1 Introduction

In this chapter we apply a financial infrastructural framework to analyze the politics of global correspondent banking relationships (CBRs), the agreements by which a bank in one country agrees to provide financial services to clients on behalf of another bank elsewhere in the world.¹ CBRs are a vital aspect of the global economy, enabling institutions and people access to global financial markets and services (Rice, von Peter, and Boar, Reference Rice, von Peter and Boar2020). They are especially important for the sending and receiving of remittances, the lifeblood of survival and development for people and countries around the world (World Bank, 2022). Banks have maintained CBRs at least since the eighteenth century and, until recently, have consistently expanded the network of CBRs globally (Schenk, Reference Schenk2021).

In the early 2010s, however, banks in the Global South began raising an alarm. Banks in wealthier economies, they said, were unceremoniously severing these agreements, leaving banks in the Global South with fewer financial service providers. The potential impact was considerable: higher costs for remittance inflows, slower service times for financial services, limited access to global capital markets, or even an inability to fund vital imports like food and energy.

For their part, banks culling these relationships blamed overzealous regulation targeting money laundering and terrorism financing. Soon, global financial institutions, the financial press, and policymakers all were echoing the same refrain. The rising costs of compliance and the risks of noncompliance meant that relatively low-profit correspondent accounts were not worth the trouble. De-risking, they argued, threatened to increase the cash intensity of developing economies, thereby reducing regulatory oversight. De-risking showed that the Anti-Money Laundering and Counterterrorism Financing (AML/CFT) regime had gone too far and was possibly even counterproductive (see, e.g., The Economist (2014) for a summary of this position).

Despite little evidence that the roots of de-banking lay in AML de-risking, the critique appeared to shift the global AML/CFT agenda. Global meetings of AML experts focused less on assessing the regime’s effectiveness and more on alleged unintended consequences. Put differently, for the first time in the AML/CFT regime’s twenty-five-year history, the AML agenda focused on whether there was too much regulation. Why was this critique so effective, when other, more established critiques about the regime’s fairness (e.g., Naylor, Reference Naylor2001; Vlcek, Reference Vlcek2018; Gilmour, Reference Gilmour2022) and effectiveness (e.g., Findley, Nielson, and Sharman, Reference Findley, Nielson and Sharman2014; Gutterman and Roberge, Reference Gutterman, Roberge, Allum and Gilmour2019) had less impact?

An infrastructural framework helps us understand this case and sheds light on the politics of global correspondent banking more broadly. Efforts to reassess regulation really only gained traction once banks flexed their infrastructural power and began cutting CBRs: de-banking under the guise of de-risking. The result was a powerful coalition: global banks seeking to roll back regulation, de-risked jurisdictions seeking to maintain access to global financial markets, and political actors in the Global North who for various reasons were interested in maintaining North–South financial relations.

The case also provides useful insights regarding the politics of global financial infrastructures. While Mann’s infrastructural framework and much of the resulting literature stresses the state’s far reach into civil society, this case, like Braun (Reference Braun2020) and Braun and Gabor (Reference Braun, Gabor, Mader, Mertens and van der Zwan2020), highlights how the state’s entanglements in liberalized global financial markets can limit the state’s own power vis-à-vis banks. Much of the global financial infrastructure is not a public good, but a private one. The social and material ties that make up financial infrastructures are not accidental but carefully cultivated and often colonial in origin and practice. They are not friction-free pass-throughs but power-laden relationships. They are not always in the background, at least for those to whom access is not guaranteed. Finally, the social relationships and material connections that comprise CBRs are path dependent. There are moments at which the right drivers can create critical junctures and generate more fundamental change in the system. We discuss two such potential drivers: alternative payment systems and the “indigenization of finance” (Griffin and Martin, Reference Griffin and Martin2023, p. 4). The overwhelming signal from the case, however, is that banks in the Global North retain primary control over the infrastructures of global finance and the financial and political power that comes with it.

