Compensatory Layering and the Birth of the Multipurpose Multilateral IGO in the Americas

Abstract International organizations come in many shapes and sizes. Within this institutional gamut, the multipurpose multilateral intergovernmental organization (MMIGO) plays a central role. This institutional form is often traced to the creation of the League of Nations, but in fact the first MMIGO emerged in the Western Hemisphere at the close of the nineteenth century. Originally modeled on a single-issue European public international union, the Commercial Bureau of the American Republics evolved into the multipurpose, multilateral Pan American Union (PAU). Contrary to prominent explanations of institutional genesis, the PAU's design did not result from functional needs nor from the blueprints of a hegemonic power. Advancing a recent synthesis between historical and rational institutionalism, we argue that the first MMIGO arose through a process of compensatory layering: a mechanism whereby a sequence of bargains over control and scope leads to gradual but transformative institutional change. We expect compensatory layering to occur when an organization is focal, power asymmetries among members of that organization are large, and preferences over institutional design diverge. Our empirical and theoretical contributions demonstrate the value a more global international relations (IR) perspective can bring to the study of institutional design. international relations (IR) scholars have long noted that international organizations provide smaller states with voice opportunities; our account suggests those spaces may be of smaller states’ own making.

organization of this kind. 2 However, as we show, the first MMIGO was in fact the Pan American Union (PAU), created at the turn of the twentieth century to foster cooperation among the United States and Latin American countries.
The PAU was not created with multipurpose cooperation in mind, at least initially. Instead, it evolved from the single-issue Commercial Bureau of the American Republics, which was itself an afterthought to a diplomatic conference in 1890. The MMIGO form emerged gradually through a sequence of bargains among the bureau's members. The bureau's initial design resembled-and indeed, intentionally mimicked-the Europe-based public international unions that proliferated in the late nineteenth century. 3 Headquartered in Washington under the direction of the US State Department, it had a single, narrow purpose: to collect and distribute commercial information. Within a few years, however, its development diverged from the European template. By 1910, this process had culminated in the world's first MMIGO.
Latin American states accepted the incipient institution as a focal point for hemispheric cooperation but contested Washington's outsized control over the organization. In response, the United States compromised by offering to expand the bureau's mandate and partially revise decision-making procedures in accordance with Latin American preferences. This process, which we call compensatory layering, gave rise to the PAU.
The MMIGO's roots in the Americas have received little attention in the history of global governance and less among international relations (IR) studies of institutional design. This is a missed opportunity for three reasons. First, studying the PAU's emergence contributes to the development of global IR, which seeks to incorporate perspectives, histories, and agencies from the global South. 4 Explanations of institutional innovation remain stubbornly Eurocentric, crediting the origins of IGOs to a handful of North Atlantic powers. The PAU's development offers a corrective, demonstrating unmistakable traces of Latin American influence in the design of the first MMIGO. 5 As such, the case can advance the so-far limited conversation between global IR and the study of institutional design.
Second, the development of the PAU defies existing theoretical accounts that explain the origins of the MMIGO either as the deliberate outcome of great-power negotiations at "critical junctures," such as major war, 6 or as an evolutionary response MMIGO because both issues of scope (multipurpose) and decision-making procedures (multilateralism) were central to organizational development.
2. Ruggie 1992, 553. 3. Buzan and Lawson 2015;Osterhammel 2014. 4. Acharya and Buzan 2019. 5. A growing number of studies document Latin American agency in the development of global governance, including Fajardo 2022;Fawcett 2012;Finnemore and Jurkovich 2014;Garcia Vargas 2012;Hanrieder 2015;Helleiner 2014;Long 2018;Scarfi 2021;Sikkink 1996Sikkink -97, 2014Thornton 2021;Tourinho 2021. 6. For example, Ikenberry 2001Ikenberry , 2011 to the functional pressures that arose from growing interdependencies. 7 Indeed, many quite different explanations for institutional creation, choice, and design agree that stronger states devise institutions' properties and purposes. Our account illustrates how relatively weaker states gain greater influence over institutional design than is commonly thought. In doing so, we open theoretical space for understanding how non-great powers engage with processes of institutional design and development. Third, the PAU's development opens possibilities for conceptual clarification and hypothesis generation regarding pathways of institutional development. The PAU did not begin its life as a MMIGO-or even as the Pan American Union. However, as we show, the PAU gradually evolved beyond its original institutional design through a series of conditioned, iterative bargains among actors with diverse preferences and power. We argue that compensatory layering best explains the ensuing transformation from single-issue Commercial Bureau to multipurpose, multilateral PAU.
Compensatory layering is a mechanism whereby a sequence of bargains over control and scope leads to gradual but transformative changes to institutional design. 8 The concept advances the growing synthesis between historical and rationalist accounts of institutional design and development. As the term suggests, compensatory layering is closely related to historical institutionalist mechanisms of incremental change, especially layering. 9 But where layering tends to be endogenous and at the margins of institutional functions, our mechanism connects the impetus of intergovernmental bargaining-more closely associated with institutional choice theories-with transformative changes in international organizations. Historical institutionalist accounts of layering and rational institutional accounts of bargaining suggest that institutional outcomes will favor the most powerful actors; in contrast, compensatory layering illustrates a pathway through which weaker actors gain considerable-if gradual-influence over the shape and functions of institutions.
We develop the concept of compensatory layering to explain a single, inherently important case. Not only do existing theories fail to explain the emergence of the MMIGO's design, but because the PAU was the first of its kind, the case allows us to discard mechanisms of diffusion from the core. Although emulation shaped the Commercial Bureau of American Republics, the forerunner to the PAU, these mechanisms could not have produced the PAU's multipurpose, multilateral design. And while the case is unique in that sense, it has broader implications. Most directly, we point to instances where the PAU's example influenced the design of later institutions. In addition, our study of the case allows us to identify the scope conditions under which compensatory layering should occur elsewhere. Iterative bargains will drive transformative institutional change when an organization is focal, power asymmetries among members of that organization are large, and preferences over institutional design diverge. 7. Buzan and Lawson 2015;Murphy 1994. 8. Our use of "control" and "scope" draws on Koremenos, Lipson, and Snidal 2001, 770-72. 9. On layering, see Hanrieder 2015;Thelen 2003;van der Heijden 2011. The remainder of the article proceeds as follows. We review the IR literature on the origins of multilateral institutions, especially accounts that explain the diversity of institutional designs. Building on that literature, we conceptualize the process of compensatory layering and specify the conditions under which we expect it to occur. We then examine the roots and development of the MMIGO in the Americas at the turn of the twentieth century and consider the impact of the PAU's design on subsequent organizations, most prominently the League of Nations. Finally, we underscore the implications of our findings for global IR and the study of institutional design and change.

Explaining the Origins of Multipurpose Multilateralism
Today MMIGOs play a central role in facilitating cooperation by bringing together diverse issues under one roof and set of rules. However, this institutional form is of relatively recent origin. Its predecessors were single-issue international organizations, the so-called public international unions. 10 European states set up the first IGO, the Central Commission for the Navigation of the Rhine, at the Congress of Vienna in 1815. The number and diversity of European-based IGOs grew rapidly in the second half of the nineteenth century, with myriad organizations each covering one of a wide range of issues. 11 The International Telegraph Union (1865) and the Universal Postal Union (1874) became the first organizations to develop an increasingly global membership. 12 But although public international unions, conferences, and bodies proliferated, until 1919, European IGOs remained devoted to promoting cooperation within single-issue areas. The public unions also lacked many of the features we associate with multilateralism today. In general, each union was organized and operated under the auspices of a single government. 13 Why, then, did American states develop a novel, multipurpose, and multilateral institutional design?
