Transnational governance as strategy? Mapping and explaining the European Union’s participation in transnational public-private governance initiatives

ABSTRACT

The European Union (EU) is involved in a range of global governance institutions. Its participation in transnational public-private governance initiatives (TGIs) is particularly pronounced. What explains the EU’s involvement in TGIs? Why does the EU participate in some initiatives but not others? I draw on theories of orchestration, access of transnational actors to intergovernmental organizations, and legitimacy of global governance institutions to identify factors that shape EU involvement in TGIs. Using new data on EU participation in 636 TGIs, I provide the first systematic mapping of the Union’s involvement in transnational governance initiatives and test my hypotheses using statistical analysis. My results suggests that the EU is more likely to become involved in TGIs with many participants and an institutional design that provides the initiative with organizational structure and operational capacity. My findings are relevant for research on transnational governance, inter-organizational interactions in global governance, and EU foreign policy.

KEYWORDS:

• European union
• transnational public-private governance initiatives
• intergovernmental organizations
• inter-organizational interaction
• institutional complexity
• orchestration

Introduction

Intergovernmental Organizations (IGOs)¹ are important actors in world politics. They pursue ambitious goals, such as containing the use of violence, facilitating free trade, or promoting human rights (Wallace and Singer 1970; Shanks, Jacobson, and Kaplan 1996; Pevehouse, Nordstrom, and Warnke 2004; Pevehouse et al. 2020). Their activities influence the behavior of governments and multinational corporations in a broad range of issue areas and affect the lives of millions of people around the globe (Abbott and Snidal 1998; Barnett and Finnemore 2004). Yet, the resources IGOs have at their disposal to work towards the achievement of their goals are limited even for the most resourceful organizations, such as the European Union (EU) or the World Bank (Hawkins et al. 2006; Abbott et al. 2015). To reduce this disparity between goals and resources and become more effective governors, IGOs have developed strategies that include other governance actors, including governments, other IGOs, business, and civil society organizations, in various ways (Keohane and Nye 1974; Johnson 2014; Hale and Roger 2014; Abbott et al. 2015; Andonova 2017).

One such strategy is the participation in transnational public-private governance initiatives (TGIs) in which IGOs cooperate with governments, business, and civil society organizations to address global or transnational problems (Westerwinter 2021). Many of the most prominent IGOs are involved in transnational governance initiatives. The World Health Organization (WHO), for example, has facilitated the creation of a range of public-private partnerships in the health domain and engages in joint funding and project implementation with these institutions (Hanrieder 2015). Likewise, the United Nations Environment Programme (UNEP) is orchestrating and otherwise collaborating with a host of governance initiatives with different mixtures of participants to fight climate change and other environmental problems (Hale and Roger 2014; Van der Lugt and Dingwerth 2015). By participating in TGIs, IGOs obtain access to resources and capabilities that they would not have otherwise which, in turn, allows them to be more effective in addressing the global problems they were designed to tackle. Becoming engaged in TGIs and collaborating with business and particularly civil society organizations may also enhance the legitimacy of an IGO (Abbott and Snidal 2009).

The EU is part of this club of IGOs that are frequently involved in transnational governance initiatives. Supporting and participating in TGIs has become an important foreign policy instrument for the Union that it uses to pursue its interests within the complex institutional architecture of contemporary global governance (Abbott et al. 2015; Blauberger and Rittberger 2015; Commission of the European Communities 2001).² A prominent example is the Kimberley Process in which the EU represents its member states with a single voice to contribute to the regulation of the global trade in rough diamonds and to ensure that illegal profits from diamond mining do not spur civil war (Haufler 2015). Despite its importance for EU foreign policy, little is known about the scope and nature of the EU’s involvement in TGIs. How often does the EU participate in TGIs? What explains the EU’s involvement in TGIs? Why does the EU participate in some initiatives but not others? This paper begins to fill this research gap.

Today, TGIs are an essential part of the governance architectures in a broad range of global policy domains, including environmental protection, human rights, health, trade, finance, and security (Börzel and Risse 2005; Kaul 2006; Abbott and Snidal 2009; Westerwinter 2021). Some TGIs are among the most important governance actors in their issue area. In the security domain, for example, the International Code of Conduct for Private Security Service Providers’ Association, which brings together states, private security companies, and non-governmental organizations (NGOs), is the central, and to date, only effective institutional platform for regulating private security service providers at the global level (Avant 2016). Similarly, the global governance of ‘conflict diamonds’ depends on the Kimberley Process in which states, the EU Commission, the diamond industry, and NGOs set, monitor, and enforce standards for the import and export of rough diamonds (Haufler 2015). Finally, in the health domain, TGIs, such as the Global Fund to Fight AIDS, Tuberculosis and Malaria, are cornerstones of the institutional architecture of global governance that dwarf the relevance and financial capabilities of the WHO and other IGOs (Hanrieder 2015). As these examples indicate, TGIs are a central part of global governance in many issue areas of world politics. They often possess the authority to make and enforce decisions that change the behavior of governments, IGOs, and multinational corporations. They can also command considerable financial, informational, and other resources that are important assets to govern global problems. Thus, for IGOs that seek to improve their performance or to enhance their legitimacy, TGIs constitute attractive options for inter-institutional cooperation (Marx and Westerwinter, this issue; Biermann and Koops 2017).

