Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Soft governance : why states create informal intergovernmental organizations, and why it matters Roger, Charles Barclay 2016

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


24-ubc_2016_september_roger_charles.pdf [ 11.17MB ]
JSON: 24-1.0307284.json
JSON-LD: 24-1.0307284-ld.json
RDF/XML (Pretty): 24-1.0307284-rdf.xml
RDF/JSON: 24-1.0307284-rdf.json
Turtle: 24-1.0307284-turtle.txt
N-Triples: 24-1.0307284-rdf-ntriples.txt
Original Record: 24-1.0307284-source.json
Full Text

Full Text

!      SOFT GOVERNANCE: WHY STATES CREATE INFORMAL INTERGOVERNMENTAL ORGANIZATIONS, AND WHY IT MATTERS   by  CHARLES BARCLAY ROGER  B.A., Concordia University, 2007 M.Sc., London School of Economics and Political Science, 2008     A DISSERTATION SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF   DOCTOR OF PHILOSOPHY  in  THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES  (Political Science)       THE UNIVERSITY OF BRITISH COLUMBIA  (Vancouver)  July 2016   © Charles Barclay Roger, 2016  ii Abstract  Informal intergovernmental organizations have become a prominent feature of the global landscape. Yet it remains unclear why states create informal organizations in some instances and formal organizations in others. Thus far, scholars have argued that states choose to create informal organizations when they offer an “efficient” solution to certain kinds of cross-border cooperation problems. However, such functionalist arguments are underspecified and rest on weak evidence at present. Existing research suggests that functionalist theories can indeed explain certain cases, but numerous anomalies arise when we look at others. This dissertation argues that this is because functionalists do not take into account how domestic politics, distributional conflict and state power can decisively influence the kinds of organizations that are likely to appear. It offers an alternative account of the emergence of informal organizations that incorporates these variables. The theory advanced emphasizes how domestic politics and institutions structure state preferences over organizational form, and how the distribution of preferences and state power then shape the organizations that subsequently emerge. Specifically, it argues that informal organizations arise when either a) policymakers in powerful states face significant domestic constraints, or b) autonomous bureaucrats are given responsibility for “leading” cooperation on the behalf of powerful states. In order to test this theory, a variety of methods are used. First, the theory is evaluated quantitatively through a statistical analysis of an original dataset of formal and informal organizations. Second, the theory is evaluated qualitatively through process tracing of the “emergence” of the International Monetary Fund, the General Agreement on Tariffs and Trade and the International Competition Network. Overall, the analysis provides powerful support for the central thesis of this dissertation: while certain aspects of the cooperation problems states face do play a role, domestic politics and state power are the most important determinants of organizational form. The dissertation’s findings are argued to have implications for theories of rational design in the field of International Relations, for our understanding of the overall rise of informal organizations in the global system, as well as for policy debates about the desirability of this new breed of international institution.            iii Preface  This dissertation is the original, unpublished, independent work of the author, Charles Roger. The research was approved by the University of British Columbia’s (UBC) Behavioral Research Ethics Board (BREB), BREB Number H15-01249.                                        iv Table of Contents  Abstract ........................................................................................................................................... ii Preface ............................................................................................................................................ iii Table of Contents ........................................................................................................................... iv List of Tables ................................................................................................................................. vi List of Figures ............................................................................................................................... vii Acknowledgements ...................................................................................................................... viii Chapter 1. Introduction: The Rise of Informal Organizations ........................................................ 1 1.1 Introduction ........................................................................................................................ 1 1.2 A Preview of the Argument ............................................................................................. 10 1.3 Alternative Explanations ................................................................................................. 21 1.4 What is at Stake? ............................................................................................................. 24 1.5 Methods and Limitations of this Study ............................................................................ 27 1.6 Outline of the Study ......................................................................................................... 30 Chapter 2. Conceptualizing and Mapping Organizational Forms ................................................ 34 2.1 Introduction ......................................................................................................................... 34 2.2 The Key Features of Formal and Informal Organizations .................................................. 35 2.2.1 The Properties and Implications of Formal and Informal Organizations .................... 42 2.3 Mapping the World of Informal Organizations .................................................................. 54 2.3.1 Previous Efforts ........................................................................................................... 55 2.3.2 Constructing the Database ........................................................................................... 61 2.3.3 Preliminary Findings .................................................................................................... 67 2.4 Conclusion .......................................................................................................................... 70 Chapter 3. Theorizing Informal Organizations ............................................................................. 72 3.1 Introduction ......................................................................................................................... 72 3.2 Previous Efforts .................................................................................................................. 73 3.2.1 A Critique of Functionalist Theories ........................................................................... 83 3.3 A Liberal Theory of Organizational Form .......................................................................... 89 3.3.1 Step 1: A Model of State Preference Formation .......................................................... 91 3.3.2 Step 2: A Model of State Preference Aggregation ..................................................... 100 3.4 Conclusion ........................................................................................................................ 109 Chapter 4. Analyzing the Determinants of Organizational Form ............................................... 114 4.1 Introduction ....................................................................................................................... 114 4.2 Defining a Sample of Intergovernmental Organizations .................................................. 115 4.3 Explanatory Variables ....................................................................................................... 121 4.3.1 Measuring Institutional Autonomy ............................................................................ 122 4.3.2 Measuring Domestic Constraints ............................................................................... 131 4.4 Quantitative Analysis ........................................................................................................ 138 4.4.1 Simple Analyses of the Data ...................................................................................... 138 4.4.2 Accounting for Alternative Explanations .................................................................. 142 4.4.3 Addressing Selection and Endogeneity Concerns ..................................................... 158   v 4.5 Conclusion ........................................................................................................................ 163 Chapter 5. The Emergence of the IMF and GATT ..................................................................... 165 5.1 Introduction ....................................................................................................................... 165 5.2 Case Selection ................................................................................................................... 168 5.3 Probing the Validity of Rational Design Hypotheses: A Congruence Analysis ............... 171 5.4. Examining Causal Mechanisms: Within-Case Analyses ................................................. 180 5.5 The Emergence of the IMF as a Formal Organization ..................................................... 182 5.5.1 Demand and Institutional Selection ........................................................................... 182 5.5.2 The Formation of Preferences Over Organizational Form ........................................ 188 5.5.3 The Negotiation and Establishment of a New Intergovernmental Organization ....... 195 5.5.4 Summary .................................................................................................................... 199 5.6 The Emergence of the GATT as an Informal Organization ............................................. 202 5.6.1 Demand and Institutional Selection ........................................................................... 202 5.6.2 The Formation of Preferences Over Organizational Form ........................................ 207 5.6.3 The Negotiation and Establishment of a New Intergovernmental Organization ....... 213 5.6.4 Summary .................................................................................................................... 220 5.7 Conclusion ........................................................................................................................ 222 Chapter 6: The Emergence of the ICN ....................................................................................... 224 6.1 Introduction ....................................................................................................................... 224 6.2 Case Selection ................................................................................................................... 226 6.3 Probing the Validity of Rational Design Hypotheses: A Congruence Analysis ............... 229 6.4 Examining Causal Mechanisms: A Within-Case Analysis ............................................... 239 6.4.1 Demand and Institutional Selection ........................................................................... 240 6.4.2 The Formation of Preferences Over Organizational Form ........................................ 248 6.4.3 The Negotiation and Establishment of a New Intergovernmental Organization ....... 261 6.5 Conclusion ........................................................................................................................ 268 Chapter 7. Conclusion: Findings, Implications, and Future Research ........................................ 272 7.1 Introduction ....................................................................................................................... 272 7.2 Theoretical Implications ................................................................................................... 275 7.3 Empirical Implications ...................................................................................................... 282 7.4 Policy Implications ........................................................................................................... 289 7.5 Future Research ................................................................................................................ 296 7.6 Conclusion ........................................................................................................................ 300 Bibliography ............................................................................................................................... 302 Appendix 1 .................................................................................................................................. 331 Appendix 2 .................................................................................................................................. 336      vi List of Tables  Table 1 Properties of Formal and Informal Organizations ........................................................... 53 Table 2 Criteria, Indicators and Evidence Determining Inclusion in the Dataset ........................ 67 Table 3 Functionalist Determinants of Organizational Form ....................................................... 78 Table 4 Simple Bivariate and Multivariate Regressions ............................................................. 140 Table 5 Bivariate and Multivariate Regressions ......................................................................... 151 Table 6 Robustness Checks ........................................................................................................ 156 Table 7 Summary of Rational Design Predictions ...................................................................... 179 Table 8 Summary of Rational Design Predictions ...................................................................... 238 Table 9 List of Intergovernmental Organizations in the Sample ................................................ 331 Table 10 Descriptive Statistics: IGO Dataset Versus Sample .................................................... 337 Table 11 Descriptive Statistics: Domestic Constraints ............................................................... 337 Table 12 Descriptive Statistics .................................................................................................... 338 Table 13 Correlation Matrix ....................................................................................................... 339 Table 14 Collinearity Diagnostic ................................................................................................ 340 Table 15 Balance Diagnostic ...................................................................................................... 340 Table 16 Estimation Results ....................................................................................................... 341                             vii  List of Figures  Figure 1 Intergovernmental Organizations in the Global System ................................................ 69 Figure 2 Model of the Preference Formation Process ................................................................ 100 Figure 3 Visualizing the Distribution of State Preferences ........................................................ 104 Figure 4 Bargaining Scenarios With Divergent Preferences ...................................................... 107 Figure 5 Temporal Patterns in the Sample of Intergovernmental Organizations ....................... 120 Figure 6 Autonomy Distribution .................................................................................................. 130 Figure 7 Constraints Distribution ............................................................................................... 137 Figure 8 Differences in Means .................................................................................................... 139 Figure 9 Marginal Effects on the Probability of Informality ...................................................... 142 Figure 10 Average Marginal Effects ........................................................................................... 153 Figure 11 Balance Density Plot .................................................................................................. 162 Figure 12 The Impact of Changing Cooperation Problems, 1950-2005 ..................................... 288 Figure 13 The Impact of Institutional Autonomy and Domestic Constraints, 1950-2005 ......... 289 Figure 14 Matched and Mismatched Intergovernmental Organizations ..................................... 295 Figure 15 Post-1950 Temporal Trends in IGO Dataset .............................................................. 336 Figure 16 Post-1950 Temporal Trends in the Sample ................................................................ 336      viii Acknowledgements  As with all research, this project has benefited from the insights and support of a great many people. Above all, I would like to recognize the contribution of my doctoral supervisor, Peter Dauvergne, and my committee members, Katia Coleman and Yves Tiberghien. They have each read through numerous drafts of the dissertation and offered valuable comments on each of them that improved the conceptual framework, theory and empirics immeasurably. They have also each been great supporters of my work throughout my doctoral studies, and for all of that I am deeply thankful. Second, I am indebted to Felicity Vabulas and Duncan Snidal for providing much of the inspiration for this work. Their pioneering research has offered a wellspring of useful ideas to engage with, and I have benefited from the comments they have offered on this project. Third, I am thankful that a number of colleagues and friends have also taken time out of their busy schedules to read through and comment on portions of this work, or who provided valuable support in other ways. These individuals include Liliana Andonova, Jessica Green, Philipp Genschel, Thomas Hale, Derek Hall, Virginia Haufler, David Held, Eric Helleiner, Kathy Hochstetler, Matthew Hoffmann, Marieke Kleine, Rahul Mediratta, Joost Paulwyn, Stacy VanDeveer, Robert Wade, Nicole Weygandt, and Kevin Young. Fourth, I have also benefited from the friendship and scholarship of my colleagues at the University of British Columbia, many of whom have gone above and beyond the call of duty in terms of their support for this project. These include Justin Alger, Jen Allan, Michael Byers, Arjun Chowdury, Brian Job, Li Xiaojun, Jane Lister, Miriam Matejova, Richard Price, Paul Quirk, Allen Sens, Lisa Sundstrom, Brent Sutton, Chris Tenove, Carla Winston, and Linting Zhang. Finally, of course, this project could not have been completed or even begun without the unfailing support of my family and close friends. They have been there at each and every stage and know far more about studying informal intergovernmental organizations, as well as the ups-and-downs and ins-and-outs of completing a PhD, than anyone outside of academia ever should.  Beyond these individuals, I also wish to acknowledge the financial support that I have received. My doctoral studies were funded by a Joseph Armand Bombardier Canada Graduate Scholarship, the John S. Montalbano Scholars Fellowship, and the Olav A. Slaymaker Fellowship in Environment. I also received support for this project, in particular, from the Department of Political Science and the Faculty of Graduate Studies at UBC, as well as the Liu Institute for Global Issues.  1   Chapter 1. Introduction: The Rise of Informal Organizations  1.1 Introduction  The financial crisis of 2007-08 was a great test for global governance. States were faced with the most significant economic turmoil since the Great Depression: economic output plummeted, financial institutions crumbled, and international trade and cross-border capital flows collapsed. With the global economy on the brink of disaster, states called upon international institutions to help avert a catastrophe. The Group of 20 (G20) achieved new prominence as the “international steering committee of the world economy.”1 Leaders from both industrialized and emerging economies convened in 2008 and 2009 to coordinate their national responses. The International Monetary Fund (IMF) swung into action to provide emergency loans, with a massive infusion of funds bolstering its capacity.2 The Financial Stability Forum—established in the wake of the East Asian financial crisis that had occurred ten years earlier—was reconstituted with a new charter, expanded membership and a new name: the Financial Stability Board (FSB). United States (US) Treasury Secretary Timothy Geithner would refer to it as the “fourth pillar” of the global economic governance architecture—alongside the IMF, World Bank and World Trade Organization (WTO).3 The Group of 10 (G10) and the Basel Committee on Banking Supervision (BCBS) recognized the inadequacy of Basel II and within two years concluded a new capital adequacy standard for banks, Basel III.4 The activities of other institutions in the field of global finance, such as the International Organization for Securities Commissions (IOSCO) and the !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!1 Bergsten 2006; Cooper and Thakur 2013. 2 Helleiner 2010; Blustein 2013; Helleiner 2014c. 3 Griffith-Jones, Helleiner, and Woods 2010. 4 Young 2011; Vestergaard and Hojland 2011.  2   International Association of Insurance Supervisors (IAIS), also expanded significantly. 5 Together, these bodies mounted the global response to the crisis and helped to chart a path to what was hoped would be a more stable world.  Some have argued that one of the most surprising stories of the financial crisis is that these institutions were largely successful—that the system worked. 6  Initially, economic output, international trade and capital flows had fallen far faster than they had in 1929. Yet they recovered much faster as well. Even if growth remained sluggish thereafter, a second Great Depression was clearly prevented. Daniel Drezner has attributed this, at least in part, to the fact that international institutions performed tolerably well during and after the crisis. “The open global economy survived,” he argues, “because of ‘good enough’ governance.”7 Others have disputed this, arguing that global governance was not “good enough”—that the system failed. David Zaring, for instance, has concluded that many of the most important institutions “had little to say in response to the [financial shocks].” “Those responses that were forthcoming,” he claims, “were either ineffective or counterproductive.” 8  Eric Helleiner has echoed this assessment, to some degree. According to him, “the G20’s contribution to the financial management of the crisis ended up being much less significant than advertised […] And the FSB’s capacity to act as a kind of fourth pillar of global economic governance turned out to be very limited.”9 The achievements of other bodies, like the BCBS, IOSCO and IAIS, have been !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!5 Pagliari 2012. 6 Drezner 2014. 7 Ibid, p.124. 8 Zaring 2010, p.475. 9 Helleiner 2014c, p.3. Also see Helleiner 2012b; Helleiner 2012a.  3   similarly downplayed.10 Such observers believe that while it is certainly true that there was a flurry of activity at the international level and that the worst was avoided, the actual contribution of international institutions to this outcome was negligible.   The debate between these two views is an important one, and remains unresolved. It highlights the fact that we need to think carefully about the way institutions operate, and how they can be improved in the future—even if their response was “good enough” this time around.11 In many ways, the systemic risks that we face, from the environment to the economy to global health, are too great. Yet there is a third story here—one that many could be forgiven for missing, but which may shed some light on the issues at stake. That story is about the major changes that we are presently witnessing in the nature of global governance. Specifically, the global response to the financial crisis underlined the extent to which informal, or “soft,” intergovernmental organizations have come to dominate our approach to important global issues. The G20 was, for instance, central to the global response to the crisis, and its rise has been heralded as one most important developments in international affairs in recent years.12 Yet it is unlike many of the intergovernmental organizations we have “traditionally” relied on and studied in the field of International Relations (IR).13 While it has a centralized organizational structure, it has no permanent secretariat and no formal status under international law. The FSB has a secretariat but it primarily comprises officials that have been seconded by states, and, again, its charter is a “soft” or “non-binding” agreement. The same is true of the G10, BCBS, IOSCO and IAIS. Of !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!10 For similar views, see Brummer 2012 and Blustein 2013. 11 See Helleiner and Pagliari 2011 and Kahler 2013 for useful discussions of some of the larger issues raised by the financial crisis for global governance and IR research. 12 Cooper and Thakur 2013; Kirton 2013. 13 Kahler 2000; Kahler and Lake 2003; Abbott 2015.  4   course, more traditional—or formal—intergovernmental organizations, like the IMF and WTO, played an important role in the response to the crisis as well, but the extent to which informal organizations occupied the news headlines was striking. Even in the highly legalized European context the trend was evident. In response to the debt crisis, for instance, three institutions have been especially prominent in Europe: the European Central Bank (ECB), the European Commission and the Euro Group. The first two are formal intergovernmental organizations; the last has, until quite recently, had no formal legal basis.14  The pattern is hardly confined to the field of global finance. Informal organizations15 can be found in a variety of issue areas. On climate change, for instance, alongside the United Nations, there is the Clean Energy Ministerial, the Group on Earth Observations, the Global Methane Initiative and the Carbon Sequestration Leadership Forum. Until its demise in 2011, the Asia Pacific Partnership on Clean Development and Climate also played a role. In the field of money laundering, there is the Financial Action Task Force (FATF), the Asia/Pacific Group on Money Laundering (APG), the Eurasian Group (EAG) and the Egmont Group. In the field of consumer health and safety, there is the International Cooperation on Cosmetics Regulation (ICCR), the Intergovernmental Forum on Chemical Safety and the Global Harmonization Task Force, among many others. Lest the impression be given that informal cooperation is confined to “low politics” issue areas, in the field of international security there is the Australia Group, the Nuclear Suppliers Group, the Proliferation Security Initiative (PSI), the Missile Technology Control !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!14 This changed with the signing of the Lisbon Treaty in 2009. 15 Throughout this study, the terms “informal organization” and “formal organization” are used interchangeably with “informal intergovernmental organization” and “formal intergovernmental organization.” When I use the term “intergovernmental organizations,” without qualification, I mean to refer to both formal and informal organizations.  5   Regime, and the Wassenar Arrangement (formerly known as CoCom, the Coordination Committee for Multilateral Export Controls). Like the G20 and FSB, each of these institutions has few of the features that we associate with “traditional” intergovernmental organizations. Their institutional or bureaucratic footprint is “light” and their standing under international law is uncertain, at best. Yet, on the whole, these new informal institutions—and many others just like them—are impacting the lives of nearly everyone around the world. Further, their importance seems to be growing.  Few would dispute that informal organizations have been playing a more central role in global affairs. Yet, curiously, few attempts have been made to demonstrate this fact conclusively, explain why it has occurred or sought to understand whether this is a good thing. Has there really been a shift in the nature of global governance? If so, how dramatic has the change really been? If a major shift has indeed occurred, what caused this? Why do states create formal organizations in some instances and informal organizations in others? And what, ultimately, does our answer to this question imply for our ability to govern important global issues? The purpose of this study is to answer these questions. It does so by developing a theory that aims to explain why states choose to create informal organizations, testing it using a range of methods and assessing what this says about the desirability of this new breed of international institution. However, I argue that a great deal needs to be done before this can be accomplished. At present, we still have only a limited understanding of how to conceptualize and think about informal organizations. Thus, there is significant need for clarity about what really “counts” as an informal organization, and how such bodies are different from formal organizations, like the United Nations or World Bank. Finally, but relatedly, we have only a limited understanding of the actual empirical patterns of  6   governance that we see. Currently, we have fairly good data on the formal intergovernmental organizations that exist, or which have existed, but the data on informal organizations is quite limited. Our empirical understanding of—and ability to explain—their rise is therefore deficient. We have impressions and anecdotes that give us a “sense” of what is going on, not data or statistics.   This study makes a number of important contributions to our understanding of informal organizations. First, it offers a much needed conceptual clarification of their main features, properties, and implications, explaining exactly how they are similar to and different from the kinds of formal organizations that have been the traditional focus of scholarship in the fields of IR and International Law (IL). In recent years, scholars of IR and IL have made some effort to explain what is special about informal organizations, and their work has paved the way for more ambitious research.16 This study builds on their contributions. Yet, I argue, prevailing efforts still suffer from the fact that they miss or misunderstand some of the key characteristics and consequences of formal and informal organizations. It therefore attempts to specify exactly what informal organizations are, and how we can know them when we see them. It explains in considerable detail how informal organizations are different from formal organizations, and clarifies what this means in practical terms. In line with existing research, for instance, it shows how formal and informal organizations are able to do different things after they have been established. I refer to these unique abilities of formal and informal organizations as their “international properties,” since they relate to how an organization functions or performs. Yet the study also shows that creating each kind of organization can have different practical implications !