Abstract
Multilateral cooperation in international organisations is characterised by regime complexity. The literature usually adopts a policy-focused perspective studying the properties, effects, and dynamics within given regime complexes for different policy areas. Yet few accounts of why states drive regime complexity have been provided in the literature. Therefore, we adopt a state-focused perspective and observe how states differ in the extent to which they foster complexity through overlapping memberships and policy competencies in international organisations. In order to explain this variation, we extract state motivations from the regime complexity literature, but also incorporate the role of geopolitical opportunity structures for complexity as well as interactions between both elements. The empirical analysis reveals that the power to pursue self-interests leads to duplicated policy competencies, whereas duplicating international organisation memberships by creating new international organisations or joining existing ones is costly and a less favoured route towards pursuing substantive gains. The motivation to gain external reputation also positively influences the overlap in membership and policy competencies. Moreover, the number of neighbouring states and the disappearance of deep-rooted ideological cleavages are important opportunity structures for states furthering complexity. Opportunity structures also reinforce the positive effect of power to pursue self-interests and external reputation motivations on complexity. Thus, we contribute to regime complexity research in showing that not all states equally foster regime complexity and this relationship is dependent on a specific context.
Keywords
Introduction
States have created international organisations (IOs) as arenas in which they could collectively address pressing problems. Since the end of World War II (WWII), states participate in ever more IOs and equip these organisations with an increasingly broad range of policy competencies, in areas as diverse as trade, security, human rights or the environment. Thereby, states contribute to a rise in regime complexity. Raustiala and Victor (2004: 279) define regime complexity as an ‘array of partially overlapping and nonhierarchical institutions governing a particular issue-area. Regime complexes are marked by the existence of several legal agreements that are created and maintained in distinct fora with participation of different sets of actors’.
From a state-perspective, creating complexity is puzzling. On the one hand, joining multiple IOs with at least one common policy competence requires material and political resources in the form of negotiation costs, membership fees, and sovereignty costs. On the other hand, complexity potentially reduces the efficiency and effectiveness of IOs through a waste of resources and an increase in transaction costs (Bond, 2010; Brosig, 2011), incompatible policy output and outcomes (Gebhard and Galbreath, 2013; Gómez-Mera, 2015), a weakened sense of legal obligation (Drezner, 2009), and non-compliance (Panke and Stapel, 2018b). Thus, states that are strongly exposed to complexity are unlikely to reap the full benefits from cooperation.
In order to shed light on the puzzle of complexity, we adopt a state-centred perspective. The negative externalities of regime complexity are likely to increase the more overlapping organisations a state has joined and the more competencies a state covers simultaneously in its different IOs. Accordingly, we examine why states join multiple IOs that have at least one policy competency in common but are not in a hierarchical relationship (overlapping membership) as well as why states equip their IOs with duplicated policy competencies (overlapping competencies). In other words, by adopting a state-centred rather than an organisation-centred perspective, complexity can be broken down into two components; overlapping memberships as well as overlapping policy competencies, both of which could endanger the effectiveness of cooperation beyond the nation-state. Distinguishing between membership duplication and policy competency duplication as two dimensions of complexity, we pose the following research question: Why do states further complexity through overlapping memberships and by equipping organisations with overlapping policy competencies, and why are some states more prone to fostering complexity than others?
To address this question, we first draw on regime complexity approaches to situate our argument and develop a set of hypotheses on why states foster complexity. In addition to the regime complexity literature, we also include geopolitical opportunity structures, namely the number of neighbouring states and the presence or absence of deep-rooted cleavages linked to the Cold War, as well interactions between substantive motivations and opportunities for cooperation. We expect that state motivations, such as being dissatisfied with the status quo, as well as the opportunity structures with which they are confronted (number of neighbours and presence or absence of deep-rooted cleavages) increase their inclination towards complexity, both individually and by interaction with each other.
The yearbook of IO and the Correlates of War dataset together represent the universe of cases. In order to illustrate whether and how both dimensions of complexity have evolved over time, we need to select a sample of IOs. To this end, one could select all IOs operating in one specific policy field, or a specific type of IOs (developmental IOs, task-specific IOs, global IOs, regional IOs). As we are interested in competency duplications next to membership duplications, zooming into one policy field is not sufficient. Thus, we need to select one type of IOs. We opt for regional IOs as they are usually multi-purpose organisations and engage in several policy fields (Lenz et al., 2015). We draw on a novel dataset that includes 198 states, 73 regional IOs and 344 policy competencies for the time period between 1945 and 2020. The descriptive analysis illustrates that complexity has increased over time. Both dimensions of the dependent variable ‘complexity’, the overlaps of memberships and the overlaps of policy competencies, remained limited at first but have grown over time and especially during the 1990s and 2000s. In this period, states have amplified complexity and the potential negative consequences attached to it. Moreover, states vary with regard to the number of overlapping memberships and overlapping policy competencies. While some states have duplicated more than 10 memberships and 150 individual policy competencies, others are far less inclined to foster complexity.
