6   The Changing Institutional Landscape of Financial Governance and Developmental Finance in Emerging Market and Developing Economies

The crises of the 1970s and 1980s generated demands for “South-South” development institutions that would be largely autonomous from the BWIs. In contrast, the Asian and the global crises spawned a new pragmatism reflected in the view that EMDE institutions could complement and, to a much lesser extent, substitute for the BWIs. EMDE institutions are evolving in ways that allow them to fill persistent gaps in the global financial architecture—such as by increasing their ability and commitment to provide long-term project and infrastructure financing and by expanding liquidity support during periods of instability. In short, we find institutional proliferation and expanding mandates that place the EMDEs at the center of an evolving landscape of what I have referred to throughout as financial governance and developmental finance.

As I have by now argued at length, the Asian and especially the global crises precipitated discontinuities in the global financial governance architecture. The willingness and ability of EMDEs to undertake ad hoc and uncoordinated innovation in financial governance in the absence of a theoretical and political consensus around a new paradigm is one of the most important legacies of the recent crises. The innovations are best understood as uneven, partial, experimental, contested, and incomplete. In this chapter, I focus on salient discontinuities—namely, the way in which crises have contributed to decentralized institutional innovations in the domains of reserve pooling and project and infrastructure finance in EMDEs. The innovations do not coalesce around a unified global architecture that rivals the BWIs. Instead, we find an expansion in the reach and mission of disparate institutions that pre-dated the global crisis, as well as the creation of entirely new ones. Taken together, they increase the heterogeneity and density of the financial governance landscape and seed the ground for a more complex, decentralized, and resilient global financial system (Chin 2015; Grabel 2013a; 2013b; Helleiner 2010; Huotari and Hanemann 2014; Mittelman 2013; Riggirozzi and Tussie 2012).¹ Equally important, the initiatives also promise to enhance policy autonomy and thereby widen policy space for development. At the same time, the emerging institutional landscape complicates the terrain in which the BWIs operate; affords EMDE policymakers greater voice than is available to them at the BWIs; increases EMDE exit options and leverage at the BWIs by increasing the opportunity for forum shopping; and creates opportunities for unscripted learning by doing and learning from others. When approached from this vantage point, the developments surveyed in this chapter emerge as enormously promising institutional experiments. The perspective on offer follows Hirschman in refusing to discount small, disconnected, ad hoc innovations for which the pedigree cannot be traced back to a grand, utopian scheme. Instead, these incremental innovations have taken root in the concrete demands facing policymakers, who are adjusting pragmatically to the changing circumstances and challenges they face.

Crises and Interest in Regional and Transregional Financial Architectures in EMDEs

Interest in regional financial architectures in EMDEs is rooted in the development models and economic crises of the past several decades (Ocampo 2006, chap. 1). In South America, a first wave of financial regionalism was associated with the “Grand Designs” of the 1950s and 1960s, a second with the “New Regionalism” or “Open Regionalism” of the 1980s and 1990s, and a recent wave with the “post-hegemonic, post-neoliberal” politics in countries that up until recent elections featured populist and left-leaning governments (Riggirozzi and Tussie 2012). In EMDEs and Europe, economic and financial crises spurred the creation of regional reserve pooling arrangements and safety nets (Rhee, Sumulong, and Vallée 2013). For instance, the creation of the European “currency snake,” Arab Monetary Fund, ASEAN Swap Arrangement, and Andean Reserve Fund is connected to the instability associated with the collapse of the pegged exchange rate system, the oil shocks of the 1970s, and the Latin American debt crisis of the 1980s.

The Asian Crisis

The East Asian and other crises in the 1990s stimulated interest among scholars and practitioners in regional financial mechanisms to deliver liquidity support and long-term project financing through institutions that were to some extent independent of the BWIs. As discussed in chapter 3, the AMF proposal was motivated by the Asian crisis. But the focus on regional financial architectures extended beyond the region. For example, Mistry (1999) argued for the expansion of regional crisis management capacities that would support national and global institutions; Agosin (2001) for a Latin American Fund capitalized by a modest proportion of the region’s reserve holdings (possibly backed by contingency credits from the international banking system), which could supplement IMF initiatives and provide a degree of protection from regional financial crises; and, in the 2002 “Monterrey Consensus,” the UN advanced a case for a multilayered financial architecture in which regional and subregional institutions play a greater role (UN 2002, para. 45).

The Asian crisis and subsequent global economic conditions played a pivotal role in the rise of reserve accumulation strategies by EMDEs. The success of this strategy provided the political confidence and material support for the institutional innovations of the global crisis.

The Global Crisis

The global crisis focused more attention than the Asian crisis on the role and capacity of regional and transregional financial arrangements in EMDEs. The UN’s Stiglitz Commission and a second UN-appointed commission called for a new global monetary system built from the bottom up through a series of agreements among regional arrangements, echoing the embrace of financial regionalism that followed the Asian crisis (UN 2009, chap. 5; 2014). UNCTAD (2011, 115), too, emphasized the role of regional financial architectures. It usefully offered a vaguely articulated concept of “developmental regionalism” to frame discussions of South-South cooperation. Scholars such as Ocampo (2010c) argued that improving economic and social governance requires a multilayered network in which regional and subregional institutions mediate between global and national financial arrangements.² In a similar vein, Wade (2008) discusses institutional “middleware”—software that allows different families of operating systems to communicate with one another in one integrated platform—as an analogy for what is currently lacking in the global financial architecture.

Support for regional financial architectures also came from unexpected actors. In contrast to the hostility with which the IMF greeted the AMF proposal, Strauss-Kahn argued during the global crisis that the Fund should “look at ways to collaborate with regional reserve pools. We … do not see such funds as ‘competitors’ … collaboration could … include Fund resources … as a backstop to regional pools” (Strauss-Kahn 2010b).³ Statements released by the IMF’s IMFC in October 2010 and April 2011 also highlighted the importance of IMF cooperation with regional arrangements (Rhee, Sumulong, and Vallée 2013, 227; Miyoshi 2013, 4). For its part, in April 2009 the G-20 promoted regional financial arrangements when it dispersed a portion of the IMF’s new funds to the main regional development banks (MRDBs) to support new lending (Chin 2010). At the time, Indonesia proposed that a portion of the IMF’s new resources be allocated to the AsDB (ibid.). In November 2010, the G-20 charged finance ministers and central bank governors with exploring “ways to improve collaboration between regional financial arrangements and the IMF,” and a year later it endorsed a broad set of modest principles for cooperation (see G-20 2011b).⁴ As we will see, under Lagarde’s leadership, the IMF has continued to deepen its engagement with regional (and transregional) reserve pooling initiatives.

The global crisis has stimulated heterogeneous financial regionalisms across EMDEs.⁵ Some observers, including myself, see the crisis as having even a broader effect on the financial landscape, moving it in the direction not just of more regionalisms but toward a more complex, fragmented, and pluripolar direction—driven in large part by initiatives from below rather than from above.⁶

Analytical Framework

The diverse institutions that I survey in what follows can be corralled under the following framework: institutional stasis; capacity expansion; hybridization; and institutional creation.

Institutional stasis refers to maintaining rather than expanding precrisis institutional activity and identity. Capacity expansion refers to enhancements in the scale of activity of existing institutions. It is most simply achieved through increased funding by participating governments but also through new revenue streams, expanded geographical reach, or the introduction of novel mechanisms or programs toward achievement of traditional or newly identified objectives.

Hybridization can occur purposely, when an institution decides to reach beyond its existing mission, but also unintentionally, when an institution seeks to maintain its traditional focus but its actions ultimately blur aspects of the institution’s identity. For instance, a development bank that traditionally provides project financing might begin to provide countercyclical financial support during a crisis, or its project support might come to play an important countercyclical role during the crisis. Finally, institutional creation involves transformation of proposals or aspirations into concrete institutions by existing or new parties.

A Selective Survey of Architectural Innovations in EMDEs

There are far too many innovations to examine here, and I make no claim for comprehensiveness or even exploratory depth. Instead, I provide a view from 30,000 feet of a sample of institutions—transregional, regional, subregional, national, and multilateral—the evolution of which is emblematic of developments and aspirations elsewhere. I first consider arrangements that involve reserve pooling. I then turn to development banks that provide longer-term project and/or infrastructure financing. As discussed in chapter 1, these two forms of support constitute what I term developmental finance.

For reasons of space and also because I am focused on initiatives that take the BWIs as their critical point of departure, I do not consider two other types of financial architectural innovations, namely, payment systems and macroeconomic, monetary, and exchange rate integration schemes (on these, see Fritz and Mühlich 2014; 2015).⁷ These, too, evolved considerably during the global crisis, reflecting other modalities of experimentation and discontinuity in the global financial architecture.

Tables 6.1 and 6.2 summarize the results of this survey and anticipate the discussion that follows. Table 6.1 highlights the chief goals and practices of the institutions and arrangements that I survey. Table 6.2 maps their stasis and change during the global crisis.

Table 6.1

Chief Institutional Goals or Practices

Table 6.2 Mapping Stasis and Types of Change during the Global Crisis
Institution or Arrangement		Institutional Stasis		Capacity Expansion		Hybridization		Institutional Creation
CMIM				✓				✓
FLAR				✓		✓
ArMF				✓
EFSD						Created as hybrid		✓
CRA								✓
CAF				✓		Project loans and “fast disbursement and contingent operations” played a countercyclical role
NDB								✓
AIIB & Belt and Road Initiative/Silk Road Fund								✓
BNDES				✓		Support played a powerful countercyclical role
CDB				✓		Support played a powerful countercyclical role
MRDBs				✓		✓
Source: Author’s analysis.

Reserve Pooling Institutions and Arrangements

Reserve pooling, also known as liquidity sharing, is initiated for the purpose of providing precautionary liquidity and/or countercyclical forms of support to members of a pooling arrangement in the event of currency, liquidity, and/or balance of payments pressure. Support is intended to stave off or mitigate stress in a member economy. It can also reduce the severity of a variety of contagion effects experienced by participating nations, while protecting policy autonomy, especially during periods of economic instability. Reserve pooling may take place through a formal institution or agreements among participating central banks. In some cases, participating central banks may physically transfer reserves to an institution that manages them in a common pool and disburses them following application by a member and evaluation of the member’s request. Alternatively, central banks may self-manage their funds domestically and commit themselves legally to contribute funds to a common pool when circumstances warrant.⁸

Decisions by reserve pooling arrangements (and development banks) are made by boards that are either residential or nonresidential. Nonresidential boards are more cost-effective and are widely understood to be more agile and responsive to the particular needs of member countries.

Table 6.3 summarizes key aspects of the reserve pooling arrangements examined here.

Table 6.3

Reserve Pooling Institutions and Arrangements (US$Billions; 2016 Data Unless Otherwise Noted)

The Chiang Mai Initiative Multilateralisation

In 2000, the Chiang Mai Initiative (CMI) built on the failed AMF proposal and a 1977 bilateral currency swap agreement among five Association of Southeast Asian Nations (ASEAN) central banks.⁹ The CMI ultimately involved the ten central banks of ASEAN and the ASEAN +3 countries. The CMI was conceived as “bridge financing and a supplement to IMF-led bailouts” (Grimes 2015, 149). The initial value of the swaps was US$30 billion, and the value of these commitments increased several times. Initially, and at the behest of creditor countries, disbursals in excess of 10% of the credits available to a borrowing country required that it submit to an IMF surveillance program. The threshold for IMF involvement was raised to 20% in 2005.

The global crisis induced the deepening and expansion of the CMI. The arrangement is now intended to address potential and actual balance of payments and short-term liquidity difficulties among its members. In 2009, ASEAN+3 finance ministers agreed to “multilateralize” the arrangement, an idea they had considered since 2005 (Cohen 2012, 44). The arrangement was accordingly renamed the Chiang Mai Initiative Multilateralisation (CMIM). Multilateralization entailed several things. Decisions on disbursing funds from a US$120 billion virtual currency pool would be made collectively. Members pledge reserves to CMIM, which are owned by and remain with member nations. Assistance under CMIM would be disbursed as dollar–local currency swaps among the ASEAN+3 and Hong Kong. Multilateralization was also reflected in the decision to establish an independent secretariat, the ASEAN+3 Macroeconomic Research Office (AMRO). After politically challenging discussions, AMRO was established in Singapore in 2011 and began to operate in January 2012. It is the regional surveillance unit of the CMIM. It “monitors and analyses regional economies and contributes to the early detection of risks [through preparation of quarterly reports on the macroeconomic conditions of the region and member countries], provides policy recommendations for remedial actions, and ensures the effective decision-making of the CMIM” (AMRO website). AMRO’s mandate is more ambitious during crises. It is to provide a timely “analysis of the economic and financial situation of the CMIM Swap Requesting Country; to monitor the use and impact of the funds disbursed …; and monitor the compliance … with any lending covenants to the CMIM Agreement” (Rhee, Sumulong, and Vallée 2013, 233).

The link between CMIM support and IMF surveillance has been a matter of controversy among members from the start. Initially some members, most notably Malaysia, opposed any IMF link (Sohn 2012). But China’s and Japan’s representatives argued that the IMF link was essential to CMIM’s credibility during its formative period. The link also allayed concerns about repayment and reflected the challenging politics of regional surveillance.¹⁰ At the outset, China reportedly favored linking all CMIM disbursals to IMF conditionality (Amyx 2008, 121; Hamanaka 2016).¹¹ Japan also favored the IMF link, since it keeps the IMF (and hence the United States) engaged in the region (Cohen 2012, 52). Paradoxically, this link may have inhibited members from drawing on available resources, given their experience with the IMF during the Asian crisis (Sussangkarn 2011). For committed futilists, the link undermines the significance of CMIM. But to the extent that the threshold for IMF involvement has been raised several times, and that CMIM and AMRO continue to evolve and deepen their relationships with regional and transregional bodies, the significance of the IMF link can be expected to diminish over time.

As the global crisis worsened, CMIM members wrestled with and deepened the arrangement a second time. In May 2012, the size of the swap pool was doubled to US$240 billion, the maturity of the IMF-linked and delinked swaps was lengthened, and the arrangement’s original crisis resolution facility was renamed the Stability Facility. Moreover, in a gesture that signified increasing CMIM autonomy, the threshold for IMF involvement (including for the new precautionary line of credit) was raised to 30% in 2012, with a plan to increase it to 40% in 2014, “subject to review should conditions warrant” (Grimes 2015, 150, fn8).¹² The move to 40% was deferred, and though raised in a speech by the finance minister of the Philippines during a May 2017 ASEAN meeting, no action was taken (Caraballo 2017). Observers suggest that the adjustment to 40% is ultimately likely since the governments of China and Japan reportedly will support it once CMIM’s screening, surveillance, and conditionality are regularized (Katada and Sohn 2014). A new Precautionary Line was also introduced in May 2012. The Precautionary Line is intended as a crisis prevention facility available to members that are at risk of experiencing a funding crisis despite “responsible” macroeconomic and financial management. Precautionary support is to be extended without strict ex ante conditionality. However, members would have to meet a five-point qualification criterion that is based on the “Economic Review and Policy Dialogue Matrix” for all ASEAN+3 members. The indicators under this matrix focus on external position and market access, fiscal policy, monetary policy, financial sector soundness and supervision, and data adequacy (ASEAN+3 2016; AMRO 2016a). Disbursal criteria under the Precautionary Line are under development, as noted in the communiqué of a May 2016 meeting of ASEAN+3 finance ministers and central bankers, which states that the body was “looking forward to further development” of disbursal criteria (ASEAN+3 2016). The communiqué also notes that disbursal would be linked not just to criteria under the Review and Dialogue Matrix but also to reviews of the economic reports of the country and analyses by AMRO, AsDB, and the IMF. These measures indicate the institutional cooperation and complementarity that is an explicit part of the CMIM/AMRO vision.

China, Japan, and Korea provide 80% of CMIM resources, while ASEAN countries provide the remaining 20%.¹³ Poorer countries have privileged borrowing access. Multilateralization also involves an agreement on voting weights, a matter that was contentious. Japan reportedly favored tying voting weights to contributions, mirroring the practice at the IMF and some regional financial institutions (Ciorciari 2011, 939–940). Negotiations ultimately yielded a compromise after smaller ASEAN members raised concerns about dominance by large members. Decisions regarding renewals and disbursals from CMIM are decided on the basis of a weighted two-thirds majority voting system in which each country receives 1.6 basic voting shares plus additional voting shares based on contributions. Reflecting the power and wealth dynamics of the region, China and Japan have the same voting weight, Korea half that weight, and ASEAN countries the residual. Despite China’s and Japan’s significant block of votes, neither alone can veto disbursals.

