Explaining coherence in international regime complexes: How the World Bank shapes the field of multilateral development finance

Foundations of multilateral development finance

The WB was established to mobilize financing for the reconstruction of war-torn countries and for the long-term economic development of countries with limited access to private capital markets. Its Articles of Agreement (AoA) and activities during the formative years reveal the specific notion of multilateral development delegates had in mind when they met in 1944. They designed a system that was characterized by the legalization of multilateral development under the roof of one focal IO committed to multilateralism and economic embeddedness. These governance features are enshrined in institutional design and are transposed into actual behavior by MDBs and their member states. This implies that the application of governance features can vary across institutions and time. When this variation crosses certain thresholds, features are no longer implemented and rivaled by conflicting governance features. At this point, regime complexes become incoherent. We specify these thresholds in the following section.

Multilateralism is a key feature of international organizations (Helleiner, Citation2019). Ruggie (Citation1992) argued that what is distinctive about the feature is not so much the coordination of national policies in groups of three or more but that this coordination occurs on the basis of generalized principles ordering relations among states, without regard to particularistic interests or strategic exigencies.

At Bretton Woods, delegates embraced this notion when they designed the WB as a truly international organization in which all members had some stake and felt responsible for decision-making and guaranteeing the provision of development finance (Gavin & Rodrik, Citation1995). At first, the WB was intended to become an intergovernmental organization, tightly controlled by member states, which would provide guarantees for private loans. However, this initial design was quickly refined for two reasons. First, private investors and WB management preferred direct lending over guarantees (Mason & Asher, Citation1973). To this end, the organization had to be able to raise capital in private markets. Hence, the first WB Presidents, Eugene Meyer, and John J. McCloy demanded more independent management and less intervention by national governments. Instead of a fully state-run organization, the WB became an independent IO where daily work was implemented by an international secretariat that aimed to appear technocratic and apolitical. Second, the universal aspirations of the Bretton Woods system were quickly undermined after the war When the USSR and its allies refused to join, the system came to be rivaled by an alternative order centered around central planning and nonliberal values (Wiles, Citation1968).

Taking these initial limitations into account, Bretton Woods multilateralism in development finance had four components. First, it acknowledged public responsibility for development and was backed by (creditworthy) states. Second, WB membership was open to all states acknowledging this governance feature. Third, decision-making rules gave a voice to both debtor and creditors countries and created a common sense of responsibility. Finally, WB management was comparatively independent of national governments, focusing on low-risk profitable lending, and, as a result, able to generate lendable capital on financial markets (Humphrey, Citation2014). The regime complex would become incoherent if development financing became more transnational, giving private actors similar rights and duties, or more exclusive, restricting state membership based on narrow political or economic grounds.

Legalization is the second governance feature of the system. According to Abbott et al. (Citation2000), legalization describes institutions defined by obligations, precision, and delegation. The WB featured such obligations by formally binding its members to provide paid-in capital, guarantee callable capital, and support the collective effort to mobilize development finance. It defined the purposes of this finance and formalized the conditions under which it could be provided. Similarly, the WB formalized its project lending approach, stipulating the rules on project application, appraisal, implementation, and evaluation. This has contributed to legalizing multilateral development finance as the key principles, including project supervision, end-user controllability, international procurement, and conditionality became formal rules (Bradlow & Fourie, Citation2013). Overall, the system developed in Bretton Woods went beyond ad hoc informal development finance and established a formal organization with a legal mandate and an independent bureaucracy to implement its rules. The regime complex would become incoherent if these formal rules of international lending and formal loan agreements were no longer implemented by formal international institutions, turning to informal institutional designs, practices, and procedures.

Focality defines the extent to which IOs are the single and uncontested governance leader in an international policy space (Abbott et al., Citation2015, p. 24). This implies two possible scenarios. The first describes a situation in which an organization is the primary governance actor in a given policy field. In this situation, an IO is free to implement its mandate because there are no other actors capable of playing a meaningful role. The second scenario describes a situation in which a policy space is populated by more than one relevant actor but these actors acknowledge the superiority of one organization. In this scenario, an IO can still be focal because other actors in the field accept the governance features of the focal IO, aim to cooperate with it, and try to learn from the focal IO’s expertise.

