This symposium presents a rich set of concepts, data, and testing of relationships regarding the apparent explosion of informal governance activities in world politics. The authors theorize and describe a wide variety of modes of informal governance. They explore both informal institutions themselves and informal governance within formal institutions. They present new datasets and perform preliminary statistical analysis of those data. They also study the relationship between domestic regime types and state choices to participate in different forms of informal governance. The articles in this symposium lay out an exciting research agenda.

In this concluding note, I will briefly summarize some of the key contributions of these articles. In the second section, I emphasize the heavy reliance on ideal-type theorizing in this body of work, address some of potential pitfalls of this approach, and suggest an alternative. The third section raises questions about whether states always choose informality strategically, or whether sometimes informality is simply the result of contextual roadblocks that prevent states from achieving desired levels of formality. The final section discusses possible research agendas.

FormalPara Summary and contributions to the study of informal governance

The study of informal governance in world affairs is not a new idea. The influential International Regimes volume (Krasner 1983) was driven in large part by a desire to steer the study of international organization in a direction that could encompass informal regimes. However, as the contributors to this symposium rightly note, work on international organization in the 1990s and later returned to a focus on formal international organizations. The recent shift of attention back to informal institutions is welcome and productive.

Contributors to this symposium provide insight into both the role of informality in formal organizations and the growth of various types of informal institutions. Kersting and Kilby (forthcoming) build on a strong body of work that examines the exercise of informal influence within the World Bank (and other international financial institutions). Their novel contribution is to bring in the role of domestic politics in the United States, particularly divided government. That the United States informally shapes the World Bank’s lending patterns is well understood. What is less apparent is how the exercise of this influence varies. If the United States were to fully maximize its influence at all times, this would be an exercise in rent extraction that would likely undermine the World Bank as a functional organization. Instead, the United States needs to be strategic about when and how to exercise influence. Kersting and Kilby argue, somewhat counterintuitively, that the returns to exercising influence are greatest when the executive branch faces divided government, since then its ability to influence domestic aid flows is more limited. Their article provides powerful support for this hypotheses.

Carlson and Koremenos (forthcoming) turn to the puzzling role of absolute monarchs in international institutions, and in international agreements more generally. Others have argued that absolute monarchies, many of which are Islamic law states, may be constrained in their ability to fully participate in formal international agreements by their interpretation of Sharia law (Powell 2015). Carlson and Koremenos challenge this understanding, arguing instead that these states strategically choose to cooperate with one another in numerous informal ways. They point out that secret agreements allow monarchies to cooperate with one another to the advantage of state leaders, acting in a cartel-like fashion. While the high correlation between Islamic law states and absolute monarchies complicates efforts to disentangle the effects of religious beliefs versus strategic choices, Carlson and Koremenos present intriguing evidence to back up their argument and set a new research agenda for the study of authoritative institutions.

Westerwinter et al. (forthcoming) presents a new, important dataset on transnational governance initiatives (TGIs). Careful coding elucidates just what TGIs are and how they have proliferated in recent years. Westerwinter demonstrates the utility of this dataset by exploring the relationship between domestic regime type and participation in TGIs, finding that democracies participate at significantly higher rates. Contrasting Westerwinter’s and Carlson and Koremenos’ work sets up an important set of questions about the dimensions of informality that are most attractive to different regime types.

Reinsberg and Westerwinter (forthcoming) further develop the exploration of TGI data by focusing on international development, an issue-area that has seen rapid growth in this form of international cooperation. They spell out four theoretical perspectives that might explain why TGI is an attractive form of organization. Subjecting the data to rigorous tests that take into account selection effects, they find that the choice of TGI depends heavily on context. That is, the existing institutional architecture determines when actors will choose TGI as a form of governance. This work demonstrates the explanatory potential of systematic work on informal institutions in global governance.