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!16 Klabbers 2001a; Pauwelyn, Wessel, and Wouters 2012; Vabulas and Snidal 2013.  7   for states as well. I refer to these as the “domestic implications” of formal and informal organizations, since they concern what it really means to establish them in the first place. To some degree, of course, the domestic implications and international properties of these organizations are related and interlinked, but this distinction—previously overlooked—turns out to be quite important for how we explain why a formal or informal organization emerges in a given context.   The second contribution of the study is to offer an empirical assessment of the phenomenon by collecting new data on the number of informal organizations that have emerged over time. As mentioned, existing efforts to measure the number of intergovernmental organizations in the global system primarily focus on formal bodies. This means that we have no reliable way of knowing how many informal organizations exist and whether we really are seeing something new.17 It could be that the hype about informal organizations is overblown. Perhaps, for instance, they have been around for some time, and we have a misleading perception about the frequency of informal organizations because they have only attracted more systematic attention in recent years. Perhaps, as well, there are very few informal organizations, and therefore we can safely overlook them without too much loss. Inspired by a similar effort by Felicity Vabulas and Duncan Snidal, the study rectifies this deficit by developing new coding rules that operationalize the concept of an informal organization and creating an original dataset that helps expand our understanding of intergovernmental organizations in important ways.18 This dataset employs the same source material and methods used to generate the most comprehensive existing dataset on !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!17 It also implies that we are, as shown later on, seriously underestimating the total number of intergovernmental organizations. 18 See Ibid.  8   formal organizations—the Correlates of War Intergovernmental Organizations Database (COW-IGO)—and can therefore be reliably combined with the latter to compare the number of formal and informal organizations over a period of nearly 200 years. The data demonstrate, conclusively, that informal organizations have indeed become a more prominent feature of the global political landscape over the last 50-60 years. In 1945, it shows, they were relatively rare; today, they comprise about a third of all active intergovernmental organizations. Perhaps even more importantly, however, the data also open up new research frontiers by allowing scholars to quantitatively analyze the determinants of organizational form for the first time—a task undertaken in Chapter 4.  The third, and most important, contribution of the study is to move beyond conceptualization and description to offer a more compelling explanation of informal organizations than we have at present. To the extent that they exist, prevailing accounts are largely functionalist in nature, couched in the rational design tradition of IR.19 These hypothesize, generally, that informal organizations emerge when they offer benefits to states that their formal counterparts do not.20 As such, they have some intuitive appeal, and, as will be shown, some explanatory power as well. Yet, in the final analysis, they remain weak and incomplete. First, such theories have yet to be tested in a rigorous manner. Existing case studies show that the arguments that have been advanced are indeed plausible. But their ultimate value remains uncertain, since they focus on “easy tests” and do not demonstrate that the functionalist account can explain a wide range of cases. Worryingly, some contradictory evidence suggests that they may not, or that they miss !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!19 Koremenos, Lipson, and Snidal 2001; Koremenos 2013b; Koremenos forthcoming. 20 Whytock 2005; Eeilstrup-Sangiovanni 2009; Bradford 2011; Vabulas and Snidal 2013. Important exceptions include Bach 2010; Verdier 2013.  9   important parts of the story at least. This relates to the second concern, which is that rational design arguments tend to neglect the role of domestic politics, distributional conflict and power in driving an organization’s level of formality. This is a problem that plagues theories of rational design more generally, but it seems especially problematic when it comes to explaining organizational form and may help to account for such empirical anomalies. The explanation that I offer attempts to incorporate these variables, and shows how they systematically change the outcomes we would otherwise expect. Further, I test this theory—as well as rational design hypotheses—in a more rigorous fashion and demonstrate that the factors it identifies do indeed play an important role in determining the kinds of organizations that emerge. Indeed, the evidence provides powerful support for this study’s central thesis: while characteristics of a cooperation problem do shape the form of an organization to some extent, domestic politics and state power are in fact the primary drivers.  Finally, the study uses these findings to contribute to the normative debate about informal organizations. Many commentators have, thus far, been quite optimistic about these new institutions, arguing that they are better suited to the complex cooperation problems that states face at present.21 Informal organizations, they believe, have a number of advantages that make them more appropriate for addressing the new kinds of issues we confront in an increasingly globalized world. These arguments coincide with a more general disillusionment with formal institutions. Prominent intergovernmental organizations, such as the United Nations, IMF, World Bank, and WTO, as well as others like the European Union (EU) and World Health Organization !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!21 Slaughter 1997; Reinicke and Deng 2000; Slaughter 2004; Patrick 2014; Roger and Dauvergne 2016.  10   (WHO), have been subject to withering criticism in recent years. 22   Both scholars and policymakers have lamented inter alia their slow bureaucratic pace, their lack of transparency and legitimacy and their inability to adapt to new challenges. This is not to say that informal organizations have gone without criticism. Others are more pessimistic and worry that informal organizations are, at best, a distraction from serious solutions to the world’s problems.23 At worst, they actively undermine effective and accountable global governance. The debate between these rival views overlaps with the debate about the response to the financial crisis, mentioned above, and takes on increasing significance in a world where informal organizations have come to dominate efforts to resolve some of the most pressing issues that we face—financial crises being just one. What it lacks, at present, is a theoretically informed and evidence-based assessment of the value and appropriateness of informal organizations. By undertaking a rigorous investigation of the determinants of organization form and drawing out the policy implications of the findings, this study aims to put the debate about the desirability of these institutions on a new footing.   1.2 A Preview of the Argument  As mentioned above, the central question that this study addresses can be simply stated: why do states create formal organizations in some instances and informal organizations in others? By developing our understanding of what drives states to create informal organizations in the first place, I argue, we can start to explain why informal organizations have become more common !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!22 Stiglitz 2002; Blustein 2013; Patrick 2015; Garrett 2015. Also see Hale, Held, and Young 2013. 23 Karlsson-Vinkhuyzen and Vihma 2009; Verdier 2009; Zaring 2010; Blustein 2013; Helleiner 2014c.  11   over time and gain insights into whether—or when—they are appropriate for resolving global problems. To do so, it undertakes two main theoretical tasks. First, it offers a conceptual argument about what informal organizations really are and how they differ from their formal counterparts. It makes the case that informal organizations have certain features that make them quite unique, but also, as described above, that they have certain international properties and domestic implications, which may make them more or less desirable in different contexts and for different kinds of actors. Second, I advance an explanatory theory that attempts to model the processes of state preference “formation” and interstate preference “aggregation” that are hypothesized to jointly determine the form that an organization is likely to take.24 The theory emphasizes the role of domestic politics and institutions in shaping state preferences over organizational forms and highlights how the distribution of preferences and state power then determine the final outcomes that we see. The argument is laid out in full later on (in Chapter 3), and is elaborated upon in the Conclusion to explain the temporal trend that we see. Here, I offer a brief summary.  First, consider the conceptual argument: that formal and informal organizations differ in important ways and that these differences shape both how they are established and how they function or perform. To see this, it is helpful to contrast the two and to start by identifying the main features of formal organizations, which most readers will be familiar with. Formal organizations are, of course, a central feature of the global political landscape, and a key unit of !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!24 Note that I differentiate between preference “formation” at the domestic level and preference “aggregation” at the international level. I use these particular words to emphasize the distinctness of each stage. However, there is a sense in which the domestic preference formation process involves a process of preference “aggregation” as well, albeit with respect to the various domestic actors active in an issue area. While true, I maintain the distinction in order to facilitate exposition of the theory.  12   analysis in the field of IR. They come in many shapes and sizes, and address different issues. This diversity makes it inherently challenging to define them precisely. Nevertheless, most political scientists and international lawyers believe that they share at least three basic characteristics.25 First, formal organizations are created by sovereign states. Though other actors may be involved in various ways, organizations created by non-state actors (multinational corporations, non-governmental organizations), sub-state actors (cross-border associations of municipalities, for example) and other international institutions (so-called “emanations”) are not generally regarded as formal intergovernmental organizations. Second, they have bodies that are functionally separate from member states. Usually this would imply an autonomous body that is delegated responsibility for undertaking certain tasks on the behalf of states: a secretariat with a budget, staff and offices. Often, these can be quite large bureaucracies. Finally, they are established by an agreement under international law. That is, to be considered a formal intergovernmental organization a body must be founded by an international agreement that establishes precise legally “binding” obligations, i.e. an international treaty. 26  Formal organizations are, therefore, distinguished by the fact that they are a highly legalized form of international cooperation.27  With this definition in hand it is easier to see the similarities and differences. Like formal organizations, for instance, informal organizations are created by states. This is what qualifies !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!25 Singer and Wallace 1970; Schermers and Blokker 2003; Pevehouse, Nordstrom, and Warnke 2004; Klabbers 2009. 26 For the purposes of this dissertation, I refer to all agreements that establish international legal obligations as “international treaties,” even if they go by different names or include some “non-binding” elements in practice. On the distinction between legally “binding” and “non-binding” international agreements, see Schachter 1977; Aust 1986; Aust 2000.  27 Abbott et al. 2000.   13   them both as intergovernmental organizations. However, informal organizations are, generally, less likely to have bodies that are clearly separate from their members. The G20 and the FSB, as noted above, were created by states but have no independent secretariats of their own. Secretarial services are, instead, provided directly by one or more members, or by another intergovernmental organization. If they do have an independent secretariat, it is likely to be quite small; the sprawling bureaucracies characteristic of many formal organizations are almost entirely absent. Nevertheless, in order to count as an intergovernmental organization, an informal body must still have a persistent, if limited, organizational identity. At the very least, there must be regular meetings, policy outputs, and a corporate personality of some kind. The third and most important feature of informal organizations is their “twilight” status beyond the traditional boundaries of international law.28 This is, essentially, what makes them informal intergovernmental organizations. In contrast with their formal counterparts, which are established by international treaty, informal organizations are typically founded by “non-binding” agreements—a Memorandum of Understanding (MOU), “gentleman’s agreement,” or the like.29 The Terms of Reference of the APG offers a striking example of this. It describes the APG as a “multilateral organization” that is “established by agreement among its members and is autonomous.”30 Yet, at the same time, it explains, the organization is “voluntary and cooperative !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!28 Schachter 1977; Zaring 1998. 29 Note, again, that the terms used to refer to “non-binding” agreements are actually quite flexible in practice. Deciding whether an agreement is “legally binding” or not cannot simply be determined by the name itself. Ultimately, what matters is the intentions of the actors involved, which can only be determined by a holistic assessment of each agreement, including the language and surrounding practices of states. A more detailed discussion of these markers of informality is contained in Chapter 2 where I lay out specific criteria for determining whether an agreement is informal or not. Note, also, that in order to reduce confusion I refer to all non-binding agreements as “MOUs,” or as “non-binding agreements,” even if particular agreements go by different names. 30 APG 2012, p.6.  14   in nature,” and that its authority “does not derive from an international treaty.”31 Informal organizations are, therefore, a less legalized form of international cooperation.32  The distinction between formal and informal organizations would not be particularly relevant or interesting if it did not make a major difference in practice. Yet it does. Formal and informal organizations have different properties and implications that make them more or less useful, or more or less desirable in different contexts and for different actors. First, they have different international properties—that is, they can do different things. Given the nature of their constitutive agreements and the fact that they have only a limited organizational structure, informal organizations are generally more flexible, more agile, and offer greater confidentiality. That is, they can be changed more easily, they can move more swiftly, and they can be much less transparent. Formal organizations, by contrast, can be larger and more autonomous. That is, their bureaucracies can be more advanced and they can be granted greater operational independence from member states. These properties, it will be shown, may make these different kinds of organizations more or less efficient solutions to different kinds of problems—a fact which rational design theorists emphasize. However, these are not the only aspects of formal and informal organizations that matter for states. In fact, they may actually be of secondary importance in many instances. A second set of implications is domestic in nature—that is, creating one or the other kind of organization has very different practical implications for states—and, as will be discussed, for different actors within states. Formal organizations, for !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!31 Ibid. 32 The distinction between formal and informal organizations can, in fact, be restated in the terms developed by Abbott et al. 2000. While formal organization generally exhibit high levels of precision, obligation and delegation, informal organizations generally display lower levels of precision, obligation and delegation.   15   instance, typically have high ratification requirements, entail significant monitoring arrangements and require larger budgetary allocations. In some circumstances, these implications may actually benefit certain domestic actors. Informal organizations, by contrast, require few of these and we can expect other kinds of actors to prefer them for this reason. These domestic implications or consequences of formality and informality, it is worth noting, are emphasized by my own theory.  Now, consider the explanatory argument. The model that I advance takes the cooperation problem that states face as its starting point. Cooperative efforts begin with an increase in the “demand” for governance in a given issue area, which, it is assumed, creates an incentive for the creation of an intergovernmental organization of some kind.33 The theory predicts that the demand for governance is related to externalities that increase as interdependence between states deepens. However, these also create an incentive for particular actors within states to initiate or “lead” cooperation as a result of the fact that they have a stake in a particular issue. In this sense, the problem not only creates demand for cooperation, it also “selects” particular actors to become involved. The selection process matters—and, I argue, leads to unique observable implications—because different actors within states can reasonably be anticipated to have different “baseline” preferences over an organization’s level of formality. That is, some will tend to prefer a formal organization, while others will prefer an informal organization. Specifically, I distinguish between two different types of actors within states: political institutions, which are predicted to prefer formality, and independent agencies, which are predicted to prefer informality. While the former have been the “traditional” kinds of actors within states and !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!33 Keohane 1982; Moravcsik 1997.  16   involved in international affairs, the latter have become increasingly common over time and more actively involved in acting on the behalf of states internationally.34 We can therefore expect either of these to be “selected” by externalities created by a cooperation problem in any given situation and this leads us to expect different state preferences.  Why are these actors expected to have different preferences over organizational form? This is the most important logical step in the argument. The key is the fact that these actors occupy different positions within domestic institutional structures. Political institutions are distinguished by the fact that they are directly managed by or directly accountable to elected politicians. This means that actors leading these institutions are heavily influenced by electoral politics and their tenure is relatively uncertain—that is, they can lose their jobs if, for instance, other politicians or citizens are unsatisfied with their performance. This situation, I argue, leads these institutions towards a strong preference for greater formality for two reasons. First, since international treaties are seen to be more prestigious domestically (compared to soft “non-binding” agreements), greater legalization helps politicians to demonstrate leadership, bolster their reputations and improve their probability of being reelected. 35 In the words of Anthony Aust, “ministers prefer to be seen signing treaties rather than MOUs,” since they help to signal political strength to domestic audiences.36 Second, since greater legalization makes it more difficult for a state to “exit” an agreement (or intergovernmental organization) by raising the reputational costs of doing so, creating a more formal organization helps to “lock in” the policy preferences of a !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!34 Kennan 1997; Thatcher 2003; Slaughter 2004; Singer 2007; Bach 2010; Jordana, Levi-Faur, and i Marin 2011; Verdier 2013; Helleiner 2013. 35 See Aust 2000; Brewster 2004; Guzman 2008. 36 Aust 2000, p.38.  17   politician to a far greater extent.37 They will, therefore, be more certain that their policies will remain in place, even if they are removed from office at some point in the future. The domestic implications of formal intergovernmental organizations are, therefore, a key asset for actors within political institutions.  Independent agencies occupy a different position within a domestic institutional structure and this causes them to have a very different preference. Above all, they are distinguished by the level of autonomy they possess, which insulates them to a greater extent from electoral politics. The tenure of officials in independent agencies is also relatively more certain, by comparison, since they typically serve defined terms in office and can only be removed with great difficulty. Their incentives can, therefore, be expected to be very different. In contrast with political institutions, for instance, independent agencies highly value the autonomy they possess and will likely take actions that will limit the extent to which other domestic actors—legislators, lawyers, diplomats, etc.—intervene in their affairs. 38 As David Andrew Singer has noted, “political intervention, in its various guises, is the bane of a regulator’s existence. When politicians attempt to influence regulatory policy directly […] they threaten the agency’s autonomy and prestige.”39 “Regulators will,” therefore, “use all strategies at their disposal to minimize the possibility of intervention.”40  Informality helps them to do so—and, again, this is due to the domestic implications highlighted above. Since creating an informal organization does not require extensive ratification procedures, monitoring arrangements or budgetary allocations, they allow independent agencies to both engage in intergovernmental cooperation to reduce the externalities !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!37 Brewster 2004; Guzman 2008; Moravcsik 2000a. 38 Damro 2006b; Singer 2007; Bach 2010. 39 Singer 2007, p.22. 40 Ibid.  18   they face—which may, in turn, lead to political intervention if left unchecked41—and maintain a higher level of control over their affairs than would otherwise be possible. The domestic implications of informal intergovernmental organizations are, therefore, a key asset for actors leading independent agencies.  That said, political institutions and independent agencies cannot achieve their preferred outcomes at all times. The former are unlikely to be able to pursue a formal approach when they face significant domestic constraints. Creating a formal organization requires a larger shift from the prevailing domestic “status quo,” since the domestic requirements for signing an international treaty are much more onerous. The prospects for formal cooperation are therefore likely to be heavily influenced by factors that make it more or less difficult for policymakers to secure the domestic support necessary. When opposition to government treaty-making is high, for instance, formality becomes a less achievable objective. When this occurs, a political institution may then switch to an informal approach, which may satisfy some of the constituents in favor of international cooperation while circumventing those that oppose it. An informal organization is a second-best solution, but is likely viewed as better than the alternative: no organization at all.42 Independent agencies, by contrast, are unlikely to achieve their first preference when their actions become politicized. This may occur under several different conditions, but it most likely happens when an independent agency is seen to be failing or underperforming in some important way by other domestic political actors. 43 An issue that arises might also simply be too important !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!41 On this point, see Damro 2004; Singer 2007. 42 It is, also, theoretically quite easy to “scale-up” an informal organization at a later point in time, see Abbott and Snidal 2004. And, in fact, we do see a fairly significant number of informal organizations being reconstituted as formal organizations at a later point in time.   43 See Damro 2006b; Singer 2007.  19   or controversial to be left up to bureaucrats alone, particularly if there are significant domestic distributional consequences.44 Politicians may then demand a say. Either way, when this happens more political actors and institutions may threaten to intervene or actively intervene in the affairs of an independent agency. If so, then the agency will lose its ability to control foreign policymaking, and the probability that a more formal approach to international cooperation will be pursued can be expected to increase correspondingly.   This is, of course, a “stripped down” model of what goes on within a single domestic political arena. The real world is, no doubt, more complicated. However, the model still admits a great deal of complexity: the domestic preference formation process described above can play out in different ways depending on the state of politics and the institutional structures prevailing within relevant states at the time an intergovernmental organization is to be established. Indeed, the model suggests that it is entirely possible that the preferences of states may conflict, even when they are dealing with the same issue. The second step in the model seeks to explain how such conflicts are resolved. It starts by taking state preferences as given and shows how they “aggregate” to produce either a formal or informal organization using a simple bargaining model. In the model, states bargain over a continuum that ranges from mostly informal to mostly formal. If, after the preference formation process plays out (the first “stage” described above) the preferences of states converge on one side of the continuum, then, of course, we should expect the commonly preferred outcome to arise. If two states both want an organization to be highly formal in nature, then neither needs to hold up cooperation or coerce or cajole others to achieve its first preference. They may disagree about other dimensions of institutional design and engage !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!44 See Bach 2010.  20   in hard bargaining over those issues. But, when it comes to organizational form, there is little disagreement. Since neither can do better by coercing the other or opting out, the commonly preferred outcome is in equilibrium.  By contrast, when state preferences diverge—that is, when one state prefers informality and the other formality—bargaining power is likely to be a much more important consideration. In this case, states are assumed to agree that an intergovernmental organization of some kind is desirable but they disagree over the form it should take. When this occurs, an organization may be unlikely to materialize, since the preferences of states may be so far apart that there is no scope for agreement. But, more importantly, if they do at least agree that an organization of some kind is necessary, neither can attain its preferred outcome without compelling the other in some way. The final outcome is, therefore, likely to reflect the preference of the state with the greatest bargaining power. Here, in line with classic theories of bargaining, a state is expected to have bargaining power if its preference is closer to the prevailing status quo or “reversion point.”45 When this is the case, an actor is more willing to say “no” to any agreement it does not like and can more credibly hold out or “veto” cooperation if it does not get what it wants. Its more-dependent bargaining partner is then more likely to defer to a point closer to the more powerful state’s preference in order to secure an agreement of some kind. The final outcome should, then, reflect the preference of the state with greater veto power. However, bargaining power is not simply a matter of whose preference is closer to the original status quo; other actions are possible that can change the outcome quite dramatically. For instance, if another state can threaten to “go it alone” with others or take actions—applying sanctions or providing “side benefits”—that have !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!45 Nash 1950; Rubinstein 1982; Muthoo 1999; Richards 1999.   21   the effect of moving the status quo, then an opponent is more likely to defer to a point closer to its preferred outcome—even if its initial preference was closer to the status quo.46 In this case, by contrast, the final outcome will reflect the preference of the state with the greater ability to go-it-alone, sanction or buy-off others, all of which may be related to factors “outside” the bargaining scenario itself, such as the size of a state’s market. This means, all else being equal, that states preferring informality may have a bargaining advantage, since their preference is typically much closer to the prevailing status quo; but the theory also allows for other possibilities if a state’s bargaining partners can take actions that remove the initial status quo or “reversion point” as an option. Things, in other words, are not always equal, particularly if states aren’t.  1.3 Alternative Explanations  In this study, I refer to the two-step model outlined above as the “liberal” theory of organizational form, since it suggests that formality is primarily a product of domestic politics, institutions and state power—variables that liberal theories of IR have traditionally emphasized.47 As indicated above, it is quite different from the rational design arguments prevailing in the literature at present. These alternatives generally leave domestic politics and power out of the causal equation. They instead hypothesize that, when states create institutions of various kinds, they fashion them to resolve the particular problems they face in the most efficient way. The character of a cooperation problem is itself the most important causal factor. The process they envision can be thought of as akin to the construction of a building, with states !