Sage Journals: Discover world-class research

Abstract

Regime complexity is often assumed to be either detrimental or conducive to global governance. We argue that the effects of complexity vary depending on why actors turn institutional overlaps into manifest conflicts over the prevalence of different norms and rules (so-called interface conflicts) and how these conflicts are managed. This paper conjectures that two types of interface conflicts shape the effects of complexity in distinct ways. First, non-instrumental interface conflicts are side-effects of institutional differentiation that emerge when new events or issues evoke incompatible response strategies by different institutions. Since none of the actors involved in such conflicts necessarily have an interest in undermining the other, they are likely to be handled in a cooperative way and thus have integrative effects for institutional order. Second, instrumental interface conflicts are purposeful creations of actors seeking political advantage by positioning one norm or rule against another. Here, overlaps are utilized to substantively contest existing institutions, making non-cooperative conflict management, and eventually the fragmentation of institutional order more likely. To analyze these conjectures, we first conduct cluster analyses of 77 interface conflicts which reveal “benign,” “malign,” and “hybrid” clusters of interface conflicts. We further analyze temporal and causal sequences in qualitative case studies for each cluster.

Keywords

Global governance interface conflicts international order multilateralism norms regime complexity

Introduction

It has become common knowledge that the post-Cold War proliferation of formal and informal, public and private, hybrid and noncompound international institutions has increased institutional density in global governance (Raustiala, 2013). As a consequence, a rich research program has grown around the concept of international regime complexity, denoting sets of partially overlapping and non-hierarchically integrated institutions that intersect in the governance of a particular issue area or subject matter (e.g. Alter and Meunier, 2009; Orsini et al., 2013). While contributors to this research program agree that regime complexity alters the opportunity and incentive structures for states and non-state actors in global politics, the literature remains divided on precisely how it does so and with what consequences for the effectiveness and integrity of global governance (see also Henning and Pratt, 2023).

One group of scholars initially claimed that regime complexity would undermine intergovernmental cooperation and compliance with international legal obligations. Institutional overlap, they argued, allowed states to exploit double structures and legal ambiguities to play institutions against each other (Benvenisti and Downs, 2007; Drezner, 2009). The result would be the fragmentation of international law and a disintegration of global order (Drezner, 2013; see also Gómez-Mera, 2016).

By contrast, later assessments foregrounded not only the potential functional benefits of densely populated governance spaces but also states’ long-term interest in stable multilateral institutions. Functionally, regime complexes were said to provide more flexibility and attract greater resources as well as expertise than single institutions, allowing them to achieve global governance goals more effectively than less complex regimes (Keohane and Victor, 2011; Lesage and van de Graaf, 2013). Moreover, even where powerful states could exploit institutional overlaps to attain short-term goals to the detriment of regime integration, they keep a prevailing long-term interest in multilateral cooperation. Therefore, the expectation is that institutional overlap leads to a division of labor or deference among the institutions of a regime complex in the long run (Gehring and Faude, 2014; Morin et al., 2017; Pratt, 2018).

Both camps can point to empirical observations that underpin their respective stance. Some regime complexes show signs of conflict and exploitation of overlaps, such as those for plant genetic resources (Raustiala and Victor, 2004) and European security (Hofmann, 2009). Others, such as the regime complexes of development banking (Faude and Fuß, 2020) and climate change (Keohane and Victor, 2011), seem to operate cooperatively. Given this variety in outcomes, regime complexity per se can hardly determine conflictive or cooperative outcomes. While regime complexity scholarship has started to investigate variation in the effects of complexity, it often focuses on institutional structures and legal configurations as explanatory factors (Eilstrup-Sangiovanni and Westerwinter, 2022: 250; Henning and Pratt, 2023). In this article, we shift the focus to political agency.

In our view, an essential but often overlooked question is how actors navigate overlapping rule sets and, in particular, how they activate and manage conflicts arising in this context. Therefore, we emphasize the concept of interface conflicts, defined as “incompatible positional differences between actors about the prevalence of different international norms or rules” (Kreuder-Sonnen and Zürn, 2020: 243). To better understand the various institutional outcomes under regime complexity, we argue it is important to foreground why different actors—governments, international organizations (IOs), other non-state actors—activate interface conflicts and how they are managed.

First, we submit that the management of interface conflicts within a regime complex can vary and is decisive for the integrity of the order. When conflict management is highly conflictive, it likely undermines normative coherence and thus cooperative outcomes in a regime complex. When conflict management is cooperative, however, it should increase coordination and may thus even strengthen the governance capacity of a regime complex.

Second, we hold that the mode of conflict management in regime complexes is strongly influenced by the type of conflict we are witnessing, especially regarding its origin. We distinguish two ideal-typical ways in which interface conflicts come about. On the one hand, interface conflicts can emerge as unintended side-effects of divergent problem-solving logics across individual institutions or issue areas. As unforeseen events occur or new issues emerge, different institutions might suggest different responses. Sometimes, the solution to the problem in one area contradicts institutionalized goals in another. In such cases, actors leveraging norms or rules to achieve outcomes at odds with other norms or rules follow the logic of their institutional field. They do not act to undermine another institution or pursue ulterior motives. We label such emergent collisions non-instrumental interface conflicts (NIICs). On the other hand, the norm collisions underlying interface conflicts can also be willfully created by state or non-state actors to challenge particular norms and institutions, circumvent legal obligations, or contest institutional privileges of other actors. Actors exploit or strategically construct norm inconsistencies not only to justify behavior inconsistent with rules of institution A by reference to the rules of institution B, but also to challenge the political goals and authority of existing institutions. We label such deliberate conflicts instrumental interface conflicts (IICs).

We argue that IICs and NIICs are likely to trigger different kinds of conflict management. Whereas emergent NIICs may expose substantial disagreements about norm hierarchies and governance goals, the involved actors should be open to compromise and have a predisposition for cooperative conflict management. Deliberate IICs, on the other hand, are more difficult to conclude in mutually agreeable terms. By definition, at least one party to the conflict intentionally takes a position that is incompatible with that of others. Conflict management is thus more likely to be dominated by parochial interests rather than a desire for a smooth functioning of the regime complex. Non-cooperative conflict management is expected to prevail initially; cooperative conflict management may develop over time but is not guaranteed. By implication, we furthermore suppose that NIICs, through their cooperative management, will contribute to more integration or coordination in regime complexes, whereas IICs are more likely to lead to further fragmentation and disorder.

We probe this argument against empirical evidence from the Interface Conflicts 2.0 Dataset¹ (ICD) which contains information on 77 interface conflicts in various issue areas that occurred between 1970 and 2018. We start by conducting a quantitative cluster analysis with several variables of interest included in the ICD. It allows us to detect the systematic common appearance of features and hence to see the regularity with which instrumental and non-instrumental conflicts come with cooperative or non-cooperative conflict management and integrative or disintegrative effects on the regime complex, respectively. The results mainly support our theoretical conjecture: NIICs are virtually always managed cooperatively, while non-cooperative conflict management and fragmentation outcomes prevail in many IICs. However, we also see a subset of cases which is characterized by deliberate IICs that are managed cooperatively and eventually lead to regime complex integration. This additional cluster points to processes through which initially uncompromising stances can be overcome and yield cooperative solutions.

We complement this quantitative analysis with three qualitative case studies from each cluster. The case studies confirm the temporal and causal sequence of our variables of interest and provide examples of the mechanisms that link the origin of an interface conflict to the conflict management type, and its management type to order effects. The first case study focuses on a typical case to illustrate our main theoretical conjecture. The second case study particularly explores the conditions under which IICs may be managed cooperatively despite their confrontational predisposition. The third study explores what we call a “messy” case to test the boundaries of our theoretical concepts.

The article provides three main contributions. First, our focus on manifest conflicts and their management increases our understanding of the mechanisms through which regime complexity enhances or undermines international cooperation and how the architectures of individual regime complexes change over time (see also Eilstrup-Sangiovanni, 2022; Eilstrup-Sangiovanni and Westerwinter, 2022: 243–248). Second, we theorize a key variable that helps explain the direction of this change. By distinguishing different types of conflicts likely to produce different types of outcomes, we bridge the divide between regime complexity “optimists” and “pessimists.” Third, we provide new empirical evidence on the micro-level of regime interaction and conflict management, enriching debates about regime complexity often marked by meso- or macro-level observations. This evidence sheds light on how actors shape patterns of regime complexity toward more consonant or dissonant forms.