In Bretton Woods, the WB was deliberately created to be the focal IO managing multilateral development finance. Incoherence in the field could have emerged from the United Nations (UN) specialized agencies. Both the WB’s AoA and the UN Charter require the WB to cooperate with these bodies. Yet, the Bank was fearful that too close a relationship with the UN would subject it to undesirable political influence and hurt its credit ratings. Hence, WB management used its financial independence and growing expertise to negotiate a general cooperation agreement with the UN, which declared WB independence rather than establishing UN superiority (Mason & Asher, Citation1973, p. 58). On this legal basis, the WB became focal in the UN system of development. Cooperation took place in technical assistance and project preparation. Because borrowers often did not have the expertise to develop projects fit for WB financing, UN specialized agencies and the WB cooperated in pre-investment studies. Often, the WB took the lead, financing and implementing these studies, whereas UN experts provided occasional support on topics such as agriculture and education.

This high focality, together with its strong financial capabilities, gave the WB important first-mover advantages, consolidating its prominent role and governance features for the decades to come. First, the organization employed highly skilled experts on higher salaries, hired more consultants, commissioned more studies on developmental topics, issued more publications, and held more seminars than its UN counterparts (Kapur, Lewis, & Webb, Citation1997, p. 9). Second, the WB Young Professionals program established in 1949 has by now trained more than 1,700 people who went on to hold positions at the Bank or other development institutions (World Bank, Citation2018). The WB’s external education and research program, the WB Institute was established in 1950. It has trained more than 39,500 people from developing countries and other international development institutions (World Bank, Citation2008).

Third, in addition to these educational facilities ‘by which it can act as a focal point for the study and discussion of development problems’ (World Bank, Citation1955, p. 25), the ‘Bank wishes to make its own knowledge and expertise in the development field as widely available as possible’ (World Bank, Citation1967, p. 17). To this end, it began publishing scientific reports and disseminating its data to member states and IOs (World Bank, Citation1973, p. 67). This early investment in knowledge and expertise, its financial independence, and the limited number of competitors in the field made the WB a respected provider of development expertise able to define and shape the features of the system (Clemens & Kremer, Citation2016; Park & Vetterlein, Citation2010). A fundamental change in this feature of the system would require other MDBs to openly compete with the WB, trying to promote a concept of development that conflicts with that of the WB.

Market Embeddedness is the final governance feature. It describes the idea that international institutions define a set of rules but give governments room for maneuver to intervene and cushion the cycles of private markets (Helleiner, Citation2019; Ruggie, Citation1982). In the case of multilateral development finance, this implies that states are provided with both financial resources and political leeway to implement economic development strategies. These strategies need not necessarily follow a purely Keynesian economic model. Yet, they should embrace the role of the state in economic development. Purely neoliberal approaches focusing exclusively on the wisdom of markets are beyond the notion of embeddedness.

The WB was envisioned as embodying this framework in two ways. First, the organization was designed to channel capital to countries with limited access to private markets. This was intended to empower developing countries to pursue state-led development strategies (Helleiner, Citation2014). Second, the WB’s decision-making structure and loan conditionality aimed to ensure that development interventions were in line with donor and WB management preferences on how best to advance economic development (Gilbert & Vines, Citation2000).

During its formative years, the WB implemented a narrow version of this governance feature. When the organization opened its doors, management began working on its development tasks from a rather blank slate. Beyond a vague task description, such as the provision of resources for ‘the development of productive facilities’ or ‘productive purposes’ (World Bank, Citation2012: Art. 1), the AoA were silent on the allocation of resources. Neither specific countries, nor sectors, or other specifications of how to define and achieve development were provided. Similarly, there was scant guidance from the fledgling discipline of development economics, or from a body of widely accepted procedures and practices (Kapur et al., Citation1997, p. 93). Hence, management had wide discretion in interpreting its mandate. The second president McCloy and his successor Black used this discretion to secure funds for the Bank by convincing private investors that ‘its loans will be sound business risks’ (World Bank, Citation1947, p. 15). WB management embarked on this task in two ways.

First, lending focused on projects and sectors that were considered economically sound and productive, producing earnings that would help repay the loan. As a result, early WB development lending was overwhelmingly directed toward capital infrastructure, including railways, communication facilities, power plants, and ports, excluding other sectors, such as agriculture, education, and human development (Mason & Asher, Citation1973, p. 134).

Second, the WB not only promoted this approach among the growing community of development agencies and practitioners: its management also went to great lengths to convince the banking community that this approach was creditworthy. In 1946, the WB established a marketing department in New York to promote and launch its bond issues among the Wall Street community, and high-ranking management embarked on a vigorous campaign targeting the investment community.