In this paper, I provide the first systematic analysis of the EU’s participation in TGIs across a large number of governance initiatives. I empirically map the variation of EU involvement in TGIs using a new dataset on 636 transnational governance initiatives created in the period between 1885 and 2017 (Westerwinter 2021). In order to develop a first set of explanations for the observed variation in EU involvement in TGIs, I draw on an emerging theoretical debate about the interactions between IGOs and other types of global governance institutions (Abbott et al. 2015; Hale and Roger 2014; Andonova 2017; Tallberg et al. 2013). Specifically, I draw on theories of orchestration, access of transnational non-state actors to IGOs, and legitimacy of global governance institutions. I start from these theories because their focus on inter-organizational interactions is directly related to my research question and allows me to generate a first set of hypotheses to guide my empirical analysis.³ While theories of orchestration emphasize the importance of TGI resources and the similarity of the governance functions that TGIs and the EU provide to their members as facilitators of interactions, studies of the legitimacy of global governance institutions highlight the legitimacy enhancing effect of civil society involvement in global governance processes. Research on the formal access on transnational non-state actors to IGOs identify both resource and legitimacy advantages for IGOs as drivers of inter-organizational interactions.

I empirically probe these hypotheses using statistical analysis. My findings suggest that the size of a TGI in terms of its participants, the number of major powers involved, as well as the level of organizational structure and capacity are positively related to the likelihood with which the EU becomes involved in an initiative. I also find that the Union is particularly attracted to TGIs in which a large number of states, other IGOs, and business actors are engaged. By contrast, a large number of NGO participants reduces the likelihood that the Union becomes involved in a TGI. These findings suggest the need for the development of more nuanced theoretical explanations of the EU’s involvement in transnational governance and point toward additional quantitative and qualitative analysis in future research.

The paper makes two main contributes to the special issue and the broader literature on global governance. First, drawing on the theoretical framework developed in the introduction (Marx and Westerwinter, this issue), I provide the first systematic mapping of the EU’s interactions with TGIs across a large number of governance initiatives and issue areas. This is an important contribution because it enables researchers to gauge the scope of the EU’s engagement with TGIs as well as the features of the initiatives with which the EU cooperates. Furthermore, in combination with the papers by Roger, Tokhi, and Jordana et al., my contribution provides the basis for a comparative analysis of the intensity and form of the EU’s interactions with different types of global governance institutions (Marx and Westerwinter, this issue). Second, I contribute to the nascent literature on the interactions between IGOs and other types of global governance institutions more generally. I use extant theories of orchestration, access of transnational actors to IGOs, and legitimacy of global governance institutions to derive a first set of candidate explanations for the observable variation in the EU’s involvement in TGIs and test these hypotheses using statistical analysis. This analysis demonstrates the usefulness of existing theoretical arguments but also identifies potential for further theory development and testing based on qualitative case studies and more nuanced statistical analysis. My work also emphasizes the added value of quantitative data and methods in studying the EU’s role in global governance and provides a starting point for combining these methods with qualitative analyses.

I proceed in five steps. First, I empirically map the phenomenon of IGO participation in transnational governance initiatives highlighting especially the EU’s involvement as well as variation in the Union’s TGI participation that existing case studies have so far been unable to detect. Second, I draw on studies of orchestration, transnational actor access to IGOs, and legitimacy of global governance institutions to develop a first set of theoretical explanations for the EU’s involvement in TGIs. Third, I present the data and methods that I use to explore the empirical implications of my theoretical discussion. The next section presents the findings of my empirical analysis. The final section concludes.

Mapping EU participation in transnational public-private governance initiatives

Although scholarship on transnational governance initiatives has grown considerably over the past two decades (Schäferhoff, Campe, and Kaan 2009; Roger and Dauvergne 2016), little research exists that examines the interactions between transnational governance initiatives and IGOs, such as the EU. The few studies that have started to explore these interactions are predominantly based on single case studies (Abbott and Snidal 2009; Hanrieder 2015; Haufler 2015). By contrast, research that uses large-n databases is rare. Two exceptions are Hale and Roger (2014) and Andonova (2017). Hale and Roger’s analysis maps the involvement of IGOs and governments as orchestrators in 75 governance initiatives (2014). Andonova (2017) examines the relationship between the UNEP, the World Bank, the United Nations Children’s Fund (UNICEF), and the WHO and 347 global partnerships that have emerged around these organizations. While these studies make important contributions to our understanding of the relationship between IGOs and new forms of transnational governance, they are limited in their empirical focus on a specific issue domain (climate change) or particular IGOs (UNEP, World Bank, UNICEF, WHO). This raises questions about selection bias and the generalizability of findings. Furthermore, theoretically they only begin to explore the involvement of IGOs in transnational governance initiatives and do not systematically test hypotheses about the factors that make IGOs more likely to engage with these new governance arrangements. No study exists that systematically explores the involvement of the EU in new modes of transnational governance.