In the following section we briefly review the literature on global financial infrastructures, stressing those aspects on which this research can shed the most light. In Section 3, we provide a primer on global correspondent banking and an analysis of the debate over de-risking, emphasizing the predictions made about its impact and how those claims shifted the political agenda relating to AML/CFT. In Section 4, we show that the predictions regarding de-risking were not very accurate. In Section 5, we analyze this disconnect – unrealized impact that led to an agenda shift – as a case of infrastructural power. We also examine what responses have developed that might lessen the control over finance that Global North banks currently have. In Section 6, we conclude by considering the lessons of the case of de-risking for our understanding of global financial infrastructures.

2 Financial Infrastructures and Infrastructural Power

As the editors of this handbook note in Chapter 1, financial infrastructures “make possible the movement of other elements that are vital for the economy, including services, capital, or know-how” (Westermeier, Campbell-Verduyn, and Brandl). The servers and computers and networks that comprise the informational backbone of global finance are a vital part of the story. Procedures or processes are part of the infrastructure, as are relationships. In other words, the global financial infrastructure comprises a set of relationships, made active by sets of material objects, relationships, and ideas that define how finance can and should flow.

Bernards and Campbell-Verduyn (Reference Bernards and Campbell-Verduyn2019, p. 777) argue that infrastructures, which “are as much a heuristic device as a specific set of objects,” display five key characteristics.

1. 1. Facilitation – they make other actions possible.

2. 2. Openness – they are nonexcludable.

3. 3. Durability – they persist over time.

4. 4. Centrality – they shape how markets function.

5. 5. Obscurity – they operate in the background and generally remain unnoticed.

Insofar as infrastructures are the way things get done, change must happen in and through infrastructures. At the same time, both the physical and social aspects of infrastructures generate positive feedback dynamics, creating path dependence and some inherent resistance to change – especially radical change – as new technologies (and new ideas and new participants) must interface with preexisting ones (Bernards and Campbell-Verduyn, Reference Bernards and Campbell-Verduyn2019).

Given the basis in Mann’s (Reference Mann2012) foundational work, much of the literature on global financial infrastructures emphasizes the power of the state. Control of, and through, infrastructures is one of the key ways that modern, liberal states exercise power beyond what older, more overtly despotic forms of the state were able to exercise (Carruthers, Reference Carruthers1999). Financial infrastructures can protect or even extend hegemonic power (Konings, Reference Konings2010). In this volume, Coombs distinguishes among four types of infrastructural power, all of which underscore the power of the state to shape the actions of market actors and, thus, the market itself.

The case of global CBRs also highlights, however, that the accrual of power by the state is only part of the story. To begin, control of infrastructure can lead to an accrual of power by nonstate, that is, market, actors, as well. This power gain is especially dramatic when it entails control over “public” goods, or perhaps better said, over goods on which the public depends (Mann, Reference Mann2012; Busemeyer and Thelen, Reference Busemeyer and Thelen2020). When a firm or group of firms become necessary in the provision of supposedly public goods, those firms can restrict and monetize access to that necessary infrastructure. This power is not dependent on a firm’s complete control of the infrastructure, but is likely more a question of whether market actors are functionally dependent on a firm’s services. For example, to be maximally competitive, market actors need to pay national stock exchanges for access to a series of mutually reinforcing products, making the exchanges “global providers of financial infrastructures” (Petry, Reference Petry2021, p. 576) and granting them significant market power.

This enhanced role for market actors can also mean that state actors face limitations on their ability to act as they might otherwise prefer. For example, the actions of central banks give rise to financial innovation and the increasingly important shadow banking system, creating an “infrastructural entanglement” (Braun, Reference Braun2020, p. 396; see also Braun and Gabor, Reference Braun, Gabor, Mader, Mertens and van der Zwan2020). Market governance is now dependent on that shadow banking, especially as market-based banking has become a more prominent tool of market governance (Braun, Reference Braun2020; Braun and Gabor, Reference Braun, Gabor, Mader, Mertens and van der Zwan2020). The state’s strategies for shaping market actors simultaneously create a dependence on those same actors. The result is a curtailment of the state’s reach and an extension of the reach of market actors. The case of CBRs, then, at least raises the possible outcome that market actors can control global financial infrastructures and thus gain power vis-à-vis the state. The politics of global financial infrastructures is not, in other words, only a question of which state has power or how much power a given state has vis-à-vis civil society, but rather which actors – state or nonstate – have what power and over whom?

