The coral reefs of global governance: how formal IOs make informality work

ABSTRACT

Why have states increasingly relied on informal IOs as instruments of cooperation? Most research on this question has focused on identifying incentives that lead states to establish informal bodies. But attention is turning to ‘supply-side’ factors – those making it easier for informal IOs to operate. I advance this strand of research by explaining how formal IOs make informality work. Specifically, I theorize how the presence of formal institutions in a region or issue area – particularly, those capable of providing key resources and services to informal IOs – changes calculations about their viability and leads policymakers to embrace informality more readily. Support for this argument is provided through an analysis of formal IO assistance to informal bodies, and case studies of the EU’s impact on two organizations – COST and EUREKA – to unpack the mechanisms at work. Overall, the article offers insights into the drivers of informality, presents original data on inter-institutional interactions, and speaks to broader arguments about the EU, organizational ecology, and regime complexity.

KEYWORDS:

• Informality
• European Union
• organizational ecology
• regime complexity
• institutional interactions
• international organizations

1. Introduction

Informality has become a prominent feature of contemporary global governance. Increasingly, when states set down international rules and provide public goods, they have relied less on treaties and relatively more on soft law, which deliberates eschews legally binding obligations (Pauwelyn, Wessel, and Wouters 2012). Similarly, across many issue areas, in the place of formal international organizations (IOs), like the United Nations (UN) and the World Trade Organization (WTO), there has been a proliferation of informal IOs that exist in a ‘twilight zone’ beyond the traditional boundaries of international law (Vabulas and Snidal 2013; Roger 2020). Within Europe, this shift has been especially pronounced. European states make extensive use of informal instruments across numerous domains (Kleine 2013; Bickerton, Puetter, and Hodson 2015; van Heumen and Roos 2019). And, within the otherwise high-legalized integration and foreign policy processes of the European Union (EU), in particular, traditionally informal bodies like the European Council, the Eurogroup, and the E3 have become particularly important.

This shift has attracted considerable attention from International Relations (IR) scholars, and the theories offered to explain it have been quite diverse. Some have demonstrated how certain ‘problem structures’ and conflicting state preferences can generate an incentive for them to establish informal institutions (Stone 2013). When issues require speedy decision-making, flexible implementation, or confidentiality, for instance, informality appears to have certain advantages. The presence of such problems is therefore expected to increase the likelihood that states will adopt an informal approach. Others have looked within states, observing how the involvement of independent agencies in negotiations and the political constraints policymakers often confront can lead them to embrace informality too (Roger 2020; Slaughter 2004). Together, existing accounts suggest that globalization, growing tensions, domestic polarization, and the broader transformations associated with the rise of the regulatory state have shifted ‘demand’ for informal bodies, leading them to be created more frequently.

Through such studies, our understanding of the origins of informality has undoubtedly advanced. Yet, as Manulak and Snidal (2021) have observed, current accounts only tell part of the story, since they neglect the related question of ‘supply’ – the factors that have made it easier for informal IOs to operate. Their argument is that the growth of informality is not only a product of changes that increase the incentive for states to cooperate informally, but that the cost of supplying governance through informal IOs has also fallen, making such bodies more viable solutions – and formal IOs less necessary – than they were. Their analysis shows that this shift has been partly driven by the declining costs of communicating and travelling across borders. In the past, when these costs were high, formal IOs offered meeting points for diplomats and helped to efficiently transfer information. They were crucial for lowering transaction costs so that states could make mutually beneficial policy adjustments (Keohane 1984). But technological changes have made it easier for officials to coordinate with counterparts abroad. As a result, cross-border governance can increasingly be supplied with minimal institutions, and the number of informal IOs has grown in response.

In this article, I argue that in addition to falling information, communication, and transportation (ICT) costs and shifts in the ‘demand’ for informality, one of the crucial factors underpinning the rise of informal bodies has been the prior proliferation of formal IOs – especially, ones that are well-resourced, with open-ended mandates. This is because an international environment already replete with such institutions is likely to change calculations about the costs and benefits of informal designs. One of the key advantages of informal institutions is that they can be created relatively easily – at ‘low cost’ – and, with their light institutional footprints, can flexibly respond to changing problems (Abbott and Faude 2022). These features may be desirable in many instances, as scholars have shown. But informality also has costs. On their own, informal arrangements usually lack basic capabilities that most formal IOs possess, like secretariats, legal privileges and immunities, financial resources, and enforcement capabilities. Changing ICT costs may have obviated the need for some of these things. But, equally, when formal IOs are plentiful, states can rely on these already-existing institutions to extend their unique resources and services to informal bodies, helping to plug organizational gaps and make them more viable solutions to cross-border problems.