There has been little empirical attention to this question, and existing theoretical accounts offer insufficient answers. Discussions of early IGOs occasionally namecheck the PAU, but the broader importance of the Western Hemisphere organization has been downplayed. 14 Instead, the literature on the historical development of multilateral organizations focuses on the functional advances of the nineteenth-century European public unions and places the roots of the League of Nations and the United Nations in the aftermath of the two world wars. 10. Ad hoc forms of multilateralism-the coordination of relations among three or more states on the basis of commonly recognized principles-have been a long-standing feature of international relations, but formal IGOs date back only two centuries. Ruggie 1992, 574-83. 11. Eilstrup-Sangiovanni 2020Murphy 1994;Reinalda 2009;Yao 2019, 340. 12. Howland 2015Ravndal 2020. 13. Murphy 1994Reinalda 2009, 3. 14. See, for example, Buzan and Lawson 2015, 88-89;Reinalda 2009, 131-35. Herren 2016 suggests that pan-American institutions merely followed European models. More common are sporadic references to "Pan-Americanism" without mentioning the organization, as in Acharya 2011, 115;Acharya and Buzan 2019, 57-59. This historical focus echoes a broader pattern in IR's explanations of institutional creation and design. Broadly speaking, these tend to emphasize power-based explanations to account for the dynamics of institutional creation and functionalist logics to explain forms of institutional design and evolution. We first discuss these approaches before turning to more recent historical institutionalist explanations of endogenous mechanisms of change. 15

Power-Based Explanations
Power-based approaches explain the emergence of international orders and associated organizations as the result of great-power bargains, especially in the aftermath of major wars. Although most commonly associated with realism, a wide range of analytical traditions share a similar view, including English School, 16 liberal, 17 and constructivist theories. 18 According to this explanation, an institution's design-its organizational structure, decision-making procedures, and mandate-reflects framers' preferences at the time of its creation. In perhaps the most influential account, Ikenberry shows how different forms of institutional order result from the bargains among victorious great powers. Since 1815, these forms of order have gained an increasingly "constitutional" structure that includes more, and more formal, institutions. 19 These are understood to reflect the preferences of powerful states at the moment of creation, though institutions may outlive the intentions of their founders. Following a similar logic, Ruggie emphasizes that "it was less the fact of American hegemony that accounts for the explosion of multilateral arrangements [after World War II] than it was the fact of American hegemony." 20 More recently, Mitzen discusses multilateralism as the outcome of great powers' efforts at problem solving, while Lascurettes points to designers' emphasis on excluding potential rivals. 21 Despite the evident differences among these theoretical accounts, they agree that institutions' rules and mandates depend on the preferences of the powerful.
Power-based accounts highlight several critical turning points for the design of international organizations, most prominently the Congress of Vienna (1815), the 15. Sociological institutionalists argue that organizations often adopt pre-existing designs because they regard them as appropriate, as in Finnemore 1996. The literature on regional institutions emphasizes emulation as a diffusion mechanism, especially regarding mimicry of the European Union: Börzel and Risse 2012;Jetschke and Lenz 2013. Diffusion mechanisms such as emulation are less apt for describing the first instance of the MMIGO, although its forerunner, the Commercial Bureau, was modeled on a European public union.
16. Clark 2005. 17. Ikenberry 2001. 18. Mitzen 2013. More typically, though, constructivists consider a wider set of actors, both state and nonstate, in the shaping of collecting expectations and the "social purpose" of organizations : Acharya 2004;Reus-Smit 1999;Sikkink 2014. 19. Ikenberry 2001. 20. Ruggie 1992, 568. 21. Lascurettes 2020Mitzen 2013. Paris Peace Conference (1919), and Bretton Woods and the UN conferences . 22 Following the preferences of their founders, European great powers defined the contours of the international order in the aftermath of the Napoleonic Wars, with a view toward reaffirming dynastic legitimacy. 23 Similarly, the League of Nations emerged from a compact between Britain and the United States. 24 After the US failed to join, the British and French redirected the league's design toward the preservation of imperial or quasi-imperial relations as a preferred mode of governance. 25 In the creation of the United Nations, the balance shifted toward the United States, which sought to lock in liberal values internationally, while setting up a competition with the Soviet Union. 26 Power-based explanations largely adopt the "punctuated equilibrium model," which considers institutions as stable unless an exogenous shock disrupts them. 27 Within these windows of opportunity, decisions about institutional scope and membership follow the preferences of great powers, who retain institutional control.

Functionalisms
Functionalism underpins the second heterogeneous group of explanations. 28 In the broadest sense, functionalism argues that institutions can be explained in terms of the tasks they perform. Functionalism comes in many forms. Historical sociologists, for example, contend that the first IGOs developed to meet demands for cooperation in the wake of the Industrial Revolution. Changes in communication and transportation technologies generated interdependencies, which then ushered in the creation of the specialized public international unions. 29 As global interdependencies grew, so did international institutions, the designs of which responded to these functional needs for cooperation. The high density of interactions between European societies, and their hub position within colonial networks, explains why these organizations first emerged there. A functionalist logic also informs rational design and institutional choice theories. 30 In this view, states deliberately design international institutions as mechanisms to mitigate or solve collective-action problems. IGOs are created as "focal points" that provide platforms for repeated interaction, 31 logrolling, 32 and 22. These feature prominently in accounts by Ikenberry 2001, Lascurettes 2020, and Ruggie 1992 issue linkages, 33 reducing transaction costs and incentivizing cooperation. The characteristics of the cooperation problem (ideally) determine institutional design. This allows states to make concessions in one area in exchange for gains in another. Koremenos, Lipson, and Snidal emphasize five facets of design over which states often bargain: membership rules, issue scope, the centralization of tasks, control over decision making, and flexibility of arrangements. 34 In this view, an institution's scope should increase as the heterogeneity of actors and their preferences increases, when distribution problems are more severe, and when the enforcement of agreements is more difficult. 35 "This leads rational states to increase scope until the marginal cost of adding another issue roughly equals the marginal benefit." 36 More recently, institutional choice theorists argue that the environment in which organizations operate also shapes their design. 37 Jupille, Mattli, and Snidal argue that the presence or absence of alternatives conditions the decision of states to use, select, change, or create institutions. 38 Lipscy emphasizes that policy-area characteristics explain whether new institutions emerge or whether existing institutions concentrate functions. 39 In policy areas where outside options exist, powerful states can push for changes to institutional design that reflect their interests by threatening to withdraw (an option that is not equally available to smaller states); however, where no credible exit option exists, path dependencies will likely ensue, and change becomes less likely as the threat of withdrawal is ultimately hollow.
Critics, most notably realist scholars, have argued that functionalist accounts ignore power asymmetries. Current institutional choice theories, which developed in part in response to this criticism, share an emphasis on distributional bargaining with power-based accounts. 40 In line with the underlying logic of functionalism, institutional choice understands institutions first and foremost as responses to collective-action problems. However, in situations where states could choose from many theoretically optimal institutional outcomes, the preferences of powerful states matter most.
