Brokering Climate Action: The UNFCCC Secretariat Between Parties and Nonparty Stakeholders

Our article aims to better understand the role of the secretariat of the United Nations Framework Convention on Climate Change (UNFCCC) in the increasingly complex global climate governance structure. We employ an innovative approach to addressing this issue by systematically examining the climate secretariat’s relations with the main groups of actors involved in this policy domain, in particular with nonparty actors. In a first step, we use social network analysis (SNA) to examine the secretariat’s relations with nonparty and state stakeholders and to identify its position in the UNFCCC policy network. An understanding of where the climate secretariat stands in the global climate governance network and which actors it interacts with most allows us to draw preliminary conclusions about the ways in which it connects with other stakeholders to influence global climate policy outputs. In a second step, we conduct thirty-three semistructured interviews to corroborate the results of the SNA. Our findings lend support to the argument that the climate secretariat may gradually be moving from a rather neutral and instrumental stance to playing a proactive and influential role in international climate governance. It aims to increase its political influence by establishing strategic links to actors other than the formal negotiation parties.

Stripple 2008, 368) have taken ownership of the implementation of "a universal, ambitious climate agreement that is differentiated, fair, lasting, dynamic, balanced, [and] legally-binding" 1 in an attempt to stay below 2°C.
In this article, we explore the role of the UNFCCC secretariat within this unique global governance structure. One of the climate secretariat's duties is to promote and coordinate the so-called Nationally Determined Contributions (NDCs), domestic mitigation measures pursued by the parties (Article 4, paragraph 2, 12, Paris Agreement), via the continuous exchange of information. Based on the notion that the climate secretariat "strives to keep all stakeholders informed on the negotiating process … through a variety of communication products" (United Nations Framework Convention on Climate Change [UNFCCC] 2019), while also facilitating the NDCs, we argue that the climate secretariat makes use of the multiactor and multilevel structure of the global climate regime to gauge national positions and mobilize climate action by nonparty stakeholders to achieve the goal of the Convention. This helps the member states reach compromises by extending the policy debate to different actors inside and outside of government who would otherwise probably not connect.
The literature on international bureaucracies and treaty secretariats has mostly examined the important and influential roles secretariats can and do play in international negotiations (see, e.g., Bauer et al. 2012;Ege 2016, 2017;Busch 2009;Jinnah 2014;Johnson 2013). We still lack knowledge about the role and position of secretariats in their respective networks and how they interact with and likely even influence different kinds of stakeholders. Moreover, the literature on the global climate governance regime has focused mainly on the interaction between negotiation parties and nonparty actors. An important strand of this literature has studied the authority of nonparty stakeholders and their influence on decision makers and negotiation outcomes (Böhmelt 2013;Böhmelt and Betzold 2013;Lund 2013;Moussu 2015;Rietig 2014;Nasiritousi et al. 2014Schroeder and Lovell 2012;Tallberg et al. 2013). Despite shedding light on the important role of international treaty conventions, such as the UNFCCC, and the activities of nonparty stakeholders that might contribute to societal transformation in global climate governance, these studies have neglected the link between convention secretariats and party and nonparty stakeholders.
The climate secretariat needs to master the unique architecture of multilevel global climate governance despite its narrow formal mandate that emphasizes its logistical and informational role while explicitly exempting it from taking a more active part in multilateral negotiations (Hickmann et al. 2019;Jörgens et al. 2016;Kolleck et al. 2017b). Consequently, when communicating and cooperating with different kinds of stakeholders to guarantee the successful realization of the measures agreed upon in the Paris Agreement, the secretariat always strives to act in a balanced and impartial way (Well et al. 2020). By scrutinizing the position of the climate secretariat within the UNFCCC regime and how it engages with negotiation parties and nonparty stakeholders, this article seeks to contribute to the literature on secretarial interaction patterns as a first step toward a better understanding of the role played by the climate secretariat in the global climate governance regime.
We apply a mixed methods design to gain deeper insight into the ways in which the climate secretariat engages with others. In a first step, we employ tools of social network analysis (SNA) to study how the climate secretariat interacts and cooperates with the entire range of actors operating in the climate policy realm. Based on an original data set derived from a large-N survey among organizations in the field of global climate governance, our SNA maps networks of policy-specific communication and cooperation among diverse actor groups and assesses the position that administrative organizations occupy within these networks. The choice of this first approach is based on the fact that existing studies usually focus on just one actor type, which allows them to draw only limited conclusions about the relationships between various actor types within the same negotiations (notable exceptions include Böhmelt and Betzold 2013;Lund 2013). In a second step, we conduct document analysis and thirty-three structured interviews with party and nonparty stakeholders as well as with members of the climate secretariat between 2015 and 2018-during and between the Conferences of the Parties (COP) 21 to 23-to narrow down our assumption and answer our research question. Interviewees were asked about the nature of the relationship and extent of interaction. The combination of a survey-based SNA with interviews and document analysis enables us to reconstruct the policy-specific information flows and identify the actors who hold positions that allow them to connect different groups of actors.
