The Legitimation of International Organizations: Introducing a New Dataset 1

: This article introduces a new dataset on how international organizations (IOs) justify their authority. For a long time, IOs were believed to derive legitimacy from member-state consent and technocratic problem-solving capacities. Over recent decades, the growing politiciza-tion of IOs, political polarization within Western democracies, and power shifts in the international system have spurred IOs’ efforts to justify their right to rule, using a variety of legitimation practices. While research on the theory and practice of IO legitimation has grown considerably over the past decade, much of this work builds on case studies of prominent global and regional IOs. As a result, we lack data suitable for systematic comparative analyses across time, IOs, and world regions. The Legitimation Strategies of Regional Organizations (LegRO) dataset aims to narrow this gap, providing data on the standards, intensity, and modes of legitimation for twenty eight regional IOs from 1980 to 2019. These variables inform theoretical and policy-relevant research on contemporary global governance by providing the ﬁrst systematic overview of IOs’ legitimation practices. Resumen


Introduction
Since the Battle of Seattle in 1999 and the failed referendums on the European constitution in 2005, students and practitioners of global governance have realized that international affairs are no longer the preserve of political elites.With the end of the permissive consensus on international cooperation and, more recently, its increasing contestation by rising powers and populist parties in established democracies, scholars and policymakers have sought to understand the causes and consequences of these developments.It is now widely recognized that the politics of legitimation and delegitimation have become a central element of contemporary Downloaded from https://academic.oup.com/isp/advance-article/doi/10.1093/isp/ekad008/7179500 by Universitat Lueneburg user on 25 May 2023 global governance and that international organizations (IOs) are one of its central pillars.Yet the research on this topic is limited to (comparative) case studies, making it difficult to systematically map how different dimensions of IO legitimation vary across organizations and over time.The focus on the legitimation of prominent global IOs--such as the United Nations (UN) or the International Monetary Fund (IMF)--and a few regional organizations ( Binder and Heupel 2015 ;Gronau and Schmidtke 2016 ;Hensell 2022 ) has led to incomplete and potentially biased views about IO legitimation.
In this article, we introduce the Legitimation Strategies of Regional Organizations (LegRO) dataset, describe its features, and explain how International Relations (IR) scholars and the policy community can use it.The dataset contains information on the legitimation of twenty eight regional IOs from 1980 to 2019.The data build on a content analysis of annual reports issued by an IO's general secretariat and communiqués of meetings of heads of state and government.Based on the coding of almost 1,500 documents, we present a set of quantitative measures that can be used to map IO legitimation within and across organizations and explore its causes and consequences.
IO legitimation denotes practices that seek to enhance audiences' belief in the normative appropriateness of an IO's authority.It is distinct from delegitimation, which describes practices that challenge belief in the appropriateness of an IO's authority ( Gronau and Schmidtke 2016 ;Tallberg and Zürn 2019 ).The LegRO dataset focuses on practices of legitimation performed by IO representatives.We define IO legitimation as generalizable norm-based justifications of an IO's authority presented in the organization's public communication.This communication results from negotiations among member states and international bureaucrats as legitimation agents.We assume that it primarily addresses audiences that directly work with an IO, can provide resources, or have a stake in IO policies.
Systematic data on IO legitimation can illuminate essential questions in contemporary global governance.First, they help us understand IOs' behavior and institutional development.IR scholars have devoted considerable attention to explaining the institutional design, policy output, and behavior of IOs ( Koremenos, Lipson, and Snidal 2001 ;Gutner and Thompson 2010 ).This research focuses predominantly on states' material preferences as the main driver of these phenomena.We suggest that IOs' quest for legitimacy is an important but underexplored explanation for why IOs take specific institutional forms, how they translate formal institutional designs into concrete policies, and why they change over time.Our data allow researchers to examine the normative roots of IOs' institutional design, policy output, and behavior.This could contribute to the development of a theory of normative institutional design that complements extant work building on the idea of rational design.
Secondly, our dataset helps uncover the norms that underpin global governance.Scholarship on the norms that shape international cooperation and the normative challenges the international order faces in a multipolar world has grown substantially ( Finnemore and Sikkink 1998 ;Ikenberry 2020 ;Lake, Martin, and Risse 2021 ).If norms matter in international cooperation, then it is imperative to understand which norms are voiced by relevant actors and who supports them.While the existing research examines the normative expectations and challenges of national governments, civil society, and the broader public, the LegRO dataset illuminates the normative foundations and political ideologies of IOs.These data enable a systematic analysis of the drivers of IOs' activities, if and how they react to normative challenges, and how responsive they are to their social environment.Beyond individual IOs, they also offer insights into the normative structure of the international system more broadly, since legitimation discourses and their respective authoritativeness derives from what Bukovansky ( 2002 ) calls the "international political culture" (see also Nuñez-Mietz 2018 , 732-33).Thirdly, the dataset facilitates research on the drivers of IO legitimacy.In times of multiple and overlapping crises, the world needs strong multilateral institutions capable of finding solutions to pressing problems, such as the climate crisis, the global pandemic, or Russia's invasion of Ukraine.To play this role, IOs require support from governments, civil society, the business sector, and citizens.The literature on popular legitimacy identifies IOs' legitimation practices as a potential determinant of individual legitimacy beliefs, alongside individual-and organizational-level factors such as a general social trust or organizational performance ( Bexell, Jönsson, and Uhlin 2022 ;Dellmuth et al. 2022 ;Dellmuth and Tallberg 2023 ).This research stands to benefit from systematic data on IO legitimation.
Finally, systematic data on IO legitimation is of great relevance to policymakers.Since IOs generally lack coercive enforcement mechanisms and rely on voluntary support and compliance, belief in their legitimacy on the part of those who work with an IO, those who can provide resources, and those who are expected to comply with their policies is essential.When legitimacy beliefs are positive, rulership tends to be cheaper, compliance higher, and support more forthcoming ( Hurd 1999 ;Tallberg and Zürn 2019 ).Mass protests against global economic IOs, citizens' rejection of the European constitution, or member-state withdrawal triggered by mass-level dissensus have put IOs across the world on alert.Their representatives are often concerned about these developments and seek ways to win their critics over ( Sommerer et al. 2022 ).Our data address policymakers' demand for comparative insights into legitimation practices and how IOs around the world communicate the normative value of international cooperation.
The following section reviews existing research and data on IO legitimation.In the third section, we introduce the building blocks of the dataset, including the IO sample, the data sources, and the measurement strategy.In the fourth and fifth sections, we illustrate the versatile applicability of the data in two steps.First, we present the main descriptive patterns.Secondly, we explore the association between our data and IO authority, a variable commonly considered to shape IO legitimation.We conclude by summarizing how scholars and practitioners can apply the dataset.