A first wave of historical institutionalism centered on positive feedback loops and path dependencies. These mechanisms lock in institutional constraints and generate vested interests that make initial institutional designs hard to change. In this model, path-dependent processes tend to favor institutional stasis, and some form of exogenous shock is needed to overcome these constraints, creating opportunities for change at "critical junctures." 42 Later historical institutionalist approaches have placed growing weight on endogenous, incremental institutional change. 43 A central mechanism in this regard is layering, which refers to "the processes through which institutional arrangements are renegotiated periodically in ways that alter their form and functions." 44 When faced with institutions they cannot change directly, actors may circumvent bureaucratic resistance and member-state opposition by adding new elements to existing arrangements. 45 Historical institutionalists expect that layering leads to small incremental adjustments rather than comprehensive revisions of organizational structures. According to Hanrieder, the decentralized governance structure of the World Health Organization (WHO) evolved from Latin American states' opposition to the integration of the pre-existing Pan American Sanitary Bureau into the WHO in 1948. 46 Because decision-making rules favored smaller states, frictions arose between institutional strictures and material power. 47 Frustrated in their attempts to reform the WHO, powerful donors later leveraged their external resources to create new institutional layers in the form of special programs. 48 These diverse accounts all expect powerful states to exercise disproportionate influence over institutional design and development. However, the case of the PAU demonstrates that this does not always happen. Layering is not always a way for the powerful to bypass status quo players. When institutions act as focal points under broader conditions of power asymmetry, layering can lead to unanticipated transformative change that reflects the preferences of weaker actors. 41. Pierson 1996, 136. 42. Capoccia andKeleman 2007;Lundgren, Squatrito, andTallberg 2018. 43. Mahoney andThelen 2015. Early historical institutionalist research in IR drew heavily on neofunctionalism; see Schmitter 1970. On the agency of bureaucrats in driving endogenous institutional change, see Barnett and Finnemore 2004;Hall 2016. 44

Compensatory Layering
Compensatory layering is a process through which transformative institutional change emerges from sequential bargaining over organizational design. This mechanism adds to a growing synthesis between historical and rationalist institutionalisms. Rationalist explanations of institutional design benefit from greater attention to sequencing, temporality, and unintended outcomes. 49 In turn, rationalism's clear account of strategic interactions helps historical institutionalists respond to critiques that they lack an adequate theory of agency. 50 Building on this synthesis, compensatory layering embeds iterated bargaining within historically contingent processes of institutional development. States' deals regarding institutional design often produce unintended results, which in turn condition further rounds of bargaining. Over time, compensatory layering can lead to designs that were neither anticipated by hegemonic preferences nor a response to functional needs. We expect compensatory layering to occur only when three conditions are met: when states cooperate within a focal institution, when power asymmetries among states are large, and when preferences over organizational design diverge. The first condition, focalness, refers to the convergence of actors' expectations that an institution represents the natural site for cooperation. Schelling introduced the idea of focal points to explain how people select among multiple equilibria in their daily lives. When faced with several similarly beneficial possibilities, actors tend to choose an option based on its "prominence" or "conspicuousness." Schelling is agnostic about the sources of focalness, suggesting that in practice it "may depend on imagination more than on logic." 51 We follow Jupille, Mattli, and Snidal, who argue that two factors shape an institution's focalness. First, actors' past decisions and beliefs create shared expectations that an institution is a "default" or "natural" site of cooperation. Second, focalness depends on an organization's institutional environment; all else being equal, an institution's focalness increases as the number of viable outside options decreases. 52 Institutions with a high degree of focalness will pull in actors even when they are dissatisfied with its design. 53 Second, power asymmetries stem from the material and positional disparities in relationships between two or more states. On the one hand, power asymmetries create tensions over institutional control and the distribution of costs and benefits from cooperation. On the other, they allow powerful states to act as paymasters, bearing the costs of institutional setup and side payments. Asymmetries make it difficult for weaker states to create viable alternatives. In this way, asymmetry also can contribute to focalness-especially the (non)existence of exit options-well beyond 49. Jupille, Mattli, andSnidal 2013. 50. Hay andWincott 1998;Keohane 2017. 51. Schelling 1960; see also Manulak 2022, 51. 52. Jupille, Mattli, andSnidal 2013, 40-43. 53. In a more sociological approach, Pouliot 2020, 744 argues that agents are "pulled" toward actions by webs of practices in their social environment; focalness could be seen as having similar effects. the moment of institutional creation. Weaker states may struggle to entice the powerful state away from status quo institutions, or to offer substantial benefits to other weaker states.
Third, for compensatory layering to occur, there must be preference divergence across two or more dimensions of institutional design. There are many reasons why states may disagree on the terms of cooperation. 54 The divergence of preferences can be exacerbated by power asymmetries, 55 given that weaker states may fear powerful states' control of institutions. In a focal organization, actors who are dissatisfied with the status quo will seek to revise the institution rather than abandoning it. The result is a sequence of bargains that is repeated until actors are satisfied, or until the institutional arrangement becomes fundamentally transformed or superseded.
The process of compensatory layering only occurs when all three conditions are present. In the first step of the sequence, a powerful state (A) supports the creation of an international organization to promote cooperation in a defined set of issue areas. As the sponsor, A prefers an institutional design that allows it to maintain a high level of formal control as cooperation increases. Less powerful states (BCD) are also members of the organization. They will join the institution when they find membership preferable to the status quo or remaining isolated outside a new institution. BCD likely bear few costs and stand to achieve some cooperative gains; lacking the capacity to sponsor an effective alternative organization, they defer to state A's initial institutional design. If preferences over the control and scope of cooperation between A and BCD converge, the sequence ends. However, if preferences over institutional mandate and control diverge, BCD will contest the initial institutional design so long as it retains its focalness.
In the second step of the sequence, A and BCD renegotiate the institutional design. Even as continued use makes the organization more focal, divergent preferences over control contribute to BCD's growing dissatisfaction. Although A is reluctant to surrender control over the institution, it may be more amenable to BCD's preferences in other areas, offering adjustments to the organization's scope. As with issue linkages, broadening the scope widens the bargaining space and allows concessions in one area to be offset by gains in another. 56 Whereas issue linkage is a one-shot bargain that leads to preference convergence, compensatory layering continues when actors disagree about institutional control despite the expansion of the organization's scope. The sequence ends when the new institutional design meets BCD's minimal threshold of acceptability or the institution loses its focalness. 57 Otherwise, the previous step is repeated within a revised institutional setting. 54. Frieden 1999;Lundgren, Squatrito, and Tallberg 2018, 553, 562. 55. Voeten 2004, 747. 56. Poast 2012. Theoretically, the sequence would also end if power asymmetries were ameliorated, broadening BCD's exit options and their opportunities for greater institutional control. In most cases, however, power asymmetries will change only slowly.
Each repetition of the sequence leads to gradual adjustments in organizational design. Sequential agreements yield the creation of institutional layers that increase the organization's scope. Those layers, in turn, change the institutional status quo, with later bargains now occurring within the renegotiated context. It is often noted that IGOs provide important voice opportunities for non-hegemonic states; 58 this process suggests those spaces may be of those smaller states' own making. Over time, these incremental modifications accumulate into a transformative change that may favor weaker states. Although we develop compensatory layering from a single case, the mechanism should be observable elsewhere when the scope conditions are met.