Despite the evolution of CMI/CMIM since 2000, its development has been described by some observers as “excruciatingly slow” and episodic, driven as it is by crises and irregular contact among officials (Cohen 2012, 61). Skeptics see it as mostly symbolic (Cohen 2012, 45; Grimes 2015, 150). For some, the 2012 decisions were disappointing (Grimes 2015; Haggard 2013). For instance, critics cited the paradoxical nature of the Precautionary Line. On the one hand, it represents a maturation of CMIM. On the other, it was seen to threaten the still limited regional political solidarity by creating two classes of ASEAN members—those with and those without responsible policies (Grimes 2015, 154, 157). It also potentially places China and Japan in the position of determining which countries fall into which category (ibid.).

Some analysts have suggested that great power rivalries, mutual suspicion, and regional security tensions run so deep that the IMF is viewed as a necessary “neutral third party” in the region, especially by China and Japan (Grimes 2011; 2015, 149; see also Cohen 2012). Historical and political mistrust among CMIM members compounds challenges within CMIM and is reflected in the tradition of noninterference in one another’s affairs and in the cautious, institutionally shallow integration often referred to as the “ASEAN way” (Katada and Sohn 2014). Some see the IMF link as reflecting the dominant role of the United States and the IMF (Parisot 2013). Others note that CMIM’s swaps have never been activated, and that as the global crisis emerged, members turned to bilateral swaps between their central banks and those of the United States, China, and Japan (Helleiner 2014b). Later in the crisis and after the CMIM’s Precautionary Line was introduced, Japan, China, and Korea each became parties to large, new bilateral swaps with several ASEAN member states, and Japan and China each opened swaps with Korea (see Grimes 2015, table 1; Mühlich and Fritz 2016). These developments are taken as evidence that CMIM failed its first real test (Cohen 2012; Emmers and Ravenhill 2011; Grimes 2009b; 2009c; 2011; Ravenhill 2010). Others dismiss CMIM since larger members continue to stockpile reserves while the swap pool (especially the portion that is not linked to the IMF) remains small relative to the potential needs of larger countries during crises (Cohen 2012; Kim and Yang 2014).¹⁴

Several observers are also skeptical as to whether AMRO will ever evolve into a true regional surveillance body (Azis 2011; Brown 2011; Grimes 2011; Kawai 2010). Critics cite sensitivities among members, lack of clarity about how its autonomy might be ensured and how its analyses might be used, and what standards it will apply (Grimes 2015; Haggard 2013). Grimes concludes that “progress in AMRO seems to have stalled,” though he also notes that participants in AMRO surveillance procedures report that they have been given privileged access to member economic data and that undisclosed discussions of economic policy have been productive (Grimes 2015, 153; Siregar and Chabchitrchaidol 2013).

Not all observers are skeptical about CMIM and AMRO. Chin (2012, 7) is cautiously optimistic; he sees AMRO’s progress as a possible “second step” in a gradual loosening of the CMIM-IMF link. AMRO is in fact developing in notable ways. As of 2016, 25 members of its 46 person staff were on its surveillance team.¹⁵ Its staff—though still small relative to its mandate of reviewing the economies of 13 member countries—is refining surveillance capabilities. AMRO completes a surveillance paper on each member country every year, though these are not yet publicly available. AMRO, the IMF, and national central banks and finance ministries are developing processes for working together in various ways. To date, the IMF has participated in all key AMRO meetings; AMRO has reached an agreement with the Fund to observe all of its meetings with individual countries; Fund staff have engaged in “outreach and dialogue” with AMRO (Miyoshi 2013, fn25); and central bank governors and finance ministers now meet at AMRO. These processes are likely to enhance the capabilities of AMRO staff to conduct surveillance independent of the Fund.¹⁶ In the interim, this contact could render the IMF-CMIM link more palatable to CMIM members should AMRO represent member country interests with the Fund in an effective and vigorous manner. The maturation of AMRO’s surveillance capabilities and authority could very well reduce or even eliminate the need for an IMF link (Sussangkarn 2011, 213–218). Finally, as an indication of the evolving character of CMIM, ASEAN+3 members signed an agreement in November 2014 to upgrade AMRO to an international organization. The agreement went into effect in February 2016.

It would be naïve to imagine that the CMIM will take the place of the IMF in the region. It is not intended to do so. But this hardly suggests that the CMIM is fated to remain a marginal or subordinate player. The key is to recognize complementarities that can enhance the CMIM’s stature, influence, and relative autonomy over time in ways that promote regional financial stability. From this perspective, for instance, the large national reserve accumulation in CMIM countries, alongside CMIM’s financial resources, increases the capacity and creates productive redundancy in the global and regional safety nets, with vast potential benefits to global financial resilience. Smaller CMIM members may also benefit from the opportunity for forum shopping during crises that CMIM affords. Moreover, IMF-linked swaps through CMIM might eventually be associated with adjustment programs that look substantially different from those negotiated in cases where AMRO officials do not have a seat at the table with the IMF. If, as some now worry, AMRO is unable to acquire influence over the IMF, CMIM and AMRO officials might very well continue to weaken the IMF link by further raising the threshold for IMF involvement in CMIM affairs.¹⁷

These concerns should be taken seriously, but if the global crisis reveals anything, it is that unexpected developments happen just when the need arises.¹⁸ The decisions made in 2009 and 2012, and the ongoing discussions in CMIM and AMRO, underscore the dynamism of the arrangement and policymakers’ commitment to push its institutional boundaries, even if just gradually. What some have described as a disappointingly slow process should be recognized as promising and productive when one considers the historical and geopolitical factors that would seem to doom the enterprise from the start. In any event, it is premature to conclude that CMIM and AMRO will fail to adapt as new demands are placed on them. Contrary to the view of skeptics, CMIM’s limited resources in comparison with the potential needs of larger member countries or the vast resources of some of the region’s central banks, or compared with China’s vast crisis-response package of November 2008, does not indicate insignificance. CMIM is part of an evolving liquidity-support architecture within which its contributions could be consequential. It is not (yet) intended to substitute for other institutions, but the learning, trust, bargaining, and socialization by finance officials that takes place through CMIM may very well create the conditions for more significant cooperation and further institutional development in this and other regions during future crises. The point is that the identities and practices of fledgling institutions are not set in stone from the get-go and are not dictated once and for all by their charters. Instead, they often evolve and adapt in ways that were not anticipated by their founders, especially when they start at a manageable scale and develop in line with their expanding internal resources and the challenges they confront. There is no reason to think that CMIM and AMRO are incapable of that kind of development. In fact, they seem to be evolving along these lines.

The costs of the EU’s failure to address regional surveillance and the Troika’s heavy-handedness were certainly not lost on CMIM members as they gathered in 2012 to expand and deepen CMIM arrangements. That lesson continues to inform CMIM thinking and decision making. Indeed, in 2016, a study of “Troika Financial Assistance Programs in the Euro Area for CMIM’s Future Reference” (ASEAN+3 2016) was completed under CMIM’s auspices. In addition, in May 2016, CMIM and AMRO conducted “test runs” of the IMF delinked portions of CMIM funds under various scenarios, and later in the year conducted the first test run of the crisis resolution facility that is linked to an IMF program. The test revealed important inadequacies, which are now being addressed.¹⁹ These initiatives indicate that CMIM and AMRO continue to deepen their capacities in the postcrisis context.

CMIM’s structure and procedures have been watched closely by policymakers in Latin America (AsDB and IADB 2012), inspiring the 2014 decision to launch a similar initiative by the BRICS, as we will see. Moreover, representatives from AMRO, ESM, the Latin American Reserve Fund, the Arab Monetary Fund, the Contingent Reserve Arrangement of the BRICS, the Eurasian Development Bank, the G-20, and the IMF met for the first time on the sidelines of the fall 2016 meetings of the IMF–World Bank (AMRO 2016b). The meeting served as a forum to discuss views on cooperation among these regional and transregional bodies and with the IMF and regional financial arrangements. The decision was made to convene for further discussion annually (AMRO 2016b), suggesting increased cooperation, the deepening of networks, and the gradual emergence of an increasingly complex financial architecture across the globe.²⁰

The Latin American Reserve Fund

Latin America has the longest history of regional integration efforts and the greatest number of regional and subregional financial arrangements in the developing world (Ocampo and Titelman 2012). The Andean Reserve Fund was founded in 1978. In 1988, the organization changed its name to the Latin American Reserve Fund (Spanish acronym FLAR) when it decided to admit non-Andean nations, the first of which was Costa Rica in 2000. Like CMIM, the FLAR is a regional reserve pooling arrangement that lends to member central banks.²¹ It is designed to respond to transitory liquidity issues in member states. In the event of more enduring structural problems, the FLAR may provide what ultimately can be seen as “bridge finance” while a member seeks support from another institution, particularly the IMF (Perry 2015, 28).

The FLAR maintains five credit facilities, including balance of payments, liquidity, and contingency loan facilities. Contingency loans provide precautionary access to funds for up to six months (with possible renewals) to address internal or external shocks. Access is capped at twice a member’s paid-in capital. There is no provision for prequalification under the contingency line. Balance of payments credits, which provide funding for up to three years and are capped at 2.5 times a member’s paid-in capital, are the most frequently used of the FLAR’s credit facilities, closely followed by the liquidity credit line, which provides resources for one year up to paid-in capital (Ocampo 2015a, 160).²² Poorer FLAR members Bolivia and Ecuador have somewhat higher access to all of the FLAR’s credit facilities (ibid., fn7). The FLAR also supports member central banks and other public institutions through assistance in managing reserves by improving the liquidity of and return on international reserve investments and facilitating public debt restructuring.

Headquartered in Bogotá, Colombia, the FLAR has eight members.²³ FLAR capital comes primarily from capital subscriptions by member central banks, though it twice issued bonds (US$150 million in 2003 and US$250 million in 2006). FLAR members deposit funds as capital contributions with the institution (in contrast to CMIM), which FLAR manages. As of June 2016, the FLAR had subscribed capital of US$3.9 billion, of which almost US$2.8 billion is paid-in. Four members (Colombia, Costa Rica, Peru, and Venezuela) account (equally) for almost 67% of its subscribed capital; the smaller countries account for 8.3% each.

The FLAR is older, has a broader mandate, has a larger permanent staff, but is less well capitalized than CMIM.²⁴ The FLAR is not hampered by some of the leadership and governance issues that plague CMIM. Each FLAR member is assigned one vote. A supermajority of 75% of those present is required for key decisions, such as disbursal of funds. The supermajority scheme becomes more demanding when decisions involve modification of credit limits, terms, and existing agreements. The institution’s straightforward and equitable governance is vital to its ability to respond rapidly and flexibly to support requests, particularly in comparison with more cumbersome IMF support procedures (Ocampo and Titelman 2009–2010, 262; Rosero 2014). The average FLAR approval time for balance of payments and external debt restructuring is 32 days (Carrasquilla 2015). However, the institution’s voting system has at times delayed decisions. The selection of the last few executive presidents, for instance, took several years (Rosero 2014).

FLAR lending decisions appear markedly evenhanded despite stark differences in country size and political ideologies among members (Rosero 2014). Lending is not linked to the IMF. This fact and the FLAR’s equitable voting system contribute to its legitimacy among members. There has never been a default on a FLAR loan despite the absence of traditional conditionality, even though some members have been in arrears to the IMF or have defaulted on commercial loans (Kawai and Lombardi 2012). The extraordinary repayment record reflects the effectiveness of peer pressure among members and their ownership of the institution (Rosero 2014, fn9). FLAR members treat the institution as a preferred creditor. These factors have resulted in a sterling credit rating (indeed, the best in Latin America, not including the IADB) (Ocampo 2015a).

The FLAR has deepened its surveillance capabilities over time. Since 2011, the FLAR has had a macroeconomic monitoring unit, the Division of Economic Studies, which reviews and monitors member performance and economic prospects. A central bank seeking balance of payments support or external debt restructuring is required to present to the FLAR executive president a written report on how it will mitigate the problem that motivates the support request through monetary, credit, exchange rate, fiscal, and/or trade measures. A decision on whether to grant support is then made by the nonresident Board of Directors following consideration of the remediation program by the Division of Economic Studies. To this point, the FLAR has not denied support on the basis of the plans presented, and has generally not required additional adjustment measures beyond those proposed by a central bank requesting support (Velarde 2015, 150).²⁵ Hence, there is no conditionality in the traditional IMF sense.²⁶ But as part of its loan contract, the borrowing country agrees not to impose measures that affect the imports from another FLAR member as part of its balance of payments restructuring process (Rosero 2014, 46). Division of Economic Studies staff assess the balance of payments situation and repayment capacity of countries receiving support (Titelman et al. 2014, fn28). Staff may also make technical visits to the country’s institutions and require reports to the FLAR’s executive president and board (ibid.). The review is expedited in the case of short-term support, such as liquidity and contingency credits. The executive president is authorized to approve these requests without involvement of the Board of Directors (Rosero 2014, 65).²⁷ The streamlined approval process permits very rapid short-term support disbursement.

Over its lifetime (and through September 2016), the FLAR has made 47 disbursements, amounting to roughly US$6.4 billion (Mühlich and Fritz 2016). The FLAR has lent to all of its members except Uruguay and Paraguay. In some cases, the FLAR contributed stabilizing resources when the IMF did not, or when members declined to engage the Fund (Ocampo and Titelman 2012). The FLAR has disbursed balance of payments loans to members during crises (namely, during the debt crisis of the 1980s, the period that followed the Asian crisis, and during the global crisis) in amounts that exceed IMF disbursals to FLAR members (with the only exception being the period 1989–1993) (Ocampo 2015a, 164). Prior to the global crisis, FLAR lending was significant in comparison with IMF lending; indeed, from 1978 to 2003, FLAR loans of US$4.9 billion were almost 60% of the size of the US$8.1 billion in loans from the IMF to countries in the Andean Community (Chin 2010, fn41). Excluding Venezuela, the FLAR disbursed 22% more than the IMF to FLAR member countries from 1978 to 2013 (Ocampo 2015a, table 1). Though FLAR resources are relatively small in the aggregate, they are significant relative to the needs of smaller member states, and lending has been redistributive subregionally (Ocampo and Titelman 2009–2010, 262). The institution’s main direct beneficiaries have been Bolivia and Ecuador (Ocampo 2015a, 163). But mitigation of crises in smaller members induced by balance of payments issues has also benefited the region’s other economies by stabilizing trade flows (Kawai and Lombardi 2012), as has the commitment required of borrowing countries not to take measures that interfere with intra-FLAR trade. For instance, Colombia, an important trade partner, has gained indirectly from repeated FLAR support to Ecuador (ibid.).

The FLAR’s presence reduces the pressure on smaller countries to accumulate reserves and hence the opportunity cost of doing so (Eichengreen 2010). The FLAR has provided important savings to members by making funds available at better terms than are available on international markets to countries under stress (Rosero 2014). In some instances, FLAR resources have been leveraged as part of broader support programs.²⁸ FLAR membership has also been beneficial to members since reserves committed to it have yielded greater returns than those maintained in national reserve portfolios (Perry 2015, 27).

In terms of lending, the FLAR largely maintained rather than expanded its role during the global crisis. Compared with the tumultuous period of 1978–1991, there has been a sharp falloff in the number of loan requests and loans granted (Rosero 2014). The decline reflects Latin America’s relative vitality during the crisis rather than any failure on the part of the FLAR.²⁹ During the global crisis, the FLAR received and acted on requests for assistance to support balance of payments or liquidity from only two members. It provided support to Ecuador on four occasions, for a total of US$1.8 billion, and it approved a loan of US$482.5 million to Venezuela in July 2016.

Looking beyond lending as a metric of change, we find evidence of gradual FLAR evolution during the global crisis and up to the present. Membership broadened, with Uruguay joining in 2009 and Paraguay in 2015. In 2015, Guatemala and the Dominican Republic were formally invited to begin the process of accession (FLAR website, annual report 2015, 25). FLAR members approved a 40% increase in subscribed capital in 2012 (Titelman et al. 2014, fn9). Uruguay and Costa Rica prepaid their entire subscribed capital in 2011 and 2015, respectively; Paraguay increased its paid-in capital in 2015, and in the same year, Costa Rica doubled its subscribed capital (Ocampo 2015a, 160). In recent years, the FLAR has begun to play a more important role in improving the investment conditions of members’ reserves, giving it a role as a regional financial intermediary (Ocampo 2015a, 160). In addition, after more than a decade of dialogue, the FLAR and the IMF agreed to allow a portion of the capital paid in to the FLAR to count toward their international reserves with the IMF (FLAR website, annual report 2015, 27). This double counting reduces the cost of FLAR membership for new (especially small) economies.