Afterwards, we put the theoretical expectations regarding motivations, opportunity structures, and interactions between motivations and opportunity structures to an empirical test. Given the large number of observations, we use regression analysis and plot interaction effects. The quantitative analysis reveals that powerful states further complexity with regard to the competency dimension while there is no significant effect concerning the membership dimension, indicating that creating and joining new IOs is costlier than equipping existing IOs with additional policy competencies. Reputation seeking governments join more IOs and equip them with policy competencies, by which they create overlaps in membership and competencies. In addition, there is strong evidence that geopolitical opportunity structures for cooperation also matter for the development of complexity. Especially states surrounded by many neighbours feature high levels of membership overlaps, while states not hampered by deep-rooted ideological cleavages duplicate memberships as well as competencies. There is also empirical evidence for interaction effects between motivations and opportunities as they mutually reinforce each other.
In sum, this paper contributes to regime complexity research in showing that not all states are under all circumstances equally inclined to foster complexity. Especially states that are well equipped to navigate existing complexity – powerful states – push for complexity with respect to competencies and do even more so when opportunity structures are present. Opportunity structures also matter for dissatisfied states: an increase in dissatisfaction fosters increasing complexity, the more pronounced the geopolitical opportunity structures. Likewise, reputation-seeking states that turn to IOs are also affected by opportunity structures as they are more prone to produce overlaps in both respects, the more neighbours they are exposed to and when they are not hampered by deep-rooted ideological cleavages. Moreover, this paper shows that the nature of complexity is not confined within a policy field, as states are members of numerous IOs that overlap with respect to different policy areas and not all states are members in the same IOs. Based on these findings, the concluding section discusses implications for the possibilities of effective coordination between overlapping IOs. Given the increasing regime complexity since the turn of the millennium, it is unlikely that states set up cooperation and coordination among overlapping IOs due the massive transaction costs needed. At the same time, states are unlikely to abolish IOs en masse. This suggests that overlaps persist in the near future and states need to come up with novel ways to effectively pursue their interests under the condition of high regime complexity.
Theorising complexity
Due to the increase in the number of IOs and in their policy scopes, regime complexity, which is characterised by member state and policy competency overlaps between IOs, has risen over time. As a result, research on regime complexity is thriving. 1 Scholarship is well advanced in explaining and understanding how states deal with complexity, for instance through forum-shopping (Hafner-Burton, 2009) and chess-board politics (Alter and Meunier, 2009). However, the literature is less developed in understanding the motivations and opportunities of states to foster regime complexity in the first place (exceptions include Alter and Raustiala, 2018; Benvenisti and Downs, 2007; Faude and Fuß, 2020; Panke and Stapel, 2021; Raustiala and Victor, 2004). To contribute to knowledge on states as drivers of regime complexity, this section draws on regime complexity scholarship and also includes approaches on international cooperation more generally. On this basis, we theorise how state motivations and opportunity structures as well as their interactions influence variation in overlapping competencies as well as in overlapping membership.
We start from the micro-foundation that states, as core actors in international relations, pursue their interests when joining IOs and equip them with policy mandates. They do so either in order to achieve substantive self-interests (e.g. wealth) or to increase their legitimacy or external reputation (Allison, 2008; Bach, 2016; Börzel, 2016; Herbst, 2007; Malamud, 2003; Mattli, 1999; Söderbaum, 2004). We also take into consideration that state action takes place in geopolitical contexts with different opportunity structures for cooperation, such as the number of neighbouring states and deep-rooted cleavages linked to the Cold War. Hence, we first theorise how different motivations and opportunity structures influence the extent to which states foster complexity. In the second step, we further advance the theoretical debate by developing hypotheses about interaction effects between motivations and opportunity structures.
Realist-inspired regime complexity research has argued that becoming members of more than one IO at a time and equipping additional organisations with similar policy competencies is a rational strategy for strong states to pursue. Regime complexity allows powerful states to subsequently influence policies across multiple organisations and thereby pursue their own substantive interests (Alter and Meunier, 2009; Drezner, 2009; Panke, 2020) or use their power to engage in forum-shopping for a broad range of issues (Benvenisti and Downs, 2007; Henneberg and Plank, 2020; Hofmann, 2019). While in individual cases, such as in the intellectual property domain, some small states have also used forum-shopping (Helfer, 2009), it is the powerful states that have the resources to actively engage across multiple arenas at the same time which is key to exert influence across the board. In other words, creating complexity is a way to increase the margins of state manoeuvre, for instance with respect to economic policies (Pratt, 2021; Puig, 2014). A large body of rationalist cooperation approaches points out that states as strategic rational actors are oriented towards maximising gains and are more successful to pursue their respective aims if they hold more power (e.g. Axelrod, 1984; Keohane and Nye, 1977; Mattli, 1999). This also applies to IOs. The more powerful a state is in an IO, the better its bargaining position and the more likely it succeeds in pursuing its substantive interests, for instance, because it is able to offer side-payments, create issue-linkages or threaten others with pursuing alternative outside options or engaging in unilateral activities (e.g. Moravcsik, 1998). As powerful states are more likely able to maximise their self-interests in IOs than smaller states, they should not only be highly motivated to foster complexity but can also be expected to be more often successful in doing so. Accordingly, hypothesis 1a expects: The stronger states are, the more likely they are to foster complexity with respect to overlapping memberships and overlapping policy competencies.
