1 Introduction

The current international security architecture is institutionally complex. States deal with security issues unilaterally and bilaterally, but also in regional and in international organizations.

Especially after WWII, states created several international security organizations (ISO), focusing on international peace and security issues. Among the most prominent ISOs are the Security Council (SC), the Conference on Disarmament (CD), the Organization for the Prohibition of Chemical Weapons (OPCW), the International Atomic Energy Agency (IAEA), the Arms Trade Treaty (ATT) regime, and the First Committee of the United Nations General Assembly (UNGA-C1). In these ISOs, states negotiate the international security order in creating security norms such as nuclear non-proliferation (Müller et al. 1994; Thayer 1995; Lodgaard and Maerli 2007), the responsibility to protect (Evans and Sahnoun 2002; Williams and Bellamy 2005; Cooper 2009), the regulation of arms trade and the prohibition of illicit trade with weapons and ammunition (Krause 2002; Bromley et al. 2012; Erickson 2015).

In addition to cooperating in ISOs, states also work together in regional groups and organizations in which membership is based on geographical criteria (‘regional actors’, RA). After WWII and again after the end of the cold war, the numbers of RAs have increased considerably. Today, more than 60 RAs are located all over the globe (Panke and Stapel, forthcoming). More often than not, RAs were initially created to enhance regional trade and create common markets, but their policy scopes have been expanded over time and today, 19 explicitly cover security policies (Panke and Stapel, forthcoming).

Regional actors, such as Economic Community of West African States (ECOWAS), the European Union (EU) or the Arab League, do not hold voting rights in ISOs, but are at best observers in ISOs. Nevertheless, RAs are not necessarily inactive in ISOs as their members can be active on their behalf when international security norms and rules are negotiated. Articulating collective regional positions instead of individual national ones is a means for states to increase their power in international security negotiations and increase the chances of effectively influencing negotiation dynamics and outcomes.

Although the activity of RAs in ISOs can culminate in the regionalization of today’s international security negotiations and although it can have important implications for both the effectiveness and legitimacy of the international security architecture, we do not know much about the behavior of regional actors in ISOs. Thus, this article sheds light on dynamics within today’s multi-layered institutional architecture and addresses this research question: why and when do regional actors get active in international security negotiations?

To provide an answer to this question, Sect. 2 examines how active regional actors participate in the security-focused IOs. It shows that not all regional actors are equally vocal in the negotiations, which is puzzling given that negotiation activity is important for negotiation success. To explain the variance in regional actor vocality, Sect. 3 draws on international cooperation theories and develops hypotheses about the activity of regional actors in international negotiations, which are tested with quantitative methods. It is striking that even in the state-dominated policy field of security, regional actors are getting vocal in the process of negotiating peace architectures. However, the role of regional actors varies, depending on their size, status in the ISO negotiation arena and their coverage of security policies as well as the size of the ISO and the formal decision-making rule in the international negotiation arena.

2 Regional Actors in International Security Negotiations: The Empirical Puzzle

States often cooperate on a regional basis and have formed RAs as institutionalized arenas for cooperation of at least three states with geographic proximity. Today’s more than sixty RAs are spread all over the globe (Panke and Stapel, forthcoming),Footnote 1 19 of which explicitly have competencies regarding security issues (ArabLeague, ASEAN, AU, CARICOM, CEEAC, CIS, COMESA, CSTO, ECOWAS, EU, GUAM, IGAD, NATO, OAS, OSCE, PIF, SADC, SCO, and UNASUR) (Panke and Stapel, forthcoming).

The United Nations (UN) were created in 1945 to maintain international peace and to develop and sustain friendly relations among sovereign states, as well as to foster social and humanitarian progress and to promote human rights (Cede 2001; Schlesinger 2011; Panke 2013b). According to the UN Charter, in two of the six principle UN organs, international security and disarmament issues are negotiated. These are the first committee of the United Nations General Assembly (UNGA-C1) as well as the Security Council (SC). In addition, there are several international organizations and regimes created and subsumed under the UN umbrella, which also deal with questions of international security and disarmament. These include the ATT regime, the Conference on Disarmament (CD), the Organization for the Prohibition of Chemical Weapons (OPCW), and the International Atomic Energy Agency (IAEA). In these ISOs, none of the regional actors with security-related competencies is a full member. Yet, several ROs are registered as observers (for example, the African Union in the IAEA or the EU in the UNGA’s first committee) and, therefore, have access to negotiations without being able to vote. Besides voicing regional positions directly through RA delegates, RAs can also be indirectly active, whenever one of its member states speaks on its behalf in international security negotiations.