A conflict model of regime complexity

In this section, we outline the core concepts of a conflict model of regime complexity. We assume that institutional overlaps and norm collisions are ubiquitous in regime complexes and densely populated institutional orders more generally. Where two or more institutions, norms, or rules apply to a governance object, social actors may activate or construct a conflict between them—or choose not to (Kreuder-Sonnen and Zürn, 2020). The mere presence of overlapping institutions and potentially conflicting rule sets thus says little about order levels in regime complexes. To describe the degree of (dis-)integration and understand changes over time, we argue that focusing on actual conflict and its management is key.

Our unit of analysis is the interface conflict, which occurs when actors in international politics express—discursively or behaviorally—positional differences about which one of two or more overlapping international norms or rules should take precedence. Operationally, interface conflicts are either directly identified through statements expressing conflicting views on norm precedence or indirectly through justifications of behavior that is inconsistent with norm A, but consistent with norm B. The concept of interface conflicts thus leaves aside norm inconsistencies that are not activated. Interface conflicts, as they are perceived and acted upon, will spur some form of conflict management (see Rittberger and Zürn, 1990)—attempts by conflicting or third parties² to resolve perceived incompatibilities, such as establishing hierarchies, dividing tasks, or introducing collision rules (Zelli, 2011: 207). Such attempts need not be sincere or successful. Outcomes of conflict management can vary widely depending on the legal setup and the parties’ willingness to settle. Interface conflicts may lead to greater integration through coordination and differentiation, or productive “legal entanglement” (Krisch, 2021), but also to fragmentation if management worsens contradictions or triggers further conflicts. Figure 1 summarizes the baseline framework.

Figure 1.

Analytical framework.

A key question is when conflict management leads to more integration or fragmentation. Put simply, functional institutionalists expect management to rectify collisions and foster integration (Gehring and Faude, 2014; Henning and Pratt, 2023), while power-based approaches suggest that insincere attempts will lead to fragmentation (Benvenisti and Downs, 2007; Drezner, 2009). We suspect that both perspectives are valid under certain conditions. Elucidating these conditions should help bridge the divide between skeptics and optimists of regime complexity, providing a tool to understand empirical variation in integration and fragmentation (see also Henning and Pratt, 2023).

Our first key argument is that the origin of an interface conflict matters for its management. Different conflict origins provoke different responses. We distinguish IICs and NIICs, leading to non-cooperative and cooperative management, respectively. These conflict types reflect two dimensions of regime complexity that are often not analytically separated: the challenge of navigating overlapping institutions (Alter and Meunier, 2009) and the opportunities for cross-institutional strategizing to advance parochial interests (Morse and Keohane, 2014). The distinction of these conflict types is not novel, but we claim that it can be utilized to explain differences in regime complex evolution.³

NIICs emerge as an unintended side effect of growing institutional complexity. Especially when new topics arrive on the global agenda, partially overlapping institutions with distinct goals and mandates may simultaneously suggest different ways of handling the issue. Similarly, different rule-addressees may resort to different institutions and justify their positions with reference to different norms when first encountering a new political problem. An example of such an emergent NIIC can be found in the dispute over the World Bank’s obligation to consider indigenous and environmental rights in its projects (Krisch et al., 2020). While advocacy groups and the Office of the High Commissioner for Human Rights (OHCHR) claimed that the World Bank needed to respect a wide range of human rights norms, the Bank itself held that it was bound by its own Articles of Agreement that prohibit “political” considerations to affect project design. Neither conflict party instrumentally sought a norm collision to undermine the norms or authority of the other side. Rather, the different objectives and functional logics of the development finance and the human rights regimes suggested different ways of addressing a problem of growing political relevance. To be sure, conflict activation will always contain intentionality when expressing positional differences regarding the prevalence of different norms. It is the emergence of the underlying norm collision that is unintentional in NIICs.

For deliberate IICs, a conflict is not simply emergent but actively created to challenge a particular rule, norm, or institution. Actors dissatisfied with the goals or obligations of one institution may instrumentalize another institution’s norms to assert their political preferences.⁴ They might activate previously dormant norm collisions or create new ones to generate conflict. A notable example is the clash between the African Union (AU) and the United Nations (UN) Security Council during the 2011 Libyan crisis. Embedded in a long-standing power struggle between the UN and AU within the African security regime complex (Brosig, 2015), the Libyan civil war served as a trigger and opportunity for the AU to challenge the UN’s authority. AU representatives mobilized norms of sovereignty, non-interference, and subsidiarity to contest the UN’s norms of human security and the Responsibility to Protect, seeking to undermine the UN’s authority (Moe and Geis, 2020b).

These differences in conflict origin have implications for the expected type of conflict management. We distinguish between cooperative and non-cooperative conflict management, following Keohane’s (1984) definitions. Cooperative management involves attempts where parties agree to follow procedural norms and accommodate each other’s preferences at least somewhat in their respective position. Non-cooperative management occurs when parties seek to resolve the dispute solely in their favor, disregarding their opponent’s preferences or procedural norms.

Our key proposition is that NIICs are more likely to be managed cooperatively, whereas IICs tend to lead to non-cooperative management, at least initially. In NIICs, conflict parties may still want to uphold their positions (otherwise there wouldn’t be conflict), but they also value the stable provision of public goods by the involved regimes (see also Gehring and Faude, 2014). This mix of motives should encourage some willingness to compromise and/or adhere to procedural norms, leading to cooperative management. For instance, in the conflict pitching human rights norms against the rules of the World Bank, the latter preferred to stick to its policies. Still, it generally also supported human rights and environmental protection, making compromise possible. Consequently, conflict management proceeded cooperatively, with both parties using norm-based arguments and partially accommodating each other’s positions (Krisch et al., 2020).

In contrast, IICs are more likely to lead to non-cooperative conflict management, where at least one party is unwilling to compromise or follow procedural norms. If the conflict was activated to challenge an unwanted institution, it is unlikely that this desire fades during management. Similarly, when parties aim to avoid obligations by leveraging different institutions, they are unlikely to pursue mutually accommodating solutions. IICs thus often result in non-cooperative management, with the conflict parties trying to assert their positions uncompromisingly.⁵ This was the case in the conflict between the AU and the UN. The UN’s response to the AU’s open challenge was basically to ignore the contestation and shut the AU out of the process. While the AU had prepared a roadmap for peace and was to send mediators into Libya, the UN Security Council imposed a no-fly-zone over the country, effectively staging a foreign military intervention (Hehir, 2013). In the aftermath, neither the UN nor the AU undertook any steps toward alleviating the conflict (Williams and Boutellis, 2014).

Our argument does not imply that every NIIC is managed cooperatively or every IIC non-cooperatively. While NIICs may rarely result in non-cooperative management for idiosyncratic reasons, IICs should offer greater variation in outcome. NIICs resemble “coordination games” resolved relatively easily, while IICs resemble “collaboration games,” where management may remain non-cooperative but can shift toward cooperation over time (Snidal, 1985; Zürn, 1992). Ongoing non-cooperative behavior may lead to suboptimal outcomes for both parties, fostering an interest in finding cooperative solutions. While initial IIC management is often non-cooperative, it may evolve into a cooperative approach depending on the conflict’s dynamics and the parties’ interest constellations (see Axelrod and Keohane, 1985).

Finally, we argue that cooperative and non-cooperative conflict management have different effects on the institutional order within a regime complex. Cooperative management contains elements conducive to greater regime complex integration. By working toward mutually agreeable solutions, potentially involving third-party assessments that serve as precedents, cooperative management can increase coordination, collaboration, and even institutionalization of conflict management systems to address future norm collisions. In the case of the World Bank, for example, cooperative management led to an integrative outcome: The Bank now interprets its Articles of Agreement as mandating, not prohibiting, human rights considerations in project implementation. The norm collision eventually resulted in mutual incorporation (Krisch et al., 2020). In contrast, non-cooperative conflict management tends to increase fragmentation. Parties with uncompromising stances who disregard procedural norms are more likely to exacerbate inter-institutional contradictions. Their behavior may also draw other actors into the conflict, forcing them to choose sides, which further destabilizes the regime complex. This dynamic was evident in the AU-UN conflict, where mistrust grew, and the “spirit of partnership” between the institutions diminished, worsening the fragmentation of the African security complex (Moe and Geis, 2020a; Williams and Boutellis, 2014).