These early lending practices, the organization’s promotional activities, and its success – reflected in AAA credit ratings – defined the field of multilateral development finance, its procedures, and practices for decades (Kapur et al., Citation1997, p. 116). It entrenched a narrow understanding of embeddedness both among development practitioners and in the investment community, setting a precedent for future MDBs. The regime complex would become incoherent if this feature of project lending for productive purposes was to be replaced, for instance, by an approach restricting rather than enabling developing members to implement their development strategies, or that shifted MDB activities to producing development knowledge instead of financing projects.

These governance features are the benchmark against which the innovations of subsequent decades are to be measured. We can diagnose a coherent system if today’s regime complex still acknowledges all four features. A conflictive regime complex would be characterized by a departure from one or more of the features. The more features are rivaled, the more conflictive the regime complex becomes.

Innovations within the World Bank

The WB has experienced a series of transformations. These changes in institutional design and behavior are the subject of a vast literature, which we cannot take fully into account here (for an overview see, for instance, Clegg, Citation2013; Park & Vetterlein, Citation2010; Phillips, Citation2009). Instead, we argue more modestly that these transformations have not substantially challenged the four features of the Bretton Woods system of development finance. While some features have been transformed others have become even stronger and have improved the WB’s capacity to spread its features to other MDBs (Kapur et al., Citation1997; Marshall, Citation2008; Winters & Kulkarni, Citation2014).

With time, the multilateral character of the WB has become more pronounced. Membership has grown across the globe, making it a truly universal IO with currently 189 member states. Today, public responsibility for development is acknowledged by almost all states (Kramarz & Momani, Citation2013; Stone, Citation2013).

Change in this governance feature has been brought about by the more transnational design of the WB. In response to public pressure during the 1990s, the WB became more transparent to the public, more open to civil society organizations, and more accountable to private stakeholders (Heldt, Citation2018; Park, Citation2010; Yi-Chong & Weller, Citation2009). These reforms opened up the WB to transnational influence. However, change was limited and did not result in a conflicting governance feature. Because the rights and duties of private actors are restricted and less formalized, private actors still play a minor role in the operations of the WB.

A similar diagnosis holds for legalization. With the creation of the IFC in 1956, the IDA in 1960, the ICSID in 1967, the MIGA in 1988, and with the growing fiduciary management of trust funds, the WB has substantially broadened the scope of its operations and the legalization of multilateral development finance. The IFC has extended the formalized project lending approach of the WB to private enterprises (Bradlow & Fourie, Citation2013; Mason & Asher, Citation1973, p. 335–379). With the creation of the IDA, the WB began to expand beyond its narrow focus on infrastructure and thus brought its lending regulations to other sectors, including agriculture, education, and social development, and the sphere of concessional lending to low-income countries (Libby, Citation1975; Morrison, Citation2013). Moreover, the WB’s management of trust funds, that is to say, earmarked contributions by donor countries to support specific development purposes (Reinsberg, Michaelowa, & Knack, Citation2017), seems to extend the organization’s lending rules and procedures to bilateral development finance.³

An especially important innovation, deepening legalization, was the creation of the ICSID. As a legal dispute settlement body designed to resolve conflicts between foreign investors and host governments, the ICSID brought more legal certainty to the system of multilateral development finance (Parra, Citation2017; Schreuer, Citation2016). Similarly, the MIGA has contributed to legalization of the field by providing legal guarantees to investors and lenders (Baroudi, Citation2017).

These innovations have also affected the governance feature of embeddedness. Most fundamentally, the creation of the IDA transformed the lending practices of the WB. As IDA resources are provided by member states instead of bond markets, pressure for financial soundness is lower. As a result, WB lending became broader, including agriculture, human development, as well as institution building and good governance. Over time, the WB drifted away from its focus on economically viable infrastructure projects towards more social and political issues (Best, Citation2014). In this context, Lyne and colleagues (Citation2009) show, for instance, that the share of social projects funded by the IBRD and the IDA has grown from 10% in the 1980s to around 30% in the 2000s. Similarly, Güven (Citation2012) demonstrates that during the 2000s nearly half of WB lending commitments emphasized social sustainability, poverty alleviation, and human development as core objectives. This innovation has deepened the embeddedness of WB lending, since it addressed criticism from the developing world that the WB did not provide funding for projects it deemed necessary.

However, this broader scope was accompanied by more conditionality, a trend that reached its peak with the Washington Consensus (Williamson, Citation1993) and the WB’s structural adjustment lending during the 1980s and 1990s. As the policy conditions attached to WB loans were increasingly shaped by neoliberal ideas focusing on the wisdom of markets and the limited role of the state, including fiscal discipline, cutting subsidies, trade liberalization, privatization, and deregulation, this shift in lending practices clearly limited the embeddedness of the system (Babb, Citation2013; Biersteker, Citation1990). Developing states were drawn into a neoliberal economic model that narrowed their room for maneuver instead of broadening it (Rodrik, Citation1990). During this period, the WB came close to losing embeddedness. However, this period came to a close with the 1990s, and the WB began to restrict its conditionality (Nelson, Citation1996; Rodrik, Citation2006).