FormalPara The pitfalls of ideal-type theorizing

The literature on informal institutions in world politics has grown rapidly, and much of this work develops new ideal types to describe informal institutions. Abbott et al. (2016) introduced the concept of private transnational regulatory organizations (PTROs). They find that PTROs have grown rapidly in the climate change policy space, with their flexibility and ability to fill niches a strong advantage over formal intergovernmental institutions (IGOs). Andonova (2017) examines the growth of interactions between IGOs and a variety of private actors, leading to the creation of many public-private partnerships. Vabulas and Snidal (2013) focus on informal intergovernmental organizations, while Avant and Westerwinter (2016) examine transnational public-private governance initiatives. Westerwinter et al. (forthcoming) introduces a comprehensive dataset on TGIs.

Authors contributing to this literature share common assumptions and theoretical approaches. Like the contributors to this symposium, they all sense that global governance increasingly is about much more than formal IGOs, and want to map the terrain of informality. They develop new ideal types to describe the forms of governance that they observe, and measure the growth of these types. They develop theoretical expectations about when such types are likely to thrive or to be chosen by states, and have begun subjecting those hypotheses to systematic tests. This work has the hallmarks of a productive research agenda.

Yet the heavy reliance on ideal-type theorizing brings with it pitfalls that could potentially undermine the research agenda. How are these various forms of informality related to one another? Given the complexity of global governance arrangements, is it really possible to draw clear lines separating these types from one another? It seems more likely that these types overlap with one another in various ways. It also seems likely that, when studying a particular informal institution, classifying it as one of the types to the exclusion of others may be difficult.

Ideal-type theorizing has the advantage of being a fairly straightforward way to generate hypotheses (Hekman 1983). For example, one of the most influential (still) books in the modern study of international relations, Power and Interdependence (Keohane and Nye 1977), relied almost entirely on the introduction of ideal types and using them to explore change in world politics. Focusing on an ideal type allows the theorist to strip an entity down to its essentials and relate those essentials to the phenomenon to be explained. An ideal type, as a starting point, introduces a dichotomy: either an object being studied is an example of the ideal type, or it is not. As Westerwinter et al. (forthcoming) discusses, at the early stages of a research program such dichotomies allow for systematic measurement and categorization.

Of course, no real-world example ever precisely meets the conditions of the ideal type. As the research program progresses, scholars identify objects – institutions, in this case – that are closer to or further away from the ideal type. Rather than thinking in terms of categories and dichotomies, theorists begin to specify the underlying dimensions that together constitute the ideal type. More nuanced theoretical contributions can then emerge.

I would argue that the research agenda on informal institutions is now at the stage where it would be productive to move away from ideal-type theorizing to begin conceptualizing the underlying dimensions. Rather than relying on discrete categories – is this particular entity a PTRO, a TGI, or possibly both? – theorists could place the entity in a multidimensional space based on the most important underlying dimensions. We could then begin to observe which institutions are close to one another in this space, and which are far away from one another. We could analyze which dimensions can be studied on their own, and which are not actually independent of one another but depend on each other in important ways.

The articles in this symposium provide insight into what might be some of the most important dimensions along which informal institutions can be arrayed. A list of potentially productive dimensions to study would include:

  1. 1.

    Formalized decision-making procedures, ranging from completely ad hoc to highly explicit and formal.

  2. 2.

    Bindingness, ranging from agreements that impose no obligations on parties (purely voluntary and aspirational) to those that are legally binding.

  3. 3.

    Transparency, ranging from completely secret agreements (see Carlson and Koremenos forthcoming) to completely public.

  4. 4.

    Participation of non-state actors, ranging from purely private to public-private partnerships to solely intergovernmental.

  5. 5.

    Hierarchy, ranging from a horizontal network structure to hierarchic, top-down institutions with managing directors and executive boards.

  6. 6.

    Rigidity of institutional structure, ranging from completely fluid to open to periodic change to completely inflexible (think of the United Nations Security Council (UNSC)).

We can begin to map the concepts used by authors in this issue on these six dimensions. Westerwinter et al. (forthcoming) and Reinsberg and Westerwinter (forthcoming) study TGIs, which they define in part by their reliance on public-private partnerships. Thus, on dimension 4, TGIs fall in the intermediate range. This probably also implies that they are not highly hierarchical (dimension 5) or completely rigid (dimension 6). However, TGIs vary substantially along the first three dimensions, as Westerwinter’s work demonstrates.

