The Causal Mechanisms of Interaction between International Institutions

Thomas Gehring and Sebastian Oberthür 

Abstract:

The paper examines how international institutions may exert causal influence on each other’s development and effectiveness. For this purpose, four general causal mechanisms are developed that elucidate the distinct routes through which influence can travel from the source institution to the target institution and reveal the role of various actors in this process. They rely on the transfer of knowledge, the commitments entered into under the source institution, the behavioural effects generated by the source institution, and the functional linkage of the ultimate targets of governance of the institutions involved. More specific Weberian ideal types of institutional interaction help assess systematically the effects of institutional interaction for governance within the international system. The causal mechanisms and ideal types of interaction are mutually exclusive models that help analyse real-world interaction situations. They may also serve as a basis for the systematic analysis of more complex interaction situations. 

Prof. Dr. Thomas Gehring is professor of international politics at the Department of Social and Economic Sciences of the Otto-Friedrich-University Bamberg/Germany; e-mail: thomas.gehring@sowi.uni-bamberg.de

Dr. Sebastian Oberthür is academic director at the Institute for European Studies of the Vrije Universiteit of Brussels/Belgium; e-mail: sebastian.oberthuer@vub.ac.be 
 

11.344 Words including references

 

 
The Causal Mechanisms of Interaction between International Institutions

 
Introduction

The importance of interaction between international institutions is increasingly recognised by scholars of international relations. Whereas international institutions have traditionally been analysed in isolation from each other (Haas et al., 1993; Victor et al., 1998, Miles at al., 2002), they may in fact influence each other’s development and effectiveness. Due to the growing “regime density” (Young, 1996), scholars have identified a risk of “treaty congestion” (Brown Weiss, 1993: 679). So far, most contributions have examined specific interaction situations such as the problematic relationship between the World Trade Organization (WTO) and several multilateral environmental agreements with trade restrictions (Tarasofsky, 1997; Brack, 2002), overlapping jurisdictions of human rights institutions and related courts (Tistounet, 2000), or the interaction between the International Labour Organization (ILO) and the WTO (Compa and Diamond, 1996; Moorman, 2001). However, conventional wisdom has over-emphasised conflict between international institutions, whereas a more systematic assessment might reveal that synergy dominates in certain areas of international governance such as international environmental relations (Oberthür and Gehring, 2006a).

This article addresses the conceptual issue of how international institutions may exert causal influence on each other’s development and effectiveness. For this purpose, it develops a conceptual framework focusing on the exploration of patterns and causal mechanisms of institutional interaction. Conceptual knowledge about the causal mechanisms transferring influence from the source institution to the target institution has remained sharply limited to date (see the overview in Stokke, 2001a: 1-8). Existing approaches focus on an interaction situation or a “regime complex” (Victor and Raustiala, 2004) without systematically analysing the causal relationship between the institutions involved, or they mainly attempt to systematise and categorise phenomena of institutional interaction (e.g. Young, 2002; Young et al., 1999; Stokke, 2000; 2001b; Rosendal, 2001; Chambers and Kim, forthcoming).

We argue, first, that institutional interaction is best studied on the basis of cases with a unidirectional cause-effect relationship which are driven by clear-cut causal mechanisms. A case of interaction involves a single source institution from which influence originates, a single target institution thereby affected, and a single causal pathway through which influence runs from the source to the target. The identification of clear cause-effect relationships is the prerequisite for the exploration of causal mechanisms of institutional interaction. This approach builds on the established causal analysis of regime consequences (Bernauer, 1995; Underdal and Young, 2004). Accordingly, complex interaction situations should be decomposed analytically into their component bilateral cases. The fuller picture of a complex interaction situation can be regained by recombining the separate cases belonging to it.

On this basis, we develop four causal mechanisms that individual cases of institutional interaction can follow. The causal mechanisms elucidate the distinct routes through which influence travels from the source institution to the target institution and reveal the role of various actors in this process. In cases of Cognitive Interaction, members of one institution adjust their preferences to knowledge originating from another institution. In cases of Interaction through Commitment, obligations originating from one institution influence the constellation of interests and thus the decision-making process in another institution. In cases of Behavioural Interaction, an institution induces actors to change their behaviour that is directly or indirectly relevant to the implementation and performance of another institution. Impact-level Interaction will occur, if the effects of an institution on its ultimate target of governance such as the ozone layer influence the ultimate target of another institution such as the global climate.

More specific Weberian ideal types of institutional interaction help assess systematically the effects of institutional interaction for governance within the international system. Each of these types follows its own theoretically reconstructed rationale that produces particular effects on the target institution. Cases of Cognitive Interaction have rather different logics depending on whether they are intentionally triggered by the source institution or not. Cases of Interaction through Commitment may result in the transfer of commitments from one institution to another, or lead to the mobilisation of new instruments to enforce obligations, or create a demand to delimit the jurisdictions of the institutions involved. Cases of Behavioural Interaction lead to predictable outcomes for the target institution depending on whether the interaction is primarily based on different objectives, memberships or governance instruments of the institutions involved.

We hold that the identified causal mechanisms and ideal types of interaction are mutually exclusive and generalizable. They constitute theoretically derived models that help analyse real-world cases of institutional interaction, similar to the familiar game-theoretic models that help understand co-operation problems. Although we illustrate our argument primarily with cases from international environmental governance, each causal mechanism and ideal type relies on its own rationale that is independent of any specific policy area. The models are thus applicable to international institutions at large.

The conceptual framework developed here can serve as a basis for the systematic analysis of more complex interaction situations. In contrast to “integrationist” approaches (Young, forthcoming) to institutional interaction, our bottom-up approach provides the basis for the systematic reconstruction of the structures and governance implications of various forms and complexes of institutional interaction. Reconstructing such interaction complexes from individual cases of varying types promises to capture their emergent properties and fosters the analysis of the interlocking structures of international governance. It turns out that cases following different causal mechanisms and types may form causal chains and clusters of varying patterns and composition. Overall, the causal mechanisms and more specific ideal types help structure and understand the multifaceted realm of interaction between international institutions, provide systematic insights into the conditions of governance prevailing in this realm, and constitute a first step toward the development of a theory of institutional interaction.

 
Cause-Effect Relationship and Causal Mechanisms of Institutional Interaction

 

The Unit of Analysis: a Case of Institutional Interaction

Institutional interaction will exist only if one institution (the source institution) affects the development or performance of another institution (the target institution). Otherwise, we would merely observe co-evolution, i.e. the parallel, but causally unrelated, development of two or more institutions. Causation means that an effect observed within the target institution or its issue-area is attributable to another institution so that we would not expect the effect to occur in the absence of the source institution. To establish an incident of institutional interaction, we must identify (1) the source institution and more specifically its particular component(s) or decision(s) from which influence originates as the independent variable; (2) the target institution and more specifically its particular component(s) that are subject to influence originating from the source institution as the dependent variable; and (3) a cause-effect relationship between the source institution and the target institution accounting for the identified effect. There is no institutional interaction without an effect within the target institution or the issue-area governed by it, be it observable or anticipated (Gehring and Oberthür, 2004).

To allow serious causal analysis, interaction situations must be decomposed into their component cases of interaction. Decomposition will be especially required in three types of situations. First, two institutions may be involved in numerous cases of interaction at the same time. For example, the Montreal Protocol for the protection of the ozone layer indirectly promotes the use of certain greenhouse gases (hydrofluorocarbons, HFCs) regulated under the Kyoto Protocol to the UN Framework Convention on Climate Change, and its non-compliance procedure provided a precedent for the elaboration of a similar component within the climate change regime (Oberthür, 2001). Second, an interaction situation may involve a whole complex of regimes. For example, the Baltic Sea is affected by several global environmental regimes addressing, inter alia, oil pollution from ships and dumping of wastes at sea, an important regional regime (HELCOM), and overall arrangements such as the United Nations Convention on the Law of the Sea (UNCLOS) (Young, 1996). It is likely that these institutions interact with each other in various forms. Third, two or more institutions may “co-evolve” over time so that neither of them would exist in its current state in the absence of the other. If influence runs back and forth between the institutions, analytical decomposition of the co-evolution process into sequential cases helps examine clear cause-effect relationships (Archer, 1985; Carlsnaes, 1992). For example, the co-evolution of the global Basel Convention on the transboundary movement of hazardous wastes and several related regional regimes may be analysed as a sequence of phases. After the Basel Convention triggered establishment of several regional regimes, the regional regimes with their tighter regulatory approaches influenced the development of the Basel Convention (Meinke, 2002: 249-349).