Yet, poor states also have self-interests when they operate in IOs (Hey, 2003; Ingebritsen et al., 2006; Panke, 2010; Steinmetz and Wivel, 2010). IOs depend on the compliance of their member states for the implementation of IO decisions and policies, yet compliance cannot per se be taken for granted by all states alike – especially if the decisions and policies overwhelmingly reflect the substantive interests of powerful states (Keohane, 1989; Lake, 2009; Moravcsik, 1998; Powell, 1999). In such a situation, poorer states can strive to extract side payments form wealthy states in exchange for compliance. In line with these substantive interests, small states may join many IOs and support the expansion of their policy scopes, thereby creating overlaps (Ngongang, 2009). Based on these considerations, hypothesis 1b stipulates: The weaker states are, the more likely they are to foster complexity with respect to overlapping memberships and overlapping policy competencies as this offers the chance to extract side-payments from larger member states.
The literature on competitive regime creation links dissatisfaction of states to the creation of regime complexity. Several contributions argue that states are more likely to be dissatisfied when their general power or wealth has increased considerably since joining an IO while their influence in the IO has not (Lipscy, 2017; Pratt, 2021; Weiffen et al., 2021). The latter is usually the case as IO rules of procedures and IO primary law cannot be changed by the state gaining general power or wealth alone as a result of which such a state does not obtain more voting shares or permanent seats in the IO. Morse and Keohane (2014), in turn, argue that states which are dissatisfied with their IO’s status quo are more likely to engage in competitive regime building in order to obtain outlets better suited for pursuing their substantive interests, while such rising powers are not expected to broadening the range of competencies of IOs in which they are member of but not satisfied. Hence, hypothesis 2 is derived: The more dissatisfied states are within an existing IO, the more prone they are to foster complexity with respect to the membership dimension, but not the competency dimension.
Hypothesis 1a expects that powerful states value complexity and create it because this multiplies the chances to successfully pursue substantive aims. Hypothesis 1b is directly competing with H1a as it expects that smaller states seek to extract side-payments from large states and thereby foster overlapping memberships and competencies. Hypothesis 2 is complementary as it focuses on rising powers that are dissatisfied with the status quo and therefore want to renegotiate regional orders. Which, if any, of these expectations holds, is an empirical question addressed in section 4.
One additional self-interest of states in IOs relates to international legitimacy (Herbst, 2007; Panke, 2021; Söderbaum, 2004). Joining additional IOs and expanding policy competencies of IOs of which their state is already a member allows governments that recently assumed office to showcase their commitment to international cooperation and to bolster their external reputation (Panke and Stapel, 2021; Pevehouse, 2005; Riggirozzi and Grugel, 2015). Thus, hypothesis 3 considers external reputation-seeking as a substantive interest of newly incumbent governments, which foster overlaps in membership and competencies. Hypothesis 3 states: The stronger the external reputation-seeking motivation of states, the more likely they are to foster complexity with respect to overlapping memberships and overlapping policy competencies.
Pursuing self-interests interests through multiplying IO membership and policy competencies presupposes that these states have the opportunity to do so, which is not usually discussed in the regime complexity literature. Opportunity structures are time-and context-dependent and can take many forms (Mucciaroni, 2001; Olsen, 2001). With respect to complexity, geopolitical opportunities to cooperate with other states in IOs are essential.
If a country is surrounded by many neighbours, it has many opportunities to interact with other countries and to engage in cooperation in multiple manners, including with different sets of states and in multiple policy realms (Keohane and Nye, 1977; Nye, 1965). These cooperation opportunities foster complexity when being realised. For instance, a state with many neighbours can create or join a higher number of organisations than a geographically more isolated state. In addition, the higher the number of neighbours, the more likely it is that a state pursues different aims in different policy fields with different partners. This, in turn, is inclined to increase the overlap of competencies. Hence, hypothesis 4 focuses on geopolitical opportunities and expects: The more neighbours a state is surrounded by, the more prone it is to complexity with respect to overlapping memberships and overlapping policy competencies.