How vocal are different regional actors and how can variation in the active participation of regional actors in ISOs be explained?

This question is important, not only since the international security architecture is largely a negotiated one; essential features have incrementally evolved through numerous international negotiations in a broad variety of institutional contexts (Buzan and Hansen 2007, 2009; Buzan and Waever 2008; Daase 2008; Rathbun 2011). Hence, this section maps the negotiation activity of all regional actors and all UN member states in the UNGA-C1, the SC, the ATT, the OPWC, the IAEA, and the CD. These ISOs all provide public access to the necessary records (protocols, verbatim records, or detailed minutes that encompass information on who took the floor during negotiations) and convene negotiations on a regular basis in the period under observation (2008 and 2012). This period is selected to avoid biases concerning external events. For the same reasons, four negotiations per yearFootnote 2 have been selected, covering the core policy area of the respective ISOs.

This article uses a content analysis to construct a large-N dataset on the active participation of states and regional actors in international security negotiations. The official negotiation records (in verbatim protocols, minutes, press releases) were hand coded. This captures how often each state or delegate of a regional actor speaks up in a negotiation, as well as whether a state exclusively expressed a national position or spoke on behalf of a regional actor. The resulting dataset includes 102 individual negotiations taking place within the six international security organizations. Together, these negotiations include more than 14,000 data points on the negotiation activity of ISO member states and regional actor delegates.

It is often argued that international security is the domain of the sovereign nation-state (Waltz 1959; Benson 2012; Collins 2013; Mangold 2013; Dannreuther 2014; Hirst 2014). Yet, it is remarkable that even in the realm of international security, regional actors get active and voice their positions in international negotiations taking place in ISOs. In the 102 international security negotiations under scrutiny, 3.38 % of all speeches made were regional in character, while in 96.62 % states expressed national positions. Table 1 illustrates that there is variation between institutional contexts. Relative to the national statements made, regional actors are most vocal in the ATT, followed by the OPCW and the first committee of the UNGA, the SC and the CD and least vocal in the IAEA.

Table 1 Absolute numbers and percentages of regional statements in ISO negotiations
Full size table

Figure 1 allows one to zoom into the distribution of regional actor statements in the negotiations taking place in the six ISOs. The boxplots show that the speeches made by regional actors are not distributed evenly across the negotiations within each of the ISOs. In ATT negotiations, the average number of regional positions voiced is six and the median is four. In the UNGA’s first committee, the mean number of regional actors statements is 2.2 and the median is 2. In the CD, the average number of regional statements is 1.6 and the median is 1.0, while there are outlier negotiations in which regional actors were considerably more active (four statements).

Fig. 1
figure 1

Boxplots of regional voices in international security negotiations

Full size image

Which regional actors are most active in international security negotiations? Not all of the 19 regional actors with security competencies participated actively in the 102 ISO negotiations (cf. Table 2). The EU, ASEAN, ECOWAS, UNASUR, and PIF have competencies in the security policy area and participate actively in ISOs. However, the Arab League, AU, CEEAC, CIS, COMESA, CSTO, GUAM, IGAD, NATO, OAS, OSCE, SADC, and SCO also have security competencies (Panke and Stapel, forthcoming), but remain silent. Moreover, some regional actors, such as Mercosur, SICA and the CoE as well as UN regional groups, such as the African Group (ARGUN), the Eastern European and Others Group (EEGUN), the Latin American and Caribbean group (LACGUN) and the Western European and Others Group (WEOGUN), voiced regional positions at least once in spite of not formally having security policy competencies. Table 2 further illustrates that the regional actor activity varies considerably, with the EU being the most vocal in international security negotiations. The ARGUN, CARICOM and Mercosur all voice positions in between 22.5 and 15.7 % of the 102 negotiations.

