States in international organizations: Promoting regional positions in international politics?

States, RIOs, and international relations

Traditionally, international relations are characterized by interactions between states. Early International Relations scholarship examined the conditions under which states are increasingly likely to wage war with one another and when and how peace could be pursued. ¹ In the 1970s and 1980s, the puzzle of cooperation under anarchy was increasingly addressed. ² Hence, researchers examined the role of power, interests, and institutions for interstate cooperation. ³ The major focus was on why, how, when, and where states cooperate with one another, either on an ad hoc basis or in institutionalized arenas. ⁴ IOs, as institutionalized arenas for cooperation between three or more states, in which membership is not defined by geographic criteria, not only increased in numbers in this period but also turned into essential fora for negotiating the rules and norms of an international order. ⁵

With the end of the Cold War, a wave of regionalization took place, as the number of RIOs, as institutionalized arenas for cooperation between at least three states from the same geographical region, increased to a total of more than 70 today. ⁶ Initially, RIOs were in most cases created to foster cooperation between their member states internally, in areas such as creating free-trade zones or common markets and cooperating in agriculture or technology sectors. ⁷ Yet, most of them are now also engaged in external affairs, such as neighbourhood policies and peace and security issues. ⁸ In the last decades, comparative regionalism scholarship and EU researchers have studied how the RIOs turned into international actors. ⁹ One strand of scholarship has pointed out that today RIOs play de facto important roles in areas that were formerly in the realm of nation-states, ¹⁰ such as leaving regional imprints on international norms through voting alignments, ¹¹ and through participating in international negotiations themselves. ¹²

RIOs became full members with speaking and voting rights in IOs only in a few exceptional instances, such as the EU in the WTO and the UNFCCC. More often, RIOs can register themselves as observers in IOs and thereby obtain access to international negotiations, whereas only states with full IO membership status have speaking and voting rights. ¹³ However, it is not the case that RIOs are unimportant for international negotiations in IOs even when they have no formal status at all. Case studies have demonstrated that, even under such conditions, RIOs play a role. ¹⁴ Member states can speak on behalf of their RIO (see the third section, “Regionalized international negotiations”) and can also vote as a bloc in IOs. ¹⁵ Thus, the increase in regime complexity in general and the rise of RIOs in particular could suggest that, in the course of the last decades, a regionalization of international relations has taken place and also that RIO member states play an important role in this process (for details see the third section, “Regionalized international negotiations”). Nevertheless, we do not know much about the conditions under which states turn into agents of regionalization and voice regional rather than national positions in international negotiations.

Regionalized international negotiations and the role of RIO member states—the pattern

To examine the role of states in pushing a regionalization of international relations, a dataset on international negotiations is needed that captures who speaks up in formal IO negotiations and whether the position expressed is national or regional in character. We compiled a dataset that encompasses 21,194 observations, tracing the activity of 205 states and state-like actors ¹⁶ within 27 international negotiation arenas and covering a broad variety of different policy areas (a list of IOs is provided in Table A1 in the Appendix). To examine the prevalence of regional positions in international negotiations, IOs must provide detailed information on who spoke up in negotiations and what the speech made was about. Unfortunately, such information is not available for the large majority of IOs covered in the Correlates of War (COW) dataset. Thus, we included in our dataset those intergovernmental IOs that maintained a functioning and updated website, had IO legislative bodies in regular session, and in which states negotiate and pass IO outputs (e.g., decisions, regulations, resolutions) on a frequent basis in the period under examination (2008–2012). In addition, these intergovernmental IOs needed to provide sufficiently detailed negotiation information (e.g., in verbatim protocols, detailed records, or reports).

The dataset relies on a subsample of IOs that are active rather than those that are no longer or only rarely used by member states, in order to negotiate international rules and norms or engage in operational activities. Thus, the findings of this paper do not apply to “zombie” or inactive IOs. ¹⁷ Similarly, this paper includes only IOs that exceed a certain degree of transparency of negotiations and omits IOs that do not allow public access to information on who was active in putting forward which positions in negotiations. Accordingly, inferences to nontransparent IOs, which do not provide sufficiently detailed information on negotiations, need to be avoided as well.