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!46 Gruber 2000; Moe 2005. 47 On the liberal tradition of IR, see Zacher and Matthew 1995; Slaughter 1995; Moravcsik 1997.   22   acting as “architects.”48 The structure that an architect will design will be different if the intention is to create a residence rather than a place of work. If there are earthquakes or high winds in the location it is built, then certain reinforcements will be put in place to ensure stability. If the local weather is generally colder, insulation and heating will be added. Similarly, rational design theorists argue that when states attempt to cooperate with one another they face a range of problems that differ in various ways. These problems make certain institutional designs more appropriate than others. They then argue that informal organizations have certain properties that contrast with the properties of their formal counterparts and make one or the other more or less advantageous for addressing particular problems.49 The specific aspects of institutional design they emphasize—in contrast, with the theory just presented—are the “international properties” described above. Their general hypothesis is, then, that states choose to create informal organizations when doing so is the “best” way of solving the particular kind of cooperation problem they confront in a given context.   Rational design theories therefore suggest that we can predict what kind of organization will be created if we understand the particular features of the problem that create incentives for states to choose different institutional designs. To do so, they develop a number of specific hypotheses about what these features are. An informal organization is expected to be preferred, for example, when there is high uncertainty about future states of the world, since the greater flexibility that an informal organization offers means it can be adjusted more easily in response to new contingencies. In contrast, a formal organization should be expected when there is significant scope for opportunism—when, for example, a problem has the structure of a Prisoner’s !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!48!This helpful metaphor comes from Kleine 2013.  49 Eeilstrup-Sangiovanni 2009; Bradford 2011; Vabulas and Snidal 2013.  23   Dilemma—since the greater operational independence that formality can afford may offer certain advantages: a more autonomous organization might be better at providing credible information about the behaviour of other states or might even be able to sanction states that engage in opportunistic behaviour to maintain a cooperative equilibrium. By comparison, when the scope for opportunism is low, such mechanisms are not necessary and an informal organization should be preferred. Each of these hypotheses is discussed in more detail in Chapters 3 and 4. For now, it is simply important to note the general thrust of the functionalist argument: that, on average, states are believed to create organizations that are “matched” with the problems they are supposed to solve, and that we should therefore expect to find a strong correlation between particular aspects of the problems states confront and different institutional designs.  This study takes these arguments seriously and even attempts to test their observable implications more thoroughly than has been done thus far. The reason for this is simple. When designing an intergovernmental organization, states no doubt try to create institutions that are well suited to the problems they wish to address. It would be surprising if they did not, and there is certainly evidence that this occurs. Rational design arguments, this study will show, have some validity. At the same time, there are good reasons to believe that they are incomplete and do not adequately explain the emergence of informal organizations on their own. Some of these reasons have been briefly mentioned above, and further evidence is provided throughout the study to support this claim. The liberal theory is advanced as a deliberate response to the theoretical and empirical inadequacies of rational design arguments and leads to quite different conclusions about the emergence, rise, and, ultimately, desirability of informal organizations, as will be discussed next. Further, the evidence offered in the study shows that the factors that the liberal  24   theory emphasizes appear to be the most important determinants of organizational form. Nevertheless, it is important to keep in mind that while I go to great extents to demonstrate the utility of the liberal theory relative to rational design arguments, it is in fact complementary to such accounts. The theory adds nuance, explains additional variation and has different implications, but it is not, strictly speaking, envisioned as an “alternative.” Although the study may adopt such language at times, the real world is complex and motivations are diverse, and while it is essential for any theory to demonstrate its value relative to other competitors, it is important to also admit a degree of analytical eclecticism as well in recognition of this fact.50   1.4 What is at Stake?  Some may be inclined to ask why any of this matters. Why should we care if rational design theorists are right, or if the liberal theory presented in this study offers a more compelling explanation? One answer is that it matters because we value the truth and want to know how the world really works. For many social scientists and, especially, scholars of IR, this may be enough. As individuals who are curious about IR and wish to better understand evolving patterns global governance, we care about the validity of our conjectures and about the puzzle itself. There are, also, important theoretical stakes within the field of IR, since, in some respects, this study can be construed as a broader challenge to “pure” theories of rational design. The rational design project represented an important innovation in IR when it first appeared, and the specific theories it subsequently spawned have achieved some prominence in the discipline over the past 15 years.51 A number of important methodological innovations have followed as well, which this !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!50 Sil and Katzenstein 2010b; Sil and Katzenstein 2010a; Lake 2013. 51 Koremenos, Lipson, and Snidal 2001; Koremenos 2013b.  25   study readily exploits. Researchers in this tradition have, therefore, provided a number of important insights and advanced the field in tangible ways. Further, the approach has been particularly useful for fostering links between the fields of IR and IL—areas of research that have been studied separately for far too long, but which contain important insights that each other can learn from.52 Yet they have also been criticized for neglecting domestic politics, distributional conflict and state power as important variables, as noted earlier. Part of the value of this study is, therefore, to show the extent to which this is true. By deliberately evaluating the rational design and liberal explanations side-by-side it demonstrates that we cannot fully understand what drives institutional design without taking these factors into account.   For others who may wonder what the practical value of investigating these questions is, there is a second response: it matters because the answer has important policy implications. As mentioned, there is a vigorous debate about the merits of the various institutions we rely on to govern global issues, and many believe that informal organizations hold some promise for providing “good enough” governance, especially since so many formal organizations seem to be failing us. If true, the rational design approach would seem to support those who are more optimistic about these new institutions. If states do in fact design institutions that are “matched” with the problems they face, then this would suggest that the rise of informal organizations has been a rational response to the changing nature of the issues we face and that these new bodies are well positioned to supply effective cross-border governance. If, on the other hand, the factors that the liberal theory identifies have been the primary determinants of organizational form, then more pessimism about these institutions may be warranted. This is because the liberal theory !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!52 Abbott 1989; Abbott and Snidal 2013; Koremenos 2013a.  26   suggests that informal organizations are often chosen for less salutary reasons: because they help independent bureaucrats in powerful states to maintain autonomy over their international activities, and because they help politicians to outflank domestic opponents of their foreign policies. If such factors are behind their rise, then there would be no necessary relationship between the kind of problem states face and the form of an intergovernmental organization. Many may then be poorly “matched” with the pressing global challenges we face today. While it would, of course, be possible that some informal organizations are well suited to the particular issue they are supposed to address, nothing would guarantee this outcome. Ultimately, therefore, if the liberal theory is correct, we would have to be much more skeptical about the more optimistic claims made on the behalf of many informal organizations.  This is, essentially, what I claim in the Conclusion of this study. The evidence presented in this study indicates that these domestic drivers have played the most important role in determining organizational form and that this has, at least in some cases, led states to create institutions that are poorly “matched” with the problems they are supposed to solve. In fact, in the Conclusion, I also provide some evidence showing that informal organizations are more likely to be poorly matched with the underlying issues they aim to address than their formal counterparts. I argue that there is indeed, then, an urgent need to rethink the desirability of these organizations and to develop our understanding of the conditions under which they are able to offer a viable solution to the complex issues that we face today. Further, given that many are likely to be so dysfunctional, it is essential for us to identify the problematic cases that may be most in need of reform. Towards this end, I provide in the Conclusion a new way of systematically evaluating the extent to which organizations may be matched or mismatched with the underlying problems  27   they are designed to resolve. The approach that I present is preliminary and mainly intended to be illustrative, but could potentially be developed into a new tool that would allow policymakers to look across a broad range of institutions and focus their attention on those that are most likely to be in need of change. In this way, the project not only calls our attention to the fact that at least some institutions may be dysfunctional, it also provides a way of shifting the current policy debate towards a more theoretically-informed variety of normative institutional analysis.   1.5 Methods and Limitations of this Study  In practice, any social scientific study faces a tension between two overarching goals: maximizing external validity and maximizing internal validity. Achieving external validity means that we conduct a study in such a way that it ensures that the causal inferences we draw from a given body of evidence can be generalized. Generalization refers, in turn, to the idea that any conclusions that we reach extend beyond the particular cases that we look at, and ideally, across a wide range of cases. A study that maximizes internal validity, by contrast, ensures that a theory explains particular cases well, typically on a micro-level. In an ideal world, a study would be able to achieve both of these goals simultaneously. Often, however, achieving one comes at the expense of the other.53 One way of attaining high external validity is to study a large number of cases. But, in order to do so, we must narrow our analytical focus to a few quantifiable indicators and forgo a detailed analysis of the causal process that leads to an outcome. This obviously does an injustice to the actual complexity of the world and increases the likelihood that our measures of key explanatory variables are inaccurate because they fail to take into account the context of a particular case. Further, with observational (as opposed to experimental) !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!53 Coppedge 1999.  28   quantitative methods, we can never completely eliminate the possibility that the relationships we find within a large sample are spurious—caused by some third unobserved factor, which a more detailed analysis might reveal. By contrast, if we chose to maximize internal validity by conducting a study that examines particular cases in considerable detail in order to take complexity into account and increase the accuracy of our measures this frequently comes at the price of limiting the number of cases we can realistically look at. The external validity of our study is thereby reduced.  There is, unfortunately, no sure way of resolving these tensions in any one study. But it is possible to strike a healthy balance by using a variety of analytical strategies. This is the approach taken here. In order to maximize external validity, the study undertakes a quantitative analysis of a sample of formal and informal organizations. At around 90 observations, this is still a fairly small sample by conventional standards. However, by testing the observable implications of the two main theories, the analysis helps to ensure that the explanation does indeed hold across a diverse range of cases. Once this is done, the study then “zooms in” on a few cases in order to ensure that the causal processes do indeed resemble those posited by the theory when observed up close. This more fine-grained qualitative analysis helps to ensure internal validity. In doing so, however, the study also makes an attempt to maximize the generalizability of the case study findings through careful case selection. This is accomplished by examining organizations that are “crucial cases” for the theories in question.54 Generally, there are two varieties of crucial cases: “most likely” cases and “least likely” cases. The former are those where a theory must be found to explain a case if we are to have confidence in it because the !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!54 Eckstein 1992. Also see Gerring 2007; Bennett 2008.  29   conditions are especially propitious, and where failure would significantly lower our confidence in a theory if it fails. The latter, by contrast, are cases where our confidence in a theory rises significantly if a theory can explain it well, since the conditions are not especially propitious, but, for this same reason, should not lead us to significantly lower our confidence in a theory if it does not. External validity is maximized in the qualitative analyses by examining cases that are “most likely” for rational design theories, and, at the same time, “least likely” cases for the liberal theory. Doing so, gives us significant analytical leverage. If we find that rational design predictions are not borne out, then our confidence in that theory should be reduced significantly; and, if we find that the liberal theory performs well, then our confidence in it should increase significantly—indeed, we should expect it to apply in cases beyond those examined.   The study therefore makes a significant effort to adopt analytical strategies that can make the most of a limited number of cases. However, it inevitably has a number of theoretical and empirical limitations, and it is worthwhile noting what these are from the start. First, the study has a somewhat narrow geographical scope. This is because it primarily investigates intergovernmental organizations created by relatively well-governed democratic states. The sample used in the quantitative analysis, for instance, mainly includes intergovernmental organizations created by OECD states. Organizations that were created with the involvement of non-democratic states are, for the most part, excluded from the analysis. Obviously, this limits to some degree the external validity of any conclusions we reach, since the organizations that are subjected to analysis are not representative of the overall population of intergovernmental organizations, many of which were created with the assistance of developing countries or were even created exclusively by them. Yet this is not as problematic as it first appears. For reasons  30   that will be explained in more detail at the end of Chapter 3, the liberal theory is primarily expected to apply, at present, to intergovernmental organizations created by relatively well-governed democratic states. Further, the case studies include organizations that have involved developing countries to a greater extent. Second, the study has a somewhat narrow temporal scope. This is because it primarily investigates intergovernmental organizations that have been created since the end of the Second World War. Due to data constraints, for instance, the sample used in the quantitative analysis only includes organizations that are created between 1950 and 2005. Further, the case studies mainly focus on organizations that were constituted after 1944. Again, this limitation is somewhat mitigated by the fact that the liberal theory is mainly expected to apply within this time period. However, the narrow temporal scope necessarily limits the extent to which we can generalize to earlier periods. In the future, as discussed further in Chapter 4 and in the Conclusion, efforts will have to be made to extend the theory and empirics beyond these limits in order to fully understand the rise of informal organizations. Yet the analysis undertaken here represents a necessary first step towards that goal. By showing that the theory explains the cases that it looks at well, it suggests that efforts to build upon the account presented here and more thorough testing will be rewarding.  1.6 Outline of the Study  With these limitations in mind, it is now helpful to provide a general outline of how this study will progress. The analysis begins in the next chapter (Chapter 2) with an effort to conceptualize and map the “universe” of formal and informal organizations. In the first half of the chapter, the distinction between formal and informal organizations is explained in considerable detail, and their unique international properties and domestic implications are reviewed. The second half of  31   the chapter then discusses problems with existing efforts to measure informal organizations, explains how I have operationalized the concept of an informal organization and outlines the data collection methods. It concludes by presenting descriptive statistics from the database that I have developed, showing the unprecedented rise in the number of informal organizations over the past 50-60 years. Chapter 3 focuses on theory. In the first part of the chapter, I provide readers with an overview of the logic behind rational design theories, tease out the specific predictions that rational design theorists have advanced and explain why they fall short, both theoretically and empirically. The chapter then presents the liberal theory advanced in this study in considerable detail. It begins by explaining the logic of the preference “formation” stage and outlining the different conditions in which we would expect states to prefer either a formal or informal organization. The second preference “aggregation” stage is then explained using a simple bargaining model, showing when we should expect a formal or informal organization to emerge. Finally, the chapter concludes by noting the scope conditions of the theory, which are expected to determine when it is likely to operate as expected.  The following chapters focus on testing the empirical predictions of the rational design and liberal theories. Chapter 4 presents the study’s quantitative analysis of the determinants of organizational form. It starts with an overview of the sample selection methodology and the sample itself. It then develops new measures of explanatory variables for testing the observable implications of the liberal and rational design theories. Finally, the chapter presents the results of the quantitative analysis, and conducts a number of additional tests in order to check the robustness of the results and to address potential selection and endogeneity problems. Overall, the analysis provides powerful empirical support for the study’s central thesis. Chapters 5 and 6  32   then conduct cases studies of individual organizations. Chapter 5 investigates two particularly prominent intergovernmental organizations created during and after the Second World War: the IMF (a formal organization) and the General Agreement on Tariffs and Trade (GATT, an informal organization). Chapter 6 investigates a much more recent informal organization: the International Competition Network (ICN). In each of these chapters, the study begins with a review of the case selection methodology and discusses probative value of each case. I then conduct “congruence” analyses that seek to evaluate the extent to which rational design conjectures are borne out, and find, generally, that the cases are not well explained by the variables rational design theorists believe to be important. This is followed by more detailed “within-case” analyses that attempt to test the extent to which the causal process leading to the emergence of each the organizations resembles that predicted by the liberal theory. Again, these analyses demonstrate that the variables identified by the liberal theory are powerful predictors of organizational form.  Finally, in Chapter 7, the various threads are drawn together. First, the chapter presents a brief summary of the main findings of the study. Second, the chapter reflects upon the broader theoretical, empirical and policy implications of the study. It discusses what the findings mean for the study of IR, and especially for theories of rational design and legalization. It also explains what the results suggest about the underlying causes of the extraordinary rise of informal organizations, and provides some evidence to demonstrate that the theory offers a plausible account of the broader changes that are afoot. And, finally, it explores what the study implies about the desirability of informal organizations, and how we can start to approach the more practical question of organizational reform. The last portion of the chapter is then devoted to  33   proposing a number of profitable avenues for future research. These include: studying the process of organizational change, which is not addressed to any significant extent in this study; extending the analysis to examine the determinants of organizational form across the developing world, and investigating other potential causal mechanisms, such as policy diffusion and institutional isomorphism, which are potentially important causal factors but are not extensively explored in this study; and, finally, revisiting, refining and extending our more general theories of intergovernmental organization by taking informal organizations into account.  34   Chapter 2. Conceptualizing and Mapping Organizational Forms   2.1 Introduction  The overall aim of this study is to understand why states create informal organizations and what this means for our ability to solve global problems. In order to do so, however, we must first be clear about what they are, how we can know an informal organization when we see one, and what general organizational patterns we actually see in the world. This is important because informal organizations can take a wide variety of forms, and it is often difficult to understand what unifies them as a group. Indeed, until recently, we have not even had good terms for speaking and thinking about them. Largely for this same reason, we have only limited data that can allow us to gauge broader trends. Most research has focused on case studies of single organizations; few efforts have been made to take a more synoptic view. Conceptualizing and mapping the universe of informal organizations is therefore an essential first step for any careful analysis of their origins and impact. Unless we are able to define them, understand what they have in common, and clearly identify a set of informal organizations in the world that we can subject to comparative analysis, it will be impossible to develop meaningful causal explanations.   Towards this end, the present chapter has two main goals. First, I describe at a general level the defining features, properties and implications of formal and informal organizations. I do so by initially describing formal organizations. This is because readers are likely to be more familiar with formal organizations and because there is already some consensus among IR and IL scholars about their core characteristics. This then allows us to more clearly define and elucidate the main features of informal organizations, since we will be able see precisely how they depart  35   from the formal model. In doing so, I draw upon the existing literature in IR and IL on informal intergovernmental organizations, especially the pioneering work of Jan Klabbers, Felicity Vabulas and Duncan Snidal, and I illustrate important points with real world examples.55 Building on this conceptual groundwork, the second part of the chapter then discusses and presents a new database of informal organizations. I begin by briefly explaining how the database was constructed and how it compares with previous and related efforts. I then present a number of descriptive statistics derived from the dataset that help us visualize the universe of cases that a good theory of organizational form needs to explain.   2.2 The Key Features of Formal and Informal Organizations  When we think about “international relations” as a discipline and as a practical activity, intergovernmental organizations are among the most immediate entities that come to mind. Typically, however, when we think about them we tend to think of formal intergovernmental organizations; that is, we think of organizations like the United Nations (UN), WTO, World Bank, or IMF. The reason for this is simple: these organizations command our attention. Hardly a day goes by without an important decision being made by the UN Security Council, a critical dispute settlement ruling by the WTO, or a major loan or surveillance report by the World Bank or IMF. But these kinds of formal organizations are only the most prominent examples of a much broader phenomenon. There are in fact hundreds of formal organizations active across a wide range of issue areas.56 Though most operate in relative obscurity and rarely make the news !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!55 Klabbers 2001a; Vabulas and Snidal 2013.  56 Shanks, Jacobson, and Kaplan 1996; Pevehouse, Nordstrom, and Warnke 2004; Volgy et al. 2008.  36   (except perhaps when they close their doors!57), today there is hardly an activity that is not in some way regulated by one formal organization or another.58   Looking across this landscape, it is clear that formal organizations vary considerably in terms of inter alia their structure, size, and purpose.59 This variation makes it inherently challenging to define them precisely. In fact, it is probably impossible to do so, since formal organizations are fundamentally social constructs.60 Nevertheless, there are at least three defining characteristics that are fairly widely agreed upon in practice.61  First, it is generally agreed that formal organizations are created by sovereign states; or, more precisely, by individuals authorized to act on the behalf of sovereign states. This differentiates them in an important way from other kinds of organizations that operate across borders, such as multinational corporations (MNCs) and international non-governmental organizations (INGOs), which are mainly created by non-state actors. The criterion also distinguishes formal organizations from bodies created by other kinds of public actors, such as sub-state actors (regional governments, municipalities, etc.), or through the legislative acts of other intergovernmental organizations.62 Of course, there are occasionally exceptions to this rule. Some organizations include non-state actors or intergovernmental organizations as members and founders. The governing body of the International Labour !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!57 The Economist 2010. 58 For a review of the literature on intergovernmental organizations in the fields of IR and IL, see Abbott 2015. 59 See, for example, the exchaustive comparative analysis by Schermers and Blokker 2003. 60 See the discussion on this point in Klabbers 2009, p. 6-12. It is worth recalling Kant, “Nothing entirely straight can ever be fashioned from the crooked wood of which humankind is made,” see Kant 2009, p.16. 61 Compare Schermers and Blokker 2003; Alvarez 2005; Amerasinghe 2005; Brolmann 2007; Klabbers 2009. Also, see Pevehouse, Nordstrom, and Warnke 2004. 62 Both varieties are discussed further below. See Shanks, Jacobson, and Kaplan 1996; Pevehouse, Nordstrom, and Warnke 2004.  37   Organization (ILO) has a tripartite structure that includes representation from states, workers and employers, and the European Community (EC) was a founding member of the WTO. Some organizations have also been founded by non-state actors and subsequently “internationalized” by states.63 Thus, on its own, the criterion does not offer a hard and fast rule, but for an intergovernmental organization to be regarded as such states should generally be the primary actors involved in its creation.  Second, formal organizations usually have bodies that are functionally separate from their members, or with a “distinct will.” Usually this would imply a body of some kind that is delegated a degree of operational independence and responsibility for undertaking certain tasks. Often, these can be quite large bureaucracies with offices spread out across the world. The World Bank, for instance, employs around 12,000 staff and has offices in over 100 countries. Many of have been delegated considerable independence and make impactful decisions on a day-to-day basis.64 Other organizations, like the European Court of Justice (ECJ), seem to have escaped the control of states altogether.65 Again, there are important exceptions. One of the oldest formal organizations still operating today—the Commission for the Navigation of the Rhine, established in 1815—employs only 20 people. Many others, such as the secretariat of the UN Framework Convention on Climate Change (UNFCCC) seem to possess little scope for autonomous decision-making. 66  As with many treaty secretariats, its “function” is primarily to support interstate negotiations. Yet, even here, such secretariats have a physical headquarters, exist !