Table 2 Regional voices in ISO negotiations
Full size table

It does not have to be a delegate of the regional actor itself (e.g., the European Commission) who is voicing the regional position in international security negotiations, but each member state can also speak on behalf of its regional group. Out of the 4673 formal negotiation contributions made by states in the 102 international security negotiations, 158 or 3.38 % are regional in character. Thus, although all states can potentially leverage up through voicing regional positions, the frequency with which a state acts on behalf a regional actor and articulates regional positions varies considerably between countries (cf. Table 3). Antigua and Barbuda, Barbados, and Honduras are not very vocal as such, expressing only 1, 3 and 1 regional positions, respectively, but did not express exclusively national positions at all. Consequently, 100 % of their statements were regional in character. Suriname voiced two regional and one national position, while Belgium made 10 national as well as 13 regional statements in the ISOs under scrutiny.

Table 3 Regional positions expressed in ISOs by member states
Full size table

3 Theory

Why are regional actors not equally vocal in all ISOs? Why are the positions of some regional actors more often voiced than of others? Why are some states more inclined to speak on behalf of regional actors than others? To answer these questions, this section draws mainly on IR theories on cooperation and conflict to formulate testable hypotheses.

On the international level, regional actors are not equally vocal in all ISOs (cf. Sect. 2). Institutional design approaches allow one to develop hypotheses about the propensity to which international security negotiations are regionalized (e.g., Goodin 1995; Koremons et al. 2001). These approaches contend that institutional rules impact actor behavior through setting incentives and disincentives for particular pathways of action (e.g., Mitchell 1994; Panke 2006). The larger the ISOs are, the longer negotiations take when every member state voices a position. Thus, larger ISOs tend to adopt either formal rules of procedure that delimit the number and duration of speeches that can be made in international negotiations or develop informal norms on how to achieve high efficiency. One way to limit the number of positions voiced and increase the efficiency of international negotiations is to bundle individual positions in collective ones (Panke et al. 2015). Accordingly, international negotiations should attract a higher number of regional positions, the more states are member of the respective ISO (hypothesis 1). In addition, ISOs vary in the extent to which they are formally open to grant RAs formal access to negotiations, which enable RA delegates to articulate regional positions. Since formal institutional opportunities are likely to influence actor’s conduct (Jepperson 1991; Powell and DiMaggio 1991; Hall and Taylor 1996; Peters 1999; Blavoukos and Bourantonis 2010), one could expect that espcially those ISOs attract more regional negotiation contributions in international security negotiations, which are institutionally open to RAs and grant many of them formal status (hypothesis 2) (Wetzel 2011; Gehring et al. 2013; Debaere et al. 2014; Orsini 2014). Another institutional design feature of ISOs that could be important for the frequency to which they attract regional positions in negotiations is the decision-making rule (Delreux et al. 2012). When international security norms and rules are passed unanimously, each ISO member state turns de facto into a veto player that could block the decision (Goodin 1995; Tsebelis 2002; Panke 2006). While it is a sensible strategy for ISO member states in institutional negotiations that operate on the basis of majority rules to speak on behalf of their RAs to maximize their influence over international negotiation dynamics and outcomes by pointing to the number of combined votes one speaks for (Habeeb 1988; Panke 2013b), consensus ISOs do not lend themselves to collective bargaining moves. Accordingly, in ISOs in which the formal decision-making rule is majority voting, the share of regional positions voiced in international security negotiations should be higher than in ISOs with unanimity rule, as leveraging up through RAs is more effective in the former (hypothesis 3).