This political dynamic is especially true of the global payments infrastructure, which is “profoundly political and was historically built through colonial violence and political struggle” (de Goede, Reference de Goede2020, p. 353; see also Westermeier and de Goede and Atme, this volume). These power dynamics have only intensified with the mutual financialization of security and the securitization of finance (de Goede, Reference de Goede2020; see also Amicelle and Jacobsen, Reference Amicelle and Jacobsen2016; de Goede and Sullivan, 2016). Especially relevant to our case study, de Goede, citing Bedford, argues that infrastructures “operate as ‘hard-wired’ forms of regulation, that ‘incarnate’ legal measures and give them force beyond normative codification” (Bedford, cited in de Goede, Reference de Goede2020, p. 354). CBRs are the backbone of this system. A necessary question, then, is who, if not the state, controls the global payments infrastructure?

Much has been made of the changing landscape in global payment systems, to be sure. The advent of blockchain technology might boost the potential for decentralized finance, enhanced capacities for person-to-person lending, and the disintermediation of banks. We return to this later in this chapter, but note here that these changes arguably still represent more potential energy than kinetic.

A second contender for the largest landlord in the global payment space is SWIFT, the Society for Worldwide Interbank Financial Transactions. Scholars very recently have begun taking up the task of understanding SWIFT, especially from an infrastructural perspective (see, e.g., Robinson, Dörry, and Derudder, this volume). As the group’s involvement in sanctions enforcement against Iran and Russia show, SWIFT is not a “neutral payment network,” but rather “is inscribed with power and security politics from their beginnings, and remain important but overlooked sites of political-economic power” (de Goede, Reference de Goede2020, p. 352). The case of SWIFT, de Goede argues, is a reminder that infrastructures have agency, embody political rationality and sediment power relations, and construct and enable collectivities. Particularly relevant to our research, de Goede finds that: “Infrastructures hard-wire core-periphery structures, routings, and dis/connections, with profound effect on the space of (political) community” (de Goede, Reference de Goede2020, p. 356). More specifically, SWIFT, through an unequal internal power structure, reinforces preexisting external power structures, lending additional power to just a few international financial centers (Dörry, Robinson, and Derudder, Reference Dörry, Robinson and Derudder2018).

While we do not dispute that SWIFT reinforces preexisting power asymmetries, the case of de-risking reminds us that banks, distinct from the states in which they operate, are also powerful actors and remain so even in an era of shadow banking (Nosrati et al., Reference Nosrati, Kern, Reinsberg and Sevinc2023). In fact, banks ultimately were able to use their infrastructural power to push states in a direction that states previously had resisted. In the case study that follows, we consider all of these dynamics: private ownership of a global infrastructure, the political power stemming from that ownership, and which actors, ultimately, have the greatest share of that political power.

3 The Debate over De-risking: Claims and Consequences

Correspondent banking relationships are “an arrangement under which one bank (correspondent) holds deposits owned by other banks (respondents) and provides payment and other services to those respondent banks” (BIS, 2003, p. 16). Put most directly, “correspondent banking requires the opening of accounts by respondent banks in the correspondent banks’ books and the exchange of messages to settle transactions by crediting and debiting those accounts” (BIS, 2016, p. 9). The relationships are generally reciprocal and normally entail different currencies. They are essential elements of the global economy, facilitating investor access to global financial markets, and ensuring global trade and remittances can continue to flow. For decades, reaching back to the Herstatt crisis in 1974 (Mourlon-Druol, Reference Mourlon-Druol2015), the system has worked without any apparent crisis caused by correspondent banking, even as various rounds of globalization and technological shifts meant a massive increase in the system’s complexity.