Figure 2 summarizes our argument. The model comes with certain caveats. First, the analytical distinction between IICs and NIICs may be clear in theory but harder to pinpoint empirically, as actor intentions can be difficult to discern. Motivations may also mix elements of both, so empirical cases will fall along a spectrum rather than fitting neatly into categories. This raises methodological challenges we address in the following section and in our case selection for the qualitative analysis. Second, the model is not deterministic. Each arrow in Figure 2 represents the most likely consequence, but alternative configurations are possible, and the empirical analysis will be attentive to any systematic patterns.

Figure 2.

Conflict model of regime-complex evolution.

Quantitative analysis: origin, management, and order effects across conflicts

In this section, we begin the empirical exploration with a quantitative analysis of data contained in the ICD 2.0 developed by Zürn et al. (2024). The section first presents the dataset, explaining sampling and coding procedures. Second, we introduce cluster analysis to explore the similarities and differences between types of interface conflicts. We begin with a simple cluster solution based on the three main variables we identified theoretically (conflict origin, conflict management, and effect on order). Subsequently, we provide an extended cluster solution with a broader set of variables to offer a more encompassing characterization of interface conflicts.

The interface conflicts dataset

The ICD is the first to provide comprehensive coverage of a multitude of interface conflicts and their management across time and issue areas.⁶ It covers 77 interface conflicts whose starting points lie between 1970 and 2018 and affect institutions in virtually all policy domains. The dataset is unique in that it relies on input from a relatively large group of scholars coding complex items based on in-depth qualitative case knowledge. Data collection was carried out by six teams of researchers collaborating in the joint project “Overlapping Spheres of Authority and Interface Conflicts in the Global Order (OSAIC).” Given the focus on actor perceptions and behavior in the conceptualization of interface conflicts, cases are not readily identifiable. To be uncovered, interface conflicts typically require in-depth knowledge of given regime complexes and norm collisions around which conflicts unfold in various ways. The project thus began without prior knowledge of the universe of cases—foreclosing avenues of strategic sampling. It was made sure, however, that the individual research projects would cover various issue areas. They cover regime complexes in the fields of Internet governance (Flonk et al., 2020), drug prohibition, health, and indigenous rights (Gholiagha et al., 2020), African security governance (Moe and Geis, 2020a), international economic governance (Krisch et al., 2020), and international judicial bodies (Birkenkötter, 2020). Overall, the different projects complement each other in their case selection in terms of actors, policy fields, and procedural characteristics (Fuß et al., 2021: 13–15).

The ICD exhibits some limitations that are important to bear in mind. First, being based on qualitative studies of issue areas, the identification of interface conflicts is delimited by researchers’ perceptions of the contours of issue areas and the actors and norms that are relevant.⁷ Data collectors may thus have omitted conflicts arising from institutional overlaps that did not fit into preconceptions of the respective subject matters. Second, the sample is not systematically stratified to proportionally reflect certain characteristics of interest. Instead, case inclusion in the sample was essentially lottery-based. The main determinants were issue-area-specific interests of the different research groups, which resorted to snowball-sampling within their fields of expertise, potentially prioritizing more visible (openly conflictive) cases over more contained ones. Hence, we cannot rule out a structural over- or underrepresentation of any specific type of case.

While the sampling procedure is thus admittedly not ideal for statistical analysis, we suggest that the dataset does not introduce major structural or systematic biases regarding our research question. For one, the sample spans virtually all issue areas, looks at all conceivable conflict parties, includes cases with and without third-party involvement, and covers broad variation in our key variables of conflict origin, conflict management, and effects on order (Fuß et al., 2021: 14–15). For another, it also grasps most possible types of regime complexity and the conflicts arising within them (Eilstrup-Sangiovanni and Westerwinter, 2022; Gehring and Faude, 2014; Orsini et al., 2013): institutional overlaps within an issue area (e.g. the regime complex of climate change), overlaps among institutions rooted in different issue areas (e.g. the trade-environment nexus), and overlaps between issue-specific regimes and those that potentially cut across all issues (such as the human rights regime).⁸

The coding of the cases reflects the principles of structured expert surveys, that is, the group’s coordination unit prepared a coding scheme that the project teams filled in based on their extensive expert knowledge of the cases they studied in depth. The concepts measured in the database were developed by the research group collectively over several years, ensuring a congruent understanding of the concepts used. All coders followed a set of coding rules, which provide general guidance on determining the separation of cases, and lay out specific requirements for coding single categories of the conflict, conflict management, and its outcome. Because coding interface conflicts requires high levels of case-specific expertise, recoding all cases to conduct formalized inter-coder reliability tests would have incurred prohibitive costs. For a small subset of five cases, however, the project team “RESPONSES” double-coded all items by different researchers independently. The coders reached an agreement of 84% (Fuß et al., 2021: 16).

The ICD contains codes for 86 items ranging from the starting year and conflict parties to conflict management attributes and outcomes. For this paper, we summarize in Table 1 those variables and their coding possibilities that are most relevant to our research question. Reflecting the theoretical model, they start with the conflict origin, go via conflict management, and end with effects on the institutional order. The table also includes three additional variables we introduce below—the object, spheres, and outcome of a conflict.

Table 1.

Overview of key variables in the Interface Conflicts 2.0 Dataset.

Variable	1		0		N
Origin: Intentionally created (1) vs unintended by-product (0)	42	(54.5%)	35	(45.5%)	77
Management: Cooperative (1) vs non-cooperative (0)	59	(76.6%)	18	(23.4%)	77
Oder effect: Integration (1) vs fragmentation (0)	46	(60.5%)	30	(39.5%)	76
Object: Interest (1) vs values or means (0)	30	(39%)	47	(61%)	77
Spheres: Across (1) or within spheres of authority (0)	52	(67.5%)	25	(32.5%)	77
Outcome: Symmetric (1) vs asymmetric gains/costs (0)	21	(28%)	54	(72%)	75

Cluster analysis

We use cluster analysis to demonstrate how interface conflicts are characterized by their origin, management, and order effects. Cluster analysis is a well-established method to group observations based on their similarities (Everitt et al., 2011; Kaufman and Rousseeuw, 1990). We use agglomerative hierarchical cluster analysis based on Gower’s general dissimilarity coefficient and Ward’s minimum variance clustering method (Gower, 1971; Ward, 1963).⁹ Because hierarchical cluster analysis does not require a probability model, it is a good match for our data—relatively few cases, described by many variables, derived from a non-random sample.

While cluster analysis is not designed for hypothesis testing or drawing causal inference, we argue that if the clustering of cases follows our expectation, the method can provide initial evidence for our argument that conflict origin affects conflict management and ultimately the order of regime complexes. In conjunction with our theoretical model and the case studies in the next part of the paper, we are confident of the support we find for our causal claim.

We proceed in two steps: First, we conduct a simple cluster analysis by using our three main variables for clustering and demonstrate that their distribution largely conforms to theoretical expectations. Second, we add complexity and expand the set of clustering variables to include important alternative explanations.

Simple cluster solution

Figure 3 shows the results of the initial cluster analysis in a dendrogram. The tree-like structure of the graph illustrates the hierarchical nature of clustering, where smaller clusters are nested within larger ones. The dissimilarity of clusters is indicated by the height of the bifurcations of the tree—the higher these are located on the y-axis, the more dissimilar the two clusters they connect.

Figure 3.

Interface conflict clusters, agglomerative hierarchical clustering, complete linkage method.

Figure 3 shows that interface conflicts can be divided into three major clusters, indicated in the graph by gray shading of the three branches. There are two bigger clusters and one smaller cluster: cluster 1 contains 35 cases, cluster 2 contains 26 cases, and cluster 3 contains 16 cases. We include a full list of cluster members in the online appendix. What sets the three clusters apart is the value distribution of the three clustering variables among cluster members. Figure 4 illustrates these distributions by plotting the proportion of members of the clusters scoring 1 on each of the variables, indicated by the size of the plotted points.

Figure 4.

Distribution of variable scores within clusters.

The three major clusters largely concur with our expectations regarding the association of conflict origins, management, and effects on sectoral order. Cluster 1 consists entirely of NIICs, which are in 94% of the cases followed by cooperative conflict management and display high levels of integration (74%). We might call this the “benign cluster.” By contrast, cluster 3 exclusively contains IICs that are always followed by non-cooperative management and have almost exclusively fragmenting effects on the international order (94%). We might call this the “malign cluster.” Cluster 2 points to the possibility that deliberate IICs can eventually also be handled cooperatively. It exclusively contains IICs which are exclusively handled through cooperative conflict management and mostly lead to integration (77%). We might call this a “hybrid cluster” of malign conflicts turned benign.¹⁰

This simple analysis corroborates our theoretical expectations. Emergent NIICs are regularly followed by cooperative conflict management, which has a predominantly positive effect on regime complex integration. Deliberate IICs can go both ways, but non-cooperative conflict management is dominant.