On the regional level, Sect. 2 illustrates that not all regional actors are equally vocal. In the last decades, RAs increased their activities beyond their own borders, which led to a vibrant field of research on regional organizations as external actors.Footnote 3 With respect to operating in IOs, research has pointed out that some RAs are formal observers in more IOs than others (Wetzel 2011; Gehring et al. 2013; Debaere et al. 2014; Orsini 2014). This also applies to IOs dealing with international security. Obtaining a formal status in an ISO is resource- and time-intensive for RAs, as they have to take the initiative to apply for accreditation as observers as a means to get formal access to international security negotiations. Since time and resources are not endless, incentives are important for political conduct (Segal and Jeffrey 2000; Ringquist et al. 2003; Thomson and Stokman 2006). It makes sense that RAs only register themselves formally at an ISO, when they have strong incentives to engage in collective action in the ISO negotiation arenas, and refrain from getting a formal status when security issues are not of high importance for the RA in question. Thus, one can expect that RAs are especially vocal in ISO negotiations, when they have a formal status within the ISO compared to no formal role in the ISO (hypotheses 4). To be able to act externally, competencies of RAs are important (Smith and Hill 2005; Kissack 2010; Jørgensen et al. 2011; Oberthür 2011; Delreux 2013). If RAs possess security competencies, this increases the chance that the RA members have already agreed on some security-related norms and rules on the regional level. Therefore, RAs with security competencies are in a good position to develop a regional position for the international norms and rules at stake, which the RA members can subsequently articulate in international security negotiations. Accordingly, if an RA also has competencies in the field of security policy, it should be more active in ISO negotiations (hypotheses 5). In addition to having incentives to become active in international security negotiations and being likely to be able to formulate regional positions, the chances for an RA to become active in an ISO also depend on RA-specific opportunity structures (Panke 2013a). Rather than RA delegates directly voicing regional positions (such as the European Commission in case of the EU), it is usually the RA member states who speak up on behalf of their regional group in international security negotiations. Thus, the fact that RAs differ in membership size is important, since larger RAs have more member states that can articulate regional positions. Hence, the more member states of an RA are also members in the ISO, the more likely it is that the RA’s position becomes articulated in the international security negotiations (hypotheses 6).

Finally, with respect to the state level, some of the most prominent theories with focus on international security are (neo)realism (Morgenthau 1948; Waltz 1959, 1979) and liberal institutionalism (Keohane 1984, 1986; Keohane and Nye 1989). Both regard states as essential actors on the international level and, therefore, provide a good starting point to theorize which states are most likely to voice regional positions in international security negotiations taking place within ISOs.

Realism and Neo-realism regard state power as an important variable, impacting state conduct. In power politics, alliances are a means to achieve either power-oriented (Morgenthau 1948) or security-oriented goals (Waltz 1979), but joining them is never an end in itself (Walt 1985, 1987, Snyder 1990). Comparing smaller states with militarily powerful states, all else being equal, the former should rely more heavily on alliances, while the latter are in a better position to act on their own (Rothstein 1963). A similar logic should also apply to international negotiations. In international security negotiations, bigger states with significant military personnel are less likely to voice regional positions, while smaller states with small armies should speak more often on behalf of RAs to gain leverage in the ISO (hypothesis 7).

According to liberal institutionalism, states cooperate in regional and international institutions when the potential vulnerability of states arising from interdependencies is high (Keohane 1989; Keohane and Nye 1989). Joining regional or international institutions is not an end in itself, but a means to cope with negative externalities that arise through interdependence in the first place through cooperation (Axelrod and Keohane 1986). Thus, states that are members of many regional organizations or groups with security focus should be more likely to express regional positions in international security negotiations to maximize their own influence in the ISO (hypothesis 8).

4 Empirical Analysis

This article examines the prevalence of regional positions in international security negotiations on three levels.

On the international level, the dependent variable (DV) measures the number of times regional positions were voiced in each of the 102 ISO negotiations. Thus, the first DV is count data. Since mean and standard deviation are not far apart (mean is 1.54902 and the standard deviation is 1.843519), Poisson regression models are appropriate for the analysis of the international-level hypotheses.

On the regional level, RA activity is operationalized by a bivariate variable, 0 standing for an RA being silent in an international security negotiation, 1 standing for an RA being vocal in an ISO negotiation. Consequently, the regional level hypotheses are examined with logit regressions.

On the state level, the dependent variable is count data, as it captures the number of times a state speaks up in each of the negotiations. The state-level hypotheses are empirically analyzed with negative binominal regression analysis to accommodate the fact that variance and mean are not close in value (with 0.1742356 and, respectively, 0.0249839).

