1 Introduction

Over a decade ago, cultural anthropologist and sociolegal scholar Bill Maurer called for attention to “the act and infrastructure of value transfer” – that is, to payment, a domain long neglected in social studies of finance (2012a, p. 19). Payment, he argued, was important to understand both because it was a key cultural, economic, and technological form and because its cultures, economies, and technologies were rapidly changing. In other words, it served the infrastructural functions of money.

In the first decades of the 2000s in the United States, payment infrastructures are rapidly becoming more powerful. As scholars such as O’Dwyer (Reference O’Dwyer2023) and Westermeier (Reference Westermeier2020) demonstrate, finances are increasingly becoming platformized, with large, data-driven companies working to embed payments within their platforms and seeking to profit from access to users’ transactional data.

Understanding payment as an act and infrastructure underscores that money is a communication medium, a way of transmitting information that produces shared meaning (Swartz, Reference Swartz2020). Payment technologies – cash, cards, checks, or apps – do not simply transmit value. They communicate the nature of the relationship between two parties and reveal information about how we see ourselves and how we are viewed by powerful institutions (Zelizer, Reference Zelizer1997). As a communication infrastructure, payment binds us together in a transactional community, a shared economic world, and this shores up new and existing inequalities.

Payment infrastructures are increasingly being understood as a form of “social media.” In the broader communication context, there is a sense that the era of mass media (characterized by unidirectional broadcast and print technologies) has given way to one of social media (digital media that is niche, participatory, peer-to-peer, globalized, and surveilled). So too, mass money media have shifted to social money media. If, as geographer Emily Gilbert suggests, state currency was designed to enact “mass” transactional communities at the scope of the nation-state, what kinds of transactional communities do new “social” payment systems entail?

Of course, there is no way to neatly segment the two; most critical scholars are uncomfortable even using the terms “mass media” and “social media” (see, e.g., Papacharissi, Reference Papacharissi2015) as though they describe discrete phases. Similarly, the shift from mass money media to social money media is more complicated than it sounds. The mass medium of cash has always been accompanied by other money tokens, including foreign currency, coupons, and checks (Carruthers and Babb, Reference Carruthers and Babb1996; Henkin, Reference Henkin1998).

Yet social media money is helpful shorthand for an industry, a way of describing a certain set of technologies and a series of norms and engagements. To be sure, money has always been social and money has always been media. As a media technology, payment infrastructure is currently being redesigned to look more like social media, largely by Silicon Valley.

But this redesign brings up new questions: Who gets to control payment? Communication technologies come with constraints that can exclude potential users from the transactional communities produced by those forms of payment. Despite being a state technology, cash is difficult to control or surveil and has a low barrier to entry. New money media, created in the image and footsteps of social media, will not be equally accessible. As we move from mass transactional communities to social transactional communities, what are the implications of this shift? Who will monitor, control, and restrict these new payment rails?

2 Transactional Memories: Social Payments and Data Economies

The media studies scholar Josh Lauer (Reference Lauer2017) demonstrates that the concept of “financial identities” is centuries old: credit bureaus of the nineteenth and early twentieth centuries collected information and stories to determine borrowers’ ability to repay. These reports, Lauer says, made modern people into “legible economic actors” (p. 35).

But in recent years, there has been a shift from these aggregations of financial data to what I call transactional identities, shaped by how, where, when, and with whom we transact. If the former constituted identity via credit, the latter constitutes identity via payment infrastructure. We perform these transactional identities through payment, and they shape how others view us, whether as Amex Black card holders or electronic benefits welfare card users.

These transactional identities are being shaped in real time by social money media, as exemplified by Venmo, a peer-to-peer mobile payment app that is especially popular in the United States, where bank transfers are expensive and cumbersome. Venmo includes a social “feed” of transactions, visible to friends, similar to a Facebook news feed or Instagram post. The app requires users to annotate their transactions with notes and encourages them to use emojis: pizza, taxi, clinking wine glasses.

In this way, Venmo illustrates what sociologists Alya Guseva and Akos Rona-Tas (2017) call the “new sociability of money”: The ability of digital money technologies to “preserve the details of economic transactions, to capture our geographic movements, to infer our tastes and routines” (p. 204). The app also gives rise to new forms of social communication, including playful interactions and coded messages (Acker and Murthy, Reference Acker and Murthy2018). Its social streams reinforce, memorialize, and even potentially strain social relationships (Drenten, Reference Drenten, Belk and Llamas2022).