In the early 2010s, this long period of stable growth hit a pothole. Financial institutions in the Global South voiced concern over a pattern that saw banks in the Global North cutting off CBRs or, in some cases, pulling out of countries or territories entirely, even in cases where there was long tradition of business relations.² Banks in the Global North pointed to what they claimed was overzealous anti-money laundering regulation: regulators and bank examiners required not just “Know Your Customer” policies, but also, they argued, that banks “Know Your Customer’s Customers.” In addition, banks argued that the fines for violations were increasing dramatically. The Customer Due Diligence required to meet such a standard, and the risks of failing to do so, were simply too costly to make maintaining CBRs worth it.

Not everyone accepted this connection. The global standard setter for the AML/CFT regime, the Financial Action Task Force (FATF), provided a direct counternarrative. FATF defined de-risking as “terminating or restricting business relationships with clients or categories of clients to avoid, rather than manage, risk in line with the FATF’s risk-based approach” (FATF, 2014). It also underscored the variety of possible drivers beyond AML: “De-risking can be the result of various drivers, such as concerns about profitability, prudential requirements, anxiety after the global financial crisis, and reputational risk. It is a misconception to characterise de-risking exclusively as an anti-money laundering issue” (FATF, 2014, emphasis added).

Despite FATF’s counterclaims, ultimately the discourse linking AML to de-risking took hold. Think tanks like Oxfam International and the Center for Strategic and International Studies echoed the claim (Durner and Shetret, Reference Durner and Shetret2015; MacDonald, Reference MacDonald2019). It was repeated in hearings on de-risking held by the US House Financial Services Committee (FSC Majority Staff, 2018). The international financial institutions joined in, although they also allowed for the possibility of other drivers (World Bank, 2015, p. 9). In the summer of 2016, when the de-risking discussion was arguably at its peak, the World Bank and the Association of Certified Anti-Money Laundering Specialists (ACAMS) hosted a two-day stakeholder workshop to discuss the problem. Their report on the findings and recommendations of that workshop reflect all of the discussions mentioned (World Bank and ACAMS, 2016). ACAMS and the World Bank hosting the workshop is reflective of the dominance of the discourse that de-risking was a problem and AML was the cause.

Not only did this discourse become a consensus in public discourse, it changed policy discussions, too.³ In the wake of the de-risking debate, efforts started in both Europe and the USA – the two strongest proponents of the AML/CFT regime – to reconsider elements of the project. Following two calls for comments in 2020 – one on a proposed rationalization of AML regulation, the other on the drivers and impact of de-risking – the European Banking Agency published an opinion on de-risking which explicitly links de-risking to poor AML risk management and acknowledges that there are ways in which regulatory and supervisory authorities can discourage de-risking with clearer guidance and less categorical judgements (EBA, 2022). In the United States, also in 2020, a law ostensibly designed to strengthen AML/CFT regulation – the Improving Laundering Laws and Increasing Comprehensive Information Tracking of Criminal Activity in Shell Holdings Act (aka the “ILLICIT CASH Act”) – begins with a discussion of the dangers of de-risking and calls for a study of how best to address the problem. Since then, the US Treasury has published a de-risking strategy (Department of the Treasury, 2023). Perhaps the best example of the diffusion of this discourse is FATF itself. In 2021, FATF began focusing on the regime’s “unintended consequences,” with de-risking noted as a key item.