To advance our understanding of the participation of the EU in transnational governance initiatives, I use new data from the Transnational Public-Private Governance Initiatives in World Politics (TGIWP) database (Westerwinter 2021). This dataset contains detailed information on a broad range of characteristics of 636 TGIs created between 1885 and 2017 including the participation of IGOs in these initiatives. I combine this data with information from the Correlates of War (COW) Project’s data on IGOs (Pevehouse et al. 2020). This dataset enables me to map the participation of the EU in a large number of TGIs across issue areas and compare it to the involvement of other IGOs.⁴ It also allows me to begin examining the factors at the TGI and EU-TGI level that are related to the EU’s involvement in transnational governance initiatives.

The TGIWP data allows me to treat individual IGOs as the unit of analysis and examine their participation in TGIs. To identify IGOs that participate in transnational governance initiatives, I start from the COW definition of IGOs as international institutions that have the following characteristics: ‘(1) be a formal entity, (2) have states as members, and (3) possess a permanent secretariat or other indication of institutionalization such as headquarters and/or permanent staff’ (Pevehouse et al. 2020, 494). However, in contrast to the COW IGO data, I also include sub-bodies or special agencies of larger IGOs (e.g. sub-agencies of the United Nations or EU) as individual IGOs.⁵ This is in line with extant works on the orchestration of transnational governance arrangements by IGOs which cast the net wider than the relatively narrower COW IGO data and also examine IGO-TGI interactions that involve special agencies and programs of larger IGOs (Hale and Roger 2014; Van der Lugt and Dingwerth 2015).

I define TGI participation as the involvement of an IGO in a TGI in a broad sense, which may or may not include voting rights. An IGO may, for example, participate in a transnational governance initiative by orchestrating its activities (Hale and Roger 2014; Abbott et al. 2015), engaging in joint projects, or funding its activities. Organizations that have observer status or are otherwise involved in the work of a TGI on a regular basis (e.g. by being members of a working group of an initiative) are also considered as participants. For example, the World Bank works on joint projects with a range of TGIs, including the Global Water Partnership and Sustainable Energy for All. Similarly, the EU is a full member of the Kimberley Process with voting rights and has chaired the initiative’s working group on monitoring for many years. EU participation in TGIs is a form of inter-organizational interactions as discussed in the introduction to this special issue (Marx and Westerwinter, this issue). It may take formal as well as more informal forms and establishes the basis for regular contacts between the EU and the governance initiatives in which it becomes involved.

In total, my data identifies 375 IGOs that are involved in at least one of 567 initiatives that were operational in 2017. These IGOs include large, global multilateral organizations, such as the World Bank and the WHO, regional organizations with broad mandates, such as the EU and the African Union, as well as specialized entities, such as the International Organization of Vine and Wine and the International Olive Oil Council. Together, these organizations had 2,401 TGI involvements in 2017. On average, the IGOs in the dataset participate in 6.40 TGIs with a standard deviation of 17.64, indicating considerable variation in transnational governance participations across organizations.

Figure 1 shows the IGOs that are most involved in TGIs in 2017. Notably, the World Bank and the EU, followed by UNEP clearly lead the way in terms of TGI involvements. In 2017, they participated in considerably more TGIs than the Food and Agriculture Organization, UNICEF, and the WHO, which are the fourth, fifth, and sixth most involved IGOs in my dataset. With their 177 and 166 participations, they together account for about 14% of all IGO-TGI involvements in the data. Other organizations that are among the twenty most involved IGOs include the Organization for Economic Cooperation and Development, the International Labour Organization, and the World Food Programme. Several development banks, such as the Inter-American Development Bank, the Asian Development Bank, and the African Development Bank, and regional organizations other than the EU, such as the African Union, are also involved in a large number of TGIs. These findings resonate with existing studies that identify the World Bank, WHO, UNEP, and UNICEF as frequently involved in transnational governance initiatives (Hale and Roger 2014; Hanrieder 2015; Andonova 2017). More importantly, they move beyond these studies by showing that the phenomenon of IGO involvement in transnational governance is much broader than these few prominent cases suggest. These first insights also identify, for the first time, rich variation in IGOs’ involvement in transnational governance initiatives. IGOs that are frequently involved in TGIs operate in a broad range of issue areas, differ in their size and mandates, and include universal as well as restricted membership organizations.

Figure 1. IGO participation in TGIs, 2017. International Bank for Reconstruction and Development (World Bank), European Union (EU), United Nations Environment Programme (UNEP), Food and Agriculture Organization (FAO), United Nations International Children’s Emergency Fund (UNICEF), World Health Organization (WHO), Inter-American Development Bank (IADB), United Nations Development Programme (UNDP), Asian Development Bank (ADB), African Development Bank (AfDB), Organization for Economic Cooperation and Development (OECD), United Nations Educational, Scientific and Cultural Organization (UNESCO), United Nations Human Settlements Programme (UN Habitat), International Fund for Agricultural Development (IFAD), United Nations (UN), World Food Programme (WFP), International Finance Corporation (IFC), United Nations Development Fund for Women (UN Women), International Labour Organization (ILO), United Nations Industrial Development Organization (UNIDO).