Venmo reveals that money has always been social, but it also encloses that sociality within its platform and records it for perpetuity (O’Dwyer, Reference O’Dwyer2019). Scholars across fields have argued that money is a technology of memory (see, e.g., Kocherlakota, Reference Kocherlakota1998). Social media is also a technology of memory; it is part of what the media scholar Jordan Frith has called “a new memory ecology” assembled on mobile phones (2015, pp. 90–91).

These transactional memories can also be used for surveillance and control. Nigel Dodd writes that “a device for remembering cannot be divorced from the criticism that it is also a vehicle for political and commercial surveillance, above all, as long as the technology involved is controlled by corporations and states” (2014, p. 296, original emphasis). Kelsi Barkway (Reference Barkway2023) documents how even seemingly benign technologies (in this case, benefits cards for distributing welfare payments) can be perceived as tools of surveillance and social control, inciting fears in users that their spending is being monitored and they will be judged as undeserving.

Digital transactions are, in fact, being used for control and punishment. Police in Atlanta have used access to an activist’s PayPal account to bring criminal charges against a bail fund that has supported protests against the construction of a large police training facility (Lennard, Reference Lennard2023). Consumers may not be aware that their transactional memories can be used to cause them trouble, as in the case of mobile-phone borrowers in Kenya, whose phones offer them access to credit while simultaneously generating a stream of financial and personal data that can harm their credit scores. Social scientists Kevin P. Donovan and Emma Park (Reference Donovan and Park2022a, 2022b) note that many of these borrowers end up seeking expensive short-term credit to prevent damage to their credit scores, ensnaring them in a cycle of debt that is difficult to exit, a form of “predatory inclusion.”

In fact, many of our experiments with digital payment technologies are being enacted on the world’s most vulnerable people. Aaron Martin (Reference Martin2019) documents how mobile money platforms facilitate both familiar and novel forms of surveillance of users by government entities and service providers. Similarly, anthropologist Margie Cheesman notes that companies and aid organizations providing support in refugee camps are testing out web3 technologies such as blockchain wallets to distribute various kinds of payments. In this environment, where users have limited choices, rights, and protection, forced use of these technologies requires them to generate financial data over which they have little control (Cheesman, Reference Cheesman2022a). Thus, she recommends that web3 technologies not be used experimentally among vulnerable populations (Cheesman, Reference Cheesman2022b).

3 Chokepoint Power: How Controlling Payment Infrastructure Controls Users’ Lives

The systems that allow us to get paid, like many other critical infrastructures, are largely invisible until they stop working (Star, Reference Star1999; Edwards, Reference Edwards, Misa, Brey and Feenberg2003). And when those systems stop working, it often comes as an account that is frozen without warning, perhaps for opaque reasons. Users may have little recourse, and the consequences of an account freeze can be severe.

In response to the 2008 financial crisis, the US Department of Justice and the Financial Fraud Enforcement Task Force launched Operation Choke Point. The name of the project is notable: The task force had the power to constrain merchants’ ability to get paid, targeting fraudulent institutions by “choking them off from the very air they need to survive” (Zibell and Kendall, Reference Zibell and Kendall2013). To fully participate in a transactional community – to be a “citizen” of that community – you need unobstructed access to a system of payment, because a system that can suddenly cut you off is more dangerous than not having access at all.

In recent decades, the business of getting paid has been changing in critical ways. In the United States, for example, payment acquisition systems are shifting from traditional independent sales organizations or ISOs (middlemen organizations that serve as payment service wholesalers), to tech startups looking to disrupt the payments system (see Figures 24.1 and 24.2). Payment cards were originally designed for an economy in which the line between buyers and sellers was clear; modern payments companies facilitate peer-to-peer payments in a geographically dispersed communication system.

Figure 24.1 Traditional and platform models of how payments are acquired.

Source: Created by Meghan Macera.

Figure 24.2 The communication of a card payment.

Source: Created by Meghan Macera.

In the 1990s, an emerging set of payment service providers (PSPs) overlaid new systems on existing infrastructure, bridging old and new technologies and policies (some more successfully than others). The first of these providers, and likely still the most successful, was PayPal, which created parity between users. Its primary innovation was to bypass the old payment acquisition system by keeping money in a closed loop on its platform for as long as possible. This has become the predominant model for PSPs coming out of the tech sector, such as WePay, Square, Venmo (now owned by PayPal), and most of the embedded social media payment systems, such as Facebook Messenger Payments.

However, in the midst of this shift, the way that payment providers manage risk has also changed, in ways that do not benefit – and indeed, can imperil – consumers. This is a shift in what sociologists call “riskwork” – how risk is imagined and managed – that can lead to payment shutoffs for vulnerable users. Managing payments means managing risk, and managing risk is inherently political.