Toward the First MMIGO
The immediate predecessor of the first MMIGO was not initially designed to be multipurpose, nor especially multilateral. Instead, the design of the Commercial Bureau emulated that of European public unions, with a focus on a single taskthe collection and distribution of commercial information-conducted under the auspices of a sponsoring government. Compensatory layering helps explain why the Commercial Bureau took such a distinct path. Drawing on a combination of secondary and primary sources, including conference proceedings, diplomatic correspondence, and participants' memoirs, our case study illustrates the process of compensatory layering in the transformation of the single-issue bureau into the multipurpose, multilateral PAU.
The PAU's development occurred under a specific set of conditions. First, the bureau emerged as a focal institution for regional diplomacy. Second, preferences over the bureau's design soon diverged, driven by the dissatisfaction of Latin American states with potent US influence over the institution. Third, power asymmetries among the actors both limited the emergence of alternatives and produced piecemeal concessions over successive bargains. Early institutional changes in control and scope, implemented in previous rounds, conditioned subsequent asymmetrical negotiations. The result was an unplanned transformation-and the creation of the first MMIGO.
There is little evidence that the development of the first MMIGO reflected spillover in response to functional needs; economic interdependence within the Western Hemisphere was still more aspirational than real. Nor does the process correspond to an evident hegemonic design. Initially, the United States proposed the Commercial Bureau to foster trade and investment through the collection and distribution of commercial information. But the United States did not drive the transition from single-issue bureau to multipurpose Union. Instead, the poorly consolidated bureau served, somewhat unexpectedly, as a focal point for asymmetrical bargaining 58. Hurrell 2005;Krasner 1981;Panke 2017. over broader political matters. As the only pan-American institution with such a broad membership, the bureau's focalness exceeded the importance of its limited initial mission. Given the lack of institutional alternatives, states faced a choice between using the bureau or returning to ad hoc-and bilateral-cooperation. The bureau became a hub for regional diplomacy, both between Latin America and the United States and among Latin American diplomats. The resulting use of the institution further increased its focalness.
Bargains over institutional design were also shaped by power asymmetries. Greater material power granted the United States centrality in the operations of the bureau and, later, the PAU. The United States hosted the institution's headquarters, contributed most of its budget, and strove to maintain a high level of control. Whereas the United States required broad participation to achieve its policy aims, Latin American states lacked the resources to establish alternative fora. Despite this convergence of expectations on the site of cooperation, US and Latin American diplomats had heterogeneous preferences over institutional control and the content of cooperation. Latin Americans contested high-handed US management of the bureau and insisted on sovereign equality as a basis for Pan-Americanism. Debates over control were hardfought-and concessions, piecemeal. Preferences for institutional scope also diverged, with Latin American states favoring a broader agenda for the pan-American institution. They were strikingly successful in expanding the bureau's initial single-issue scope. Over less than two decades, the narrowly tailored Commercial Bureau of American Republics became increasingly multipurpose and multilateral.

From Conference to Bureau
The Commercial Bureau was designed during the First International Conference of American States (ICAS), held in Washington, DC, in 1889-90. ICAS was the first conference with near-universal participation of the existing states in the Western Hemisphere. However, it followed from a long tradition of efforts to construct regional unity among the former Spanish colonies in the Americas, starting with the 1826 Congress of Panama. 59 During numerous summits and in drafting frequently unratified treaties-very often without, and sometimes in opposition to, the United States-Latin American states enunciated and consolidated regional preferences for sovereign equality. The region's diplomatic traditions emerged from shared republicanism, interlinked independence struggles, and fears of foreign intervention. 60 Brazil's transition in 1889 from monarchical empire to republic removed a perennial irritant (Canada remained outside the system for decades). Over several decades, Latin America's conference system and diplomatic practices grew more 59. Finnemore and Jurkovich 2014;Scott 1926;Yepes 1955. 60. de la Reza 2000Long and Schulz 2021. geographically universal and were marked by an emphasis on arbitration, an aspiration for greater commerce, and a tendency toward legalization. While the Americas had a multifaceted diplomatic agenda, none of this was well institutionalized. The antecedents for regionalism in the Americas were ad hoc, and there was little to suggest the region as the site for the emergence of the first MMIGO.
The US sponsorship of ICAS represented a significant departure in this long tradition of Latin American-led regional cooperation. While mutual interest in pan-American cooperation increased during the late nineteenth century, so did material asymmetries between the rapidly industrializing northern power and its Latin American neighbors. Yet Latin America was hardly a unified bloc. Although Latin American states were apprehensive of US dominance, many were also attracted by the opportunities of the US market. Argentina worried about competition with its exports to Europe, whereas Mexico considered ICAS an opportunity to consolidate its close political and commercial ties with the United States. Mexico's delegate, Matías Romero, was irritated by the posturing of what he described as an ill-informed and vain Argentine delegation. 61 Brazil likewise favored closer ties with Washington to expand its trade and balance its South American rivals. 62 Chile, in turn, only reluctantly engaged with "political questions" at the conference because of the lasting fallout of the 1879-83 Pacific War with Peru and Bolivia. 63 The conference produced an ambitious series of recommendations on commercial and technical issues. These were ultimately mostly ignored. Overall, ICAS was seen as a frustrating experience, marked by logistical problems and inter-American tensions. 64 However, the conference did produce three concrete results. First, it laid the foundation for the Pan American Railway Commission, which promoted the expansion of intercontinental rail connections. 65 Second, it spurred a series of successor conferences. Finally, the Committee on Customs Regulations recommended the creation of the Commercial Bureau. 66 The conference agenda and invitation included no suggestion for creating such a body; instead, it emerged from practical discussions about the need for greater "distribution of useful information" for commerce. The committee cited two models: the Universal Postal Union and the International Union for the Publication of Customs Tariffs. The latter was being negotiated in Brussels since 1888 with the participation of the United States and several Latin American countries. It provided the explicit template for the operation of the bureau, 67 which was tasked to gather, publish, and distribute such statistics "for the common benefit and at the common expense." 68 The new bureau's scope was 61. Romero 1890, 362. 62. Lessa 2002. 63. Petersen 2022Wilgus 1922, 668. 64. Smith 2000. 65. Caruso 1951Scott 1931, 11. 66. Scott 1931. International American Conference 1890, 535. Thirteen of the forty signatory states of the Brussels convention were from Latin America.
68. Kelchner 1930, 335;Scott 1931, 27-28, 36. very specific: following the model of European public unions, control sat squarely with the sponsoring United States, and membership was open to all conference attendees. The bureau accomplished its mission via the publication and circulation of the Bulletin, with early issues including directory information and a survey of hemispheric coffee production. The USD 36,000 annual budget-fronted by the United States but later to be paid by all member-states proportionally according to population-would cover employment of a director, secretary, accountant, clerk, typewriter, three translators, one messenger, and one porter, along with printing expenses for the trilingual Bulletin.
In Europe, interdependencies led to the creation of institutions to address externalities related to trade and transportation. In contrast, the Commercial Bureau was established to promote commerce where it was scarce. Thus, functionalist arguments based on the European experience offer limited insight into the PAU's origins. Although the founding of the single-issue bureau is consistent with power-based explanations that exalt the role of a regional "paymaster" in creating and maintaining institutions, the PAU's evolution followed a different logic. In its early years, the bureau suffered "recurring moments of confusion as to its mission, anxiety over its lack of a clear public constituency, and doubt as to its permanence." 69 The bureau's existential crises were often linked to weak and inconsistent US support and opposition in Congress and from industry groups. Its early existence was tenuous, and there was no clear US impetus to expand its scope.