This can happen in two ways. First, when states design institutions, they may opt not to create formal IOs – even if, in a different context, this choice may have been warranted – when they can rely on those at hand to supply missing institutional features. This is more likely when a number of well-resourced formal IOs are available to ‘lend’ resources and provide services to a new body, making states more confident that low-cost institutions can effectively operate as substitutes for formal ones. The stock of formal IOs, in this case, shifts calculations about the viability of this approach and leads states to embrace informality more readily. Second, the presence of formal IOs may allow informal bodies to avoid formalization later on. Sometimes, after states create low-cost institutions, organizations confront new challenges and states need to expand their capacities in response. In such instances, there may be an incentive to reconstitute an organization on a new legal basis, as when the General Agreement on Tariffs and Trade (GATT) was converted into the WTO. However, if basic capabilities can instead be ‘borrowed’ from a pre-existing institution, this reduces pressure to do so. An organization may then remain informal, which would not have in another context.

Overall, I argue, when there are numerous, well-resourced formal IOs operating in a region or issue area they can operate like a ‘coral reef’ that makes it possible for a large number of informal IOs to grow, thrive, and survive.¹ Collectively, like falling ICT costs, they help to make informality work, relieving some of the need for states to create more formal IOs. In contrast with Manulak and Snidal’s argument, however, they are hardly less necessary. Indeed, when seen from this perspective, formal organizations have actually been crucial to the proliferation of informal institutions. The rest of this article develops this theoretical insight and provides a preliminary demonstration of its ability to illuminate patterns of informality. I begin, in the next section, by situating my argument within the literature and explaining how the spread of informality is likely to be partly endogenous to the prevailing institutional context – namely, the number of formal IOs available to plug organizational gaps. Second, I demonstrate that aggregate patterns of governance provide evidence consistent with this account. I show that formal IOs have provided extensive assistance to informal bodies, and that well-resourced, general-purpose institutions have been especially important – the EU in particular. Third, I zoom in on the European experience to unpack the causal mechanisms at work, focusing on the EU’s involvement in two informal bodies: the European Cooperation in Science and Technology (COST) and the European Research Coordinating Agency (EUREKA). The conclusion then extends the paper’s findings, explaining how they speak to ongoing debates in IR, help to illuminate the impact of the EU on global governance, and how scholars can develop this line of research in the future.

2. Theorizing the supply of informal organizations

The rise of informal IOs – defined as organizations, created by states, that are constituted by non-binding international agreements – has been much puzzled over in recent years.² Once thought of as exceptional, scholars have increasingly documented their extraordinary growth in the postwar period, as shown in Figure 1.³ Today, they are central instruments of cooperation, constituting 30–40% of all active IOs (Roger and Rowan 2021a). Their rise has inevitably raised questions about the drivers and implications of this trend, and numerous answers have already been provided. Some functionalist explanations cast informal organizations in a favourable light. According to these accounts, the prevalence of informal IOs in specific regions and issue areas is linked to the problems states face – whether there is demand for fast, flexible, and confidential institutions, for instance (Vabulas and Snidal 2013). States have increasingly relied on informal bodies, on this view, because they are better ‘matched’ with the new challenges states have confronted as interdependence has deepened. Others explanations paint a somewhat darker picture. Power-based approaches, like Randall Stone’s (2013), suggest that states have opted for informality because their relationships have become more conflictual, making it difficult to agree on the design of formal institutions; while, in my own work, I have explained that growing domestic constraints and the involvement of independent agencies have led states to embrace informality, even when this is suboptimal from a broader policymaking perspective (Roger 2020).

Figure 1. The growth of informal IOs.

These accounts are quite different in nature. However, there is a common thread that links them: all focus on the shifting ‘demand’ for informality. In each, scholars have explained the growth of informality by pointing to changes in the global political landscape that have increased the incentive for states to establish such bodies. Undoubtedly, this is an important element of the story. Yet it has become apparent that it is not the only one, and that current accounts neglect the corresponding question of ‘supply.’ This view has been forcefully expressed by Manulak and Snidal (2021), who have argued that falling ICT costs have also made it easier for officials within states to meet, communicate, and cooperate with one another, and that this development has diminished the benefits of ‘organizational centralization’ and independent secretariats.⁴ With less need for the benefits that formal IOs traditionally provide, policymakers have little reason to pay the high costs – both financial and political – needed to attain them. Informal designs constitute more plausible avenues for cooperation that can supply similar levels of governance at much lower cost.

Manulak and Snidal offer a valuable perspective on current debates. They are correct in their observation that such mechanisms operate differently from those scholars have focused on so far. The transportation and communication revolutions have also, certainly, enabled new ways of cooperating across borders. The growing interactions between public officials occurring outside the channels provided by formal IOs – in Europe and the transatlantic area, for instance – offers a testament to this idea (Hopkins 1976; Slaughter 2004). I argue, though, that ICT costs are not the only factor augmenting the supply of informal governance, and, specifically, that Manulak and Snidal’s approach overlooks the crucial role that formal IOs have played in making informality work. Far from diminishing in importance, they have actually been one of the central factors underpinning the proliferation of informal institutions. This is so, I explain, because formal IOs can extend crucial services and resources to informal bodies that increase their viability as ‘low-cost’ solutions to cross-border problems. And, when a large number of institutions play this kind of role, this changes state calculations about the costs and benefits of informality, relieves pressure to use a treaty, and leads states to create more informal bodies.