Extended cluster solution

In the second step of our analysis, we extend the selection of variables entering the cluster analysis. We include three variables that each reflect an alternative explanation for the type of conflict management or the type of order effect (or both). If the cluster analysis still yields the three types just identified, the argument about the structuring power of underlying types of conflicts is further strengthened.

First, we include the variable object which describes the substance of the conflict as either about the (power) interests of the involved actors or about incompatible values or the means by which these values should be realized (see Rittberger and Zürn, 1990). It could be intuitive to expect that, irrespective of the conflict origin, conflict management tends to be non-cooperative in interface conflicts about power interests where bargaining logics dominate and cooperative in conflicts about values or means where normative arguments matter and create room for compromise. Second, we include the variable spheres. It denotes whether an interface conflict occurs within one or across two or more spheres of authority, defined as “a governance space with at least one domestic or international authority, which is delimited by the involved actors’ perception of a common good or goal at a given level of governance” (Kreuder-Sonnen and Zürn, 2020: 255). Conflicts within a sphere might be more likely to yield cooperative outcomes than conflicts across spheres since normative clashes are supposedly less fundamental and actors’ incentives to compromise for the sake of cooperation gains are higher (Gehring and Faude, 2014). Finally, we include the variable outcome which pertains to the symmetry in the distribution of costs and gains of a settlement among conflict parties. An asymmetric distribution could have a fragmenting effect on order, whereas a symmetric distribution could have an integrative effect. Including these three variables allows us to see whether our hypothesized associations hold or are overlain by the most likely countervailing factors.

The findings of the extended cluster analysis are illustrated in the dendrogram in Figure 5. Including the additional variables does not substantively alter our main finding and conclusion. We can see three main clusters that are, by and large, identical to the ones from the simple cluster solution in size and composition. The internal structure of the clusters is more varied—with more sub-clusters and hierarchical levels of clustering—because of the higher number of examined variables. However, the central threefold division remains.

Figure 5.

Interface conflict clusters, extended variables set.

Most importantly, the distribution of our three main variables is very similar to the initial simple cluster analysis. That is, interface conflicts still divide into a “benign,” a “malign,” and a “hybrid” cluster. Figure 6 shows the distributions, including the three new variables. There are only slight changes among our three main variables compared to the simple cluster analysis. Most notably, cluster 1 now contains more instrumental conflict origins (21% as opposed to 0%), while cluster 2 contains fewer (78% vs 100%). Everything else remains largely the same: cooperative management and integrative effects barely changed (cluster 1: 95% and 71% as opposed to 94% and 74%, respectively; cluster 2: 100% and 82% as opposed to 100% and 77%, respectively). Cluster 3 is identical to the simple solution.

Figure 6.

Distribution of variable scores within clusters, extended cluster solution.

The additionally included variables do not significantly alter the clusters. Object and outcome do not seem to vary in any meaningful way across the three clusters. Interests (as opposed to values and means) as the object of conflict occur in a roughly similar frequency across clusters and characterizes the majority of cases in each (62%, 65%, and 88%). Conversely, outcomes with a symmetric (as opposed to an asymmetric) distribution of costs and gains are also spread in a roughly similar manner across clusters representing a minority of cases in each (24%, 45%, and 12%). However, the variable spheres may offer some additional insights into the existing clusters: in cluster 1, all interface conflicts occur within a sphere of authority; in cluster 2, almost all (96%) conflicts occur across spheres; and in cluster 3, the distribution between conflicts within and across spheres is even (50%). Whereas the malign cluster is thus not shaped by spheres, it is noteworthy that the benign cluster is characterized by within-sphere conflicts and the hybrid one by conflicts occurring across spheres.

In sum, the cluster analyses lend support to our theoretical argument about the relationship between conflict origin, conflict management, and order effects. They show the existence of a benign, a malign, and a hybrid cluster, whose internal cohesion is determined by the three variables of interest. The benign cluster is about cases of NIICs that spur cooperative management and integrative effects. The malign cluster is about IICs that spur non-cooperative management and fragmenting effects. These two conform entirely to the baseline theoretical expectation. The hybrid cluster of IICs with cooperative management and integrative effects is less straightforward and confirms the expectation that IICs—given their collaboration game character—may evoke cooperative conflict management over time. The results prompt two further questions we seek to address through qualitative case studies in the following section: (1) Since the cluster analysis per se does not say anything about causal connections, can we corroborate the temporal sequence and causal mechanisms assumed for the clusters? (2) In the hybrid cluster, can we see a turn from non-cooperative to cooperative conflict management over time, and what explains this turn?

Qualitative analysis: from conflict to management and order effects within cases

This section illustrates the internal logic of the three clusters and explores the causal mechanisms associated with the variables in our model based on qualitative case studies. Instead of simply exemplifying our two main theoretical propositions with fitting cases, we select cases that reflect the variety of causal connections between our variables and problematize the cluster boundaries. We look at one IIC (colliding refugee norms mobilized by Israel and the US to contest the UN Relief and Works Agency (UNRWA)) that leads to non-cooperative conflict management and fragmentation (cluster 3), another IIC (the creation of the International Renewable Energy Agency (IRENA) whose norms and rules opposed the existing International Energy Agency (IEA)) that has the opposite effect, leading to cooperative conflict management and integration (cluster 2), and finally a borderline case of NIIC (a conflict between private investors and the government of Zimbabwe over the preponderance of human rights versus investment norms) where instrumental and non-instrumental motivations partially overlap and (therefore) the types of conflict management and order effect also blur (cluster 1).

Of these three cases only one is a “typical” case from our theoretical vantage point (refugee norms), where an IIC leads to non-cooperative conflict management and eventually to fragmentation. Such a typical case allows us to illustrate the theoretical argument. The second case (energy norms) is one which contradicts the baseline assumption: an IIC leads to cooperative management and integration—exactly the opposite of the first “typical” case. While we theoretically presaged the possibility for such cases to occur and the “hybrid” cluster thus did not come as a surprise per se, our theory is not able to fully explain when and how IICs turn benign eventually. Studying this atypical case thus allows us to further develop the theory. Our third case (human rights vs investment norms) exhibits lots of gray tones when it comes to assessing the variables of interest. While coded as a NIIC (in the binary coding scheme of the ICD), it is a case that also shows some signs of rather instrumental reasoning. While coded as cooperative conflict management in the ICD, there are also some aspects in the conflict management that look rather non-cooperative. And while coded as having integrative order effects in the ICD, the qualitative analysis also reveals greater complexity here. This case thus allows us to problematize (our own) binary conceptions of interface conflicts while reinforcing the theoretical expectation that the type of interface conflict matters for its management and order effects.

The malign cluster: pitting UNHCR against UNRWA norms

UNRWA was founded in 1949 through UN General Assembly Resolution 302, ending the Arab-Israeli conflict of 1948. The agency’s mandate comprises the provision of education, relief services, as well as microfinance and housing assistance to registered Palestinian refugees in Palestine, Syria, Lebanon, and Jordan. According to UNRWA, a Palestinian refugee is defined as “a person whose normal residence was Palestine for a minimum of two years preceding the outbreak of the conflict in 1948 and who, as a result of this conflict, lost both his home and his means of livelihood” (Barakat, 1973: 147).

There is an evident linkage to the work of the Office of the United Nations High Commissioner for Refugees (UNHCR). Arguably, there is also some normative inconsistency between the two, especially as far as UNRWA’s classification procedure is concerned. In contrast to UNHCR, UNRWA allows Palestinian refugees to transmit their refugee status from one generation to another. Moreover, UNHCR relies on Article I (c)(3) of the 1951 UN Convention and Protocol Relating to the Status of Refugees, providing that an individual no longer has refugee status if they have “acquired a new nationality and enjoy the protection of the country of their new nationality.” In contrast, UNRWA’s definition of a Palestinian refugee includes no such provision.