In addition to the frequency of the EU’s involvement in TGIs, we can also begin to examine in what initiatives the EU participates. In terms of size, as Figure 2 shows, the 131 TGIs in which the EU is involved have on average 160 participants. They are also relatively strongly populated by democratically governed countries, as indicated by an average Polity score (Jaggers and Gurr 1995) of 6.84 among the states that participate in the TGIs in which the EU is involved. Furthermore, the EU tends to be involved in TGIs with governance task-issue area profiles that are relatively dissimilar from its own functional and issue area scope (with an average of 0.18 on a similarity measure that varies between 0 and 1). Finally, the TGIs in which the EU participates are on average relatively informal in terms of their organizational structures. Only 28% of the initiatives in which the EU is involved have both a secretariat and a founding document.

Figure 2. Features of TGIs in which the EU participates.

These observations demonstrate rich variation in the EU’s participation in transnational governance initiatives. They raise an important question: What motivates the EU to become involved in some TGIs but not others? In the next section, I begin to develop theoretical explanations that help to address this question.

Explaining EU participation in transnational public-private governance initiatives

The previous section showed that the EU participates in some types of TGIs more than others. In this section, I draw on an emerging debate about the interactions between IGOs and other types of global governance institutions in order to begin to develop explanations for this variation. I particularly focus on arguments about orchestration, access of transnational actors to IGOs, and IGO legitimacy. While orchestration theory emphasizes IGO access to the resources of transnational governance initiatives and the similarity of the functions that IGOs and TGIs perform (Abbott et al. 2015), research on IGO legitimacy highlights the potential legitimacy enhancing effect of civil society involvement in global governance (Tallberg, Bäckstrand, and Scholte 2018). Studies of the formal access of transnational non-state actors to IGOs examine both the resources and legitimacy that transnational actors may contribute to the work of IGOs (Tallberg et al. 2013, 2014). These arguments are grounded in different theories that make specific assumptions about the behavior of IGOs, states, and other governance actors. Despite some differences in their underlying assumptions, I argue that they are not mutually exclusive but can complement each other in facilitating our understanding of the EU’s involvement in transnational governance.⁶ The theories from which I derive my hypotheses are not the only ones that are relevant for studying the EU’s participation in TGIs. However, because of their specific focus on inter-organizational interactions they speak most directly to my research question. As I elaborate in greater detail in the conclusions, future research may fruitfully apply other theories in order to develop more nuanced explanations for the EU’s engagement with transnational governance initiatives.

Throughout my theoretical discussion, I conceptualize the EU as an IGO. This is in line with extant theories and datasets of IGOs (Tallberg et al. 2013; Hooghe and Marks 2015; Pevehouse et al. 2020). While this approach may under emphasize some of the special features of the EU’s institutional architecture, it enables me to compare the Union to a broader set of organizations and draw on theories developed to understand the actions and interactions of these organizations. It also allows me to contextualize my empirical findings and use insights gained from studying the EU’s involvement in transnational governance initiatives to inform debates about the interactions of IGOs with TGIs more broadly.

Orchestration is a special form of inter-organizational interaction between IGOs and other governance actors, which serve as intermediaries between the IGOs and their governance targets (Abbott and Snidal 2009; Hale and Roger 2014; Abbott et al. 2015; Abbott, Green, and Keohane 2016; Abbott 2018). In IGO orchestration, intermediaries may be other IGOs, transgovernmental networks, business, and TGIs. When an IGO, such as the EU, orchestrates a TGI, the IGO provides the TGI with support and ideational resources, which empower the TGI and increase its ability to pursue its goals successfully. Given the shared goals that the orchestrating IGO and the intermediary TGI seek to attain, increasing the intermediary’s effectiveness also makes the IGO more likely to perform well. In fact, orchestration is a strategy for IGOs to achieve their goals effectively in light of a shortage of their own capabilities (Abbott et al. 2015). Thus, orchestration has the potential to enhance the effectiveness of intermediaries, such as TGIs, as well as the effectiveness of the orchestrating IGOs. In light of this potential of orchestration to enhance IGO effectiveness and to facilitate the achievement of IGO governance goals, it might be that IGOs that lack the resources to govern effectively on their own may be particularly prone to become involved in TGIs to obtain access to relevant governance resources that they would not have otherwise.

While orchestration holds the potential to enhance an IGO’s capabilities, it also requires substantial amounts of resources to begin with. Participating in a TGI is costly. This applies to all actors involved including IGOs. IGO staff needs to spend some of their scarce time obtaining and processing information about the governance activities of a TGI and attend meetings. If the goal is to shape the activities of a TGI in accordance with IGO preferences, IGO representatives in TGIs also need to engage in some way in agenda-setting, proposal making, document drafting, lobbying for their positions, negotiating, implementing decisions, and monitoring. Because of their tripartite nature, negotiations in TGIs are often complicated and protracted (Abbott and Snidal 2009). In addition, TGI activities often occur across the globe. Together, this implies that engaging in the operations of a TGI and effectively orchestrating it for the purposes of achieving governance goals can be costly for an IGO. Thus, while the orchestration of TGIs may be an appealing governance strategy for IGOs with small budgets, small staff, and little resources in general, it might in fact be that only IGOs which possess a certain amount of resources and organizational capacity are able to pursue this strategy effectively. At the same time, IGOs that control large amounts of resources on their own may have no need to engage in orchestration to achieve their goals. They may instead rather pursue governance strategies that provide them with tighter control over processes and outcomes (Abbott et al. 2015).