For each of the international negotiation arenas in the dataset in each of the four years (2008–2012), four negotiations on substantive issues typical for the policy area of the respective IO were included. IOs differ in the number and volume of output documents they pass. For IOs in which legislative bodies convened fewer than 16 times or produced fewer than 16 output documents, we included as many negotiations as possible. On the whole, the dataset includes 512 negotiations.

Depending on the IO’s reporting system and institutional structure, we collected information on which actor spoke and what type of position they expressed from officially available reports, minutes, and press releases from the negotiations taking place in the IO’s main legislative arena (usually the assembly and respective committees) in which the bulk of negotiations takes place. The speech acts of a specific negotiation were coded manually, based on a content analysis, including silent but present actors, voiced national positions of states, and speech acts by states on behalf of RIOs. We operationalized negotiation activity on the speech level. For example, if a state representative declares that he/she is speaking “on behalf of” an RIO, that “the RIO agrees to/disagrees with/objects/supports/proposes/recommends, etc.,” the speech is counted as voicing a regional position, while the absence of a reference to the regional position or preference is counted as voicing national positions. Thus, our analysis captures negotiation activity as well as the type of intervention for each actor.

The dependent variable of this study, the negotiation behaviour of RIO members, has three empirical expressions: 0, 1, and 2. Values 1 and 2 indicate that the state in question has made a formal speech in the IO negotiation, whereas 0 indicates that a state is a member of an IO but remains silent in a negotiation. If a speech made by a state is national in character, it is coded with 1, whereas a speech in which a state expresses an RIO position is coded with 2. ¹⁸

Of the 21,194 observations in the dataset, the actors were silent in 16,292 instances and spoke up in 4901 instances. Of the 4901 speech acts made, 409 expressed RIO positions (8.3% of all speeches made). As Table 1 shows, not all states spoke up equally often and not all states were equally inclined to voice RIO positions instead of national ones. In absolute terms, France voiced a total of 25 RIO positions, whereas Zimbabwe spoke on behalf of an RIO only once. In between are countries such as Mozambique, Uruguay, and Vietnam (four RIO positions each) or Brazil, Barbados, and Singapore (five RIO positions each). Somalia, Uzbekistan, and Togo did not negotiate on behalf of an RIO at all.

OPEN IN VIEWER

Table 1.