The independent variables are operationalized in the following manner. Information on the formal role of RAs in ISOs stems from the ISO’s homepages. Observers and formal administrative or organizational roles were coded with 1, while all RAs without formal status in an ISO negotiation were coded with 0. Similarly, for each ISO year, the number of RAs with formal status was counted. The size of ISOs is measured by the number of member states in the period 2008-2012. The data stem from ISO homepages. Finally, the formal decision-making rules of ISOs were coded in a categorical manner, 1 representing consensual decision-making rules and 0 representing majority decisions. The information stems from the ISO founding treaties and the rules procedures. State power is operationalized through GDP (in billion USD for the period 2008–2012) and the data stem from the World Bank. Data on security policy competencies of RAs have been captured based on the Yearbook of International Organizations and the respective homepages of RAs (Panke and Stapel, forthcoming). On this basis, the number of RA memberships with security competencies per state has been captured by counting the RAs with security competencies that each state belonged to in the period 2008–2012. The military power of a state is captured by the size of the armed forces (on active duty) per country as an average of the years 2007–2011 and stems from the World Bank.Footnote 4 The descriptive statistics of all independent variables are reported in the Appendix (cf. Table 5).

Table 4 reports the findings of the regression analysis. All models are designed parsimoniously and avoid problems arising from multicollinearity.

Table 4 Regression analysis
Full size table

Model 1–2 in Table 4 presents the Poisson regressions, which examine the international level, and sheds light on the plausibility of hypotheses 1–3. As expected by H1, an increase in the number of member states of an ISO significantly increases the likelihood for the articulation of regional actor positions in the international negotiation. Acting as a group rather than each IO member state articulating a position on its own renders multilateral negotiations more effective, which is especially important when IOs are large in terms of membership size and when negotiations take place in the high politics area of international security. The empirical results for hypothesis 2 are also in line with our expectations. The more regional actors have observer status in an ISO, the more likely it becomes that regional positions are voiced in the respective negotiations on international security rules and norms. Thus, the fact that some IOs are more open to grant RAs formal access to negotiations matters, as RAs use formal pathways to actively participate in international security negotiations when they exist. Table 4 provides insights into the plausibility of hypothesis 3 as well. Consensual decision-making rules in ISOs reduce the chance for RA activity in a significant and robust manner. Whenever there a consensus rule applies, each individual IO member is de facto a veto player and can prevent outcomes from being passed. In such settings, states cannot benefit much from working through groups, as leveraging group voting power by pointing to the number of votes an RA combines only makes sense, when majority voting takes place.

Model 3 in Table 4 puts the regional level expectations to an empirical plausibility probe. The dependent variable is RA activity (0 standing for an RA being silent in an international security negotiation and 1 standing for an RA being vocal in an ISO negotiation). Model 3 shows that RAs with a formal status in an ISO negotiation are significantly more likely to voice regional positions, compared to the RAs without formal status at all (as expected by H4). Thus, the fact that some RAs but not others invest the time and resources necessary to become accredited as registered observers is indicative of the willingness of these collective actors to participate in international security negotiations. Also, regional positions have a greater chance of being articulated in international negotiations on security issues, when the RA in question has security competencies (H5). RAs with explicit security competencies can often already return to prior agreements between the member states. With respect to the security norm or rule on the international negotiation agenda, ROs are, therefore, in a better position to develop a common regional position that can subsequently be articulated in the ISO. Finally, model 3 in Table 4 also shows that RA size is positive and significant. This is in line with hypothesis 6. Larger RAs feature a higher propensity for becoming vocal in ISO negotiations, since there are more actors who can speak on behalf of their regional group.

Model 4 features the negative binominal regression analysis of the state-level hypotheses (Table 4). Hypothesis 7 needs to be refuted as militarily weaker states are not systematically more likely to speak on behalf a regional actor in an ISO in an effort to leverage their position in the RA in international security negotiations. In fact, the correlation points to the opposite direction: the chance of a regional position being articulated increases significantly for states with larger militaries. This indicates not only that all states can—in principle—leverage their position in RAs in international security negotiations by referring to groups, but also that militarily powerful states might be especially active within regional groups when security topics are on the agenda and regional positions are developed. They might subsequently be especially motivated and able to push these positions in international security negotiations. Hypothesis 8 expected that states are more inclined to voice regional positions in international security negotiations, if they are members in a higher number of RAs with security competencies. Model 4 supports this expectation as an increase in the security RA memberships a state holds goes hand in hand with an increased propensity to articulate regional positions in ISO negotiations.

