.osia.mil/osialink/dtirp/pamph01.html). Back.

Note 14: According to former CIA Director and Secretary of Defense, James Schlesinger, this treaty is a "godsend" for foreign spy networks. Back.

Note 15: The MGP was initially established as the "Monitoring Procedure" but was renamed via Resolution VI.14 of the Brisbane Conference of the Parties in 1996. Back.

Note 16: Recommendation C.4.7, Annex 1, par. 1. Back.

Note 17: In the case of the MGP for Trinidad and Tobago, the listing and visit seem to have been arranged largely by a staff member of the Wildlife Section of the Forestry Division of the Ministry of Agriculture, Land and Marine Resources (Ramsar Convention Bureau, 1996). Back.

Note 18: The section heading plays on Fred Ikle's 1961 article entitled "After Detection, What?" (Ikle, 1961). Back.

Note 19: This "internationalization" of the problem corresponds in some respects to "venue shopping" in domestic judicial systems, and to what Oran Young describes as "realigning domestic interests." Back.

Note 20: See, for example, the recommendations made in response to the MGP of the Azraq Oasis in Jordan, described in (Davis, 1994, 55-57). Back.