In sum, orchestration theory suggests that IGOs that lack governance resources have incentives to become involved in transnational governance initiatives. At the same time, we can argue that because participation in TGIs is costly it might only be those IGOs that have a sufficient amount of resources at their disposal that are able to pursue this strategy to further enhance their governance effectiveness. With respect to the EU this suggests that we should expect the EU to be more often involved in TGIs than other, relatively less resourceful IGOs. At the same time, if part of the Union’s motivation to become involved in TGIs is access to additional resources, then we would expect the EU to be particularly prone to join initiatives with large amounts of resources at their disposal, as for example expressed in their number of participants and particularly resourceful participants, such as states that are considered major powers in the international system.

Theories of orchestration also emphasize the general fit of the governance goals of an orchestrating IGO and the orchestrated TGI (Abbott et al. 2015). Since IGOs have no formal hierarchical authority over transnational governance initiatives and other types of potential intermediaries, they rely on shared goals and missions as basis for cooperation. This suggests that IGOs would choose their TGI partners according to the extent to which their missions or mandates overlap with their own goals. Initiatives that pursue the same tasks in the same policy areas would be attractive candidates for joint governance efforts. By contrast, IGOs could also seek to partner with initiatives that perform functions that are complementary to their own. For example, a standard setting IGO could have an interest in collaborating with an implementing TGI that focuses on executing projects on the ground. This would suggest that IGOs try to partner with TGIs that perform different functions in either the same or different policy fields. One condition that could determine whether governance function and issue area fit makes IGO participation in TGIs more or less likely could be the amount of resources or governance capacity an IGO has at its disposal. Organizations with little resources or capacity may, for example, focus on initiatives that complement their governance activities, while more resourceful organizations may be able to focus on initiatives that match their own functions and policy focus more narrowly. Given the resources that the EU has at its disposal compared to other IGOs, we may expect the Union’s TGI involvement to resonate with the latter pattern.

Theories that focus on the access of transnational non-state actors to IGOs highlight the potential resource and legitimacy advantages for IGOs that are related to providing non-state actors with access to the decision-making bodies and the broader operations of an IGO (Tallberg et al. 2013). Allowing non-state actors, such as NGOs and business, to participate in IGO deliberations and decision-making provides the IGO with access to information, funds, technical capacity, and other resources, to which it would have otherwise no access. This resonates with the emphasis on the governance enhancing effects of TGI involvement in orchestration theory. In addition, including non-state actors in IGO decision-making can also be a strategy to increase the democratic legitimacy of an organization. IGOs are often confronted with criticism about their lack of transparency, inclusiveness, and accountability (O'Brien et al. 2000; Tallberg, Bäckstrand, and Scholte 2018). The United Nations Security Council and international financial institutions, such as the International Monetary Fund and the World Bank, are prominent examples. One way to address these concerns is to give transnational non-state actors the opportunity to become involved in IGO decision-making. They may be allowed to provide information to government representatives and IGO staff, attend meetings, and make statements during negotiations (Tallberg et al. 2013). In turn, such participation from non-state actors may increase the perceived legitimacy of the resulting decisions and the IGO as a whole.

This legitimacy seeking strategy of IGOs that is reflected in the granting of formal access of non-state actors to IGO decision-making and deliberations may also be a driver of IGO involvement in TGIs. By becoming involved in a TGI, an IGO obtains access to and develops collaborations with NGOs, business, and other non-state actors. In doing so, the IGO may develop a reputation as particularly non-state actor friendly and open toward their involvement which, in turn, might positively affect its perceived legitimacy. In fact, pursuing TGI participation and providing transnational actors with formal access to its decision-making may be two sides of the same coin both geared towards increasing an IGO’s legitimacy. The fact that the World Bank is both a champion in terms of TGI participation and has improved its accessibility for transnational actors in the recent years points into this direction. Transnational access and TGI participation may also be mutually reinforcing. IGOs that provide transnational actors with access may become involved in TGIs because of the advice and lobbying of non-state actors. On the other hand, IGOs that are involved in TGIs may over time decide to formally open up toward transnational actors based on positive learning experiences within multi-stakeholder initiatives and the forming of trust and good working relationships with non-state actors. As a consequence, we may expect that IGOs with high levels of transnational access are also frequently involved in TGIs. In addition, we might expect that factors that positively affect transnational access, such as the level of democracy among the member states of an IGO (Tallberg et al. 2014), also facilitate the TGI participation of IGOs.⁷

Data and methods

I explore the link between EU participation in TGIs and a set of explanatory variables at the TGI and EU-TGI level based on the information on IGO participation in TGIs in the Transnational Public-Private Governance Initiatives in World Politics data (Westerwinter 2021). I combine this data with information from the COW Project’s data on IGOs (Pevehouse et al. 2020)⁸ and new data on IGO governance functions. For my statistical analysis, I focus on the EU and all TGIs in the TGIWP dataset that have been created up until 2014 and were not terminated by then. In total, I include 567 TGIs in my analysis. I use data on the EU’s TGI participations in 2014 because of limited data availability on the independent variables that I include in my analysis.