Another type of opportunity structure relates to the dissolution of deep-rooted ideological cleavages. During the Cold War, system competition between capitalist, free-market economies and communist, centrally planned economies divided the states between two camps, led by the United States and the Union of Soviet Socialist Republics (USSR; Immerman and Goedde, 2013). Due to these ideological cleavages, opportunities to cooperate with other states, to create and join IOs and to equip them with competencies, were limited to those belonging to the same camp (Rosecrance, 1991). With the weakening and eventual disappearance of the cleavage after the end of the Cold War, the opportunity structure for cooperation changed (Hogan, 1992). As a result, states were no longer prevented from cooperating with each other on ideological grounds. When states of the formerly different camps joined each other’s IOs, complexity accelerated. On this basis, hypothesis 5 states: If strong ideological cleavages disappear, complexity increases with respect to overlapping memberships and overlapping policy competencies.
In order to contribute to the regime complexity literature, we not only include opportunity structures into the analysis, but also interaction effects between motivations and opportunities for cooperation. Apart from individually influencing complexity, opportunity structures could also interact with substantive interests. The positive effects of motivations and opportunities should reinforce each other. With an increase in motivation, the actual usage of a given opportunity is more likely, leading to more complexity. Likewise, with a given motivation and an increase in opportunities, the chances that at least one of these opportunities is actually used and brings about complexity increases. Hence, hypothesis 6 expects: The positive effect of opportunity structures on complexity is the stronger, the more pronounced states’ interests in enhancing complexity with respect to overlapping memberships and overlapping policy competencies are. As there are two different opportunities (number of neighbours, absence of cleavage) and three different motivations (maximising self-interests (H1a, H1b), dissatisfaction with the status quo (H2), reputation-seeking (H3)), there are altogether six different variants of hypothesis 6. Hypothesis 6a and 3b focus on maximising self-interests and opportunities, hypothesis 6c and 3d on dissatisfaction and opportunities, and hypothesis 6e and 3f with reputation-seeking and opportunities.
The empirical variation of complexity
According to Alter and Raustiala (2018) organisations forming a regime complex can be functionally or territorially defined. The literature usually addresses functionally-defined organisations in a particular policy field. A territorial perspective is lacking as of yet. In order to contribute towards filling this gap, we adopt a territorial approach and examine multi-purpose organisations operating in particular geographical spaces, namely regional IOs.
Regional IOs are defined as general-purpose organisations with a set of primary rules, headquarters or a secretariat, that consist of at least three states that cooperate in more than one specific policy area. Unlike global IOs, the membership criteria of regional IOs relate to geographical considerations (Panke et al., 2020). Studying regional IO overlaps provides important insights into the territorial dimension of regime complexity. Regional IOs are more prone to competency overlaps than global IOs as they are – due to their multiple-purpose character –equipped with competencies in many different policy fields. By contrast, regional IOs are less likely to face membership overlaps than global IOs because they have fewer member states. As a consequence of these two opposing characteristics, insights from regional IOs also travel to global IOs as the level of overlap is not per se higher or lower in regional IO than in global IO regime complexes.
Altogether, 73 organisations qualify as regional IOs following this definition (see Table 2 in the Appendix for an overview). 2 The new versions of the Regional Organisations Competencies (ROCO) datasets (Panke et al., 2020, version 2.0) cover all 73 regional IOs for the time period between 1945 and 2020 or since their establishment and provide information on the organisations states have joined and the policy competencies with which states have equipped their organisations. Empirically, this approach allows for including the universe of cases over an extended period of time.
The ROCO datasets retrieve information from different sources. Membership information stems from official IO sources and secondary literature. Information about policy competencies of IOs were coded on the basis of IO primary law (founding treaties, treaty changes, protocols, annexes). While policy competencies do not determine their policy-making and operational activities, IOs cannot make or enact rules and norms if they do not possess the formal competencies to do so in the first place. The dataset distinguishes between 11 different policy fields (agriculture, development, economy and trade, energy, environment, finance, good governance, health, migration, security and defence, as well as technology and infrastructure) in the internal and external realm, respectively. 3 Between 14 and 17 specific policy competencies were coded for each of these 22 policy areas. Hence, the ROCO dataset includes annual information on whether or not a particular organisation was equipped with competencies concerning 344 different polices (for a detailed presentation see Panke and Starkmann, 2019).
Regional regime complexity is based on the number of state memberships in IOs and the number of policy competencies with which the IOs are equipped. Being in more than one IO is unproblematic for the actors concerned if these IOs differ in the set-up of their policy competencies, and therefore have different scopes. Yet once a state’s IOs have one or more identical policy competencies, complexity occurs and the overlap between IOs can become problematic, potentially compromising the very effectiveness of regional governance. To address this puzzle, it is key to adopt a member state perspective and investigate how and why complexity has developed with respect to both dimensions.