The article is structured as follows. We first discuss the changing perception of the role of international public administrations (IPAs)that is, the administrative bodies of international organizations (IOs)in international treaty negotiations in general and then formulate our hypothesis. Taking into account the unique characteristics of the global climate governance structure, we then describe the interaction between the three types of actors-the climate secretariat, negotiation parties, and nonparty stakeholders. The analysis of our findings allows us, inter alia, to test our assumptions and discuss our findings on the ways in which the climate secretariat interacts with other stakeholders.

The Changing Conceptualization of International Treaty Secretariats
International treaty secretariats are established by states as formal bodies to provide the parties to an intergovernmental convention with a common knowledge base irrespective of national capacities (Depledge 2005). They shall support governments and nonparty stakeholders in subsequent rounds of issue-specific negotiations within multilateral treaty regimes through the provision of technical, legal, and procedural expertise-as well as normative and diplomatic knowledge (Bauer 2006;Jörgens et al. 2016;Kolleck et al. 2017a). Scholars in the field of International Relations and IPA have long treated IOs and their bureaucracies as instruments of nation-states rather than as actors in their own right (Ness and Brechin 1988). Consequently, until recently, research on IOs has been primarily concerned with exploring whether IPAs represent a challenge to state power and the political control of bureaucracies.
More recently, this focus has shifted to an empirical examination of the degree to which international secretariats, and IPAs more generally, exert autonomous influence on politics and policies. The expectation that IPAs may constitute partially autonomous and potentially influential actors of global governance is based on a series of assumptions. First, IPAs are often perceived as the institutional memory of their respective policy regime (Bauer 2006), that is, to have superior "informal knowledge about the history and evolution of institutional processes" (Jinnah 2010, 62). Moreover, they are said to often have an informational advantage on technical and legal issues over their political masters (Barnett and Finnemore 2004;Derlien et al. 2011). IPAs are no longer understood as just epiphenomena of national governments but rather are regarded either as self-interested actors using information asymmetries to their advantage vis-à-vis their principals, namely, the nation-states, or as agents of the global common good, whose actions go at least partially beyond national interests. Against this backdrop, scholars have studied the agency and influential role of IPAs in multilateral negotiations by inquiring whether, how, and to which degree they exert influence on international policy making (see, e.g., Bauer et al. 2012;Bauer and Ege 2017;Busch 2009;Jinnah 2014).
Secretariats of multilateral environmental conventions may try to mobilize support to advance their own proposals and to build momentum for agreement (Abbott and Snidal 2010;Jörgens et al. 2016;Kolleck et al. 2017b). Some IPAs, for example, the desertification and the biodiversity secretariat, have framed discourses and problem perceptions in line with their governance preferencesdespite narrow and issue-specific mandates, close supervision by their principals, and relatively limited scientific and administrative capacities (see, e.g., Bauer 2006;Conliffe 2011;Depledge 2007;Jinnah 2011;Jörgens et al. 2016;Kolleck et al. 2017a;Siebenhüner 2009). They have raised their convention's profile, set items on the agenda, introduced amendments to draft proposals, and promoted the institutionalization of their conventions (Bauer 2009;Michaelowa and Michaelowa 2017;Pallavi 2011). Moreover, many of the initiatives and compromises proposed by the chairs or presidencies of multilateral frameworks are traceable back to the secretariat (Depledge 2005). Eckhard and Ege (2016, 961) conclude that IPAs act as "autonomous actors with some degree of influence on global public policy." Today, it is not only the state signatories of a convention who contribute to processes of multilateral decision-making. 2 Rather, administrative and political actors interact horizontally, vertically, and diagonally with one another, leading to blurred lines and competencies between national and supranational as well as direct and indirect administrative activities. In this regard, Kingsbury et al. (2005, 5) argue that "much of global governance can be understood and analyzed as administrative action: rule making, administrative adjudication between competing interests, and other forms of regulatory and administrative decisions and management." A global administrative space is said to evolve (Wessel and Wouters 2007, 281) in which states are no longer the single determinant but rather one of many. Wessel and Wouters (2007, 281) therefore call "for the recognition of a global administrative space in which international and transnational administrative bodies interact in complex ways."