I use an undirected dyad setup in which each observation in the data is a pair of the EU and one TGI. In total, my data consists of 567 undirected EU-TGI dyads. My main dependent variable, EU participation, is an indicator variable that captures the participation of the EU in a given TGI_i in 2014. Since my dependent variable is binary, I use logistic regression methods to estimate my statistical models (Long 1997). I explore the relationship between EU participation in TGIs and two sets of explanatory variables that capture features of TGIs as well as properties of EU-TGI dyads.

I include a range of TGI covariates in my analysis. First, I include the natural logarithm of the total number of TGI participants. Based on the theoretical discussion above, we may expect that larger TGIs provide the EU with a larger pool of resources to tap into and are therefore more attractive partners if EU participation is motivated by gaining access to resources controlled by other governance actors. I also break the total participant count down into the number of state, IGO, company, and NGO participants a TGI has in order to examine whether it is particular groups of TGI members that the EU is especially interested in reaching out to. Furthermore, I include the number of major power states participating in a TGI to assess whether the EU is more attracted by TGIs that include major power countries in their memberships which are likely to be more resourceful than other countries.

The average level of domestic democracy of state members of TGIs is another variable that I incorporate in my models. This variable serves as a proxy for the democratic qualities and good governance qualities of a TGI which may be attractive for the EU when it searches for partner institutions (Tallberg et al. 2013). Including a variable that captures whether a TGI has a founding document and a secretariat, I capture the level of institutional formality of TGIs (Westerwinter, Abbott, and Biersteker 2021). The data for this variable comes from the institutional design information contained in the TGIWP data. This variable provides a proxy for the level of organizational structure and capacity of a governance initiative. Finally, I control for TGI founding year as well as the number of issue areas and governance functions that an initiative addresses.

At the EU-TGI dyadic level, I use a variable that captures the similarity of the EU’s and a TGIs’ governance functions and policy scopes. This variable is based on the cosine similarity of governance functions and issue areas of the EU and a given TGI and is computed as the product of these two similarities. It ranges from 0 to 1 with 0 indicating an EU-TGI pair with completely different governance functions and policy scope and 1 referring to identical institutions. I use this variable to examine whether the EU selects the TGIs in which it participates based on considerations about complementarities or differentiation of functions and policy domains. Table 1 reports descriptive statistics for all variables included in my analysis.

Table 1. Summary statistics dependent and independent variables.

	Mean	Std. dev.	Min.	Max.	Observations
EU participation	0.231	0.422	0.000	1.000	567
TGI start year	2001.125	14.367	1885.000	2017.000	567
Log num. TGI participants	3.915	1.180	1.386	9.485	520
TGI major powers	2.427	2.349	0.000	10.000	478
TGI average democracy	7.426	3.547	−10.000	10.000	478
Formal TGI	0.192	0.394	0.000	1.000	567
Governance-issue similarity	0.185	0.141	0.000	0.612	551
Log num. state participants	1.686	1.137	0.000	5.257	544
Log num. IGO participants	1.293	0.818	0.000	4.043	550
Log num. company participants	2.225	1.660	0.000	9.031	528
Log num. NGO participants	2.462	1.331	0.000	7.924	534

Results

Table 2 presents my first set of results. Models 1–3 include first TGI characteristics and then add the governance-issue area similarity measure at the dyadic level. We see that the EU is more likely to join a TGI as the size of the initiative in terms of participants grows. In other words, TGIs with larger memberships are more likely to attract EU involvement. This relationship is also substantively relevant. If we increase the Log TGI participants variable by one standard deviation beginning at its sample mean, we observe an increase in the predicted probability of EU participation from 0.26 to 0.31 or by 19%. Similarly, a one-standard deviation increase in the number of major power participants of a TGI increases the predicted probability of the EU becoming involved by 16%. These findings suggest that the EU seeks out transnational governance initiatives that bring together a large number of public and private actors and constitute a rich pool of material and ideational governance resources.

Table 2. Logistic regression estimates EU participation in TGIs.

	(1)	(2)	(3)
TGI start year	$-{0.004}^{\ast }$	$-{0.003}^{\ast }$	$-{0.003}^{\ast }$
	$(0.002)$	$(0.002)$	$(0.002)$
Log TGI participants	${0.051}^{\ast \ast }$	${0.044}^{\ast }$	${0.042}^{\ast }$
	$(0.018)$	$(0.018)$	$(0.019)$
TGI major powers	$0.016$	${0.018}^{\ast }$	${0.018}^{\ast }$
	$(0.009)$	$(0.009)$	$(0.009)$
TGI average democracy	$-0.007$	$-0.007$	$-0.006$
	$(0.006)$	$(0.006)$	$(0.006)$
Formal TGI		${0.141}^{\ast \ast }$	${0.153}^{\ast \ast }$
		$(0.052)$	$(0.053)$
Governance-issue similarity			$-0.071$
			$(0.168)$
BIC	$529.147$	$527.786$	$537.905$
Observations	$446$	$446$	$435$

Standard errors in parentheses. All significance tests two-tailed.