The relative share of RIO positions voiced by countries^a

country	% silent	% RIO positions	country	% silent	% RIO positions	country	% silent	% RIO positions	country	% silent	% RIO positions
FRA	50.96	15.92	ECU	61.98	1.65	BOL	75.79	0.00	MAR	63.36	0.00
SWE	63.64	13.64	CHL	61.29	1.61	BRN	95.00	0.00	MCO	95.24	0.00
BEL	73.85	10.00	ZAF	56.82	1.52	BWA	84.38	0.00	MDA	86.14	0.00
ESP	60.63	8.66	USA	31.34	1.49	CAF	95.79	0.00	MDG	91.00	0.00
CYP	85.19	8.33	Total	77.29	1.38	CHE	50.00	0.00	MDV	83.52	0.00
HUN	82.50	7.50	TON	90.79	1.32	CHN	35.94	0.00	MEX	50.74	0.00
CZE	79.65	7.08	BTN	92.41	1.27	CMR	83.76	0.00	MHL	96.51	0.00
DNK	74.17	6.67	GRD	92.05	1.14	COG	88.68	0.00	MKD	91.95	0.00
THA	50.00	6.14	PNG	87.91	1.10	COK	96.49	0.00	MLI	87.25	0.00
JAM	78.57	6.12	DMA	94.57	1.09	COL	62.18	0.00	MNG	84.31	0.00
SGP	76.92	5.49	KWT	84.78	1.09	COM	93.18	0.00	MRT	96.36	0.00
BRB	79.79	5.32	VCT	95.70	1.08	CPV	95.65	0.00	MUS	88.68	0.00
SVN	84.35	5.22	BHR	81.91	1.06	CUB	58.25	0.00	NER	91.92	0.00
POL	75.00	5.17	MNE	87.37	1.05	CUW	95.65	0.00	NGA	63.78	0.00
IDN	43.22	5.08	ARM	85.42	1.04	DJI	90.00	0.00	NIU	96.49	0.00
TTO	82.86	4.76	LTU	91.67	1.04	DZA	60.66	0.00	NPL	70.00	0.00
GUY	93.33	4.44	GEO	85.57	1.03	EGY	67.21	0.00	NRU	98.63	0.00
VNM	74.76	3.88	KHM	93.88	1.02	ERI	93.41	0.00	PAK	56.25	0.00
MYS	53.33	3.81	HND	85.86	1.01	FJI	86.32	0.00	PAN	82.57	0.00
MOZ	84.91	3.77	OMN	91.92	1.01	FRO	95.65	0.00	PCN	76.27	0.00
BRA	42.03	3.62	SRB	82.00	1.00	FSM	93.06	0.00	PHL	56.52	0.00
NLD	72.27	3.36	AZE	82.18	0.99	GAB	86.84	0.00	PLW	96.25	0.00
ARE	79.35	3.26	BFA	84.16	0.99	GHA	81.58	0.00	PRK	78.38	0.00
SUR	91.30	3.26	DOM	78.43	0.98	GIN	88.68	0.00	ROU	84.35	0.00
AUS	39.02	3.25	SLV	79.61	0.97	GMB	90.53	0.00	RUS	51.88	0.00
DEU	61.29	3.23	ETH	80.00	0.95	GNB	93.62	0.00	RWA	90.82	0.00
URY	65.60	3.20	ZWE	70.75	0.94	GNQ	95.35	0.00	SDN	69.00	0.00
MMR	80.85	3.19	EST	89.81	0.93	GRL	77.78	0.00	SEN	74.17	0.00
GBR	52.31	3.08	SVK	87.96	0.93	GTM	78.38	0.00	SLB	94.37	0.00
BLZ	91.26	2.91	TZA	72.48	0.92	HKG	96.23	0.00	SLE	84.47	0.00
FIN	80.00	2.61	IRL	75.68	0.90	IND	38.02	0.00	SMR	96.47	0.00
VEN	63.64	2.48	COD	86.73	0.88	IRN	44.59	0.00	SOM	93.20	0.00
NOR	55.73	2.29	ZMB	79.65	0.88	IRQ	69.23	0.00	SSD	96.97	0.00
BHS	91.30	2.17	ISL	86.09	0.87	ISR	62.71	0.00	STP	98.84	0.00
SWZ	88.30	2.13	PER	71.30	0.87	JOR	79.05	0.00	SXM	95.65	0.00
MLT	82.11	2.11	CIV	85.34	0.86	JPN	36.72	0.00	SYC	92.31	0.00
ATG	93.81	2.06	TUR	54.31	0.86	KAZ	81.51	0.00	SYR	75.00	0.00
PRY	81.44	2.06	NZL	52.10	0.84	KEN	66.39	0.00	TCD	88.66	0.00
HTI	87.76	2.04	ARG	60.16	0.78	KIR	96.47	0.00	TGO	88.12	0.00
QAT	77.55	2.04	CAN	42.19	0.78	KNA	94.32	0.00	TJK	94.51	0.00
MWI	81.37	1.96	KGZ	88.42	0.53	KOR	52.00	0.00	TKM	96.84	0.00
GRC	82.52	1.94	AFG	75.76	0.00	LAO	88.89	0.00	TLS	86.08	0.00
SAU	77.67	1.94	AGO	78.85	0.00	LBN	76.00	0.00	TUN	76.47	0.00
UKR	73.33	1.90	ALB	92.71	0.00	LBR	89.36	0.00	TUV	96.25	0.00
NAM	76.85	1.85	AND	100.00	0.00	LBY	76.47	0.00	TWN	84.78	0.00
PRT	80.53	1.77	BDI	89.58	0.00	LCA	92.05	0.00	UGA	80.19	0.00
AUT	71.30	1.74	BEN	86.36	0.00	LIE	91.30	0.00	UZB	98.85	0.00
NIC	85.22	1.74	BGD	58.41	0.00	LKA	62.39	0.00	VAT	44.68	0.00
HRV	82.76	1.72	BGR	87.39	0.00	LSO	83.50	0.00	VGB	94.44	0.00
ITA	73.11	1.68	BIH	90.63	0.00	LUX	89.52	0.00	VUT	96.67	0.00
CRI	71.67	1.67	BLR	66.96	0.00	LVA	94.74	0.00	WSM	90.24	0.00