This success in changing the framing on AML/CFT is surprising. FATF was founded in 1989 and the regime it guides has advanced steadily since, creating tighter standards, developing more tools for enforcement, and working to diffuse the standards on a global scale. In short, proponents have steadily deepened and widened the regime over thirty years. This has been true despite critiques that evidence for the regime’s effectiveness was lacking. This intensification of the regime increased after the 9/11 attacks in the United States, when members added counterterrorism financing to the regime’s goals. The regime’s securitization made it difficult to critique efforts. Adding to the puzzle, FATF members in 2012 decided to move away from tracking only whether the right laws were on the books and focus also on whether states’ regulatory systems were actually effective at preventing laundered money from entering the financial system. So how did de-risking move to the top of the agenda? In the following section, we show that the actual impact of de-risking is likely not the answer, as the impact was not as dire as many predicted. These muted outcomes and the actual patterns of de-risking suggest that something other than concern over AML was driving the de-risking agenda.

4 The Impact of De-risking: Much Ado about a Little?

One possible explanation for the rise of de-risking to the top of the AML agenda is that de-risking was, in fact, having the predicted impact. That is, it was widespread, it was costly, and it was caused by AML rather than some other more pedestrian cause. In this understanding, moves across the regime to consider reform are, therefore, positive and surprising signs of responsiveness. To consider this explanation, we draw here on the plethora of reports that were generated as a result of this concern. The patterns of de-risking found in that body of evidence, we argue, suggest that the predicted impact of de-risking ultimately was more dramatic than its actual impact.

5 Discussion

Predictions about the impact of de-risking were dire. To be sure, a few specific cases seem to live up to these predictions and they are important in their own right. But the evidence given suggests that the systemic impact of de-risking was either overstated, largely averted, or some combination of the two. Those dire predictions laid the blame for this coming doom squarely at the feet of AML: standards and consequences that were too high, especially in Europe and the USA; systems that were too weak in the de-risked jurisdictions. These predictions hit their target. The AML/CFT regime had been largely immune to critique since its inception, but especially after members securitized the regime following the 9/11 attacks in the United States. The fears over de-risking shifted the agenda.

An infrastructural perspective suggests that global banks were able to achieve this by leveraging their control of CBRs. The threat of exit from the world’s most valuable markets as a means of protesting AML was never a credible one. The threat of big banks to exit CBRs, however, was.

This strategy only works, however, if other important actors were interested in ensuring that banks in the Global South had continued access to CBRs with banks in the Global North. In other words, someone had to care. There are surely varying rationales for those concerns. For some, especially those on the receiving end of de-risking, the potential economic harm was the motivation. Their concern might have been necessary, but it surely was not sufficient to shift the agenda.⁷

For people and institutions in the Global North, other rationales might be at play. On the micro-level, former Chairwoman of the United States House of Representatives Committee on Financial Services, Maxine Waters, has a long-standing interest in US–Caribbean relations. As a result, de-risking was and remains a particular concern for her and she uses her institutional power to keep the issue high on the agenda. For example, it was under her watch that the Committee held multiday hearings on de-risking. In April 2022, Waters co-chaired, with Barbados Prime Minister Mia Mottley, the “Roundtable Discussion on De-risking and Correspondent Banking” (Salmon, Reference Salmon2022).

There are other interpretations. It is arguably in the interest of advanced economies to avoid economic meltdowns and promote sociopolitical stability in their respective neighborhoods. There are postcolonial interpretations of these interests, too. Ensuring the continued financial dependence of other countries on US and European banks allows the USA and European Union (EU) to exercise oversight and control over those countries. In this sense, building and maintaining CBRs constructs political power for states, but also for banks who can deny those services.⁸ De-risking represented a threat to that system of control. In Europe, the European Commission is funding a correspondent banking program to address issues within the EU and collaborate with the EU AML/CFT Global Facility, which provides capacity building to non-EU countries, with meetings such as a new regular regional conference on correspondent banking in Eastern and Southern Africa (EU AML/CFT, 2022).

The ability of banks to leverage CBRs toward shifting the agenda on AML/CFT shows that banks, and especially banks in North America and Europe, still hold the lion’s share of power in global financial relations. While scholars correctly have paid more attention to SWIFT in recent years, the case of de-risking shows that banks retain substantial agency. The de-risking story is also a reminder that global financial infrastructures are not, in fact, global public goods; they are not nonrivalrous and nonexcludable. It may be the case that banks are unable to completely shut off the taps; it is especially the case that no one bank can on its own exclude access, at least in the long term. Cutting off CBRs can reduce the flow to a trickle, however, unless and until alternative arrangements can be made. In global finance, this means higher costs, lower competitiveness, and less growth.