These divergent normative backgrounds and concomitant practices co-existed for several decades without causing much concern. It was only in 2017, when the nationalist Israeli government of Benjamin Netanyahu, backed by the newly elected Republican US President Donald Trump, took a stab at UNRWA by claiming that the organization’s refugee classification was inconsistent with the broadly recognized and more legalized refugee definition of the UNHCR and should therefore be abandoned. Prime Minister Netanyahu deliberately activated the interface conflict in order to weaken UNRWA which had been a thorn in Israeli nationalists’ side for a long time. This motivation is reflected in the fact that Netanyahu did not suggest ways to reconcile UNRWA’s practice with UNHCR’s legal basis, but instead called for UNRWA’s responsibilities to be entirely transferred to UNHCR and for its funding to be cut (Beaumont and Holmes, 2018). Hence, we treat this as a case of IIC.

The conflict created two opposing actor coalitions with Israel and the United States (and a few additional countries) on one side and the Palestinian authorities, the European Union (EU) (including most member states), and the UN Secretariat on the other. Management of the conflict was entirely non-cooperative, with first the United States and later the EU taking unilateral actions to assert their preferences without regard for the other. On the side of the United States and Israel, there was no interest in rectifying the norm collision other than through the complete abandonment of UNRWA. In turn, EU and UNRWA had little incentive to compromise.

On 31 August 2018, the United States announced that it would not provide any additional funding to UNRWA in the future.¹¹ In July 2019, the Netherlands and Switzerland announced to suspend their funding to UNRWA as well. The withdrawal of US funding, in particular, left UNRWA in a precarious financial situation. The EU, the largest contributor to UNRWA, urged the United States to reconsider its decision (European Parliament, 2018). It said it would continue its assistance to the agency and discuss funding alternatives with other partners. Antonio Guterres, the UN’s Secretary General, also expressed his “full confidence” in UNRWA, and urged other countries “to help fill the remaining financial gap” (UN News, 2018). Despite Israeli and US pressure, the UN General Assembly’s Special Political and Decolonization Committee (Fourth Committee) renewed the mandate of the agency on 15 November 2019 for another 4 years (UNRWA, 2019). In a statement published by the official Palestinian news agency WAFA, Palestinian President Mahmoud Abbas welcomed the vote as “a triumph for international law and the rights of the Palestinian refugees, until their issue is finally resolved in accordance with United Nations resolutions” (WAFA, 2019).

In line with our analytical framework, the temporality of this case is straightforward: the deliberate IIC led to non-cooperative conflict management whose effect on the regime complex has been more, not less, fragmentation. The stand-off between supporters and critics of UNRWA remained virulent throughout the tenure of Trump and Netanyahu. To date, no agreement has been reached and none of the various actors involved seems to be willing to compromise. The war on Gaza that started in October 2023 has overshadowed the interface conflict with a much more drastic and violent component.

The hybrid cluster: competing norms in global energy governance

The second case concerns a conflict over the prevalence of competing norms in global energy governance embodied by the IEA, which promotes oil and fossil fuels as the main sources of global energy, and the IRENA, which promotes renewable energies. Even though the conflict between the two was created strategically by environment-oriented actors to challenge the fossil orientation of the IEA, conflict management remained cooperative, and its outcomes further integrated the global energy regime complex. By uncovering the reasons for this outcome, this case provides a first step toward theorizing the conditions under which IICs may be managed cooperatively.

Initially, the IEA was the focal institution for global energy governance. It was created in response to the 1973 Oil Crisis, aiming to ensure a steady amount of oil supply (Willrich and Conant, 1977). Already beginning in the 1980s, however, dissatisfaction grew with the IEA’s predominant focus on and promotion of fossil energy sources. Germany and other states with the agenda to push renewable energy internationally criticized the IEA with growing intensity. There was dissatisfaction with “the lasting contradiction between the perceived dangers resulting from the use of nuclear and fossil fuel energy and the dearth of political initiatives to introduce renewable forms of energy with which to avert those dangers.”¹²

With the IEA remaining unresponsive, a coalition of like-minded states decided to create an alternative institution promoting the global use of renewable energy sources (Wright, 2011). IRENA was established in January 2009 with the sole objective to “promote the widespread and increased adoption and the sustainable use of all forms of renewable energy.”¹³ Its objectives squarely contradict those of the IEA (van de Graaf, 2013). As a consequence, major industrialized states and emerging economies took incompatible positions on which institutional goal to prioritize (Faude and Fuß, 2020: 277–279). The main supporters of IRENA had thus strategically created an interface conflict with the goal to challenge the IEA’s policy orientation. Both the IEA leadership and its main supporting member states were initially strongly opposed to IRENA (Esu and Sindico, 2016; van de Graaf, 2015). For instance, the United States, Australia, Japan, and the United Kingdom first sought to limit IRENA’s budget “in order to protect IEA’s role as the central energy organization of the OECD states, but also because of own energy interests and preferences for certain consumption paths and patterns” (Roehrkasten and Westphal, 2013: 13).

Despite the instrumental origin of the conflict and an initial period of mistrust and non-cooperation among the two organizations, conflict management eventually became cooperative. From the perspective of IRENA supporters, the competitive creation soon seemed to pay off as the IEA apparently felt the need to make some policy adjustments in favor of renewables: In September 2008, the IEA upgraded its renewable energy unit into a division, staffed by nine full-time analysts, as a sign that the agency was aware of the urgent need to accelerate the large-scale diffusion of renewable energy technologies to reduce carbon emissions. In addition, the IEA has expressed itself unusually positively about solar energy in two reports published in the mid-2010s (van de Graaf, 2013).

Following the rhetorical shift on the side of the IEA, the two organizations signed a partnership agreement in January 2012, including a letter of intent identifying potential areas of cooperation such as in technology and innovation or the sharing of renewable energy data (Urpelainen and van de Graaf, 2015). On 19 April 2019, the IEA and IRENA signed a memorandum of understanding, which includes joint activities in line with both agencies’ mandates (IRENA, 2019). Hence, the deliberately created interface conflict pitting the colliding norms of IEA and IRENA against each other has been managed cooperatively over time. Eventually, both organizations signed agreements to “foster cooperation and develop synergies rather than conflict.” (Esu and Sindico, 2016: 242)

How was this possible? Why did IEA and its most powerful member state supporters not resist IRENA’s entry to the governance sphere more forcefully, given that it challenged its focal position in the regime complex and sought to subvert the prevalence of conventional energy sources that IEA was established to protect (Fuß, 2023: 138; von Allwörden and Schuette, 2024). The main reason is that the normative position promoted by the supporters of IRENA resonated increasingly well with trans- and international audiences amid growing awareness of a global climate emergency linked to the consumption of fossil energy. Put differently, the successful delegitimation of IEA’s uncompromising opposition to renewable energy eventually “forced” the IEA to show openness to the political goals of IRENA as it otherwise faced a legitimacy crisis. One key development in this direction was the United States’ accession to IRENA in 2011 (U.S. Department of State, 2011). For the Obama administration that entered into office in 2009, supporting and later joining IRENA was a welcome opportunity to align with the normative stance in favor of renewables and sending a signal of global support for multilateral efforts to curb climate change (Colgan et al., 2012). As a core member of IEA simultaneously supporting IRENA, the United States had a keen interest in reducing the incompatibility between the two and facilitated IEA’s cooperative stance.

In sum, the global energy regime complex went from a phase of disorder through cooperative conflict management to coordination and even integration. The case study of IEA versus IRENA demonstrates that not all IICs lead to non-cooperative management attempts and resulting fragmentation. What becomes clear is that IICs may lead to cooperative management if the conflict is created for the purpose of achieving certain policy goals that have normative appeal and are thus seen as legitimate among larger audiences. To the extent that such a challenge is politically successful and creates traction, there are incentives for the supporters of the challenged institution to accommodate and adapt the institution. They will want to keep the original institution as it still better reflects their interests, but they will also seek to minimize the inter-institutional contradictions with the challenger to (re-)gain legitimacy among critical audiences.

The benign cluster: human rights versus investment law in Zimbabwe

Zimbabwe’s post-2000 land reform gave rise to an interface conflict between human rights and international investment law. In general, human rights law and investment law are relatively distinct fields. While human rights law commonly imposes obligations on state rather than non-state actors, Bilateral Investment Treaties (BITs) focus primarily on the protection of the rights of private foreign investors (Taillant and Bonnitcha, 2011). Under BITs, states usually delegate the legal power to resolve potential investor-state disputes to ad hoc arbitration courts under the Convention of the Settlement of Investment Disputes Between States and Nationals of Other States (ICSID Convention). While local communities in host states are often directly affected by impacts of foreign direct investment (FDI), they do not normally have access to arbitral tribunals except through third-party amicus submissions (Steininger, 2018; Vadi, 2019).