The FLAR is insufficiently capitalized to respond on its own to the needs of larger economies, especially during crises that affect several members simultaneously. The FLAR’s potential to expand its capacity is limited by the absence of some of the region’s largest economies, including Brazil, Mexico, Chile, and Argentina. Observers have consequently argued for broadening its membership and deepening its resources (Ocampo 2015a; Titelman et al. 2014). The recent inclusion of and invitations to smaller economies is indicative of an interest in membership expansion. Observers have also argued for instituting larger paid-in quotas for members (Eichengreen 2010; Rosero 2011); establishing contingent credit lines with member central banks and private banks; intermediating funding from or cooperating with the IMF (Rosero 2011; 2014); and connecting it with other subregional, regional, and multilateral institutions (Ocampo and Titelman 2012).

Broadening FLAR membership and capitalization poses challenges. It is difficult to cultivate and sustain ownership in an institution as its membership increases. A larger institution might risk loan defaults with consequent effects on the institution’s credit rating, and may require both new forms of monitoring and even some type of prenegotiated ex ante conditionality (Titelman et al. 2014, 21, fn23). This could be politically challenging. For this and other reasons, Titelman et al. (2014) argue that any prenegotiated conditionality should not involve the IMF or any other outside entity. It may also be difficult to maintain the institution’s governance structure, particularly if larger countries maintain higher paid-in quotas. On this point, Ocampo (2015a, 168–170) argues that the FLAR should introduce new membership categories for the largest and smallest countries and move to weighted voting for larger loans.³⁰

Even if expanded, institutions like the FLAR should be viewed as complementary insurance mechanisms that are part of a global patchwork of financial cooperation. In extreme cases, the IMF could leverage FLAR capital to mobilize a larger pool of resources, or the FLAR could take action in conjunction with other regional institutions (Titelman et al. 2014, 17). We might envision a capacity-based division of labor in which regional mechanisms like the FLAR provide support to small- and medium-sized countries and act independently during localized economic disturbances—something it has already done—while the IMF provides support to large countries and partners with the FLAR during large-scale crises, though without IMF-driven conditionality (as per Ocampo 2006, chap. 1; 2015a, 170).³¹

The FLAR has pursued “strategic alliances” with a range of other institutions, including AMRO, the Development Bank of Latin America, and the BIS. The result is an emerging cross-cutting network of cooperation that stands to enhance the capacity of these partners while generating cross-institutional learning. Connecting FLAR more directly to the IMF, on the other hand, may be politically unpalatable to some members, particularly those few that still have populist governments (Armijo 2012). Cooperation between the FLAR and the IMF to date has involved the participation of FLAR officials in IMF training programs (Miyoshi 2013, 21), the decision to count a portion of FLAR commitments as part of their IMF reserves, and participation (as previously noted) along with other institutions in an October 2016 meeting at the IMF, now to be held annually. More generally, the relationship between the IMF and regional financial arrangements has always been uneasy and complex (Eichengreen 2015, 133–134), as Europe’s recent experience underscores. There is no reason to expect that this would not be the case with the FLAR.³²

The Arab Monetary Fund

The Arab Monetary Fund (ArMF), headquartered in Abu Dhabi, United Arab Emirates, was founded by central bankers from the Arab world and began operating in 1977.³³ It began as the monetary arm of the Arab League. Today it has 22 members and a small amount of subscribed capital, approximately US$3.8 billion as of the end of 2015.³⁴ As with the FLAR, the ArMF takes deposits from member central banks. The ArMF can borrow from members and from Arab and foreign institutions and markets, and can issue securities. Like the FLAR, the ArMF has a broad developmental and financial stability responsibility. ArMF policy mandates include the provision of financial support to members experiencing balance of payments problems; the promotion of exchange rate stability, monetary policy coordination, financial market deepening, intraregional trade, and current account liberalization; and the eventual establishment of a common currency (an early, unrealized political goal of the Arab League).

The ArMF has several lending facilities. Four of its facilities focus on balance of payments needs, three on supporting particular sectors (viz., government finance, the financial system, trade, and oil), and, since 2009, a facility supports countries facing short-term liquidity problems caused by difficulties accessing international financial markets. ArMF loans have varying access limits and are disbursed with varying degrees of oversight. The institution appears from the one relevant study to be extraordinarily nimble. McKay, Volz, and Wölfinger (2011, 21) report that some types of loans, such as those available under the new short-term liquidity facility and what are termed automatic balance of payments loans, are disbursed very quickly and carry no requirement of a country mission or conditionality. Management makes a decision following rapid preparation of an internal report, with later notification to the Executive Board (see also Mühlich and Fritz 2016, 15). Other types of loans are generally approved in one to six weeks and require an adjustment program agreed to by the member state and the ArMF (McKay, Volz, and Wölfinger 2011, 21). For example, loans under the extended loan facility support structural balance of payments problems, provide the largest amount of support, and have a repayment period of up to seven years. Extended loans involve consideration by the Executive Board and require a mission, an adjustment program, conditionality and monitoring by the ArMF, and supplementary support from other regional and multilateral institutions, such as the IMF (Corm 2006; McKay, Volz, and Wölfinger 2011, 21). Conditions on ArMF loans tend to be less stringent than those associated with the IMF (Corm 2006, 309; UNCTAD 2007,122; UNCTAD 2015b, 74).

The ArMF has a technical staff that observers consider highly competent (McKay, Volz, and Wölfinger 2011; Miyoshi 2013).³⁵ Staff members conduct reviews of member country economic conditions and financing needs (ibid.). However, some analysts question whether monitoring is sufficiently stringent (McKay, Volz, and Wölfinger 2011), though loan arrears remain small and are associated with countries facing particularly difficult political and social conditions (e.g., Somalia, Syria, and Sudan).

The ArMF’s governance structure is not unlike that of the BWIs (and the MRDBs). Decisions of the eight-member Executive Board are by absolute majority, with votes weighted by size of member contribution. Three countries (Saudi Arabia, Algeria, and Iraq) together hold 38.5% of the votes. That these countries are overrepresented underscores the point that governance of EMDE regional institutions is not inherently more egalitarian than that of the BWIs.

From its establishment through the end of 2015, the ArMF has made 174 loans to 14 member nations, totaling US$8.2 billion (ArMF website, annual reports).³⁶ Average drawing volume tends to be very small, and smaller, oil-importing members have been the most frequent users of lending facilities (especially in the 1980s) (Mühlich and Fritz 2016). About 63% of the loans were related to balance of payments pressures. The ArMF was faced with growing demands on its resources stemming from the challenge of the global crisis, the Arab Spring, and rising food and falling oil prices. During 2009 the ArMF made five loans totaling US$470 million via its new short-term liquidity facility. Between 2009 and 2015, the institution approved a total of 33 loans to eight countries, totaling US$3.5 billion. Moreover, the dollar value of loans extended during each year of the period 2009–2013 exceeded that for any other year (except 1988) since the institution began to operate. The US$800 million in loans extended in 2015 represents a new peak for the institution.

The ArMF has no formal relationship with the IMF. The IMF has provided technical assistance to the ArMF on domestic bond market development (Rhee, Sumulong, and Vallée 2013). The ArMF also coordinates with the IMF on technical workshops (ArMF website, annual reports, table B-2) and took part in regular meetings of the IMF and World Bank during 2015. The institution’s Articles of Agreement charge it with providing “complementary” lender of last resort financing for some types of loans. Borrowers are expected first to withdraw their reserve tranche from a regional organization, the IMF, or another multilateral body before seeking support from the ArMF’s ordinary, extended, and structural adjustment facilities (Miyoshi 2013, 31–32). This explicitly complementary role is necessitated by the ArMF’s small capitalization and is reflected in the frequent parallel use of the IMF and ArMF. Since its creation, parallel use has occurred on 22 occasions, mostly during the Arab Spring (Mühlich and Fritz 2016, 23). The ArMF’s resources and lending could obviously be increased significantly to provide more support to its poorer members, given the vast assets possessed by some of its oil-exporting members.

The Eurasian Fund for Stabilization and Development

The member countries of the Eurasian Economic Community created the Eurasian Fund for Stabilization and Development (EFSD) in June 2009.³⁷ Until 2015, it was known as the Eurasian Economic Community Anti-Crisis Fund. Since its founding, the Fund has operated as a hybrid that involves features of reserve pooling and development banking. The EFSD serves as a regional safety net that extends what it terms financial credits to governments to offset the effects of the global crisis; funds stabilization programs by supporting budgets, balance of payments, and currencies; and ensures the long-run economic stability of member nations. The EFSD also provides what it terms investment loans to governments and firms for large interstate projects that support regional integration or national investment and has a new program of grants aimed at supporting social programs.

The EFSD was established with subscribed contributions of US$8.5 billion by six countries.³⁸ It has paid-in contributions of US$2.8 billion, most of which comes from Russia, its largest member.³⁹ Paid-in contributions to the EFSD come from pooled member resources via budget contributions. At present, it has no capacity to issue bonds or to otherwise tap financial markets (Rhee, Sumulong, and Vallée 2013). Votes at the EFSD are weighted by capital contributions (as per the ArMF, the BWIs, and the MRDBs). Russia holds 85% of the votes and consequently holds veto power.

The Eurasian Development Bank manages EFSD resources and conducts surveillance of EFSD borrowers (Rhee, Sumulong, and Vallée 2013, 224).⁴⁰ There are no automatic disbursals of financial credits from the EFSD, and all disbursements are tied to a heavily and regularly monitored adjustment (i.e., conditionality) program. Financial credits are followed by consultations intended to determine the likelihood of borrower success in implementing reforms or stabilization programs that are funded by the EFSD. Recipients are not required to work with the IMF, though the EFSD claims that it is “guided” by the IMF in matters relating to financial credits. It also uses IMF benchmarks when assessing various matters, such as corporate governance. Indeed, an EFSD annual report notes that the manager “consulted with the IMF on a regular basis regarding economic policy guidelines for Armenia, Belarus, Kyrgyz Republic, and Tajikistan” and that EFSD officials have been discussing coordination initiatives with the AsDB, World Bank, and IMF since 2014 (EFSD website, annual report 2014,12). The EFSD does not extend financial credits to countries that are in arrears to the IMF, other multilateral institutions, or EFSD members. However, in the case of Belarus, the EFSD extended a financial credit to the country when the IMF declined to do so. Decisions on financial credits by the EFSD are rapid—available evidence suggests that internal decisions on loan disbursements are made in two to eight weeks (Mühlich and Fritz 2016, 15).

To date, the EFSD has extended only four financial credits, totaling almost US$3 billion (to Tajikistan in 2010 and 2016, Belarus in 2011, and Armenia in 2015). Its largest extension of financial credits to date was to Belarus—its support package of US$2.6 billion was equal to almost 6% of the country’s GDP.⁴¹ The 2010 support package of US$70 million to Tajikistan was equal to about 1% of its GDP. The case of Belarus suggests that EFSD surveillance has teeth: disbursal of a sixth tranche of funding was postponed from 2013 to 2015 because the country missed stabilization targets established in its agreement with the EFSD. Support was reestablished in 2016.

The Contingent Reserve Arrangement of the BRICS

The BRICs acronym was developed in 2001 by Goldman Sachs’ Jim O’Neill, who saw Brazil, Russia, India, and China as having broadly similar characteristics. O’Neill predicted that these countries would have higher levels of economic interaction with one another (O’Neill 2001). The marketing label had performative force: it is credited with spurring coalition building, the creation of networks of influence inside and outside the group (e.g., within the IMF and the G-20), and institutional innovations (Ban and Blyth 2013).

Beginning in 2006, Russia promoted informal meetings of diplomats from the four countries.⁴² The first recorded meeting among the group’s policymakers took place when their foreign ministers met on the fringes of the UN General Assembly in the fall of 2006. Since 2006, BRICs officials have deepened, broadened, and formalized discussions of financial governance. The first BRICs leaders’ meeting was held in Japan on the eve of the 2008 G-8 summit, and the first stand-alone BRICs Leaders’ Summit took place in Russia in 2009. Since then, leaders have been meeting annually at the group’s Leaders’ Summits and informally on the sidelines of international meetings, while heads of ministries have also met regularly. BRICs executive directors at the BWIs also meet informally. Intensified intra-BRICs cooperation in international finance after 2009 led to broader dialogue and the development of structures to institutionalize cooperation among officials in other policy areas, such as agriculture, national statistics institutes, and foreign policy. South Africa was invited to join the group at the third Leaders’ Summit, in China in 2011, and the group has since been known as the BRICS.

Since 2011, the BRICS has moved rapidly to develop plans and launch initiatives to create new financial institutions. The first BRICS-level financial initiative was the announcement in mid-2011 of plans to establish joint listing of one another’s stock index futures and other derivative indices on their respective stock and securities exchanges. This came to fruition in March 2012, when the five founding members of the BRICS Exchanges Alliance began cross-listing benchmark equity index derivatives. In 2011, BRICS leaders promised to strengthen cooperation among their national development banks. Plans for financial cooperation became more ambitious in 2012, when the group began to discuss formation of a development bank that would supplement existing institutions. In 2012, BRICS finance ministries also agreed to encourage trade between members, denominated in bilateral currencies.

Intra-BRICS cooperation took a step forward at the July 2014 Leaders’ Summit in Fortaleza, Brazil. In what became known as the Fortaleza Declaration, the group announced that it had reached agreement on two initiatives—the founding of a reserve pooling arrangement called the Contingent Reserve Arrangement (CRA) and the New Development Bank (NDB) (the latter will be discussed later). Long-standing frustration with the BWIs was explicit in the Fortaleza Declaration, which stated that, “International governance structures designed within a different power configuration show increasingly evident signs of losing legitimacy and effectiveness” (BRICS 2014b). Notwithstanding these frustrations, the declaration also made clear that both the CRA and the NDB were to be complements to, and not substitutes for, the BWIs.

The CRA is a reserve pooling arrangement meant to provide liquidity protection (including precautionary support) through currency swaps to members during balance of payments crises.⁴³ China has pledged US$41 billion to the CRA’s US$100 billion pool; Brazil, India, and Russia have each pledged US$18 billion; and South Africa has pledged US$5 billion. Pledges by China, Brazil, India, and Russia to the CRA are nearly equal to each of their IMF quotas. At this point, most CRA capital is callable. No single member is to have effective veto power over fundamental changes in the CRA. As of this writing, the criteria to be used in decisions pertaining to qualification for support under both the liquidity and the precautionary facilities are still under development. Support decisions will be made by a “Standing Committee” comprising five directors appointed by each of the member country central banks. Countries applying for support from the CRA (including for precautionary support) in amounts above 30% of their eligibility must be in compliance with the surveillance and disclosure obligations of the IMF’s Article IV (sections 1 and 3) and Article VIII (section 5), and they may not be in arrears to BRICS nations or to regional or multilateral institutions (Henning 2016, 125–126; BRICS 2014c, Article 14(b)(v)). The most controversial aspect of the CRA rests precisely in the decision to replicate the CMIM-IMF link.

It is unrealistic to treat the BRICS as a serious challenge to the roles of the United States and the IMF in global financial governance.⁴⁴ Instead, it should be seen as a group that occupies an “intermediate space” in global interstate power (Armijo, Katada, and Roberts 2015; Armijo and Roberts 2014) and as creating a “parallel order” rather than one that rivals the United States and the BWIs (Stuenkel 2016a; 2016b). The group’s ability to cooperate is rooted in part in its potential to serve as a counterweight to traditional powers.

That said, important sources of tension that sometimes yield conflict persist among BRICS members (Chin 2014a). The BRICS group has often had to overcome or, more accurately, work around important differences and persistent fissures to reach consensus on some matters of global financial governance. The Russian government, in particular, has not always or easily fallen into line with positions taken by other members, and it has used its membership in the BRICS instrumentally in its conflicts with other powers (e.g., as a way to insulate itself from clashes with the United States and Europe over its policy in Ukraine). Indeed, China and Russia oppose adding the other BRICS to the permanent membership of the UN Security Council because this would reduce their own influence (Bond 2016, 615).