Existing Data on IO Legitimation
IO legitimation denotes practices that seek to enhance audiences' belief in the normative appropriateness of an IO's authority.IO delegitimation, by contrast, describes practices that challenge belief in the appropriateness of an IO's authority ( Gronau and Schmidtke 2016 , 540;Tallberg and Zürn 2019 , 588).This conceptualization implies (1) that various actors, including international bureaucrats, civil society actors, and governments, can engage in legitimation; (2) that these legitimation agents can use discursive, institutional, and behavioral practices to shape audiences' legitimacy beliefs; and (3) that audiences encompass actors who work directly with an IO, provide resources for an IO's work, or are expected to comply with an IO's policies.
Research focusing on IOs' legitimation practices has increased significantly over the past decade.Yet despite growing interest in what IO representatives do and say to convince IO staff, governments, civil society, business, and citizens of the normative appropriateness of their organization, we lack systematic data that enable comprehensive temporal and cross-sectional comparative analyses.The existing empirical work is dominated by analyses of the legitimation practices of single IOs or small-N comparative case studies.For a few prominent organizations, including the European Union (EU) ( Hurrelmann, Gora, and Wagner 2013 ;Biegon 2016 ;Hensell 2022 ), the UN ( Binder and Heupel 2015 ;Kentikelenis and Voeten 2021 ;Tørstad 2023 ), and the IMF ( Halliday, Block-Lieb, and Carruthers 2010 ;Gronau and Schmidtke 2016 ;Clift and Robles 2021 ), the legitimation efforts are well documented.
Work that moves beyond individual IOs is rare and tends to offer comparisons of the legitimation of a few organizations ( Dingwerth et al. 2019 ;Schmidtke 2019 ;Rauh and Zürn 2020 ).The main exceptions are Bexell and colleagues' ( 2022 ) study of the legitimation practices of sixteen global governance institutions, including IOs and nongovernmental or hybrid organizations, and Dingwerth and colleagues' ( 2020 ) analysis of the democratic legitimation discourse of twenty global IOs.
Beyond research focusing directly on IO legitimation, we find several large-N studies that overlap with legitimation research because they perceive certain institutional design elements to be driven by legitimacy concerns.Most prominently, work on democratic control suggests that IOs establish transparency and accountability mechanisms ( Grigorescu, 2007( Grigorescu, , 2010 ) ), civil society access ( Steffek, Kissling, and Nanz 2008 ;Tallberg et al. 2014 ), and parliamentary bodies ( Lenz, Burilkov, and Viola 2019 ;Schimmelfennig et al. 2020 ) for legitimation purposes.In addition, research on public outreach shows that IOs have begun to create centralized communication units ( Ecker-Ehrhardt 2018 ), make their messages more accessible to the general public ( Rauh 2022 ), and increase geographic representation among their staff ( Parizek and Stephen 2021 ;Badache 2022 ) to foster belief in their legitimacy.While these studies build on comprehensive IO samples and thus contribute representative data, they are empirically selective, focusing on institutions that functionalist approaches have difficulty explaining.
Overall, the empirical research on IOs' legitimation efforts has increased significantly.It suggests that IO legitimation is multifaceted and varies across several dimensions.First, IO representatives use different modes of legitimation to shape legitimacy beliefs, including linguistic justifications, behavioral adaptations, and institutional reforms.Secondly, the intensity of legitimation --i.e., the frequency with which IOs apply these practices--varies across organizations and over time.Thirdly, IO representatives seem to rely on a broad set of standards of legitimation to substantiate their claims.These standards include, for instance, technocratic norms related to expertise, problem-solving capacity, and neutrality; liberal norms that emphasize democracy, human rights, and the rule of law; and communitarian norms connected to collective identity or national sovereignty.Yet, studies documenting this fascinating variation have three important limitations.First, they mostly rely on qualitative research designs that zoom in on individual IOs or small-N comparisons, while large-N studies comparing the different dimensions of IO legitimation across multiple organizations and over time are rare.Secondly, these studies are predominantly concerned with a few prominent global and regional IOs, fueling doubts about external validity.As we show in this contribution, this work overlooks widespread empirical phenomena, such as legitimation via an IO's purpose, communitarian norms, or policy adaptations.Thirdly, existing studies focus on specific aspects of legitimation, such as the modes or standards of IO legitimation.Consequently, we lack generalizable knowledge about all three dimensions of legitimation--a situation that the LegRO dataset seeks to rectify.