The consequences of a case of institutional interaction may be beneficial, adverse, or neutral for the target institution. The main effects of institutional interaction occur in the target institution and can be assessed against this institution’s prime objective. While established research on the effectiveness of international institutions has also used an institution’s prime objective as the major yardstick for assessing its consequences (e.g. Young, 1999; Haas et al., 1993), the causes and effects of institutional interaction are located in the domains of different institutions so that the prime objective of the target institution is used as the relevant yardstick. If the effects of institutional interaction support the objectives of the target institution, they create synergy between the two institutions involved. If they contradict the target’s objective, they result in disruption. The effects of an interaction may also be indeterminate or neutral, if they do not clearly hamper or reinforce the target institution’s pursuit of its objective.  

The Search for Causal Mechanisms of Institutional Interaction

The identification of causal mechanisms of institutional interaction is directed at elucidating how causal influence is transferred from the source institution to the target institution. A causal mechanism is a set of statements that are logically connected and provide a plausible account of how a given cause creates an observed effect (Schelling, 1998). It opens the black box of the cause-effect relationship between the institutions involved (Elster, 1989: 3-10). The analytic concept of causal mechanisms is well established in the social sciences (Coleman, 1990: 1-23; Hedström and Swedberg, 1998: 21-23).

In its basic form, a causal mechanism of institutional interaction comprises three separate stages, as illustrated in Figure 1. First, the source institution affects the preferences or behaviour of relevant actors within its own domain. Influence originates from the structural level and is directed at the actors' level. Second, this effect leads to a change of preferences or of individual behaviour of actors relevant to the target institution. Third, individual action produces the effect observed within the target institution or its issue-area. Consider the following example. The Kyoto Protocol provides an incentive for carbon sequestration in forests. If states and private actors plant fast-growing trees in response, they might encroach upon biodiversity, which is protected under the Convention on Biological Diversity (CBD) (Pontecorvo, 1999; Jacquemont and Caparrós, 2002). To establish the causal relationship between the two regimes, one would have to demonstrate (1) that relevant actors react to the adoption of the Kyoto Protocol by increasing carbon sequestration in forests; (2) that these behavioural changes encroach upon biodiversity-rich habitats such as tropical rain forests; and (3) that the Convention on Biological Diversity is affected by this change because its performance is undermined. Effects observed within the target institution are not necessarily attributable to changes in the behaviour of states; other types of actors, like non-governmental organisations, industry, or the secretariats of international institutions, may play an important role (see also Selin and VanDeveer, 2003). Moreover, interaction may already occur, and be responded to politically, if actors are only expected to engage in increased carbon sequestration, and if this action is only possibly affecting the target institution.  

Figure 1: Causal Mechanism of Institutional Interaction 
 
 

Causal mechanisms provide a micro-foundation for the analysis of institutional interaction by revealing how actors matter in the process (George and Barnett, 2005: 135-145). Like any other theory in the social sciences, a concept of institutional interaction requires a reliable micro-macro link (Buzan et al., 1993: 104; Alexander and Giesen, 1987), which has been discussed in International Relations in the form of the agent-structure problem (Wendt, 1987; Carlsnaes, 1992). In cases of institutional interaction, both the independent and the dependent variables (the source institution and the target institution) are located at the macro-level. However, an international institution will rarely influence another institution directly without intermediate adaptation by relevant actors. The exploration of causal mechanisms of institutional interaction constitutes a cornerstone for the development of a theory of institutional interaction.

The identification of the causal mechanisms of institutional interaction has been informed by the three levels of effectiveness of governance institutions that are well-established in the literature on international regimes (Underdal, 2004: 34). To be effective, a governance institution must produce an appropriate output in the form of collective knowledge or norms prescribing, proscribing or permitting behaviour. The output may generate some form of behavioural change of relevant actors, the outcome. Finally, changes of behaviour might have an impact on the ultimate target of governance.

Institutional interaction may be driven by four mutually exclusive causal mechanisms, each of which is based upon a clearly distinguishable rationale and will be discussed in detail in the next sections. Asking how, and under which circumstances, an institution can be influenced by another institution, we identify two causal mechanisms located at the output level that exert influence on the decision-making process of the target institution. A third causal mechanism is located at the outcome level and involves changes of behaviour of relevant actors operating within the issue-areas governed by the institutions involved. Finally, one causal mechanism is located at the impact level. The latter two mechanisms do not modify decision-making of the target institution, but its effectiveness within its issue area.

Since the causal mechanisms are so general that cases following one of them may reveal significantly different characteristics, we identify, in the second step, more specific ideal types of institutional interaction. These Weberian ideal types are also mutually exclusive and follow their own rationales. They specify how influence is transferred from the source institution to the target institution and elucidate the consequences of particular settings of institutional interaction for international governance. Thus, they draw attention to systematic implications of institutional interaction for policy-making.

 
The Causal Mechanism of Cognitive Interaction

The Rationale of Cognitive Interaction

The decision-making process of an international institution may be influenced, if information, knowledge or ideas (Haas, 1992; Risse-Kappen, 1994; Yee, 1996) produced within the source institution modify the perception of decision-makers operating within the target institution. For such Cognitive Interaction to occur, the source institution generates some new information such as a report revealing new scientific or technological insights or an institutional arrangement solving a particular regulatory problem. Then, some actor (e.g. a member state or a non-governmental organisation) feeds the information into the decision-making process of the target institution. Subsequently, this information changes the order of preferences of actors relevant to the target institution. Finally, this modification of actors’ preferences has an impact on the collective negotiation process and the output of the target institution. For example, the members of the target institution may discover, and decide to adopt, an institutional innovation introduced within the source institution, such as a non-compliance procedure or a particular arrangement for providing assistance to developing countries.

Cognitive Interaction is based on the premise that the rationality of actors is “bounded” because relevant information is not entirely available or because information processing capacity is limited (Keohane, 1984: 100-115; Simon, 1972). Under these circumstances, actors will be prepared to adapt their preferences to new information (Checkel, 1998; Risse, 2000). If they were fully aware of their interests in a given situation, as is frequently assumed in rational choice co-operation and regime theory (see Martin, 1993; Hasenclever et al., 1997), Cognitive Interaction could not be expected to occur.

Cognitive Interaction is purely based on persuasion and may be conceived of as a particular form of inter-institutional learning (similarly Stokke, 2001a: 10). It can occur between any two institutions whether or not their memberships or their issue-areas overlap. Moreover, it is the least intrusive of all causal mechanisms because learning cannot be imposed on the target institution. Consequently, the source institution does not exert any pressure on the decision-makers of the target institution. If relevant actors adapt their preferences, the decision-making process of the institution can nevertheless be significantly affected and the effects will be felt even by those members of the target institution that have not been convinced.

Cognitive Interaction occurs in two forms. If it is intentionally triggered by the source institution, members of the target institution use an institutional arrangement or policy idea of the source institution as a model. If the source institution intentionally initiate a learning process, it deliberately draws the attention of the members of the target institution to the effects of their decisions on the source institution. 

Policy Model

If Cognitive Interaction is unintentionally triggered, members of the target institution voluntarily use some aspect of the source institution as a Policy Model. For example, the compliance system under the Montreal Protocol for the protection of the ozone layer influenced the negotiations on the compliance system under the Kyoto Protocol on climate change because it provided a model of how to supervise implementation and deal with cases of possible non-compliance (Oberthür and Ott, 1999: 215-222). The members of the Montreal Protocol did not establish the model in order to influence the Kyoto Protocol. They did also not have the ability to impose their model on the target. Instead, the Montreal Protocol presented an institutional arrangement which the members of the Kyoto Protocol conceived of as a useful source of inspiration.