^aThe country abbreviations follow the ISO-3 code, accessible under https://unstats.un.org/unsd/tradekb/Knowledgebase/50347/Country-Code

Table 1 shows cross-country variation. States vary in the extent to which they became active in international negotiations to which they had access as well as the extent to which they negotiated on behalf of an RIO. When France, Sweden, and Belgium spoke up in IO negotiations, they expressed RIO positions (in this case, EU positions) more than 10 percent of the time. Denmark, Thailand, Jamaica, and Singapore voiced RIO positions about 6% of the time each of the countries spoke up. Whereas Denmark spoke for the EU, Thailand and Singapore voiced positions of the Association of South East Asian Nations (ASEAN), and Jamaica negotiated on behalf of the Caribbean Community (CARICOM). Overall, the positions of 25 different RIOs were expressed in the international negotiations (see Table A2). The EU position is most prevalent, followed by CARICOM and ASEAN, with about 30 speeches. The South African Development Community (SADC), the African Union (AU), the Arab League, the Gulf Cooperation Council (GCC), and Mercosur have 10 or more positions voiced.

States are drivers of regionalization in IOs. This could suggest that states can benefit from acting in concert rather than negotiating on their own. Yet, if this were true, it is puzzling that not all states make use of the opportunity to use RIOs as a means to leverage up in IO negotiations. To shed light on this issue and examine the factors behind the phenomenon that states drive international-level regionalization, the subsequent sections address the following questions: Why do states speak for RIOs in international negotiations and why do they differ in the extent to which they do so? These questions will be answered in the remainder of this paper.

Accounting for variation in states as drivers of regionalization

This section theorizes the conditions under which states are likely to speak on behalf of RIOs in IOs. We use rationalist liberal International Relations theory in order to first theorize under which conditions states are likely to be silent in international negotiations, and, in a second step, theorize when non-silent states voice regional rather than national positions. Liberal internationalist approaches open the black box of unitary states and theoretically capture in a first step whether and how state positions are developed domestically. ¹⁹ In a second step, liberal internationalist approaches capture how states, as strategic rational actors, pursue their stances on the international level. ²⁰

Liberal theories of international cooperation argue that a state’s foreign policy positions are not naturally given and not simply “out there,” but need to be formulated in the domestic realm first. ²¹ Only when the domestic construction of a national position is completed, is the position communicated to the diplomats at the IO negotiation tables and a state can start to actively voice its positions. ²² For the development of national positions in governments and foreign ministries, capacities are essential. With respect to policy expertise, the better resourced a state is financially, the better the internal governance processes work in order to formulate national positions swiftly for all issues that are on an IO’s negotiations agenda at a given point in time. ²³ In contrast, states with insufficient governance capacities are more likely to encounter situations in which the domestic formulation of national positions is either delayed or—worse—cannot be accomplished for all issues on the negotiation agenda of IOs. ²⁴ As a consequence, diplomats of states with lower governance capacity are more likely to face situations in which their capitals did not provide a national position and are, consequently, less often authorized to articulate a position in an IO. As a consequence, states in which diplomats do not know the national position are more often silent in international negotiations than states with higher levels of governance capacity. ²⁵ Thus, hypothesis 1 expects that states with lower governance capacity are more likely to be silent in international negotiations in IOs.