The logic behind this argument begins from the observation that informal IOs are often not optimal when considered in isolation. They are certainly ‘cheap’ to create, which is a real advantage in many instances. But this does come at some expense, since IOs created without binding instruments often lack crucial features that can leave them ill-equipped in the face of many problems. Lower ICT costs may have reduced the need for these by making it easier to coordinate without complex institutional arrangements. However, this is not always so. Across many domains, states still need centralized, independent, and relatively powerful regulatory bodies, and numerous informal IOs have been ‘formalized’ at later stages (Hawkins et al. 2006; Roger 2020). So, it appears that organizational gaps may remain. One study by Patrick (2015) gives a sense of this. He explains that while informal arrangements do have advantages, their success has sometimes been limited because they often lack key capabilities. Verdier (2009) has also shown that several financial institutions have been ‘mismatched’ with the underlying problems they are asked to address. And, in other work (Roger 2020), I have performed a similar analysis on a wider set of bodies, finding that informal IOs are much more likely to be mismatched than formal ones.

These studies caveat some of the more grandiose claims that have been made about informal governance. What they do not account for, however, are the resources and services that formal IOs often provide, supplying some of the institutional backbone informal organizations would otherwise lack. Many informal bodies do not possess their own meeting facilitates, for example, yet they can make use of formal IOs’, benefiting from the centralization they offer (Vabulas 2019). Informal bodies also frequently lack independent secretariats. However, formal IOs have sometimes played a surrogate role, providing the benefits that an independent staff would usually afford (Ibid). In these cases, informal institutions establish close relationships with formal ones and profit from their resources, privileges, and immunities: costs are defrayed, capabilities are enhanced, and legal protections are obtained. But many have benefited from looser relationships too. Formal IOs often provide funding and analytical support, and have sometimes facilitated the implementation of informal rules. The OECD, for instance, has supported several informal IOs, like INTOSAI and the Deauville Partnership, and the Bretton Woods institutions have promoted compliance with the Basel Committee’s rules (Martens and Jakobi 2010; Brummer 2020, 127).

These types of direct and indirect relationships augment the capabilities of informal IOs. Certainly, individual states can – and frequently do – provide secretariats, meeting spaces, and other resources themselves, or on a rotating basis (Vabulas 2019). But when formal IOs do so on their behalf, informal bodies benefit from their independence, their legal privileges and immunities, and their unique regulatory capabilities. They do not become formal IOs in the process, but informal IOs can operate as if they were. Knowledge of this possibility can, accordingly, be expected to change state calculations about the costs and benefits of informal designs. First, it may increase the likelihood that states will choose informality when they are establishing new institutions. When policymakers know that formal IOs are available to ‘lend’ resources and services to plug organizational gaps, this increases confidence in their viability and reduces the need to rely on formal mechanisms. Even if ‘demand’ for formality was constant, then, we could expect reliance on informality to grow over time. Second, in contexts where the demand for formality increases after an informal body is created – perhaps, due to the shifting nature of a problem or an expansion of regulatory objectives, as in the GATT/WTO case – the availability of formal IOs should reduce the need to fully ‘formalize’ later on. Rather than converting an informal body into a formal IO using a treaty, states can upgrade the capabilities of an informal organization by ‘borrowing’ resources and services from pre-existing institutions. This lowers the cost of producing governance and may allow an organization to maintain an informal design, even if this choice would have been suboptimal in a different context.

Of course, not all formal IOs can play this role. Larger, better equipped institutions are more likely to have the wherewithal to engage in these kinds of activities. Those with bigger facilitates, larger budgets, and more staff should find it easier to lend space, provide resources, and offer unique services to informal bodies (Heldt and Schmidtke 2017; Hale and Roger 2014). Mandates may matter too. ‘General-purpose’ organizations should be more likely to regard these kinds of activities as somehow advancing their diverse objectives, and may support a wider range of organizations (Johnson 2014.Hooghe, Lenz, and Marks 2019) Overall, as well, formal IOs with these characteristics are likely to be unevenly distributed across the global landscape. We can expect them to cluster, specifically, in regions or domains where richer, more tightly-integrated, and more democratic states have tended to predominate – like Europe and North America – since these have historically been the top creators of formal institutions (Pevehouse et al. 2020). Ultimately, this has implications for divergent patterns of informality across regions and issue areas, since the calculus of cooperation should shift more decisively when this is the case. In places where more formal IOs are available to fill out gaps in informal bodies, confidence in the viability of such solutions should increase: low-cost institutions will not only be cheap but also more ‘cost-effective’ governors than they would otherwise be. And, as the costs of producing governance fall – as the stock of formal IOs accumulates – this should be followed by a rising number of informal bodies.