First, the number of overlapping IO memberships captures the membership dimension of complexity as it focuses on the number of a state’s overlapping IOs. The number of overlapping memberships is defined as the number of IOs that a state has joined and which have at least one policy competency in common. Accordingly, the conceptual maximum number of overlapping IO memberships would be 72 as there are altogether 73 regional IOs that existed between 1945 and 2020. Yet not all IOs during the whole time period but they were either founded later (e.g. Economic Community of West African States (ECOWAS) in 1975 or Shanghai Cooperation Organization (SCO) in 2001) or they were dissolved (e.g. Southeast Asia Treaty Organisation (SEATO) or Organisation of Central American States (ODECA)). Also, in practice, not every state can join every regional IO since they follow geographical membership criteria per definition. In Africa, the number of regional IOs is 23, while there are 16 in the Americas, 18 in Asia, and 16 in Europe.
Empirically, the maximum number of overlapping memberships is 10 (Russia since 2005, see Figure 1) and the minimum is zero (Israel, North Korea, and Timor-Leste). In between are countries such as Nigeria (4 since 2001), China (4 since 2006), or the USA (6 since 1996). The United Kingdom is also in the mid-field but an interesting case because its IO membership duplications increased to 6 by the mid-1970s and declined to 2 by the year 2020 due to the dissolution of SEATO in 1977, leaving the SPC in 2005, the dissolution of the Western European Union (WEU) in 2011, and the exit of the UK from the European Union (BREXIT) in 2019.
The membership and policy dimensions of complexity for exemplary states.
Second, the policy competency dimension of complexity is represented by the number of overlapping policy competencies with which a state has equipped its IOs. This dimension captures how many of the total of 344 different IO policy competencies a state has transferred to more than one of its IOs (ROCO 2.0 dataset). Hence the conceptual maximum of overlapping competencies is 344. Empirically, the variable ranges between 0 and 180. Russia covers 180 different policy competencies in at least two of its IOs since 2014 (Figure 1). The United States, Finland, Kirgizstan, Rwanda or Thailand can also be found at the upper end of the empirical spectrum (with more than 100 different overlapping competencies), while states such as Switzerland, South Africa, Nepal or Paraguay are located at the lower end of the spectrum (less than 40), and others duplicate less than 20 different competencies in IOs (e.g. the United Kingdom with 15 since 2020).
In addition to the cross-country variation, there is considerable variation over time concerning both dimensions of complexity (see Figure 3 in the Appendix). While the average number of overlapping memberships was lower than one until 1974 and around one in the 1980s, it increases strongly after 1991, reaching a mean of 3.7 in 2020. Similarly, a state duplicated less than ten different competencies on average in its IOs until 1983 and this increased to more than 10 after the 1990s. The competency duplications reached an average of 61.3 different in 2018 (slightly decrease afterwards to 60.9 in 2020). Given the negative potentials of complexity, it is puzzling why states create or join IOs that overlap with respect to policy competencies in the first place, and why they have duplicated competencies in their IOs over time.
Analysis and discussion
Why do states join multiple IOs that have at least one policy competency in common (overlapping membership) and why do states equip their IOs with duplicated policy competencies (overlapping competencies)? By adopting a state-centred rather than an organisation-centred perspective, complexity can be broken down into two components: overlapping memberships and overlapping policy competencies. Both components could endanger the effectiveness of cooperation beyond the nation-state. Accordingly, the dependent variables count the number of overlapping memberships in IOs per state and year and overlapping competencies in IOs per state and year (see section 3). As both dependent variables are not continuous, ordinary least square (OLS) regressions are not suitable. Instead, the count data call for logistic regressions with negative binomial specifications as both dependent variables are over-dispersed. Due to the annual observations for each of the states, we have a cross-sectional time-series data structure. Hence, a general linear cross-sectional time-series model with a negative binomial family and a log link is the appropriate model specification. Moreover, since state observations are not independent from each other at different points in time, we use clustered standard errors on the state level. 4
The independent variable of H1a and H1b is operationalised by the power of a state, and we rely on real gross domestic product (GDP) data (in trillion US dollar, constant 2017) which we collected from Penn World Table data (Feenstra et al., 2015). Hypothesis 2 captures that rising powers, as states strongly gaining in general power or wealth, become dissatisfied in their IOs. In line with Lipscy (2017), we operationalise the extent to which states are dissatisfied with the status quo by the relative gain in economic power over time. We again used real GDP (in trillion US dollar, constant 2017) data from Penn World Table and calculated the change in GDP percentage in a state’s moving average of 10 years. Hypothesis 3 considers the external reputation-seeking motivation of states. As newer governments are expected to be more interested in gaining external recognition, we operationalise the variable by counting the years the chief executive has been in office based on information from the Database of Political Institutions (Cruz et al., 2021). To assess geopolitical opportunity structures, two factors are included. The independent variable of hypothesis 4, the number of neighbours, is operationalised by the total number of borders per state and year. The data were retrieved from the Correlates of War Project (Pevehouse et al., 2020). To measure the disappearance of ideological cleavages over time (hypothesis 5), we include a dummy variable for the post-Cold War era. To allow for the causal mechanisms to unfold, the independent variables are time lagged for 3 years in all models.