The Policy-Model type of Cognitive Interaction is framed by actors in the target institution. It is difficult to foresee which kind of information or decision originating from the source institution might prompt interaction of this kind, because any institutional arrangement, decision, or scientific or technological information from any other institution might serve as a Policy Model. Also, numerous types of actors may pick up the information or idea and feed it into the decision-making process of the target institution, including one or more member states, an interested non-governmental organisation, the secretariat of the target institution, or even relevant individuals. These actors do not even have to participate in the source institution, because information and ideas can also be obtained from the outside by surveying the field, reading reports, or examining institutional arrangements.

The Policy-Model type of interaction highlights that members of an institution can enhance the effectiveness of their governance efforts by learning from occurrences taking place within other institutions. Any institution may learn in this way from any other institution that displays appropriate precedents and deals with relevant problems. Accordingly, actors operating in the framework of international institutions frequently look for precedents. Learning from a Policy Model can generally be expected to strengthen the effectiveness of the target institution, because it requires that target members collectively consider the policy model to be useful. Frequently, a model is modified and adapted so as to fit the particular needs of the target (”complex learning;” see Haas, 1990). Thus, the negotiators of the Kyoto Protocol were inspired by the compliance system of the Montreal Protocol, but they adapted and strengthened the arrangement according to their particular needs (Werksman, 2005). In rare cases, actors may learn how to hamper rather than enhance effectiveness. From the Montreal Protocol, negotiators of the Kyoto Protocol learnt that a powerful institutional arrangement for feeding scientific and technical knowledge into policy-making has the potential to significantly advance the decision process. Therefore, parties opposing such advances blocked a similar arrangement under the Kyoto Protocol (Oberthür, 2001: 360-1).  

Request for Assistance

Cognitive Interaction may also take the form of an intentional request by the source institution for assistance of the target institution. In conformity with the core characteristic of Cognitive Interaction, the source institution cannot impose any change upon the target institution. However, it can draw the attention of actors within the target institution to a particular aspect of the outside world which they had so far not taken due account of – at least seen from the perspective of the source institution. For example, more specific wildlife treaties have repeatedly requested the Convention on International Trade in Endangered Species of Wild Fauna and Flora (CITES) for assistance. In response, CITES has enacted related trade restrictions or called upon its global membership to protect certain endangered species. In turn, the World Customs Organization (WCO) adapted its customs codes in response to a request by CITES, thus supporting the implementation and enforcement of the latter’s trade restrictions (Lanchbery, 2006).

In cases of Request for Assistance, the source institution largely frames the learning process. In contrast to learning from a Policy Model, it intentionally adopts a decision without which interaction of this type would not occur. The secretariat of the source institution will usually formally transfer a Request for Assistance to the secretariat of the target institution that will officially feed it into the decision-making process of the latter. Since the source institution cannot force the target institution to adapt, some learning resulting in an adaptation of preferences is required to produce a reaction within the target institution. Actors may learn that an adaptation of their institution could either mitigate undesirable side-effects on, or further strengthen the effectiveness of, the source institution.

Whereas a Request for Assistance will generally produce synergistic or at least neutral effects for the target institution, it is intended to create a positive feedback effect on the source institution. While it is hardly conceivable that the members of the target institution adopt measures that harm their own institution, they may find it difficult to reject Requests for Assistance with indifferent effects. Thus, the World Customs Organization responded favourably to the request of CITES to adapt its customs codes although it hardly benefited from this adaptation. However, Requests for Assistance are always intended to reinforce the effectiveness of the original source institution by triggering a reverse case of interaction. Hence, CITES requested the World Customs Organization to adapt, because it expected this change to facilitate the enforcement of its own policy. In this second case, the direction of influence is reversed, and influence is transmitted through a different causal mechanism, namely Behavioural Interaction.

Requests for Assistance provide an instrument for furthering effective international governance. They enable an institution to draw on other institutions in order to enhance its own effectiveness, even if it cannot force the target institution to adapt its rules. To this end, members of international institutions might actively search for opportunities to instigate learning processes within other institutions. Since the success of a Request for Assistance largely depends on the positive reaction by the target institution, the endeavour will be particularly promising if the requested adaptation supports the objectives of the target institution or is at least easily compatible with them. As in the case of the interaction between CITES and the World Customs Organization and INTERPOL, Requests for Assistance may be the starting point for the establishment of a more regular coordination between the institutions involved (Lanchbery, 2006)

 
The Causal Mechanism of Interaction through Commitment

The Rationale of Interaction through Commitment

The decision-making process of an international institution will also be influenced, if the commitments entered into by the members of the source institution modify the preferences of these actors regarding issues related to the target institution. For Interaction through Commitment to occur, the preferences that are modified as a result of the obligations under the source institution must influence the collective decision-making process of the target institution and its output. For example, the WTO commitment not to discriminate imported goods according to the methods by which they have been produced in their countries of origin renders it difficult for WTO members to adopt trade sanctions within international environmental regimes or within the International Labour Organization (ILO) to reinforce the effectiveness of these institutions (Brack, 2002). Whereas Interaction through Commitment ideally presupposes that actors are bound to the obligations of the source institution when participating in decision-making within the target institution, anticipated commitments may also trigger the mechanism, if they induce an actor to adapt its preferences in a concurrent decision-making process.

Interaction through Commitment is based on the premise that international obligations create at least some binding force for their addressees. According to the constructivist “logic of appropriateness” (March and Olson, 1998), actors may generally be assumed to adhere to valid norms, because such behaviour is legitimate and reduces the costs of instrumental decision-making. According to the rationalist “logic of consequences”, behaviour in conformity with valid norms supports desired co-operative projects. Actors might also endeavour to preserve a reputation of keeping their promises because possible future co-operators would otherwise be less inclined to enter into agreements with them (Keohane, 1984: 105-106). In addition, commitments may be costly, so that actors gain an interest in subjecting competitors to similar obligations. In all these cases, the commitment entered into within the source institution affects the preferences of actors relevant to the target institution.

Interaction through Commitment requires some overlap of the memberships and issue-areas of the interacting institutions, but also significant differences between them. Without overlapping memberships, the target institution would remain unaffected because none of its members would be subject to relevant commitments under the source institution. Without overlapping issue-areas, commitments could not modify the preferences of states regarding issues related to the target institution. However, the institutions must also differ in some important dimension to create momentum for interaction because two institutions with identical properties could not be expected to exert significant influence on each other. The three ideal types of Interaction through Commitment developed in the following vary with respect to one of three dimensions: the objectives of the institutions involved, their memberships or their instruments of governance. The underlying rationales of these three types of interaction differ profoundly because inter-institutional influence relies on exactly one distinct key difference for each ideal type.  

Jurisdictional Delimitation

Jurisdictional Delimitation cases arise from distinct, frequently incompatible, priorities of the institutions involved. The rationale of the ideal type presupposes that the same group of member states addresses the same issues employing identical governance instruments within two institutions with different objectives. International institutions are throughout established to improve the status quo within a given issue-area in a particular direction (Gehring, 1994: 433-49). Accordingly, every institution disposes of its own criteria to judge policy measures and the related behaviour of states and non-state actors. As a result, different institutions may appraise a policy measure differently. Consider that the WTO and its predecessor GATT aim to abolish trade obstacles and liberalise international trade, whereas other institutions with a very similar membership elaborate and maintain regulatory standards for the protection of the environment. Accordingly, policy measures such as environmentally motivated trade restrictions will be considered either as effective means for enforcing international environmental regimes, or as undesired interference with international trade.