3. Analyzing aggregate patterns of formal IO support

Already, there is some evidence consistent with this theory. We know, for instance, that informal institutions have expanded in the wake of the tremendous growth of formal IOs since World War II. Further, as Figure 1 showed, there appears to have been an important substitution effect, since the rate of formal IO creation visually levels off when the informal growth rate picks up. Available statistical analyses also suggest that there is more behind this than a shift in demand or changing technology – with formal IOs playing an important role. In my earlier analysis of the origins of informality (Roger 2020, 97–98), for example, I found that design choices are partially endogenous to the prevailing institutional context: as formal IOs grow, states are more likely to adopt informal designs. Finally, if we look at geographical patterns, we find relevant trends as well. European and North American states are the most significant participants in formal IOs, overall, and have created the largest number of informal bodies (Pevehouse et al. 2020; Roger 2020). Other regions, where formal IOs are less numerous, appear comparatively less involved.

These findings lend some support to the argument I advance. They indicate that there is, at least, an important association between the number of formal IOs in operation and subsequent design choices. To probe the theory’s merits further, I start by unpacking the aggregate dynamics surrounding this relationship, exploring patterns of formal IO support for informal organizations. To do so, I draw on the dataset of informal IOs I have developed with Sam Rowan (Roger and Rowan 2021a). Within this set, I focus on the 200 or so ‘currently active’ in 2010, and for each organization I examine the direct and indirect assistance informal IOs have received, plus the formal IOs providing this support. To code instances of ‘direct’ support, I examined IO websites, official documents, and secondary sources to determine whether a formal IO provides or hosts the secretariat of an informal body. This is usually clearly stated and easy to observe. Less easily observed are instances of ‘indirect’ support. To code these, I leveraged the short descriptions of IOs in the Yearbook of International Organizations (UIA 2019), which include self-reported information on ‘IGO relations’ for each body. These provided an initial list of associations, which were then cross-referenced with IO websites, official histories, and annual reports. An organization was then recorded as having received indirect support if a listed formal IO was found to: i) provide funding or technical assistance to an informal IO; and/or, ii) be explicitly involved in the implementation of projects or rules, usually through joint programs.⁵ These data were then used to generate a list of the formal IOs involved, and information on the total support offered.

The results, shown in Figure 2, indicate that informal IOs receive substantial levels of assistance. More than half (51%) have a close relationship with at least one formal IO. Of these 107 organizations, 47 have received direct support. Thus, nearly a quarter of all informal IOs have secretariats that are provided or hosted by a formal IO. This level of support is only a minimum bound, however. It signals a particularly tight connection between a formal and informal IO. But an informal body may receive considerable indirect assistance beyond this, including funding, technical assistance, and help with the implementation of programs or rules. The Financial Stability Board (FSB) offers an example. It is based at the Bank for International Settlements (BIS) and ‘counts’ as an informal organization receiving direct support. Meetings usually take place in Basel, and it receives a variety of secretarial and analytical services. Yet it also has a substantial relationship with the IMF, which helps enforce its financial standards (Brummer 2020).

Figure 2. Assistance received by informal IOs in 2010.

In other cases, relationships are looser. But these amount to a substantial number too. In total, 60 organizations (30% in total) have enjoyed indirect support from at least one formal IO. This usually means a body has received funding and technical support from a formal institution, of the kind just described, but without direct secretarial support. An example, here, is the Versailles Project on Advanced Materials and Standards, an organization that harmonizes measures and establishes best practices in the area. It operates separately from other institutions, but it has received analytical support from the International Energy Agency and the International Bureau of Weights and Measures, two formal IOs. Often, though, indirect assistance extends far beyond this. As with the FSB, we frequently see formal IOs cooperating with informals in the implementation of projects or helping to enforce the soft law they create. The International Organization of Securities Commissions, for example, has an independent secretariat in Madrid, but formal IOs, like the World Bank, have assisted with both the creation and implementation of its rules. The capabilities of informal IOs have, therefore, been considerably enhanced even when relationships are more distant.

This clearly indicates how heavily formal IOs have been relied upon to augment informal bodies. When doing so, however, it also appears that they have relied on a particular set of formal organizations that aligns with the theory advanced in the previous section. Indeed, if we now turn to the set of formal IOs involved, the distribution of responsibilities has clearly fallen unevenly across institutions. In total, around 80 formal organizations were found to be assisting informal IOs in 2010. This represents roughly a fifth of all those active at this time. Within this group, too, levels of activity vary considerably. Most organizations provided only one instance of support, while the top-20 shown in Figure 3 were responsible for almost two-thirds of the total, along with most instances of direct support. The EU – responsible for 31 instances of direct and indirect support – stands out as the most significant provider, by far. It is followed by the UN, the World Bank, OECD, IMF, and BIS, and a number of other well-known IOs.⁶

Figure 3. Support provided by formal IOs in 2010.