Over time, the WB has embraced different versions of market embeddedness, some coming close to abandoning this governance feature (Humphrey, Citation2016). But varying emphasis on market forces or state intervention and on infrastructure financing or the more social dimensions of development, including good governance, health, education, and the environment has not produced a fundamentally different system. Rather, the WB has shifted between different varieties of embeddedness but has not switched to an entirely new paradigm (see also Helleiner, Citation2019).

The broader scope of lending, accompanied by rapid growth in financial resources and personnel, further empowered the WB (Heldt & Schmidtke, Citation2017) and thus increased its focality. With time, the organization developed a set of mechanisms to shape the field of development finance. First, involved in a broad variety of development projects across the globe, the WB began orchestrating an increasing number of development actors. To this end, it took the lead in various coordinating groups in which governments and IOs coordinate assistance to particular developing countries (World Bank, Citation1968, p. 18). In these aid consortia, such as the Aid to Pakistan Consortium, the WB set the agenda and invited participants, including regional MDBs like the Asian Development Bank (ADB) (Haas, Citation1974).

Second, the WB broadened its research activities. Influential publications, including the World Development Report and the Doing Business reports, made the organization ‘the single most important external source of ideas and advice to developing-country policymakers’ (Gavin & Rodrik, Citation1995, p. 332).

Third, the above-mentioned training activities of the WB and its continuous interaction with development practitioners gave the WB a ‘tutelage’ function through which it could set the agenda and shape the ideas of many actors in the field (Krueger, Citation1998). For example, in a survey of 6,731 government officials and development practitioners from low and middle-income countries, who were asked to evaluate the influence of 57 aid agencies, the WB was rated as having the highest agenda-setting influence (Custer, Rice, Masaki, Latourell, & Parks, Citation2015, p. 48).

In sum, the innovations at the WB amount to the transformation but not removal or replacement of governance features. Changes to all four features do not reflect a fundamental departure from the purpose and principles developed in Bretton Woods. This does not imply that many of the changes discussed above did not have important consequences for the design and behavior of the WB. Rather, we suggest that these transformations are compatible with the four governance features envisioned in Bretton Woods.

Innovations in the emerging regime complex

Beginning in the mid-1950s, the Bretton Woods system of development finance and its legacy organization confronted a growing external challenge. Regional coalitions began to establish new MDBs. This development has continued to this day, resulting in a population of 27 MDBs operating in the policy space once the exclusive preserve of the WB. This section describes the formation of the regime complex. Second, it turns to the focality of the WB to show that the organization was directly involved in the design of many new MDBs, that it invested in coordinating the emerging regime complex, and that many MDBs learned from WB experience. Third, we show that these mechanisms have indeed contributed to the coherence of the regime complex by spreading the governance features of the Bretton system to many MDBs.

Figure 1 shows the growth of the regime complex from the creation of the WB to the present (Pratt, Citation2018a). The upper plot shows the geographical scope of state coalitions forming new MDBs. We define global coalitions as groups of states from two or more continents. The remaining MDBs were designed by regional coalitions from one continent. The lower plot classifies the same set of MDBs according to geographical lending scope. Globally active MDBs provide support to countries from two or more continents. MDBs with a regional focus support countries from one continent. Sub-regional organizations concentrate on even smaller groups of countries (Delikanli, Dimitrov, & Agolli, Citation2018). An example of the latter is the Caribbean Development Bank (CDB). Formed in 1970 by Barbados, Canada, Dominica, Grenada, Guyana, Jamaica, St. Kitts & Nevis, St. Lucia, St. Vincent & Grenadines, Trinidad & Tobago, and the United Kingdom, this MDB supports projects in the Caribbean region.

Figure 1 shows that, for more than a decade, the WB was the unrivaled IO in development finance. The regime complex began to emerge in the mid-1950s with the emanation of the Council of Europe Development Bank (CEB) and the European Investment Bank (EIB) from the European integration process. Over the following decade, coalitions from all major world regions formed similar regional development banks (RDBs). In 1959, states in the Western hemisphere created the Inter-American Development Bank (IADB). It was followed by the Central American Bank of Economic Integration (CABEI) in 1960. The African Development Bank (AfDB) was formed in 1965. The creation of the ADB in 1966 completed this first wave of RDBs (Ben-Artzi, Citation2016). This first set of MDBs was strongly supported by the United States, being a founding member of the ADB and joining and financing the IADB and AfDB. The hegemonic power of the US during this period should contribute to a coherent set of governance features as the US was already the driving force behind the institutional design of the WB.