The LegRO Dataset
We construct the LegRO dataset around three attributes that contribute to reducing the gaps identified in the literature.The first attribute is broad geographical coverage.Most analyses focus on prominent global and regional IOs, thus risking selection bias.We know little about legitimation by less "visible" regional organizations, especially from the Global South, such as the League of Arab States (LoAS) or the Economic Community of West African States (ECOWAS).To address this gap, we have compiled a sample of regional IOs from different world regions.
Secondly, comprehensive analyses comparing standards and modes of legitimation are scarce and tend to omit the critical empirical phenomena mentioned above.To provide a fuller picture of IO legitimation, the dataset operates with Downloaded from https://academic.oup.com/isp/advance-article/doi/10.1093/isp/ekad008/7179500 by Universitat Lueneburg user on 25 May 2023 typologies for the standards and modes of legitimation that deductively synthesize existing conceptual work and inductively integrate empirical observations that emerge from the more comprehensive geographical coverage.This design allows for variable comparisons--for instance, of different standards and modes of legitimation across IOs and over time.
Thirdly, the dataset prioritizes IO communication as the data source.We suggest that practices aimed at justifying an IO's authority, irrespective of the specific mode, are invariably accompanied by legitimation discourse.This is because justifications require language--i.e., explanations of why something is right or reasonable.Behavior and institutions do not speak for themselves; only justificatory discourse makes them a legitimation practice.This approach enables a more comprehensive picture of IO legitimation as it decouples the empirical research from the auxiliary assumption that IO behavior and institutions can only serve legitimation if they do not have functional roots.

Sample and Data Sources
Our IO sample draws on one of the most comprehensive datasets on the institutional design of IOs to enable access to potential explanatory variables and, thus, cumulative empirical insights.The Measure of International Authority (MIA) dataset ( Hooghe et al. 2017 ) encompasses forty one regional and thirty five global IOs.As hand-coding was required to guarantee fine-grained and valid data, we had to balance the two fundamental goals of representativeness and feasibility.To this end, we used the following selection criteria.
First, we focused on regional organizations, as the legitimation literature has largely ignored this important subset of IOs.2 Organizations from the Global South are particularly underrepresented, thus raising doubts about the representativeness of previous findings.Given the literature's focus on global IOs and select prominent regional IOs, this sample promises to generate discoveries.By comparing our data to findings on the legitimation of global IOs, we can learn more about the differences and similarities between the two types.Furthermore, increasing multipolarity in global governance makes regional international cooperation more important and potentially more distinct ( Garzón 2017 ;Hurrelmann and Schneider 2015 ).Hence, the legitimation of regional IOs offers novel insights into the normative foundations of a multipolar world and adds to the literature on comparative regionalism. 3Secondly, to achieve a representative sample of regional IOs, we selected IOs from four major world regions (Africa, Americas, Asia-Pacific, and Europe) and cross-regional IOs.Thirdly, we included general-purpose and task-specific organizations for each region.Next we stratified the group of task-specific organizations by including IOs from economic, cultural, and political issue areas.Finally, we excluded organizations such as NATO and NAFTA because they do not regularly publish annual reports and communiqués.Based on these criteria, the dataset covers the legitimation efforts of twenty eight regional IOs between 1980 and 2019 (see Online Appendix A1 for the sample).The selected time frame is sufficiently long to capture regional and global political, economic, and social developments that may have shaped legitimation.At the same time, the period is a pragmatic choice because going further back in time imposes growing constraints regarding the accessibility of data sources.
We analyzed the legitimation of the selected organizations using two of their public communication outputs: annual reports and the communiqués from meetings of heads of state and government.These documents have four advantages over other publications such as press releases, speeches, or social media communication.First, the documents reveal discursive legitimation and capture different legitimation modes because they are intended to document what an organization has done and will do in the future.In contrast to speeches or press releases, which are issue-driven and context-specific, these documents provide general overviews of an IO's activities.As such, the selected documents are rich data sources because they systematically "record the legitimation warrants of the IO" ( Halliday, Block-Lieb, and Carruthers 2010 , 84).Secondly, a survey among the IOs in our sample shows that member states and staff perceive these documents to be important communication media that have a wide reach and draw the attention of important audiences, including IO staff, governments, and civil society organizations. 4 The drafting process generally encompasses various IO bodies, the bureaucracy's head, and member-state representatives.Thirdly, in contrast to other potential sources, the selected documents have a comparable structure and content, are published regularly, and are available for many IOs over extended periods.Consequently, our data can be linked to research that draws on similar sources and analyzes other IOs.Finally, these documents reflect the views of the two most important agents within an IO--member states, represented by heads of state and government, and the organizations' administration--who may each set their own priorities regarding legitimation.
Ideally, this approach yields one annual report and one communiqué for the 974 IO years in our sample.Yet the number of documents, we analyzed is lower for two reasons.For one, not all IOs have published a communiqué and an annual report every year.For example, during the civil wars in Central America in the 1980s, the Central American Integration System (SICA) did not issue annual reports and barely held meetings of heads of state and government.Such interruptions reduced the number of IO years for which both documents should be available to 837.Moreover, working with less prominent regional IOs comes at the cost of limited data availability.The online archives of some organizations are incomplete, and even through direct communication with the respective IOs, we were sometimes unable to obtain documents. 5Overall, we coded 670 annual reports and 789 communiqués.For twenty four IO years, we found neither an annual report nor a communiqué.In the following analysis, we pool data from both sources for the sake of parsimony and because member states and IO staff contribute significantly to drafting both types of documents.Nevertheless, we also provide disaggregated data to researchers interested in comparing legitimation across sources.
Given the length of the documents, we focused on sections that provide an overview of normative commitments, present the organizations' identity and desired public image, and showcase achievements. 6These sections encompass general overviews, summaries, forewords, introductions, and conclusions.Since the number of paragraphs in the selected sections--our coding unit--varies, we calculated a 25 percent range around the mean number of paragraphs in these sections.We coded a minimum of sixteen and a maximum of twenty eight paragraphs per document.Given the two types of documents per IO year, the legitimation of an IO is represented by a minimum of thirty two and a maximum of fifty six paragraphs. 7his procedure yielded 32,675 coded paragraphs.