A visual inspection of the institutions in Figure 3 suggests that the major supporters are all relatively well-resourced, general-purpose organizations, as anticipated. To confirm this statistically, Table 1 presents results from a simple analysis using Hooghe et al.’s (2019) Measure of International Authority (MIA) dataset, which covers more than 70 formal IOs.⁷ The dependent variable is a count of all instances of formal IO support – both direct and indirect – provided to informal bodies in 2010. The independent variables drawn from Hooghe et al.’s dataset include whether an organization is general-purpose or task-specific (Scope), budget size (Resources), independence (Authority, an average of Hooghe et al.’s pooling and delegation variables), membership (Size), and its Age. In addition, I include controls for the mean level of GDP and democracy (Polity) across state members. These scores are all from 2009. The coefficients, estimated using negative binomial regression, show that an IO’s Scope and Resources are, with Size, the most important determinants of their support for informal bodies. All have a positive, statistically significant relationships with the dependent variable, and their effects are almost equally large.⁸ Other variables and controls point in predictable directions, but only Polity – members’ level democracy – is statically significant. Although not shown here, the results are also robust when we exclude the EU and when we replace Authority with its two components.

Table 1. Main results.

Variables	Coefficient β	Mfx
Scope	1.219 (0.434)***	0.602
Resources	1.051 (0.282)***	0.565
Authority	0.529 (1.661)	0.062
Size	0.008 (0.003)***	0.556
Age	−0.009 (0.008)	−0.256
GDP (mean)	0.031 (0.019)	0.378
Polity (mean)	0.089 (0.048)*	0.351
Constant	13.71 (15.56)
Alpha	0.084 (0.365)
Observations	74
Log Likelihood	−121.4
Wald chi2(7)	38.84
Prob > chi2	0.000

Robust standard errors in parentheses.

* p < 0.1; ** p < 0.05; *** p < 0.01.

Finally, what does the global landscape look like? To investigate this question, I first examine membership patterns in these formal institutions. Specifically, I use membership data from the Correlates of War IGO dataset (Version 3) to generate a score for each state that indicates the number formal IO ‘supporters’ that they are members of, weighted by the total amount of support provided by each body (Pevehouse et al. 2020). These are displayed in Figure 4, which shows, as expected, that European and North American states are members of the largest number of particularly active formal IO supporters, indicating that significant ‘coral reefs’ have emerged in these areas.⁹ Asia and South America fall in the middle of the distribution, suggesting that somewhat smaller coral reefs are present in these regions, while African and the Middle Eastern states have the lowest levels of membership in these bodies. Second, I examine how these scores are related to patterns of informality. To do so, I compare these figures with data on the share of informal bodies in states’ total IO membership portfolios (Roger and Rowan 2021a). This reveals that the two measures are strongly correlated (Pearson’s r = 0.64, p < 0.01). Thus, states with significant membership in the biggest formal IO supporters also tend to be particularly involved in informal institutions.

Figure 4. The coral reefs of global governance.

4. Case studies: EU support for COST and EUREKA

This analysis confirms several of my argument’s observable implications, providing a base of supportive evidence. However, it does not show how important formal IO assistance has been for choices about informal designs, either by reducing the need to create a formal body at the outset or by relieving pressure to formalize when problems change later on. To unpack the mechanisms at work, I therefore examine two cases: COST and EUREKA. These were selected for two reasons. First, because they offer examples of each type of assistance – direct and indirect – by the most prominent formal IO supporter, the EU. They should, then, give us a sense of its importance for their operations, albeit in different ways. Second, because both involve similar groups of states, are each involved in coordinating European funding on research and development (R&D) for ICT projects, especially, and have broadly similar designs. Thus, the cases are relatively alike in key respects, allowing us to exclude other possible explanations. These are useful characteristics. However, it is important to be aware that this approach has some drawbacks, since both organizations are from the same region – one where formal and informal organizations are particularly common – and involve support from a uniquely well-resourced, general-purpose formal IO. As a result, while the cases may be especially helpful for probing the basic validity of the causal mechanisms I identify, their ability to illuminate dynamics in regions and issue areas where such institutions are absent is less clear. More research will be needed to do so, as outlined in the conclusion. But my analysis suggests that such efforts would be worthwhile.

4.1. European Cooperation in Science and Technology (COST)

COST is an international organization set up by ministers of research in 1971. It has traditionally focused on precompetitive or non-commercial projects of public interest that bring together cross-national research groups in the fields of inter alia informatics, telecommunications, transportation, oceanography, metallurgy, environmental protection, and meteorology. The organization operates according to a non-binding framework where members propose ‘actions,’ where participation occurs on a voluntary basis through the signing of MOUs between variable groups of states, and where project management is coordinated internationally but funding occurs through national frameworks. As far as IOs go, it has had a relatively low profile. Yet it has consistently played an major part in pan-European scientific and technological collaboration, and its flexible, ‘à la carte’ formula has often been regarded as a success (Roland 1988). Less often noted, though, is the crucial role the European Community/EU has played in making its limited institutional structures work.