The roots of the interface conflict pitting international investment law against human rights, especially the rights of indigenous people, can be traced to Zimbabwe’s colonial era before 1980. Then-Southern Rhodesia was under British rule and consisted mainly of territories seized by the British South Africa Company, whose post-independence successor, Border Timbers Limited (BTL), was protected by a BIT between the United Kingdom and Zimbabwe, same as other foreign nationals such as the von Pezold family whose investments in Zimbabwean agricultural estates since 1988 were protected by BITs with Germany and Switzerland. These private actors thus had legal claims to property in Zimbabwe.¹⁴ At the same time, however, displaced indigenous communities returning to their former homelands claimed that the territories legally owned by foreign investors were originally theirs according to traditional—if unwritten—law and that it was their indigenous right to be adequately compensated (Krajewski, 2017). This norm collision was fully activated and turned into a manifest interface conflict when the government of Zimbabwe nationalized these properties in 2005 with the justification to enhance the indigenous population’s access to land and to rectify colonial injustice. In response, the von Pezolds and others filed a claim for ICSID arbitration under the respective BITs arguing that the nationalization represented undue expropriation.

Arguably, none of the conflict parties instrumentally created a collision between investment law and human rights with the goal of challenging an unwanted norm or institution in principle. It resulted from two parallel historical processes that none of the affected parties had direct control over: decolonization and the progressive expansion of human (and especially indigenous) rights norms. Within the concrete context of Zimbabwe, the two norm sets suggested different ways of dealing with the issue of expropriation and restitution. Accordingly, the Mugabe government justified deviations from investment protection norms as necessary and legitimate to honor norms of indigenous empowerment (Tzouvala, 2022: 228–230). Conversely, the von Pezolds were dissatisfied with their expropriation and sought to rectify it by recourse to applicable investment protection law. For these reasons, the interface conflict appears mainly as non-instrumental (and is coded as such in the ICD (Zürn et al., 2024)). However, reflecting our understanding of real-world cases falling on a spectrum between entirely instrumental and entirely non-instrumental conflicts, the present case is certainly not the “ideal” NIIC. After all, it also seems that the Mugabe government, acting out of anti-colonial motivation, deliberately sought to mobilize indigenous rights to weaken international investment protection. A certain degree of instrumentality—at least by one conflict party—thus needs to be acknowledged.

Broadly in line with our conflict model, the mainly non-instrumental nature of the interface conflict provoked conflict management that was mainly cooperative, but also showed traits of non-cooperative behavior. The key reason to regard conflict management as cooperative is that, throughout the phases of conflict management, the conflict parties largely stuck to procedural norms as provided by third-party arbitration and thus showed a general willingness to accommodate the other position somewhat. While the conflict parties were not ready to find mutually agreeable solutions in negotiations, they submitted the core of the norm dispute to a third party for adjudication and followed its procedural requirements. While ICSID arbitration is skewed in favor of investment law—and thus often in favor of investor interests—which makes it potentially undesirable from a human rights perspective, the conflict parties still chose not to simply impose their political interest uncooperatively. Instead, they committed to institutionalized procedures allowing for outcomes they did not fully control and that might go against their interests.

In 2010, the von Pezold family and BTL sued Zimbabwe under its BITs with Switzerland and Germany. They challenged Zimbabwe’s actions during the land reform program, in particular the expropriation of their property as well as the government’s failure to protect them from vandalism, violent riots, and illegal settlers on their land. The government of Zimbabwe argued that the expropriation of the German and Swiss property represented proportionate conduct under the margin of appreciation.¹⁵ In May 2012, the chiefs of four indigenous communities of Chimanimani, along with the European Centre for Constitutional and Human Rights (ECCHR), submitted a joint petition to participate in the proceedings as amici curiae on behalf of the host state (Emmert and Esenkulova, 2018). They declared that the ICSID could not reach a decision without taking the internationally recognized indigenous peoples’ right to self-determination and their land rights into account. To do so “would be to produce or make inevitable a violation of (our) fundamental human rights under international law” (ECCHR, 2012: 8). While hinting at a possibility for human rights law to be applicable in ICSID arbitration, the tribunal eventually rejected the amicus curiae application, noting that potential violations of international human rights norms pertaining to indigenous people fell outside the scope of the case as constituted by the parties.¹⁶

Zimbabwe lost the dispute in July 2015 on the grounds that the expropriation had uniquely targeted foreign and white landowners (see Tzouvala, 2022 for a critique). The arbitrators found that the acquisition of the property without compensation amounted to expropriation and a violation of fair and equitable treatment standards and ordered the government of Zimbabwe to return the territory to the von Pezolds. In addition, it was ordered to pay US$65 million plus interest for the lost value.¹⁷ The state challenged both awards but lost the appeals in November 2018.¹⁸

Despite a letter of assurances issued by the Zimbabwean government in March 2016 pledging to honor the arbitral awards if not annulled, the state has yet to implement the ICSID decision.¹⁹ The claimants currently seek to enforce the awards in other ICSID member states, such as the United States and Malaysia, working toward the seizure of Zimbabwean assets abroad.²⁰ There thus seems to be a non-cooperative turn in conflict management by Zimbabwe on the final stretches of the process which puts a question mark behind the government’s cooperative approach overall. Yet after the resignation of Robert Mugabe as President in 2017, Zimbabwe has shown a strong willingness to come to terms with the legal and normative fall-out of the land reforms through political compromise. Under President Mnangagwa, the country formally discontinued the practice of seizing land from predominantly white farmers and quickly started negotiations on compensation for evicted farmers. In July 2020, the government reached a historic deal with the Commercial Farmers Union representing over 4,000 former white farmers to pay a total of USD 3.5 billion for lost land and income (Kwaramba, 2020). Overall, conflict management can be described as “mainly cooperative”—a somewhat “tainted” outcome reflecting the equally “tainted” input of a non-ideal NIIC.

The conflict management outcome was relatively one-sided and distributed costs and gains quite asymmetrically. The episode of conflict management also did not substantively resolve the collision between human rights and investment law. However, the legal proceedings that both sides subscribed to provide at least some clarity about the relationship of the norms in question, if only by showing that investment rules take precedence according to the ICSID decision. In addition, the arbitral analysis also provided hints on the procedural and substantive conditions under which indigenous rights could be applicable in ICSID proceedings. Taking into account the attempts at a political settlement, the interface conflict pitting investment rules against norms of political self-determination and indigenous rights led from an initial moment of uncertainty and disorder among the overlapping institutions to emerging patterns and rules of engagement to accommodate the antagonistic normative objectives. More broadly, the case also added to political discussions at the level of states and IOs to better reconcile investment and human rights law in general, and indigenous rights in particular (Coleman et al., 2020). Notably, there seems to be an increase in the number of cases in which arbitrators confronted with collisions of indigenous rights and investment law seek to balance the different values rather than myopically stick to the single applicability of one set of rules (Vadi, 2019). Arguably, this interface conflict thus eventually led to increased coordination in the regime complex.

Conclusion

NIICs are mostly benign. They are handled cooperatively and lead to integrative effects in regime complexes. By contrast, IICs show a mixed pattern. Roughly, two-thirds of them are malign (leading to non-cooperative conflict management and fragmentation), while one-third appear to be hybrid (evolving from non-cooperative to cooperative management and eventually leading to more coordination). We conclude that the often-opposed strands of scholarly literature on regime complexity and legal fragmentation—one rather optimistic, one rather pessimistic about prospects of cooperation and governance capacities—point, in essence, to two substantially different dynamics of interface conflicts.

NIICs are unavoidable side-effects of institutional differentiation. Contrary to norm conflicts in the nation-state, however, they cannot be handled in a constitutionalized mode. Instead of reverting to a hierarchy of legal principles, conflict management prevails in a horizontal mode, either in form of legal entanglement or integrative bargaining. The effects of this type of interface conflict do not necessarily improve the global governance system, but they can be considered benign. IICs, on the other hand, are mainly utilized to foster the interests of the more powerful rule-makers. In the majority of cases, such interface conflicts are detrimental to an integrated global governance system and can thus be considered “malign.”