Initially, the bureau seemed more likely to wither than to expand and thrive. The United States controlled it, both by design and as a consequence of its greater resources. Although the great power wished to maintain that control, it also needed Latin American participation to keep the bureau afloat. The bureau's role as the only space with a broad pan-American membership gave it an unexpected focalness. Under these circumstances, divergent preferences and Latin American demands catalyzed iterative negotiations over control and scope of the focal institution. 70

From Commercial Bureau to MMIGO
Despite its inauspicious beginnings, in the following years the Commercial Bureau's centrality to inter-American affairs only deepened. The bureau's expansive membership was key to its focalness, as was its location in one of the few cities where most Latin American states had diplomatic representation. What is most notable, however, is the gradual transformation of the bureau's institutional design. The driving force of this evolution was bargaining over control and scope. The first turning point came in 1893. The State Department's control over bureau operations emerged as a sore point for Latin Americans. Frustrated by perceived marginalization from decision making, 69. Vivian 1974, 565. 70. Kelchner 1930 Argentina, Colombia, and Mexico expressed their intention to leave the bureau; the loss of these leading Latin American powers would have made the organization unsustainable. Previous regional traditions had helped consolidate Latin American preferences for procedural rules founded in sovereign equality, even where states' material interests diverged. Latin American statesmen insisted that the operations of the bureau should reflect this principle. 71 In response to these calls, Clinton Furbish, second director of the bureau, suggested structural reforms to allow greater consultation with member-states. 72 Changes in practices began informally; this was enough to keep the protesting states in the organization but not enough to dispel their persistent dissatisfaction. In 1896, the Latin American diplomatic corps, presided over by Mexican minister Matías Romero, pressured US Secretary of State Richard Olney to consider institutional reform. Romero added to this pressure by leaking word of the Latin American discontent to the US press. 73 This episode resulted in the creation of a five-member committee, which recommended the institutionalization of a permanent body tasked with operational oversight of the bureau. 74 These partial concessions of control laid the foundation for what would become the more multilateral governing board of the PAU. Although the committee's monthly meetings were chaired by the United States, rotating Latin American representatives gained substantial influence and could call meetings by majority vote. "The Union, for its part, became more truly representative as bureau governance was clarified and more closely defined. Latin American diplomats played a more active part in this process than is commonly recognized." 75 The committee's management authority over staff and operations was further strengthened in 1899, including the addition of de jure authority over the appointment of director William W. Rockhill. As these changes slowly developed, Latin American diplomats continued to use the bureau despite their dissatisfaction with its design. In part they did so because the bureau provided a platform that did not exist elsewhere for coordinating Latin American positions. This use and the lack of readily available alternatives contributed to actors' converging expectations that the bureau was a focal point for addressing disparate concerns. 76 Although Latin American dissatisfaction initially focused on control, bargaining was not limited to this single dimension. An expansion of the bureau's mission and functions accompanied the Latin American push for a proto-multilateral structure. The same 1896 meetings that forced partial concessions of institutional control also produced a recommendation "that the activities of the bureau be extended 71. Herrera 1946, 95. Casey 1933 similarly emphasizes "the demands of certain of the Hispanic American representatives in Washington." 72. Kelchner 1930, 335. 73. Carrillo Reveles 2018, 52. 74. Inman 1965, 51. 75. Vivian 1974, 566. 76. Carrillo Reveles 2018Inman 1965, 51;Kelchner 1930, 355-57. to include all subjects relating to the economic life and growth of the countries" rather than just the dissemination of commercial information. 77 Negotiations over the nature of the bureau repeatedly shifted between questions of control and scope. The bureau's development coincided with the start of the most blatantly imperial period of US involvement in Central America and the Caribbean. In this context, Pan-Americanism occupied a dual role as the "friendly face of US imperialism" and as an opportunity for Latin American countries to voice their concerns. 78 Secretary of State John Hay entered office on the heels of the US victory over Spain in the War of 1898. Hay brought with him a "conception of the bureau as a great agency for the development of fraternal intercourse and the promotion of common interests and aspirations." 79 As a result, US interest and willingness to dedicate resources to the bureau grew. However, US expansionism also underscored divergent preferences: US imperial misadventures spurred great Latin American opposition and attempts to constrain US unilateralism. 80 In late 1899, the United States began to encourage Mexico to propose and host a second ICAS. 81 This conference would provide an opportunity to expand and formalize many of the changes to the bureau's institutional design. As conference organizer, Mexico suggested that the bureau, rather than the host country (as was customary in Europe at the time), should draw up the agenda. This solved two problems at once: it limited the influence of the State Department over the content of the conference, and it allowed Mexico to maintain a position of formal neutrality in the political spats among participating states. 82 By the time of the second ICAS, held in Mexico City in 1901-02, the bureau was already playing a role in preparing the conference agenda and managing logistics-a substantial expansion of its original mandate of sharing commercial information. The bureau even set up an advisory branch in Mexico City.
The evolving bureau retained close links with the international conferences that had created it; as a result, the ICAS served as another venue for sequential intergovernmental bargains over institutional change. Continuing discontent over the workings of the bureau led to a Guatemalan proposal at the second ICAS for greater reorganization-a proposal that diverged from hegemonic designs. A committee at the conference followed Guatemala's plan closely and drafted a resolution to multilateralize bureau operations. 83 In this new round of bargaining over institutional design, the United States pressed to retain underlying control while making compensatory concessions on consultation and scope. The US secretary of state retained 77. Kelchner 1930, 335. 78. Sheinin 2000; see also Petersen andSchulz 2018, 113-16. 79. Kelchner, 336. 80. Friedman andLong 2015. 81. The initial suggestion was made under President McKinley, but Theodore Roosevelt embraced the idea after McKinley was assassinated. Carrillo Reveles 2018, 54-57. 82. Noel 1902Wilgus 1922, 30-31. 83. Casey 1933Second International American Conference 1902, 248-54. chairmanship ex officio; however, participation in the bureau's chief executive organ, now renamed the governing board, was broadened to include all diplomatic representatives of the member states accredited to the US government. 84 At the second ICAS, American states approved the creation of new layers that responded to Latin American preferences, extending the institution's mandate further. An international library was created, and technical cooperation increased. Furthermore, the resolution stated that "the Bureau shall be charged especially with the performance of all the duties imposed upon it by the resolutions of the present International Conference." 85 The organization's purpose had shifted from collecting and disseminating commercial information to more broadly facilitating the operation of the system of American states. "This sweeping provision made the bureau the permanent administrative organ of the International Conferences of American States." 86 This was particularly important as the second ICAS had itself become even more multipurpose than the first. The conference also shaped an explicit link with emerging global institutions by passing a protocol of adherence to the 1899 Hague Conventions. Drawing on earlier shared international legal cooperation, Latin American states pressed to go even further, with mandatory (instead of voluntary) arbitration of commercial disputes, and an assertion of domestic legal primacy, also known as the Calvo clause. 87 This set the stage for continuing battles over the inclusion of nonintervention and formal equality as core values of the emerging hemispheric institutions. 88 The direct bureau-ICAS links underscored the importance of interstate bargaining, in which Latin Americans repeatedly pushed for more multilateral structures under the principle of sovereign equality and a broader political and technical scope. These demands won partial concessions of control while widening the bureau's scope to noncommercial issues. In recognition of these changes, member-states adopted a more conspicuously multipurpose name for the old Commercial Bureau: the International Bureau of the American Republics.