Measuring IO Legitimation
We used content analysis to map IO's legitimation practices based on their public communication.The core of the analysis is the legitimation statement, defined as a generalizable, norm-based justification of an IO's authority made by IO representatives (see Online Appendix A2 for detailed coding rules).In the following, we introduce the coding scheme before explaining how we used it to map the standards, intensity, and modes of legitimation.Legitimation statements assume two basic forms, which we captured with the OES and OIS grammars.O stands for legitimation object, S for normative standard, E for evaluation, and I for identity (see table 1 ). 8The legitimation object refers to the organization in general but not its officials or policies. 9In the OES grammar, this object is evaluated positively through terms such as "good", "great", "improve", or "important milestone", and the statement emphasizes a normative standard .We did not code negative evaluations because they do not fall under our definition of legitimation as a justification of an IO's authority. 10The OIS grammar describes what the organization stands for, what it aims to achieve, and what norms guide the organization in pursuing these goals ( Schmidtke and Nullmeier 2011 ;Dingwerth, Schmidtke, and Weise 2020 ).Such identity statements are descriptive and do not require a positive tone.A paragraph can contain no, one, or more than one legitimation statement.The number of statements we coded per paragraph is based on the number of distinct normative standards.For example, a paragraph that suggests that an IO is legitimate because it has democratic procedures yields one legitimation statement.A paragraph arguing that an IO is legitimate because of its democratic procedures and contribution to protecting human rights yields two legitimation statements.
Four expert coders participated in the coding.We examined documents in seven languages (English, French, Spanish, Portuguese, Arabic, Russian, and Danish).When the coders did not know a language well enough, we had the selected paragraphs translated into English. 11We tested intercoder reliability for the identification of legitimation statements and all element of the legitimation grammar.Reliability tests built on a random sample of 585 paragraphs from thirty documents (approximately 2 percent of the corpus).For all steps of the coding process, we achieved Krippendorff's α of 0.69 or higher, which denotes an acceptable level of intercoder reliability ( Lombard, Snyder-Duch, and Bracken 2002 , 593).

Component One: Standards of Legitimation
The dominant strand of IO legitimation research focuses on the standards of legitimation, understood as the "normative yardsticks that underpin evaluations of international organizations" ( Dingwerth et al. 2019 , 5).Scholars ask, for instance, whether IOs emphasize the democratic quality of their procedures ( Rocabert et al. 2019 ;Dingwerth, Schmidtke, and Weise 2020 ).To advance research on legitimation standards, we build on and move beyond the literature by structuring legitimation standards according to two dimensions: a normative dimension, which captures the benchmark for justification (technocratic, liberal, and communitarian), and an institutional dimension that specifies the respective focus of legitimation (procedure, performance, and purpose) (examples include Bernstein 2011 ; Brassett and Tsingou 2011 ;Tallberg and Zürn 2019 ;Stappert and Gregoratti 2022 ).
In the institutional dimension, we build on Scharpf's ( 1999 ) distinction between input and output legitimacy and thus include procedural and performance standards that are predominant in the literature ( Nuñez-Mietz 2018 ; Strebel, Kübler, and Marcinkowski 2019 ; Mace 2020 ).We widen this dimension by including purpose legitimation, i.e., justifications in which IOs invoke aims and objectives ( Oelsner 2013 ;Oates 2017 ;Sangiovanni 2019 ).
The resulting typology encompasses established and novel conceptualizations.It provides the starting point for empirical research that compares standards of legitimation within IOs, across IOs, and over time.Table 2 shows that our coding further distinguishes between more specific standards within each of the nine categories.For example, we have coded economic welfare, peace and security, international influence, and functional capability as technocratic-performance legitimation.This allows the data to speak to overarching topics and detailed questions about specific standards.Based on this coding, the LegRO dataset provides information on how frequently IOs use certain standards and the share of each standard per IO year.This enables versatile empirical analyses and constitutes the most comprehensive approach to capturing legitimation standards in the literature.