During subsequent decades, the regime complex continued to grow. By 1993, states had created 24 MDBs. These additions to the regime complex were more variegated than its predecessors. They include, for instance, the first sub-regional MDBs, like the Arab Fund for Economic and Social Development, financing projects in the Arab region. This more diverse group of MDBs was partly designed to challenge the dominance of Western donors in multilateral development finance. Especially sub-regional institutions, like the Andean Development Cooperation (CAF) or the Eastern and Southern African Trade and Development Bank (ESATDB), aimed to challenge the system (Park & Strand, Citation2016).⁴ This is also reflected in their membership. While the US was a key actor in the creation of the first set of MDBs, it only joined the International Fund for Agricultural Development (IFAD), and the European Bank for Reconstruction and Development.

The final wave of institutional proliferation came during the late 2000s when rising powers took the lead in creating a series of new MDBs. In 2006, Russia together with Kazakhstan created the Eurasian Development Bank. In 2013, the BRICS (Brazil, Russia, India, China, and South Africa) coalition founded the NDB as an explicit alternative to the WB. In 2014, 21 Asian states joined a Chinese-led initiative to create the AIIB, which has since also attracted major non-regional financial powers, including Great Britain and Canada. Since many observers see the AIIB and the NDB as fundamental challenges to the Bretton Woods system and the WB, we return to these two most recent additions to the system in the following section.

Has the WB shaped the coherence of this regime complex by providing a blueprint for the design and behavior of new MDBs? To address this question, we first turn to WB focality within the regime complex. According to our theory, the governance features of the WB should dominate the design and behavior of new MDBs if the WB was able to advocate similar institutional designs, orchestrate MDB lending, and set precedents from which new MDBs learn.

We find evidence for all three mechanisms leading to regime coherence. First, according to the WB’s own account presented in annual reports and historical narratives provided by new MDBs, the WB was directly involved in the creation and design of more than one-third of newly created MDBs. This engagement in the design of new MDBs began already during the first wave of RDBs and continues to this day. Already in 1960, the WB reported that it provided ‘the IADB with assistance during its organization stage’ with the purpose of finding ‘future opportunities for cooperation’ (World Bank, Citation1960, p. 7). Similarly, the WB assisted in the creation and set-up of the AfDB to work out arrangements ‘for studying and financing possible projects of common interest, as well as for coordinating the activities of the two institutions more generally’ (World Bank, Citation1965: 14).⁵ In some cases, including the AfDB, the East African Development Bank (EADB), and the CDB, this early engagement in institutional design and set-up also included the exchange and training of personnel. The WB has, for instance, ‘seconded one of its staff members as adviser to the President of the African Development Bank, while two staff members of that Bank are spending a year at the World Bank’s headquarters, gaining experience in techniques of development finance’ (World Bank, Citation1971, p. 33), it has ‘provided in-service training for staff members of the AfDB and the East African Development Bank’ (World Bank, Citation1973, p. 67), and it has been recruiting staff for the CDB (World Bank, 1970, p. 27). More indirectly, RDBs like the ADB used the WB as a blueprint, for instance, to develop their financial model and lending practices. The principle of financial soundness was thus at the core of the ADB’s lending practices because its management had learned from WB experience (White, Citation1970). These efforts suggest that states forming new MDBs frequently follow the path embarked upon at Bretton Woods, ask the WB for direct input, and use the WB as a blueprint from which they can learn best practices and institutional designs.

Second, the World Bank reports orchestration and coordination activities ‘with a number of other international agencies in the field of economic development, helping to avoid duplication of effort and to assure the mutual reinforcement of their independent activities’ (World Bank, Citation1973, p. 65). Overall, we find that 21 of the 27 MDBs have interacted regularly with WB since their creation. During the formative period of newly established MDBs, this interaction has given the WB the power to spread and popularize its governance features due to its first-mover advantages of recruiting talented and experienced personnel. For many issues, the WB has greater embodied experience in the room – people with personal and professional experience of the issue – than most other development agencies (Clemens & Kremer, Citation2016, p. 64). These coordination efforts range from regular ‘exchanges of information and staff visits’ (World Bank, Citation1971, p. 32) and ongoing cooperation between the WB’s regional missions and RDBs (World Bank, Citation1965, p. 15) to formal agreements on the provision of data and information by the WB, such as data on external debt (World Bank, Citation1968, p. 23), economic reports on countries of common membership, and appraisal reports on Bank projects (World Bank, Citation1973, p. 67); a broad variety of cofinancing efforts; and even the direct provision of services, including project appraisal and loan administration, to other MDBs (World Bank, Citation1978, p. 94). This suggests that WB management and member states expect the WB ‘to play a large and guiding role in the further improvement of machinery for the coordination of development assistance’ (World Bank, Citation1971, p. 32).