Carlson and Koremenos use secrecy as their proxy for informal cooperation. Thus, the agreements they study fall in the low range of transparency (dimension 3). They also argue that these informal agreements are not legally binding, so fall in the low range of dimension 2; and that the agreements also exhibit personalistic rather than formal decision-making procedures (dimension 1); no participation by non-state actors (dimension 4); no hierarchy (dimension 5); and a fluid institutional structure (dimension 6). By thinking about how the informal agreements coded by Carlson and Koremenos fall in these six dimensions, it becomes clear that they are focusing on a very narrow slice of the informal governance world, for example compared to those studying TGIs. The agreements they study score in the very low range on five dimensions of formality, but on the high end in terms of exclusion of non-state actors. This specification of ideal type in turn raises the question of whether these agreements should really be considered informal institutions at all, or if this space is actually just reflective of a lack of institutions. Their work also raises the question of the independence of these six dimensions, as secrecy seems to necessarily imply particular scores on other dimensions.

Kersting and Kilby (forthcoming) study an interesting hybrid of formality and informality. The institution they study – the World Bank – scores high on many of these six dimensions, having a hierarchical institutional structure and formalized decision-making. At the same time, their findings – as well as those of Kilby’s other studies of the World Bank (e.g., Kilby 2013), Randall Stone’s (2011) study of the International Monetary Fund, Kleine’s study of the European Union (2013), and Vreeland and Dreher’s (2014) study of the UNSC – reveal the importance of flexibility and various types of informality in the actual operations of these formal IGOs. While decision-making is formal on paper, in practice the role of powerful states varies depending on the situation.

To bring in one final example, Abbott et al. (2016) focus on PTROs. PTROs provide an intriguing contrast to the secret, informal agreements that Carlson and Koremenos code. These two apparently very different types of agreements/institutions in fact share many characteristics. They would both score in the low range in terms of hierarchy (dimension 5), rigidity of institutional structure (dimension 6), formality of decisionmaking procedures (dimension 1), and bindingness (dimension 2). However, they are on opposite ends of the spectrum when considering the participation of state vs. non-state actors (dimension 4) and transparency (dimension 3). Conceiving of this set of agreements as having many characteristics in common, while varying dramatically in terms of other characteristics, would allow more productive and precise theorizing. For example, Carlson and Koremenos argue that secret agreements are particularly conducive to the pursuit of cartel-like arrangements among absolute monarchies. PTROs would seem in some respects to be the exact opposite ideal type – they are private and largely transparent – but share similar cartel-like tendencies on other dimensions. Is it possible that analysis of secret agreements among absolute Islamist monarchies could provide unexpected insight into the functioning of private regulatory agreements on climate change in the West? Or vice versa?

FormalPara The strategic choice of informality versus impediments to formality

Contributors to this symposium largely adopt a strategic-choice approach to explain institutional design. This work follows in the tradition of Abbott and Snidal’s (1998) approach to informal versus formal institutions, and the Koremenos-Lipson-Snidal approach to institutional design (Koremenos et al. 2001). A strategic-choice approach to institutional design, or more generally to international relations, includes the following elements (Lake and Powell 1999):

  • Actors who are goal-oriented and strategic;

  • Actors’ goals are defined by their desire to maximize their payoffs, which can be defined broadly as long as they meet basic mathematical conditions such as transitivity;

  • Actors anticipate the likely actions and reactions of other actors in their strategic environment;

  • Actors’ realized payoffs are determined by interaction with the choices of other actors and by the strategic environment;

  • The strategic environment includes the structure of actor interaction; information conditions; and constraints on actor choices.