The launch of the CRA triggered an avalanche of commentary that broke down along the lines of Hirschman’s possibilists and futilists. Futilists dismissed the “empty symbolism” of the CRA, emphasizing the decision to replicate the CMIM-IMF link, the small size of CRA resources relative to potential demands, and the dollar-based funding commitments to the CRA that reinforce the currency’s dominant global role.⁴⁵ More broadly, skeptics emphasize what they see as fatal internal tensions that will continue to disrupt the group’s cohesiveness and its potential to transform financial governance.⁴⁶ Others emphasize the “sub- or neo-imperial” tendencies of the BRICS, while still others dismiss the significance of the BRICS because their growth prospects have slowed.⁴⁷

Possibilists are not persuaded. In the possibilist view, the CRA (warts and all) is part of an evolving, fragmenting global financial landscape in which institutional experimentation is becoming the “new normal.” From this perspective, the CRA is understood to complement existing institutions and advance the growing disbursal of economic power while holding the potential to increase the voice of EMDEs in the global financial governance architecture either directly or through the leverage associated with forum shopping.⁴⁸ For possibilists, the CRA is one among many parallel experiments that provide opportunities for learning, problem solving, and deepening networks of influence.

Possibilists insist that it is premature and unhelpful to dismiss in advance the CRA and other BRICS initiatives. It is of course true that the impact of the BRICS and their various initiatives will be uneven and even contradictory, reflecting enduring tensions within each of its member states, among its members, and between those members and other actors (states and institutions). But that is equally true of all complex institutions and their endeavors—they are not adequately described by exclusive reference to their formal mission statements or just one aspect of their practice. It is also true that critical issues must be addressed before the CRA begins to disburse funds. China’s economy is larger than all of the other BRICS combined (Chin 2014a). It will be important for China’s voice at the CRA not to dominate those of other members, something that may be complicated by the size of its contributions (Chin 2014a; Desai and Vreeland 2014). In addition, the CRA (like the CMIM) will have to address the difficult issues of surveillance and its relationship to the IMF.⁴⁹ But these obstacles are not insurmountable, and the motivation to overcome them is high given the degree to which the CRA promises results that are not presently achievable elsewhere. There is good reason to expect that the CRA will ultimately develop independent, well-resourced, and technically competent surveillance capacity over time, and as that occurs, the IMF link may lessen or be eliminated. For these reasons, the CRA carries significant potential to catalyze widespread change through its own internal performance, through competition or cooperation with other pooling arrangements, and through the example it sets for other institutions (Griffith-Jones, Fritz, and Cintra 2014).

Development Banks

Development or infrastructure banks are specialized financial institutions that channel large sums of long-term financing for investment in capital-intensive projects, and public and private enterprises and particular sectors (such as infrastructure) that generate positive externalities but that are likely to be underfunded by private banks.⁵⁰ Development banks are generally mandated to provide credit at terms that would render industrial and infrastructure investment viable. They also perform other functions, such as providing technical assistance to borrowers. Development banks differ in many ways. For example, they diverge in terms of the size of their assets, overall significance in the economy, source of funds, types of projects that are funded, and whether they focus on long-term lending to large industries and infrastructure or on realizing special policy mandates.

Development banks in EMDEs and the MRDBs were created over fifty years ago to address the shortage of project and infrastructure finance. EMDEs host the majority of development banks. A recent study found that as of 2015 there were over 250 national development banks across the world, holding at least US$5 trillion in assets (Gallagher and Sklar 2016). The Asia and Pacific region hosts the largest number of national development banks (119), followed by Latin America and the Caribbean (63), Africa (61), the Middle East (45), and Europe and North America (15) (ibid.).

In the 1950s and 1960s, over 70% of the World Bank’s IBRD and IDA operations were directed toward infrastructure. But by 1999 only 19% of IBRD and IDA operations targeted infrastructure (Humphrey 2015a, 3). The shortfall in infrastructure spending and financing emerged and widened in the 1980s and 1990s as Western bilateral donors and the MRDBs focused instead on social and policy-related lending (e.g., poverty alleviation) (Chin 2014a, 368; Humphrey 2015a; 2015b). Over the last decade, IBRD and IDA lending for infrastructure rebounded to 30%–40% of activity (Humphrey 2015a).⁵¹ Despite this rebound (from a low level), the exigencies of the global crisis and the freezing of private markets aggravated the infrastructure financing gap (Chin 2014a). In the crisis context, the World Bank, MRDBs, and some national, subregional, and regional development banks based in EMDEs took on roles that we traditionally associated with institutions that focus on liquidity support, such as the IMF and FLAR. Development banks introduced and, where such facilities previously existed, significantly increased disbursements of shorter-term loans and other forms of financing (such as trade credits) that had countercyclical effects. Indeed, a recent World Bank survey of 90 development banks across the world highlights the important and often overlooked countercyclical impact of these institutions (de Luna-Martínez and Vicente 2012).

The inadequate provision of infrastructure financing remains a critical deficiency of the global financial architecture (Chin 2014a). A widely cited study of the infrastructure investment gap in EMDEs estimates it at US$1–1.5 trillion per year over the next twenty years (Bhattacharya and Romani 2013).⁵² Other studies suggest that by the 2020s EMDEs will need US$2–3 trillion annually to support sustainable infrastructure investment that would be sufficient to achieve the SDGs (Bhattacharya, Oppenheim, and Stern 2015; Bhattacharya and Holt 2017).⁵³ To put the infrastructure investment gap into context, note that in fiscal year 2014 the World Bank Group spent US$24 billion on infrastructure, which represented an increase from US$16.7 billion in the previous year (UNCTAD 2015b, 172). The recent creation of two new development banks that focus on sustainable infrastructure and a range of other infrastructure-focused initiatives led by China can only be understood in the context of the infrastructure gap.

The world of development banking has its own terminology. Authorized capital represents the maximum amount of capital that an institution can raise, as determined by the maximum number of shares that it can issue, as set forth in its Articles of Agreement. Member nations commit callable capital to the institution as a guarantee in the event that these resources are needed. The World Bank introduced callable capital as a feature of multilateral development bank operations, and today the World Bank (and the IADB) has a higher share of callable capital than do regional development banks based in EMDEs. This allows the World Bank to issue bonds at some of the lowest rates in the world, for sovereign or nonsovereign issuers (Humphrey 2014). Paid-in capital is the amount of up-front capital actually contributed by member nations. Subscribed capital is composed of paid-in and callable capital. Paid-in capital, together with reserves that are built up as a consequence of the institution’s operations, constitutes the institution’s equity. Finally, the ensuing discussion sometimes refers to the World Bank Group, which includes all five of the institution’s organizations, and sometimes to the IBRD or IDA, the World Bank Group’s two main lending arms, which together constitute the World Bank. We use the less cumbersome and more common term World Bank in cases where the lack of precision is not consequential to the discussion.

The terrain of development banks is vast, and hence the discussion of particular institutions in what follows is intended to be illustrative rather than comprehensive. Table 6.4 summarizes key aspects of the development banks examined here.

Table 6.4

Development and Infrastructure Banks and Initiatives (Selected, US$Billions, Unless Otherwise Noted)

The Development Bank of Latin America

The Development Bank of Latin America (formerly the Andean Development Corporation; Spanish acronym CAF) was created to support development and integration of the Andean Community countries. The renaming of the institution is consistent with its broadening reach throughout Latin America and the Caribbean.

The CAF was launched in 1970.⁵⁴ Its membership and the focus of its loan programs have broadened considerably over time. The CAF is a multilateral, regional development bank that focuses on mobilizing medium- and long-term lending for productive investment and fostering regional integration. It lends to governments, the private sector, and public institutions. The CAF raises funds from shareholder countries and through commercial paper activities in the United States and Europe. A large majority of its funds (69% in 2015) come from bond issues sold in international and member country capital markets. In 2015, the CAF raised US$3 billion from bond sales.

Today the CAF has 19 member countries from Latin America and the Caribbean, plus Spain and Portugal.⁵⁵ Its newest member, Barbados, became a shareholder in 2015, and Trinidad and Tobago completed the process of becoming a full member in the same year. The CAF is headquartered in Caracas, Venezuela, and has 12 other offices and a nonresident Board of Directors. Member countries are both clients and shareholders of the CAF. The CAF is owned almost exclusively by EMDEs. Members own most of its assets, which is in marked contrast to the ownership structure of other regional multilateral lenders. Borrowing shareholder countries hold nearly all of the votes at the CAF. By comparison, just 37% of the votes at the World Bank are held by borrowing shareholder countries (Humphrey 2014, 616). CAF administration has emphasized that AEs will never be allowed to hold more than 10%–15% of its shares to preserve its governance structure and mission (ibid., 620).

At year-end 2015, the CAF’s total assets were valued at US$32.5 billion, an increase of 51% over 2011. Shareholders’ equity in the institution increased by 50.7% over the same period (CAF website, author calculation). Authorized capital in 2015 was US$15 billion and subscribed capital US$5 billion. The CAF is among the most dynamic of all multilateral development banks in terms of lending volume. The CAF was relatively marginal in regional lending during the 1970s and 1980s, when the World Bank dominated lending to the region, followed by the IADB.⁵⁶ This began to change in the 1990s, as World Bank and (to a lesser extent) IADB shares of lending to the region fell. CAF lending grew fourfold between 1991 and 2007; during the same period, lending by the IADB doubled and World Bank lending to South America grew by only 40%. Between 2000 and 2011 CAF contributed 56% of total multilateral financing to Andean Community members.

The CAF lends broadly throughout its membership. Andean countries receive a substantial share of CAF loans. From 2011 to 2015, Bolivia, Colombia, Ecuador, Peru, and Venezuela received 56% of loan disbursals. In contrast to the FLAR, a significant percentage of recent CAF loans have gone to larger countries. From 2011 to 2015, for instance, 30.3% of its disbursed loans went to Argentina, Brazil, and Mexico (CAF website).

In 2015, the CAF funded almost as much infrastructure in Latin America as the IADB. At year-end 2015, 62.3% of the CAF loan portfolio was in infrastructure. In 2015, the CAF approved $3.3 billion in loans for infrastructure (which represented 27% of approved loans), whereas the IADB approved US$4.3 billion for infrastructure and environment (which they report as a combined category). Almost half of CAF-approved disbursements from 2010 to 2015 were in the form of medium- and long-term loans. Given the scarcity of medium- and longer-term finance in EMDEs, the CAF’s role is significant. CAF lending was particularly important during the global crisis insofar as funds for longer-term project finance in EMDEs contracted severely, especially as private lenders fled these markets.⁵⁷

CAF finance has important countercyclical and developmental impacts since it provides stable funding to members. CAF financing dampened instability during the EMDE financial crises of the 1990s and the global crisis, when CAF loans remained high. Annual loan approvals ranged from US$9.1 billion in 2009 to US$11.7 billion in 2014. During 2011–2015, the CAF averaged US$11.1 billion in lending annually, compared with US$8 billion from 2006 to 2010. In 2015, the CAF increased its countercyclical activity through what it called fast disbursing and contingent operations of US$2.4 billion. In 2015, the CAF approved a record volume of loans of US$12.3 billion (and disbursed US$5.9 billion), while the IADB approved loans of $US11.1 billion (and disbursed US$10.2 billion).

Despite CAF dynamism, it has by no means displaced the World Bank and the IADB in regional crisis response capacity (see Table 6.4). In 2008, 2009, and 2010, CAF loan approvals were US$7.9 billion, US$9.1 billion, and US$10.5 billion, respectively. By contrast, in the same years, the World Bank approved loans to Latin America of US$4.6 billion, US$14 billion, and US$13.9 billion, while the IADB approved loans of US$11.2 billion, US$15.5 billion, and US$12.4 billion (Ocampo and Titelman 2012, 15). Later in the crisis, the gap between CAF loan approvals and those of the World Bank and the IADB to all of Latin America and the Caribbean narrowed. In 2012, for instance, the CAF approved US$9.3 billion, while the World Bank approved US$6.6 billion and the IADB US$11.4 billion in loans to the region (Humphrey 2014, 615).

As with the FLAR and some ArMF lending facilities, the CAF loan approval process is very rapid compared with that of the World Bank and the IADB.⁵⁸ The average time to loan approval by the CAF is usually under three months (Humphrey 2015a, 17). Rapid approval stems from the institution’s streamlined procedures, in which loans and technical assistance up to fairly large limits are approved by the CAF’s senior management rather than its full membership, following a rigorous project evaluation at several stages in the project life cycle. Agreed-on borrower commitments are designed during this process. However, failure to meet these commitments generally does not result in the suspension of loan tranches. Rather, failure may trigger additional technical assistance by the CAF. The CAF tends to emphasize those actions that can be implemented by a country’s executive branch or that are likely to achieve legislative approval.

Also like the FLAR, country ownership (both literally and figuratively) in the CAF accounts for its very high loan recovery rate. During the financial crisis of the late 1990s, Ecuador’s government continued to service its CAF debt while disappointing other creditors, and when Peru’s government limited debt service payments to 10% of export earnings during the 1980s, it continued to service debt to the CAF in full (Griffith-Jones, Griffith-Jones, and Hertova 2008). Indeed, there were very few defaults on CAF loans from 1999 to 2003, despite the fact that the region faced severe difficulties. As a consequence, the CAF’s credit rating is investment grade and is higher than those of its individual member countries. Moreover, the speed of loan approvals and the close relationship with borrowing countries accounts for the proclivity of Andean member countries to borrow from the CAF, even though its loans are somewhat higher in cost than those offered by the World Bank and IADB (Humphrey 2014, figure 1). The higher cost of CAF loans stems from the CAF’s somewhat lower credit rating than the World Bank and IADB, which in turn derives from the fact that the World Bank and IADB have a high share of callable capital guaranteed by highly rated, nonborrowing AEs (ibid., 618). The CAF often cofinances loans with the World Bank and the IADB.

Member nations’ commitment to the CAF is also apparent in the ease with which the institution raised capital from its members during the global crisis. One might have expected national policymakers to withdraw from multilateral commitments. Instead, early in the crisis, CAF officials were able to tap shareholders for additional paid-in capital contributions. Shareholders quickly and unanimously approved a US$2.5 billion paid-in capital increase in August 2009 (Humphrey 2014). A CAF Treasury official spoke of the ease of securing this capital increase: “We asked shareholders for US$1 billion … they gave us US$2.5 billion. We can increase the capital base whenever we need to, because the structure of our shareholders is so different. It’s not like that with the IADB and the World Bank, where some shareholders contribute capital and others benefit by taking out loans” (quoted in Humphrey 2014, 628). Humphrey observes that “including paid-in capital from new members, CAF raised US$4 billion in paid-in capital following the 2008 crisis, almost as much as the vastly larger World Bank (US$5.1 billion) and more than twice as much as the IADB (US$1.7 billion)” (ibid.). In 2015, CAF shareholder countries again approved an increase in paid-in capital (of US$4.5 billion) (CAF website, annual report).

The CAF issues a large percentage of bonds in Latin American currencies, which are held by regional and international investors. For example, in June 2004, the CAF issued bonds in Colombian pesos (which was a first for Latin America), and it did so again in December 2008 and April 2009. More recently, it issued bonds in Peruvian, Mexican, Venezuelan, and Uruguayan currencies.⁵⁹ In 2007, 32.9% of CAF bonds were issued in Latin American currencies; in comparison, the IADB issued just 14.9% of its bonds in Latin American currencies in the same year (Ocampo and Titelman 2009–2010, table 2). During the global crisis, the CAF introduced and utilized two new financial products to support infrastructure finance, “Collateralized Infrastructure Debt Obligation,” a securitized debt obligation for which infrastructure loans serve as collateral, and “Debt Funds for Infrastructure.” The latter were used in Colombia in 2014 (issued in Colombian pesos and indexed to inflation) and Uruguay in 2016 (issued in Uruguayan pesos and dollars, and also indexed to inflation). Local currency bonds reduce exchange rate risk for the CAF and borrowing countries. More importantly, it promotes the development of local currency bond markets, something that has positive spillovers from the perspective of financial resilience, stability, and access to long-term credit. It also mitigates the locational mismatch that plagues so much lending to EMDEs. The CAF has also signed cooperative agreements with the Green Climate Fund and the Global Environment Facility, reflecting its increasing emphasis on sustainable financing.

The New Development Bank

The BRICS Fortaleza Declaration of 2014 also announced the launch of the NDB. As with the CRA, the NDB was motivated by deep frustrations with the governance of the BWIs.