Overall, the WB has shaped the regime complex both directly by assisting the design process of MDBs and orchestrating their behavior and indirectly by providing ideas and information. We now turn to the question of whether these efforts have indeed contributed to the coherent application of the governance features of multilateralism, legalization, and market embeddedness.

A review of membership rules and governance structures shows that the regime complex is strongly shaped by multilateralism. First, the vast majority of MDBs has no or fairly limited restrictions on state membership. With the exception of MDBs emanating from regional integration projects, such as the EIB, which is part of the European Union, or the OPEC Fund for International Development, which is attached to OPEC, and some small sub-regional organizations, MDBs are open to members from different world regions and with different income levels. Second, the governance structure of all MDBs is virtually the same. With minor variations, all organizations follow the design of the WB, including a board of governors, a board of directors, and a relatively independent international bureaucracy. Third, the composition of these bodies is highly state-centric. With the exception of three sub-regional banks (EADB, ESATDB, and BSTDB), decision-making rights are restricted to state representatives. Under these rules, both debtors and creditors have voting rights, thus creating the common sense of state responsibility for development that was also at the core of negotiations in Bretton Woods.

A similar picture emerges for legalization. First, all organizations in the regime complex are formal IOs, established by internationally recognized treaties, with a permanent secretariat and other institutionalizations, including headquarters and permanent staff. Hence, more informal structures similar to those identified by Fioretos (Citation2019) in international monetary politics or in global governance more broadly (Vabulas & Snidal, Citation2013) do not exist in the regime complex of multilateral development finance. Second, like the WB, other MDBs have developed formal rules on their lending procedures and conditions. A review of lending tools and procedures shows that all MDBs have developed formal criteria on eligibility, project and program appraisal, due diligence, and conditionality. Hence, legalization is a strong governance feature in the regime complex. This does not imply that the content of these formal rules is similar across the regime complex. As the following section on market embeddedness shows, lending practices vary across MDBs.

As regards market embeddedness, the picture is more diverse. MDBs emphasize different world regions, sectors, and borrowers; provide a variety of financing mechanisms; and demand different levels of conditionality. However, this diversity is in line with the broader understanding of development and market embeddedness the WB developed over time. While the early WB focused on economically sound infrastructure projects, the creation of new lending institutions within the WB and a broader understanding of what shapes development in different social, economic, and political contexts has led the WB to spread its lending across a broader field of sectors and debtors (see above). A review of MDB mission statements, project portfolios, and recent annual reports reveals similar diversity. Today, all MDBs finance development projects in more than one sector. The IADB reports the broadest lending scope with projects in no less than 16 sectors (2017: 6), including infrastructure, trade, social, environmental, and institutional projects. With projects in three sectors, namely infrastructure, agriculture, and social development, the Arab Bank for Economic Development in Africa (Citation2017, p. 18) has the narrowest lending portfolio. The most popular sectors are energy, financial institutions, and infrastructure, as the majority of MDBs reports that the largest share of their funding is directed towards these sectors. The approach of the CEB is rather specific, as it concentrates its lending on supporting small and medium enterprises in creating and preserving jobs (2017, p. 43). However, such direct lending to private actors is also a common feature of the regime complex. Like the WB’s IFC, 17 MDBs have established financing facilities that lend directly to the private sector. Overall, the regime complex focuses strongly on project lending for productive purposes; none of the MDBs in the complex has shifted to a radically new approach.

Similarly, MDBs vary in their demand for safeguards and conditionality. While a small set of sub-regional MDBs, including CAF and ESATDB, either does not report the use of such requirements or leaves their concrete formulation and implementation to borrowing countries (Humphrey & Michaelowa, Citation2013), most regional and global MDBs have established formal conditionality rules. These demands are generally less stringent than the WB’s structural adjustment approach. Nevertheless, the vast majority of MDBs does not finance every project application with the assumption that it will automatically contribute to development.

As a result, the regime complex is marked by different varieties of embeddedness in line with the different periods of lending at the WB. The low or no conditionality lending practices of sub-regional MDBs have the potential to undermine WB development goals, leading to a race to the bottom within the regime complex (Pratt, Citation2018b). However, the general governance feature of market embeddedness is strong and continues to shape the multilateral development finance regime complex.