This is somewhat ironic given that COST emerged from a failed effort to establish a single Community science and technology policy in the 1960s (Peterson and Sharp 1998). At that time, there was a growing sense that technological developments were rapidly changing and becoming fundamental to future economic growth (Sandholtz 1992). Further, without a major increase in R&D spending, and a collaborative approach, European states risked falling behind the United States (US). This realization led to proposals for a European Technology Community, and, eventually, the creation of a Sub-committee on Scientific and Technology Policy (PREST) under the Council of Ministers in 1965 to identify a path forward.¹⁰ PREST produced two reports with recommendations for new institutional arrangements and identified several areas where high-tech collaboration could occur. However, disagreements within the expert groups set up to finalize the details proved insurmountable, particularly over the issue of participation by non-member states. After the summit of European heads of state and government in The Hague in 1969, an intergovernmental conference scheduled to establish a new organization was then abandoned. Yet in the wake of this development, a Committee of Senior Officials (CSO) envisaged in the original proposals was created in 1970 through an exchange of letters between ministers. This body was delegated responsibility for fleshing out the ‘legal framework and the administrative and financial arrangements to be adopted’ and would itself become the nucleus of COST, persisting after its creation as the organization’s executive body (Council of the European Communities 1981, 48).

Within the CSO, similar issues plagued negotiations. Initially, officials envisioned a legalized, bureaucratic arrangement – a formal IO – that would pool project resources. But the divisions between countries meant that outcome was increasingly unrealistic. In response, the CSO considered a ‘less specific and less binding’ option (Roland 1988, 13). This reflected the lowest-common denominator across states, and eventually won out. The issue with this approach was that a flexible arrangement built upon incongruent national systems, without any supranational element, entailed real costs that ‘had to be compensated for’ (Ibid). On its own, an informal body would lack the administrative capabilities and the legal personality regarded as necessary (Aked and Gummett 1976, 280). The only way these could be provided without creating a formal IO, policymakers believed, was through support from a well-resourced institution. As Antonio Sacchettini (1974, 449) – a legal advisor closely involved in these discussions – observed shortly after its creation, COST represented: ‘a form of international organization whose extremely simplified structures could not justify its autonomous existence without a direct attachment to other organizations with an already consolidated political and administrative structure.’ In practice, this meant the European Community. COST was, under the CSO, evolving into an independent institution with its own membership and governance structures, but due to its genesis as a Community initiative the Council and European Commission maintained close involvement in its work. They stepped in to serve as COST’s ‘implementing agents’ (COST 2011).

COST was given definitive shape through a resolution at a ministerial conference. This set out an operational framework that was non-binding, easy to conclude, and offered considerable flexibility. But, from the outset, this loose structure was combined with the capabilities of a formal IO. Basic costs and services were covered by the Council Secretariat, and the Commission was involved in managing and monitoring the implementation of specific actions. Ultimately, this was critical to its design. A ‘pure’ informal approach was clearly regarded as inadequate – some level of formality was needed. In a different context, this may have forced states to create a formal IO; or, plausibly, without confidence in COST’s viability, cooperation may simply have ended. However, the ability to rely on an already-existing institution relieved policymakers of the need to do so by making them more confident an informal approach could work. Essentially, the Community’s involvement stabilized COST’s design, becoming a permanent feature for decades afterward. And, even when the EU’s role ended in 2003 – when the European Science Foundation stepped in – it continued to provide substantial indirect support, which can be just as important for an informal IO.

4.2. European Research Coordinating Agency (EUREKA)

EUREKA offers a compelling example of the impact indirect support can have on questions of institutional design. At a basic level, the organization is similar to COST, as already noted. Created in 1985, EUREKA is a non-binding and flexible institution that coordinates research funding according to an analogous ‘à la carte’ formula. Its fields of activity overlap considerably – informatics, telecommunications, transport, etc. – and it has largely been regarded as a successful model of cross-border collaboration too. But there are differences. For instance, in contrast with COST, which provides funding for basic research by institutes, research centres, and universities, EUREKA supports R&D by European firms aimed at bringing new products to market. It has had a somewhat higher profile as well. Whereas COST has largely flown under the radar, performing a ‘useful, if low key’ role, EUREKA has generated some controversy throughout its history (Peterson and Sharp 1998, 31). And, finally, the role of the EU in supporting its activities has been less direct. Yet, in key respects, it has been no less crucial for the way EUREKA has evolved over time.

The motivation behind EUREKA stemmed from renewed concerns among European states about the possibility of a technology gap between them, the US, and Japan (Peterson 1993). By the early 1980s, Japanese firms were having a major impact on global markets for high-tech products. And, in 1984, the US announced the Strategic Defense Initiative (SDI), which, in addition to its troublesome implications for arms control negotiations with the Soviet Union, promised to inject new funds into US high-technology firms and sap markets away from Europe. Almost immediately, French President Mitterand proposed a new organization aimed at countering the civilian dimension of SDI. It quickly gained support from Chancellor Helmut Kohl and Hans-Dietrich Genscher, as well as ministers from other European states and a host of industry groups. But there were differences over its design between the French (who proposed ‘an autonomous European agency’) and the Germans (who were concerned the ‘proposal wouldn’t have got through the German parliament’ and pushed for a less ambitious approach) (Riesenhuber 2015, 15). Deferring to Germany’s preference, negotiators ‘agreed to develop a flexible and open form of cooperation’ and launched EUREKA in Hannover within a year after the idea was first proposed (Ibid).