In the traditional ISO model, issuers represent the interests of cardholders, while acquirers represent the interest of merchants and there is a marketplace for payments in high-risk industries. In contrast, within the new payments model, the PSP’s client is the platform, not the parties who are transacting. Risk is managed, not through a marketplace model, but a standard tech-industry mechanism: the terms of service (TOS) agreement, which users must agree to (but don’t usually read) when they sign up for an account, TOS agreements can change at any time. and there is no compelling interest for tech companies to find a way to manage risk when they can simply ban any transactions they deem “too risky.”

As is common in the realm of social media, these peer-to-peer payment systems use surveillance and automation to enforce TOS agreements and mitigate risk. Surveillance scholars have noted that the power of surveillance extends beyond watching to identifying, classifying, and assessing (Gandy, Reference Gandy1993; Lyon, Reference Lyon2002), making surveillance a form of “social sorting.” As Fourcade and Healy point out, the “classification situations” produced by the wrangling of “big data” are “presented, and experienced, as moral-ized systems of opportunities and just deserts.” They “have learned to ‘see’ in a new way and are teaching us to see ourselves that way, too” (2017, p. 10).

Although the tech industry could, theoretically, develop systems that profit from varied appetites for risk, like traditional ISOs, there has instead been a shift toward probabilistic modeling and monitoring, using machine learning to monitor users’ social media presence to flag “high-risk” transactions and ban them.

But this gives rise to a variety of mistakes, like bans on users who tag a Venmo purchase “Cuban” for a sandwich, or playfully use a bomb emoji. Predictive analytics systems are always experimental and designed to live in “perpetual beta,” in which products are “developed in the open, with new features slipstreamed in on a monthly, weekly, or even daily basis” (O’Reilly, 2005). This makes it even harder for users to predict what might earn them an account freeze.

To make things more confusing for users – and more perilous for users with fewer resources – TOS agreements tend to be unevenly enforced. As the internet researcher Tarleton Gillespie (Reference Gillespie2018) points out, platforms of all kinds routinely make seemingly arbitrary calls about what does and does not violate TOS. For banned users, often their only form of recourse is a byzantine and ineffective process, while, in the meantime, they have lost access to their stored funds, as well as to the transactional community of the platform.

This has been notably true for sex workers. As of 2018, the sex-worker activist Liara Roux has collected dozens of examples of discrimination against sex workers by financial services companies (Lake and Roux, Reference Lake and Roux2018; see also Blue, Reference Blue2015). Legislation and policies intended to reduce human trafficking also constrain sex workers, denying them access to the websites they use to work and make a living (Blunt and Wolf, Reference Blunt and Wolf2020).

We still haven’t created payment infrastructures that move at the pace of our modern world but maintain all of the key affordances of cash. Cash is anti-surveillant, self-clearing, immediate, and reliable (O’Brien, Reference O’Brien2017; Scott, Reference Scott2022). But the most vulnerable are often forced to choose between payment channels that are unreliable or totally inappropriate for the digital nature of their work. We are still not getting payments right – not for everyone, and not all of the time.

4 Money (and Everything Else): Increasingly Private, Segregated, Siloed

In recent decades, the card-issuing business in the United States has become increasingly competitive and stratified, producing niche transactional identities. Although payment card products are mandated to look similar and use the same infrastructure, they are imbricated in different infrastructural, economic, and discursive assemblages (see Deville, Reference Deville, Gabrys, Hawkins and Michael2013; Gießmann, Reference Gießmann2018). Some cards pay users back; others charge usage fees. Some are more expensive than others for merchants to accept. The architecture of the modern card network is marked by hierarchy, difference, and communication.

Merchants agree to pay slightly more in interchange fees to receive payment from rewards cards and other luxury credit cards designed for the most “desirable” consumers than they do from standard cards. But because merchants then increase their costs to account for interchange, some consumer advocates argue that customers wind up paying for their own – or other people’s – rewards (see, e.g., Schuh, Shy, and Stavins, Reference Schuh, Shy and Stavins2010). As Maurer (Reference Maurer2012b) has pointed out, this system doesn’t exactly fit the picture of the capitalist economy; it is a rare situation where competition among issuers for the “best customers” drives prices up for everyone.

Universally accepted payment cards are relatively new (see Swartz and Stearns, Reference Swartz, Stearns, Nelms and Peterson2019; Swartz, Reference Swartz2020). The earliest precursor to credit cards, Charga-Plates, emerged in the 1930s. Resembling dog tags, these metal rectangles could be used by department stores to quickly imprint a customer’s account information on a payment slip.