Based on our analysis, we cannot say for sure whether these malign interface conflicts are an inherent part of growing institutional density in the absence of a constitutionalized political system (via the provision of institutional opportunities) or whether they are just another tool in the struggle about global governance and a liberal world order that takes place independently of interface conflicts. If anything, our cases support the latter interpretation. It stands in line with Clark’s (2021) argument that pooling and coordination between IOs take place when member states are geopolitically aligned, while competition prevails when member states show significant political differences. When IOs and their underlying norms are pitted against each other, the drivers often seem to be member states that stand in political opposition to each other. Malign interface conflicts seem to be an instrument to reject strong forms of global governance, while benign interface conflicts happen inadvertently and are curbed in their effects by actors interested in maintaining global governance arrangements. Accordingly, there would be less IICs if global governance was less contested in general. Interface conflicts, in this view, are not the causes of contestation, but a means to express it.

Our observations invite further research in at least two directions. First, the hybrid cluster of interface conflicts needs further exploration. While it may be theoretically plausible that even deliberate IICs are sometimes managed cooperatively and have integrative effects on the regime complex as a whole, we still require a better understanding of the conditions under which a conflict’s non-cooperative impetus may be overcome to lead to cooperative forms of management. The case study involving IEA and IRENA provided first insights, but a more encompassing analysis of such hybrid cases is warranted to corroborate and expand on them. Second, the expanded cluster analysis suggested that whether an interface conflict occurs within one or across two or more spheres of authority might be systematically relevant for the benign and hybrid clusters. It would be worthwhile to study if and through what mechanisms this variable shapes those interface conflicts and how it contributes to management types and ordering effects. For example, why it is particularly often conflicts across spheres that initially bring about non-cooperative management attempts which then turn cooperative, is an empirical puzzle to be explored.

Footnotes

Acknowledgements

We thank our colleagues in the Research Group “Overlapping Spheres of Authority and Interface Conflicts in the Global Order (OSAIC)” whose collective output this paper builds on. The paper idea goes back to a presentation by Michael Zürn at the OSAIC final conference in December 2021. Earlier versions were presented at the 2022 American Political Science Association Congress in Montreal, the 2023 International Studies Association Congress also in Montreal, the 2023 German International Relations Section Conference in Friedrichshafen, and the workshop “The Evolution of International Regime Complexes” at Technical University Munich in December 2023. We thank the participants in these forums and in particular our discussants Orfeo Fioretos, Sören Stapel, Thomas Sommerer, and Miles Kahler for their important comments and feedback. A special thank you goes to Frederick Pena Sims and Constance Nothomb for research assistance.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The project received funding from the German Research Foundation (DFG) under grant agreement FOR 2409.

Ethical considerations

Not applicable

Consent to participate

Not applicable

Consent for publication

Not applicable

ORCID iDs

Christian Kreuder-Sonnen

Alexander Schmotz

Michael Zürn

Data availability

The Interface Conflicts 2.0 Dataset is publicly available at https://doi.org/10.7802/2716. Replication code and alternative cluster analyses can be found in the online appendix.

Notes

Author biographies

Christian Kreuder-Sonnen is Professor of Political Science and International Organizations at Friedrich Schiller University Jena and co-director of the MA International Organizations and Crisis Management (IOCM). In the academic year 2024–2025, he was a Jean Monnet Fellow at the European University Institute.

Alexander Schmotz is a Senior Fellow in the research unit Global Governance at the WZB Berlin Social Science Center. He works on authoritarian international politics and international drivers of democratic backsliding, primarily using quantitative methods.

Michael Zürn is Director of the research unit Global Governance at the WZB Berlin Social Science Center and Professor of International Relations at the Free University Berlin. From 2019 to 2023, he was co-spokesperson of the Cluster of Excellence “Contestations of the Liberal Script” (SCRIPTS).

References

Alter

Meunier

(2009) The politics of international regime complexity. Perspectives on Politics 7(1): 13–24.

Axelrod

Keohane

(1985) Achieving cooperation under anarchy: strategies and institutions. World Politics 38(1): 226–254.

Barakat

(1973) The Palestinian refugees: an uprooted community seeking repatriation. International Migration Review 7(2): 147–162.

Beaumont

Holmes

(2018) US confirms end to funding for UN Palestinian refugees. The Guardian, 31 August. Available at: https://www.theguardian.com/world/2018/aug/31/trump-to-cut-all-us-funding-for-uns-main-palestinian-refugee-programme (accessed 10 August 2022).

Benvenisti

Downs

(2007) The empire’s new clothes: political economy and the fragmentation of international law. Stanford Law Review 60(2): 595–632.

Birkenkötter

(2020) International law as a common language across spheres of authority? Global Constitutionalism 9(2): 318–342.

Brosig

(2015) Cooperative Peacekeeping in Africa: Exploring Regime Complexity. London: Routledge.

Clark

(2021) Pool or duel? Cooperation and competition among international organizations. International Organization 75(4): 1133–1153.

Coleman

Cordes

Johnson

(2020) Human rights law and the investment treaty regime. In: Deva

Birchall

(eds) Research Handbook on Human Rights and Business. London: Edward Elgar Publishing, pp. 290–314.

10.

Colgan

Keohane

van de Graaf

(2012) Punctuated equilibrium in the energy regime complex. Review of International Organizations 7(2): 117–143.

11.

Drezner

(2009) The power and peril of international regime complexity. Perspectives on Politics 7(1): 65–70.

12.

Drezner

(2013) The tragedy of the global institutional commons. In: Goldstein

Finnemore

(eds) Back to Basics: State Power in a Contemporary World. New York: Oxford University Press, pp. 280–310.

13.

ECCHR (2012) Human Rights Inapplicable in International Investment Arbitration? A commentary on the non-admission of ECCHR and Indigenous Communities as Amici Curiae before the ICSID tribunal. Available at: https://www.ecchr.eu/fileadmin/Kommentare_Konferenzberichte_Weiteres/Kommentar_ICSID_tribunal_-_Human_Rights_Inapplicable.pdf (accessed 11August 2022).

14.

Eilstrup-Sangiovanni

(2022) Ordering global governance complexes: the evolution of the governance complex for international civil aviation. The Review of International Organizations 17(2): 293–322.

15.

Eilstrup-Sangiovanni

Westerwinter

(2022) The global governance complexity cube: varieties of institutional complexity in global governance. Review of International Organizations 17(2): 233–262.

16.

Elster

(1989) Nuts and Bolts for the Social Sciences. Cambridge; New York: Cambridge University Press.

17.

Emmert

Esenkulova

(2018) Balancing investor protection and sustainable development in investment arbitration–Trying to square the circle? DOI: 10.2139/ssrn.3260265.

18.

Esu

Sindico

(2016) IRENA and IEA: moving together towards a sustainable energy future—competition or collaboration? Climate Law 6(3–4): 233–249.

19.

European Parliament (2018) UNRWA: MEPs debate US decision to cut funding for UN agency for Palestinian refugees. Available at: https://www.europarl.europa.eu/news/en/headlines/world/20180927STO14536/unrwa-meps-debate-us-decision-to-cut-funding-for-un-agency (accessed 10 August 2022).

20.

Everitt

Landau

Leese

, et al (2011) Cluster Analysis. Hoboken, NJ: Wiley.

21.

Faude

Fuß

(2020) Coordination or conflict? The causes and consequences of institutional overlap in a disaggregated world order. Global Constitutionalism 9(2): 268–289.

22.

Flonk

Jachtenfuchs

Obendiek

(2020) Authority conflicts in internet governance: Liberals vs. sovereigntists? Global Constitutionalism 9(2): 364–386.

23.

Fuß

(2023) The Survival of Focal Organizations: Unpacking Response Strategies of Existing International Organizations to New International Organizations. Berlin: Freie Universität Berlin.

24.

Fuß

Kreuder-Sonnen

Saravia

, et al (2021) Managing regime complexity: introducing the interface conflicts 1.0 dataset. Available at: https://bibliothek.wzb.eu/pdf/2021/iv21-101.pdf (accessed 31 July 2025).

25.

Gehring

Faude

(2014) A theory of emerging order within institutional complexes: how competition among regulatory international institutions leads to institutional adaptation and division of labor. Review of International Organizations 9(4): 471–498.

26.

Gholiagha

Holzscheiter

Liese

(2020) Activating norm collisions: interface conflicts in international drug control. Global Constitutionalism 9(2): 290–317.

27.

Gómez-Mera

(2016) Regime complexity and global governance: the case of trafficking in persons. European Journal of International Relations 22(3): 566–595.

28.

Gower

(1971) A general coefficient of similarity and some of its properties. Biometrics 27(4): 857.

29.

Hehir

(2013) The permanence of inconsistency: Libya, the security council, and the responsibility to protect. International Security 38(1): 137–159.

30.