The NDB is designed to finance investment in infrastructure projects and more sustainable development (including sustainable infrastructure) in the BRICS, with an eye toward allowing other low- and middle-income EMDEs to buy in and apply for funding in the future. It has authorized capital of US$100 billion and subscribed capital of US$50 billion (with each of the five founding signatory countries contributing US$10 billion); 20% of subscribed capital is paid in (and will be paid in over a seven-year schedule), and the rest is callable. The initial size of the NDB’s resources is notable. By comparison, the IBRD had subscribed capital of US$263 billion as of June 30, 2016, of which only $15.8 billion is paid in. China’s contribution of US$10 billion to the NDB is not much smaller than its contribution to the World Bank, and the contributions by each of the other BRICS to the NDB exceed their contributions to the World Bank.⁶⁰

The NDB approved its first loans in May 2016. By year-end 2016, it had approved a total of seven loans, collectively amounting to US$1.5 billion. Each of its member nations was approved for one of its first five loans, and an additional two loans were approved later for China and India. The loans were extended to public sector entities in each of the countries to support small-scale renewable energy and transportation-related projects, and were financed by “green” renminbi (RMB)-denominated bonds issued in the Chinese market. As of November 2016, the institution reported that it had received approval from member governments to develop local currency bond offerings in the Indian, Russian, and South African markets (NDB 2016a; 2016c).

The five members of the NDB hold equal votes in the institution, and its presidency will be held for five years and will rotate among the five members. The first president is Indian; the chair of the Board of Governors is Russian; a Brazilian chairs the Board of Directors; the headquarters is in Shanghai, China; and Johannesburg, South Africa, will host the African regional center. There are also plans for a regional office in Brazil. The location of the NDB’s headquarters in China was reportedly a matter of controversy, but the decision was reportedly made after China agreed to contribute a large share of the CRA’s initial funds (Chin 2014a).

The Articles of Agreement of the NDB are in important respects similar to the governance model of the World Bank and the MRDBs. This is notable in light of what its founders identify as its break with prevailing norms (Humphrey 2015b, 23). For example, each capital share translates into one vote (ibid., 8). However, unlike the legacy institutions, the NDB has a nonresident board (BRICS 2014a, article 12(g)). The BRICS countries have assured themselves of a dominant position in the institution, even as new members join. Contributions by new members cannot reduce aggregate voting power by the BRICS to below 55% or provide any new member with more than 7% of aggregate voting shares (ibid., article 8). Nonborrowing members can never have more than 20% of total voting shares (ibid.). Moreover, the institution’s president and vice president must be from the BRICS (ibid.). This governance structure may inhibit expansion of the institution’s membership. So, too, might the likelihood that the NDB will not be awarded a AAA credit rating based on the bond ratings of the BRICS countries themselves and the fact that it does not have any AEs among its shareholders (Humphrey 2015b).⁶¹

Despite these obstacles, some analysts suggest that NDB loans could dwarf those of the World Bank in the next several decades, especially if membership is broadened and the institution cofinances loans with governments and private investors (Desai and Vreeland 2014). In terms of cofinancing possibilities, the NDB signed memoranda of understanding with the CAF and the World Bank in September 2016, and with BNDES in April 2017. The NDB’s loan portfolio capacity is projected to reach about US$45–65 billion by 2025 (Humphrey 2015b, figure 5).

Along with the initiatives led by China, the creation of the NDB should be understood in part as a response to the vast need for infrastructure spending and finance. Taken together, these initiatives are apt to have catalytic effects on the World Bank and the AsDB (and other MRDBs). Certainly, the rush to create infrastructure finance facilities within the legacy institutions and the G-20 suggests that this is the case, particularly since attention has returned to the infrastructure gap as a central obstacle to development.⁶² Chin (2014a), for example, suggests that the World Bank may have renewed its attention to infrastructure finance partly in response to the perceived threat posed by the NDB.⁶³

Risk management is a key challenge for the NDB. The NDB must ensure that its loan portfolio is of high quality; that loans are carefully chosen and monitored for their developmental, social, and environmental impacts; that finances are sustainable; that the cost of borrowing from it is not prohibitively high; and that China’s influence does not dwarf all others (Chin 2014a; Desai and Vreeland 2014). That the institution’s first loans in 2016 were for renewable energy is notable; maintaining such a focus is essential to the institution’s credibility and progressive impact. As of August 2016, the institution disclosed an “Environmental and Social Framework,” though it was developed without civil society input or consultation. Civil society groups note that while the principles embodied in the framework document are commendable, the document does not advance sustainability criteria or robust social and environmental safeguards (Coalition for Human Rights in Development 2016). Moreover, the matter of incorporating new members into the NDB will have to be addressed. That is a lot to manage, and there is good reason to expect a series of failures and missteps as the NDB matures and expands. That is simply the nature of institutional development.⁶⁴

Initiatives Led by China: The Asian Infrastructure Investment Bank and the Belt and Road Initiative/Silk Road Fund

Simultaneous with its involvement in BRICS initiatives, the Chinese government has created an ambitious new institution, the Asian Infrastructure Investment Bank (AIIB), the equally ambitious “One Belt, One Road,” or simply the “Belt and Road” initiative, and at least 13 regional or bilateral funds that will radically increase Chinese development and infrastructure finance abroad (Gallagher, Kamal, and Wang 2016, 1). The Silk Road Fund is one of the 13 new funds. Taken together, the funds are projected to contribute up to US$116 billion in project financing. Three funds, totaling US$22 billion, are earmarked for Africa; four funds, totaling US$37 billion, for Latin America and the Caribbean; and two funds, totaling US$6 billion, for Eurasia. In addition, there is a Green Silk Road Fund and a China-ASEAN Fund (with the AsDB), totaling almost US$6 billion, a US$3.2 billion South-South Climate Fund, and a US$2 billion South-South Cooperation Fund (ibid., table 1). The funds express the foreign policy ambitions and economic objectives of China’s leadership but also reflect frustrations with the governance of the BWIs, the dominance of the United States and especially Japan at the AsDB, and the failure of the World Bank and other MRDBs to respond to unmet infrastructure needs (Dollar 2015).⁶⁵ It is clear that China is “poised to be the largest development [and infrastructure] lender in the world” (Gallagher, Kamal, and Wang 2016, 1). In what follows, I discuss the AIIB and the Belt and Road/Silk Road Fund.⁶⁶

In terms of capitalization and number of members, the AIIB represents the largest of China’s contributions to the changing institutional landscape. It is to focus on infrastructure (and cross-border infrastructure connectivity), sustainable economic development, and the promotion of regional cooperation by working with bilateral and multilateral institutions (AIIB 2016b). China’s leadership announced the AIIB in 2013, and it was founded in October 2014. The bank is headquartered in Beijing. During 2015, China invited 57 nations to join as founding members, and by December 2016 all had signed the organization’s Articles of Agreement. As of May 2017, there are 77 approved members.

Among G-7 countries, only the United States and Japan are absent from the AIIB. In public statements, U.S. government officials expressed disappointment that some of its key European and Asian-Pacific allies, including the United Kingdom, Germany, South Korea, Australia, New Zealand, and India, joined as founding members (Parker, Chassany, and Dyer 2015). The U.K. government was singled out for what an unnamed senior U.S. administration official termed its “constant accommodation of China” (quoted in Perlez 2015). Chinese officials, along with those of the IMF, World Bank, and AsDB, have made it clear that they see the AIIB as complementary to legacy institutions (McGrath 2015). The decision by the United States and Japan not to join suggests that they see it differently. U.S. officials defended their decision on grounds that critics found hypocritical—that the AIIB would neither embrace norms around transparency and anticorruption nor adopt the social and environmental safeguards of the BWIs and AsDB.

The AIIB was established with an authorized capital base of US$100 billion, with initial subscribed capital by China of US$29.8 billion. As with the NDB, 20% of the subscribed capital is paid in. The AIIB will initially have US$20 billion of usable capital, which could sustain US$100 billion of total lending (Griffith-Jones, Xiaoyun, and Spratt 2016, 18). The size is significant: the AIIB’s authorized capital amounts to almost 68% of the authorized capital base of the AsDB (US$147.5 billion) and is equal to 38% of the IBRD’s subscribed capital. In its first year of operation, 2016, the AIIB approved nine projects across seven countries, totaling US$1.7 billion. All but one of these loans is to be cofinanced with legacy institutions (e.g., World Bank, AsDB, EBRD).⁶⁷

By 2025, the AIIB is conservatively projected to have a loan portfolio capacity of US$70–90 billion (Humphrey 2015b, 15). Under less conservative scenarios, its loan portfolio could reach US$100–120 billion or more, making it the second-largest development bank in the world (Griffith-Jones, Xiaoyun, and Spratt 2016, 26). By comparison, US$120 billion was more than the individual lending by the AsDB, IADB, or AfDB in 2014, and almost as high as the total stock of lending by the IBRD in 2014 (which was US$140 billion) (ibid.). The AIIB will largely cofinance projects with other multilateral lenders in its first years of operation. With cofinancing, the scale of the infrastructure projects to which the AIIB contributes could reach US$240 billion by 2025 (ibid., 27). Many legacy institutions (namely, the AsDB, EBRD, EIB, and World Bank) have signed cooperative agreements with the AIIB. With 16 of the world’s largest economies as shareholders, the AIIB is likely to receive a AAA bond rating, lowering its cost of capital and making its loans attractive to middle-income countries (Humphrey 2015b). The AIIB’s openness to membership, reflected in part in its permitting founding members to participate in negotiations over its statutes, distinguishes it from the NDB and gives the AIIB operational advantages (ibid.).

The AIIB reflects a widespread tension in the emerging financial architecture between continuity and innovation (as argued by Chin 2016). In terms of continuity, the first AIIB president is from China (the largest shareholder), while future presidents must be from a regional member nation. Voting shares at the institution depend on several factors, including GDP. Nonregional members of the AIIB are limited to 25% of the institution’s aggregate voting share, ensuring the dominance of Asia’s and particularly China’s voice. China will hold 26.06% of the votes at the institution, providing it with de facto veto power on major policies and decisions that generally require a supermajority.⁶⁸ In terms of discontinuities, the AIIB has a dual nonresident board comprised of a large Board of Governors and a 12-person unpaid Board of Directors (nine from the region, three from outside the region). The directors will meet periodically during the year to supervise operations and management. The AIIB’s structure is expected to speed decisions and lower operating costs.⁶⁹ The AIIB also has no branch offices in borrowing countries. Most importantly, the institution will almost certainly be a key actor in the emerging pluripolar distribution of financial power in the world economy.

The AIIB widely circulated a document for comment that outlines its “Environmental and Social Framework for Lending.”⁷⁰ By February 2016, its Board of Directors had approved an “Environment and Social Policy.” Many observers nonetheless raised concerns about social and environmental standards and the institution’s support for coal-fired plants in places lacking access to power (He 2016, 13–15). In October 2016, the AIIB took on these (and other) matters in an “Energy Strategy” document that it presented for public comment.⁷¹ The document notes that “coal and oil-fired power plants would be exceptionally considered if cleaner technologies are not available for well-founded energy security or affordability reasons” (AIIB 2016a). While that statement generates concern about its commitment to sustainability, it does introduce a bar that coal- and oil-fired plants will have to clear and opens the institution to close surveillance by environmental organizations.⁷² In addition, it should be noted that many AIIB staff have extensive World Bank experience. This suggests that the new institution is informed by and has the opportunity and perhaps also the intent to draw on the World Bank’s (and the AsDB’s) safeguards in this domain (Chin 2016, 15). Not all observers are persuaded that the World Bank represents the best model of environmental stewardship, however.⁷³

In 2014, at about the same time as it unveiled the AIIB, China launched another initiative focused on infrastructure (and energy). It involved the creation of two modern-day “Silk Roads”—the overland “Silk Road Economic Belt” and the “21st Century Maritime Silk Road.”⁷⁴ The Belt and Road initiative has roots in ideas and practices that have existed at the national and subnational levels since the 1980s (Summers 2016). It will facilitate transportation and communications between China and some 65 countries across Central Asia, the Middle East, Russia, and Europe, and through investments in maritime infrastructure, it will speed access to East Africa, the Red Sea, and the Mediterranean. Anxious observers of China’s rise see the Belt and Road as its most ambitious effort to extend economic diplomacy and its grand strategy through its own Marshall Plan. It promises to enhance the country’s dominant role in Asia, buy influence, and secure privileged access to markets and resources (Clover and Hornby 2015).

The Silk Road Fund was created to support projects of the Belt and Road initiative, which will also be supported by the AIIB and the China Development Bank (CDB). The fund has an initial capitalization of US$40 billion. Support for the Silk Road Fund comes from four government agencies. In 2015, the PBOC transferred US$82 billion to three state-owned banks for Belt and Road projects (Economist 2016b). The State Administration of Foreign Exchange (SAFE, which is the PBOC unit that manages the renminbi) will hold a 65% stake, China Investment Corporation (a sovereign wealth fund) and the China Export-Import Bank will each have 15%, and the CDB will have the remaining 5% (Kozul-Wright and Poon 2015). The fund is open to investors from other countries as well (Gallagher, Kamal, and Wang 2016). Official figures indicate that 900 deals, worth US$890 billion, are planned or under way. By comparison, the postwar Marshall Plan was valued at US$130 billion in current dollars in 2015 (Economist 2016b).

Given the large environmental footprint of the energy-related loans made by China’s older policy banks (such as the CDB), it is by no means certain that the new Chinese-led initiatives will have a greener footprint in practice than those of their predecessors (see Gallagher, Kamal, and Wang 2016; Tabuchi 2017). The first four loans made by the AIIB were for projects that prima facie do not ensure sustainability—an electricity grid, slum development, and highway and road construction—while the second batch of loans are for hydropower and a gas turbine. Nevertheless, these (and other) Chinese-led financial initiatives seem particularly fruitful in their “potential to transform multilateral lending more broadly” (Kozul-Wright and Poon 2015). As Kozul-Wright and Poon (2015) note, China’s experience with experimental and incremental development strategies makes it particularly suited for this leading role in transforming the institutional landscape. China’s initiatives also increase the possibility for forum shopping, with attendant effects on voice in existing institutions. The Chinese initiatives have already placed and will likely continue to place pressure on the private sector, the G-20, the World Bank, and the MRDBs to increase infrastructure spending.

National Development Banks: The Brazil National Bank of Economic and Social Development and the China Development Bank

Brazil’s National Bank of Economic and Social Development (Portuguese acronym BNDES) and the China Development Bank (CDB) have been described as “mega banks” since they have assets of greater than US$100 billion, which distinguishes them from the over 50% of development banks with assets of less than US$10 billion (Chandrasekhar 2015b). The activities of BNDES and the CDB during the global crisis provide dramatic examples of the evolving nature of national development banking.

BNDES

BNDES was founded in 1952.⁷⁵ It was created as a public (federal) development bank to provide long-term and/or high-risk capital primarily to private Brazilian firms whose needs for capital were not being met by the country’s private banks (Hochstetler 2014a, 361). BNDES is part of an extensive public bank infrastructure in Brazil, though we focus here on this institution only.⁷⁶

In terms of its assets, BNDES eclipses all other national lending institutions in Latin America. As of June 30, 2016, its assets totaled US$291 billion (BNDES website). BNDES is significantly larger than the IADB (which had assets of US$116.5 billion in 2015 and subscribed capital of US$170.9 billion in 2016) and close to the size of the IBRD (which had assets of US$343 billion and subscribed capital of US$253 billion as of June 30, 2015). BNDES annual disbursements more than quadrupled during former president Lula da Silva’s tenure in office (2003–2010), growing to three times the size of the World Bank’s annual lending in 2010 (Hochstetler 2014a, 360). BNDES disbursed US$41.8 billion in 2015, which was a 15.6% reduction in disbursements compared with the previous year. BNDES disbursements for the first two quarters of 2016 totaled $11 billion, which suggests continuation in the downward trend in disbursements. But even with the reduction in disbursements, BNDES disbursements in 2015 still dwarfed those of the IADB by a factor of over four to one, and were close to those of the World Bank Group, which disbursed US$45 billion worldwide in fiscal year 2015 (and US$49 billion in 2016). Indeed, the only development bank based in an EMDE whose assets and disbursements exceed those of BNDES is the CDB.