Component Two: Intensity of Legitimation
In contrast to work on standards, studies on the intensity of legitimation are rare.Yet, questions about how much of their public communication, behavior, and institutional design IO representatives devote to legitimation are critical.The answers to these questions provide insights into an IO's organizational culture, identity, and coherence.IOs that feel a constant urge to legitimize are often characterized by low internal cohesion and conflicting identities ( Billerbeck 2019 ).Similarly, We operationalize intensity by dividing the count of legitimation statements by the total number of coded paragraphs because the number of coded paragraphs varies across IOs and over time.As one paragraph can include more than one legitimation statement, the measure has an absolute minimum of zero, indicating the absence of legitimation, and a theoretical maximum equal to the number of normative standards in our typology.

Component Three: Modes of Legitimation
Research on the modes of legitimation--defined as the types of practices IOs use for legitimation ( Bexell, Jönsson, and Uhlin 2022 , 31-2)--is in its infancy.Existing research identifies three modes of IO legitimation: discursive legitimation relies on language; behavioral legitimation rests on actions in day-to-day work; and institutional legitimation involves the basic rules and formal structures that guide an IO's operations ( Gronau and Schmidtke 2016 ; Bäckstrand and Söderbaum 2018 ).While research on discursive legitimation has developed a sophisticated toolkit of quantitative and qualitative methods ( Schmidtke and Nullmeier 2011 ;Rauh and Zürn 2020 ), it is less clear how we can systematically map institutional or behavioral legitimation.
To promote this research agenda, we propose a novel approach that builds on our coding of legitimation statements, enables us to distinguish between different modes of legitimation, and decouples the analysis from assumptions about the functional roots of IO behavior and institutions.This method assumes that even behavioral and institutional modes require a communicative proposition that a specific practice is intended to legitimize an IO's authority.Unlike legitimation which builds exclusively on discourse, other modes of legitimation entail (1) a communicative legitimacy claim and (2) an instantiation of this claim that Downloaded from https://academic.oup.com/isp/advance-article/doi/10.1093/isp/ekad008/7179500 by Universitat Lueneburg user on 25 May 2023 highlights behaviors or institutions, translating a legitimation standard into concrete action.We classify legitimation statements that do not represent an instantiation as discursive legitimation.An example of such discursive legitimation is APEC's ( 2007 , 1) claim that "we will continue to strengthen our community and build a sustainable future".This statement suggests that APEC is legitimate because it aims to protect the environment, but does not mention what the organization does to achieve this goal.By contrast, the statement by the Association of Southeast Asian Nations (ASEAN) that "the Leaders adopted the Statement on Joint Response to Climate Change as a concrete manifestation of ASEAN's collective commitment to address climate change" ( 2010 , 3) exemplifies a mode of legitimation that moves beyond discourse.This statement highlights a concrete behavior with the leaders' adoption of "the Statement on Joint Response to Climate Change" to instantiate the claim that ASEAN is legitimate because it aims to protect the environment.Hence, we conceive of this statement as behavioral legitimation.
We distinguished between four modes of legitimation (see table 3 ).In addition to discursive legitimation and based on the literature, we identified behavioral and institutional modes ( Gronau and Schmidtke 2016 ;Bexell, Jönsson, and Uhlin 2022 ).Finally, we integrated our empirical material and identified a fourth mode that emphasizes policy output.In contrast to behavioral legitimation, which emphasizes daily practices, policy legitimation highlights a formal course of action or principle adopted by an IO.While this mode has received little attention in the literature, our data show that it is empirically relevant and can contribute to research on the drivers of IO policymaking.In coding these modes, we assumed that an IO can instantiate a legitimation statement through multiple modes.Hence, the share of each mode per IO year has an absolute minimum of zero, indicating the absence of this mode of legitimation, and an absolute maximum of one, suggesting that each legitimation statement is accompanied by this mode.

Patterns in IO Legitimation
This section illustrates the versatile applicability of the dataset by presenting patterns in the standards, intensity, and modes of legitimation.We show that our data help narrow gaps, correct working assumptions in the literature, and provide policyrelevant insights.