The Climate Secretariat as Knowledge Broker?
Building on studies that take the behavior of international bureaucracies-rather than the principal-agent relationship between states and IOs-as a starting point, this article assumes that not all bureaucratic behavior in global politics is government-imposed and that international secretariats can play a constitutive role in shaping party and nonparty stakeholder preferences. Moreover, we believe that the climate secretariat's potential for influence relates to the global climate governance structure and its transnational networking and mobilization capacities.
As stated in the introduction, the global climate governance regime is characterized by a multisectoral and multiactor network structure. It is a system in which skilled action can enable actors such as international secretariats to mobilize support for their policy preferences (Jänicke 2015;Ostrom 2010). They do so by interacting with party and nonparty stakeholders at different levels of government. Jinnah (2014), for example, shows that treaty secretariats position themselves at the center of transnational communication flows that surround official multilateral negotiations to provide policy-relevant information to negotiators from the outside. Abbott and colleagues (Abbott et al. 2015;Abbott and Snidal 2010) further conceive of IOs and their secretariats as "orchestrators" who follow a complementary strategy of "reaching out to private actors and institutions, collaborating with them, and supporting and shaping their activities" to achieve their regulatory goals and purposes (Abbott and Snidal 2010, 315). Finally, Carpenter (2001), in his study on the autonomy and influence of regulatory agencies in the United States, points out that administrative agencies 2. In this regard, a representative of a think tank quoted by Reschke (2016) refers to Ostrom and Ostrom's (1965) notion of polycentricity, which emphasizes the multifaceted nature of humanecosystem interaction. Namely, it explains the variety of relationships between governmental units, public agencies, and private businesses coexisting and functioning in a public economy that can be coordinated through patterns of interorganizational arrangements. may try to influence public policy making by publicly promoting their preferred policy options in issue-specific discourses outside the political system. Against this backdrop, several authors conceptualize secretariats as knowledge or policy brokers (see Bauer and Weinlich 2011;Biermann and Siebenhüner 2009;Eckhard and Ege 2016;Jinnah 2014;Jörgens et al. 2016;Kolleck et al. 2017a). Understood by Sabatier and Jenkins-Smith (1993) as actors who mediate and seek compromise between competing advocacy coalitions, actors possessing a brokerage position "bring parties together" and create an "enabling environment" (Lepoutre et al. 2007, 10) among actors who lack "access to or trust in one another" (Marsden 1982, 202). They can potentially negotiate the stream of information and "bring together ideas that emerge within the network" (Kolleck 2014, 55), to "facilitate exchange, identifying potential options for multi-actor agreement, and helping to craft patterns of communication as well as multilevel and multi-actor governance arrangements" (Bressers and O'Toole 2005, 141). The concept of knowledge brokerage hence emphasizes that information dissemination plays a key role in exerting influence in political processes.
Since Max Weber, studies in the field of public administration have found that bureaucracies derive authority from their superior expertise (see, e.g., Barnett and Finnemore 2004;Derlien and Böhme 2011). Giving meaning to information, bureaucracies are able to shape social reality, prompt action, and exert cognitive influence. International public administrations in general and convention secretariats in particular are no exceptions to this (Bauer and Weinlich 2011;Jinnah 2010). Scholars of international bureaucracies have shown that international treaty secretariats can be powerful actors that wield (independent) influence in global policy making. For example, Barnett and Finnemore (2004) explain with reference to Weber that bureaucratic power includes control over information (meaning bureaucrats have information that others do not have) but also the ability to transform information into knowledge, that is, to structure perceptions. International bureaucracies exert influence, inter alia, through the use of their central position in actor networks, their privileged access to information, their professional authority, and technical expertise (Bauer and Ege 2016;Jörgens et al. 2016;Kolleck et al. 2017b;Widerberg and van Laerhoven 2014;Jinnah 2014).