In Jurisdictional Delimitation cases, the members of the institutions involved are in a “mixed motive” situation resembling the game-theoretic constellation of the Battle of the Sexes (Stein, 1982; Keohane, 1984). On the one hand, they possess a general interest in some sort of separation of jurisdictions in order to avoid fruitless regulatory competition and a reduced effectiveness of their respective institutions. Neither side will be served if the institutions involved mutually undermine their effectiveness due to incompatible obligations which may be interpreted by addressees at will. On the other hand, the constituencies of the institutions have conflicting preferences regarding the appropriate solution that make it notoriously difficult to solve such problems. In the conflict between trade and environment, actors prioritising liberal international trade will advocate regulation by the WTO, while countries struggling for protection of the environment may prefer jurisdiction of the environmental regimes involved.

Commitments of the earlier institution will almost automatically limit the room for manoeuvre within the later institution by strengthening the actors preferring the earlier institution’s objectives. Like the game-theoretic Battle of the Sexes situation, the ideal type of Jurisdictional Delimitation elucidates when and how overlapping jurisdictions create a demand for their separation, but not how the balance is eventually struck in a particular case. However, it is well-known that an equilibrium found in Battle of the Sexes situations will be fairly stable, because neither side can expect to gain from resumption of conflict. This implies that the earlier institution will possess a “first-mover advantage” (Héritier, 1996; Mattli, 2003). In the trade and environment debate, the older GATT/WTO has been viewed as “chilling” the negotiations on environmental regimes because negotiators have frequently shied away from even discussing measures that might be in conflict with GATT/WTO rules (Palmer et al., 2006: 186). In this case, the first-mover advantage of the WTO delimits the range of options available to negotiators within the environmental regimes. 

If an international institution challenges an established distribution of jurisdictions, Jurisdictional Delimitation issues will lead to open conflict. An institution’s jurisdiction can be challenged by agreeing within another institution on incompatible commitments. Creation of such “strategic inconsistency” (Victor and Raustiala, 2004: 301) will be particularly relevant, if new regulatory objectives such as environmental protection shall be promoted in a field already governed by an existing institution. Thus, international environmental regimes established since the 1970s have almost automatically encroached upon the jurisdiction of the international trade regime established since the late 1940s, whenever they have restricted trade for environmental purposes. Hence, environmental regimes have gradually pushed back the jurisdiction of the WTO and partially reversed the latter’s “chill effect”.

Jurisdictional Delimitation cases pose the governance challenge to arrive at a delimitation of jurisdictions that balances the diverging interests and realises the common interest. Conflict management will tend to identify those measures which honour the basic objectives of each institution, while being least detrimental for the objectives and operation of the other institution. This does not necessarily require an overarching institutional structure. Balancing of competing objectives such as free trade and environmental protection can occur within either of the institutions involved. As observable in negotiations on environmentally motivated trade restrictions (Oberthür and Gehring, 2006c), actors that are members of both institutions will play a major role, because they will usually have the strongest interest in making contradictory obligations compatible.  

Nested Institutions

Interaction between Nested Institutions is driven by the different memberships of two international institutions that pursue identical objectives and employ the same governance instruments. If the membership of one institution forms part of the membership of another institution, two formally independent institutions with similar objectives and regulatory means are “nested” into each other (Aggarwal, 1983; Young, 1996). For example, the global Basel Convention of 1989 and a number of regional agreements, most importantly the African Bamako Convention of 1991, regulate transboundary movement of hazardous waste (Clapp, 1994). It is well-known from negotiation analysis that institutions with divergent memberships may arrive at differing obligations, even if addressing identical problems (Sebenius, 1983; 1992).

Interaction between Nested Institutions rests on the following components. First, it is typically easier to reach agreement within a smaller (e.g. regional) than in a larger (e.g. global) institution because a higher number of participants usually implies a greater heterogeneity of interests (Snidal, 1994). Hence, African countries could agree to ban completely the transboundary movement of hazardous wastes to Africa under the Bamako Convention, while similar agreement proved initially impossible under global Basel Convention due to fierce resistance by industrialised countries. In some cases, greater homogeneity of interests may also follow from factors other than the size of the institution, such as more equal economic development. Consequently, a similar momentum might also develop between institutions of a similar size but with partially overlapping memberships. Second, commitments agreed within the source institution streamline the preferences of its members, where heterogeneous preferences may have prevailed before. As a result, the members develop a common interest in expanding obligations to other countries – be it to commit competitors to costly obligations, to preclude free riding, or to reinforce the effectiveness of their agreement. Third, based upon their common commitment, members of the (smaller) source institution form a natural coalition during negotiations within the (larger) target institution. It will be more difficult to persuade a whole group of states than a single country to change its pre-agreed negotiation position or to offer sufficient side-payments. Hence, the probability increases that the co-ordinated position of the coalition constitutes some “focal point” (Schelling, 1960: 100) around which expectations for an agreement converge. Indeed, the formation of a firm coalition of African countries and other developing countries advocating a total ban of waste exports to the South made subsequent agreement on the ban in the global context possible (Meinke, 2002: 271-331). 

Interaction between Nested Institutions constitutes a mechanism for policy diffusion within the same policy field and provides opportunities for forum shopping. Actors striving for regulation of a particular issue may choose whether to promote their proposals predominantly in the smaller or in the larger institution. The rationale of interaction between Nested Institutions suggests that effects will largely support the effectiveness of the target institution, and occasionally also of the source institution, while disruption is unlikely to occur. The identical objectives of the institutions generate compatible priorities and render disruptive effects on target institutions highly improbable, if not impossible. Evidently, policy diffusion will not always occur, because adverse constellations of interests prevailing within the larger institution may effectively preclude adoption of a policy measure agreed upon in a smaller setting. 

Additional Means

Interaction through Commitment may also take place between two institutions with identical memberships that pursue the same objective, if the governance instruments (means) available within these institutions differ significantly. Frequently, international institutions do not control the full spectrum of possible governance instruments. While some institutions rely almost exclusively on “soft” law, others resort to legally binding (“hard”) international law (Abbott and Snidal, 2000). The European Union even controls supranational law that is, under certain conditions, directly applicable within the legal systems of the member states and supported by the particularly stringent supervisory and judiciary mechanism of the EU (Alter, 2000). Some institutions dispose of financial assistance mechanisms (Keohane et al., 1996) or non-compliance procedures (Victor et al., 1998), while others do not.

Interaction activating an Additional Means occurs in two stages. First, actors committed to an obligation within the source institution transfer the commitment to the target institution. This will be comparatively easy because of their previous agreement on the same obligation within the source institution. However, such simple diffusion of an obligation alone cannot be assumed to change the situation significantly for actors that are members of both institutions. In the second stage, incorporation of the transferred obligation must therefore mobilise an additional governance instrument, such as a particular form of law or a specific enforcement or assistance mechanism, that provides an additional incentive to implement the obligation. The interaction between the regime for the protection of the North-East Atlantic (OSPAR) and the International North Sea Conferences established in the 1980s may serve as an illustration. While relying on declarations that were formally non-binding, the Conferences took place at a high political level and generated political pressure. Having politically agreed on the phase-out of certain substances and activities, it became difficult for the same countries to resist the adoption of substantively identical obligations within the framework of OSPAR. Since OSPAR relies on hard international law, the originally soft obligations were transferred into legally binding hard law (Skjærseth, 2006).

Interaction of the Additional-Means type will regularly raise the effectiveness of both institutions involved. If the diffusion of an obligation activates an additional governance instrument, it will support the effectiveness of the target institution. Since the new obligation is, by definition, in line with the target institution’s own objective, implementing the obligation will support this objective by activating the target’s governance instrument in addition to the instruments of the source institution. At the same time, activating an Additional Means automatically contributes to a more effective implementation of the source institution. In the just-mentioned example, the more effective implementation of hard-law OSPAR obligations to protect the Northeast Atlantic automatically helped achieve the goals of the North Sea Conferences. The adoption of the obligation in the target institution can thus be expected to trigger a positive feedback effect on the source institution that follows a different causal mechanism, namely Behavioural Interaction.