In addition, we control for two factors. First, the peacefulness of states as the absence and presence of inter-state conflicts. The information is provided by the Uppsala Conflict Data Programme. Second, we also control for the age of states. The information on the year of independence stems from official country websites, which we used to calculate the number of years the state existed during the time period of observation (1945-2020). Table 3 in the Appendix 1 provides the summary statistics and Table 4 features the marginal effects of the five explanatory variables.
In a first step, we discuss the findings for the motivation hypotheses (H1-3) and opportunity structure expectations (H4-5 respectively). We then zoom into the interaction effects (H6).
The first hypothesis (H1a) focuses on the role of state power as a means to further self-interests in and through IOs, thereby fostering complexity. Table 1 illustrates that all models feature the expected positive sign. However, an increase in power has no significant effect on overlapping IO memberships (Models 1 and 2), whereas there is a strong significant effect on overlapping policy competencies (Models 3 and 4). Hence, hypothesis 1a is supported by empirical evidence with respect to the competency dimension of complexity, but not the membership dimension of complexity. The former suggests that powerful states further the overlap of policy competencies as a means to pursue their own interests. This finding resonates well with qualitative studies of regime complexity arguing that especially powerful states benefit from complexity as it allows them to pursue their own substantive interests in different issue-areas (Alter and Meunier, 2009; Benvenisti and Downs, 2007; Drezner, 2009). The findings that power (H1a) and increases thereof (H2, see below) do not significantly translate into more duplicated memberships might be due to the fact that creating new IOs or joining additional IOs is politically and materially costly and therefore less often pursued than duplicating competencies.
Regression analysis.
Clustered standard errors in parentheses: * p < 0.05, ** p < 0.01, *** p < 0.001.
By contrast, hypothesis 1b is not supported empirically. It is not the case that weaker states are more likely to foster overlapping memberships and overlapping policy competencies. The coefficient is not negative in any of the models (Table 1, models 1-4). This suggests that weak states do not create overlaps to extract side-payments from larger member states in exchange for compliance. Empirical studies have illustrated that IO decisions and policies often reflect the interests of powerful states (Moravcsik, 1998) and that compliance with IO outputs is more likely when they are in line with the substantive interests of states (Chayes and Handler-Chayes, 1993). While weak states might seek to extract side-payments from larger states in individual instances (Ngongang, 2009; Panke, 2010), they do not systematically foster overlaps in order to use additional IOs or additional policy areas as sources to extract side-payments form powerful states in exchange for compliance. This interpretation is in line with the enforcement literature that points out that weaker states comply because they cannot afford to become subject to sanctioning (Börzel et al., 2010; Fearon, 1998).
There is no empirical evidence for the plausibility of hypothesis 2, which expected that rising powers engage in competitive regime creation. Neither in the models without control variables nor in the more extensive models an increase in relative power translates into creating IO membership overlaps (Models 1 and 2). Turning to overlapping policy competencies, hypothesis 2 did not expect a systematic positive relationship, which is in line with Models 3 and 4 as the coefficients are neither robust nor significant. Hence, being dissatisfied with the status quo does not generally lead to more overlapping memberships, while it does not systematically impact duplicated policy competencies. Although Brazil and China as rising power have been dissatisfied with the status quo and also push for institutional changes through Union of South American Nations (UNASUR) and the SCO, respectively, that result in increasing overlaps (Stephen, 2021; Weiffen et al., 2021), creating new outlets for cooperation requires considerable resource investments. Yet, against this background, it is not surprising that the quantitative analysis does not provide systematic evidence for competitive regime creation across the board.
Hypothesis 3 sheds light on reputation-seeking motivations of states and expects that younger governments are more interested in gaining international recognition by joining additional IOs or broadening the policy competencies of IOs they are already member of than governments that have been in the office for longer periods of time. All models feature the expected negative sign and are statistically significant, with the exception of model 4, which controls for additional variables and does not meet the threshold for statistical significance. Taken together, we can conclude that the empirical evidence points towards the plausibility of hypotheses 3. States that strive for international recognition join additional IOs, as several studies have evidenced (Dimitrova and Pridham, 2004; Mansfield and Pevehouse, 2006; Mehmeti, 2017). These states increase membership overlaps and, in tendency, also policy competency overlaps.