The changes also cemented the evolving organization's focalness as it became the default site for discussing a growing range of issues. New initiatives included a series of medical, scientific, and juridical conferences, as well as the creation of a monetary commission and a major Pan-American Exposition, held in Buffalo, New York, in 1901. Both the ICAS and the International Bureau of the American Republics gave these cooperative initiatives greater institutional form. In 1902, both the International Sanitary Bureau and the Columbus Memorial Library were created under the bureau's auspices. 89 The contrast to the European institutional environment 84. That the representatives were those accredited to the United States would become a point of contention when the State Department used nonrecognition to exclude representatives of governments it opposed -a long-running issue with revolutionary Mexico after 1910.
85. Casey 1933, 442-44;Kelchner 1930, 338;Resolution of the Second ICAS, 1902, in Scott 1931, 93. 86. Kelchner 1930, 338. 87. Schulz 2017. 88. Morgenfeld 2010, 105-10. 89. Dumont 2022Petersen 2022, especially chapter 2 and Appendix 1. The Sanitary Bureau would later be renamed the Pan American Sanitary Bureau, which is today the secretariat for the Pan American Health is notable; there, cooperation expanded through the addition of myriad public unions. The presence of several rival powers capable of sponsoring institutions led to the proliferation of single-issue organizations. As a result, no single institution gained focalness on par with the incipient Pan American Bureau.
In 1903, in the final major round of bargaining toward the MMIGO, the governing board approved plans for a permanent building to house the bureau. The new building would physically remove the bureau from the State Department's auspices-its focalness was now literally set in stone. The building was needed to house a staff that had expanded from five in 1890 to twenty-four in 1906. The growing staff foreshadowed the development of an international civil service, with an assistant director from Latin America starting in 1898, employees from across the hemisphere, and requirements for staff to report to the governing board. 90 While structures were increasingly multilateral, practices in the bureau were often less than egalitarian and reflected ultimate US organizational control. Asymmetric reliance on US beneficence was made clear when USD 850,000 of the USD 1 million budget for the construction of the new headquarters came through a donation from Andrew Carnegie. The US government paid roughly 70 percent of the bureau's operating budget.
While US financial dominance was largely uncontested, Latin Americans continued to question US prerogatives within the bureau. The bureau's growing focalness meant that actors were invested in its operation despite their divergent preferences over its design. Negotiations again took center stage in Rio de Janeiro at the third ICAS, in 1906. Mexico complained about the US's "exaggerated predominance compared to other American countries" within the bureau. Bolivia and others chafed at US efforts to control the agenda of the ICAS, especially to keep denunciations of the US's Caribbean occupations off the table. 91 At the third ICAS, Ecuador emphasized the juridical equality of states to argue that members "should have the right to elect the [bureau's] chairman by a majority vote," instead of deferring to the United States. 92 While such a change was not adopted in 1906, Latin American pressure for more multilateral control of the bureau grew. The American states approved further institutional reform. This included granting the bureau an explicit role as a "Permanent Committee" for the ICAS, tasking it with serving as a repository for hemispheric agreements, planning a wide array of pan-American conferences, shaping conference agendas, and "to assist in obtaining the ratification of the resolutions and conventions adopted by the Conferences." 93 The focus on recording and promoting ratification was an institutional innovation intended to overcome shallow implementation of regional agreements in the Americas. The bureau thus Organization. The Columbus Memorial Library remains the library and archive for the Organization of American States.
90. Kelchner 1930, 344. 91. Carrillo Reveles 2018, 67-68. 92. Casey 1933. Resolution on the Third ICAS, 1906, in Scott 1931 became the first international organization to function as a treaty depositary, later a central task of the League of Nations. Diverging preferences over control and scope within the focal institution culminated in a series of changes approved at the 1910 ICAS in Buenos Aires. Latin American dissatisfaction with weak multilateral governance structures had been developing in an environment of abrasive asymmetry and increasing Latin American protests against US imperialism. In Buenos Aires, member-states renamed the Bureau the Pan American Union and made clear it was intended to be a permanent body that would encompass the growing array of issues and conferences under Pan-Americanism's big tent. Most importantly, the conference made changes to institutional design that enhanced multilateral control. Many Latin American delegates again urged that the US secretary of state should no longer preside over the governing board ex officio; instead, the chairman was to be elected by members of the board. A resolution advancing this organizational change was adopted at the long-delayed Fifth ICAS, held in 1923. To strengthen multilateralism in practice, Latin American states also approved greater financial contributions to the institution's budget and urged the creation of national committees tasked with monitoring the ratification of ICAS conventions. 94 The changes in the organization's purpose were reflected in its Bulletin-the bureau's original raison d'être-which now sought to be a "dignified magazine" that carried news about the organization and the multiplying pan-American conferences and exhibitions, updates on diplomatic appointments, and articles and photos from across the Americas. 95 Instead of a clearinghouse for commercial statistics, the PAU was an entrepôt for the diversifying array of technical and scientific cooperation in the Americas, a point of interaction among American diplomats, and a lasting, physical representation of the ICAS.
While the PAU's multilateral practices still reflected profound asymmetries of power, its organizational structure had evolved substantially, from single-state control modeled on European public international unions to a multilateral organization that stressed the juridical equality of states and sought to promote cooperation across a swath of international affairs. From a small office created to print a bulletin of commercial statistics emerged a novel institutional form.

The Institutional Influence of the PAU
Though regionally focused, the PAU was the world's first MMIGO. The design would soon become a prominent feature of the international institutional landscape. What does the MMIGO's development in the Americas tell us about its emergence elsewhere? In what follows, we first ask whether the PAU's institutional form directly influenced the design of subsequent MMIGOs, especially the League of Nations; 94. Carrillo Reveles 2018, 68-69;Casey 1933, 447-53;Inman 1965, 76-77. 95. Bulletin of the International Union of the American Republics, 1910, 1. second, we probe the plausibility that similar processes of compensatory layering account for the growth and transformation of international organizations elsewhere.