Because of its synergistic effects on the source and on the target institution, interaction activating Additional Means allows actors operating within the source institution to enhance the effectiveness of international governance. As in the case of Nested Institutions, the option to trigger interaction provides opportunities for forum shopping because actors can choose in which of the institutions available to launch a particular regulatory initiative. As the interaction between the North Sea Conferences and OSPAR illustrates, actors may even establish an institution in an area already governed by another institution for the sole purpose of activating an additional governance instrument.

 
The Causal Mechanism of Behavioural Interaction

Institutional interaction will influence the performance of an international institution within its own domain, if behavioural changes triggered by the source institution become relevant for the implementation of the target institution. Interaction following this causal mechanism is located at the outcome level. All international governance institutions are designed to influence the behaviour of relevant actors in order to achieve their objectives such as protecting the environment or liberalising international trade (Levy et al., 1995; Young, 1992). In some cases, behavioural changes occurring within the domain of one institution exert influence on the domain of another institution. If the Kyoto Protocol creates incentives to plant fast-growing trees in ways that encroach upon biodiversity, it undermines the performance of the Convention on Biological Diversity. The rules of the target institution may remain totally unaffected, and if they were modified, this adaptation occurred as an additional response to the original effect. 

Behavioural Interaction comprises the following steps. First, the source institution must produce an output such as a set of prescriptions or proscriptions. Then, relevant states or non-state actors adapt their behaviour significantly in response to the output. Relevant behavioural changes include unforeseen side-effects and deviating behaviour, such as increased smuggling in response to trade restrictions adopted to protect endangered species. Subsequently, the behavioural changes within the source institution’s issue-area have an impact on the behaviour related to the implementation of the target institution, either because they are relevant for both issue-areas, or because they prompt further behavioural changes in the issue-area governed by the target institution. Finally, this behavioural effect exerts influence on the performance and effectiveness of the target institution.

Behavioural Interaction does not depend on a collective decision within the target institution, because it occurs as the aggregate result of the behaviour of actors operating within the two issue-areas involved. Behavioural Interaction is thus, unlike Cognitive Interaction and Interaction through Commitment, almost entirely controlled by the source institution. It might even come about unnoticed by the members of the institutions involved.

We can identify three ideal types of Behavioural Interaction. Whereas the different types of Cognitive Interaction and Interaction through Commitment differ as to the particular activities needed within the source institution and the target institution, the types of Behavioural Interaction are characterized by specific differences between source and target that support particular hypotheses as to the effects of interaction in the issue area of the target institution. In general, Behavioural Interaction, like Interaction through Commitment, requires that the two institutions involved are close enough to matter for each other and at the same time distinct enough to create momentum for interaction. The three types developed in the following vary as to whether the institutions involved differ in respect of their objectives, or their memberships or their governance instruments.

Behavioural Interaction characterized by different objectives of the institutions involved will tend to result in disruption of the target institution. In this case, the same group of actors ideally addresses the same issue employing identical means of governance within two institutions that pursue different objectives. As a result, behavioural changes of states and non-state actors triggered by the source institution may easily be at odds with the objectives of the target institution and may thus undermine the latter’s performance. It is difficult to think of a situation in which they could be systematically synergistic because they neither expand commitments to other actors nor activate other means of governance. In the just-mentioned example, behaviour responding to incentives created under the Kyoto Protocol to plant fast-growing trees as sinks for carbon dioxide serves to protect the global climate, while possibly undermining the objective of preserving global biodiversity. Behavioural Interaction driven by different objectives is closely related to Jurisdictional Delimitation cases located at the output level. Indeed, cases of both types appear frequently together in a complex interaction situation. For example, the conflict between the WTO and international environmental institutions with trade restrictions is not limited to the rules made at the output-level, it extends to the implementation of these rules by addressees at the outcome level. Whenever a member state implements a trade restriction originating from an environmental regime, it will implicitly undermine the effectiveness of the WTO.

Behavioural Interaction relying upon different memberships of the institutions involved will always create synergy. In this case, different groups of actors, ideally, address the same issue employing identical means of governance within two institutions that pursue like objectives. Due to the matching objectives, behavioural changes triggered by the source institution will automatically be in conformity with the policy direction of the target institution. Thus, this type of Behavioural Interaction constitutes the corollary of interaction between Nested Institutions at the output level. Compared with the latter, however, the direction of influence is reversed: Significant interaction effects can only be expected to originate from the larger (global) institution affecting the smaller (regional) institution, because extra behavioural effects result from the larger membership of the global institution. Again, cases located at the output level may occur together with cases located at the outcome level, occasionally even in the form of causal chains. Whereas the regional Bamako Convention on the transboundary movement of hazardous wastes influenced the normative development of the global Basel Convention, the latter subsequently helped implement the Bamako Convention, because it restricted the exports of such wastes from industrialised countries to non-OECD countries.

Behavioural Interaction driven by different means of governance available to the institutions involved will also virtually always create synergy. In this case, the same group of actors ideally addresses the same issue within two institutions that pursue the same objectives, but dispose of different means of governance. Due to matching objectives, behavioural changes triggered by the source institution will once again almost automatically serve the ends of the target institution. Thus, this type of Behavioural Interaction constitutes the corollary to the Additional Means type located at the output level. As already mentioned, interaction activating an Additional Means of governance, such as harder law or a sanctioning instrument unavailable within the source institution, will frequently trigger a reverse case of Behavioural Interaction in which the original target institution helps implement the original source institution. Requests for Assistance  frequently trigger a similar causal chain. For example, the action taken by World Customs Organisation (WCO) in response to the Request for Assistance from CITES set off a feedback case of Behavioural Interaction through which the WCO supported the implementation of CITES.

 
The Causal Mechanism of Impact-Level Interaction

Finally, interaction may be located at the impact level. In this case, the ultimate governance target of one institution, such as international trade or the ozone layer, is directly influenced by side-effects resulting from the effect of another institution on its ultimate target of governance. A stylised example that we owe to Arild Underdal may illustrate this least intuitive causal mechanism. Consider that protection of the stocks of cod and herring are the ultimate targets of two separate international institutions. As cod eats herring, successful protection of cod, resulting in a growing population of this species, will unintentionally decrease the population of herring. In this case, the two institutions involved are not linked at the level of output, because neither the norms nor knowledge produced within the cod regime influence the norms protecting herring, nor through behavioural changes, because decreased fishing of cod does not directly influence the fishing activities related to herring. They are “functionally linked” (Young, 2002: 23; 83-109) at the impact level because the effect of the source institution on its ultimate regulatory target (increased population of cod) directly influences the ultimate regulatory target of the target institution (decreasing population of herring).

Whereas Impact-Level Interaction involves human action, inter-institutional influence is not channelled through human behaviour. The causal mechanism of Impact-level Interaction involves the following pathway. First, the source institution produces an output. Second, states and non-state actors operating within this institution’s issue-area adapt their behaviour in response to this signal. Subsequently, these behavioural changes affect the ultimate target of governance of the source institution. Finally, this impact exerts influence on the ultimate governance target, and thus on the effectiveness, of the target institution. is the most intrusive of all causal mechanisms. Like Behavioural Interaction, it is dominated by the source institution, and it does not depend on any action within the target institution or its domain. Instead, it depends on the non-societal, functional link between the ultimate governance targets of the two institutions involved.

 
Mutual Exclusiveness and Generalizability of Causal Mechanisms and Types

The general causal mechanisms presented in the preceding sections constitute mutually exclusive models. Being deductively derived, such models must be theoretically coherent, but they cannot be empirically right or wrong (Snidal, 1985). They are not intended to provide precise descriptions of all properties of relevant interaction cases, but to focus on the relevant components of the different causal pathways that a case of interaction may follow. Hence, they may, or may not, fit a given case of interaction. Mixed cases may exist but are rather unlikely to occur because the mechanisms are located at three different levels of effectiveness and governance, and they involve different actors. While related cases may occur concurrently at different levels, a case of interaction cannot be driven by Behavioural Interaction and Interaction through Commitment or Impact-level Interaction at the same time. However, in rare cases it might be difficult to distinguish whether members of the target institution learn from the source institution (Cognitive Interaction) or adapt their preferences according to the commitments entered into under the source institution (Interaction through Commitment). The occurrence of such mixed cases following more than one causal mechanism does not undermine the mutual exclusiveness, nor the analytical value, of the theoretically derived causal mechanisms.