The AIIB and the NDB: the end of regime coherence?

Many observers claim that the coherence of the regime complex is in jeopardy due to the rise of the China-led BRICS coalition and its aspirations in international development (de Jonge, Citation2017; Weaver, Citation2015). Because China and its coalition partners are assumed to have different preferences, the coalition is expected to design international institutions with conflictive governance features. The AIIB and NDB are thus a demanding test for our theoretical argument. Both new MDBs are equipped with the necessary financial resources to become crucial actors in multilateral development finance. Whether they are designed to use these resources to undermine the system or whether the WB has been able to spread its governance features to these potential challengers is addressed in the following section.

To this end we, first analyze WB focality for the AIIB and NDB. Was the WB involved in their institutional design? How do the three MDBs interact? Have the AIIB and NDB sought to learn from the experience of the WB? As in WB interaction with other MDBs in the regime complex, we find evidence for all three mechanisms leading to regime coherence. First, the founding documents of the AIIB and NDB indicate that they are designed to cooperate with existing MDBs (Asian Infrastructure Investment Bank, Citation2014, p. 1–2; New Development Bank, Citation2014, p. 2–3). This cooperation began already during the design of the two organizations. During a meeting with the WB management, the AIIB President (Asian Infrastructure Investment Bank, Citation2015) acknowledged:

Similarly, WB representatives noted that the 2016 cooperation agreement with the NDB ‘formalizes the Word Bank Group-NDB knowledge partnership that has accompanied the establishment of the NDB’ (World Bank, Citation2016).

Second, this direct input by the WB was paralleled by efforts to learn from WB experience (Lichtenstein, Citation2018). As a high-level Brazilian official involved in the design of the AIIB explained (as quoted in Serrano Oswald, Citation2018, p. 3).

These learning efforts are also reflected in the management of the two organizations. Senior management positions are filled by professionals with previous job experience in the WB and other established MDBs. At the AIIB, five of eight senior management positions are taken by professionals previously employed at the WB, ADB, and AfDB. The AIIB’s International Advisory Board lists a number of former members of WB senior management. At the NDB, three of the five senior management positions are occupied by individuals previously employed at the WB or the ADB (Heldt & Schmidtke, Citation2019).

Third, direct cooperation and coordination between the WB, other established MDBs, and the two potential challengers began during the formative period of the latter. Both MDBs signed memoranda of understanding with the WB and other established MDBs (Shelepov, Citation2017). These documents reveal that both organizations are open to cooperation with the WB and other established MDBs. The memoranda identify commonalities between the cooperating organizations and respective areas of mutual interest. Building on this common understanding of purposes and mandates, the organizations specify goals and ways of cooperation. These provisions include the exchange of information and knowledge in various areas of institution-building and policy-making, the cofinancing of projects, the exchange of staff, and regular high-level consultations. Notably, the memoranda between the WB and the two new MDBs include explicit sections on advisory services in policy-making and institution-building the WB will offer in the future. Furthermore, regulations on cofinancing stress that respective projects must comply with the standards of the cooperating banks. This cooperation provides the WB with a tool to orchestrate the lending practices and conditionality of the AIIB and NDB.

These cooperation contracts have also been put into practice. Since its foundation, the NDB has financed 22 projects, of which three were cofinanced with other MDBs (New Development Bank, Citation2018). The AIIB’s record is more pronounced. Since 2016, the organization has provided financial support to 32 projects. Of these, no fewer than 21 have been cofinancing endeavors with other MDBs. In 15 projects, the WB was among the cooperating organizations (Asian Infrastructure Investment Bank, Citation2018). Overall, this shows that the WB has used its focal position to shape the design and behavior of the AIIB and NDB. Whether this role as a provider of expertise helps the WB to introduce its governance features into the new banks is examined in the following.