If states are not silent in international negotiations, they can voice national positions or regional positions. Hence, the next paragraph also draws on liberal rational choice approaches to hypothesize under what conditions states are likely to speak on behalf of an RIO rather than articulate a national position. To this end, it distinguishes between incentives and resources as factors impacting the international conduct of states. ²⁶

First, IOs differ in size. Some are small and have fewer than 50 member states, whereas others are encompassing and have more than 190 members. Similar to other institutional design features, the openness for state membership can have important consequences for IO internal dynamics. ²⁷ In larger IOs, the number of members that can participate in the negotiations by voicing a position is high, which can render such multilateral negotiations time consuming and possibly also inefficient if every single state were to articulate its own positions. Thus, in negotiation arenas with many participants, states should have incentives to group their positions in order to reduce the number of speeches made and increase the efficiency of the negotiation process. If RIO member states have coordinated themselves and formulated a common regional position rather than each RIO member articulating a national position, this behaviour not only increases the speed of international negotiations, but also could increase the chances for negotiation success, as the collective bargaining power reinforcing the position is higher when a state speaks on behalf of an RIO rather than expressing a national position. According to hypothesis 2, an increasing number of negotiation participants in IOs should increase the likelihood that states voice regional positions instead of national ones.

Second, all states are members of several RIOs, ²⁸ but the number of RIO memberships varies. Some states, such as Tuvalu, have joined only a few RIOs, whereas others, such as Russia, have become members in eight or more RIOs. The number of RIOs a state has joined could influence its incentives to invest its own resources and voice a position in IO negotiations on behalf of an RIO. The rational choice literature has long ago identified the problem of free riding, ²⁹ which suggests that states’ incentives to produce a common good and articulate a regional position declines in cases where the RIO position does not adequately reflect the national position. The more RIOs a state has joined, the higher the number of potential regional positions that a state could voice and the more likely that the position of at least one RIO to which a country belongs is sufficiently close to its national position. Hence, hypothesis 3 expects that, with an increasing number of RIO memberships, states are more likely to voice a regional position than a national one.

Third, regardless which RIO states joined, they retained their sovereign decision-making authority regarding which position to pursue in IO negotiations in which they are full members. Thus, similar to national positions in states, regional positions of RIOs for issues on an IO negotiation agenda are not just out there but need to be formulated by the RIO member states. ³⁰ Usually, one RIO member has a chair position for a given period of time and organizes RIO coordination meetings either at the RIO headquarter location or at the location in which the IO is based. In these meetings, regional positions are developed by member states. The state holding the RIO chair has not only obligations in respect to the organization of group coordination meetings but also incentive to voice the regional position in subsequent IO negotiations. Thus, according to hypothesis 4, states holding the RIO chair are more likely to voice a regional position than a national one in international negotiations.

In addition to these incentives, resources could also impact the likelihood that a state voices a regional position in IO negotiations. Deciding whether to voice a regional or a national position is not capacity neutral. Rather, the former requires greater financial and staff capacity than the latter. The state that commits itself to push the RIO position in international negotiations needs to maintain a channel of communication with the other RIO member states in order to flexibly adapt the regional position to the evolution of international negotiation dynamics. In addition, the state voicing a regional position in the formal IO negotiation arena is often also using its financial and staff capacities to promote the regional position in informal venues, such as coffee breaks, bilateral meetings, and diplomatic receptions. Finally, hypothesis 5 expects that an increase in financial resources and an increase in the diplomatic size of a state increases the likelihood that it negotiates on behalf of an RIO in IO negotiations.

Empirical analysis and discussion

The independent variables are operationalized as follows: Data on government effectiveness (hypothesis 1) was provided by the World Bank, capturing the perceived quality of public services, civil service and degree of independence, quality of policy formulation and implementation, and the credibility of government commitment. ³¹ The information on the variable IO size (hypothesis 2), which captures the number of IO members participating in the international negotiations under scrutiny, for each of the IOs in the dataset, was obtained through the respective IO home pages and negotiation protocols or reports (accessed October 2013). Hypothesis 3 focuses on the number of RIO memberships a state holds. These data were obtained through the respective RIO home pages (accessed December 2014). Data for which states hold RIO chair positions and for how long (hypothesis 4) were collected based on the respective home pages of the RIOs in the dataset (accessed October 2013). Financial capacities (hypothesis 5) are measured by the natural logarithm of gross domestic product (GDP) covering the years 2008–2012 (in million US$, based on current US$), and the data stem from the World Bank. ³² Finally, diplomatic staff capacities (also hypothesis 5) capture the number of diplomats posted in New York per state as a proxy for the diplomatic capacities of states. ³³ The data stem from the United Nations (UN) bluebooks. In addition to the variables stemming from the hypotheses, the models on the probability of remaining silent also control for financial and socio-economic capacities. Whereas financial capacities capture the economic power of a country (log GDP in billion US$, see above), socio-economic capacities capture the level of socio-economic development in a country. It is measured by GDPpc, the data of which also stem from the World Bank. The descriptive statistics as well as a correlation table are provided in Tables A3 and A4 in the Appendix.