The four causal mechanisms are likely to cover the full range of relevant causal mechanisms. The decision-making process of a target institution may hardly be systematically influenced other than by Cognitive Interaction and Interaction through Commitment. Other forms of institutional output, such as financial transfers or the threat of sanctions, will hardly influence the decision-making process of another institution without first generating behavioural effects within the issue-area of the source institution. It is also highly unlikely that the decision-making process of an international institution is directly affected by changes of behaviour occurring within the domain of another institution without observable, or at least anticipated, effects on behaviour within the target institution’s domain. Likewise, interaction influencing the behavioural performance of the target institution will always originate from the behavioural effects of the source institution. It seems to be a rather remote possibility that, for example, the international ozone regime would pass international trade rules that are unrelated to the protection of the ozone layer. Nothing of that kind has been reported in the literature so far. It is also difficult to see how effects at the impact level within one institution might directly affect the behaviour of actors governed by another institution or that institution’s decision-making process without first affecting the ultimate target of governance of the target institution.

The ideal types of institutional interaction also rely on their own, mutually exclusive rationales. They are theoretically constructed and elucidate a particular set of important characteristics of cases that are related to their underlying rationales. Like the general causal mechanisms, they have explanatory power (Weber, 1904: 190-212). Since the ideal types of interaction presented above have been developed against the backdrop of a sample of somewhat more than 150 of cases (Gehring and Oberthür, 2006), we cannot exclude that more ideal types with yet other rationales exist. Moreover, in the real world, mixed cases involving two or more ideal types may occur which are difficult to classify. Two interacting institutions may, for example, differ simultaneously in regard to their memberships and their means of governance. However, the occurrence of mixed cases does not conceptually devaluate the usefulness and mutual exclusiveness of the ideal types developed, because the types are designed to reduce complexity and point to particularly relevant aspects, not to exhaust the possible variance, of real-world cases. 

Both the causal mechanisms and the ideal types presented in this article are generally applicable to the study of institutional interaction in international governance. While they have been developed against the backdrop of a sample of cases from environmental governance, and most of our illustrative examples referred to are therefore from that governance area, they reflect general rationales derived from theories of international cooperation and international relations. They can be employed to analyse cases of institutional interaction from other samples and policy fields.

 
Towards the Analysis of More Complex Interaction Situations

The approach developed in this article can be employed to systematically explore more complex interaction settings and their emergent properties. Eventually, we want to examine interaction complexes composed of several cases of interaction. These cases may influence each other in unexpected ways so that new properties of the overall situation emerge which are not inherent in its individual components. The analysis of individual interaction cases does not reveal these properties of the overall situation, much like an examination of individual trees and plants cannot grasp the emergent properties of a forest. However, the properties of the forest emerge from the particular forms of co-existence of the trees and plants, including their mutual influence on each other's existence and development. Hence, the systematic recombination of cases promises to allow for the exploration of the typical patterns and emergent effects of more complex interaction situations while retaining the analytical advantages of the decomposing approach. Causal chains and clusters are two principal patterns of more complex interaction situations.

Interaction cases may be related sequentially so that they form causal chains. In such constellations, one case of interaction gives rise to a subsequent case that feeds back on the original source institution or influences a third institution. As a result, the interaction process acquires a momentum of its own. The occurrence of causal chains is closely related to the rationale of several of the ideal types of interaction. A Request for Assistance is issued in order to generate a subsequent case of Behavioural Interaction feeding back to the original source institution. An Additional Means case will usually be followed by a positive feedback effect in the form of Behavioural Interaction. Anticipated Behavioural Interaction with disruptive effects can easily lead to attempts on the side of the original target institution to create a Jurisdictional Delimitation case. Likewise, the original target institution of a case of Jurisdictional Delimitation may respond with its own jurisdictional activity, thus influencing the original source institution and leading to the gradual separation of jurisdictions. For example, the WTO, especially through its Agreement on Sanitary and Phytosanitary Measures (SPS Agreement), constrained the use of trade measures related to genetically modified organisms under the Cartagena Protocol on Biosafety. Negotiations on the Cartagena Protocol, in turn, precluded the development of further rules on the appropriate risk assessment for genetically modified organisms within the WTO (Oberthür and Gehring, 2006c).

Cases of interaction may also “cluster” around certain issues and institutions. While causal chains address causation between sequential cases of interaction, clusters address settings of parallel cases of interaction without requiring causation between cases. In this case, several institutions concurrently address a particular issue in complementary or competitive ways. Accordingly, we may expect to find competitive interaction clusters (e.g. interaction between the WTO and several environmental regimes) as well as co-operative clusters in which particular forms of inter-institutional division of labour and coordination develop. For example, CITES has gradually become the centre of a cluster of institutions in the area of wildlife protection, thus reflecting the evolution of a more centralised arrangement for inter-institutional management (Lanchbery, 2006). The International North Sea Conferences have even been established not least to provide for a co-ordinating mechanism, and to set priorities, for future activities in various institutions relevant for the protection of the North Sea (Skjærseth, 2006).

Causal chains and clusters are two typical patterns of interlocking structures of international governance institutions. They draw attention to the fact that institutional interaction generates broader patterns and structures of emergent informal rules and norms. While these interlocking structures (Underdal and Young, 2004: 374-75) reflect the normative expectations of international society, they also define the division of labour among the institutions involved in a given problem area of international relations. Interlocking governance structures evolve in ways that are distinct from the evolution of the sector-specific institutions of which they are composed (Victor and Raustiala, 2004: 279). So far, they have hardly been “rationally designed” (Koremenos et al., 2001) - not least because of the lack of an overarching institution capable of managing interaction within the international system. Shedding light on the causal mechanisms by which efforts to enhance the effectiveness of international governance in different issue areas may affect each other, the conceptual framework presented in this article helps understand the origins of emergent interlocking governance structures and provides a basis for their further systematic study.

 
Conclusion

The conceptual framework of institutional interaction developed in this article elucidates how an international institution can influence the normative development and the performance of another international institution. While other approaches seek to describe and classify complex interaction situations (Young, 1996; 2002), it focuses on the causal relationship between the institutions involved. While actor-centred approaches attribute interaction to forum-shopping activities of relevant actors (Victor and Raustiala, 2004), it points to the institutionally created opportunities and restrictions for such forum shopping and demonstrates how actors transmit influence from the source institution to the target institution. It also captures numerous instances of institutional interaction that are not intentionally triggered and thus reaches well beyond intentional forum-shopping.

The conceptual framework helps structure and understand the multifaceted realm of interaction between international institutions and provides systematic insights into the conditions of governance within the international system. It demonstrates that the broad area of institutional interaction comprises different phenomena that must be carefully analysed to grasp their precise consequences for effective international governance. The four causal mechanisms elucidate how influence can generally travel from the source institution to the target institution, and they draw attention to the varying roles of actors, their preferences and behaviour in this process. The more specific ideal types of interaction reflect particular characteristics of different subsets of cases following the same causal mechanisms and provide substantive hypotheses about the governance implications of institutional interaction.

The general causal mechanisms and more specific ideal types provide an analytical apparatus for the systematic analysis of interaction between international institutions. Based upon their specific rationales, they constitute abstract models stripped of the empirical complexity of real-world situations. Their analytical power originates from their ability to draw attention to the underlying rationales of cases of interaction. They provide the analyst with a tool-box of possible causal relationships between international institutions that facilitate the exploration of real-world situations of institutional interaction. This toolbox might support the analysis of interaction situations in a similar way as the well-known game theoretical models help analyse the formation, operation and design of international institutions.