The regime complex of multilateral development finance is marked by a multilateralism that acknowledges public responsibility for development, gives a voice to both creditor and debtor states accepting this basic feature, and is comparatively independent of member states. By and large, the AIIB and NDB embody this governance feature. First, like the WB, the two organizations define their purpose in terms of fostering economic development and thus acknowledge public responsibility for this task (Asian Infrastructure Investment Bank, Citation2014, p. 1–2; New Development Bank, Citation2014, p. 2–3). Second, membership rules are similar to those of the WB and most other MDBs as they are open to all states and do not specify exclusive conditions (Kaoutzanis, Poast, & Urpelainen, Citation2016). The NDB is open to all members of the United Nations (New Development Bank, Citation2014, p. 5) and the AIIB accepts members of the WB and the ADB (Asian Infrastructure Investment Bank, Citation2014, p. 3). Third, the governance structure of AIIB and NDB is similar to that of the WB (Wang, Citation2017). Both organizations are managed by a board of governors, executive directors, a president, a management team, and staff. These bodies have similar rules on composition, powers, and voting rights, thus, reaffirming the rights and duties of both creditor and debtor states (Asian Infrastructure Investment Bank, Citation2014, p. 21; New Development Bank, Citation2014, p. 10). Both new MDBs define the board of governors as their supreme governing body, composed of one governor per member, and they delegate authority for day-to-day operations to the board of executive directors. As power-based institutionalist theories suggest, representation and voting rules advantage powerful founding members (Kaya, Citation2015). While the WB grants reserved seats and additional votes to the members who contribute the most capital, the AIIB and NDB reserve seats for regional and founding members, respectively, and provide these members with additional votes or veto rights. Finally, the AIIB and the NDB are even less transnational than the WB. Membership is restricted to states, and civil society participation mechanisms are informal and restricted to exchanges of information and ideas (Heldt & Schmidtke, Citation2019).

Despite slightly more exclusive membership rules, the AIIB and the NDB uphold the governance features of multilateralism. Especially the AIIB’s rapid growth in membership, including 44 regional and 26 nonregional members, suggests that China and the BRICS coalition do not intend to establish an exclusive system of multilateral development finance.

As regards legalization, we have argued that the system is characterized by formal organizations with a legal mandate and formal lending rules and practices. The AIIB and the NDB do not challenge this feature. Both organizations have a formalized mandate, governance structure, and bureaucracy. Their constitutional treaties have a structure and content similar to those of the WB, and lending rules and practices are codified in by-laws, rules of procedure, and similar formal documents. In his detailed comparison of the AIIB and the WB, Wan (Citation2016) shows that both organizations are not only formalized but that the AIIB has drawn directly on the WB AoA. He shows that the majority of regulations in both founding documents are similar and, at times, even use the same wording. Overall, these similarities indicate that neither the NDB nor the AIIB is designed to challenge the governance feature of legalization.

As we have shown above, market embeddedness has evolved through innovations both internal and external to the WB. Neither the WB nor its regional and sub-regional counterparts focus exclusively on providing long-term finance for national growth strategies, pursuing industrial modernization via infrastructure projects, as the delegates in Bretton Woods intended. Rather, the approach to multilateral development finance has broadened in two important ways. First, financial support no longer exclusively emphasizes infrastructure programs but includes social development, environmental targets, and good governance. Second, financing via direct loans to governments is accompanied by other instruments, including guarantees and cofinancing initiatives with private actors. The AIIB and the NDB are fully in line with this approach. Despite their public emphasis on capital infrastructure projects, both organizations have a multi-sector lending portfolio that supports state and non-state borrowers. The AIIB reports lending projects in six sectors, including water supply, transportation, and energy as the most frequent areas, but urban development and social projects are also part of the portfolio (Asian Infrastructure Investment Bank, Citation2018). Of the 32 projects approved by the end of 2018, 28% did not focus exclusively on classical infrastructure projects but on broader issues of human development. Similarly, the NDB reports projects in five sectors, focusing most strongly on transport and energy but also including agriculture, good governance, and sanitation projects (New Development Bank, Citation2018). Of the 23 projects approved by the end of 2018, almost 35% were directed towards human development and good governance.

Also in line with the WB’s application of the market embeddedness feature, these loans are not unconditional. Both organizations have developed environmental and social safeguards that debtors are required to follow in implementing funded projects. While the NDB primarily supports the use of country systems in the management of environmental and social risks (Serrano Oswald, Citation2018), it has also formulated its own requirements, including, for instance, the protection of minorities and vulnerable groups (New Development Bank, Citation2016). Even more like the WB, the AIIB uses an elaborate system of safeguards, which even explicitly refers to WB conditionality as a source on which the organization has drawn (Asian Infrastructure Investment Bank, Citation2016).

Overall, the new development banks appear to be following a path reminiscent of those pursued since Bretton Woods (Kahler, Citation2016). In contrast to prominent statements by their founders, current institutional design and behavior do not constitute a new type of bank, challenging the Bretton Woods system. Rather, they largely mimic the governance features of the WB (Acharya, Citation2017). Nonetheless, only the future can tell whether current similarities will indeed translate into a similar functioning of the AIIB and NDB.