Henning

Pratt

(2023) Hierarchy and differentiation in international regime complexes: a theoretical framework for comparative research. Review of International Political Economy 30(6): 2178–2205.

31.

Hofmann

(2009) Overlapping institutions in the realm of international security: the case of NATO and ESDP. Perspectives on Politics 7(1): 45–52.

32.

IRENA (2019) IRENA and IEA strengthen cooperation for a secure and sustainable energy future. Press Release, 9 April. Available at: https://www.irena.org/news/pressreleases/2019/Apr/IRENA-and-IEA-strengthen-cooperation-for-a-secure-and-sustainable-energy-future (accessed 12 February 2023).

33.

Kaufman

Rousseeuw

(1990) Finding Groups in Data: An Introduction to Cluster Analysis. Hoboken, NJ: Wiley-Interscience.

34.

Keohane

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

35.

Keohane

Victor

(2011) The regime complex for climate change. Perspectives on Politics 9(1): 7–23.

36.

Krajewski

(2017) Ensuring the primacy of human rights in trade and investment Policies: model clauses for a UN Treaty on transnational corporations, other businesses and human rights. Available at: https://www.cidse.org/wp-content/uploads/2017/03/CIDSE_Study_Primacy_HR_Trade__Investment_Policies_March_2017.pdf (accessed 11 August 2022).

37.

Kreuder-Sonnen

Zürn

(2020) After fragmentation: norm collisions, interface conflicts, and conflict management. Global Constitutionalism 9(2): 241–267.

38.

Krisch

(ed.) (2021) Entangled Legalities beyond the State. Cambridge: Cambridge University Press.

39.

Krisch

Corradini

Reimers

(2020) Order at the margins: the legal construction of interface conflicts over time. Global Constitutionalism 9(2): 343–363.

40.

Kwaramba

(2020) Zimbabwe: president seals historic land deal. Available at: https://allafrica.com/stories/202007300206.html (accessed 27 May 2024).

41.

Langlet

Vadrot

(2024) Negotiating regime complexity: following a regime complex in the making. Review of International Studies 50(2): 231–251.

42.

Lesage

van de Graaf

(2013) Thriving in complexity? The OECD system’s role in energy and taxation. Global Governance: A Review of Multilateralism and International Organizations 19(1): 83–92.

43.

Moe

Geis

(2020a) From liberal interventionism to stabilisation: a new consensus on norm-downsizing in interventions in Africa. Global Constitutionalism 9(2): 387–412.

44.

Moe

Geis

(2020b) Hybridity and friction in organizational politics: new perspectives on the African security regime complex. Journal of Intervention and Statebuilding 14(2): 148–170.

45.

Morin

Pauwelyn

Hollway

(2017) The trade regime as a complex adaptive system: exploration and exploitation of environmental norms in trade agreements. Journal of International Economic Law 20(2): 365–390.

46.

Morse

Keohane

(2014) Contested multilateralism. Review of International Organizations 9(4): 385–412.

47.

Orsini

Morin

J-F

Young

(2013) Regime complexes: a buzz, a boom, or a boost for global governance? Global Governance 19(1): 27–39.

48.

Pratt

(2018) Deference and hierarchy in international regime complexes. International Organization 72(3): 561–590.

49.

Raustiala

(2013) Institutional proliferation and the international legal order. In: Dunoff

Pollack

(eds) Interdisciplinary Perspectives on International Law and International Relations: The State of the Art. Cambridge: Cambridge University Press, pp. 293–320.

50.

Raustiala

Victor

(2004) The regime complex for plant genetic resources. International Organization 58(2): 277–309.

51.

Rittberger

Zürn

(1990) Towards regulated anarchy in East-West relations. In: Rittberger

(ed.) International Regimes in East West Politics:. London: Pinter, pp. 9–60.

52.

Roehrkasten

Westphal

(2013) IRENA and Germany’s foreign renewable energy policy: aiming at multilevel governance and an internationalization of the Energiewende? Available at: https://www.swp-berlin.org/publications/products/arbeitspapiere/Rks_Wep_FG08WorkingPaper_2013.pdf

53.

Snidal

(1985) The game theory of international politics. World Politics 38(1): 25–57.

54.

Steininger

(2018) What’s human rights got to do with it? An empirical analysis of human rights references in investment arbitration. Leiden Journal of International Law 31(1): 33–58.

55.

Taillant

Bonnitcha

(2011) International investment law and human rights. In: Gehring

Segger

M-C

Newcombe

(eds) Sustainable Development in World Investment Law. Alphen aan den Rijn: Kluwer Law & Business, pp. 57–80.

56.

Tzouvala

(2022) Invested in Whiteness: Zimbabwe, the von Pezold arbitration, and the question of race in international law. Journal of Law and Political Economy 2(2): 226–251.

57.

UN News (2018) Donors come to UN agency’s aid with extra $118 million to help Palestine refugees. Available at: https://news.un.org/en/story/2018/09/1021202 (accessed 24 July 2025).

58.

UNRWA (2019) Overwhelming vote to extend UNRWA mandate at United Nations fourth committee meeting. Available at: https://www.unrwa.org/newsroom/press-releases/overwhelming-vote-extend-unrwa-mandate-united-nations-fourth-committee (accessed 10 August 2022).

59.

Urpelainen

van de Graaf

(2015) Your place or mine? Institutional capture and the creation of overlapping international institutions. British Journal of Political Science 45(4): 799–827.

60.

U.S. Department of State (2011) The United States joins the International Renewable Energy Agency (IRENA). Available at: https://2009-2017.state.gov/r/pa/prs/ps/2011/03/157728.htm (accessed 19 March 2024).

61.

Vadi

(2019) Natural resources and indigenous cultural heritage in international investment law and arbitration. In: Miles

(ed.) Research Handbook on Environment and Investment Law:. Cheltenham: Edward Elgar, pp. 464–479.

62.

van de Graaf

(2013) Fragmentation in global energy governance: explaining the creation of IRENA. Global Environmental Politics 13(3): 14–33.

63.

van de Graaf

(2015) The IEA, the new energy order and the future of global energy governance. In: Lesage

van de Graaf

(eds) Rising Powers and Multilateral Institutions. London: Palgrave Macmillan, pp. 79–95.

64.

von Allwörden

Schuette

(2024) Internationale Organisationen am Existenz-Limit: ‚IO-Survival Politics‘ als ‚Kontestationsmanagement‘. Zeitschrift für Internationale Beziehungen 32(1): 43–60.

65.

WAFA (2019) President Abbas welcomes UN resolution extending UNRWA mission by 3 years. Available at: https://english.wafa.ps/page.aspx?id=zwXnova114431166696azwXnov (accessed 19 March 2024).

66.

Ward

(1963) Hierarchical grouping to optimize an objective function. Journal of the American Statistical Association 58(301): 236–244.

67.

Williams

Boutellis

(2014) Partnership peacekeeping: challenges and opportunities in the United Nations-African Union Relationship. African Affairs 113(451): 254–278.

68.

Willrich

Conant

(1977) The international energy agency: an interpretation and assessment. The American Journal of International Law 71(2): 199–223.

69.

Wright

(2011) The international renewable energy agency: a global voice for the renewable energy era? Renewable Energy Law and Policy Review 2(4): 251–268.

70.

Zelli

(2011) Regime conflicts and their management in global environmental governance. In: Oberthür

Stokke

(eds) Managing Institutional Complexity:. Cambridge, MA; London: The MIT Press, pp. 199–226.

71.

Zürn

(1992) Interessen und Institutionen in der internationalen Politik: Grundlegung und Anwendungen des situationsstrukturellen Ansatzes. Opladen: Leske + Budrich.

72.

Zürn

Kreuder-Sonnen

Fuß

(2024) Interface conflicts 2.0. GESIS data archive. DOI: 10.7802/2716.

The complex effects of regime complexity: a conflict model

Abstract

Keywords

Introduction

A conflict model of regime complexity

Quantitative analysis: origin, management, and order effects across conflicts

The interface conflicts dataset

Cluster analysis

Simple cluster solution

Extended cluster solution

Qualitative analysis: from conflict to management and order effects within cases

The malign cluster: pitting UNHCR against UNRWA norms

The hybrid cluster: competing norms in global energy governance

The benign cluster: human rights versus investment law in Zimbabwe

Conclusion

Footnotes

Acknowledgements

Funding

Ethical considerations

Consent to participate

Consent for publication

ORCID iDs

Data availability

Notes

Author biographies

References