Emulation is a recurring mechanism for the diffusion of institutional forms. 96 As we have seen, the incremental development of the PAU's first-of-a-kind design did not result from the emulation of a pre-existing organization. Although the League of Nations followed the PAU, there is little evidence that the league's framers sought to model the world organization on the PAU's multipurpose multilateralism. Early historiography exalts the league's originality and downplays any link to its Western Hemisphere predecessor. These accounts were written by the league's advocates, emphasizing the role of bureaucrats in what Robert Cecil, Britain's chief negotiator of the Covenant, called the "first great experiment" in international government. 97 Egon Ranshofen-Wertheimer, the league's "first chronicler," 98 argued that the organization in Geneva departed radically from others. Ranshofen-Wertheimer, who worked in the secretariat in the 1930s, believed that earlier organizations lacked a genuinely international civil service; this also led him to discount the importance of the PAU. "The experience of the Pan American Union might have purveyed some interesting lessons," he noted, but "there is no indication whatsoever that the draftsmen of the Covenant and the Secretary-General were aware of the supranational experience gained in Washington over a period of nearly twenty years." 99 Despite these assertions, there is strong evidence that Pan-American principles did inform the negotiation of the Covenant. Those in the United States who most influenced the league's creation were intimately aware of Pan-Americanism, although they were reticent about the PAU at the Paris Peace Conference, to avoid European interference in their "sphere of influence." Meanwhile, the efforts of some British diplomats to belittle the PAU suggest they saw it as a poor precedent precisely because it granted too much voice to supposedly inconsequential Latin American states. 100 Historians have long argued that the Western Hemisphere served as a "testing ground" for the United States to hash out its ideas about world ordering. 101 Wilson famously invoked the Monroe Doctrine in his "Peace Without Victory" speech to Congress in 1917; he later insisted on its incorporation into the League of Nations Covenant to forestall domestic opposition. 102 Early in the war, the US government proposed a Pan-American Pact for the creation of a collective security system based on republican forms of government, sovereign equality, and territorial 96. Börzel and Risse 2012;Finnemore 1996;Jetschke and Lenz 2013. 97. As Gram-Skjoldager and Ikonomou 2019, 423 note, the first histories were written by people who either advocated or worked in the league, exalting their achievements and excusing the league's failures. More recent scholarship reclaims an emphasis on bureaucratic agency in the evolution of the league's mandate and institutional form : Clavin 2013;Mazower 2012, 145;Pedersen 2015, 46-47. 98. Gram-Skjoldager and Ikonomou 2019, 433. 99. Ranshofen-Wertheimer 1945. See the discussion of Leonard Woolf, later. 101. Gilderhus 1986, 134-36;Grandin 2012, 86;Knock 2019, 81-84;Mazower 2012, 90-91. 102. Scarfi 2016 integrity. The pact-advanced on the initiative of Wilson's close advisor, Edward "Colonel" House-ultimately failed due to Latin American concerns over US interventionism. 103 House, however, continued to consider it as a model for the postwar order. 104 As the key figure in shaping US plans for the Paris Peace Conference, House liaised with the British, drafted the first US proposal for a League of Nations, and advised Wilson during the negotiations. 105 It is no coincidence, then, that the phrasing of the league's Article X on respect for the territorial integrity and the independence of member-states closely follows the wording of the Pan-American Pact. 106 The intellectual influence of Pan-Americanism on the postwar order did not reflect US ideas and interests in isolation. For Grandin, the inclusion of territorial integrity in the Covenant reveals "Latin America's importance in generating Wilsonian liberal internationalism." And further, "the League itself-Wilson's famous fourteenth point-was directly modeled on the Pan American conferences that the United States had been participating in since 1889 and Spanish Americans had been convening since Bolívar's 1826 Panama Congress." 107 Mazower concurs. In his view, the PAU was "an earlier forerunner of Wilson's League." 108 Claims of the PAU's influence echo the arguments of Latin Americans during the formation of the league and, later, the United Nations. 109 One prominent Latin American jurist and diplomat wrote of the league: "Nothing similar existed in Europe then. This was, then, a typically American institution, with a physiognomy that belonged to the Western Hemisphere and that, as such, can be considered one of the great contributions of the Americas to the evolution of international politics." 110 Although US policymakers were aware of the PAU, there is little indication that the drafters at the Paris Conference considered its multipurpose design as a model for the League of Nations. Instead, the primary origin of the League of Nations' design was British rather than American. Whereas the Wilson administration focused on general principles, Britain carefully studied institutional precedents and arrived in Paris with a clearly formulated plan for the world organization. 111 Wilson adopted the 103. Gilderhus 1986. 104. Raffo 1974. The director of the study group tasked with preparing the postwar settlement requested a confidential copy of the Pan-American Pact from Secretary of State Robert Lansing on 19 November 1917, with the explicit intent to "use it as a possible form of general international agreement, without indicating that it was in contemplation an agreement for this hemisphere." Mezes to the Secretary of State, 19 November 1917, in Department of State 1942; see also Armstrong 1982, 8. 106. Raffo 1974. The editor of House's private papers comes to a similar conclusion: the Pan American Pact "was designed not merely to bring the American states more closely together, but also to serve as a model to the European nations when they had ended the war. Both in its specific language and in its general intent, the Pan-American Pact is the immediate prototype of the Covenant of the League of Nations." Seymour 1926, 238. 107. Grandin 2012, 86; see also Gilderhus 1986, 136. 108. Mazower 2012, 91. 109. de Oliveira Lima 1921Padilla 1945. 110. Yepes 1930, 126. 111. Cecil 1941 organizational structure of the league as presented to him in the British drafts. 112 The first time the multipurpose design of the league was proposed was in the influential Smuts Plan of December 1918, 113 which also introduced the distinction between council, assembly, and secretariat, and the mandate system. Early British plans fused European congress diplomacy with Britain's experience with colonial administration. They rarely mention the Americas, and when they do, Latin American contributions figure as cautionary examples, as in The Hague Conference of 1907. 114 In the same vein, Leonard Woolf's influential Fabian proposal of 1916 criticized the PAU's structure of "equal voting" as infeasible and disproportionately empowering of small states, arguing for a council reserved to great powers so as not to be "swamped" by Latin American states. 115 This suggests the league's British architects were indeed aware of the PAU model but considered its design overly multilateral. Although David Mitrany initially shared this skepticism, writing during World War II he suggested that the PAU might provide a positive example for bringing coordinated specialized agencies together under one organizational roof. 116 In a different vein, the PAU's multipurpose and multilateral structure did directly influence at least one regional MMIGO-its successor, the Organization of American States (OAS). In 1948, the OAS consolidated and formalized prewar, Pan-American antecedents to create a new organization that closely resembled the previous one. 117 Even slightly earlier, the PAU appears to have been considered as a model for the 1945 creation of the multipurpose Arab League. 118 The existence of these bodies shaped Article 51 of the UN Charter, which set out the framework for later regional organizations. Elsewhere, early accounts of the Organization of African Unity note that both its organizational charter and structure resembled those of the OAS. 119 These similarities were not coincidental. "The Ethiopian plan [for the OAU] drew heavily upon the seminal ideas of the Organization of American States … and was indeed drafted in large measure by Mr. Truco, Chile's representative to the OAS." 120 Although European institutions are often assumed to be the global model, postwar regional organizations bore greater resemblance to the structures of the PAU and OAS than their European contemporaries-for example, in the prevalence of equal versus weighted voting in governing bodies. The OAS has continued to be an important model for regional organizations outside of Europe. 121 In this way, 112. MacMillan 2001, 99;Raffo 1974, 167. 113. Miller 1928, 43-44. 114. Wilson 1928, 118. 115. Woolf 1916, 236, 239. 116. Mitrany 1943, 36. 117. Long 2020, 243. 118. Macdonald 1965 writes that "the Egyptian Premier, Nahhas Pasha, is known to have studied reports on the organization and activities of the Union of American Republics during this period." On functional areas, see ibid., 171, 180-81.
the PAU quite plausibly exercises a lasting, if indirect, effect on the physiognomy of regional MMIGOs. As a mechanism for institutional design, compensatory layering differs notably from the plausible instances of emulation suggested earlier. In emulation, architects draw on existing designs regarded as beneficial or appropriate. In contrast, compensatory layering explains transformative, but largely unplanned, changes to existing institutions resulting from iterative bargaining. While case studies of compensatory layering in other institutions are beyond the scope of this article, we suggest there are plausible examples where the mechanism helps explain institutional change. We expect compensatory layering to occur where states' preferences over the site of cooperation converge, where preferences over control and scope diverge, and where asymmetries among member-states are large.