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!63 Klabbers 2009. The International Development Law Organization is a good example of a non-profit association that was subsequently reconstituted by states as a formal intergovernmental organization. 64 Haftel and Thompson 2006; Hawkins et al. 2006. 65 Alter 1998. 66 Busch 2009.  38   independently of any member and they do make some decisions in an autonomous manner—even if they do so within the narrow confines set down by states.67 Autonomy may be a matter of degree, but it is autonomy nonetheless.  Finally, but perhaps most importantly, formal organizations are constituted by agreements under international law; that is, they are constituted by international treaty.68 Such agreements establish the objectives of an organization, rules of procedure, the rights and obligations of members, and so on. But, as treaties, they are agreements governed by a body of formal and customary international (as well as domestic) law, such as the 1969 Vienna Convention on the Law of Treaties, which specifies how agreements establishing formal legal obligations must be concluded, interpreted, amended, and, if necessary, terminated. 69  Again, not all intergovernmental organizations are established in quite this way. As will be discussed further below, some intergovernmental organizations are established informally, without recourse to international treaty. But many others, as mentioned above, have also been established by the formal legislative acts of other intergovernmental organizations. The European Agency for Safety and Health at Work was constituted by several EC Council regulations, while the UN Development Programme (UNDP) was created by a resolution of the UN General Assembly. One study has even found that such “organizational progeny” constitute the vast majority of the intergovernmental organizations active today. 70  However, in IR such organizations—often referred to as “emanations”—are regarded as a fundamentally different kind of body, given their !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!67 Churchill and Ulfstein 2000; Beach 2004. 68 For discussion of the concept of treaty in international law, see Klabbers 1996 and Aust 2000. 69 Klabbers 1996; Aust 2000; Shaw 2008. 70 Shanks, Jacobson, and Kaplan 1996. The term “organizational progeny” comes from Johnson 2014.  39   unique mode of establishment.71 Here, and in keeping with this usage, I reserve the use of the term “formal organization” or “formal intergovernmental organization” to those bodies established solely through international treaty, and not via a legislative act. Now, having outlined the central features of formal organizations, we are able to see more clearly the ways in which informal organizations are both similar to and different from their more formal counterparts. In terms of similarities, the most obvious feature that both share is the first: informal organizations are the progeny of states. While there is perhaps more variation in terms of the kinds of actors involved in creating them, they remain fundamentally different from private bodies, such as MNCs and INGOs. It is also worth reemphasizing here that they are different from organizations created by sub-state public actors, such as municipalities or regional governments, which are increasingly common forms of informal cross-border cooperation.72 Examples of such organizations include the C40—Cities for Climate Leadership or the World Mayors Council on Climate Change. Because these organizations share some features with informal organizations—they both involve public actors and are constituted outside of international law—they might create some conceptual confusion. Yet they are not immediately analogous to intergovernmental organizations of any kind (formal or informal), since, except in certain exceptional circumstances, the actors that create them rarely have the ability to act on the behalf of a state.73 Although there are certainly valuable comparisons, these organizations should for now be regarded as essentially different creatures. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!71 Singer and Wallace 1970; Pevehouse, Nordstrom, and Warnke 2004.  72 Lee 2013; Bulkeley et al. 2014. 73 See the discussion of this point in Aust 2000, chapter 4.  40   Second—and here we begin to see some differences—informal organizations are less likely to have bodies that are clearly separate from their members. In general, they tend to be very small in size—only a few staff—and the sprawling bureaucracies that are often found in the set of formal organizations are almost entirely absent. Even the UNFCCC Secretariat with only 200 or so staff would have to be considered large by the standards of most informal organizations. IOSCO—arguably one of the larger informal organizations—employs fewer than 30 people, including a number of purely administrative staff. However, some have no dedicated staff or headquarters at all. 74  A notable feature of many informal organizations is their “virtual” character, whereby secretarial services are supplied by a single state, either on a permanent or rotating basis. Both the G7/8 and G20, for example, lack independent secretariats; secretarial services are instead provided for the group by the states that “host” the rotating annual summits. Similarly, the French Ministry of Finance provides secretarial services for the Paris Club. These kinds of informal organizations, which I refer as “virtual organizations” since they have no independent physical existence or permanent staff, are clearly different from formal organizations in this respect, but nevertheless possess members, a persistent institutional structure and identity, and rules procedure that make them clearly identifiable as organizations.75 Third—and this is where formal and informal organizations differ most significantly—informal organizations are constituted by agreements outside of international law. Instead of a treaty, they !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!74 Vabulas and Snidal in fact define informal organizations partly in terms the absence of a secretariat, see Vabulas and Snidal 2013. I differ from them on this point by emphasizing that delegation is a matter of degree, not kind. Informal organizations tend to have smaller secretariats (or none) while formal organizations have larger ones. 75 For an excellent general discussion of the nature of organizations and of organization theory from the perspective of a sociologist, see Scott 2001.  41   are typically founded by “non-binding” agreements, such as a MOU.76  Though they contain most of the same constitutive provisions as a treaty establishing a formal organization (that is, they lay out goals, a structure, rules of procedure, and so on) they are officially regarded as an “arrangement” or “understanding” between states, not international law.77 Occasionally, this is signified by the fact that such constitutive agreements refrain from language implying that signatories intend to generate international legal obligations; they use “should” instead of “shall,” for example.78 But most often this fact is made explicit in the constituent agreement itself through some statement explaining that any commitments made by parties are of a precatory rather than legally obligatory nature. An especially clear example can be found in Article 4 of the Terms of Reference of the APG. The APG describes itself as a “multilateral organization” that is “established by agreement among its members and is autonomous.” Yet, at the same time, it states that the organization is “voluntary and co-operative in nature,” and its authority “does not derive from an international treaty.”79 Some informal organizations are also constituted by “oral agreements” between policymakers. This occurs when there is no written constitutive agreement !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!76 For a useful discussion of MOUs and how they differ from treaties, see Aust 2000, chapters 2-3. Also see Schachter 1977; Aust 1986; and Guzman and Meyer 2010. 77 Other names used in the literature include “declaration,” “political commitment,” “de facto agreement,”  “non-legal agreement” or “gentlemen’s agreement.” Note that, as mentioned earlier, despite the variety of titles they go by, I generally refer to all such agreements in this study as “non-binding agreements” or “MOUs.” 78 Aust 2000, chapter 3. In fact, the intention to establish or not establish legally binding commitments is the critical factor determining whether an agreement constitutes international law, or not. Determining intentions is, of course, challenging, since they cannot be directly observed, but they can nevertheless be inferred from both the text of an agreement and the practices and actions of those that negotiate it. Further discussion of how we can infer whether a constitutive agreement is intended to be legally binding or not can be found later in the chapter.  79 Article 4, APG Terms of Reference 2012.  42   but a set of institutions and practices nevertheless develops that gives rise to an intergovernmental organization, as has been the case with the G7/8.80  2.2.1 The Properties and Implications of Formal and Informal Organizations  Overall, then, there are a number of important similarities and differences between formal and informal organizations. But these are not simply differences of style; these are differences of substance that have a range of distinct, practical consequences. Some of these consequences, as noted in the Introduction, can be characterized as the “international properties” of formal and informal organizations in that they concern how an intergovernmental organization will tend to operate and evolve over time. These are the properties that the existing literature has general emphasized, and I outline them below. However, creating and working through formal and informal organizations can also have distinct “domestic implications” as well—that is, creating such organizations will tend to affect states in different ways. These characteristics of formal and informal organizations have, by comparison, been underemphasized by the existing literature, and I elaborate on them in detail. Ultimately, this is important because, as we will see in the next chapter, some of these implications may be beneficial in some contexts and not in others, and they will tend to benefit some actors at the expense of others. Further, I argue that understanding this fact is important for explaining why formal and informal intergovernmental organizations arise. Thus, in this section, I consider both the properties and implications of formal and informal organizations, and begin with a discussion of their “international properties.” !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!80 Dobson 2006; Hajnal 2007.  43   First, formal and informal organizations differ in terms of their flexibility—how easy it is to change or adapt an organization once one has been created. Here, informal organizations have the upper hand, since there are generally fewer impediments to doing so.81 There are two reasons for this. One, in comparison with formal treaties, non-binding agreements are much easier to change because there are fewer requirements for doing so under international law and the costs of renegotiating a non-binding agreement are lower.82 Although treaties may vary in terms of how flexible they are, a treaty will typically specify precise procedures for adjusting or amending a constitutive agreement, and these criteria must be followed or else a state or group of states may violate international law and suffer reputational consequences. Typically, the criteria for adjusting a constitutive agreement—in terms of the number of votes needed, for instance—may also be quite high, and given that any changes may result in significant domestic legal effects states may also be more reluctant to agree to amendments. A non-binding agreement, by contrast, can be adjusted much more easily, since there are fewer requirements. Given the relatively limited legal effects of such treaties, states may be more willing to accept changes. Further, if they do not agree with any changes, certain members may simply decide to follow a different practice, or even “exit” an agreement, without suffering any reputation consequences. They will not be found to be violating international law by doing so.83 Two, because formal organizations generally involve a higher degree of delegation, resistance by entrenched bureaucracies may also make change more difficult, since the larger, more complex bureaucracies associated with formal organizations can be expected to be more capable of thwarting reforms. The recurring failures to reform the World Bank as a result of bureaucratic !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!81 Koremenos 2005; Vabulas and Snidal 2013. 82 Abbott and Snidal 2000. 83 Aust 2000; Guzman and Meyer 2010.   44   opposition offer a particularly poignant example of the challenges states often face.84 Informal organizations, which involve only limited levels of delegation, face no such obstacles. Second, informal organizations are generally believed to differ in terms of their agility—how quickly they can arrive at decisions and take actions. Working through an informal organization, in comparison with formal organizations, generally allows officials to avoid a range of diplomatic formalities that can sometimes slow down and stifle interactions between officials. There is, for instance, much less need for interactions between officials to be mediated by diplomats and heads of state when commitments are framed as “voluntary” and “non-binding,” posing no threat to a state’s sovereignty.85 As a result, many believe that it is easier for officials to form cordial relationships with their counterparts abroad, which can facilitate speedy decision-making.86 The confidentiality that an informal organization affords—discussed next—also means that officials can engage in franker and freer dialogue than they otherwise would, which again may mean that officials can arrive at decisions swiftly and without public posturing.87 Finally, because the terms of a constitutive agreement are also non-binding, it is also possible for members of an informal organization to take actions even if others disagree.88 In a formal organization, if a state or group of states decided to oppose a particular action, acting as a !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!84 Weaver 2008. 85 Aust 1986; Zaring 1998. It is for this reason that many informal organizations will therefore simply list components of states as their “members.” The members of IOSCO are, as its name suggestions, mainly national securities commissions, and members of the FSB are primarily central banks and ministries of finance. The members of formal organizations are almost always “states,” and the signatories of their constituent agreements are usually diplomats or heads of state. 86 Hardt 2014. 87 Puetter 2006. 88 Eeilstrup-Sangiovanni 2009.  45   “spoiler,” others may be unable to move forward independently without violating the terms of an agreement. However, within an informal organization, a “coalition of the willing” may be able to take action in this kind of situation, since they will be less worried about violating any legal obligations by taking independent action. An informal organization may, therefore, be able to act more swiftly because of the inherent malleability of its basic rules. The third way in which formal and informal organizations differ is in terms of their confidentiality—how secretive an organization may be. Since they are not constituted by international treaty, informal organizations can generally be much less “public” than their formal counterparts. Any treaty constituting a formal organization must, for example, be registered with the UN under Article 102 of the UN Charter, and published in the United Nations Treaty Series, a publicly accessible compendium of international agreements.89 Thus, anyone with an Internet connection can easily access a range of information about nearly any formal organization. With informal organizations this is much less true, since there is no obligation under international or domestic law for their constitutive agreements (or any other information, for that matter) to be made public. 90 It is, essentially, up to the members of each organization to decide what information to disclose. Some, such as the G7/8, do not even have websites or written agreements setting out their goals and how they work. It may also be possible for an informal organization to be more secretive because the body has only a limited institutional presence. In the case of a “virtual organization,” for instance, there is no office or staff that can be contacted !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!89 Domestic requirements also usually mandate the publication of all legally binding agreements that a state enters into. In the US, the Department of State keeps a record of all treaties and executive agreements that the country enters into. In the UK, the Foreign and Commonwealth Office maintains a freely accessible database of all treaties that the government has entered into since the 1830s. Typically, non-binding agreements do not have to be entered into such databases, see Garcia 2015. 90 Aust 1986; Aust 2000.  46   to attain information about an organization. One cannot simply phone the G20 in the way one can call the IMF or a UN office. For the same reason, informal organizations are less likely to be susceptible to the kinds of pressure that can be placed on a formal organization to gain access to its deliberations. It is much more difficult to protest or lobby an organization that has no physical existence, or which has meetings that can be moved around the globe as needed. The terms of access will, therefore, be solely at the discretion of the organization itself. Fourth, these organizations differ in terms of their independence from member states. Here, formal organizations clearly have the upper hand. Formal organizations generally have bodies that are functionally separate from their members and may be granted a high degree of operational independence that is enshrined in the formal rules of an organization. A simple example of this is the longstanding practice of granting officials of formal organizations certain privileges and immunities under international law that are not normally afforded to other kinds of actors, and which may help to preserve their independence and capabilities over time.91 However, organizations may also be granted special constitutional protections that explicitly limit the extent to which states can intervene in their activities. This is the case, for instance, with judges in various international courts who are seated for defined terms and cannot be removed once in office or pressured by states.92 Informal organizations, by contrast, are granted few of these powers. At the extreme—in the case of a virtual organization, for instance—no delegation occurs at all, since an organization has no independent secretariat. In other cases, as with the FSB or the Paris Club, the officials comprising an organization’s secretariat are simply state officials seconded for a brief period of time, or secretarial services are provided by one of the !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!91 See Klabbers 2001a. 92 Alter 2014.  47   organization’s member states. Finally, in cases where delegation to an independent secretariat does occur, as is the case with IOSCO or APEC, the secretariat’s powers and scope for autonomous action are highly constrained. Fifth, formal and informal organizations differ in terms of their scale and scope—how large they can be. Many formal organizations, as noted above, have huge numbers of staff that can take on highly complex tasks and provide a much larger range of services to states—from administration and arbitration to peace-keeping and policing. The European Commission, for example, is divided in to 28 Directorate-Generals—akin to government departments or ministries—that each comprise hundreds, or even thousands, of highly specialized staff. The largest, the Directorate-General of Translation has over 2000 employees. The United Nations Secretariat employs roughly 8,000 people, spread across 20 specialized Departments. In general, the capabilities of such an organization are therefore very high. Informal organizations, by contrast, are much smaller and typically provide a much narrow range of services to states. Some of the largest informal organizations, such as APEC or IOSCO, employ around 30-40 staff, at most; most others employ only a handful, or, as in the case of a virtual organization like the G20, none at all.  Further, the role of their staff seems to be much more basic. The tasks that they are asked to perform are, for the most part, of a purely administrative nature—mainly, facilitating discussion and negotiation and providing basic kinds of information. The capabilities of informal organizations are, therefore, very low. Now consider the domestic implications of organizational form, which can be equally, if not more significant. First, because they are constituted by international treaty, creating a formal organization typically initiates a quite intrusive, complex and demanding domestic political  48   process. Drafting an international treaty, for instance, typically requires considerable input from lawyers and diplomats in order to ensure that the agreement is precisely worded, follows the standard process, and is balanced against broader foreign policy goals.93 In the United Kingdom (UK), for instance, any department or agency that has responsibility for an issue may propose to enter in to and lead negotiations on that issue with another country.94 However, in doing so, it must consult with and seek advice from the Foreign and Commonwealth Office (which has primary authority over treaty-making), including its Legal Advisors, on the contents and form of the agreement.95 Treaties must also, typically, be drafted in consultation with a legislature, which can often have considerable influence. To take an example, again, from the UK, the Kingdom Act requires a list of treaties being negotiated to be submitted to Parliament so that they can be discussed. Further, once an agreement has been concluded, it must be submitted to Parliament for a period of 21 days in accordance with the “Ponsonby Rule” before the agreement can be ratified. Even then, in many cases, treaties require implementing legislation for them to fully come into effect, which means that departments rarely proceed in negotiations without first attaining the requisite legislation. These procedures are relatively similar across many Commonwealth states, where Parliamentary systems generally prevail.96 In Presidential systems, of course, legislatures are also very influential—sometimes even more so. In the US, international treaties—if they are regarded as “treaties” under domestic law97—must be written in !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!93 Aust 1986; Martin 2000; Hollis, Blakeslee, and Ederington 2005. 94 Sinclair, Dickson, and Maciver 2005, p.732. 95 Ibid. 96 See Hollis 2005 for a summary review of different national approaches to treaty-making, and Hollis, Blakeslee, and Ederington 2005 for detailed reviews of these different national approaches. 97 If an international treaty is considered an executive agreement—which does not require ratification by the Senate—then according to the Case Act, the agreement must at the very least be reported to Congress within sixty days of the agreement coming into force. Such agreements,  49   consultation with the Senate and ultimately ratified by two-thirds Senate majority before they can come into force. The latter requirement, especially, presents a significant impediment to the creation of formal organizations, as ratification of their constituent agreements requires congruent interests between the Senate and executive, as well as significant expenditure of political capital by the latter.98 Overall, the point is that, in most political systems, creating intergovernmental organizations that are constituted by international treaty, i.e. a formal organization, typically requires greater input from a host of domestic actors and opens up avenues for various legal or political challenges by them that may delay, complicate or even stymie cooperation under certain circumstances.  Informal organizations, by contrast, do not confront these problems. If an agreement does not bind a state under international law—if, for instance, all one aims to do is “network” with their homologues abroad, or create a “virtual” forum for creating other “soft law” agreements—then there is less need to consult with a cabinet, lawyers or diplomats.99 Such agreements do not need to be stated in precise legal language, can be concluded without diplomatic ceremony and, since they are inherently more flexible, do not need to be so carefully balanced against other goals.100 In the United Kingdom, Canada and France, for instance, non-binding MOUs and arrangements administratifs may be concluded solely by the government agencies involved; the FCO or Ministry of Foreign Affairs need not be consulted.101 Equally, it is typically not necessary for !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!of which there are several types, are quite common but remain controversial. Further, despite the lack of a ratification requirement, the Senate is often consulted as necessary and able to shape such agreements to a significant extent, see Martin 2000.  98 Kelley and Pevehouse 2015a. 99 Klabbers 2001a. 100 Aust 1986. 101 Hollis 2005, p.22.  50   such agreements to be disclosed to or ratified by a legislature.102 In the US, for example, according to the Case Act and various State Department regulations, it is generally not necessary to consult or even inform Congress about agreements that are not intended to be “legally binding.” 103  Of course, domestic requirements related to the conclusion of non-binding agreements vary quite considerably across states, but it can safely be said that the process of concluding them is generally much less demanding and intrusive than the equivalent process for more formal ones.104 Organizations constituted by informal agreements can, therefore, be created with less input from other domestic actors, and it should be harder for many actors within the state to control, delay or veto a state’s participation in them.105  Second, since formal organizations typically have bodies that are functionally separate from their members and are, generally, more “public” than their informal counterparts, much greater effort will usually be expended monitoring participation on an ongoing basis. As a large literature in IR has noted, organizations that are capable of operating independently can perform important tasks on the behalf of states, but this independence can also be problematic because an agent with greater freedom of action may not always behave in ways that a principal desires.106 A higher degree of independence, therefore, potentially leads to agency costs, and this possibility can be expected to be mitigated, all else equal, through more vigilant institutional monitoring and control over the modalities of a state’s participation, as well as of the organization itself. Often this will occur through the creation of oversight committees that scrutinize policy related to a !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!102 Klabbers 2001a. 103 Aust 1986; Garcia 2015. 104 Aust 1986; Raustiala 2005; Hollis, Blakeslee, and Ederington 2005; Hollis 2005. 105 One indicative study finds, for instance, that the “legality” of an agreement has a significant, positive effect on whether a treaty is ratified, although the notion of “legality” used in the study is somewhat different from the notion of “formality” as used here. See Spilker and Koubi 2014.  106 For an overview of the issues, see Vaubel 2006.  51   specific organization on an ongoing basis, or requirements that information on treaties and participation in formal intergovernmental organizations be routinely submitted to a legislature.107 It may also occur when a legislature explicitly retains certain rights of control. The US Congress, for example, has delegated to the executive branch considerable power over IMF and World Bank policy but has reserved authority over certain dimensions of US involvement that it considers particularly important.108 Since informal organizations are less public and, typically, less autonomous than formal organizations, institutionalized oversight and control of participation by a legislature is typically much lower.109 In most cases, it is simply the department most directly involved in an informal organization that controls the modalities of participation. Monetary authorities and central banks have, for instance, routinely operated quite independently from legislatures and other political actors with respect to their activities within the G10 and BCBS.110 To the extent that a legislature is capable of controlling involvement in an informal organization, this influence will frequently to be confined to power that can be exerted through the budget process, and the enactment of statutes that can place limits on a state’s activities. But even this—the third major “domestic implication” of different organizational forms—may be limited. Formal organizations are typically functionally separate from their members and much larger than informal organizations. As a result, they usually require separate financing that a legislature can often manipulate. Indeed, in the United Kingdom and Canada (but in most other states as well), a major financial component to any treaty typically necessitates consultation with a !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!107 Pahre 1997. 108 Broz and Hawes 2006; Lavelle 2011. 109 Hollis 2005; Hollis, Blakeslee, and Ederington 2005. 110 Goodhart 2011.  52   legislature, as well as implementing legislation.111 This, in turn, increases a legislature’s ability to intervene in the negotiating process and shape any agreement to participate in an intergovernmental organization, since the executive departments and agencies involved will ultimately depend on approval by the legislature to secure the funds that are needed.112 This influence can then extend beyond the negotiation process, since funding typically must be renewed on a regular basis. In the US, for instance, Congress’ power over contributions to the IMF remains one of the key ways in which it is able to exert influence over the organization and other departments within the government that may have different interests.113 Congress has also periodically withheld the country’s assessed contribution to the budget of the UN and other intergovernmental organizations.114 In many cases, participation in informal organizations may be similarly susceptible to interventions through the appropriations process.115 However, their small size and informal status often means that they can frequently be financed through alternative measures that can be more difficult to control. The many states’ contributions to the GATT secretariat were, for years, undertaken through roundabout ways—typically by utilizing funds that individual departments already controlled. 