The activity of BNDES has varied as national development strategies have evolved, but it has consistently remained a central actor in Brazilian finance (Hochstetler 2014a). It has been so central to economic policy that Armijo (2017) termed it a symbol of the country’s “new developmentalism.”⁷⁷ The expanded mission that BNDES took on under President Lula after 2005 necessitated a dramatic increase in its resources (Hochstetler 2014a). New activities included supporting the internationalization of Brazilian firms by enlarging the export-financing programs established in the previous decade, supporting outward FDI of Brazilian firms, and offering trade credit and working capital to Brazilian firms operating abroad when traditional sources of financing dried up (Armijo and Echeverri-Gent 2014; Tavares de Araujo 2013). The internationalization of Brazilian firms induced modest internationalization of BNDES (Hochstetler 2014a). In August 2009, it opened its first branch office in South America, in Montevideo, Uruguay, followed by a holding company in London, and in 2013 an office in Johannesburg. As Hochstetler notes, international lending by BNDES should not be seen as a matter of “South-South” solidarity: “Brazil’s President Luiz Inácio Lula da Silva [spoke of] … the link between BNDES’ international lending and Brazil’s foreign policy interests [as driven by its] … being a ‘big country’ promoting its strategic economic and political interests, along with those of its loan recipients” (ibid., 360–361).

BNDES rhetoric concerning internationalization overstates its practice. BNDES disbursements in South America are more qualitatively and quantitatively limited than both the bank’s champions and critics contend (Hochstetler 2014a; Hochstetler and Montero 2013). Structural and legal factors restrict its international activities in ways that may keep many of the bank’s resources inside the country and that tightly link finance to Brazilian firms and to employment generation at home (Hochstetler 2014a; Hochstetler and Montero 2013). Notwithstanding these constraints, 16% of the loans made by BNDES during fiscal year 2016 were denominated in foreign currency (on par with levels in the previous two years).

Under President Lula, BNDES began to provide countercyclical finance. During the global crisis, BNDES increased disbursements, coordinated actions with private banks to support distressed firms, and took other measures to channel liquidity to small and medium-sized banks that were under stress (Armijo 2017; Tavares de Araujo 2013; Torres Filho 2011). The government used the public bank infrastructure to orchestrate a rapid countercyclical response to the global crisis, but the majority of disbursals went through BNDES (Armijo and Echeverri-Gent 2014; World Bank 2013b, 106). Enabled by a generous capital injection from the government, in 2009 BNDES extended special credit facilities and loans at subsidized rates (World Bank 2013b, 106). BNDES accounted for a third of the growth in credit to the private sector during 2008 and 2009 (Hochstetler and Montero 2012). In 2010, annual lending by BNDES amounted to about 70% of long-term credit in the country (Chandrasekhar 2016). It played a critical role in providing financing when private domestic lenders in Brazil contracted their operations in 2008 and all but froze lending from September 2008 to January 2010 (Chandrasekhar 2016; Torres Filho 2011).

Countercyclical interventions by BNDES required an unprecedented increase in disbursements. BNDES disbursements averaged just over 2% of GDP from 2000 to 2007. They then ranged from 3% to 4.5% of GDP between 2008 and 2014 (Tavares de Araujo 2013, figure 5; Gottschalk 2016). As of late 2009, BNDES accounted for 19% of the stock of outstanding loans from the country’s financial system to households and firms (Armijo 2017). By the end of 2011, the bank’s loan portfolio was equal to 10% of GDP and its total assets represented about 15% of GDP (ibid.). The expansion of public credit was central to the country’s ability to emerge from a recession after only a six-month downturn, and to the stability of manufacturing in the face of a 16% fall in exports from September 2008 to March 2009 (Hochstetler and Montero 2012). Notably, in December 2009, 67% of bank loans in the country with a maturity of over five years were made by BNDES (Torres Filho 2011). A 2016 report states that at present over 90% of loans with a maturity of greater than three years were extended by the country’s three public banks, and BNDES is by far the largest provider of long-term finance in the country (Studart and Ramos 2016, 21). Under these circumstances, it is not surprising that BNDES is a key player in overall Brazilian infrastructure finance (ibid., 25).

Observers have long criticized BNDES on many grounds, including the damaging environmental footprint of the projects it supports and the concentration of its loans among a relatively small number of large firms and sectors, particularly electricity and gas (Armijo 2017; Hochstetler and Montero 2013).⁷⁸ However, research on BNDES also shows that it is almost the sole source of domestic funding of sustainable infrastructure (Studart and Ramos 2016) and that support for innovation went to smaller and newer firms (Hochstetler and Montero 2013). In addition, it developed new programs to assist micro, small, and medium-sized firms (Armijo 2017). In fact, 31% of total disbursements in 2014 went to small and medium-sized firms (Gottschalk 2016).

In recent years, BNDES has begun to cooperate with other multilateral and regional development banks. It has partnered with the World Bank; in 2011, it signed a Financial Cooperation Agreement with the presidents of development banks of China, India, Russia, and South Africa as part of its continuing engagement with the BRICS, and, in the fall of 2015, it signed a memorandum of understanding designed to strengthen cooperation with the NDB.

CDB

China has three major “policy banks”—the CDB, the Agricultural Development Bank of China, and the Export-Import Bank of China. These public banks were established in 1994 to provide financing to high-priority government projects. The CDB supports the government’s macroeconomic policies and development objectives, and our discussion focuses primarily on this institution.⁷⁹ It is a primary source of long-term finance for infrastructure and large-scale investment in basic and heavy industry (UNCTAD 2015b, 169). In fact, US$1.5 trillion of CDB loans went to infrastructure and basic industry from 2006 to 2014 (Wang 2016, 4). As of year-end 2014, about 53% of the CDB’s outstanding loans were allocated to infrastructure (ibid., 14).

The CDB is the world’s largest development bank in terms of assets, which remained high and grew steadily during the global crisis. The CDB’s assets were US$108 billion in 2001 and US$560 billion at the start of the crisis in 2008. By 2013, they had risen to US$1.3 trillion (Wang 2016). At year-end 2015, they stood at roughly US$1.7 trillion (a 22% increase over 2014), about five times the 2015 assets of the IBRD.⁸⁰

The CDB undertook strongly countercyclical initiatives during the crisis by lending actively in the domestic market and providing important support for the country’s export performance. Lending by the bank grew markedly: at year-end 2008, outstanding loans were valued at US$460 billion, and at year-end 2014, they stood at US$1.2 trillion. The loans extended in 2014 were valued at US$276 billion, which represented a 20% increase over 2013 lending (CDB website, annual report 2014, 12). As signs of an economic slowdown and financial fragility became apparent during the summer and fall of 2015, the CDB responded with new countercyclical support that supplemented other government measures.

During the crisis, China launched a variety of bilateral financial initiatives in Asia, Africa, Latin America, and the former Soviet bloc countries through its policy banks, especially the CDB, but also through the Export-Import Bank. Cross-border loans and lines of credit are driven by many of the same objectives as the country’s currency swaps, particularly access to key resources and markets and support for SOEs and other firms. At year-end 2015, the CDB’s foreign currency loans totaled US$276 billion, which represented 19% of its total loan portfolio and a ninefold increase in foreign-currency lending compared with 2007 (CDB website, annual report 2015, 60).⁸¹ Many of these loans support infrastructure development in EMDEs and China’s access to raw materials (UNCTAD 2015b, 169–170).

China committed around US$132 billion in financing to African and Latin American governments between 2003 and 2011, and more than 80% of this financing was committed since the start of the global crisis.⁸² About half of these loans (US$75 billion) were resource secured, involving exports of oil, cocoa, and diamonds. Interest rates on loans from China’s policy banks are not generally out of line with rates available on global markets, and interest rates on loans to Latin America are roughly comparable to those of the World Bank, IADB, and CAF (Gallagher 2016, 78). The lack of transparency around Chinese loans is cause for concern by policymakers in EMDEs. China’s loans do not carry the same policy conditionalities as loans from the BWIs. But these loans are not without “strings,” such as those that tie loans to the procurement of goods and construction services from Chinese firms. CDB conditionality illustrates that loans made by EMDE lenders are not always more advantageous to recipients than loans from multilateral sources. But the evidence on these conditions is not unequivocal. Contrary to popular perception, Bräutigam and Gallagher (2014) find no evidence that China’s loans involve specific requirements to employ workers from the country or that they lock in low commodity prices. Moreover, they find that borrowers have room to negotiate on employment, training, and local content when this is seen to benefit the project from the perspective of its lenders.

Today, the CDB and the Export-Import Bank of China provide more financing to Latin American governments than the World Bank and the IADB, and more to Asia than the World Bank and AsDB (Gallagher 2015b). The CDB and the Export-Import Bank of China provided US$21 billion in finance to Latin American governments and state-run firms in 2016 (Myers and Gallagher 2017). China’s cross-border loans dwarf those of BNDES. From 2008 to April 2015, the CDB extended loans to 40 countries, valued at US$233 billion (and representing 60% of the bank’s total credit portfolio) (Wang 2016, table 3). The international holdings of the CDB and the Export-Import Bank are estimated to be about US$684 billion, which is very close to the combined US$720 billion in international assets held by the World Bank, IADB, AfDB, and AsDB (Gallagher, Kamal, and Wang 2016, figure 1). However, it is critical to note that the vast majority of the US$117 billion in energy-related loans made by the CDB and the Export-Import Bank to EMDEs in 2007–2014 are in oil and gas exploration, and even more are in power plants, 66% of which are coal fired (ibid.).⁸³ Thus, international loans by the CDB and Export-Import Bank of China cannot be seen as supporting the environmental concerns embodied in some of the SDGs (Tabuchi 2017). Moreover, this lending surge is not without other costs. The risks associated with overleveraging in China and overlending to (now) slowing EMDEs are only now becoming apparent.

The significance of lending to EMDEs by Brazil’s and especially China’s development banks (and those of the other BRICS as well) extends beyond the economic value of the loans.⁸⁴ Loans by China and Brazil enhance their individual influence and also enhance the collective international stature of rising powers vis-à-vis the AEs. In addition, lending by China and other emerging powers is imparting complexity to the traditional Bretton Woods–era architecture of project finance, when the line between AE lending and EMDE borrowing was clearly drawn.

The Main Regional Development Banks

The MRDBs are a type of multilateral development bank that deserves some attention here, though they are BWIs and not creations of the EMDEs. They, too, evolved, though modestly, during the global crisis. Among the MRDBs, I focus here on the AsDB, IADB, and the African Development Bank (AfDB).⁸⁵ Nonborrowing AEs are shareholders in these three institutions, and are the dominant decisionmakers in the AsDB and IADB since voting shares are allocated according to capital contributions by member nations. The AsDB and the AfDB are headquartered in EMDEs, whereas the IADB is headquartered in Washington, D.C.

Generally speaking, the MRDBs are vehicles for poverty reduction and the provision of public goods. The global crisis demonstrated the crucial countercyclical function that the MRDBs serve when private financial markets freeze. Prior to the global crisis, the AsDB was already lending more than the World Bank inside the Asian and Pacific region, and the IADB and FLAR were already providing more crisis-related financing in South America than the IMF (Woods 2010). The global crisis accelerated this trend. The MRDBs enjoyed an increase in their resources (thanks to the G-20) and accordingly became more active and introduced new countercyclical programs. The AsDB, AfDB, and IADB responded to the crisis in their regions in some cases more quickly and with larger loans than the IMF and the World Bank, and they introduced new temporary rapid financing programs and countercyclical lending facilities to support developing and low-income countries (Chin 2012; Woods 2010). The MRDBs as a group increased their lending commitments to EMDEs by 71% in 2009 compared with 2008, and their disbursements grew by 45% in the same period.⁸⁶

With G-20 backing, the AsDB introduced a new countercyclical instrument, the Counter-cyclical Support Facility, to provide support of up to US$3 billion to economies in the region affected by the crisis. Between 2008 and 2009, AsDB’s lending commitments grew by 42% and its disbursements by 33%. Other MRDBs quickly followed the AsDB’s example, and they were granted a portion of the new funds committed to the IMF to establish new regional lending facilities to promote rapid countercyclical support within their regions (Chin 2012). The IADB established a US$6 billion rapid disbursal Emergency Liquidity Fund to support the countercyclical efforts of member governments. It also increased callable capital by US$4 billion, increased its commitments by 38% in 2009 (having already increased its disbursements significantly in 2008), and disbursed 60% more in 2009 than in 2008. The IADB approved loans of US$11.2 billion, US$15.5 billion, and US$12.4 billion in 2008, 2009, and 2010, respectively (Ocampo and Titelman 2012, 15). The AfDB established a US$1.5 billion Emergency Liquidity Facility during the crisis. Between 2008 and 2009, it increased its lending commitments by 137% and its disbursements by 125% (which is the largest increase in disbursements of any of the MRDBs).

The MRDB response to the crisis involved the launch of a large number of regional initiatives. In these efforts, individual MRDBs generally worked with other institutions, notably the World Bank (Ocampo et al. 2012, 65–69). The World Bank and the MRDBs have often competed with one another (ibid.). The massive needs generated by the crisis pushed the institutions to collaborate. There were nevertheless major failings in these responses. Both the World Bank and the MRDBs responded inadequately to the needs of low-income countries. In addition, despite rhetoric to the contrary, investment in infrastructure by the World Bank and the MRDBs fell as a share of total investment in recent decades (though more recently it has begun to rise modestly) (Humphrey 2015b, figure 2).

The World Bank and the MRDBs have long been criticized for the slow pace of their loan approval procedures, particularly in comparison with institutions such as the CAF, and even now in relation to the AIIB and NDB. From initial project identification to first disbursement, the average loan approval time for the IBRD and IDA is 28 months (and for sovereign loans it is 14 months). For the AfDB, the average is 13 months; for the AsDB, 11 months; and for the IADB, 5.8 months (Humphrey 2015a, 13, figure 14).⁸⁷

Despite expansion in the scale and scope of activity by the MRDBs during the global crisis, gross disbursements by the World Bank Group still outstrip those of the MRDBs. In 2015, the IADB disbursed loans of US$10.2 billion, the AsDB US$11.7 billion, and the AfDB US$4.2 billion. Those figures amount to 20.8%, 26.2%, and 9.4%, respectively, of the World Bank Group’s disbursements.⁸⁸ In contrast, the IBRD approved loans of US$23.5 billion and US$18.6 billion in 2015 and 2014, respectively. Commitments by the IBRD exceeded its historical average of US$13.5 billion a year in loan approvals from 2005 to 2008, before the global crisis. In 2015, the IBRD approved what was a record level of financing for any year except at the height of the global crisis. Notwithstanding the expansion in activity by the MRDBs and especially the World Bank, these institutions appear stalled in comparison with China-backed initiatives, those of the BRICS, the dynamism of the CAF, and even that of a less expansive BNDES. This is particularly the case when one considers the inability of the Western-backed multilateral development banks to increase their paid-in capital despite much talk of the infrastructure finance gap and the challenges of the SDGs (Gallagher, Kamal, and Wang 2016).⁸⁹ The pivot of project finance away from the West is clear when one takes account of the full range of China’s policy banks, the new family of multilateral development banks and funds financed and/or cofinanced by China, and its support for the NDB (ibid., 3).

A Caveat Regarding Africa and Two Initiatives Led by Venezuela

Neither the Asian nor the global crisis has had a major catalytic effect on reserve pooling and project finance institutions on the African continent. The only meaningful outcomes in this regard relate to South Africa’s role in the CRA and the NDB, its membership in the AIIB, and the activism of the AfDB.⁹⁰ This stands in contrast to the effects of these crises elsewhere in the global south and east, as we have seen. This type of unevenness is to be expected—indeed, it is part and parcel of the overall inconsistency at the heart of the emerging financial architecture. Where we have seen more meaningful steps toward financial architectural innovations in Africa is in the realm of regional macroeconomic coordination and monetary integration, where several subregional initiatives are in the planning stages (see Fritz and Mühlich 2014, 29–40).

Two initiatives led largely by Venezuela are worth noting briefly because they contain elements of development banking and also because they illustrate the fragility of institutional experimentation. The Bolivarian Alliance for the Peoples of Our Americas (Spanish acronym ALBA) is a multifaceted project built around a vision of Latin American and Caribbean solidarity described as “post-hegemonic regionalism” and “regionalism without regions.” ALBA encompasses a suite of financial, trade, aid, and social integration programs, one of which is a development bank, the Bank of ALBA.⁹¹ The bank was established in January 2008, is headquartered in Caracas, Venezuela, and has a branch in Cuba. It is a development bank with a small portfolio (and subscribed capital of US$1 billion) and a mission of supporting ALBA’s integration aims and its social, economic, and political programs.⁹² At the 11th ALBA Summit, in February 2012, members committed to allocating 1% of their reserves to the Bank of ALBA. Loans made by the Bank of ALBA do not involve conditionality, and discussions are reportedly by consensus (Janike 2008).⁹³

As of this writing, the prospects of the Bank of ALBA (and the other initiatives under ALBA’s umbrella) remain uncertain, though dim, owing to the severe political and economic challenges confronting Venezuela.⁹⁴ But the experience with ALBA’s many initiatives and the relationships involved in building them may lay the groundwork for future innovations that may arise as political and economic circumstances evolve and acute needs for new initiatives emerge. Failed initiatives, too, provide critical opportunities for institutional learning.