Although regime complexity approaches do not usually focus on geopolitical opportunity structures when accounting for the evolution of regime complexes, our analysis shows that these factors can be important. Table 1 illustrates that the number of neighbours has a significant and positive effect on membership overlap (Models 1 and 2), whereas there is no significant effect with respect to policy overlaps in any of the two specifications (Models 3 and 4). A higher number of neighbours increases the membership dimension of regime complexity. Hence, hypothesis 4 is empirically supported only with respect to the number of overlapping memberships. The models show a positive and significant relationship between the absence of the Cold War cleavage, on the one hand, and membership as well as the competency overlap, on the other hand. In line with hypothesis 5, the disappearance of deeply-rooted geopolitical cleavages increases both dimensions of complexity. These findings suggest that the development of regime complexes can best be understood when not only substantive motivations (Alter and Raustiala, 2018; Benvenisti and Downs, 2007; Faude and Fuß, 2020; Raustiala and Victor, 2004), but also geopolitical opportunity structures are taken into account.
Turning to the interaction effects, the generic hypothesis 6 stipulates that motivation and opportunity structures mutually reinforce each other’s effect on complexity. There is evidence in support of this expectation for all specifications. Figure 2 features the plots for the predicted margins with 95% confidence intervals of the six interaction hypotheses (H6a-f) for the membership dimension and the policy dimension of complexity, respectively.
Interaction plots.
In general, the positive effect of having power to pursue self-interests in and through IOs on complexity is further enhanced the more geopolitical opportunities a state has for cooperation. The positive effect of state power on the duplication of both the IO memberships as well as the duplication of IO policy competencies is reinforced by a higher number of neighbours and by the absence of cleavages after the end of the Cold War as both increase the opportunities for regional cooperation. These observations are in line with hypotheses 6a and 6b and apply to the competency dimension of complexity and the membership dimension of complexity.
Hypotheses 6c and 6d shed light on the interaction between being dissatisfied with the status quo and the number of neighbours and the lack of a Cold War cleavage as compared to the presence of such a cleavage, respectively. An increase in the opportunity to cooperate increases the positive effect of being dissatisfied of a state on the membership dimension of complexity and also on the competencies dimension. In other words, when a state is increasingly dissatisfied with the status quo, it is even less likely to foster complexity when it shares fewer borders with other states and during the Cold War than afterwards.
Finally, hypotheses 6e and 6f focus on the interaction effect between reputation-seeking and opportunity structures. The plots in Figure 2 support the expectation that both features are mutually reinforcing with respect to the overlaps in membership and in policy competencies. When states are increasingly focused on gaining reputation, the positive effect of opportunity structure further reinforces an increase in both dimensions of complexity. The more a state is interested in boosting its external recognition, the more likely this motivation leads to overlaps when it shares more borders with other states and after the Cold War.
In sum, substantive interests matter for states inclinations to foster complexity, but in different respects. Powerful states are inclined to broaden the policy competencies of IOs to pursue their self-interests in these areas and thereby foster policy competency overlaps. However, unlike expected by approaches on competitive regime creation, rising powers do not create more complexity as such. Yet, the external reputation seeking motivation plays out as expected. The shorter a government is in office, the more IOs it joins and the greater the policy competency overlap becomes. We also investigate geopolitical opportunity structures. This reveals that the number of neighbouring states is important for the membership dimension of complexity, while disappearance of deep-rooted cleavages after the end of the Cold War has a positive effect on overlaps of membership and competencies. Apart from these individual effects, motivations of powerful states, of rising powers and of governments that seek external recognition as well as opportunity structures in terms of a high number of neighbours as well as the absence of deep-rooted ideological cleavages are mutually reinforcing.
Conclusions
Transboundary issues call for cooperation between states, which often takes place in IOs. Over time, the number of IOs has increased. Much of the regime complexity literature studies how states navigate existing complexity and its effects. This paper sheds light on drivers of complexity and studies why not all states further regime complexity to the same extent. To do so, the paper adopts a state-centred perspective that captures that states further complexity through overlapping memberships in IOs and also by equipping their respective organisations with overlapping policy competencies. Using a novel and comprehensive database that encompasses all states between 1945 and 2020 shows that fostering complexity through duplicating policy competencies is a result of states’ power to pursue self-interests. As duplicating memberships is more costly, the positive effect of state power on this dimension of complexity is not significant but still positive. Thus, the competing expectation that small states foster overlaps to increase the prospects of extracting side-payments from rich states does not hold up to empirical scrutiny. Likewise, there is no evidence that rising powers have a systematic impact on regime complexity as such. Yet, reputation-seeking motivations increase the inclination of states to join additional IOs and equip them with additional policy competencies, which results in more pronounced overlaps. In addition, opportunity structures matter for the duplication of memberships and the duplication of policy competencies. When opportunities for multilateral cooperation increase due to the disappearance of deep-rooted ideological cleavages, states enhance regime complexity in both respects. A higher number of neighbouring states operates also as an opportunity structure, but only increases overlaps in membership. Motivations and opportunities also interact and reinforce each other’s positive effects on fostering both dimensions of complexity, overlapping IO membership and overlapping IO competencies. More neighbours as well as the disappearance of deep ideological cleavages reinforce the effect of being powerful, of being a rising power as well as of being interested to boost one’s external recognition on complexity.