116  The US contribution was, for instance, routinely “smuggled” through Congress under the Department of State’s budget for “international conferences and contingencies.”117 In many cases, no separate financing may be !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!111 Copithorne 2005; Sinclair, Dickson, and Maciver 2005; Hollis 2005.  112 Ibid. 113 Broz and Hawes 2006; Lavelle 2011. 114 Nelson 1986. For a case study of such conflicts between different arms of the US government, see Woods 2006. 115 Garcia notes that the ability of Congress to control non-binding agreements more generally is largely limited to that which can be exerted through the appropriations process, see Garcia 2015, p.11. 116 See the discussion of other states’ practices, as well as US practice, with respect to financing the GATT secretariat in US House of Representatives 1956. 117 Gardner 1969, p.xxxii.  53   required at all, as when an organization takes a “virtual” form or when secretarial services are supplied by another state.118 Overall, then, the absence of a clear budget “hook” means that a legislature will typically have much less leverage over a state’s participation in an informal organization and, by extension, over the organization itself. Table 1 Properties of Formal and Informal Organizations  Summarizing the discussion so far, it is clear that formal and informal organizations have several defining features and corresponding properties and implications. These can be seen in schematic form in Table 1. Formal organizations are created by states, typically have a body that is separate from these members, and are established by binding international agreements. Informal !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!118 This is the case, for instance, with the Paris Club, where France’s Ministry of Finance plays the role that an international secretariat typically would on the behalf of others. This would then suggest that the costs of other states participation in the Paris Club is extremely low, since there is no need to regularly finance a secretariat. The same is true of the International Competition Network, as discussed in Chapter 6, where Canada’s Competition Bureau serves as the secretariat.  Formal Organization Informal Organization Defining Features Constituent Actors States States Separate Body More Likely Less likely Agreement Binding Non-Binding    International Properties  Flexibility Low High Agility Low High Confidentiality Low High Independence High Low Scale and Scope High Low    Domestic Implications Ratification High Low Monitoring High Low Resources High Low  54   organizations are also created by states, but generally have a less independent, or even “virtual” organizational structure, and are constituted by non-binding agreements. These different characteristics then lead to important practical differences that I refer to as the international properties and domestic implications of each organizational form. Informal organizations tend to be more flexible, more agile, and more confidential. They also place fewer domestic demands on governments in terms of ratification, monitoring and resources. However, these benefits appear to come at the cost of less operational independence, and more limited scale and scope. Formal organizations, by contrast, are typically less flexible, agile, and confidential, and domestic demands are greater. Yet it is possible for states to delegate greater independence to such organizations and to create larger bodies that can provide a greater range of services.  2.3 Mapping the World of Informal Organizations  Thus far, this chapter has focused on conceptualizing formal and informal organizations by outlining their defining features, properties and implications. This has helped to clarify what they are at general level, but it leaves us without a clear sense of the actual organizations and broader organizational patterns that we see in the world. How many informal organizations are there? Have they become more common over time? The next part of the chapter attempts to answer these kinds of basic questions using a new database that maps the world of informal organizations. It is the most complete database of informal organizations available, and it serves a number of important purposes in this study. Most importantly, it defines a population of cases that can facilitate quantitative and qualitative analyses of the determinants of organizational form. But, more generally, the database should also be a valuable resource for other scholars, helping to advance research in the field. Thus far, due to the lack of accurate and reliable  55   quantitative data, scholarship on informal organizations has largely been confined to qualitative case studies of one or a small number of organizations.119 This means that we have been able to say a great deal about a few important organizations but it has been difficult to make inferences beyond these cases and this ultimately limits the kinds of questions that can be asked. By relieving this constraint, the database should greatly expand the range of analyses that can be undertaken and can help move research towards explaining broader empirical trends.  In what follows, I briefly discuss related databases of formal and informal intergovernmental organizations and how they fall short for our purposes. I also explain how my own database was constructed, and I highlight several important properties of the data that we should keep in mind whenever we use them to make descriptive inferences about the world. Finally, I present a series of descriptive statistics derived from the database that should help us to picture the broad organizational patterns that any theory of organizational form should be able to explain.   2.3.1 Previous Efforts  Scholars have been collecting data on intergovernmental organizations for some time. The first serious effort was that undertaken by Wallace and Singer, which traced patterns in a dataset of intergovernmental organizations active in the global system from 1815 to 1964.120 Prior to this time, data was incomplete and imprecise; lists of organizations were mainly illustrative, and the rules determining inclusion in these lists were poorly specified, if at all. The results were, as a result, not particularly useful for making either descriptive or causal inferences about the world. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!119 Examples include Ravenhill 2001; Prantl 2006; Cooper and Thakur 2013; Vabulas and Snidal 2013. 120 Singer and Wallace 1970.  56   By developing rigorous coding rules and mining the Yearbook of International Organizations, as well as compilations produced by the League of Nations, Wallace and Singer were able to assemble a more reliable and accurate picture of intergovernmental organizations than had previously been possible. Subsequent datasets of intergovernmental organizations have built on their efforts, and, in some cases, produced radically different estimates. Mainly, this has been a result of differences over the treatment of “emanations.”121 However, a common thread running through all of these datasets has been their focus on formal organizations.122  This situation has only changed quite recently as a result of the pioneering work of Vabulas and Snidal.123  Although they are not the first scholars to conceptualize and describe informal organizations, they are arguably the first to offer both a positive, empirically testable theory of organizational form and an estimate of the total number of informal organizations in the global system.124 To construct their dataset, they adapt the approach used to construct the COW-IGO Dataset, which is an extension of the earlier work by Wallace and Singer.125 They begin by defining informal organizations in a way that is broadly similar to—but nevertheless different from—my own. According to them, an informal organization is a body that is a) created by states, who b) participate in regular meetings but have no independent secretariat or !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!121 See the discussion of this in Pevehouse, Nordstrom, and Warnke 2004. 122 Or, more accurately, they claim to. As a matter of fact, these databases do include some organizations that are not founded by a binding international agreement and which therefore do not in fact merit inclusion according to their own coding scheme. It is also interesting that the presence of a physical, separate secretariat has not always been deemed essential to the definition of an intergovernmental organization across these various datasets, although it is common to all present-day datasets. 123 Vabulas and Snidal 2013; Vabulas and Snidal 2014; Vabulas and Snidal 2015. 124 Their theory of organizational form is discussed in detail in the following chapter. 125 This dataset is described in Pevehouse, Nordstrom, and Warnke 2004. It is the most widely used and comprehensive database of formal organizations available, see Gartzke and Schneider 2013.   57   headquarters/permanent staff, and c) have an explicitly shared expectation about purpose.126 Each of these three elements has a close counterpart in the definition of a formal organization used to construct the COW-IGO dataset, which means it is possible to directly compare trends in the two datasets.127 Indicators are then developed that can tell us when each of these criteria are met. And, finally, they generate their dataset by applying these indicators to the Yearbook of International Organizations, as well as a small number of other data sources.128 Through this method, they identify 51 organizations, which they claim is a “fairly comprehensive” list that includes the “most important” informal organizations active in the global system.129  This effort to map the world of informal organizations has been an important step forward. But, I argue, their measure suffers from a number of problems that should lead us to conclude that their estimates are likely to be characterized by significant bias.130 The problems stem primarily from differences in the way that they and I conceptualize informal organization and the specific coding rules that they employ. This causes them to systematically exclude important sets of informal organizations from their dataset and include bodies that should not be. For instance, as a result of the way they code informal organizations, they restrict their dataset to organizations whose “members” include ministers or heads of state.131 The rationale for this is never explicitly !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!126 Vabulas and Snidal 2013, p.197-9. 127 In the COW-IGO dataset, a formal organization is defined as a body that is a) created by states, that is b) constituted by a formal treaty, and c) has signs of “significant institutionalization,” such as a permanent headquarters or staff. See Pevehouse, Nordstrom, and Warnke 2004. 128 Ibid, p.202-9. 129 Ibid, p.205. 130 They admit to several biases in the dataset, many of which are unavoidable and likely to plague other datasets of formal organizations as well. However, the measurement errors that I describe here are likely to lead to biased descriptive and causal inferences.  131 Ibid, p.209.  58   stated, but this excludes all organizations that list members from other levels of government, such as non-ministerial agencies. They specifically note their exclusion of the Egmont Group, which is an intergovernmental organization whose “members” are national financial intelligence units. This is problematic, however, because such non-ministerial bodies frequently have the capacity to act on the behalf of a state.132 In the case of the Egmont Group, the financial intelligence units involved do so under the authority granted to them by their respective governments.133 They represent their governments in a way that is in many respects no different from the way that a ministry would, and it is not inconceivable that the organization could have been more formal in nature. Indeed, many formal organizations do in fact list non-ministerial agencies as founders and members. For example, the convention establishing the European Patent Organization was signed by European patent offices that did not typically have ministerial status. The “members” of the Bank of International Settlements are, likewise, primarily central banks and monetary authorities. Neither of these facts leads us to exclude these organizations from the set of formal organizations, and it is therefore unclear why this should lead us to exclude such bodies from the set of informal organizations. So long as they are components of a state, the ministerial or non-ministerial status of the members of an organization is immaterial to the definition of an intergovernmental organization. And, therefore, by excluding such organizations from their dataset, Vabulas and Snidal are likely to underestimate the total number of informal organizations.134  !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!132 See the discussion of this point in Schachter 1977, p.299; Aust 2000, chapter 5; and Hollis, Blakeslee, and Ederington 2005, p.22. It is worth noting, however, that this is not always the case. 133 See Egmont Group Report, Charter. 134 Geddes 1990; King, Keohane, and Verba 1994.  59   The second problem is that their coding rules exclude groups that have physical “bricks-and-mortar” secretariats and permanent staff. This move appears to be motivated by the fact that Vabulas and Snidal regard their absence as a defining feature of informal organizations that helps to sharply distinguish them from their more formal counterparts. However, while many informal organizations do have a purely “virtual” existence, a large number of informal bodies do in fact have permanent staff and secretariats (although these do tend to be much smaller). The lack of a secretariat and permanent staff is a common feature of informal organizations, but it is not a necessary one, as others have argued—as well as myself above.135 Vabulas and Snidal do not explicitly discuss this point, but it is likely that it is the presence of a permanent secretariat and staff that explains the absence of the Asia Pacific Economic Cooperation (APEC), the Black Sea Economic Cooperation (BSEC) and IOSCO from their list of 51 informal organizations, despite the fact that each has frequently been cited as a prominent example of an informal organization by other scholars. 136 And, if we do believe that these kinds of organizations should properly be classified as such, this means that, again, Vabulas and Snidal’s coding rules are likely to underestimate of the number of informal organizations.   Finally, by counting negotiating groups and various sub-committees as informal organizations, Vabulas and Snidal include in their database a number of bodies that should not be there.137 These are, of course, important kinds of organizations that states participate in to achieve shared goals in international affairs. Further, many of these are indeed constituted without a binding international agreement. They therefore share several important features with informal !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!135 Klabbers 2001a; Berman et al. 2012; Di Stasi 2015. 136 Ibid. Also see Kahler 2000; Bach 2010. 137 This can be expected to lead to biased descriptive inference (by overestimating the total number of informal organizations), see King, Keohane, and Verba 1994, chapter 5.  60   organizations. However, it is difficult to classify such bodies as independent organizations, since they are subordinate to or derivative of other formal or informal organizations.138 An example is the BCBS, which is included in their dataset. This is an informally constituted body that has a secretariat and state members. Yet the BCBS is properly regarded as a sub-committee of the G10—an informal organization—and not a separate organization. Although it possesses its own secretariat and has its own charter it remains formally subordinate to the latter. 139 Counting it as a separate organization elides this important fact and causes us to overestimate the actual number of informal organizations. Likewise, Vabulas and Snidal include in their dataset bodies like the Group of 77 (G77) in the UN, the Association of Small Island States (AOSIS) in the UNFCCC, and the Cairns Group in the WTO.140 These are each informally constituted, created by states and demonstrate some level of institutionalization. Further, they are not formally subordinate to another organization. They sometimes even refer to themselves as intergovernmental organizations. Yet, as officially sanctioned negotiating or caucusing groups, they only operate in the context of another organization and must therefore be regarded as in some sense derivative of that larger organization.141  Counting them as separate obscures this fact and leads us to misidentify the total number of informal organizations as well. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!138 The lack of “independence” is the primary reason why the COW-IGO database excludes bodies such as UNICEF, UNDP, UNEP and the IPCC from the set of “formal” intergovernmental organizations. Although they function in a manner that is similar to many intergovernmental organizations, they are subordinate to other bodies. UNEP, for instance, was constituted by and is formally a subsidiary body the UN General Assembly. See Singer and Wallace 1970. 139 Goodhart 2011.  140 They explicitly refer to these bodies as “IIGOs-within-FIGOs”—informal intergovernmental organizations within formal intergovernmental organizations. See Vabulas and Snidal 2013.  141 AOSIS, the Group of 77, and the African Group of Negotiators, for instance, each have an official standing within the UNFCCC and UN that makes them very different from the kinds of self-standing informal organization that we find elsewhere, see Williams 1991; Chasek 2001; Roger and Belliethathan 2016. Arguably, regarding these groups as intergovernmental  61   2.3.2 Constructing the Database  My own database is inspired by and builds upon the work of Vabulas and Snidal. Like them, my goal is create an accurate, reliable and comprehensive mapping of the world of informal organizations. I also adopt their general methodology, described above, and aim to ensure that my data are directly comparable with COW-IGO data, allowing us to integrate the two datasets. However, I attempt to improve upon their effort by operationalizing of the concept of an informal organization in a way that reduces the three sources of measurement error described above. In what follows, I set out four criteria that must be met for a body to be included in my database. These should flow directly and intuitively from the conceptual work undertaken in the first part of this chapter, accurately capturing all of the most relevant dimensions of the concept of an “informal organization” and reliably differentiating them from other kinds of organizations, including formal organizations.142 After explaining each criterion, I also review the kinds of indicators—and specific types of evidence that we can look for to generate scores—which can help us to determine when each criterion is met. In some cases, several different indicators can be used, and for the specific criterion to be met a body must score positively on all or most of these. I discuss each in turn.  First, to be included in the database an organization must primarily have members that are states, or components of states.143 As elsewhere, this kind of criterion helps us to distinguish informal organizations from international bodies created by other kinds of actors. In particular, it excludes !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!organizations is akin to regarding a party as conceptually identical to a legislature. They may share some features, but regarding them as the same kind of entity would be a mistake. 142 That is, our measure of an informal organization should have high content and discriminant validity. See Adcock and Collier 2001. 143 For our purposes, “primarily” is here taken to mean greater than 90 percent of all members.  62   those that are created by non-state actors, sub-state actors or intergovernmental organizations. However, in contrast with Vabulas and Snidal’s operationalization, it specifically does not exclude bodies that list components of states rather than states as such as members, where “components of states” are regarded as governmental bodies that are officially part of a central government. This relaxes things sufficiently to allow us to include organizations such as the Egmont Group, which list agencies and other non-ministerial governmental bodies as members. In order to determine the type of actors that are members of an organization, we will typically look to member lists or lists of “founding members” to see what kinds of organizations are included.144 In particular, we will look for specific language and words such as “states” or “governments” or we will observe that members are de facto components of central governments. The members of IOSCO are, for instance, securities commissions that are typically highly autonomous yet nonetheless components of a central government. The criterion is therefore met in this case. The G20 also meets the criterion because the members of the G20 are all “states.” By contrast, the International Electrotechnical Commission (IEC) and the Monitoring Group do not count as intergovernmental organizations because their members are private industry associations and other intergovernmental organizations, respectively.  Second, an organization must exhibit signs of (limited) institutionalization, or “organization-hood.” This criterion is specifically designed to ensure that an organization is indeed an organization and not simply an ad hoc conference or meeting of some kind. To count, we must see at minimum some indication of permanence and structure, such as regular meetings and rules !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!144 Most often these are available online. Ideally, to determine the nature of an organization I look to lists of “founding members,” such as those that signed an organization’s constituent agreement.  63   about decision-making. Meeting this criterion does not require that an organization have a physical headquarters or permanent staff, as does the COW-IGO Dataset. We can, therefore, include in the database all kinds of virtual organizations that do not have these features. But, at the same time, this does not mean that we must correspondingly exclude organizations that do. A headquarters or permanent staff can, in fact, be a strong indicator of institutionalization—although, as noted, we should expect these to be quite small. In order to determine whether an organization meets this criterion we will examine a potential case to see if three conditions are met. First, there should be evidence that the members of the organization meet at regular intervals, where “regular” usually means at least once a year. To infer this we should find evidence of public records of regular meetings between members, public statements about meetings, and so on. Second, we should observe some kind of implicit or explicit rules establishing an organizational structure and decision-making processes.145 This could be found in a constitution that specifies the various bodies that comprise the organization and the voting rules that determine how decisions are made. It could also be found in the informal practices that practitioners in fact engage in. Finally, we should see some kind of organizational output, such as new rules, resolutions, or documents summarizing the activities of the organization.  Third, an organization must be independent. This criterion—which is implicit in the current COW-IGO database coding146—requires that an organization not be formally subordinate to or !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!145 In terms of structure, we will look specifically for evidence of functional differentiation, such as a division of responsibilities between different bodies. Evidence may include a chair, a board, committees a secretariat, and so on. In terms of decision-making processes, we will look for rules about order, committee membership, voting, and so on.  146 Pevehouse, Nordstrom, and Warnke 2004. However, it is explicitly stated as a criterion for inclusion in the earlier version of the dataset compiled by Singer and Wallace 1970. Also see the discussion of this point in footnote 125 above.  64   derivative of some other organization. For this to be the case, we must find evidence that suggests the organization is “self-standing” and not just a sub-component or sub-group within another body, whether formal or informal. To infer this, two kinds of evidence are relevant. First, we should look at constitutional documents or public statements for rules or language indicating that the organization is “autonomous” and not formally subordinate or accountable to another organization. Second, we can examine statements about the goals of the organization: is the primary objective of the body to influence the policies of another organization? If an official statement of purpose explicitly mentions another organization, and/or if the goal is to “increase negotiating leverage,” “facilitate bargaining,” “coordinate negotiating positions,” and so on, then we can infer that the organization is largely derivative of another. For instance, the stated purpose of the G77 is to provide “the means for the countries of the South to articulate and promote their collective economic interests and enhance their joint negotiating capacity on all major international economic issues within the UN system.”147  By defining its aims and activities in relation to another organization (here, the UN), this signals that the group is in some sense dependent on another for its raison d'être.  Fourth, an organization must be constituted by a non-binding international agreement. This final criterion is, in practice, the most crucial one for determining whether an intergovernmental organization is in fact informal. It requires that we find evidence indicating that an organization’s constituent agreement was intended to avoid establishing international legal obligations. This can be indicated, first, through either the form of the agreement or the language or statements used to compose it. For example, if the agreement constituting an organization is unwritten and implicit !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!147 G77 2015. Emphasis added.  65   in the practices states, the agreement does not—by definition—constitute international law.148 In practice, very few organizations are likely to be founded in this way. But some are. The Group of 7/8 (G7/8) is an example, as it was not established by a written agreement.149 By contrast, if an agreement is written we must then search the document constituting an organization for language that deliberately avoids implying that it establishes international legal obligations. As previously noted, this may be stated explicitly through a clause that indicates the precatory nature of the agreement: “this agreement is voluntary,” for example. But it may also be implicit in the use of language. The agreement may be referred to as an “arrangement” or an “understanding” instead of a “charter” or “treaty,” or it may use words such as “hope” “may” or “should” rather than “will” “must” or “shall.” Finding language of this kind would imply that the document is not intended to establish legal obligations. Finally, certain activities may also reveal information about the nature of an agreement. For instance, if states register an international agreement with the UN this sends a strong signal that they in fact regard it as an international treaty. Finding that an agreement was ratified by states would also tend to indicate that an agreement was intended to be binding on members, since this is a practice normally associated with international treaty-making.  These four criteria, the indicators used to determine whether the criteria are met, and examples of the kinds of evidence that we can look for to determine the scores for each indicator are summarized in Table 2. According to this scheme, an organization will count as an informal organization and will be included in the dataset only if each of the four criteria is met. For the !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!148 On this point, see Aust 2000, chapter 2. 149 Dobson 2006.  66   criteria to be met, we must find evidence that leads to positive scores on all of the indicators.150 To generate the dataset, I have then applied this coding scheme to the Yearbook of International Organizations—as other studies have done—as well as several other smaller lists of organizations from other sources.151 I have also employed the “searching and asking” technique used by Hale and Roger to “snowball” beyond these sources to find other similar organizations that fit the criteria.152 Ultimately, this method yields a sample that gathers the most prominent examples of a phenomenon. Yet it is important to remember that the dataset is likely to remain biased in several known ways.153 First, it is likely to under-represent cases that are highly secretive or which have limited information about them. This may lead to a bias against security and signals intelligence organizations, or organizations from the distant past. Second, it is likely to over-represent relatively “successful” cases, since we will generally have less information about those that were unsuccessful (and, as a result, may have disappeared, or merged with another organization). Third, we will likely over-represent cases in the Global North, since those in the Global South may have fewer resources, are likely to be less well connected, and their affairs may not be conducted in English. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!150 The one exception is that, although evidence of an unwritten constituent agreement may be a positive indicator of informality (by definition, a treaty must be written), many informal agreements are in fact written. This means that a constituent agreement can score negatively on this criterion, but still be “counted” as informal if the other criteria are met. 151 Klabbers 2001a; Slaughter 2004; Berg and Horrall 2008; Berman et al. 2012; Hale and Roger 2014. Evidence to support the coding for each organization—as well as those considered but excluded—is on file and available for scrutiny by anyone interested. Note, also, that the specific copy of the Yearbook that I have used is the 2009/10 Volume (Parts 1-3), see Union of International Associations 2009. Each volume is updated simply by adding in new entries, so it is suitable to build a database using the most recent version. Given that the entries are updated with some time lag, I assumed that the 2009/10 Volume provides a complete record of the organizations created in 2005—the most recent year for which there are data in the COW-IGO dataset. 152 Hale and Roger 2014. 153 Vabulas and Snidal 2013; Hale and Roger 2014.  67   Table 2 Criteria, Indicators and Evidence Determining Inclusion in the Dataset Criteria Indicators Evidence 1. Membership States/State Components Member list, agreement, website. 2.Institutionalization Regular Meetings Public records, website, agreement. Structure/Rules Agreements, practices, websites, public statements. Output Meeting notes, programs, policies, reports. 