In the most broad perspective, the observation that the choice of informality as an institutional form is typically a strategic choice is clearly accurate. However, this observation may obscure as much as it reveals. Simply explaining such choices as “strategic” does not allow us to determine the actual causal mechanisms leading to institutional design. One crucial distinction is between actors’ preferences and the strategic environment. It is important to distinguish between circumstances in which actors choose informal institutions because they have a positive preference for informality; and those in which they would ideally prefer more formal institutions but are constrained by the strategic environment to settle for informality. Both are strategic choices – but the distinction is meaningful.

Consider the well-known case of the International Trade Organization (ITO). An ITO was originally conceived by the United States and others as one of the core Bretton Woods institutions that were negotiated in the aftermath of WWII. The World Bank and the International Monetary Fund (IMF) emerged from these negotiations as exemplars (still) of hierarchical, formal, state-dominated international organizations. Negotiations at Bretton Woods led to further agreements in the Havana Charter that proposed a similarly formal ITO that covered not just trade issues, narrowly defined, but related issues such as investment and labor rights.

However, the U.S. Congress balked at what seemed at the time to be the singularly intrusive international organization conceived in the Havana Charter. (Why the ITO was more objectionable than the IMF is puzzling in retrospect, but likely related to the political position of capital versus labor interests in the United States at the time.) Negotiators settled for a narrower arrangement, the General Agreement on Tariffs and Trade (GATT). The GATT was informal in some ways, yet highly state-centric, transparent, and binding.Footnote 1 Over time, through hugely consequential rounds of state-centric negotiations, the GATT developed a bureaucratic structure and eventually became the more formal World Trade Organization (WTO). The formality and intrusiveness of the WTO have led to huge protests, as in the Battle of Seattle, and have probably led to the inability of the WTO to complete any major new rounds of negotiation, even as it retains an important role as a dispute mediator.

How should we interpret this journey from the ITO to the GATT to the WTO (to now perhaps the sidelining of the WTO in favor of Preferential Trade Agreements)? The path from informality to formal bureaucratic organizations is a familiar one, as documented by the articles in this symposium and studies such as that of Vabulas and Snidal (2013). We should have no doubt that informality was a strategic state choice in most of these examples. Yet simply acknowledging the strategic nature of the choice leaves open the actual reasons for particular choices.

I would offer a speculative hypothesis: informality is often not a first-choice positive affirmation by powerful state governments, but a way-station, a concession by those governments to domestic interests or opposition by other powerful states. Processes that lead from informal agreements to powerful, intrusive bureaucratic IGOs are dramatically different from choices that create informal secret agreements or PTROs. We need to understand these differences and study them, rather than lumping them all together as a strategic choice of informality.

FormalPara Research agenda

The study of international organization has followed an interesting trajectory from an initial focus on formal voting procedures, to expansion to include informal cooperation within the concept of international regimes, to highly productive studies of formal intergovernmental organizations, to this latest turn to grapple with informality once again. While not exactly a linear trajectory, this broad research agenda has been and continues to be highly productive.

What might improve the prospects for the study of informal world governance to continue to be a productive research agenda? My suggestions follow directly from the previous two sections. I recommend that scholars move away from ideal-type theorizing to conceptualization and operationalization of underlying dimensions of governance arrangements; and go beyond claims of strategic choice to more clearly specify and test particular causal claims.

Ideal-type theorizing is an important social-scientific tool and has proved its worth in the study of global governance generally, and informal governance more specifically. Yet, relying on ideal types as a method to categorize phenomena and develop hypotheses presents real limitations. I suggest a turn to focus on the dimensions that underlie discussions of ideal types; consideration of how these dimensions might interact with one another; and development of datasets that rely on these dimensions rather than discrete, dichotomous comparisons. Rather than specifying “forms of informality,” (Westerwinter et al. forthcoming, p. 3), clarifying the underlying dimensions of informality and how they interact may prove a more productive avenue of research.

I would also suggest that we lose the catch-all language of “strategic choice,” and clarify what exactly drives state or private choices about the level of formality in institutional design. Preferences versus information conditions versus domestic constraints are all very different causal paths. Work on informal institutions in world politics is likely to have the most impact if it stops hedging its bets by covering all of these causal paths under the broad category of strategic choice, and starts diving more deeply into specific factors that drive strategic choices.