The dependent variable of this study, the negotiation behaviour of RIO member states, has two empirical expressions: 0, 1, and 2. (See also the third section, “Regionalized international negotiations.”) When a state is member of an IO but remains silent in a negotiation, this behaviour is coded with 0, whereas 1 and 2 indicate that the country in question has made a formal speech in the IO negotiation. When the speech made is national in character, it is coded with 1. If the speech is regional in character, as the state explicitly speaks on behalf of the RIO, it is coded with 2. The dependent variable data are of count nature and have ordered response variables (0, 1, and 2). As states are often silent in international negotiations, ³⁴ a zero-inflation model is suitable, as this allows us to capture whether or not a state speaks up at all or remains silent in a given negotiation. Because of the ordered nature of the variable of interest (no position voiced, national position voiced, regional position voiced), zero-inflated ordered probit models are estimated.

Table 2 presents the empirical findings of the regression analysis. The upper part of the regression table captures whether states voice regional rather than national positions, and the lower part of the table captures the probability of states not voicing any position at all but remaining silent.

OPEN IN VIEWER

Table 2.

Zero-inflated Ordered Probit regressions.

	model 1	model 2	model 3	model 4	model 5
IO size	0.002*** (0.000)	0.002*** (0.000)	0.002*** (0.000)	0.002*** (0.000)	0.002*** (0.000)
number of RIO memberships	−0.032*** (0.006)	−0.031*** (0.006)	−0.040*** (0.007)	−0.033*** (0.006)	−0.050*** (0.008)
RIO chair position	3.329*** (0.133)	3.327*** (0.133)	3.293*** (0.127)	3.327*** (0.133)	3.281*** (0.127)
financial capacities	0.165*** (0.007)	0.167*** (0.007)	0.044** (0.017)	0.166*** (0.007)	0.047** (0.018)
diplomatic staff capacities	0.005*** (0.001)	0.005*** (0.001)	0.006*** (0.001)	0.005*** (0.001)	0.006*** (0.001)
inflate – probability of being silent
government effectiveness	−0.137**(0.042)			−0.070(0.062)	−0.132***(0.028)
socio-economic capacities		−0.000***(0.000)		−0.000(0.000)
financial capacities			0.266***(0.017)		0.271***(0.017)
Constant	1.046***(0.082)	1.117***(0.089)	−0.649***(0.125)	1.082***(0.089)	−0.736***(0.124)
cut 1	1.462***(0.055)	1.479***(0.055)	0.712***(0.130)	1.465***(0.056)	0.640***(0.136)
cut 2	3.376***(0.065)	3.394***(0.065)	2.683***(0.114)	3.380***(0.066)	2.637***(0.118)
Observations	20099	19943	20306	19821	20099
BIC	21247.976	21041.901	21258.067	21001.966	21169.369
LL	−10579.400	−10476.398	−10584.400	−10451.510	−10535.142