Standards
The existing work on legitimation standards rests on case studies of prominent global IOs and suggests that IO legitimation builds predominantly on technocratic performance and liberal procedures.Our data show, first, that the focus on procedures and performance misses an essential aspect of legitimation because purposebased standards dominate across IOs and over time.Secondly, the data also provide nuanced insights by showing that liberal norms are less prominent and communitarian norms more prominent than the literature suggests.
The top row of figure 1 displays the shares of procedure, performance, and purpose standards in the institutional dimension across world regions and over time.The bottom row plots the shares of technocratic, liberal, and communitarian standards in the normative dimension.We include a trendline for the population means to visualize general trends.Figure 2 provides a cross-sectional comparison.We include an "average organization" (Ø), which displays the mean of the shares calculated for each IO.The left-hand panel displays the distribution of institutional standards.The right-hand panel shows the distribution of normative standards.
In the institutional dimension, purpose legitimation is more prominent than the literature suggests.On average, the IOs in the sample emphasize purpose in 55 percent of their legitimation statements, while only 32 percent and 13 percent of statements highlight performance and procedure, respectively.This predominance of purpose has become more pronounced over time.Except for a dip during the decade 2000-2010, the overall share of purpose legitimation increased steadily.These findings correct the literature's focus on performance and procedure legitimation.Although procedure legitimation has also become more prominent over time, it remains the least relevant institutional standard.Our sample does not include a single IO that legitimizes predominantly via procedures.Thus, legitimation continues to rest primarily on what IOs (aim to) do, not on how they do it.
In the normative dimension, technocratic standards dominate across all world regions.On average, the IOs in our sample emphasize technocratic standards in 62 percent of all statements, while only 22 percent and 16 percent draw on liberal and communitarian norms, respectively.This observation supports the widespread suggestion that technocratic standards are the bedrock of IO legitimation ( Barnett and Finnemore 2004 ;Steffek 2021 ).However, the observation that liberal legitimation is less prominent qualifies the emphasis on these norms ( Zürn 2018 ;Dingwerth, Schmidtke, and Weise 2020 ).Nonetheless, our data confirm the literature's diagnosis that liberal legitimation has become more relevant over time.Overall, the share of this standard has grown from 17 percent in 1980 to 26 percent in 2019.
Finally, the data reveal that the literature has overlooked communitarian legitimation.While not at the center of IO legitimation, we find that communitarian  standards play a role in all world regions.With a share of 20 percent, Asian and cross-regional IOs draw most frequently on this standard, while African (17 percent), American (15 percent), and European (12 percent) IOs use it more sparingly.
Overall, the data show that the literature's focus on a few prominent global and European IOs affects the validity of empirical studies.Further, these observations imply that policymakers operate in a differentiated normative environment in which purely technocratic justifications are no longer sufficient to address relevant audiences.More specifically, our data inform practitioners about which norms have become indispensable and which standards have lost relevance.Furthermore, IO representatives can also draw on these data to understand the ideological profile of their counterparts and forge partnerships based on normative similarities.In times of regime complexity and institutional overlap ( Eilstrup-Sangiovanni and Westerwinter 2022 ), understanding which IOs share one's own normative benchmarks and which do not is essential for finding reliable cooperation partners.

Intensity
Few studies examine the intensity of IO legitimation.Most comprehensively, Dingwerth and colleagues ( 2020 ) show in their analysis of twenty global IOs that the number of normative justifications IOs provide in annual reports almost doubled between 1981 and 2011.Gronau and Schmidtke's ( 2016 ) comparative study on the legitimation of the IMF and the Group of Eight (G8) suggests that legitimation intensity increases during legitimacy crises.Our data map, for the first time, the legitimation intensity of regional IOs.They confirm that intensity has increased for many organizations.Yet, they also uncover substantial regional and cross-sectional variations.
The left-hand plot in figure 3 displays legitimation intensity as the number of legitimation statements divided by all coded paragraphs across world regions and over time.We include a trendline for the population means.The right-hand plot provides the cross-sectional comparison.We include an "average organization" (Ø), which displays the mean intensity of legitimation across the sample.
In line with the growing interest of scholars and practitioners, all IOs in our sample legitimize.Irrespective of their purpose, design, or regional location, IOs use public communication to justify their authority.Despite this general tendency, we find considerable variation across regions, IOs, and time.First, the data reveal that IOs from various world regions legitimize differently.With a share of 34 percent of legitimation statements over all coded paragraphs, African IOs legitimize less intensely than their counterparts from other world regions.At the other end of the distribution, Asian IOs legitimize most intensively (48 percent).With a mean intensity of 46 percent of legitimation statements over all coded paragraphs, European IOs and cross-regional IOs (45 percent) are also above the average, whereas American IOs (38 percent) are below it.
Second, we also find variation within regions.Except for African IOs, which constitute a homogenous group of low-intensity legitimizers, other world regions are home to IOs with different levels of legitimation intensity.This is most pronounced for cross-regional IOs.The most intense legitimizer--the OSCE (75 percent)--and the least intense legitimizer--the OAPEC (11 percent)--are part of this group.IOs from the Americas also vary widely.With SICA and the OECS, this region hosts the two IOs that rank second-highest and lowest in legitimation intensity.
Third, the data reveal notable temporal trends.Except for Asian IOs, intensity increased considerably during the 1980s and 1990s.This trend flattened in the second half of the observation period, and legitimation intensity stabilized at a higher level.The legitimation intensity of Asian IOs followed the reverse path.It was  Overall, our data demonstrate that legitimation intensity is a significant category for legitimation research and practice.Although all IOs in the sample legitimize, they do so in very different ways.This variation requires further analysis that explores the strikingly similar temporal trends across world regions or the vast difference between IOs from the same region.For IO representatives, these data constitute an important reference point for understanding legitimation requirements in different world regions.The results may indicate how much legitimation is appropriate and necessary and may also inform IO representatives about peer IOs' activities.When IOs are striving for recognition and acceptance in a specific issue area, legitimation intensity can be helpful.