We assume that the climate secretariat makes use of the unique multiactor and multilevel structure of global climate governance, thereby expanding its original spectrum of activity. It can do so by connecting with a variety of stakeholders operating at different levels and linking otherwise disconnected pools of ideas, acting as an intermediary in the UNFCCC regime. To confirm this expectation, we would have to observe a number of features of a policy broker in the climate secretariat. In particular, we would have to demonstrate that it engages in activities that are typical of knowledge (or policy) brokers, such as gathering, synthesizing, processing, and disseminating policy-relevant information to a wide range of different stakeholders in an attempt to alter knowledge and belief systems ( Jinnah 2014). In addition, we would have to show that it occupies a central position within the international climate policy network, enabling it to influence the flow and content of policy-relevant knowledge and information. In the next section, we will study the role of the secretariat by analyzing its interactions with state and nonparty actors. Using SNA, we will first identify the position of the secretariat within the UNFCCC stakeholder network and then examine its interactions with state and nonparty actors in more detail.

Methodological Approach and Data Sources
We conduct a systematic empirical analysis of the policy-related information and cooperation network that has emerged between negotiation parties, nonparty stakeholders, and the climate secretariat to determine the climate secretariat's position within the network as well as the ways in which it interacts with different stakeholders operating at various levels of government.
Research on the interaction between international treaty secretariats and other-state and nonstate-actors within the context of multilateral agreements has advanced considerably over past years (see, e.g., Bauer 2006Bauer , 2009Busch 2009;Conliffe 2011;Depledge 2005Depledge , 2007Jinnah 2011Jinnah , 2014Jörgens et al. 2016;Kolleck et al. 2017b;Siebenhüner 2009). A number of scholars studied why and under which conditions states and/or IOs would be interested in information provided by nonparty actors (Böhmelt 2013;Jinnah, 2014;Rietig 2014;Tallberg et al. 2013Tallberg et al. , 2015. Some scholars analyzed the links between different treaty secretariats, including the climate secretariat (Betsill et al. 2015), and the role that international secretariats play in the management of regime overlap (Jinnah 2010(Jinnah , 2011. However, although information provided by international treaty secretariats is believed to be valued by party and nonparty stakeholders, the ways in which international treaty secretariats and the stakeholders of their conventions exchange issue-specific and policy-relevant information is still largely unknown, as is the position they hold within global policy domains, such as the climate governance regime. Our article addresses this research gap through an empirical analysis of how the climate secretariat attempts to influence the multilateral climate negotiations under the UNFCCC by means of engaging with other, notably nonparty, actors. We start by applying SNA measures to identify the position of the secretariat within the UNFCCC stakeholder network. SNA adds a relational component to the analysis of actors and their influence on policy outcomes. Instead of focusing only on actors (nodes) and their attributes, SNA shifts the focus toward the relations (ties) between actors and to the overall structure of policy networks. Applied to our case, SNA enables us to concentrate on the UNFCCC network as a whole and the relationships that have emerged between different actors and actor groups within this network. Thus we shift the unit of analysis from individual secretariats to the linkages or relations between a broad range of relevant actors, including, but not limited to, the secretariat of interest.
Data were collected between September 2015 and March 2016, approaching a wide variety of state and nonstate actors operating at different levels of the global environmental policy domain via a large-N survey of organizations in the field of global climate governance. Respondents were identified through lists of COP participants in previous years. Within every organization, we identified one person to answer our survey. We then asked this person to name their contacts concerning cooperation and information exchange. We then extended the number of respondents based on the snowball principle and data provided in open questions.
The survey was received by 2,474 persons, of whom 769 answered at least partially, representing a 31 percent response rate. The survey contained two network questions. One asked "Which organizations did you cooperate closely with regarding topics discussed under the UNFCCC during the last 12 months?" and the other "Which organizations did you receive trustworthy information from during the last 12 months?" These two questions form the basis for our SNA and were combined into one undirected network consisting of 1,021 nodes and 1,834 ties. UNFCCC stakeholders were asked, among other things, to indicate the actor groups they represent. They also answered questions concerning the type of information provided by the UNFCCC secretariat. This allows us to detect if the UNFCCC secretariat limits itself to providing procedural information or if it also offers, for example, information on policy options and the technical or scientific aspects of climate policies. 3 To build a more detailed picture of the quality of interaction between the secretariat and UNFCCC stakeholders, we also conducted thirty-three interviews and substantiated interview responses through a document analysis. Interviews 1-7 were conducted with staff of the climate secretariat (e.g., staff concerned with communication and outreach; strategy; mitigation; data and analysis; finance; technology and capacity building; legal affairs; and administrative services, such as the organization of side events), interviews 8-25 with party stakeholders (e.g., representatives of different public authorities and agencies at the local, regional, and national levels), and interviews 26-33 with nonparty stakeholders (e.g., members of IOs, research organizations, nongovernmental organizations [NGOs], and business representatives).