3. Independence Formally autonomous Agreement/rules, website, public statements. Independent objectives Agreement, public statements. 4.Non-binding Agreement Written/Unwritten Practices, public statements, secondary sources. Language Agreement, public statements. Associated Behavior UN Treaty Series, public statements and records.  2.3.3 Preliminary Findings  The database reveals many fascinating trends, many of which are discussed in subsequent chapters. Here, I focus simply on the temporal trend. Figure 1 shows the total number of active formal and informal organization active in the global system from 1945 to 2005, the most recent year for which there is data available. The red line depicts the cumulative total number of formal organizations, as measured by the COW-IGO Dataset.154 The green line depicts the number of !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!154 Pevehouse, Nordstrom, and Warnke 2004. Specifically, I have used the most recent version of the dataset (Version 2.3) that extends the data through 2005. Note, however, that I have adjusted this data somewhat, since a number of informal organizations (37 in total) are miscoded as formal organizations in the COW-IGO data. GATT, for instance, is coded as being informal and therefore is removed from the formal dataset. I have also removed a number of organizations (18 in total) from the COW-IGO dataset that are not properly regarded as independent  68   informal organizations, as measured by my own companion dataset.155 The blue line is simply the sum total of these two. It shows the total number of intergovernmental organizations (both formal and informal) active in the global system throughout this period. This tells us a number of interesting things. First, the data shows that at the end of the Second World War nearly all (more than 95 percent, or 102 out of 107) of the intergovernmental organizations active at that time were formal in nature. The number of informal organizations, by contrast, was vanishingly small (less than 5 percent, or 5 out of 107). Ten years later, little had changed. The number of formal organizations had risen to 139, while the number of informal organizations only rose to 14. Then something changed. In the early 1960s, there was then a significant decline in the total number of intergovernmental organizations, as many pre-war formal organizations disappeared, or were merged with other existing bodies.156 The number of informal organization stayed relatively similar (rising to only 21 by 1969), but their relative share of the total increased significantly as a result.   !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!intergovernmental organization by any criteria, either because they do not have state members (the Council of Patrons, for instance) or are an emanation of another organization (the European Foundation for the Improvement of Living and Working, for instance, which was constituted by the European Economic Community). I have also added a large number (111 in total) of formal intergovernmental organizations that meet all of their criteria for inclusion, but for some reason or another have been omitted (the International Visegrad Fund and the Islamic Financial Services Board, are examples). Records of all of these changes to the original dataset are on file with the author. Arguably, all of this constitutes significant improvement to the original COW-IGO dataset and the new version is the most complete record of formal intergovernmental organizations in the global system currently available. 155 My own database follows the same method as the COW-IGO database for determining institutional birth and death. The only major difference is that some informal organizations “die” when they become formal. Thus, for example, GATT is coded as an informal organization and “dies” in 1994, at which point a new organization, the WTO, appears in the formal dataset. The “date of death” is in this case is taken to be the “birth date” of the new formal organization. 156 See  Pevehouse, Nordstrom, and Warnke 2004 for a discussion of this trend.  69   Figure 1 Intergovernmental Organizations in the Global System  This shift proved to be a portent of things to come. Following the absolute and relative decline in their numbers in the early 1960s, the pace at which formal organization were created picked up again—but only briefly. In the mid-1970s, the rate of increase decreased significantly. Between 1980 and 2005, for instance, number of formal organizations increased by only a third—from 300 to 407. Other scholars—looking only at formal organizations—have noted this phenomenon with some puzzlement, given the significant deepening of interdependence that occurred throughout this period.157 However, once we take informal organizations into account, the decline seems less anomalous. Overall, it did not seem to be because existing institutions were satisfactory in some sense, or because there was less need for cooperation. If anything, the need for intergovernmental organization increased, particularly in the 1990s. The overall rate of !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!157 Ibid.  70   institutional creation actually rose somewhat at that point in time. Instead, there was a significant shift in the kind of organizations that states were creating—towards a more informal variety. Between 1980 and 2005, for instance, the number of informal organizations rose from 38 to 179—just shy of a 5-fold increase. And, by the end of the period for which we have data, informal organizations were a major force in global politics, comprising around third of all the intergovernmental organizations in existence.  !2.4 Conclusion  This chapter has accomplished two core tasks that provide a firm conceptual and empirical base for exploring the rise of informal intergovernmental organizations. First, I have defined informal organizations. Like their more formal counterparts, informal organizations are created by states, but they differ in that they tend be less autonomous and exhibit much lower levels of legalization. In general, they are less likely to have bodies that are functionally separate from their members and more likely to be constituted by a non-binding agreement, such as an MOU. I have then shown that these are not simply differences in style, but have a number of substantive implications. I accomplished this by comparing and contrasting both the “international properties” and “domestic implications” of formal and informal organizations, which affect how they function and the costs (for states) of creating and working through them, respectively. With respect to their international properties, informal organizations are generally more flexible, more agile, and more confidential than formal organizations, but are less likely to be autonomous or exhibit economies of scale and scope. With respect to their domestic implications, informal organizations generally have fewer ratification requirements, require less input from other domestic actors, require less monitoring and need fewer resources.   71    The second part of the chapter then discussed the database of informal organizations that I have developed to facilitate our ability to make both descriptive and causal inference about the world of informal organizations. I began by providing an overview of previous efforts to quantify both formal and informal organizations and explained how they fall short in various ways. Building on the conceptual work undertaken in the first part of the chapter, I then developed a number of criteria that can be used to identify formal and informal organizations in the real world. This involved developing indicators that can be used to measure whether a governance arrangement should count as an “organization,” and, if so, whether we should regard it as formal or informal. Finally, I presented a number of descriptive statistics that help to illuminate the world of informal organizations as it has evolved between 1945 and 2005. This showed that at the end of the Second World War, the vast majority of intergovernmental organizations were formal in nature. However, over the ensuing 60 years, informal organizations grew in both absolute and relative terms. Today, they constitute nearly a third of all the currently “active” intergovernmental organizations. Explaining why states have chosen to create such organizations is the main task of this study, and in the next chapter I discuss how we can do so.    72   Chapter 3. Theorizing Informal Organizations   3.1 Introduction  The previous chapter distinguished between two kinds of intergovernmental organizations, formal and informal, and reviewed their different properties and domestic implications. It then demonstrated that both kinds of organization are important fixtures of the global landscape. Having done so, we are now in a position to address the central question that this study is concerned with: why are some intergovernmental organizations formal and others informal? This is particularly important because present efforts are theoretically unsatisfying and rest on weak empirical evidence. These explanations suggest that states create different kinds of organizations because some are better at solving certain problems than others. Policymakers, on this view, are able to accurately identify the problems they face and choose the organizational form that is best tailored to resolving them. This seems reasonable: why would policymakers not want to design organizations that are well suited to the problems they confront? The proposition also seems reasonable given that scholars have in fact found some evidence that the design of organizations and social institutions seems to be related to specific problems of cooperation.158 But, as we will see later in this chapter, such “functionalist” theories fall short in ways that suggest something else must be at work.   The argument of this study is that while functionalism may be a useful starting point a truly compelling theory of informal organizations must incorporate what it overlooks: domestic !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!158 Koremenos 2013b; Koremenos 2013a. Comparable theories can be found in the fields of law and economics. See, for example, Posner 1972; Schotter 1981; Williamson 1985; Williamson 1991; Milgrom and Roberts 1992; Calvert 1995.  73   politics, distributional conflict and power. The goal of this chapter is to explain why this is so and to turn this general proposition into an empirically testable theory of organizational form. I begin by describing functionalist theories in detail, since, as the best explanation offered thus far, these serve as the natural yardstick against which any theory should be assessed.159 I show that functionalists make certain “bets” about the kinds of variables that will best explain organizational form and I tease out the specific empirical predictions that they make. I then offer some preliminary evidence suggesting that they get some important empirical predictions wrong and I show that there are good theoretical reasons for doubting their explanatory power as well. In the second part of the chapter, I outline my own theory, which makes a different bet about the kinds of variables that are likely to explain the form of an organization. Specifically, I hypothesize that the kind of organization that is likely to emerge in a given context depends on processes of state preference formation, how preferences are distributed across states (or groups of states), and the distribution of power, which determines which preference is likely to prevail when conflicts arise. The model I develop yields a number of different empirical predictions about the conditions under which we should expect informal organizations to emerge that are then tested in subsequent chapters.   3.2 Previous Efforts  Functionalist theories have a long history in the field of IR and have much to say about questions of institutional design.160 The most prominent modern day variants are largely couched in the “rational design” tradition of IR. Theorists in this school argue that institutions matter a great !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!159 Popper 1963. 160 Claude 1964; Mitrany 1966.  74   deal and that states choose designs that offer them the greatest benefit in a given situation. 161 Their argument begins from the premise that states act instrumentally, carefully evaluate their options, and choose those that are expected to yield the greatest utility. They then argue that states can often benefit significantly by cooperating with one another but are confronted with a variety of different strategic problems that inhibit their ability to do so. Unless these underlying problems can be overcome, the benefits of cooperation are difficult to attain. International institutions can be designed to address these problems and help states achieve positive cooperative outcomes that would not otherwise be possible. But, since not all problems are the same, the optimal institutional design will vary.162 Some institutional designs will “match” better than others with the underlying problems that stymie cooperation in a given area. States, seeking to use their limited resources as efficiently as possible, will therefore choose those institutional designs that best address the strategic problems that they confront. The bet made by rational design theorists is that there will be a strong, consistent relationship between certain types of cooperation problems and specific institutional designs.   This approach has the advantage of being able to draw on a wide variety of results from contract theory and game theory to develop a priori conjectures about the kinds of problems that are likely to be correlated with different institutional designs that can then be subject to empirical !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!161 Martin 1992; Abbott 1993; Koremenos, Lipson, and Snidal 2001; Trachtman 2008; Koremenos 2013b; Koremenos 2013a; Koremenos forthcoming. Current theories generally trace their lineage to Keohane 1982 and Keohane 1984. 162 As Joel Trachtman 2014, p. 167, has argued with specific reference to intergovernmental organizations: “for each international cooperation area, there is an optimal international organizational structure.” He adds, rightly, that the “optimal” organizational structure may be no organization at all. Also see Martin 1992.  75   test.163  In this way, it has been used to explain a wide variety of different features of intergovernmental organizations, from the degree of independence they possess to the amount of access granted to non-state actors.164 It has also been applied to decisions about whether to create an intergovernmental organization in the first place. Abbott and Snidal have, for example, hypothesized that states will tend to create intergovernmental organizations when doing so can help them to realize gains from cooperation.165 They make their case by pointing to two key properties of intergovernmental organizations: centralization and independence. These properties are useful to states because they can help to reduce the barriers that often block cooperative agreements by inter alia providing a stable venue for negotiations, neutral information about preferences or behavior, or impartial dispute settlement. The implicit claim made by Abbott and Snidal—made explicit in the work of others166—is that states will create intergovernmental organizations when there are significant benefits to be derived from a stream of cooperative agreements in an area but also substantial transaction costs. Keohane developed a similar hypothesis in his classic work, arguing that this is particularly likely when “issue density” is high, and this hypothesis has received support from Koremenos, who has found that states are more likely to delegate tasks in international agreements when they confront “complex” cooperation problems.167   Such arguments about the emergence of intergovernmental organizations are usually framed as being about formal organizations, but the benefits of centralization and independence help to !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!163 Koremenos, Lipson, and Snidal 2001; Koremenos forthcoming. 164 Haftel and Thompson 2006; Tallberg et al. 2014. 165 Abbott and Snidal 1998. They and others draw on theories of the firm, see Coase 1937; Williamson 1985. 166 Trachtman 2008; Trachtman 2014. 167 Keohane 1982; Koremenos 2008.  76   explain why states wish to create any intergovernmental organization, whether formal or informal.168 However, it does not explain why states choose between these two varieties. Other accounts have tried to do so, and they also couch their theories in this functionalist tradition: the basic answer is that states choose the organizational form that will benefit them most in a given context. This claim is made by arguing that informal organizations have certain properties that contrast with the properties of formal organizations—in much the same way as I did in the previous chapter.169 They then argue that states choose to create informal organizations when the structure of the problem they face—Koremenos has referred to this as the “situation structure”170—provides them with an incentive to choose organizations with those properties. This is, for example, the approach taken by Vabulas and Snidal, who arguably offer the most advanced theory of organizational form.171 After comparing and contrasting formal and informal organizations, they argue that states will choose to work through an informal organization when doing so is somehow advantageous. Their core claim is that they do so because “different issue areas—with different underlying cooperation problems—are best addressed by different levels of institutional formality.”172  Vabulas and Snidal are not the only ones to make this kind of argument. Zaring, Raustiala, Whytock and Eeilstrup-Sangiovanni have all offered essentially similar functionalist theories !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!168 Although they are not clear on this point, it seems Abbott and Snidal believe that preferences for centralization or independence are each regarded as sufficient for states to create an intergovernmental organization. But, as an empirical matter, it is worth noting that while all intergovernmental organizations (both formal and informal) are centralized to some extent, not all are independent. See Haftel and Thompson 2006; Hooghe and Marks 2014. 169 Note that they focus exclusively on what I have referred to as the “international properties” of formal intergovernmental organizations. 170 Koremenos 2013a. Also see Abbott 1993. 171 Vabulas and Snidal 2013. 172 Ibid, p. 212.  77   about transgovernmental networks, which I roughly equate with informal organizations.173 Related propositions can also be found in the literature that attempts to explain why states opt for “hard” or “soft” international law.174 In each case, the prediction is that states choose the institutional form that somehow “fits” or “matches” a given problem. Exactly what problems match with formal and informal organizations is somewhat unclear, as scholars in this area of research have offered a wide variety of different conjectures. Some even appear to contradict each other to some extent.175 But across these various theories there is some overlap. Table 3 presents a summary of the specific empirical predictions that are most commonly made.176 In the far left column (“Property”) I have listed the various dimensions along which formal and informal organizations diverge. These dimensions track my discussion of the “international properties” of different organizational forms in Chapter 2. In the second and third columns, I !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!173 Zaring 1998; Raustiala 2002; Whytock 2005; Eeilstrup-Sangiovanni 2009. A related argument has also been made by Kahler and Lake 2009. In some cases, as with Zaring’s work, the focus is explicitly on “organizations,” and not “networks” per se. Elsewhere, the focus is at least implicitly on organizations, since in practice these authors seem to equate transgovernmental networks with organizations such as IOSCO, the G20, the PSI, and so on. The models offered by these authors are also particularly striking because of their similarity to the arguments made by Vabulas and Snidal. This suggests that if there are differences between informal organizations and transgovernmental networks, functionalist theories are presently unable to tell us when one or the other would be expected to arise. In a future project, it would be useful to reflect upon the different features of informal organizations and transgovernmental networks, with the intention of revising the definition of the latter. 174 See, for instance, Lipson 1991, Abbott and Snidal 2000, Guzman and Meyer 2010 and Meyer 2014. 175 This is the case, for example, with hypotheses about the relationship between the number parties, preference heterogeneity, and institutional form. While Eeilstrup-Sangiovanni argues that informality should be more likely when agreements are reached between a small number of actors with congruent interests, Koremenos has argued that informality should increase with the number of actors and the diversity of interests. See Eeilstrup-Sangiovanni 2009; Koremenos 2013c. In my discussion, I use the modal hypothesis found in the literature—in this case, that informality should decrease with numbers and preference heterogeneity. 176 These predictions are mainly drawn from Zaring 1998, p.312-323; Raustiala 2002, p.17-26; Whytock 2005, p.117-130; Eeilstrup-Sangiovanni 2009, p.205-210; Vabulas and Snidal 2013, p.209-212. In my exposition I also draw on additional sources to bolster and expand their theoretical arguments.  78   have then listed the different conditions that have been hypothesized as being important for determining each institutional choice and which are related to the properties in the first column. Consider each in turn.  Table 3 Functionalist Determinants of Organizational Form Property Formal Organization Informal Organization 1. Flexibility Low uncertainty  High uncertainty  2. Agility   Low urgency,  Routine problems. High urgency,  Systemic crises. 3. Confidentiality   Less sensitive issues, Cooperative external relations. Highly sensitive issues, Antagonistic external relations. 4. Independence   High opportunism,  Heterogeneous interests.  Low opportunism,  Homogeneous interests  5. Scale  Large numbers,  Wide issue scope.  Small numbers,  Narrow issue scope.  First, numerous scholars draw a link between the levels of uncertainty in an issue area and levels of institutional formality. When states can achieve gains in the present by creating an intergovernmental organization but are uncertain about how participation in an organization may affect the distribution of gains in the future they should be more likely to choose an institutional design that can be changed or adapted relatively easily.177 This is particularly likely to be true when an issue is characterized by what Koremenos refers to as “persistent uncertainty.”178 Persistent uncertainty prevails when the “shocks” that can be expected to arise in an area are !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!177 Ibid.; Guzman and Meyer 2010; Meyer 2014. 178 The term comes from Koremenos 2005.  79   essentially random in nature, costly to value in advance and add up as time passes. The dynamics of an issue may, for instance, be known to fundamentally change over time and present states with unforeseen challenges where the optimal response is inherently unknowable at the time an institution is created. When this is the case, an institution that can be changed more easily offers clear advantages. Once states are confronted with an issue they did not anticipate, they can easily adapt the institution they originally established to the new circumstances. An informal organization—which, as discussed in the previous chapter, is much more flexible than formal organizations—may therefore be expected to have an edge in this kind of context, as Koremenos herself notes in her discussion of G7 macroeconomic cooperation.179   Second, nearly all functionalists have linked levels of institutional informality to situations where speed is important to successful governance.180 As noted in the previous chapter, one of the crucial advantages of informal organizations is their agility. In contrast with their formal counterparts, informal organization can be set up relatively quickly, since there is no need for ratification and states will have less reason to hold out during bargaining when the shadow of the future is shorter, i.e. when emphasis is placed on the immediate present and the future is highly discounted. This makes informal organizations an ideal choice whenever failure to reach agreement in a timely manner may result in significant losses to all parties, and states therefore urgently need to set up a venue for cooperation. This may occur when, for example, issue areas are prone to unpredictable systemic crises, as in international finance where states have been forced to respond with swift ameliorative actions. By contrast, where emphasis is placed less on the speed with which cooperative agreements are reached and more on ensuring that !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!179 See Ibid, p.552. 180 Lipson 1991; Raustiala 2002; Eeilstrup-Sangiovanni 2009; Vabulas and Snidal 2013.  80   commitments made in the present can be sustained far into the future, a more formal organization may have significant advantages. When this is the case, the time and effort involved in ratifying more formal agreements and their explicitly legal nature can actually help states to signal long-term cooperative intentions.181   Third, functionalists have hypothesized a relationship between informality and the need for confidentiality.182 In comparison with formal organizations, there is generally much less need to disclose details about the inner workings of informal organizations and actors within such organizations can more easily conceal information about their intentions or actions. Informality is therefore likely to be preferred in situations where foreknowledge of the intentions or actions of members risks subverting their effect. More concretely, this means that informal organizations may be preferred in areas where sensitive information is being exchanged, as in global competition policy, where proprietary information about businesses is routinely transferred between agencies. It may also be particularly useful in areas like signals intelligence and counter-terrorism, where it may be important for actors outside of the organization to remain unaware of what goes on within. At the other extreme, in some areas an organization’s members may want to make their intentions and actions especially visible. This may be the case when, for instance, an organization’s success somehow depends on eliciting cooperation from actors formally outside of it, such as non-state actors or other states. States would then want to make their actions especially transparent, since this would allow them to gain the trust of such actors and be seen as legitimate in their eyes. Greater formality may be one way to achieve this.   !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!181 Guzman 2008; Simmons 2010. 182 Vabulas and Snidal 2013.  81   Fourth, scholars have argued that the potential for opportunism in an issue area may be related to choices about organizational form. 183 In this context, “potential for opportunism” is meant to denote the extent to which there are incentives for states to defect from a cooperative agreement to gain an advantage over others in an issue area, as in a Prisoner’s Dilemma-type problem. When states confront such “enforcement” problems, successful cooperation often turns on the extent to which states can make their commitments credible, establish reputations, and detect and punish violations. This should, in turn, make states more likely to opt for a more formal organizational design, which can facilitate inter alia independent monitoring of compliance and credible commitment-making via costly ratification. By contrast, when there is less potential for opportunism in an issue area—when, for instance, states face a coordination problem—there is much less need for such costly and invasive mechanisms. Since states only need to coordinate on one set of rules or another and have few incentives to defect from an agreement once reached, a centralized venue for reaching agreements and basic secretarial services is often all that is needed. An informal organization may then be sufficient to achieve the level of cooperation desired by states.   A related rationale that scholars have noted is related to the heterogeneity or homogeneity of preferences in a given issue area.184 As preference heterogeneity rises, this should make it more difficult for members of an organization to strike deals with each other, since their “ideal” agreements will grow further apart and the overlapping “win-set” correspondingly smaller.185 It may also reduce levels of trust, since it is less likely that the interests of one actor will be !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!183 Martin 1992; Whytock 2005; Vabulas and Snidal 2013. 184 Whytock 2005; Eeilstrup-Sangiovanni 2009. 185 Whytock 2005; Eeilstrup-Sangiovanni 2009; Vabulas and Snidal 2013.  82   “encapsulated” by another’s.186 And, even if states do reach agreement, it can increase the likelihood that there will be serious disputes between members. States may then find it particularly useful to delegate greater independence to a secretariat in order to facilitate bargaining, to monitor and enforce commitments, and to provide for third-party dispute settlement when conflicts arise.187 A formal organization may then prove more effective than an informal one. By contrast, when preferences are more homogeneous it is relatively easy for states to strike bargains, they are willing to rely on trust to a greater extent, and there will generally be fewer disputes between members of an organization. The incentive to delegate greater independence to an organizations is therefore likely to be much lower, since there will be less need for an independent broker, monitor or dispute arbiter in this context. A minimal, low-cost informal organization may then be the institutional approach favored by states.  Finally, we can hypothesize a relationship between the number of members an organization has, issue scope, and organizational formality.188 As an organization’s membership increases, it generally becomes more difficult for states to know the preferences of others and the number of potential coalitions increases. This makes it more challenging for states to strike cooperative agreements, and provides a functional rationale for a strong secretariat that has the capacity to !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!186 For a discussion of trust in terms of “encapsulated interests” see Cook, Hardin, and Levi 2005. 