The causal mechanisms and ideal types can provide the basis for an encompassing theory of institutional interaction. We have made a head-start by elucidating the causal pathways and appearances of single cases of institutional interaction and examining their consequences for international governance. Over time, we may also develop the theoretical understanding of the conditions under which interaction occurs or fails to occur. Future studies may investigate in more detail the conditions under which the causal mechanisms of institutional interaction are triggered and become operational. Situations might be identified in which a potential for interaction exists, while the target institution does not respond so that there are no effects to be observed. Likewise, more detailed studies might investigate the responses to institutional interaction and the factors that influence the success or failure of managing interaction politically. Thus, we can expect to gradually expand our theoretical knowledge about institutional interaction

The causal mechanisms and ideal types provide a sound basis for the exploration of more complex interaction situations and the interlocking structures of international governance institutions. In order to assess the causal relationships between the institutions involved, complex interaction situations must be analytically decomposed into individual cases with unidirectional causal pathways, following a particular causal mechanism and ideal type. These cases may subsequently be recombined to causal chains and clusters in order to grasp complex interaction situations. On this basis, we may gain a conceptually founded idea of the interlocking structures of international governance institutions that emerge from complex interaction situations and shape the normative expectations of international society.

 
References

Abbott, Kenneth W. and Duncan Snidal (2000) ‘Hard and Soft Law in International Governance’, International Organization 54 (3): 421-56.

Aggarwal, Vinod K. (1983) ‘The Unraveling of the Multi-Fiber Arrangement: An Examination of International Regime Change’, International Organization 37 (4): 617-46.

Alexander, Jeffrey, and Bernhard Giesen (1987) ‘From Reduction to Linkage: The Long View of the Micro-Macro Link’, in Jeffrey Alexander, Bernhard Giesen, Richard Münch, and Neil J. Smelser (eds) The Micro-Macro Link, pp. 1-42. Berkeley: University of California Press.

Alter, Karen J. (2000) ‘The European Union's Legal System and Domestic Policy: Spillover or Backlash ?’, International Organization 54 (3): 489-518.

Archer, Margaret S. (1985) ‘Structuration versus Morphogenesis’, in S.N. Eisenstadt and Horst Jürgen Helle (eds) Macro-Sociological Theory. Perspectives on Sociological Theory, Vol. 1, pp. 58-88. London: Sage.

Bernauer, Thomas (1995) ‘The Effect of International Environmental Institutions: How We Might Learn More’, International Organization 49 (2): 351-77.

Brack, Duncan (2002) ‘Environmental Treaties and Trade: Multilateral Environmental Agreements and the Multilateral Trading System’, in Gary P. Sampson and W. Bradney Chambers (eds) Trade, Environment, and the Millennium, pp. 321-52. Tokyo: UN University Press.

Brown Weiss, Edith (1993) ‘International Environmental Issues and the Emergence of a New World Order’, Georgetown Law Journal 81 (3): 675-710.

Buzan, Barry, Charles Jones and Richard Little (1993) The Logic of Anarchy: From Neorealism to Structural Realism. New York: Columbia University Press.

Carlsnaes, Walter (1992) ‘The Agent-Structure Problem in Foreign Policy Analysis’, International Studies Quarterly 36 (3): 245-70.

Chambers, W. Bradnee and Joy A. Kim (eds) (forthcoming) Institutional Interplay. The Case of Biosafety, Tokyo: UN University Press.

Checkel, Jeffrey T. (1998) ‘The Constructivist Turn in International Relations Theory’, World Politics 50 (3): 324-48.

Clapp, Jennifer (1994) ‘Africa, NGOs, and the International Toxic Waste Trade’, Journal of Environment and Development 2 (1): 17-46.

Coleman, James (1990) Foundations of Social Theory. Cambridge: Belnkamp Press.

Compa, Lance A. and Stephen F. Diamond (eds) (1996) Human Rights, Labor Rights, and International Trade. Philadelphia: University of Pennsylvania Press.

Elster, Jon (1989) Nuts and Bolts for the Social Sciences. Cambridge: Cambridge University Press.

Gehring, Thomas (1994) Dynamic International Regimes: Institutions for International Environmental Governance. Frankfurt/Main: Peter Lang.

Gehring, Thomas and Sebastian Oberthür (2004) ‘Exploring Regime Interaction: A Framework of Analysis’, in Arild Underdal and Oran R. Young (eds) Regime Consequences: Methodological Challenges and Research Strategies, pp. 247-269. Dordrecht: Kluwer.

Gehring, Thomas, and Sebastian Oberthür (2006) ‘Comparative Empirical Analysis and Ideal Types of Institutional Interaction’, in Sebastian Oberthür and Thomas Gehring (eds) Institutional Interaction in Global Environmental Governance: Synergy and Conflict among International and EU Policies, pp. 307-371. Cambridge: MIT-Press.

George, Alexander and Andrew Bennett (2005) Case Studies and Theory Development in the Social Sciences. Cambridge: MIT Press.

Haas, Ernst (1990) When Knowledge is Power. Three Models of Change in International Organizations. Berkley: University of California Press.

Haas, Peter M. (1992) ‘Introduction. Epistemic Communities and International Policy Coordination’, International Organization 46 (1): 1-35.

Haas, Peter M., Robert O. Keohane and Marc A. Levy (eds) (1993) Institutions for the Earth. Sources of Effective International Environmental Protection. Cambridge: MIT Press.

Hasenclever, Andreas, Peter Mayer and Volker Rittberger (1997) Theories of International Regimes. Cambridge: Cambridge University Press.

Hedström, Peter and Richard Swedberg (1998) ‘Social Mechanisms: An Introductory Essay’, in Peter Hedström and Richard Swedberg (eds) Social Mechanisms: An Analytical Approach to Social Theory, pp. 1-31. Cambridge: Cambridge University Press.

Héritier, Adrienne (1996) ‘The Accommodation of Diversity in the European Policy-making Process and its Outcomes. Regulatory Policy as a Patchwork’, Journal of European Public Policy 3 (2): 149-167.

Jacquemont, Frédéric and Alejandro Caparrós (2002) ‘The Convention on Biological Diversity and the Climate Change Convention 10 Years after Rio: Towards a Synergy of the Two Regimes?’ Review of European Community and International Environmental Law 11 (2): 139-80.

Keohane, Robert O. (1984) After Hegemony. Cooperation and Discord in the World Political Economy. Princeton: Princeton University Press.

Keohane, Robert. O. and Mark A. Levy (eds) (1996) Institutions for Environmental Aid. Pitfalls and Promises. Cambridge: MIT Press.

Koremenos, Barbara, Charles Lipson and Duncan Snidal (2001) ‘The Rational Design of International Institutions’, International Organization 55 (4): 761-99.

Lanchbery, John (2006) ‘The Convention on International Trade in Endangered Species of Wild Fauna and Flora (CITES): Responding to Calls for Action from Other Nature Conservation Regimes’, in Sebastian Oberthür and Thomas Gehring (eds) Institutional Interaction in Global Environmental Governance: Synergy and Conflict among International and EU Policies, pp. 157-79. Cambridge: MIT Press.

Levy, Marc A., Oran R. Young and Michael Zürn (1995) ‘The Study of International Regimes’, European Journal of International Relations 1: 267-330.

March, James G. and Johan P. Olsen (1998) The Institutional Dynamics of International Political Orders, International Organization 52 (4), 943-969.

Martin, Lisa L. (1993) ‘The Rational State Choice of Multilateralism’, in John G. Ruggie (ed) Multilateralism Matters: The Theory and Praxis of an Institutional Form, pp. 91-121. New York: Columbia University Press.

Mattli, Walter (2003) ‘Setting International Standards. Technological Rationality or Primacy of Power’, World Politics 56 (1): 1-42.

Meinke, Britta (2002) Multi-Regime-Regulierung. Wechselwirkungen zwischen globalen und regionalen Umweltregimen. Darmstadt: Deutscher Universitäts-Verlag.

Miles, Edward L., Arild Underdal, Steinar Andresen, Jørgen Wettestad, Jon Birger Skjærseth and Elaine M. Carlin (2002) Environmental Regime Effectiveness: Confronting Theory with Evidence. Cambridge: MIT Press.