Interviewees were queried, among other things, whether and, if so, with whom they mainly cooperate and exchange policy-relevant information. Stakeholders of the UNFCCC were also asked to indicate the role and importance of the secretariat within the UNFCCC realm as well as the relationship they entertain with secretariat staff. Members of the climate secretariat, on the other hand, were requested to describe the ways they interact with UNFCCC stakeholders to provide issue-specific information and eventually build trusting relationships. This allows us to better understand and retrace the methods employed by the climate secretariat to foster the facilitation of negotiations and the implementation of the Paris Agreement.

The Climate Secretariat Within the UNFCCC Stakeholder Network
To study the secretariat's potential role within the climate regime, we first analyze its position in the UNFCCC stakeholder network. We apply two centrality measures to determine the secretariat's position: betweenness and eigenvector centrality. Betweenness centrality is often described as a measure to assess an actor's broker potential within a specific network. It is calculated by counting how often an actor lies on the shortest path between two other actors. Actors with high betweenness centrality scores are in a powerful position to control flows through the network, for example, information flows. Moreover, their absence from the network has a strong potential to lead to disruption. The second measure, eigenvector centrality, accounts for the quality of an actor's connections. It measures how many ties an actor has to other actors, which in turn have many ties. In other words, it shows us how well an actor is connected to other well-connected actors. Table 1 ranks the twenty-five most central actors in the UNFCCC stakeholder network according to their betweenness and eigenvector centrality, respectively.
The UNFCCC in general and the UNFCCC secretariat more specifically both occupy an extremely central position within the UNFCCC stakeholder network. For both centrality measures, the UNFCCC ranks fourth and the UNFCCC secretariat sixth (Table 1). These findings show that the climate secretariat is well embedded within the UNFCCC stakeholder network. As mentioned earlier, a high betweenness centrality score indicates that the secretariat has the potential to act as a broker within the increasingly complex UNFCCC regime. This means that it is a potentially powerful actor within the network and can disrupt the network easily by, for example, limiting information flows. Moreover, the eigenvector centrality score reveals that the climate secretariat is well connected to other potentially influential actors and successfully engages with other multiplier stakeholders. Other actors with high scores for both measures are IOs, such as the United Nations Development Programme (UNDP) and the United Nations Environment Programme (UNEP); globally active NGOs, such as the Climate Action Network (CAN) and the World Wildlife Fund ( WWF); and governmental organizations like Deutsche Gesellschaft für Internationale Zusammenarbeit (GIZ).
As survey responses differentiated between UNFCCC and the climate secretariat, we treated the two as separate actors. Some respondents, however, might not have clearly distinguished between the UNFCCC as a treaty regime and the climate secretariat as an international bureaucracy. To account for the possibility that respondents who indicated the UNFCCC might actually have meant the secretariat, we merged the UNFCCC and the climate secretariat into one actor, recalculated its centrality measures, and adapted the egocentric network of the secretariat accordingly (see Supplemental Appendices 1 and 2; https://www. mitpressjournals.org/doi/suppl/10.1162/glep_a_00556). In the adapted stakeholder network, the climate secretariat obtains the highest scores for both betweenness and eigenvector centrality. Nevertheless, to avoid the risk of overinterpreting our data, we base our analysis on a stakeholder network that considers the UNFCCC and the climate secretariat as separate actors. In addition to examining the centrality scores of the UNFCCC secretariat, we are also interested in the secretariat's specific interactions with groups of UNFCCC stakeholders. Figure 1 shows the egocentric network of the UNFCCC secretariat, which only includes those actors who responded that they cooperated or exchanged information with the secretariat (see Supplemental Appendix 2 for The UNFCCC Secretariat's Egocentric Network The network was created with Gephi, using the Fruchterman-Reingold layout. The node color represents the actor type: blue = IOs, green = government, red = research, pink = NGOs, yellow = business. an egocentric network of UNFCCC and the climate secretariat combined). We can see that party and nonparty stakeholders are almost equally represented. The UNFCCC secretariat is an important partner for both groups of actors.
Although the secretariat's position in the global climate governance network indicates a significant potential for influence, its actual influence depends crucially on the type of information that it passes on to other actors. Figure 2 shows that the secretariat primarily provides procedural information to surveyed governments, IOs, and nonparty stakeholders. However, the secretariat also shares information on policy options related to the climate negotiations. Among all actor types, policy-relevant information was the second most common answer, ranging from 25 percent for IOs and businesses to almost 29 percent for NGOs. Furthermore, a considerable share of research (20%), government (15%), and IOs (13%) reported having received technical or scientific information from the secretariat. This indicates that the climate secretariat not only holds a central position within the UNFCCC regime but also engages in the dissemination of policy-relevant information to state and nonparty stakeholders within the network. The fact that different groups of actors perceive the secretariat as a go-to organization for policy-relevant information is an important precondition for it to exert influence on policy outcomes and an indicator of its potential as a global climate policy broker.