Clustered standard errors in parentheses with *p<0.05, **p<0.01, ***p<0.001

The regression analysis reveals that, consistent with the expectation of hypothesis 1, an increase in government effectiveness reduces the likelihood of states being totally silent in IO negotiations. The covariate for government effectiveness is robustly negative in all models (1, 4, and 5 of Table 2) but highly significant only in models 1 and 5. Models 4 and 5 include financial and socio-economic capacities, respectively, as control variables to check for the robustness of the effect of government effectiveness. The lack of significance in model 4 is because the correlation between government effectiveness and socio-economic capacities exceeds 0.7 and brings about problems of multicollinearity. (Nevertheless, model 4 is reported here in the interest of comprehensiveness.) Hence, we can infer from the regression analysis that the more government capacity a state possesses, the more likely that state is to develop a position that the national delegate can present in the IO. Consequently, states are more likely to voice either a national or a regional position in international negotiations when they do not grapple with shortages of domestic coordination capacity. For instance, countries with very low levels of government effectiveness, such as Rwanda and Suriname, are silent more than 90% of the time. By contrast, countries with high levels of government effectiveness, such as Canada, Japan, and Austria, encounter less often situations in which their delegates cannot speak up in international negotiations because they did not receive instructions on what position to put forward. They are, therefore, more active in IO negotiations. This finding is consistent with the literature on absenteeism in IOs and negotiation participation more generally ³⁵ and accounts for the large number of zeros in the dataset.

Hypotheses 2 to 5 theorized conditions under which states are likely to adopt regional rather than national positions in international negotiations. Given that the variables do not correlate strongly (below the 0.7 threshold), all independent variables of the hypotheses are placed in each of the upper parts of the regression models at once. (All findings remain equally robust and significant when the two variables that correlate at 0.6—staff capacities and financial capacities—are placed in separate models.) The empirical analysis lends support to most of the hypotheses.

Consistent with hypothesis 2, an increase in the number of participants in IO negotiations increases the probability that a state articulates a regional position. This finding is robustly significant in all five models presented in Table 2. To provide an example in larger IOs, such as the International Labour Organization (ILO), the number of regional positions put forward is considerably higher, 49, than in the much smaller dependent variable International Whaling Commission (IWC), 4, which has fewer member states. This indicates that the extent to which international negotiations are regionalized varies between negotiation arenas; states are more inclined to voice RIO positions in larger IOs, in which a tour de table of national positions would take a very long time and hamper the efficiency of IO negotiations.

The third hypothesis also focused on state incentives to voice regional rather than national positions in international negotiations. The empirical evidence does not support hypothesis 3, as the sign points robustly in the wrong direction. States that have joined an increasing number of RIOs do not have higher propensities to voice regional positions in international negotiations. Instead, models 1–5 suggest that the probability that a state negotiates on behalf of an RIO rather than voice a national position declines as the number of RIO memberships the state holds increases. This counterintuitive finding might be due to states guarding their own—ultimately limited—capacities cautiously. States that are members of many RIOs do not speak up for all of them at once but, at most, for one RIO in one international negotiation. Furthermore, the significantly negative finding of the regression analysis might indicate that states with many RIO memberships could even engage in free-riding behaviour when they have more opportunities to do so and when they can better diffuse potential sanctions for free riding.

Hypothesis 4 formulated an institutional incentive for states to become active on behalf of their RIO: states should be more likely to voice a regional position while they hold the office as RIO chair. Table 2 lends support to this expectation. Thus, an institutionalized division of labour is at play in RIOs, according to which the state chairing the internal RIO meetings should also represent the RIO position (if there is one) in the international negotiation arena. While respective states hold the chair position in an RIO, the probability that they become active for their RIO and voice a regional position increases. For instance, while New Zealand was operating as chair for the Pacific Islands Forum (PIF) meetings in 2012, it subsequently voiced the PIF positions in the Arms Trade Treaty (ATT) negotiations.

Finally, resources matter as well. Expressing regional positions and negotiating for their RIO in IOs is capacity intensive for the states concerned. All five models in Table 2 show positively robust and significant covariates for both types of capacities. This finding is consistent with hypothesis 5 and lends support to the expectation that the more financial and diplomatic staff capacities a state possesses, the higher the chances that this state articulates a regional position in an IO. Thus, states with very small diplomatic bodies, such as Nauru or Tuvalu, or with slim budgets, such as Suriname or Sudan, are considerably less inclined to invest their limited resources for voicing regional positions than countries with larger diplomatic bodies (e.g., Japan or Brazil) and larger budgets (e.g., Germany or India).