Individually, such conditions are not uncommon in the creation of international institutions, but their convergence is quite specific. One prominent, plausible example concerns the evolution of today's archetypal MMIGO, the United Nations. In process and proposal, the UN's initial design reflected sharp asymmetries. Non-great powers were excluded from US-led preparations, including the 1944 Dumbarton Oaks meeting that produced the UN's broad structure and decisionmaking rules. 122 The earliest US plan envisioned robust great-power control over a postwar organization narrowly focused on keeping the peace. But later proposals for the UN shifted toward a more expansive membership and multilateral design to gain the acceptance of smaller states. By 1943, there were also indications the UN would be multipurpose, including a proposed council dedicated to economic and social issues under the auspices of the General Assembly.
From the start, the UN possessed great focalness. In San Francisco in 1945, smaller powers accepted the UN as the only plausible organization, despite their misgivings about elements of institutional design. Focalness did not prevent contestation, however. Smaller states expressed intense dissatisfaction with the great powers' institutional control, especially the veto. Some European and many Latin American delegates threatened a walkout in San Francisco. As a result, an early history noted that "smaller powers did succeed in many instances in introducing important changes in the Dumbarton Oaks Proposals." 123 The great powers retained ultimate institutional control through the Security Council, which ceded minimal space to the General Assembly in matters of peace and security. 124 Yet lobbying by weaker members in the San Francisco conference was important in upgrading the proposed Economic and Social Council to the status of a principal organ. 125 The literature has recognized that the UN system expanded through the proliferation of new programs and agencies, with excessive "layering" even drawing complaints from Secretary-General 122. For overviews, see Bosco 2009, 13-35;Morris 2018. 123. Goodrich and Hambro qtd. in Morris 2018, 48. See also Bosco 2009, 32-36 andTillapaugh 1978 on smaller powers and bargaining over regional security arrangements in the UN Charter. 124. Bosco 2009; Morris 2018, 50-54. 125. Sharp 1969 Boutros Boutros-Ghali in 1992. 126 Attention to compensatory layering suggests that these changes to the UN's structure may have emerged from iterative bargains. Rather than being introduced on the initiative of major powers, these layers reflect the agency of the UN's weaker members who pressed for the incorporation of their concerns into the body's institutional design. 127 For example, the UN Conference on Trade and Development and the Economic Commission for Latin America and the Caribbean are rooted in the global South's preferences for new development frameworks and in these states' pushback against marginalization from decision making. 128 Compensatory layering suggests a plausible mechanism for how these bargains transformed the UN beyond what was contemplated at Dumbarton Oaks. The UN system and the US-led, postwar order are sometimes lauded for offering representation and voice opportunities to non-great powers; 129 our account suggests that those spaces have been created in part through smaller states' own successes in shaping institutional designs, including through compensatory layering.

Conclusion
The emergence of the Pan American Union as the world's first MMIGO calls into question prominent accounts of the origins of global governance. Even though the PAU pioneered the central institutional form of today's international system, its innovations are largely forgotten. Instead, most scholarship in both IR and history centers on the development of European-based institutions such as the public international unions and the League of Nations. The PAU's path from single-issue bureau to multipurpose multilateralism diverged from European models. This historically important case contrasts with theoretical expectations about institutional design and development; studying the PAU allows us to refine explanations of how international organizations emerge and evolve. Existing explanations for the development of the MMIGO emphasize the role of great-power compacts or treat their institutional designs as the result of functional necessities. But the first MMIGO emerged in the absence of these conditions. The genesis of the PAU reflected US interests in promoting commerce in the Western Hemisphere. It did not develop in response to existing functional needs. The United States took the lead in sponsoring the incipient institution, but it did not dictate its trajectory or resulting design. Instead, changes emerged from sequential bargaining over decision making and mandate. Latin American states resented US control over the PAU, but rather than abandoning the institution they sought 126. Boutros-Ghali 1992, 100; Hanrieder 2014; Pouliot 2020. 127. On ECOSOC, see Rosental 2018, 167-68, 172-74;Sharp 1969, 2-3, 19, 28-31. Along similar lines, Mantilla 2018 points to the influence of small, Latin American, and (later) postcolonial states in shaping international laws on internal conflicts.
128. Fajardo 2022, chapter 1; Jupille, Mattli, and Snidal 2013, 139-44;Thornton 2021, 153-55, 169-77. 129. E.g., Ikenberry 2001 reforms that reflected their preferences. From the perspective of the United States, it was easier to incorporate new issues than to surrender organizational control. The ensuing bargains spurred compensatory layering, a mechanism of institutional development in which sequential bargains over an organization's scope and mandate lead to incremental but transformative change. Although we identify the mechanism in a single, historically specific case, we expect compensatory layering to shape institutional development elsewhere, provided that an organization possesses focal quality, actors' preferences over the terms of cooperation diverge, and power asymmetries among those actors are large. Although our explanation contributes to a growing and productive synthesis between rational and historical institutionalism, it entails distinct theoretical expectations. Institutionalist accounts expect both layering and the presence of focal institutions to benefit powerful states, but we show how compensatory layering empowers weaker actors.
As a result, both our theoretical and empirical contributions demonstrate the importance of incorporating a global IR perspective in the study of international organizations. The case of the PAU contrasts with IR's narratives of the nineteenthcentury development of multilateralism, which remain overwhelmingly centered on Europe. As Acharya and Buzan note, "Liberal thinking overlooked the agency of the Third World. The same can also be said of the more general understandings of multilateralism and global governance." 130 Focusing on the emergence of the MMIGO underscores the relevance of this gap. As Ravndal emphasizes, even in the nineteenth century, "non-European states were more than mere add-ons to European IOs." 131 Indeed, we show that they played central roles in the innovation of international organizations.
Neither the process nor its innovative outcome can be understood without accounting for Latin American contributions. Following independence in the early nineteenth century, Latin American countries increasingly interacted with the formalization of international order. 132 If we take seriously Latin America's place in an increasingly connected nineteenth-century world, then we should pay greater attention to the region's role in the development of multilateral practices and institutions. 133 Attention to the MMIGO's emergence, and the role of Latin American actors in its creation, demonstrates the theoretical benefits of bringing global IR into conversation with the study of institutional design. IR still needs to better incorporate the agency of peripheral actors in accounts of institutional development and, more broadly, to explain the conditions under which these states "co-constitute" global order. 134 The history of the PAU illustrates the value of assessing the development of specific 130. Acharya and Buzan 2019. 131. Ravndal 2018, 181. 132. Benton and Ford 2016Schulz 2019. 133. Legal historians recognize that Latin Americans made important contributions to the doctrinal development of international law-see Becker Lorca 2014; de la Reza 2000; Obregón 2006. We suggest this focus should be broadened to other aspects of international order.
134. Mantilla 2018, 319, 343-45;Tourinho 2021. institutions in a broader context without falling back on models of diffusion from the North Atlantic to the rest of the world. The "localization" of ideas and institutions from the "core" can lead to innovation in the "periphery," and those innovations may again be transferred to other contexts. 135 For example, European designs, through the public international unions, shaped initial institutional forms in the Americas. But these forms developed very differently in a regional context marked by stark power asymmetries and divergent preferences. With roots reaching back some 130 years to an obscure office for the publication of commercial statistics, the PAU remains relevant for our understanding of the origins of today's international organizations. While a handful of North Atlantic states have held privileged international positions, they have not constructed international order or institutions alone. The case serves as a reminder of the value of broadening the focus of IR's histories and theoretical accounts beyond Europe and the United States.