De-risking and CBRs are also a reminder that infrastructure is neither a given, nor accidental. Infrastructure is, in fact, a set of relationships. Like all relationships, these must be cultivated, fostered, and tended to. And like all relationships, there can be imbalances of power that are exploited for the benefit of one side over the other. In the case of CBRs, banks give considerable thought to CBRs. But the case also shows that states see them as a point of considerable concern. In this sense, the common refrain that infrastructures are invisible until they break down may be an overstatement, at least for those who can be easily excluded from them. Those on the weak side of these agreements are likely more sensitive to the fact that the agreements are neocolonial and simultaneously create long-term vulnerabilities and short-term opportunities; to them they are both visible and politicized (see also Langenohl’s contribution to this handbook).

In that vein, it bears watching whether any developments or initiatives may lead to a loosening of those colonial ties. We note a few possibilities and discuss those here by way of conclusion.

The continued disintermediation of banks via alternative payment systems is one such possible point of reform (see also Nölke, this volume). These systems are able to bypass SWIFT and traditional banks more generally. To date, however, they represent a small fraction of global business. In the longer term those numbers might increase. But to the extent that those technologies remain rooted in, or are routed through, international financial centers, they risk recreating or reinforcing the same power structures that shape global finance today, just as SWIFT has (Brandl and Dieterich, Reference Brandl and Dieterich2023). To the degree that these alternative systems are built around technologies that are highly energy-intensive, their use risks exacerbating the climate change that already poses an existential threat to precisely the people that the new technologies are supposed to help.

The G20 is undertaking an initiative designed to prevent this kind of problem in the future. This includes the development of multilateral payment platforms, which are designed specifically to operate across jurisdictions and thus could supplant CBRs (see BIS, 2023). It also includes the use of “liquidity bridges,” agreements among central banks to allow extraterritorial branches of a bank to draw on their local central banks using collateral held in in the home office’s territory. The theory is that this reduces the amount of collateral that a service provider is required to hold, thereby reducing the costs (see BIS, 2022). Both are part of a larger “G20 Roadmap for Enhanced Cross-Border Payment” system (Financial Stability Board, 2023).

A final note of optimism comes from what Griffin and Martin (Reference Griffin and Martin2023) refer to as bank indigenization (see also Roitman’s contribution to this handbook). In the eastern Caribbean, at least since 2009, some leaders have been calling for a “rationalization” of the regional financial system, moving from a multitude of small, independent economies dependent on large international banks, and toward a multicountry economy. Decisions by international banks to close CBRs have galvanized previously lagging efforts on this front, leading to the creation of consortia of indigenous banks, which together were large enough to buy assets from foreign banks that had pulled out of the region (Griffin and Martin, Reference Griffin and Martin2023).

That said, as Bernards and Campbell-Verduyn (Reference Bernards and Campbell-Verduyn2019) have argued, infrastructure is a process as much as material object. Efforts to find alternatives to the neocolonial politics of global financial infrastructure are bound to meet resistance (see also Kaltenbrunner and Orsi, this volume). Some of that resistance is the result of path dependence. Some of the resistance is more strategic, as the owners and users of that infrastructure seek to maintain the advantages that come from it. That includes banks writ large, but also political actors that are able to achieve political goals by borrowing an otherwise private infrastructure. This exemplifies Mann’s idea that the same infrastructure can be used simultaneously by different social powers and toward different ends. This case raises the question of what happens when two actors are using the same infrastructure for seemingly opposite ends: one to exclude, one to include. Time will tell who “wins” that tug-of-war, but the logic of an infrastructural approach suggests that the advantage rests with those who control the infrastructure.