Moorman, Yasmin (2001) ‘Integration of ILO Core Rights Labor Standards into the WTO’, Columbia Journal of Transnational Law 39 (2): 555-83.

Oberthür, Sebastian (2001) ‘Linkages between the Montreal and Kyoto Protocols. Enhancing Synergies between Protecting the Ozone Layer and the Global Climate’, International Environmental Agreements: Politics, Law and Economics 1 (3): 357-77.

Oberthür, Sebastian and Hermann E. Ott (1999) The Kyoto Protocol. International Climate Policy for the 21st Century. Berlin: Springer.

Oberthür, Sebastian and Thomas Gehring (eds) (2006a) Institutional Interaction in Global Environmental Governance: Synergy and Conflict among International and EU Policies. Cambridge: MIT Press.

Oberthür, Sebastian and Thomas Gehring (2006b) ‘Conceptual Foundations of Institutional Interaction’ in Sebastian Oberthür and Thomas Gehring (eds) Institutional Interaction in Global Environmental Governance: Synergy and Conflict among International and EU Policies, pp. 19-51. Cambridge: MIT Press.

Oberthür, Sebastian and Thomas Gehring (2006c) ‘Institutional Interaction in Global Environmental Governance. The Case of the Cartagena Protocol and the World Trade Organization’, Global Environmental Politics 6 (2), in print.

Palmer, Alice, Beatrice Chaytor and Jacob Werksman (2006) Interaction between the World Trade Organization and International Environmental Regimes, in Sebastian Oberthür and Thomas Gehring (eds) Institutional Interaction in Global Environmental Governance: Synergy and Conflict among International and EU Policies, pp. 181-204. Cambridge: MIT Press.

Pontecorvo, Concetta Maria (1999) ‘Interdependence between Global Environmental Regimes. The Kyoto Protocol on Climate Change and Forest Protection’, Zeitschrift für ausländisches öffentliches Recht und Völkerrecht 59 (3): 709-49.

Risse-Kappen, Thomas (1994) ‘Ideas do not Float Freely. Transnational Coalitions, Domestic Structures, and the End of the Cold War’, International Organization 48 (2): 185-214.

Risse, Thomas (2000) ‘Let’s Argue! Communicative Action in World Politics’, International Organization 54 (1): 1-39.

Rosendal, Kristin (2001) ‘Impacts of Overlapping International Regimes. The Case of Biodiversity’, Global Governance 7: 95-117.

Schelling, Thomas C. (1960) The Strategy of Conflict. 15th ed. 1995, Cambridge: Harvard University Press.

Schelling, Thomas (1998) ‘Social Mechanisms and Social Dynamics’, in Peter Hedström and Richard Swedberg (eds) Social Mechanisms: An Analytical Approach to Social Theory, pp. 32-44. Cambridge: Cambridge University Press.

Sebenius, James K. (1983) ‘Negotiation Arithmetics. Adding and Subtracting Issues and Parties’, International Organization 37 (2): 281-316.

Sebenius, James K. (1992) ‘Challenging Conventional Explanations of International Cooperation. Negotiation Analysis and the Case of Epistemic Communities’, International Organization 46 (1): 323-65.

Selin, Henrik and Stacy D. VanDeveer (2003) ‘Mapping Institutional Linkages in European Air Pollution Politics’, Global Environmental Politics 3 (3): 14-46.

Simon, Herbert A. (1972) ‘Theories of Bounded Rationality’, in Charles B. McGuire and Roy Radner (eds) Decision and Organization, pp. 161-76. Amsterdam: North-Holland.

Skjærseth, Jon Birger (2006) ‘Protecting the Northeast Atlantic: One Problem, Three Institutions’, in Sebastian Oberthür and Thomas Gehring (eds) Institutional Interaction in Global Environmental Governance: Synergy and Conflict among International and EU Policies, pp. 102-125. Cambridge: MIT Press.

Snidal, Duncan (1985) ‘The Game Theory of International Politics’, World Politics 38 (1): 25-57.

Snidal, Duncan (1994) ‘The Politics of Scope. Endogeneous Actors, Heterogeneity and Institutions’, Journal of Theoretical Politics 6 (4): 449-472.

Stein, Arthur (1982) ‘Coordination and Collaboration. Regimes in an Anarchic World’, International Organization 36 (2): 299-324.

Stokke, Olav Schram (2000) ‘Managing Straddling Stocks. The Interplay of Global and Regional Regimes’, Ocean and Coastal Management 43 (2/3), 205-34.

Stokke, Olav Schram (2001a) The Interplay of International Regimes. Putting Effectiveness Theory to Work. Lysaker: Fridtjof Nansen Institute. FNI Report 14/2001.

Stokke, Olav Schram (2001b) ‘Conclusions’, in Olav Schramm Stokke (ed) Governing High Seas Fisheries: The Interplay of Global and Regional Regimes, pp. 329-360. Oxford: Oxford University Press.

Tarasofsky, Richard G. (1997) ‘Ensuring Compatibility Between Multilateral Environmental Agreements and GATT/WTO’, Yearbook of International Environmental Law 7, 52-74.

Tistounet, Eric. (2000) ‘The Problem of Overlapping among Different Treaty Bodies’, in Philip Alston and James Crawford (eds) The Future of UN Human Rights Treaty Monitoring, pp. 383-401. Cambridge: Cambridge University Press.

Underdal, Arild (2004) ‘Methodological Challenges in the Study of Regime Effectiveness’, in Arild Underdal and Oran R. Young (eds) Regime Consequences: Methodological Challenges and Research Strategies, pp. 27-48. Dordrecht: Kluwer.

Underdal, Arild, and Oran R. Young (2004) ‘Research Strategies for the Future’, in Arild Underdal and Oran R. Young (eds) Regime Consequences: Methodological Challenges and Research Strategies, pp. 361-380. Dordrecht: Kluwer.

Victor, David G., Kal Raustiala and Eugene B. Skolnikoff (eds) (1998) The Implementation and Effectiveness of International Environmental Commitments: Theory and Practice. Cambridge: MIT Press.

Victor, David G. and Kal Raustalia (2004) ‘The Regime Complex for Plant Genetic Resources’, International Organization 58 (2): 277-309.

Weber, Max (1904) ‘Die “Objektivität” sozialwissenschaftlicher und sozialpolitischer Erkenntnis’, in Johannes Winkelmann (ed) Max Weber: Gesammelte Aufsätze zur Wissenschaftslehre, pp. 146-214. Tübingen: Mohr.

Wendt, Alexander (1987) ‘The Agent-Structure Problem in International Relations Theory’, International Organization 41 (3): 335-70.

Werksman, Jacob (2005) ‘The Negotiation of a Kyoto Compliance System’, in Jon Hovi, Geir Ulfstein and Olav Schram Stokke (eds) Compliance with Climate Commitments: Conditions and Mechanisms, pp. 17-38. London: Earthscan.

Yee, Albert S. (1996) ‘The Causal Effects of Ideas on Politics’, International Organization 50 (1): 69-108.

Young, Oran R. (1992) ‘The Effectiveness of International Institutions: Hard Cases and Critical Variables’, in James N. Rosenau and Ernst-Otto Czempiel (eds) Governance without Government: Order and Change in World Politics, pp. 160-94. Cambridge: Cambridge University Press.

Young, Oran R. (1996) ‘Institutional Linkages in International Society: Polar Perspectives’, Global Governance 2 (1): 1-24.

Young, Oran R. (ed) (1999) The Effectiveness of International Environmental Regimes: Causal Connections and Behavioral Mechanisms. Cambridge: MIT Press.

Young, Oran R. (2002) The Institutional Dimensions of Environmental Change. Fit, Interplay, and Scale. Cambridge: MIT Press.

Young, Oran R. (forthcoming) ‘Institutional Coordination through Inter-linkages’, in: W. Bradnee Chambers and Joy A. Kim (eds) Institutional Interplay: The Case of Biosafety. Tokyo: United Nations University Press.