187 There is some disagreement about this proposition in the more general literature. Some actually argue that delegation is more likely to occur when preferences are homogeneous, see Hawkins et al. 2006. But others have disputed this, and the bulk of quantitative evidence seems to support the idea that that delegation increases with preference heterogeneity, see Koremenos 2008; Stone 2011; Green and Colgan 2013. To an extent, the disagreement between these two views may simply be a semantic problem, as states may have heterogeneous preferences at one level and homogeneous preferences at another. This is clearly the case in many strategic problems. Consider, for example, the “battle of the sexes” game, where the actors both want to coordinate on a single agreement but prefer different agreements. 188 Haftel 2013; Hooghe and Marks 2014.  83   collect, analyze and disseminate information about preferences and to set a feasible agenda. More generally, a larger membership also requires a secretariat with greater logistical capacity and facilities that can accommodate larger groups. The greater scale of such an organization may only be possible to achieve through greater formality. The scope of the issues that an organization aims to deal with may also influence decisions about organizational form. When an organization attempts to deal with many issues at once, this makes bargaining more complicated by significantly increasing the number of potential deals that can be struck, and this provides a strong reason for states to establish a secretariat that has the ability to identify feasible “package” agreements. It also means that an organization may need to retain a larger number of specialized staff and perform a more diverse array of tasks in order to operate effectively. Both considerations are likely to lead states to increase the independence and scale of an organization and to therefore opt for greater formality.  3.2.1 A Critique of Functionalist Theories  While intuitively plausible, there are reasons to doubt the empirical validity and theoretical logic of functionalist theories of organizational form. We can, for example, think of many formal and informal organizations that do not seem to have arisen under the conditions that functionalists expect them to. Consider the example of the International Organization of Supreme Audit Institutions (INTOSAI), an informal organization established in 1953. Its primary purpose is to foster exchange and dialogue between national audit offices, and to facilitate the development of standards in the field of government audit. Yet in many ways it seems that the situation structure would predict greater formality. Uncertainty about the future in the field of government audit is likely to be very low, speed is not likely to be important as tasks are routine, and confidentiality  84   is not likely to be a great priority either, since the area thrives on publicness. The organization also boasts membership of 192 states that, of course, have highly heterogeneous interests. Next, consider the International Bureau of Weight and Measures (which goes by its French acronym, BIPM). The BIPM is a formal organization, constituted by the Metre Convention of 1875, yet the underlying problem it aims to solve again lacks some of the features that would lead us to expect this. Its purpose has been to help states establish international measurement standards (initially, just mass and length), coordinate comparisons of national reference standards, and to undertake scientific activities related to these two objectives. It therefore focuses on a very narrow range of issues where there is little incentive for opportunism. Further, it was negotiated by a small number of states (11) that were all already users of the metric system and therefore had quite similar interests. We might, therefore, expect the BIPM to be founded informally—as, in fact, some other standard-setting bodies have been—but it was not.  We can, of course, multiply examples where it seems that few of the functionalist conditions are met. Still, one might reasonably argue that I have cherry-picked these cases or that these different institutional designs are driven by relevant variation in other aspects of the issue areas each organization operates in. Another way to probe the validity of the functionalist argument is therefore to look at regional variation within issue areas. Functionalists, of course, predict that institutional design should be highly correlated with the situation structure in an issue area. We should therefore expect that organizations operating within the same issue area but in different regions should be designed similarly. However, this is not always the case. For example, organizations designed to coordinate action on money laundering exhibit considerable institutional variation across regions. In some, the primary institutions are highly formal bodies,  85   such as MONEYVAL, in Europe, and the Intergovernmental Action Group Against Money Laundering in West Africa (GIAMBA), a specialized agency of the Economic Community of West African States (ECOWAS), in West Africa. Elsewhere, these institutions are highly informal. The premier global body in the area, the FATF, and other organizations like the APG and the EAG are all informal organizations. Similar examinations of organizations active in the areas of nuclear non-proliferation, international banking and metrology, to name just a few, also reveal a host of formal and informal regional bodies designed to govern the same issues.189 But, if the issue area is indeed held constant across regions in each case, how can we explain this variation? Something else must be going on.   As these brief observations suggest, the empirical validity of functionalist explanations of organizational form is—at the very least—assailable. However, some more general theoretical problems stand out as well, which may help to explain these kinds of puzzles. First, as Koremenos has herself noted, the rational design framework “does not explicitly include domestic politics.”190  Instead, it takes preferences as given, and relies exclusively on the situation structure that states confront to explain institutional design. To some degree, of course, this theoretical move is justifiable, as it allows scholars to explore difficult empirical questions with relatively parsimonious explanatory models. It may be the case that functionalist variables can explain a large amount of the variation that we find in the world. And, to the extent that this is true, the value and policy relevance of rational design theories is evident. Yet it is also !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!189 Of course, the specific problems that states face in different regions—even within the same “issue area”—may not actually be identical. However, this probe nevertheless suggests that there may be problems with the functionalist argument and that further investigation is needed, at the very least. This is done in the next chapter. 190 Koremenos forthcoming , p.7. Italics in original.   86   problematic because domestic politics have been shown to fundamentally shape international behavior and may matter a great deal for questions of formality in particular.191 Raustiala and Guzman have, for instance, both argued that domestic politics play an essential role in determining the legal form of many international agreements.192 It is, therefore, reasonable to expect that it matters for organizational form as well. If so, different levels of formality may be determined less by the nature of the problems states confront and more by domestic processes of preference formation and how these “aggregate" across states to produce the final outcomes we see. But testing this hypothesis requires that we first theorize the process of preference formation that is applicable to the form of an intergovernmental organization—a task that is taken up in the second half of this chapter.  The second problem with functionalist theories—which is directly related to the first—is that they underemphasize conflict and thereby offer an overly consensual account of the politics of institutional design. This issue has been noted about functionalist theories more generally, which tend to focus on the collective benefits that institutions offer a group of actors at the expense of their distributive effects. 193 The particular issue is that, while taking note of the collective benefits of institutions and organizations is often essential for understanding why they exist in the first place, a wide variety of different forms can often serve the same purpose. There is, in other words, a problem of “multiple equilibria.” The real question is which of these many possibilities actors will chose. Since they may all represent improvements over the status quo, !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!191 Milner 1997; Milner 1998; Moravcsik 1998. 192 Guzman 2008; Raustiala 2005. 193 See, for instance, the powerful critique by Knight 1992. In IR, this kind of critique has often associated with Krasner 1991. Others have noted that international cooperation regularly involve both collaborative and distributive dimensions, see Fearon 1998.  87   the different possibilities are not necessarily distinguished by their ability to provide collective benefits. Instead, as Knight has explained, “the major distinguishing feature of different institutional forms is their distributional consequences.”194 Actors may agree that they can benefit from cooperation, but the various solutions may have different distributional effects and therefore they will disagree over how the “additional” benefits of cooperation should be distributed among them. The critical issue for understanding institutional and organizational forms is then explaining who benefits from these different possibilities and how conflicts over institutional design are resolved.195  This brings us to the third major problem with functionalist theories: their neglect of power. If we accept that domestic politics may determine preferences for different organizational forms and that state preference can sometimes conflict, explaining final organizational outcomes requires an account of whose preference is likely to prevail in a given situation. In theory, a wide range of different factors may be relevant, such socialization or persuasion, but, whenever deep-seated interests are at stake, power—an ability to impose one’s will on others—is likely to be the most relevant.196 The issue for functionalists is that they have not been particularly good at incorporating power into their theories. Duffield has observed, for instance, that considerations of power have only been loosely integrated into theories of rational design.197 Where it is introduced into theories, this is often done in a relatively ad hoc manner—almost as an !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!194 Ibid, p.26. 195 This omission has been noted by Pollack and Shaffer, who have recently called for a “distributive approach” to understanding the design of international institutions, see Pollack and Shaffer 2013. 196 Dahl 1957; Krasner 1991; Moe 2005. 197 Duffield 2003, p.417-18.  88   afterthought.198 This contrasts sharply with other approaches to the study of IR, above all Realism, which suggest that the exercise of power is pervasive in world politics and must therefore be placed at the core of any explanation of interstate behaviour. If true, by failing to take power sufficiently seriously, theories of rational design risk restricting their applicability to a limited set of circumstances: “those in which actors either lack, or are unwilling to exploit, capabilities to coerce or to offer inducements for the purpose of obtaining institutions that better serve their individual interest.”199   In sum, functionalist theories have a long lineage in IR, have been used to explain a range of institutional features, and efforts have been made to develop conjectures to explain the emergence of formal and informal organizations. In doing so, however, they have omitted a number of factors that have been found to be important by a great many others, and which may have important implications for our understanding of organizational form in particular. Ultimately, this observation should not be taken to suggest that their arguments are obviously wrong; they may explain institutional design quite effectively in areas where their scope conditions hold, and they may be an important part of the story behind the rise of informal organizations. The point is that they may be weak and incomplete in their present form, and that making an effort to incorporate what they leave out might help to explain the kinds of empirical puzzles described above.    !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!198 Consider, for instance, the discussion of power in Vabulas and Snidal 2013. 199 Duffield 2003.  89   3.3 A Liberal Theory of Organizational Form  In this section, I attempt to improve upon existing accounts by incorporating domestic politics, distributional conflict and power into a theory of organizational form. In doing so, I couch my argument within the “liberal” tradition of IR. Like functionalism, liberalism also has a long history in the field that generally traces its origins to classic works by Niccolo Machiavelli, Immanuel Kant, and Joseph Schumpeter.200 It has been developed over the past 30 years into a “scientific” theory of international politics that is distinguished by the fact that it grounds explanations of international relations in the actions of individuals—broadly construed—rather than states.201 Liberal theories begin by modeling interactions between domestic actors as they compete to advance their policy preferences, show how these internal conflicts are filtered through domestic institutions, and then explain how this process, which can play out very differently across polities, then changes the way that states behave in relation to each other. Such models have helped to generate some of the most profound findings in the field of IR, such as the democratic peace, and underpin the most prevalent theoretical approach in the field of International Political Economy, Open Economy Politics.202 It has therefore been used quite successfully to explain a wide variety of aspects of international cooperation and even institutional design, but it has yet to be applied in any rigorous fashion to questions about the formality of intergovernmental agreements and organizations.203   !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!200 Doyle 1986; Zacher and Matthew 1995. 201 Moravcsik 1997. Also see Putnam 1988; Evans, Jacobson, and Putnam 1993. 202 Ray 1998; Lake 2009. 203 See, for example, the discussion of this point in Pollack and Shaffer 2013. Some of the most prominent examples of liberal theories include Rogowski 1990; Moravcsik 1998; Milner 1997; DeSombre 2000; Mansfield and Milner 2012. For an example of an effort that applies similar insights to new modes of governance, see Roger, Andonova and Hale 2015.  90   As elsewhere, the liberal theory that I advance proceeds in two steps.204 In the first, states decide to cooperate through an organization and develop preferences over the form that they would like it to take. The model begins with an exogenous rise in “demand” for governance. This demand provides states with an incentive to cooperate with one another but, in doing so, also “selects” certain domestic public actors to initiate or lead cooperative efforts on the behalf of a state. Different kinds of public actors can be selected that are distinguished by their positions within a domestic institutional structure, and their “identity” matters because each will have different “baseline” preferences over organizational form that then get “translated” into state preferences, given their first-mover advantage. These actors are also each subject to competing political pressures that can sometimes modify their tactics and thereby alter a state’s preference under certain conditions. The second step of the model then explains how these preferences “aggregate” to produce final institutional outcomes. The preference formation process briefly described above can play out in different ways across states, which means that preferences can often conflict to varying degrees, even within the same issue area. One state may prefer a high level of formality while another does not. When this is the case, it is impossible for all states to attain their first preferences. And, for an organization to emerge, these distributional conflicts must be resolved somehow. This preference aggregation process is modeled as a bargaining game, where states negotiate over positions along a formal/informal continuum. The final outcome—the form of the organization that emerges in a given issue area—is a function of the distribution of preferences and power.    !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!204 Legro 1996; Moravcsik 1997.  91   3.3.1 Step 1: A Model of State Preference Formation  As noted, I begin my theory of preference formation with an account of the “demand” for governance. Without some kind of public “need” for cooperation, there would seem to be little rational reason to supply cross-border governance of any kind. But where does demand for governance come from? In the main, and in line with other liberal theories of IR, I argue that demand for cross-border cooperation arises when there are externalities that result from various social, political or economic transactions between and within polities. These tend to increase as globalization takes place.205 As economies integrate, activities within or between one or more states tend to have more significant effects that others would like to contain or limit, and can only be reduced by mutual agreement amongst those involved. 206 In many cases, it may be possible for actors to strike such agreements without much trouble, particularly if the number of actors is small and if it is easy to detect violations of the agreement.207 But with rising transaction costs—“the costs of measuring the valuable attributes of what is being exchanged and the costs of protecting rights and policing and enforcing agreements”208—it becomes more difficult to do so. When this is the case, more complex institutional arrangements, including intergovernmental organizations, can be created that can help to (at least partially) reduce transaction costs and thereby make it possible for states to make mutually beneficial policy adjustments.209   !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!205 Ibid.  206 Held 2000; Held and McGrew 2007. 207 Coase 1960; Axelrod 1984. 208 North 1990, p.27. 209 Keohane 1982; Keohane 1984.  92   Such externalities therefore increase the demand for governance but they may not be sufficient on their own. In some cases, the externality may fall directly on a public actor, negatively affecting its ability to conduct its affairs domestically. In this case, the public actor (say, a ministry) will have a strong interest in pursuing cooperation without any action from societal actors. And, if transaction costs involved in reaching an agreement are sufficient, they may aim to achieve cooperation through an intergovernmental organization. In other cases, however, societal actors must successfully lobby or agitate for a government to take action. This may, of course, require societal actors to resolve collective action problems of their own.210 In general, if the benefits from such actions are diffuse, then no one in society will have an interest in shouldering the costs of lobbying or agitating. At the very least, it will then be challenging to make their interests known to public actors. If, by contrast, the benefits are concentrated, then societal actors will have a strong incentive to make their demands felt. Either way, only when societal actors are able to effectively lobby or agitate a public actor will the latter then take up the issue and attempt to craft a cross-border solution. Thus, with respect to demand for governance much depends on the particular features of the issue in question.  But “demand” also has another dimension to it: the kind of demand “process” described above not only encourages states to engage in cross-border cooperation, it also “selects” domestic institutions to lead cooperative efforts on the behalf of a state. States are not unitary entities but are in fact composed of a wide variety of institutions with different kinds of expertise, levels of autonomy and spheres of authority.211 And, for any given issue, only a few of these institutions are likely to be selected to represent the state in international fora. If, for example, a new !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!210 Olson 1965; Wilson 1982; Tarrow 2011. 211 Posner 2005.  93   externality falls primarily on non-state actors, this will lead to societal demands being placed on a specific public actor—usually a public actor that is ostensibly “responsible” for the issue in question. An industry association may, for instance, choose to lobby the US Trade Representative (USTR) to pursue a particular trade agreement, since the USTR’s bailiwick comprises trade issues. If the public actor is responsive to these demands, then it will in turn initiate or “lead” cross-border cooperation and play a critical role in formulating a state’s approach to any subsequent negotiations. Alternatively, an externality may directly affect a particular public actor’s ability to conduct its own domestic affairs, as described above, which will also lead it to call for cross-border cooperation to bolster its capacity. It will do so, in this case, without any need for lobbying by societal actors—but it is “selected” all the same. An exogenous change in technology may, for example, negatively affect a banking regulator’s ability to ensure the stability of its domestic financial system when a banking crisis occurs in another state. It would then attempt to cooperate with homologues abroad to bolster its capacity, and, given its centrality, expertise and first-mover advantage relative to other domestic actors, will typically be granted significant authority over a state’s approach to the negotiations. Of course, responsibility is not always so clearly vested in an institution and others may play a role behind the scenes—after all, delegations will typically involve multiple different actors—but in the final analysis only a small number of institutions will generally take the lead in crafting a state’s position on a particular issue.    Why do we care who leads? The selection process matters because different kinds of actors within the state are likely to have different “baseline” preferences over organizational form. To simplify, I assume that there are two kinds of public actors that can be “selected”: autonomous  94   agencies run by bureaucrats and more traditional, cabinet-level institutions run by elected officials. I refer to these as “independent agencies” and “political institutions,”  respectively.212 A political institution is defined here as a governmental entity whose authority is ostensibly derived from “the people,” and either directly elected or directly managed by elected officials. Examples include the UK FCO or the US Departments of State and Commerce, which are cabinet-level bodies that are directly accountable to the Prime Minister and President, respectively, and frequently staffed by elected officials. Independent agencies, by contrast, are defined as governmental entities that “(a) possess and exercise some grant of public authority, separate from that of other institutions, but (b) are neither elected by the people, nor directly managed by elected officials.”213 These include the growing number of “autonomous” regulatory agencies run by unelected bureaucrats who may be appointed by elected officials but serve definite terms in office and are granted significant policymaking independence.214 Examples include the US Securities and Exchange Commission, the Bank of England in the UK, or the Canadian Competition Bureau. Both political institutions and independent agencies are public, governmental entities; they differ, primarily, in terms of the degree of autonomy that they !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!212 This distinction parallels that made by Thatcher and Stone Sweet between “majoritarian” and “non-majoritarian” institutions, see Thatcher and Stone Sweet 2003. I have chosen not to adopt this terminology because, while capturing something real and important, it is potentially misleading due to the use of the term “majoritarian,” which has a somewhat different meaning in other contexts. The distinction between independent agencies and political institutions is more intuitive and avoids any confusion. 213 Ibid, p.2. 214 It could also include semi-autonomous agencies within PIs. An example of this phenomenon is the National Institute of Standards and Technology (NIST), which is semi-autonomous agency located within the US Department of Commerce. On this phenomenon more generally, see Pollitt et al. 2005. The degree of autonomy that an agency enjoys will often be an empirical matter, and may be de facto or de sure.   95   possess. While the former are managed by elected politicians and beholden to electoral and special interest politics, the latter are—by design—more insulated from these forces.215   The degree of autonomy that an institution possesses is important because there are reasons to believe that this shapes their preferences over the formality of international cooperation, as a number of scholars have noted, and that this is likely to then shape a state’s approach to institutional design.216 If an independent agency is “selected,” for example, we can reasonably expect it to strongly prefer greater informality. This is because independent regulators highly value the autonomy they already possess and will likely take precautions to limit the extent to which other domestic actors intervene in their affairs.217 To the greatest extent possible, they like to maintain strict control over information and their activities and wish to rely upon their own expertise and professional codes for decision-making. Creating and working through a formal organization may therefore be particularly disliked as a means for resolving cross-border issues, when they arise, since this requires a high level of involvement on the part of other actors—legislators, lawyers, diplomats, etc.—who may not share the same professional codes and interests. Above all, they loathe the introduction of “politics” into their affairs, which involvement by a legislature would typically entail. An informal organization, by contrast, does not require extensive ratification procedures, monitoring arrangements and budgetary allocations, as discussed in the previous chapter. Theoretically, this then allows independent !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!215 Of course, these are ideal types; in fact, autonomy is never absolute and there will likely be variation in terms of institutions exposure to the political process across states and within states over time. Much depends on the particular institutional arrangements that prevail at any given point. It is also worth noting that autonomy itself may be formal or informal. A given agency may have a high level of (formal) statutory autonomy, but it is also possible that an agency will have low statutory autonomy but a high level of informal autonomy, see Thatcher 2003. 216 Abbott and Snidal 2000; Bach 2010; Singer 2007. Also see Deflam 2002 and Whytock 2005. 217 Miller 2000; Singer 2007; Bach 2010.  96   agencies to both engage in intergovernmental cooperation to reduce externalities they face and maintain a higher level of control over their affairs than would otherwise be possible. An independent agency will, therefore, tend to seek informal cooperation if it can.   Political institutions, by contrast, are more likely to have a baseline preference for formal intergovernmental organizations. There are at least two reasons why this is the case. First, as Anthony Aust has explained, politicians prefer to sign treaties rather than MOU due to their electoral benefits.218 When politicians are actively involved in international affairs they may look down upon informal solutions to international problems because they tend to be less visible domestically. Signing a “non-binding” MOU or creating an informal organization is much less prestigious—signaling weakness rather than strength to domestic constituencies—and therefore brings fewer political benefits to politicians, who worry about reelection and their impressions amongst the electorate in ways that unelected bureaucrats do not. Second, politicians (and their constituencies) may worry about policy instability over time and attempt to “lock in” participation in an organization by engaging in more formal cooperation. As in the domestic sphere, politicians may be genuinely worried that those holding their place in the future may not share their policy preferences for intergovernmental cooperation, and they may then prefer greater formality since the more intensive procedures involved and potential for sanctions make it more costly for future politicians to roll-back participation.219  Ratification and greater formality, in this instance, is a major benefit. Independent agencies, by contrast, tend to be less !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!218 Aust 2000, p.38. Emphasis added. 219 Moravcsik 2000a; Brewster 2004. For the domestic case, see Moe 1995.  97   worried about locking in participation, since leaders’ positions are more secure and those holding their positions in the future are more likely to share the same professional codes.220   It should be immediately noted, however, that the baseline preferences of political institutions and independent agencies are viewed here as “tendencies,” and not hard “laws.” Often, we in fact see political institutions involved in the creation of informal organizations and independent agencies participating in formal ones. But why is this so? The argument here is that this puzzling behaviour is also deeply influenced by domestic politics. For instance, it may be the case that independent agencies are unable to maintain the level of autonomy they would otherwise prefer because an issue becomes “politicized” or because they have failed to fulfill their domestic mandate somehow. When this occurs other domestic actors may intervene or threaten to intervene in the policymaking process and make it impossible for an independent agency to achieve its initial preference. That is, if issues that are traditionally within the remit of an independent agency come to the attention of legislators or heads of state as a result of domestic distributional conflicts arising from agency policymaking, they may become more deeply involved in leading a state’s approach to an issue and take the kinds of ambitious actions that will please their constituencies. Another path to formality may occur if foreign actors interfere in the domestic affairs of an agency, perhaps through the extraterritorial enforcement of their laws, or because another state has chosen to handle an issue at the ministerial level. When this occurs, the policymaking authority of an agency is directly threatened and this may also prompt interventions from other domestic actors that may lead to a different outcome than we would otherwise expect if an independent agency was acting on its own.  !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!220 Miller 2000.  