Modes
Research on the modes of legitimation is limited to (comparative) case studies.Due to conceptual and methodological limitations, this work is selective in terms of the modes of legitimation analyzed.Studies focused on institutional legitimation highlight, for instance, that IOs have established democratic control mechanisms to foster belief in their legitimacy ( Grigorescu 2007 ;Tallberg et al. 2014 ;Rocabert et al. 2019 ).Research on behavioral legitimation has begun investigating how IOs use daily interactions, symbols, and events for legitimation ( Billerbeck 2019 ;Squatrito 2021 ;Hensell 2022 ).Yet except for Gronau and Schmidtke's ( 2016 ) analysis of how the IMF and the G8 have used various modes to manage legitimacy crises, there is no research on how IOs combine modes of legitimation.Our data thus breaks new ground, showing that discursive legitimation dominates across all world regions, that the three other modes also play a role, and that the blend of modes varies across organizations and over time.
Figure 4 depicts the modes of legitimation across world regions and over time.We include a trendline for the population mean.Figure 5 displays variation across IOs.Additionally, the figure provides information on an "average organization" (Ø) to facilitate comparisons of individual IOs with the sample mean.Both figures present the share of each mode of legitimation across the total number of coded legitimation statements. 12hree observations stand out.First, discursive legitimation dominates across all world regions.With an overall share between 40 and 79 percent, this mode is preferred by almost all IOs in the sample.The Organization of American States (OAS) is the only exception, favoring legitimation via institutional substantiation.The dominance of discursive legitimation decreased from the 1980s until the 2000s, only to recover during the second half of the observation period.Although our methodology may privilege discursive legitimation to some extent, this observation suggests that IOs prefer the least costly mode of legitimation.While adapting behavior, policies, or institutions requires more significant human and financial resources and may thus constitute more consequential modes of legitimation, the costs of discursive legitimation lie primarily in the potential credibility costs if IO representatives "overclaim" ( Schmidtke and Lenz 2023 ).
Second, the three other modes--behavior, policy, and institution--are relevant across all regions.Regional specificities and slight temporal fluctuation notwithstanding, their distribution is relatively stable over time.This observation suggests that IOs from different world regions and in varying historical contexts typically combine modes of legitimation in similar ways.Despite these similarities, some temporal trends and regional idiosyncrasies are worth highlighting.On one hand, the share of policy-based legitimation has increased in each region except for the Americas and Africa, where this mode has declined.On the other hand, institutional legitimation is more common in the Americas than in other regions, with a general downward trend in recent years.Lastly, behavioral legitimation is above average in Europe and the Asia-Pacific and below average for cross-regional IOs.
Third, turning to cross-sectional variation, figure 5 reveals that the ranking of all four modes of legitimation varies more across IOs than regions or over time.Overall, the predominance of discursive legitimation and the minor role of institutional legitimation are widely reflected in individual IOs.Policy-based and behavioral legitimation compete for the second and third ranks (policy-based surpasses behavioral legitimation in sixteen IOs).Hence, only eight IOs in our sample do not use institutional legitimation most infrequently and discourse most frequently.The regional background of this group is also heterogeneous, as all regions, except for Asia, are represented.
Overall, our data suggest that mapping IOs' modes of legitimation via their communication pushes the methodological frontier forward.Our approach provides a novel way to overcome the selectivity that has plagued much of the literature and to reliably map various modes of legitimation.Further, in affirmation of the literature, the results support the idea that IOs legitimize in multiple ways.While we find support for the focus on discursive legitimation, our data show that institutional legitimation is rare and less important than behavioral legitimation.Similarly, the data demonstrate that policy-based legitimation, so far overlooked in empirical research, warrants more attention.For practitioners, these findings suggest that more strongly exploiting the potential for institutional legitimation may be a way to get an edge over competitors.

IO Authority and Legitimation
This section provides a second illustration of how the LegRO dataset can contribute to addressing important questions in the literature that explore the relationship between legitimation and the authority of IOs.This demonstrates how our data can contribute to the systematic analysis of the causes of IO legitimation.Moreover, using the LegRO data as a set of independent variables may contribute to a better understanding of the consequences of IO legitimation.
Existing research often suggests that the authority enshrined in IO mandates is an essential driver of how IOs try to shape audiences' legitimacy beliefs ( Schmidtke 2019 ;Dingwerth, Schmidtke, and Weise 2020 ;Rauh and Zürn 2020 ).Following Weber ( 1978 ), the basic argument of these studies is that "the exercise of authority requires justification" ( Zürn 2018 , 63).Proponents of the "authority-legitimacy link" ( Tallberg and Zürn 2019 ) suggest that IOs with higher levels of authority are likely to use more intensive legitimation that entails more demanding standards and modes of legitimation.
Due to the lack of systematic data across a wide range of IOs, these relationships between authority and various components of legitimation have not been tested systematically.We explore this link between IO authority and legitimation by estimating a set of multilevel regression models, including an IO random intercept.We use three measures of IO authority included in the MIA dataset ( Hooghe et al. 2017 ;Hooghe, Lenz, and Marks 2019 ).First, we refer to pooling, defined as the extent to which member-state bodies move from the unanimity principle toward various forms of majority voting across six decision areas: membership accession, membership suspension, policymaking, budgetary allocation, noncompliance, and constitutional reform.The variable is an aggregate index ranging from zero to one (low to high).Second, we use delegation , defined as the extent to which member states empower supranational agents to set the agenda and make decisions across the six decision areas mentioned above.This aggregate index ranges from zero to one (low to high).Third, we apply policy scope , counting the number of policies from a predefined list of twenty five policy areas in which an IO holds competence.This variable ranges from zero to twenty five (from narrow to comprehensive).Moreover, we control for variation in the availability of our coded material.Ideally, our data would build on one annual report and one communiqué per IO year.However, we could only locate and code one of these documents in some cases.As the dependent variables pool information from both documents, having only one of them potentially biases the results.
To test the idea that more IO authority is associated with more intense and more demanding legitimation, Model 1 examines the association between authority and the share of legitimation highlighting democratic standards, Model 2 uses legitimation intensity as the dependent variable, and Model 3 explores the link between authority and the share of institutional legitimation.Since the annual scores of the dependent variables are sensitive to individual coder decisions and the drafting process of the documents analyzed, we use a two-year rolling mean in the estimations.To avoid simultaneity, we lag all explanatory variables by one year.
Table 4 summarizes the results.It shows a clear and consistent association between pooling and all three components of legitimation.The statistically significant coefficients for pooling demonstrate that IOs that shift from unanimity or consensus to majority voting tend to intensify legitimation, emphasize democratic standards, and substantiate their claims by highlighting institutional reforms (on the link between IO authority and legitimation intensity, see Schmidtke and Lenz 2023 ).By contrast, the coefficients for delegation are not statistically significant and have negative signs in Models 2 and 3. Regarding policy scope, the results are mixed.We find a positive statistically significant association with legitimation intensity, but no statistically significant relationship with democratic and institutional  legitimation.Taken together, these results suggest that pooled authority may be an important driver of IO legitimation, whereas delegation and policy scope might play a less important role.In future research, this more specific link between IO authority and legitimation can be explored further with the help of the LegRO dataset.