${ }^{\ast \ast \ast }p<0.001$, ${ }^{\ast \ast }p<0.01$, ${ }^{\ast }p<0.05$

We also observe a consistently positive and statistically significant relationship between the level of institutional formalization of a TGI and the EU’s propensity to become involved in an initiative. This suggests that the EU prefers to partner with transnational governance initiatives that have a certain level of organizational structure and capacity which allows them to engage in governance processes effectively. Holding all other variables constant at their sample means, changing from an informal to a more formalized TGI increases the predicted probability of EU involvement from 0.23 to 0.38 or 65%. Importantly, as the marginal effects plot in Figure 3 indicates, the formality of a TGI’s institutional design is the substantively most important predictor of the likelihood with which the EU becomes involved in a transnational governance initiative.

Figure 3. Marginal effects of independent variables on EU participation in TGIs.

The average domestic democracy of TGI state members is negatively associated with EU participation in an initiative which constitutes a puzzle that requires more in-depth research. However, this relationship is not distinguishable from zero. There is also no statistically significant relationship between the governance task-issue area similarity of the EU and a TGI and the Union’s likelihood of becoming a partner in an initiative.

To further explore the relationship between TGI participants and EU involvement in transnational governance initiatives, I investigate how different participant groups of TGIs shape the EU’s involvement. In particular, I distinguish between state, IGO, company, and NGO participants of transnational governance initiatives. Table 3 reports the results of this analysis. While the overall results are in line with the findings from the simpler models, they also reveal additional insights. First, all different categories of TGI participants are relevant for the EU’s propensity to become involved in a transnational governance initiative. The number of states, IGOs, and companies that participate in a TGI increase the likelihood that the Union joins the initiative. All three variables are statistically significant in the model. In substantive terms, the number of IGO participants in a TGI has the strongest effect on EU involvement. This suggests that the EU is particularly interested in joining TGIs in which other IGOs are participating. This may be driven by the general attractiveness of certain TGIs to IGOs in terms of resources, governance functions, or institutional design features. It may also be a consequence of the EU trying to be involved in the governing activities of initiatives in which many other organizations are interested in order to exert control over agenda-setting, decision-making, and other governance activities that occur in these initiatives. It could also be an indication that transnational governance initiatives are a vehicle for IGOs to interact and collaborate with each other. More research and especially qualitative case study work is required to further explore these conjectures.

Table 3. Logistic regression estimates EU participation in TGIs.

	(4)
TGI start year	$-0.003$
	$(0.002)$
Log states	${0.052}^{\ast }$
	$(0.024)$
Log IGOs	${0.157}^{\ast \ast \ast }$
	$(0.028)$
Log companies	${0.032}^{\ast }$
	$(0.014)$
Log NGOs	$-{0.039}^{\ast }$
	$(0.018)$
TGI major powers	$0.002$
	$(0.010)$
TGI average democracy	$0.002$
	$(0.006)$
Formal TGI	${0.145}^{\ast \ast }$
	$(0.051)$
Governance-issue similarity	$-0.103$
	$(0.162)$
BIC	$512.442$
Observations	$435$

Standard errors in parentheses. All significance tests two-tailed.

${ }^{\ast \ast \ast }p<0.001$, ${ }^{\ast \ast }p<0.01$, ${ }^{\ast }p<0.05$

By contrast, and perhaps surprising given scholarly work on global governance legitimacy and policy statements from the EU about its interest in multi-stakeholder governance, a large number of NGO participants in TGIs makes the EU less likely to become involved in a transnational governance initiative. This finding in combination with the statistically insignificant result for the level of democracy in the domestic regimes of the countries that are involved in TGIs raises questions about arguments that emphasize transnational governance as a vehicle for increasing the inclusiveness and democratic legitimacy of global governance. More research is needed to explore why the EU is less involved in transnational governance initiatives with large numbers of NGO members.

Conclusions

In this paper, I made first advances in systematically mapping and exploring the European Union’s involvement in transnational public-private governance initiatives in which states, IGOs, business actors, and NGOs work together to govern global problems. Using new data on IGOs’ involvement in 636 TGIs created between 1885 and 2017, I show that the EU is together with the World Bank the leading IGO in terms of transnational governance participations. I also examine the features of TGIs in which the EU has become involved. My exploratory regression analyses indicate that the EU tends to participate in transnational governance initiatives that are large in terms of their overall number of participants and major powers as well as institutionally formalized. Exploring different types of TGI participant groups in greater detail, I further find that the EU is particularly attracted by transnational governance initiatives with large number of states, other IGOs, and companies. By contrast, the presence of large numbers of NGOs among the participants of a TGI makes the EU less likely to join a governance initiative. These findings resonate with theoretical arguments that highlight the access to governance resources that transnational governance participation provides to IGOs and raise questions about theories about the democratic legitimacy and civil society participation in global governance institutions.

My paper makes important contributions to the study of how the EU interacts with other global governance institutions as well as how it operates as a player within the complex institutional architecture of contemporary global governance. I conclude by discussing four areas for future research. First, while the EU is one of the most prolific IGOs when it comes to participating in transnational governance initiatives, it is selective about which initiatives it joins. The tendency toward joining TGIs with large numbers of state and business participants suggests that part of the EU’s motivation to become involved is based on the access to additional governance resources. This is in line with theories of transnational actor access to IGOs and especially orchestration which highlight IGOs’ ability to tap into the resources of other governance actors as a main driver of inter-organizational interactions. By contrast, the propensity to join TGIs in which many other IGOs participate points toward additional strategic considerations that go beyond mere access to material and other resources. For example, the Union may have an interest in having access to governance processes in which other IGOs relevant for its activities are involved and influencing these processes in line with its own preferences. Alternatively, the EU may also use TGIs as forums through which to collaborate with other IGOs. However, the results of my exploratory regression analyses are only indicative since model coefficients only summarize patterns of correlations in my data and do not establish causal relationships between independent and dependent variables. In-depth qualitative case studies and more nuanced statistical analysis are needed to explore the strategic motivations underlying the EU’s involvement in transnational governance initiatives and elaborate on causal mechanisms. Future research would particularly benefit from applying mixed-method research designs that exploit the strength of both quantitative and qualitative methods to investigate these questions.