98    Similarly, a political institution may pursue informal cooperation under certain circumstances. This may occur, for instance, when a state wishes to cooperate with others to establish an intergovernmental organization, but faces significant domestic constraints that make a formal organization impossible.221 Politicians may prefer formal solutions, but they come at a high price compared to informal ones, largely as a result of their more extensive domestic implications. To make them work, leaders must gain a significant degree of support from other governmental bodies—especially a legislature—and from society in a broad sense. This requirement means that other domestic actors have, ceteris paribus, more opportunities to stymie formal cooperation. The form of cooperation will therefore depend to a great extent on a government’s ability to attain the domestic support necessary. One way to think about this is in terms of veto player theory.222 Establishing a formal organization can be conceptualized as a significant move from the “status quo,” since it generally involves far more domestic changes to attain the associated benefits. It should therefore be easier for a government to pursue formal cooperation when the number of veto players is low. A larger numbers of veto players makes it more difficult for politicians to move away from the status quo. Informal organizations, by contrast, involve fewer changes to the status quo and are less open to various legal and political challenges, and should therefore become more likely under such circumstances. By opting for an informal organization, a politician will be able to satisfy at least some of those demanding cooperation. It is second best, !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!221 Other possible conjectures may be developed in the future. At present, my theory does not intend to explain all instances where a political institution chooses to create an informal organization. Rather, it develops a hypothesis about one causal pathway. This means that some variation will inevitably go unexplained, but if I am able to improve upon existing theories (i.e. explain more of the variation than they have) then this will nevertheless constitute an important theoretical and empirical advance. 222 Tsebelis 2002; Jahn 2010.  99   but better than nothing from their perspective. Further, it is possible that when domestic conditions improve a formal organization may become possible in the future.  In summary, preference formation—the first explanatory step—is a function of demand and domestic politics. The process is visualized in Figure 2. Here, demand is essentially the realm of special interest politics: a global problem leads certain societal actors to push for international cooperation. It may also prompt government actors to become directly involved if an externality falls directly upon them, i.e. without lobbying or agitation. Either way, demand creates a need for an organization, but also “selects” domestic actors to lead cooperation. The domestic actors—political institutions and independent agencies—that are selected in this way then have certain baseline preferences over organizational form as a result of their position within a domestic institutional structure. If the actor that is selected has considerable autonomy and is more insulated from electoral politics, then we can expect informal cooperation to be pursued insofar as this is possible, as this allows them to achieve cooperation but maintain autonomy. If an actor has less autonomy and is more beholden to electoral politics, then they are likely to prefer a more formal organizational form, since this is more likely to lock in policy and support their goals of reelection. The solid lines indicate these baseline preferences. However, under certain conditions state preferences are reversed, as shown by the broken lines. If a more autonomous actor is selected, but their work becomes “politicized” by domestic or foreign actors and their autonomy or authority is directly threatened, then others may intervene or play a more prominent role in the negotiations, making formal cooperation a more likely outcome. Likewise, if a political actor is selected, but they simultaneously confront significant domestic constraints that make it difficult  100   to achieve the kind of domestic “buy in” that is necessary for more formal cooperation, an informal organization becomes more likely.  Figure 2 Model of the Preference Formation Process   3.3.2 Step 2: A Model of State Preference Aggregation  Thus far, the theory I have presented does not explain actual outcomes—whether a formal or informal organization does, in fact, emerge—only state preferences over outcomes. It provides a general model for understanding how domestic politics and institutions can shape state preferences over the legal form of an intergovernmental organization. This preference formation process may play out very differently across states, depending upon the way that domestic institutions have evolved over time, as well as the particular configuration of political actors active within a state at the specific point in time that it seeks to engage in intergovernmental Political(Institution,(e.g..(Dept(of(State Independent(Agency,(e.g.(SEC,(FTC. Cooperation(Problem Formal(Organization Domestic(Constraints Politicization Informal(Organization  101   cooperation. Independent agencies, for example, are an increasingly common feature of states, appearing in a wide range of sectors, but considerable institutional diversity remains within each sector. In some states, responsibility for regulating energy or telecommunications has been delegated to an independent regulatory body, while in others it remains with a political institution, such as a cabinet-level ministry; institutional design at the domestic level—as at the international level—is not simply a function of the problems states seek to solve, but depends upon a range of contextual factors, such as state traditions, institutional structures and political leadership.223  Since this is the case, it is reasonable to expect that the same type of political actor will not be leading cooperation within all of the states involved in creating a new organization and therefore states will not necessarily share the same preference over organizational form. Equally, there are a wide variety of political systems and changing constellations of actors within those systems that, taken together, may make it more or less difficult for actors in different states to pursue certain forms of cooperation.224 Given the different constraints that policymakers face at any one point in time, some may prefer to create a formal organization while others may choose a more informal approach.  As these observations suggest, it is possible that when a group of states seeks to create an intergovernmental organization, their preferences over organizational formality may conflict—often significantly so. It is also possible that states may hold the same preference, but their reasons for doing so may be very different. In order to understand the final outcome that is likely to emerge in a given context we must therefore have some account of how states preferences are likely to “aggregate.” Drawing on bargaining theory and other work in the liberal tradition, I !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!223 Pollit et al. 2001; Thatcher 2003. 224 Lijphart 1999.  102   argue that outcomes in the preference aggregation process are likely to depend on two variables: 1) the distribution of preferences, and 2) the distribution of power. The first variable refers to the extent to which preference are congruent or conflictual. Obviously, if preferences do not conflict then there is little question of which organizational form will be chosen. If states agree that an informal organizational structure is best, then we should expect an informal organization to emerge. However, if state preferences conflict, then in order to understand the actual outcomes that emerge we must explain whose preference is likely to prevail. This is where the second variable—power—comes in. To determine which kind of organization is likely arise when preferences conflict, we must examine not only who holds what preference but also who has the power to impose their preference on others.   To develop these insights more fully, we can begin to think about the preference aggregation process by first arranging state preferences along a single-dimensional bargaining continuum, as depicted in Figure 3. The particular arrangement tells us about the distribution of preferences in a particular case. In this diagram there are two states, A and B, that have decided to create an organization of some kind and they have each previously formed preferences over the formality of the organization in the prior preference formation stage, as described in the previous section. The continuum represents different institutional designs and ranges from mostly informal to mostly formal. An organizational design on the far left hand side is highly informal and is the option that is closest to the status quo (SQ) in which there is no organization; as we move along the continuum to the right, organizational designs become more and more formal and further  103   away from the status quo. 225  Both A and B are assumed to behave instrumentally and prefer points along the continuum that are closest to their ideal points, labeled a and b. Once these points are established, the two states then bargain over the design of the organization. Obviously, this model is a simplification. It suggests that states only care about a single issue—an organization’s level of formality or informality—when of course they actually care about many issues, and their bargaining on these issues may affect their bargaining related to the form of an organization. However, in this way, we can visualize the distribution of preferences and develop hypotheses about the kinds of organizations likely to emerge in a relatively straightforward fashion.       !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!225 One might argue that modeling the bargaining space as a continuum is incongruous with my conceptualization of formality as essentially dichotomous in nature. Further, there are points in the text were I speak of “more formality” or “less formality,” suggesting that organizational form is in fact a continuous variable, which again may seem inconsistent with the exposition thus far. There is some truth in this, since legal obligation is indeed largely dichotomous, see Raustiala 2005. It might, therefore, make sense to model the bargaining scenario as a “battle of the sexes” game where states agree on a set of two Pareto-improving outcomes but disagree over which outcome is preferable. The result would, I argue, be essentially the same. However, thinking about formality as a continuous variable might also not be as problematic as it appears at first. Others have, for instance, argued that legalization is, in fact, a matter of degree, since it involves other dimensions than legal obligation, such as precision and delegation, see Abbott et al. 2000. Moreover, to some extent, my conceptualization of organizational form has a “delegation” dimension, since the extent of institutionalization among both formal and informal organizations may vary, as discussed in the last chapter. Some informal organizations, for instance, involve some delegation (as with APEC) while virtual organizations involve very little (as with the G20). Precision may also be relevant as well, although it is not discussed in detail here. Nevertheless, since this is the case, it may make sense to think about the bargaining scenario as a continuum within the conceptual and theoretical framework of this study, even if legal obligation is itself dichotomous and given analytical priority here.   104   Figure 3 Visualizing the Distribution of State Preferences   The diagram in the upper half of Figure 3 depicts a scenario in which states A and B have “convergent” preferences. They are convergent in the sense that both ideal points are relatively close together. This means that although there may still be some scope for bargaining over the particular design that will be chosen (since they are not perfectly aligned) the two states generally agree on the level formality that they desire. The particular end of the continuum that the two ideal points converge on indicates the kind of organization that we should expect to arise. In this case, two ideal points are both on the far right hand side of the figure, which tells us that both states prefer a more formal intergovernmental organization and that we should expect the final outcome to reflect their shared preference. The diagram in the lower half of Figure 3 depicts a more complicated scenario in which states A and B have “divergent” preferences. They are divergent in the sense that, while both may agree upon the need for an intergovernmental organization of some kind and therefore want to move away from the status quo, they disagree (((More(Formal((SQ( a( b(a( b((SQ(Convergent(Preferences(!!Divergent(Preferences(More(Informal( 105   over the form of the organization—how far they wish to move away from the status quo. The scope of their disagreement is signaled by fact that the two ideal points are much further apart than in the previous scenario. When this is the case, it is more difficult to say what form it is likely to take.226 Will the final design be more towards the informal end of the continuum, as preferred by A, or will it be on the formal end, as preferred by B? Unfortunately, in this case, the distribution of preferences alone does not tell us much about the final outcome we can expect. To do so, we must look towards another variable: the distribution of power.  Power is a challenging term to define precisely, particularly because it can take such a wide variety of forms.227 But, at its most basic level, it can fruitfully be conceptualized as an ability to impose one’s will on another. It exists when one actor, say, A, is able to get another, B, to do what it would not otherwise prefer to do.228 As such, it is an inherently relational concept. It is not simply a property of A that allows it to impose its will on B as it sees fit, but is derived from a particular kind of relationship that exists between the two, as well as the resources at their disposal. Much depends on the specific issue in question (A may have power over B with respect to j but not k) and the particular point in time that an actor attempts to influence another (A may have power over B with respect to j at time t but not t+1). These observations tell us something about what power is. But in order to develop a theory of power that can be used for our purposes it is necessary to be more specific about the precise sources of power that states are able to !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!226 In fact, it is even possible that no organization emerges in this instance because the scope of disagreement is so wide. See Tsebelis 2002. 227 Barnett and Duvall 2005; Lukes 2005; Baldwin 2016. It is often cited as an example of an “essentially contested concept,” Gallie 1955. 228 Dahl 1957.  106   leverage in order to get what they want in a particular scenario.229 Towards this end, it is possible to gain insights into the relevant kinds of power that states can utilize by drawing on findings from general theories of bargaining.230 Specifically, in line with these theories, I argue that there are two relevant sources of power that states may exert: first, power may be derived simply from the structure of the bargaining scenario, which can in some cases give an advantage to a particular actor; second power may come from sources “outside” the game and make it possible for actors to reshape the very structure of the bargaining scenario itself. I discuss each source of power in what follows.  First, consider the kind of power that can be exerted “within” a bargaining game. Models developed by John Nash and Ariel Rubinstein, as well as others that have built upon their work, tell us that when actors bargain over the distribution of additional benefits to be derived from some agreement, the final outcome often depends on the quality of actors’ outside options.231 An easy way to think about outside options is in terms of each state’s “best alternative to a negotiated agreement,” or BATNA.232 States with better BATNAs than others are, by definition, more comfortable with the prevailing situation then their negotiating partners. Their preference is, in other words, closer to the status quo. When this is the case, it is argued, this gives them significant bargaining power. The reason being that, when an actor has a better BATNA, they can more credibly threaten to veto an agreement if they do not get their way. Actors that have comparably worse alternatives to no agreement are, by contrast, less able to credibly threaten to veto a deal, and correspondingly more willing to make compromises to ensure that an !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!229 Ibid.; Baldwin 2016. 230 Muthoo 1999. Also see Dixit and Skeath 2004. 231 Nash 1950; Rubinstein 1982; Muthoo 1999.  232 This is the term coined by the Harvard Negotiation Project.  107   agreement—any agreement—succeeds. The basic hypothesis is, therefore, that if an actor’s preference is closer to the status quo then they will have greater ability to determine the final bargaining outcome. To illustrate, a scenario of this kind is depicted in the top panel of Figure 4. There, the ideal point of state A is much closer to SQ, which means that in comparison with state B it is relatively more comfortable with things as they are. It wants to move away from the status quo, but only so far—and much less so than B. This should, if B is unable to take any further actions, give A a bargaining advantage, since A is more capable of threatening to hold out and potentially even scuttle any agreement that is not close to its ideal point. As a result, B is more likely to defer to a point closer to state A’s ideal point, a, since it prefers a to SQ. In this instance, then, we should expect the outcome of their bargaining to be an organization with a more informal design.  Figure 4 Bargaining Scenarios With Divergent Preferences   (SQ a( b(a( (b((SQ(More(Informal( More(Formal(((P1( 108   Note that the above observations suggest that states preferring an informal design should have an inherent bargaining advantage over those that prefer a more formal one. This is because an informal organizational design is, by its very nature, much closer to the status quo, which means that a state preferring an informal design can more credibly “veto” moves towards greater formality. However, this implication only follows if others are unable to take actions that can affect the structure of the bargaining game. This may not be the case, particularly if the actors have other sources of power at their disposal than the structure of the bargaining game itself—a situation that is likely to occur quite frequently in the real world. This is where the second source of power comes in. A second proposition from bargaining theory related to the one above is that, if a party can take actions that have the effect of changing another’s outside option, or BATNA, then they are generally able to fare better in negotiations. One way to achieve this is to threaten to take unilateral actions that would have the effect of removing the status quo ante as an option—shaping the choice set of their bargaining partner.233 Such actions may include threats to use sanctions, threats shut off market access, side payments or a state’s ability to “go it alone” with others, each of which are likely to depend on factors that are extraneous to the bargaining scenario itself.234 An economically larger or more militarily powerful state, for instance, may be able to leverage another’s dependence on its market or resources to get what it wants, even if the other’s preference is closer to the status quo ante, by threatening to close off access. If so, this enables the larger state to effectively undermine the favorable position that their bargaining partner would otherwise have had, and thereby reduce the credibility of their threats to veto a deal. This kind of scenario is depicted in the bottom panel of Figure 4. There, state B has made a !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!233 Hirschman 1980; Lax and Sebenius 1985; Richards 1999; Gruber 2000; Newman and Posner 2011. 234 Martin 1994; Gruber 2000; Kim 2010.  109   threat—say, threatening to shut off access to a market that state B is dependent on235—that has the potential to shift the status quo ante, SQ, to a new point (P1) that is much closer to its preference. Since P1 is closer to its ideal point, b, the threat is credible. And, by doing this, B leaves A with a choice between b and P1. Since A prefers b to P1, this make it much more likely that it will defer to B’s ideal point. Here, this means that a more formal organization is expect to emerge, in line with B’s preference.  In sum, my theory of preference aggregation hypothesizes that the final outcome (i.e. whether a formal or informal organization does in fact emerge) depends on two variables: the distribution of preferences and the distribution of power. When preferences converge, then the outcome is expected to reflect the common preference of all relevant states. By contrast, when preferences diverge, the distribution of preferences on its own tells us little about the type of organization that will emerge. Instead, power becomes an important variable in this context, and is derived from outside options, or BATNAs, and actors’ abilities to affect others’ outside options through actions that have the effect of shaping the choice set and creating a new bargaining structure altogether. If one state or a group of states has power over others, then the final outcome is predicted to reflect their preference.  3.4 Conclusion   In this chapter, I have focused on providing an answer to the question at the heart of this study: why are some intergovernmental organizations formal and others informal? I began with a detailed overview of functionalist theories, which are the prevailing approach to explaining !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!235 Hirschman 1980; Newman and Posner 2011.  110   organizational form. Such theories argue that states “design” international institutions in a rational manner by attempting to “match” institutional designs with the particular cooperation problems that they face. They wager that we should therefore see a systematic relationship between particular designs and the types of cooperation problems states face. However, I have shown that this is not always a good bet—at least when it comes to organizational form—as we find puzzling variation across organizations operating in similar issue areas and across different regions within the same issue area. Further, I have argued that functionalist theories leave out a number of important variables that have been found to be important by a large number of other scholars, and which may be particularly relevant for explaining choices about organizational form: domestic politics, distributional conflict and power. By incorporating these, I wager, we might be able to explain away some of the empirical puzzles that plague purely functionalist approaches.   The approach to doing so that I have advanced is couched in the liberal tradition of IR. This approach begins by explicitly theorizing the way that states form preferences over outcomes and then modeling how these preferences aggregate to produce the actual outcomes we see in the world. In the first of these steps, a rise in demand compels states to engage in intergovernmental cooperation but also “selects” specific bodies within those states to “lead” efforts to achieve this. Different bodies may be “selected,” which are differentiated by their positions within a domestic institutional structure. Each may care about solving the underlying problem they face and want to design appropriate mechanisms, but they are also conscious of the domestic implications of their design choices. Given that different institutional designs have different effects for these actors, they are each likely to have different baseline preferences over organizational form.  111   When selected, they would therefore prefer organizations in line with these preferences, all else being equal. But all else is not always equal: under certain conditions, these actors may not be able to achieve their first preferences and will therefore pursue different institutional designs. The second step then takes these state preferences over organizational form as given and explains how they aggregate to produce the final outcomes we see. Drawing on insights from bargaining theory, I have argued that the final institutional designs that we see are determined by two variables: the distribution of preferences and the distribution power. In some cases, the distribution of preferences alone can tell us the kind of organization to expect; in others, we must take into account not only the preferences of actors but also their relative capacities to impose their preferences on others.  The objective of the subsequent chapters is to provide more formal tests of each of these theories. However, before doing so, it is helpful to state some of my theory’s scope conditions, since these shape the kinds of the cases we choose to examine to test it. Two are especially important. First, the theory is not expected to apply across all time periods but should hold primarily for the period after the end of the Second World War. This is because it is only during this period that a set of practices—both international and domestic—crystallized that could define precisely what it meant to be a “formal” intergovernmental agreement or organization. Prior to this time, the practices associated with international treaty-making and intergovernmental organizations were inchoate and ill-defined. 236 It was only in was only in 1969 in fact that the Vienna Convention on the Law of Treaties was concluded and finally settled to some degree what would count as an international treaty and what would not. And, while river !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!236 Denemark and Hoffmann 2008.  112   commissions and administrative unions appeared with increased frequency in the nineteenth century, the precise legal nature of these organizations only started to be seriously resolved in the interwar period.237 Given that the concept of informality can only have meaning once the corresponding practices defining “formality” have been established, it makes sense to limit the use of the two concepts to periods wherein their “borders” are clear. While it is impossible for us to specify a precise date when this occurred, it seems plausible to say that most of the relevant practices were more or less settled by the start of the post-war period and that as we move further back in time from this point, the scope conditions for the theory will quickly be exceeded.238  Secondly, the theory is not expected to hold for all states but will primarily be applicable to relatively well-governed democracies. The rationale for this scope condition is that the model of domestic politics assumes that political actors are accountable to an electorate and that institutions delegated a degree of formal autonomy are in fact able to hold preferences that are indeed independent from others. If political actors are not accountable to an electorate in some sense, then it is unlikely that the arguments made about the baseline preferences of political institutions will hold—or, at least, may not hold in quite the way depicted above, suggesting that some modification of the theory is likely to be needed. Dictators do not need to be seen to please domestic constituencies in quite the same way and they are not concerned about policy instability, since their positions are secure as long as they hold power. Further, if political actors cannot in any significant sense be constrained by other veto players within a state, then my !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!237 Bederman 1996; Slaughter 1993; Klabbers 2001b; Klabbers 2009. 238 For some empirical justification for this dating, it is also worthwhile to recall that, according to Aust, the practice of concluding informal international agreements “appears to have started in a small way after the Second World War, three being concluded in the 1950s in connection with the Treaty of Peace with Italy,” see Aust 2000, p.20-21.  113   hypothesis about the effects of constraints on preferences for organizational form will not be relevant. Likewise, in authoritarian states, or poorly governed democracies, it is also unlikely that formally autonomous bodies are actually able to hold independent preferences from other political actors, since the laws establishing their autonomy may be poorly enforced. In many developing states, for instance, where authoritarian regimes are generally more common, formally autonomous central banks are not generally able exercise this autonomy because relevant laws are routinely violated.239 When this is the case, the arguments made above about the baseline preferences of independent agencies will not hold. As we move beyond the set of relatively well-governed democracies, therefore, the model that I have elaborated will be less and less applicable, and we will likely need to develop new approaches to explaining the behaviour of these states.  !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!239 Indeed, many of the effects that scholars and policymakers expect to be associated with formal central bank independence, such as low inflation, do not generally hold in developing states, which they generally attribute to poor enforcement of the laws protecting their autonomy. See the review of the literature in Eijffinger and Haan 1996. For a related discussion focusing on differences in the independence of competition agencies across developing and industrialized countries, see Jenny 2012.   114   Chapter 4. Analyzing the Determinants of Organizational Form  4.1 Introduction  In the previous chapter, I discussed two approaches to explaining the form of an intergovernmental organization. The first, most common type of explanation is functionalist in nature. It hypothesizes that states design organizations in view of the underlying cooperation problems that they collectively confront when trying to solve some issue. States should be expected to create informal organizations when the specific kind of problem that they face makes doing so advantageous. However, I have argued that prima facie such theories do not appear to perform well and that functionalists’ neglect of domestic politics, distributional conflict and power might be at the root of this problem. The second type—my own—seeks to provide a different account that explicitly incorporates these factors into a theory of organizational form. This theory is rooted in the liberal tradition of IR and points to how the domestic impacts of different design choices shape state preferences over different kinds of cooperation and how these preferences are, in turn, translated into final organizational outcomes through a process of interstate bargaining. But