Conclusion
In this article, we have presented the LegRO dataset, which provides information on the legitimation practices of twenty eight regional IOs from 1980 to 2019.This novel data substantively improve the existing legitimation research thanks to its broad coverage of IOs from different world regions and issue areas.It introduces measures of the standards, intensity, and modes of IO legitimation that were previously unavailable.To conclude, we highlight three areas of global governance research to which the data can contribute, summarizing their value to policymakers.First, the descriptive patterns in IO legitimation presented here have implications for current debates in IR scholarship.By shedding new light on the nature of IO legitimation, the dataset may inform expectations about the drivers of IO design and behavior.The descriptive patterns with respect to the standards, intensity, and modes of legitimation indicate the normativity of IOs.IO representatives care deeply about an IO's legitimacy; aim to derive it from various standards; and may even adapt the IO's behavior, policy output, and institutional design to legitimize its authority.
Second, the dataset invites a more systematic exploration of IO legitimation as a dependent variable.IOs invest considerable resources in legitimation by adapting public communication, behavior, policy output, and institutional features.What explains the variation in IO legitimation across organizations and over time?Why do IOs use specific legitimation profiles combining different standards and modes of legitimation?Do they aim to please particular audiences, are their representatives' normative predispositions a driver, or do they learn from their organizational environment ( Lenz and Söderbaum 2023 )?Third, researchers can use the data to test theories that feature IO legitimation as an explanatory factor.For example, elite framings, such as IO legitimation, are often theorized as shaping popular legitimacy beliefs about IOs.By framing their existence and work positively, using different standards and modes of legitimation, IOs compete with other elites to shape audiences' legitimacy beliefs.LegRO provides variables that researchers can use to further explore the origins of popular legitimacy beliefs in global governance.Other areas of research for which IO legitimation is a relevant variable include institutional overlap and regime complexity, international norm research, and international institutional design.
Finally, the dataset provides important policy-relevant insights.In times of rising mass-level dissensus about international cooperation and the liberal international order, policymakers worldwide are paying increased attention to issues of legitimacy and legitimation.In this context, the data presented in this article can inform policymakers about major trends in IO legitimation, what standards and modes of legitimation can help stabilize the legitimacy of IOs, and how potentially dysfunctional IOs can be delegitimized.
Figure 1.Standards of legitimation across world regions and time, 1980-2019.

Figure 3 .
Figure 3. Intensity of legitimation across world regions, IOs, and time, 1980-2019.comparably high during the 1980s, declined until the decade 2000-2010, and began to increase when all the other IOs reached a plateau.Overall, our data demonstrate that legitimation intensity is a significant category for legitimation research and practice.Although all IOs in the sample legitimize, they do so in very different ways.This variation requires further analysis that explores the strikingly similar temporal trends across world regions or the vast difference between IOs from the same region.For IO representatives, these data constitute an important reference point for understanding legitimation requirements in different world regions.The results may indicate how much legitimation is appropriate and necessary and may also inform IO representatives about peer IOs' activities.When IOs are striving for recognition and acceptance in a specific issue area, legitimation intensity can be helpful.

Table 1 .
Legitimation grammarOES Example: "The Single Market is one of the great achievements of the Union which has delivered major benefits to Europeans.It is our main asset for ensuring citizens' welfare , inclusive growth and job creation, and the essential driver for investment and global competitiveness."(EuropeanCouncil 2018 , 1, authors' own emphasis) The Asia-Pacific Economic Cooperation (APEC) forum was established to take advantage of the growing interdependence among Asia-Pacific economies, by facilitating economic growth for all participants and enhancing a sense of community in the region .It aims to help improve trade and economic performance and regional links for the prosperity of the people in the region."(Asia-Pacific Economic  Cooperation 2007 , 13, authors' own emphasis)

Table 2 .
Standards of legitimation Sommerer et al. 2022ion intensitSuchman 1995 )er IOs are attempting to create, maintain, or defend their legitimacy( Ashforth and Gibbs 1990 , 182;Suchman 1995 ).When IOs aim to create legitimacy, legitimation tends to be intensive because IO representatives are attempting to win the support of new audiences.Once legitimacy has been established, maintaining it often requires less effort.Defending legitimacy in times of crisis is also likely to require intensive legitimation because IOs are then attempting to counter concrete delegitimation ( Ecker-Ehrhardt 2018 ;Sommerer et al. 2022).