Similar to other pieces of comparative IO scholarship, ³⁶ this paper is not based on the universe of IOs but a subsample, which has implications for the generalization of the findings. Because of the nature of negotiation information needed, this paper’s insights relate to IOs that are active and that exceed a certain degree of transparency in negotiations. For these types of IOs, the analysis of this paper suggests that states are important agents of the regionalization of international relations that we currently observe. Yet, this process is not driven by all states to the same extent. States are inclined to push RIO positions in IO negotiations when they expect benefits from doing so, which is the case when IOs are large in size. In addition, states are motivated to voice regional positions in international negotiations when they have a formal role in the RIO with respect to organizing coordination meetings between RIO member states. Also, resources matter for a state’s decision to voice a regional position. When more diplomats are available and when the budgets are larger, a state is better able to perform the extra legwork and acting on behalf of the RIO. At the same time, states are sensitive to the opportunity structures provided by multiple RIO memberships. If states can save capacities and rely on the negotiation efforts of fellow RIO members, they are less inclined to voice a regional position themselves.

Conclusions

International relations have traditionally been the prerogative of states. In the overwhelming majority of IOs, RIOs cannot acquire full membership status. In some IOs, RIOs can register as observers and gain access to international negotiations, but this comes without voting competencies and usually also without speaking rights in IO negotiations. Nevertheless, RIOs can play an important role in international negotiations taking place in IOs, as case studies have already evidenced. Yet, up to now we did not know that the regionalization of international negotiations is driven by states. Irrespective of an RIO formal status, its member states can and do speak on its behalf in international negotiations. In fact, in more than 8 percent of all speeches made, RIO positions were expressed. Yet, the regionalization of negotiations is not brought about by all states to an equal extent.

The analysis demonstrated that states are more likely to make active use of their IO memberships and participate in negotiations by making formal speeches when they possess greater governance capacity. If states cannot develop national positions for all agenda items on an IO’s negotiation table, their diplomats are more often silent and express neither national nor regional positions in international negotiations. By contrast, states with higher levels of governance capacities are less often silent in IOs but may voice either national or RIO positions. This choice is driven by incentives and resources of the states in question.

This paper shows that states are more likely to negotiate on behalf of RIOs when the IO in question is larger, as voicing group positions instead of national ones reduces the total number of positions articulated and thereby increases the efficiency of multilateral negotiations. Hence, the negotiation arena’s characteristics form an important incentive structure for states’ decisions about whether to voice regional instead of national positions. A second important incentive for states to turn into agents of regionalization in IOs is linked to the formal role they have in RIOs. Whenever a state serves as RIO chair, the likelihood increases that this particular state voices a regional position.

In addition to these incentives, state resources also play a role. Expressing a regional instead of a national position is not resource neutral for the actors involved. Negotiating on behalf of an RIO often requires several RIO coordination meetings in the course of IO negotiations in order to update the negotiation position and/or strategy in IO negotiation dynamics. This requires resources that states could save if they articulate a national position instead of a regional one. Accordingly, this paper shows that states are less likely to voice an RIO position in an IO when they possess lower diplomatic staff capacity and slimmer financial budgets.

The regionalization of international relations that we witness today can have important consequences not only for the dynamics of international interactions but also for outcomes of IO policy-making. Case studies have already demonstrated that when RIOs are active in IOs through their member states, they often leave regional imprints on international negotiation outcomes. ³⁷ For instance, The Economic Community of West African States (ECOWAS) and CARICOM are both affected by gun-related violence and therefore placed great emphasis on including ammunition into the ATT. ³⁸ The two RIOs managed to achieve this goal, thereby substantively altering the negotiation outcome. ³⁹ To give another example, in the negotiations on the Rome Treaty on Nutrition, the EU single-handedly pre-empted attempts of African states and RIOs to include the phrase that “trade forms an obstacle to food systems” into the Rome Treaty of 2014. ⁴⁰ Rather than framing trade as having negative implications for food systems, the EU successfully pushed a positive frame in this respect, and the final Treaty states that “trade is a key element in achieving food security and nutrition.” ⁴¹

Whether and how the regionalization of international relations will develop in the future, whether this leads to an increased importance of RIOs over time, and whether this phenomenon has lasting changes on the effectiveness and legitimacy of international negotiations is an open question, which is worthwhile to be studied in future work.

Diana Panke is Professor of Political Science with a Chair in Multi-Level Governance at University of Freiburg.