The French and Germans wanted to create EUREKA outside of Community structures. The EC, though more active in the field of research at this time, was therefore not granted a direct role in the fledgling initiative (Peterson and Sharp 1998). Unlike COST, therefore, a small secretariat was instead established as a non-profit, purely as a project ‘clearinghouse’ with no funds of its own. The EC was, nevertheless, an important consideration – a fact signalled by the decision to locate EUREKA’s secretariat in Brussels – and it offered significant indirect assistance (Commission of the European Communities 1986, 12–15). For example, beyond its participation in EUREKA projects as a regular ‘member,’ it provided advice aimed at heading-off possible adverse implications for market competition arising from inter-firm collaboration, and promised to support technical standardization efforts resulting from EUREKA projects. The EC also seconded an official to EUREKA’s secretariat, covered one-sixth of its operating costs, and provided technical support related to the information systems and networks that would be used. This included making available the database system that it had developed for managing its own international research programmes – interestingly, an instance where technology was supplied by a formal organization. All this helped to lower EUREKA’s start-up costs and contributed to the organization’s early successes. However, the EU’s role would increase as challenges arose later on.

By EUREKA’s 16^th Ministerial Conference in Istanbul, in 1998, high-level representatives sensed that problems were accumulating and authorized a Strategic Review to make recommendations. The subsequent report, issued in March 1999, confirmed their worries: much had changed, and the organization was not prepared (Georghiou et al. 1999). European firms faced growing competition, a ‘network economy’ had emerged, and small-and-medium-sized enterprises (SMEs) were becoming increasingly important within EUREKA’s portfolio. To remain relevant, it had to offer new kinds of support. Yet the organization seemed unable to take ‘corrective action’ due to insufficient continuity between chairs, declining political commitment, and a secretariat that was ‘insufficiently resourced to compensate for other deficiencies in the network.’ The report then outlined several possible responses, labelled Autumn, Winter, Spring, and Summer. The first two indicated how bad the situation was. Autumn entailed business-as-usual, which was expected to result in a gradual descent into irrelevance and an ‘undignified death.’ Winter involved accepting that cooperation had run its course and actively terminating cooperation. Spring and Summer, by contrast, outlined paths where corrective action was taken to improve the organization’s prospects. However, their respective emphases were quite different.

In Spring, EUREKA would attempt to ‘re-launch to re-vitalize,’ particularly to establish a secretariat with greater capacity to initiate policies, pool funding, and attain a firmer financial footing for the organization’s activities. This was, by far, the most ambitious course of action, entailing quite significant political and financial costs. But it was not entirely necessary to shoulder these. Indeed, Summer – the final option – offered an even more promising alternative and could be accomplished through a simple revision to the Hannover Declaration. Rather than relaunching, EUREKA could ‘re-orientate’ by building a stronger, more ‘synergistic relationship’ with the EU. By drawing on its resources and capabilities, and reframing EUREKA activities in ways that would benefit the larger, more formal IO, it could fill out gaps in the network and achieve greater success in the long-term. This was the path EUREKA ultimately pursued. In 2000, a joint working group was established with officials from the EC, and a host of collaborative initiatives were agreed upon, laying the basis for two major programmes that dramatically expanded EUREKA’s impact: Eurostars (focused on SME’s) and InvestHorizon (focused on deep tech companies). In this way, then, an informal body responded to the challenges it faced – charting a path between decline and death and a costly formalization process – by partnering with a well-resourced formal IO.

5. Conclusion

The COST and EUREKA cases demonstrate that both types of formal IO support can impact decision-making about institutional design, and nicely illustrate the two mechanisms described in section 2. They show, specifically, that formal IO assistance can shape the choice for informality by: a) making policymakers more confident that informality can work at the outset, as in the story of COST, where direct support was crucial for plugging key legal and administrative gaps; and, b) making it more likely that an informal design will be maintained when problems change later on, as in the EUREKA case, through the indirect support it attained from the EU. Combined with the evidence from Section 3 revealing the extensive direct and indirect support provided to informal bodies by a variety of well-resourced, general-purpose institutions, this suggests that the presence of formal IOs in an organization’s environment is very likely to have been an important supply-side driver of informality. Such bodies can, I argue, operate like coral reefs that provide key resources and services to smaller entities, creating an environment where informal organizations can flourish.