The other initiative led largely by Venezuela that bears mention is an ideologically related project to create a development bank, the Bank of the South (Spanish acronym BDS). The BDS was founded in 2007 and became a legal entity in 2009 (with US$7 billion in subscribed capital, though recent reports suggest that the figure has grown to US$10 billion).⁹⁵ The BDS project moved forward in late November 2016 after a long period during which it failed to advance beyond its legal existence. At a meeting at the headquarters of the Union of South American Nations in Quito, an executive board was installed (with representatives from Venezuela, Uruguay, Ecuador, and Bolivia). The board announced that the institution would begin its preoperative phase and that Uruguayan economist Pedro Buonomo would be the bank’s first president (El Telégrafo 2016).

The initial BDS vision entailed the principles of equal voice among members and the rejection of conditionality. The BDS had stalled long before the loss of oil-fueled revenue and confidence and before many of the region’s governments moved away from populism. The Venezuelan government turned its attention away from the BDS (and toward ALBA) once other BDS members, such as Brazil, succeeded in shaping the institution in the image of a traditional development bank rather than as a vehicle of popular socialism.⁹⁶ The announcement in 2016 that the BDS would nonetheless begin functioning was unexpected, not least because of political shifts in Argentina and Brazil, original signatory countries.⁹⁷ It has initial planned paid-in capital of US$90 million, and Bolivia, Ecuador, and Venezuela are reported to have paid in already.⁹⁸ Fundamental questions remain about where the institution’s capital will come from. Other critical challenges include the weakness of the Venezuelan economy and leadership changes in Argentina and Brazil that could prove to be hostile to the institution.

Governance, Surveillance, Conditionalities, and Institutional Linkages

Many of the reserve pooling institutions and development banks examined here are characterized by governance structures that differentiate them from the BWIs, in which AEs (especially the United States) have disproportionate weight. Many of the institutions are organized to promote greater inclusiveness, though there is quite considerable divergence in the degree to which this is achieved by design or in practice. Indeed, some of the institutions considered here hew rather closely to the BWIs in terms of governance (when it comes to the influence of countries that contribute a large portion of the institution’s capital, the role of a resident board, and other matters), whereas others have made a rather sharp break with these norms. The fact that the institutions surveyed have diverse and complicated decision-making structures reflects the necessary and real tensions between the demands of the larger countries that provide the bulk of financial support, recognition of the legitimacy of concerns about inclusiveness for smaller, poorer countries, and the complicated power politics that necessarily infuses regional and transregional initiatives.

Like governance, the matter of “getting conditionality right” continues to be a key challenge, which institutions are managing in diverse ways. Some institutions, such as the CMIM, CRA, ArMF, and FLAR, plan to or already do conduct surveillance (including country missions) and utilize conditionality or require some type of adjustment program, at least under certain circumstances. Others, such as the EFSD, require conditionality under all circumstances. Some, such as the FLAR and CAF, employ an approach to monitoring that works with borrowing governments in ways that are decidedly distinct from the top-down approach of the IMF and World Bank. Here, surveillance (and in the case of FLAR) adjustment programs are minimalist, highly country-specific, peer based, and exclude the BWIs. Some institutions are actively wrestling with these issues and involve the IMF explicitly under certain circumstances (e.g., the CMIM and CRA). In contrast, the EFSD involves the IMF implicitly through consultations and, like the CRA, abstains from lending to countries in arrears to it. The early design of the BDS and the Bank of ALBA renounced conditionality altogether. In the newest institutions, such as the AIIB and NDB, the matter of project selection and assessment is still evolving, as is the issue of how to handle nonperforming loans.

For the most part, the institutions considered here are more agile than the BWIs (and the MRDBs) inasmuch as they respond quickly to economic challenges in their field of operations. In several instances, this agility—coupled, critically, with a sense of country ownership and the appropriateness of surveillance procedures—has induced countries receiving support to treat the lending institutions as if they held preferred creditor status.

An obstacle facing reserve pooling arrangements in particular concerns the challenges posed by precautionary forms of support. Some institutions, such as the CMIM, plan to utilize prequalification criteria before support is disbursed. Others, such as the FLAR, have thus far successfully used their own forms of monitoring and dialogue to determine eligibility for precautionary support without resorting to prequalification criteria. For the CRA, the matter of qualification for both liquidity and precautionary support remains under consideration among member central banks, though at this point the CRA mirrors the link to the IMF that is a feature of the CMIM.⁹⁹ Precautionary support is always a complex matter, as we have seen in the case of the IMF’s FCL. It often involves some sort of prequalification criteria, which may mean that the candidates that meet the criteria are those that are least likely to need support, and in the case of regional and transregional bodies, it may undermine the solidarity that is an intrinsic part of these arrangements.

Conclusion

As the foregoing makes clear, the institutions surveyed diverge from one another in many ways, such as their internal governance, aspirations, size and capacity, member buy-in, and the degree to which they are explicitly or implicitly linked to legacy institutions. They do not meld into any sort of new, coherent system of financial governance architecture or developmental finance. Not all are equally likely to survive, let alone thrive, in the years ahead. Indeed, the BDS was shipwrecked for many years on the shoals of the economic challenges and conflicting visions among member nations. Despite the recent resurrection of the project, its prospects are uncertain. The fate of the Bank of ALBA is similarly in doubt. Neither individually nor collectively do any of the reserve pooling institutions considered here promise or aim to challenge the IMF as the central institution of crisis response. In the realm of development banks, the institutions considered here should also not be considered against the standard of displacement of the World Bank or the MRDBs. That said, many have the firepower to play central roles in the provision of project and infrastructure finance across EMDEs. They do not amount to a new pole of financial power that will necessarily demote AE hegemony in financial affairs. Instead, the initiatives are fragmentary and heterogeneous, some are internally fraught with rivalry and suspicion, and many are no doubt marked by the same kinds of ambiguity as the IMF, where gritty, muddled day-to-day practice conflicts with coherent, pristine mission statements. Finally, the institutions may work at cross-purposes, especially during crisis moments, undermining each other’s efforts and/or imposing cross-border spillovers that disrupt each other’s economies.

I do not take these features as fatal flaws. Instead, guided by Hirschman, I recognize the present period of institutional experimentation, expansion, and hybridization as a moment of pragmatic innovation that just might yield institutions and practices that do better than their predecessors in promoting financial stability and resilience and, as a consequence of that, provide at least the possibility for development that is more stable, inclusive, sustainable, and protective of autonomy. With Hirschman, I place emphasis on the potential inherent in unscripted adjustments that are freed from the constraints imposed by hegemonic narratives that purport to demonstrate the single path to economic security and development.

At a minimum, the flourishing of heterogeneous EMDE institutions of financial governance and developmental finance generates new opportunities for exit from unresponsive institutions and for at least a degree of forum shopping among alternatives. As a consequence, it may increase EMDE resilience, bargaining power, and voice vis-à-vis the BWIs (Helleiner 2010). To the extent that opportunities for forum shopping are realized, the BWIs may face pressure to respond to long-held concerns by EMDEs. In any event, the leverage of larger EMDEs in global and regional financial governance is certainly increasing as several of the institutions surveyed here have come to play a more prominent role during the global crisis. Redundancy and the networks of cooperation that are already emerging among institutions may increase overall resilience and enhance antifragility in the global financial system. In this connection, I note that UNCTAD calls for “more diversified financial systems,” by which is meant different institutions of different sizes and mandates (UNCTAD 2013, chap. 3), that Ocampo (2006, chap. 1) has long called for a denser financial architecture, and that Culpeper (1997) argues for the benefits of competitive pluralism among multilateral development banks on the grounds that overlap and rivalry encourage innovation and productivity.¹⁰⁰ Multiple layers and increased density have the potential to yield productive redundancy—which can reduce instability, contain and ameliorate crises, and increase opportunities to finance development. The emerging productive redundancy promises to disturb the apparent efficiency of the streamlined, top-down, centralized financial governance architecture that characterized the neoliberal era, which promised efficiency but in fact generated extraordinary risk and crisis contagion while starving most EMDEs of adequate developmental finance.

The new pluripolar landscape is already affecting formal and informal governance and decision making at the IMF, and consultations within the G-20, as we have seen. As we will see in chapter 7, the terrain on which capital controls are theorized and pursued has likewise shifted. Only the most unimaginative accounts can fail to recognize that the BWIs and leading AEs today must manage a jumble of newly empowered competing interests and that doing so is already altering and will continue to alter their rhetoric, research, and practice. When compared with the neoliberal era, the change is nothing short of extraordinary, not because neoliberal impulses and strategies have been subordinated to a comprehensive alternative but because the stifling uniformity of the neoliberal era is undermined by the proliferation of EMDE institutions that are, to varying degrees, relatively autonomous from the primary institutions of neoliberal authority.

There are no guarantees, of course, that the new opportunities afforded by institutional innovation, exit, and voice will necessarily generate a more just economic landscape. The increased aperture in financial governance may not survive as emerging powers attempt to assert hegemony over other EMDEs. Would a financial governance architecture dominated by China, say, necessarily provide greater breathing room in the long run for smaller, lower-income countries? But for now, at least, we should be attentive to the potential for progressive reform that has emerged as a consequence of the increased policy space that the evolving, incoherent system provides. Certainly in contrast to the neoliberal era, when financial governance structures, practices, and ideology represented a suffocating obstacle to innovation and experimentation, today’s leaders look out on a more heterogeneous landscape that may very well prove to be much more congenial to unscripted, locally appropriate initiatives.

The new initiatives provide Hirschmanian opportunities—for learning by doing and learning from others, parallel experimentation, and providential problem solving that only comes about as a consequence of the Hiding Hand. Progress often happens when obstacles are initially underestimated so that new initiatives appear to be viable and when practitioners are then forced to search for solutions that were previously unimaginable. The next crisis may very well propel new initiatives and a deepening of embryonic institutions and partnerships that speak to challenges that now appear irresolvable. Moreover, the proliferation of institutions, even if they are not as credible, efficient, and experienced as the BWIs or the MRDBs, is vital to the creation of new networks within countries and across national borders that can enhance indigenous and widely dispersed capacity in areas that are fundamental to economic development. We should remember in this context that even experimental failures can and often do leave in their wake vital linkages and knowledge that may be available for and enable subsequent endeavors. In this vision, few successes and failures are final—they are more typically steps along branching historical paths as actors seek to confront the challenges they face. They are best able to do that, Hirschman also reminds us, when they are free to do so unencumbered by theoretical visions and institutional monopolies that attempt to prenarrate the future. Ad hoc, pragmatic adjustments rather than a tightly constrained choreography are what Hirschman put his faith in, messy though they may be. And that is just what is emerging across the new financial governance architecture.

Notes

1. Chin (2015, 307) advances a subtle interpretation of the kinds of changes examined in this chapter. He argues that, on the surface, the “rising powers” appear to have a status quo orientation toward and effect on the global financial architecture. But he urges a deeper, more patient treatment, highlighting the potential significance of their increased commitment to bilateralism and the renewal of interest in regional and subregional arrangements that respond to transnational economic (and security) problems. See also Armijo and Roberts (2014) and Fritz and Mühlich (2014).

2. See also Helleiner (2010), Ocampo (2006, chap. 1; 2010a; 2010b; 2011), and essays in Volz and Caliari (2010).

3. I thank Luis Rosero for this point.

4. Eichengreen (2015) proposes principles regarding the relationship between the IMF and regional financial arrangements.

5. On Asia, see Chin (2012); on Latin America, see Riggirozzi and Tussie (2012) and Armijo (2012); and for surveys, see Fritz and Mühlich (2014; 2015), Grabel (2013a), McKay, Volz, and Wölfinger (2011), Miyoshi (2013), Mühlich and Fritz (2016), Rana (2013), Rhee, Sumulong, and Vallée (2013), and Volz and Caliari (2010). For an examination that pre-dates the crisis see the essays in Ocampo (2006), especially Culpeper (2006). Peter Katzenstein’s (2005) A World of Regions is a seminal contribution to the political science literature on regionalism. He argues that diverse regionalism(s) reshaped but did not displace a globalized world, an observation that resonates with the discussion of regional and other institutional innovations in this chapter.

6. See treatments along these lines by Chin (2015), Golub (2013), Grabel (2013a; 2015a), Helleiner (2016b), Huotari and Hanemann (2014), Mittelman (2013), Rana (2013), Sohn (2012), Tussie (2010), and Woods (2010).

7. See Culpeper (2006) and Ocampo (2006, chap. 1) on regional payment systems and regional macroeconomic and monetary integration in EMDEs prior to the global crisis.

8. Some of the reserve pooling entities described in this chapter are formally considered “arrangements,” while others are considered “institutions.” In the text, I will use each term as appropriate, though nothing of significance is implied by the terminological choice.

9. The discussion of the CMI/CMIM draws from AsDB (2010), Capannelli (2011), Ciorciari (2011), Cohen (2012), Eichengreen (2010; 2012), Fritz and Mühlich (2014), Grabel (2013a), Grimes (2009a; 2011; 2015), Henning (2009a), Kim and Yang (2014), Lombardi (2010), Miyoshi (2013), Rhee, Sumulong, and Vallée (2013), and Sussangkarn (2011), and personal communications with officials.

10. See Eichengreen (2012) and Haggard (2013) on the challenges of regional surveillance.

11. Given China’s long-standing discomfort with the IMF, and its own commitment to policy autonomy, its insistence on an IMF link reveals the wariness with which many policymakers approach regional surveillance mechanisms. Of course, China did not expect to need assistance, so the IMF link was not expected to compromise its autonomy.

12. The Asian Bond Market Initiative was also expanded. CMIM was not given the capacity to issue its own bonds.

13. China and Japan each contribute 32%; Korea supplies 16%.

14. In order to address the latter issue, some have called for expanding the swap pool through broader membership, including India, Australia, and New Zealand (Lombardi 2010).

15. Personal communications with officials.

16. Kim and Yang (2014) are optimistic on this score; Eichengreen (2012) is pessimistic, especially in light of the ASEAN tradition of noninterference.

17. Katada and Sohn (2014) argue that the CMIM allows Japan and China to increase their bargaining power in global and regional financial governance without directly confronting Western powers.

18. Even some skeptics note that another crisis may propel further development; see Cohen (2012) and Grimes (2015).

19. Personal communication with official.

20. The AsDB and AMRO signed a memorandum of understanding to strengthen cooperation in May 2017.

21. Discussion of the FLAR draws on the institution’s website and annual reports, and Eichengreen (2012), Fritz and Mühlich (2014), Grabel (2013a), McKay, Volz, and Wölfinger (2011), Miyoshi (2013), Mühlich and Fritz (2016), Ocampo and Titelman (2009–2010; 2012), Perry (2015), Rhee, Sumulong, and Vallée (2013), Rosero (2011; 2014), Titelman et al. (2014), Urrutia (2015), and personal communications with officials.

22. In 1984, the FLAR created a subregional currency, the Andean peso, which is an accounting unit created to facilitate payment among central banks and other authorized holders (Ocampo and Titelman 2012, 18).

23. Members are Bolivia, Colombia, Costa Rica, Ecuador, Paraguay, Peru, Uruguay, and Venezuela.

24. The FLAR has 52 permanent staff members (Miyoshi 2013, 33).

25. Lending to Peru was delayed in the late 1980s, however, while the country negotiated its arrears with other international institutions (Ocampo and Titelman 2012, 26).

26. The only exception involved two loans to Ecuador. A loan was approved in October 2005 but not disbursed until 2006, on the condition that newly elected president Palacio meet FLAR requirements that the government maintain a primary budget surplus of at least 2% of GDP during 2006–2008 and that a new central bank board be installed (Rosero 2014, 75). In 2009, the FLAR approved a US$480 million balance of payments credit to the country. But when the government rejected conditions that the FLAR usually attaches to loans, credit was extended to the central bank on the stipulation that it deposit US$250 million of reserves with the FLAR (Haggard 2013, 271).