Summing up, the findings of our SNA indicate that the climate secretariat has significant potential to influence the UNFCCC regime. First, it is able to broker information of different kinds within the UNFCCC stakeholder network between party and nonparty stakeholders, which may allow it to facilitate international climate negotiations. Second, it connects with other well-connected stakeholders, meaning that it provides information to actors that, in turn, provide information to many other actors. Third, the information it provides to other actors is not limited to procedural information but also includes substantial information on policy options and the technical or scientific aspects of climate policies.

The Climate Secretariat's Interaction
By providing policy-relevant information to both the negotiation parties and nonparty stakeholders, the climate secretariat strengthens its link to the formal climate negotiations and, to a certain degree, confers institutional legitimacy to their problem definitions and policy proposals. The official mandate of the secretariat and the demands of parties do not seem to hinder it from maintaining close links to nonparty stakeholders. Rather, our survey data and subsequent interviews suggest that it focuses more strongly on the needs of nonparty stakeholders than its mandate stipulates and tries to build support among nonparty stakeholders in the hope of putting pressure on parties and thereby advancing negotiations in the desired direction. The next sections present the results of our qualitative analysis and explore in more detail how the climate secretariat interacts with party and nonparty stakeholders.

The Climate Secretariat and Nonparty Stakeholders
On closer analysis of the relationship between the climate secretariat and nonparty stakeholders, we find that different kinds of interactions take place. As nonparty stakeholders are generally less interested in information about the negotiating process itself and rather seek to understand the interaction between the parties and the underlying political questions (interview 6), the secretariat tends to give information on the proceedings and obstacles to the negotiations as well as "the possibilities for a successful outcome" (interview 6). Moreover, the secretariat provides targeted information to nonparty stakeholders so they may better understand "what climate change means on the ground in different areas" (interview 6).
Understanding the different needs of stakeholders (interview 6), the climate secretariat reaches out to nonparty stakeholders to educate people about climate change by giving a "sense of positivity and optimism that it can be done and that we are not starting at ground zero" (interview 5). For example, at COP 22, the address made by the executive secretary of the climate secretariat to the opening plenary highlighted the need to fully include nonparty stakeholders, as "they are central to the global action agenda for transformative change" (Third World Network 2016, no. 3, 1). Defined as "anybody who is not a governmental actor and wants to contribute to the process" (interview 6) by a member of the secretariat, nonparty stakeholder involvement is said to push for the greater good in international climate politics (interview 1) and makes people aware of the direct consequences of climate change.
As changing the narrative on climate change is a costly endeavor, members of the secretariat team up with a variety of other actors (interview 5) to underscore the multiple (economic) benefits resulting from acting. Members of the secretariat, for example, worked with software developers to program a climate game, promoted by "a couple of … celebrities" (interview 5) to reach out to the public so they may in turn take action. Moreover, the climate secretariat "spent a lot of time forming partnerships with key stakeholders, like the former mayor of New York, Michael Bloomberg, and celebrities who had climate interests, such as Mark Ruffalo" (interview 5). The executive secretary also talked to religious groups and their leaders, such as the pope, as well as to research organizations and other entities, asking them to speak out on climate change. The climate secretariat connects with high-level stakeholders-so-called influencers in media and in research-to articulate the secretariat's message (interview 5).
The climate secretariat also actively extends its network of focal points to, for example, youth organizations (interview 4). It invites civil society to make requests for technical assistance "from any level, be it local government, NGOs, universities, the private sector or national ministries" (interview 28). On the issue of women and gender, for example, the secretariat facilitates implementation and capacity building via a network of approximately "260 organizations, private sector entities and other types of institutions around the world" (interview 28), such as the GEF Gender Partnership, UNEP, the Women Delegates Fund, and the Global Gender Climate Alliance (interview 28). Other cooperation partners include UNESCO and the German Development Institute.