Our study contributes to the regime complexity literature by adopting a novel analytical lens with respect to complexity that illuminates important aspects commonly overlooked. While the regime complexity literature sheds light on the working and interlinkages between international organisations in one policy field (e.g. climate change), we adopt a territorial perspective. In doing so, we shed light on a variety of organisations across policy fields and through the lens of member states. Thus, for each state and year, we examine two dimensions of complexity. We study in how many organisations each state is member of that have at least one common policy competence (overlapping membership) as well as how many policy competencies each state has covered more than once in its organisations (overlapping competencies).
Analysing complexity in state-containers rather than in policy-area containers allows to consider that, through multiple memberships in general-purpose organisations, difficulties linked to complexity do not affect policy outcomes in merely one policy field but in many different policy fields at once. Moreover, it can take into account that states are members in several multi-purpose IOs and often cover similar policy fields in their different organisations, each with a different set of partner states.
When many different IOs are involved for each policy area, which additionally differ from state to state as not all states are members in the same IOs, effective cooperation and coordination between such overlapping organisations becomes challenging, if not impossible. Effective cooperation and coordination would require to incorporate the numerous IOs that are linked to each other by at least one shared member and at least one common policy competency. Due to the number of interlinked organisations, transaction costs for coordination or even orchestration of these organisations would be massive. Thus, the state-centred perspective on regime complexity suggests that hierarchical modes of governance are unlikely to be employed to manage complexity as states are likely to shy away from the enormous transaction costs linked to such an endeavour. Hence, finding ways to effectively govern regime complexes remains a challenge for the future.
Empirically, our paper suggests that powerful states operating in favourable geopolitical opportunity structures strongly foster complexity, especially by expanding the policy scopes of IOs. Powerful states can also use complexity to their advantage to further their own substantive interests as they are in a good position to avoid incompatible IO norms and rules or otherwise costly policy outputs (Alter and Meunier, 2009; Benvenisti and Downs, 2007; Drezner, 2009). Moreover, they have the capacity necessary to implement and act upon outputs of the IOs they are member of (Ingebritsen et al., 2006; Panke, 2010). Accordingly, powerful states are likely to successfully navigate complexity.
In contrast, weaker states not only further and design complexity less strongly, but are also the ones that have to carry much of the negative externalities of complexity as they are not well equipped to shape the policy outputs of IOs in their daily operation in order to ensure that their interests are reflected and mismatches of IO outputs are avoided. To make matters worse, weaker states are also likely to grapple with capacity shortages to comply with the policy outputs of overlapping IOs, which becomes more problematic when the IO policy outputs are numerous and incompatible (Ingebritsen et al., 2006; Panke, 2010). In other words, inequality between states might become further reinforced in dense regime complexes.
However, this inequality is unlikely to deepen much further. While the opportunity structures in the form of the number of neighbours and the lack of deeply rooted cleavages have increased in the past, there is a levelling effect at play. First, the number of newly independent states that turn into new, additional neighbours has risen especially during two waves after the decolonisation and the dissolution of the USSR, but is no longer growing in a steady manner. Second, the disappearance of the capitalism-communism cleavage at the beginning of the 1990s was a single event, creating additional opportunities for regional cooperation in the aftermath that were not present before. As there are currently no other similarly strong geopolitical cleavages that prevent cooperation on ideological grounds, an increase in opportunity structures for regional cooperation cannot be expected at the moment. Taken together, this suggests that the effect of opportunity structures on complexity has reached its peak and is levelling off at present. Moreover, with the rise of nationalism and populism there are new countervailing tendencies at play, that can lead to state exits from regional IOs (such as the UK from the EU), which, in turn, reduces complexity.
Footnotes
Appendix
Marginal Effects.
| Variable | dy/dx | Std. err. | P > z |
|---|---|---|---|
| Dependent variable: duplication of memberships | |||
| Power to pursue self-interest | 0.0891 | 0.0477 | 0.061 |
| Dissatisfaction with IO status quo | −0.0598 | 0.0998 | 0.549 |
| External reputation motivation | −0.0190 | 0.0083 | 0.022 |
| Opportunity structure: number neighbours | 0.0877 | 0.0298 | 0.003 |
| Opportunity structure: absence of Cold War cleavage | 1.6203 | 0.1078 | 0.000 |
| Dependent variable: duplication of policy competencies | |||
| Power to pursue self-interest | 8.7200 | 2.2437 | 4.322 |
| Dissatisfaction with IO status quo | 0.3172 | 5.5930 | −10.6 |
| External reputation motivation | −0.5675 | 0.4833 | −1.51 |
| Opportunity structure: number neighbours | 2.9128 | 1.5936 | −0.21 |
| Opportunity structure: absence of Cold War cleavage | 127.8663 | 5.6117 | 116.9 |
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: We gratefully acknowledge funding by the German Research Foundation (DFG) (PA 1257/7-1; 421167407).