27. This is also the case with Treasury credit lines.

28. The FLAR contributed just 7% of the resources to a Colombian rescue package in 1999. Its loan was part of a larger package from the IMF, IADB, World Bank, and the Andean Development Corporation (Rosero 2014, 70). FLAR members have on ten occasions drawn on the institution’s resources while also receiving support from the IMF, though one was not conditioned on the other (Mühlich and Fritz 2016, 21). This happened most frequently in the 1980s and 1990s and involved Bolivia, Colombia, Ecuador, and Peru (ibid.).

29. During the global crisis, the IMF disbursed only one loan to a FLAR member, and that was to Ecuador. It disbursed an emergency loan to the country under the Rapid Financing Instrument following the 2016 earthquake. The IMF extended an SBA to FLAR member Costa Rica, though it was treated as precautionary and was not disbursed. In general, the IMF made relatively few traditional loans to Latin America during the crisis, though it did make a number of loans to Caribbean countries. In addition, the IMF extended FCLs to Colombia and Mexico. The former is a FLAR member, while the latter is not.

30. Under the latter, each country would hold equal basic votes, and additional votes would be allocated based on contributions to a common fund and in proportion to claims on particular credit lines (Ocampo 2015a).

31. In the view of Titelman et al. (2014), an expanded FLAR should be sized to cope with “most likely” scenarios. See also Ocampo and Titelman (2012, 26) and Kawai and Lombardi (2012).

32. See Truman (2010) and especially Eichengreen (2015) on the problems of subordinating the IMF to regional financial arrangements.

33. Discussion of the ArMF draws on the institution’s website and annual reports, and Ciorciari (2011), Corm (2006), Fritz and Mühlich (2014), Grabel (2013a), McKay, Volz, and Wölfinger (2011, 20–22), Miyoshi (2013), Mühlich and Fritz (2016), Rhee, Sumulong, and Vallée (2013), Titelman et al. (2014), and UNCTAD (2007,122).

34. Members include Jordan, United Arab Emirates, Bahrain, Tunisia, Algeria, Djibouti, Saudi Arabia, Sudan, Syria, Somalia, Iraq, Oman, Qatar, Kuwait, Lebanon, Libya, Egypt, Morocco, Mauritania, Yemen, Comoros, and the Palestinian Authority.

35. The technical staff comprises about 50 employees (McKay, Volz, and Wölfinger 2011, 21).

36. Data in this paragraph are from ArMF annual reports, except where noted.

37. The discussion of the EFSD draws on the institution’s website and annual reports, and Fritz and Mühlich (2014), Miyoshi (2013), Mühlich and Fritz (2016), Rhee, Sumulong, and Vallée (2013), and personal communications with officials.

38. The countries include Armenia, Belarus, Kazakhstan, the Kyrgyz Republic, the Russian Federation, and Tajikistan.

39. The Russian government has committed not to borrow from the EFSD. The second-largest member in the EFSD, Kazakhstan, is not likely to draw on financial credits since (like Russia) it has oil reserves.

40. The Eurasian Development Bank was founded in 2006 by Russia, and the country contributes about two-thirds of its resources.

41. A member may elect to reallocate a portion of its access to another member state. Russia’s decision to do so in 2011 made the large financial credit to Belarus possible (Miyoshi 2013, 43).

42. The discussion in this and the following paragraph draws on Armijo and Echeverri-Gent (2014), Armijo, Katada, and Roberts (2015), Chin (2014a), Ghosh (2012), HKEx (2012), Mielniczuk (2013), Stuenkel (2013, 619–620, 626), Ustinova (2012), and Wade (2011, 365).

43. Details on the CRA drawn from BRICS (2014c). See Montes (2014) for examination of the treaty that establishes the CRA, including voting procedures and prioritization of creditors based in BRICS countries.

44. See Armijo and Roberts (2014), Chin (2015), and Stuenkel (2016a) on evolving thinking on the BRICS and other rising powers in relation to the global financial architecture.

45. Examples of arguments along these lines include Bond (2016), Chandrasekhar (2014), Eichengreen (2014), Patnaik (2015), Roy (2014), and Steil (2014).

46. See, for example, Ban and Blyth (2013), Fourcade (2013), Huotari and Hanemann (2014), and Palacio (2015).

47. On the new imperialism, see Bond (2016, 615–617; 2015) and Ferrando (2014); on the BRICS’ significance, see Palacio (2015).

48. Examples of arguments along these lines include Armijo (2017), Bello (2014), Desai (2013), Grabel (2013a; 2015a), Griffith-Jones, Fritz, and Cintra (2014), Mielniczuk (2013), Montes (2014), Stuenkel (2016b), Ugarteche (2014), and Weisbrot (2014).

49. Henning (2016, 126, 134) examines the question of whether the CRA (and other regional or transregional reserve pooling arrangements) will or even should develop surveillance capacities and concludes that the IMF is the better institution to handle surveillance.

50. Economic historian Alexander Gerschenkron (1962) examined the important role of specialized financial institutions in overcoming the shortage of long-term capital in late industrializing nations in the late nineteenth century. For analyses and surveys of development banks, see Chandrasekhar (2015b; 2016), de Luna-Martínez and Vicente (2012), and Humphrey (2015b). See Amsden (1992; 2001) and Wade (2003[1990]) on development banks in a range of Asian countries; Mazzucato and Penna (2016) on the “market-correcting” and “market-shaping” roles of development banks; Humphrey (2015b) for comparisons between newly created banks and older ones; and Humphrey (2015a) on a range of multilateral development banks.

51. Trends in infrastructure finance at the IADB and AsDB are similar to those at the IBRD/IDA (Humphrey 2015a).

52. A study of the infrastructure gap by the AsDB (2017) estimates that “developing Asia” will need to invest US$26 trillion between 2016 and 2030—or more than double the US$750 billion that the AsDB estimated in 2009 (AsDB 2009). To put the figure in context, loan approvals by the AsDB in 2015 were US$15 billion.

53. The vast infrastructure investment gap and the need for sustainable infrastructure is reflected in the SDGs. Infrastructure figures explicitly in SDG number 19, which focuses on infrastructure. But it also figures in several others, such as goals 6, 7, 11, 13, and 17, which focus (respectively) on water, energy, sustainable cities, climate change, and partnerships for development. See Alexander (2014) and Bhattacharya, Oppenheim, and Stern (2015) for discussions of how infrastructure investment can facilitate inclusive, green development.

54. The acronym CAF is still used for legal reasons, so it will be used in what follows. Discussion of the CAF draws on the institution’s website and annual reports, and Griffith-Jones, Griffith-Jones, and Hertova (2008), Humphrey (2014), Ocampo and Titelman (2012), and personal communications with officials.

55. Members include Argentina, Barbados, Bolivia, Brazil, Chile, Colombia, Costa Rica, the Dominican Republic, Ecuador, Jamaica, Mexico, Panama, Paraguay, Peru, Portugal, Spain, Trinidad and Tobago, Uruguay, and Venezuela. Fourteen private banks have ownership positions in the CAF.

56. Data in the balance of this paragraph draw from Ocampo and Titelman (2009–2010, 252; 2012).

57. Data in this paragraph and the next two are from the annual reports of the CAF and the IADB, except where noted.

58. The discussion in this paragraph draws from Griffith-Jones, Griffith-Jones, and Hertova (2008), except where noted.

59. I thank Luis Rosero for information on bond issuance by the CAF.

60. Discussion of the NDB draws from BRICS (2014a, 2014b), Chin (2014a), Griffith-Jones (2014), and especially Humphrey (2015b). Data on the World Bank and MRDBs are drawn from the institutions’ websites and annual reports.

61. In September 2016, the institution’s representatives announced that they were pursuing a credit rating (NDB 2016b).

62. See Alexander (2014) and Chandrasekhar (2015a) for critical examinations of the infrastructure boom.

63. The World Bank established a Global Infrastructure Facility in October 2014 as an entity outside the World Bank. The facility was launched in April 2015 with an initial capitalization of US$100 million. To date, outcomes have lagged aspirations. Various problems have limited the scope of this facility to pilot projects and preparatory work (Arezki et al. 2016).

64. In addition to the arguments advanced against the CRA, critics argue that the NDB represents the “old development as infrastructure view” (Rodrik 2013). Others argue that the NDB is not inherently progressive insofar as it does not mark a break with traditional powers and that it is neither green, democratic, nor inclusive of EMDEs beyond the BRICS themselves. See Bond (2015; 2016), Chandrasekhar (2014), Ferrando (2014), Patnaik (2015), and Steil (2014). Possibilists acknowledge these risks but nevertheless recognize the NDB as an evolving institutional experiment with potential, even if with no guarantees. For this view, see Armijo (2017), Bello (2014), Chin (2014a), Desai (2013), Eichengreen (2014), Grabel (2013a; 2015a), Griffith-Jones (2014), and Stuenkel (2016a).

65. Japan and the United States are the AsDB’s largest shareholders, followed by China. The United States and Japan each have voting rights that are nearly equal to one another and well over twice those of China. The president of the AsDB is always from Japan.

66. The discussion of these and other financial initiatives led by China draws on the AIIB website, Chin (2016), Dollar (2015), Elgin-Cossart and Hart (2015), Gallagher, Kamal, and Wang (2016), Griffith-Jones, Xiaoyun, and Spratt (2016), He (2016), and Humphrey (2015b).

67. In April 2017 the AIIB and the World Bank signed a memorandum of understanding that signals increasing cooperation and promises increased lending capacity for both institutions.

68. In the lead-up to the launch of the bank, China was apparently willing to reduce its voting rights and forgo veto power if the United States and Japan joined. It appears that the matter of China’s voting rights remains negotiable should new members join (Chin 2016, 13; Humphrey 2015b, 26).

69. Chin (2016, 17) suggests that this board structure may have a demonstration effect for the World Bank and that it is notable that the World Bank recently began a review of its governance.

70. This contrasts with the NDB, where discussion of a similar document was confined to founding member countries.

71. A final round of public comment on a revised energy strategy document concluded in March 2017.

72. It is unclear if this commitment is more or less solid than that of the World Bank. Following U.S. president Obama’s lead in 2013, the World Bank announced in the same year that it would lend for coal-fired electricity generation projects only in “rare circumstances” (Chin 2016, 19).

73. For instance, a 2016 report by an advocacy organization finds that the World Bank’s International Finance Corporation has supported institutions that have funded at least 41 new coal projects since the 2013 decision on coal-fired plants (IDI 2016).

74. The Belt and Road initiative was announced in late 2013 and is articulated in a detailed document released in early 2015 (Ministries of Foreign Affairs and Commerce 2015).

75. The “S” for social was added in 1982.

76. The discussion of BNDES draws on the institution’s website and annual reports, and Armijo (2017), Armijo and Echeverri-Gent (2014), Hochstetler (2014a), Hochstetler and Montero (2012; 2013), Studart and Ramos (2016), Tavares de Araujo (2013), and Torres Filho (2011).

77. As of this writing, it is uncertain whether BNDES will continue to play as central a role, given Brazil’s economic and political turmoil, scandals that have damaged the institution, and the current government’s decided tilt toward the market.

78. Another controversy concerns the dependence of BNDES on transfers from the national treasury, especially after 2009 (Studart and Ramos 2016, 21–22).

79. The discussion of the CDB draws on the institution’s website and annual reports, Bräutigam and Gallagher (2014), Gallagher, Kamal, and Wang (2016), Gallagher, Koleski, and Irwin (2012), and Wang (2016).

80. Data in this paragraph and the next are drawn from the annual reports of the CDB and World Bank. The combined assets of the CDB and the country’s Export-Import Bank are over US$2 trillion, whereas the combined assets of the World Bank Group and the MRDBs are just over US$700 billion (Gallagher, Kamal, and Wang 2016).

81. The CDB has also been promoting the cross-border use of the country’s currency for international trade and investment through a variety of means, such as selling renminbi-denominated bonds outside the Chinese mainland beginning in 2007. Armijo, Katada, and Roberts (2014, 16) note that “by 2014, more than 22% of China’s trade was being settled in renminbi, up from almost nothing five years previously.” China’s central bank also used currency swaps during the global crisis as an alternative mechanism of financing and trade settlement. Ultimately it established a network of 23 bilateral local currency swap agreements at a total value of RMB2.57 trillion with Asian and non-Asian central banks. The network facilitated trade settlements by providing access to the renminbi, boosted the role of the renminbi by increasing the share of China’s trade invoiced and settled in renminbi, and promoted market confidence during a period of fragility (Chin 2014b; Henning 2016, 129; Huotari and Hanemann 2014; Jiang 2014).

82. The discussion in this paragraph summarizes research by Bräutigam and Gallagher (2014). See Bräutigam (2009) on the country’s loans (and other financial flows) to Africa prior to the crisis. See Gallagher (2016) on China’s role as a dominant lender to Latin America.

83. The World Bank and the MRDBs provided about US$118 billion in energy-related financing in the same period (Gallagher, Kamal, and Wang 2016).

84. The discussion in this paragraph draws on Hochstetler (2014b).

85. I do not discuss two European banks, the European Bank for Reconstruction and Development (EBRD) and the European Investment Bank (EIB). The EBRD operates more like a commercial bank than a development bank (insofar as around 80% of its loans are to the private sector). The EIB operates mainly in industrialized European countries, and only a small share of its operations are outside the EU (Humphrey 2015a, table 1). See Babb (2009, chap. 1) on the EBRD, and Griffith-Jones, Griffith-Jones, and Hertova (2008) and Ocampo et al. (2012) on the EIB. See Babb (2009) and Strand (2014) for analyses of the MRDBs.

86. Data and discussion in this paragraph and the next two draw on Ocampo et al. (2012, especially 65–69), except where noted.

87. Data for the World Bank are from 2016, data for the IADB are from 2013, and data for the AsDB and AfDB are from 2012 (World Bank Group 2016; Humphrey 2015a, 13, figure 14).

88. Gross disbursements by the EIB of US$69.7 billion during 2015 exceeded those of the World Bank Group. World Bank Group data are reported on a fiscal year basis; AfDB, AsDB, EIB, and IADB data are as of December 31, 2015. Data are from the institutions’ annual reports.

89. In contrast, China has been increasing the paid-in capital for both the CDB and its Export-Import Bank (Gallagher, Kamal, and Wang 2016).

90. See Bradlow and Humphrey (2016) for a survey of development banks in Africa, which they argue are mostly quite small, have limited access to finance, and insufficient capacity.

91. ALBA members include Venezuela, Bolivia, Ecuador, Cuba, Nicaragua, Dominica, St. Vincent and the Grenadines, Antigua and Barbuda, St. Kitts and Nevis, St. Lucia, and Grenada. In 2006, ALBA expanded to include the Peoples’ Trade Treaty (Spanish acronym TCP), and in 2009 it accordingly became ALBA-TCP. ALBA members Antigua and Barbuda, St. Kitts and Nevis, St. Lucia, and Grenada are not members of the Bank of ALBA.

92. The most widely known of ALBA’s projects involves the Unitary System for Regional Compensation (Spanish acronym SUCRE), which is a regional payment system created in October 2009 (see Fritz and Mühlich 2014). ALBA’s portfolio of projects broadened beyond the SUCRE and the Bank of ALBA during the time that Venezuela enjoyed large oil revenues and several members had populist governments. ALBA moved into South-South aid and concessional financing via Petrocaribe, a Venezuelan program to provide oil to Haiti at preferential prices; exchanges with Cuba in which subsidized oil is bartered for medical services, education, and pharmaceuticals; and a variety of human development and capacity-building programs (Baranyi, Feldmann, and Bernier 2015; Riggirozzi 2012a; 2012b).

93. The discussion of ALBA draws on Armijo (2012), Baranyi, Feldmann, and Bernier (2015), Fritz and Mühlich (2014), Riggirozzi (2012a; 2012b), and Rosales (2013).

94. SUCRE officials report in personal communications that as of the end of November 2016, the payment system was still operating despite the challenges facing Venezuela’s economy.

95. Members of the BDS include Argentina, Bolivia, Brazil, Ecuador, Paraguay, Uruguay, and Venezuela. The discussion of the BDS draws on Armijo (2012), Marshall (2010), Marshall and Rochon (2009), Riggirozzi (2012a), Rosales (2013), Rosero and Erten (2010), and personal communications.

96. See Rosales (2013) and Bond (2016, 614) on Brazil’s role in reshaping the BDS vision.

97. Many of the participants (myself included) at a simultaneous meeting at the same venue where the BDS reached and announced this agreement were consequently surprised by the news.

98. Personal communication with official.

99. Disbursal criteria for both forms of support available from the CMIM and CRA are currently under consideration.

100. In a related vein, Weller and Zulfiqar (2013) show that greater financial market diversity (i.e., different types of institutions serving different constituencies) can contribute to stability.