From a policy perspective, this study suggests that while informal IOs may often be problematic on their own, as some have emphasized, steps can be taken to make them more viable solutions. Often, however, this will occur via interactions with other institutions that blend their capabilities. So, to understand the advantages and disadvantages of a particular variety of institution – formal IOs, informal IOs, etc. – it is important to consider their links to other bodies and the larger governance ecosystems they operate within. From a theoretical perspective, this work speaks to ongoing debates about ‘organizational ecology’ (OE) and regime complexity (Abbott, Green, and Keohane 2016; Abbott and Faude 2021). Theories of OE, for instance, have recently grown within the literature on global governance, and have generally focused on a competition mechanism where larger organizations, like formal IOs, use up resources and give way to smaller ones that fill regulatory niches. Much research on regime complexity has emphasized, likewise, how institutions frequently conflict with and undermine one another. This argument, however, highlights how bodies with different capabilities can complement and enable one another. And, in the case of OE arguments, formal IOs may actually lower the costs of producing governance via institutions that would not otherwise represent viable solutions on their own – leading them to grow, in turn.¹¹

That said, there are several areas where further research is needed. First, it will be important to explore the role of institutions beyond the EU. This organization is the largest formal IO supporter, and, empirically, it sits at the centre of my analysis. However, the EU is sometimes regarded as a sui generis institution. It is possible, therefore, that the dynamic I have described does not extend to other organizations or regions where similar institutions are missing. Arguably, there is evidence that much smaller formal IOs, like the BIS and the OECD, have had a similar effect, suggesting that it may not be so unique. The BIS, for instance, has made the use of soft law and informal IOs a viable approach to the governance of global finance, as Brummer (2020) has argued. The OECD has played a similar role by brokering informal cooperation across many issue areas.¹² But such cases need to be studied in greater detail and the ‘coral reefs’ in other parts of the world require special attention to understand how they are linked to varying patterns of informality.

Second, more work is needed to understand the micro-dynamics of the inter-institutional interactions I have described. The thesis I have advanced focuses on how the stock of formal IOs and the promise of support may shape incentives to create informal ones. Yet, in this account, the factors driving the behaviour of individual institutions are ‘bracketed.’ Why do formal IOs provide support? It may be that informal bodies can serve an important role for formal IO by helping to bypass gridlock and permit member to engage in a kind of ‘differentiated’ integration (Bickerton, Puetter, and Hodson 2015; Schimmelfennig 2019). Equally, the decision-making of individual informal IOs may be more complex, particularly when they can form relationships with different formal IOs or with particular states, which, in turn, are likely to have their own motivations for supporting informal bodies. Unpacking these dynamics will be necessary for a full understanding the specific choices that formal and informal IOs make when they decide to interact with each other. And, in doing so, this may further illuminate the mechanisms that make informality work.

Acknowledgments

This paper benefited tremendously from the comments and suggestions of many individuals, including members of GLOBE and audiences at the International Studies Association Annual Meeting, in Nashville, and at the workshop on the EU and international regime complexes at UCLouvain. In particular, for their comments and support I would like to thank Ken Abbott, Liliana Andonova, Steve Beiderman, Tom Delreux, Joseph Earsom, Benjamin Faude, Thomas Gehring, Adam Holesch, Yi huyn Kang, Mareike Kleine, Miles Kahler, Robert Kissack, Robert Keohane, Axel Marx, Abe Newman, Lucia Quaglia, Bernhard Reinsberg, Yf Reykers, Carlos Sanchez, Aneta Spendzharova, Michael Tierney, and Oliver Westerwinter. My three research assistants have also been invaluable: Carolina Barría, Francesca Leso, and Débora Terra.

Disclosure statement

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

Additional information

Funding

This work was supported by the European Commission, Horizon 2020 Project GLOBE: ‘Global Governance and the European Union: Future Trends and Scenarios ’ [822654]; Beatriu de Pinós Programme, Agency for the Management of University and Research Grants (AGAUR), Generalitat de Catalunya [2018 BP 00276]; Spanish Ministry of Science and Innovation [PID2019-111467GA-I00].

Notes

1. This metaphor has been employed in related contexts by Tarrow (2001) and Green (2013).

2. This definition follows the approach I take in The Origins of Informality (Roger 2020), which turns on the legal nature of an institution. This feature – whether an IO is constituted by treaty, or not – sharply divides organizations into two groups, formal and informal, but also admits varying degrees of institutionalization across these bodies. That said, in this article, I show that some ‘blurring’ of these categories occurs through the kinds of inter-institutional interactions I describe. For an extensive discussion of the concept of an informal IO, see Vabulas and Snidal (2013), Roger (2020), and Roger and Rowan (2021).

3. Data is drawn from Roger (2020).

4. For a similar argument, see Hopkins (1976).

5. Minor instances of inter-organization cooperation, like permitting IO observers or having an IO as a member, were not included.

6. In this figure, the ‘UN’ refers to the Secretariat only. If we include UN sub-bodies like, UNEP, UNDP, and UNECE, the total is slightly higher – 19 instance of support.

7. For detail about each independent variable and access to the dataset, see: https://garymarks.web.unc.edu/data/international-authority

8. To facilitate comparison, marginal effects are semi-standardized.

9. To ease interpretation, the scores displayed in this figure have been rescaled to range between 0–100.

10. PREST was, formally, a sub-committee under the Committee for Medium-Term Economic Policy, created by the Council in 1964.

11. As such, the account I offer aligns with the kind of ‘mutualism’ mechanism identified by Green and Hadden (2021) and complements Abbott and Faude’s (2021) work on ‘hybrid institutional complexes’ where institutions develop links with one another that can help to strengthen and legitimize both.

12. Martens and Jakobi (2010). Also see Johnson (2014).