The climate secretariat also supports and encourages the engagement of nonparty stakeholders operating at different levels beyond the UNFCCC regime. This happened, for example, in the context of identifying synergies between the Paris Agreement and the Sustainable Development Goals. When the president of the seventy-first session of the UN General Assembly called for a "focus on new strategic partnerships and mobilizing resources from public, private, blended and alternative sources," 4 the UNFCCC executive secretary quickly "echoed the UN Secretary-General's vision for addressing challenges in an integrated manner." 5 She stated that momentum is building not only among national governments but also in the "unprecedented alliances" of companies, investors, cities, regions, institutions, and individuals (see interview 5). 6 Additionally, at the UNFCCC Bonn sessions in May 2017, she said, "All sectors of society in all nations need to be on-board and fully involved to support governments as they take forward their climate action plans into the future" (UNFCCC 2017).
To sum up, members of the secretariat actively promote interaction with nonparty stakeholders (interview 2): "Stakeholders are out in the real world. They've got boots on the ground, they are dealing with vulnerable communities, they are trying really hard to talk across ministries, across sectors" (interview 1; also interview 4). Members of the secretariat interact with nonparty stakeholders while acting as mediators and facilitators on certain issues (interviews 24 and 26) or as the "hub of a network" (interview 6), thereby pushing for cooperation via the extension of topics. As one member of the climate secretariat states, "My job … is to bring together stakeholders, like NGOs, science, different levels of decision makers, experts of every potential kind, and involve them into discussion that the parties have" (interview 1). Nonparty stakeholders themselves greatly appreciate the efforts of the climate secretariat. At an exclusive meeting between nonparty stakeholders and members of the climate secretariat at COP 23, many nonparty stakeholders thanked the climate secretariat for all of its work and for giving them "the opportunity to be heard, not only in the corridors, but also in an open dialogue" (interview 32).

The Climate Secretariat as Communication Hub Between Party and Nonparty Stakeholders
The climate secretariat seems to be very aware of the importance of communication and staying within its mandate (interviews 1 and 6), as state delegates occasionally express their reservations about the climate secretariat taking an active role during COP negotiations. While members of the secretariat might not openly voice their opinions in the negotiating space (interview 1), providing nonsensitive information in a "strictly neutral" (interview 6) fashion, party stakeholders confirm that the interactions of the members of the climate secretariat with the parties go beyond merely facilitating negotiations (interviews 7 and 19). Rather, the secretariat acts as a mediator-as a communication hubbetween parties (interview 24) and nonparty stakeholders.
Interaction between party and nonparty stakeholders often takes place via participation in convention institutions like the Adaptation Committee and in other initiatives, such as the Lima Paris Action Agenda and the Technical Expert Meetings (interview 6). This kind of cooperation is particularly evident during COP sessions, in the form of jointly conducted side events organized by members of the climate secretariat.
Side events are held in parallel to negotiation sessions and provide an opportunity to discuss policy issues beyond the realm of the negotiation. They bridge the "interactions between the formal and informal spaces of climate governance" (Schroeder and Lovell 2012, 23) as states are formally required to collaborate with an observer organization to apply via the climate secretariat for a side event and/or an exhibit slot. At side events, all speakers have the opportunity to prominently present their views on a certain topic and to advocate for a preferred policy option in tandem. While the formal view of the party hosting the side event does not necessarily need to converge with that of the nonparty stakeholders, these events are often used to demonstrate a common stance on a specific issue.
Nonparty stakeholders are very keen to register and participate as observers in UNFCCC negotiations and to organize side events to actively take part in the discussions (interview 2) to ensure they get their "fingerprints on the process" (interview 1). Party stakeholders, on the other hand, welcome the nonparty stakeholders' participation in side events, as they generate important stimuli (interview 16) as well as legitimize their actions. Side events enable capacity building, the introduction of potential items for negotiation, networking across levels and policy areas, and dissemination of information (Hjerpe and Linnér 2010;UNFCCC 2015; see also interviews 17 and 30). Side events can thus be understood as high-quality conversations that are able to foster innovation, trust, and awareness of the need for possible compromises (Hjerpe and Linnér 2010), all of which can positively affect both the negotiation of issues and the implementation of past decisions.

Conclusions
In this article, we aimed to better understand the role of the UNFCCC secretariat in the increasingly complex global climate governance structure. We developed an innovative methodological approach to addressing this issue and systematically examined the climate secretariat's relations with the main groups of actors involved in this policy domain, in particular with nonparty actors. More specifically, we used SNA to examine the secretariat's relations with nonparty and party stakeholders and to identify its position in the UNFCCC policy network. An understanding of where the climate secretariat stands in the global climate governance network and which actors it interacts with most allowed us to draw preliminary conclusions about the ways in which it connects with other stakeholders to influence global climate policy outputs. In addition, we extended our findings gathered with SNA by conducting thirty-three semistructured interviews to corroborate the results of the SNA.