Second, the negative relationship between the number of NGO participants of TGIs and the statistical insignificance of domestic democracy of TGI state members for the Union’s involvement in transnational governance question existing research on transnational actor access to IGOs and global governance legitimacy. These findings also sit uneasy with policy discussions that highlight the EU as a driver of more democratic formats of global governance. While my statistical analysis in this paper is of a primarily exploratory character, this finding suggests that more nuanced theories and empirical inquiries of the EU as proponent of a more democratic global governance are needed to fully understand the Union’s strategic engagement with other types of global governance institutions.

Third, I have begun to theorize the factors that facilitate the EU’s involvement in TGIs drawing on theories of orchestration, transnational actor access to IGOs, and the legitimacy of global governance. While these are useful starting points for hypothesis development, other theories, including multi-level governance theories and models of inter-organizational collaboration, might provide additional insights that enable further advances in theory building. Similarly, my exploratory regression analyses have only controlled for a few potentially important independent variables. Others, such as the annual budgets, staff sizes, and public support of TGIs, are additional candidate explanations for EU involvement in transnational governance initiatives that future data collection efforts may incorporate in their analysis.

Finally, I have conceptualized the EU as one of many IGOs engaged in global governance. Yet, the EU can also be considered a special organizational entity given its institutional structure, its level of delegated authority, and the amount of resources it has at its disposal. Taking these specificities of the EU as organization into account reduces the generalizability of the theoretical and empirical insights toward other IGOs but may gain additional theoretical leverage for explaining the EU’s involvement in transnational governance initiatives. Future research may explore this option in greater depth.

Taken together, these research directions promise to bring about a more systematic and nuanced understanding of how different types of global governance institutions interact and specifically how the EU navigates the complex institutional architecture of contemporary global governance. They can also inform theories of international cooperation more generally which often focus on individual global governance institutions in isolation. Importantly, this agenda can lay the groundwork for research that examines the consequences of the complex interdependencies among IGOs and other forms of global governance. Given the importance of transnational governance initiatives and inter-organizational interactions for effectively governing the most pressing global problems of our time, such advances will benefit students of international cooperation and policy-makers alike.

Acknowledgments

I would like to thank Axel Marx, Charlie Roger, and the participants of the GLOBE project for the stimulating discussions and continued feedback that facilitated the development of this paper. I am also grateful to Spyros Blavoukos, Tom Delreux, Joseph Earsom, Sandra Lavenex, Lucia Quaglia, Yf Reykers, and the participants in the “The European Union and International Regime Complexes” workshop organized at the University of Louvain for their engaging comments.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This research was funded by the Swiss Network for International Studies and the Basic Research Fund of the University of St. Gallen.

Notes

1. I use the term intergovernmental organizations to refer to global governance institutions that are constituted by states as the only members and have some form of organizational structure and agency (Pevehouse et al. 2020). The alternative term international organizations (IOs), as used by some contributions to this special issue, is sometimes used interchangeably and sometimes to refer to a broader set of global governance institutions that includes IGOs and other types of institutions. Since in this paper I analyze the involvement of the EU as an IGO in transnational governance initiatives, I use the more specific term IGO in order to increase conceptual precision and avoid confusion.

2. See also Joint Communication to the European Parliament and the Council on strengthening the EU’s contribution to rules-based multilateralism, High Representative of the Union for Foreign Affairs and Security Policy, Brussels, 17 February 2021.

3. Future research may explore the usefulness of other theories for advancing our understanding of the EU’s involvement in transnational governance.

4. As I explain in greater detail below, conceptualizing the EU as an IGO enables me to compare the Union’s involvement in transnational governance initiatives to the involvement of other organizations. It also allows me to use theories about IGOs more generally to begin to identify candidate explanations for the EU’s involvement in transnational governance initiatives.

5. In cases where the EU participates in a TGI, I take into account whether it is the EU as an IGO and/or its member states that participate.

6. See Jupille, Mattli, and Snidal (2013) for a similar pragmatic approach that combines different strands of institutionalist theory in international relations to explain the use, creation, and change of international institutions.

7. At the same time, TGI participation may be an alternative strategy to increase collaboration with transnational actors for those IGOs that would like to, but are not able to, increase formal access because they are, for example, blocked by their internal decision-making procedures and lack of state support for opening up. This would suggest that it is particularly those IGOs that lack transnational access, but would be interested in increasing it, which participate frequently in TGIs. However, while the level of transnational access is observable, the preferences of IGOs are not, which makes it difficult to operationalize this hypothesis.

8. I thank Jon Pevehouse for sharing the COW Project’s IGO data prior to its public release.