5 Conclusions

International security constitutes a part of the core of what sovereign statehood is all about (Herz 1950; Jervis 1985; Krasner 1999; Mangold 2013; Dannreuther 2014). Thus, it is telling that international security negotiations are nowadays not the exclusive domain of states. This article has illustrated that regional actors are vocal in ISOs, as about 3.38 % of all formal speeches made in the 102 negotiations under examination are regional in character instead of representing particular individual national positions.

The ATT attracts the most regional negotiation activity, followed by OPWC and by the UNGA’s first committee. 19 RAs currently have security policy competencies. Yet, it is not the case that the ArabLeague, ASEAN, AU, CARICOM, CEEAC, CIS, COMESA, CSTO, ECOWAS, EU, GUAM, IGAD, NATO, OAS, OSCE, PIF, SADC, SCO, and UNASUR are per se more active than RAs lacking such competencies. The article shows that the EU is most vocal in ISOs followed by the UN’s African Group, the CARICOM and Mercosur. Most importantly, states can speak for their RA irrespective of whether the latter has formal status in an ISO. The most vocal states in this respect were Antigua and Barbuda, Barbados, and Honduras, while many other countries, such as Mexico, Latvia and the US, only expressed national but not regional positions.

None of these patterns is self-evident. Why are some ISOs more strongly regionalized than others although they are all dealing with high politics, security issues? Why are some RAs much more active in international security negotiations than others although none of them has full member status? Why are some states more inclined to speak on behalf of their RA even though other states could benefit from referencing groups as a means of leverage in ISO negotiations?

This article answered the questions. To this end, it drew on international theories on cooperation and conflict as well as on institutional design approaches to specify a set of hypotheses for the international, the regional and the state levels. The empirical analysis reveals that international and regional level variables influence regional activity in international security negotiations. ISOs that are large in size and open to many RAs as observers attract a high level of regional actor activity. Regional actors differ when it comes to articulating regional positions in ISOs. RAs are most likely active when they have formal security competencies, possess a formal status in the ISO and have many member states that could potentially act on their behalf in international negotiations in the field of security. Furthermore, in international security negotiations, state characteristics also have decisive effects on the likelihood that states will leverage collective actors. Militarily power and the number of security RAs, a state is a member of, both increase the chance that a country voices a regional position.

These findings have a series of important implications.

First, regional actors are not only active beyond their borders when it comes to international economic or trade negotiations (e.g., Dür and Zimmerman 2007), but also in international security negotiations. This and the magnitude of their involvement are surprising, since international security is often regarded as the domain of the sovereign nation-state (Morgenthau 1948; Waltz 1959).

Second, although the EU is the most prominent RA (in Western discourse and research), regional integration has taken place all over the globe, leading to a broad array of regional actors. While the EU is very vocal in ISO negotiations, it is not the only kid in town. In fact, there are no macro-regions without RAs that also become active in ISOs. Thus, the regionalization of security topics takes place around the globe.

Third, regional actor involvement has important implications for the dynamics and outcomes of international security negotiations. This is especially true when IOs are large in size, and bundle national positions in regional group positions, which increases the effectiveness of negotiations. Moreover, case studies have illustrated that being vocal in international negotiations can translate into RA influence over international norms (Panke 2013b; Laatikainen and Smith 2006; Blavoukos and Bourantonis 2011; Kingah and Van Langenhove 2012; Zwartjes et al. 2012; Weiffen et al. 2013). This is important since RA involvement carries not only the potential to increase the effectiveness of multilateral negotiations, but also the potential to leave regional imprints on international security norms, rules and practices (Kingah and Van Langenhove 2012; Zwartjes et al. 2012; Weiffen et al. 2013).