Our quantitative and qualitative analyses show that interaction between party and nonparty stakeholders and the climate secretariat occurs in many ways. Hence, in this article, we demonstrate that the climate secretariat not only provides expertise that allows party and nonparty stakeholders to understand the complexity of the issue at hand but also successfully connects with other well-connected stakeholders (e.g., influencers) to promote the implementation of the Paris Agreement and combat climate change outside of the UNFCCC regime. In doing so, it prioritizes its actions and (selectively) connects with a variety of stakeholders. We also conclude that the climate secretariat acts as an intermediary between party and nonparty stakeholders. It deliberately extends issue-specific policy debates beyond the inner circle of official negotiation parties (i.e., the national delegations) to build transnational support for the policy issues at stake, thereby raising pressure from both within and outside to continue and successfully conclude the negotiations. The climate secretariat shows key characteristics of a knowledge broker in that it deliberately engages in gathering, synthesizing, processing, and disseminating policy-relevant information to a wide range of different stakeholders in an attempt to alter knowledge and belief systems. In addition, the central position it occupies within the international climate policy network enables the secretariat to influence the flow and content of policy-relevant knowledge and information among the actors in this network.
The results of our study lend support to the argument that the climate secretariat may gradually be moving from a rather neutral and instrumental stance to playing a proactive and influential role in international climate governance. It aims to increase its political influence by establishing strategic links to actors other than the formal negotiation parties, thereby exceeding its role as a mere provider of process-related information. While its behavior is rooted in its formal mandate, we find that the climate secretariat is increasingly involved in the generation and diffusion of climate-related policy ideas and innovations. It coordinates and moderates the interaction with numerous stakeholders operating at various levels to foster cooperation and agreement. For example, the former executive secretary of the UNFCCC, Christiana Figueres, was publicly perceived as the driving force behind the UNFCCC COP 21 negotiations in conjunction with the French presidency. The laudation of the Ewald von Kleist Award, which they received at the 2016 Munich Security Conference, states, "While many played a part in achieving agreement in Paris, it was our two award winners tonight who, primarily, made the Paris accord possible" (United Nations Secretary-General 2016).
Mastering the unique multilevel global climate governance architecture despite a narrow formal mandate that emphasizes its logistical role and explicitly exempts it from taking on a more active part in multilateral negotiations, it cooperates with different kinds of stakeholders to guarantee the successful realization of the measures agreed upon in the Paris Agreement. The climate secretariat makes a deliberate choice to use its limited resources for investing heavily in networking with different kinds of stakeholders. Our study lends support to the assumption that, being intermediaries between party and nonparty stakeholders, international treaty secretariats may be gradually moving from playing a rather neutral and instrumental role in international climate governance (Busch 2009) to being proactive and influential (see, e.g., FCCC/SBI/2016/ INF.13). While the UNFCCC explicitly stipulated that the secretariat should cooperate with different stakeholders operating at various levels to guarantee the implementation of the measures agreed upon, it might be precisely this secretariat task that blurs the borders between fulfilling its mandate and participating in shaping the political will. Barbara Saerbeck is a policy officer at the Federation of German Consumer Organizations and is responsible for the energy sector. She is also associated with the Environmental Policy Research Centre at Freie Universität Berlin. Her research focuses on the role and influence of party and nonparty stakeholders as well as public administrations in (inter-)national energy, climate, and environmental negotiations. Mareike Well is a doctoral candidate at Freie Universität Berlin. Her research focuses on climate policy, multilateral negotiations, and the role of international public administrations in global environmental governance.
Helge Jörgens is an assistant professor at Instituto Universitário de Lisboa (ISCTE-IUL) and a senior researcher at Centro de Investigação e Estudos de Sociologia in Lisbon, Portugal. His research focuses on national and comparative environmental and climate policy analysis, the cross-national transfer and diffusion of policies, and the role and influence of international public administrations.
Alexandra Goritz is a doctoral researcher at Freie Universität Berlin and a research associate at Leipzig University. Her research focuses on global climate politics, international public administrations, climate education, and social network analysis. She holds an MSc in environment and development from the London School of Economics and a BA in politics and public administration from the University of Konstanz.
Nina Kolleck is a full professor at Leipzig University (Institute of Political Science). She has developed a solid line of research dealing with methods of social network analysis in policy and education.