By the 1950s, the Diners Club charge card had emerged as the first universal third-party payment card in the US, although it did not give customers access to credit, and in fact preceded the first credit card by at least fifteen years. The club functioned like a membership organization, offering a range of services beyond credit. Merchants paid a fee to be able to accept Diners Club card payments (a closed-loop system), but were assured by the organization that these members would likely spend. Indeed, having a Diners Club card was seen as a ticket to an elite group who had access to “country club style billing” or receiving a folio bill (Sutton, Reference Sutton1958). By the end of the 1960s, these elite customers were flocking to American Express, another closed-loop charge card. An AmEx card was initially difficult to get, and the company made a name for itself over the next several decades as a product for the elite, conferring privileges and status (Grossman, Reference Grossman1987).

In the late 1960s, beginning with Bank of America’s BankAmericard, banks began to issue payment cards linked to consumer credit accounts (Evans and Schmalensee, Reference Evans and Schmalensee2001). These credit cards, unlike Diners Club and American Express, were easy for bank customers to access; even with poor credit, most Americans could be approved for some kind of credit card (Nocera, Reference Nocera1994), which meant that paying by card was no longer reserved for elite customers. The BankAmericard network was eventually licensed to other banks and became the Visa Network, an open-loop system that acted as an intermediary among a variety of banks, merchants, and cardholders. As the historian of technology David L. Stearns (Reference Stearns2011) explains, opening the loop – making it possible to pay across banks, card types, and indeed transactional identity classes – was the key innovation of the bank card system.

When regulations against interstate banking loosened, beginning in 1978, banks in open-loop networks started to issue credit cards on a national level, competing for customers on a much wider geographical scale. By the early 2000s, the payment card market had become differentiated enough to create a wide range of stratified transactional identities – from “ultrapremium” cards for wealthy (or at least choosy) consumers, to small-business credit cards, debit cards, secured credit cards for those with poor credit, and prepaid cards for consumers who were unbanked or underbanked.

Just as payment methods have become increasingly stratified, so too have our visions of the future of money. A variety of scholars, activists, and entrepreneurs have described futures in which money is digital and issued by nongovernment entities (Maurer, Reference Maurer2005; Brunton, Reference Brunton2019). In the wake of the 2008 financial crisis, many people were eager to try anything other than money as usual (Maurer, Reference Maurer2011).

One of these imagined futures is cryptocurrency. Beginning with Bitcoin, first theorized in 2008, these digital currencies intentionally create new transactional communities, rethinking the value, identity, space, time, and politics of money. Designed to be a kind of “digital gold,” Bitcoin took hold of the public imagination as a kind of “Magic Internet Money,” backed not by the government but by cryptographic scarcity, able to move at the speed of the Internet without the drawbacks of traditional payment systems like fees and surveillance (Maurer, Nelms, and Swartz, Reference Maurer, Nelms and Swartz2013; Swartz, Reference Swartz and Castells2017). Bitcoin has been joined by thousands of other cryptocurrencies, few of which have ever been accepted by vendors, but all of which are ways to reimagine nongovernment-issued money, which might outcompete and outlast state-issued currency (Swartz, Reference Swartz2018; Brunton, Reference Brunton2019).

The future of money might also look something like corporate currency, a reality that is already playing out in the forms of rewards programs and social media payments. Starbucks, for example, now issues something that hews very close to a private digital currency: Starbucks Rewards, a loyalty program in which members can earn “stars” for purchases and can load and reload funds on Starbucks gift cards for perks. As of 2023, the program had 30.4 million members and funds loaded on cards had reached US$3.3 billion (Starbucks, 2023).

Facebook has also dabbled in creating a corporate currency, first with the announcement in 2019 of Libra, envisioned as a universal, global digital currency: a one-world money. Rather than being niche and segmented, Libra was described as connecting users across currencies and payment systems. Unlike cryptocurrencies, Libra’s design involved large corporations managing its monetary policy and infrastructure. Although the project was eventually shut down, in 2019 Facebook also announced the launch of Facebook Pay (now Meta Pay), which integrated payments with the company’s suite of social media products.

Some new forms of currency seek to be more universally accessible, but whether they will be able to achieve this is uncertain. Central bank digital currencies (CBDC) are currently being “explored” by many of the world’s central banks. While designs for these currencies vary widely, they are by definition a liability of the central bank, accessible to the public. However, since central banks do not have the digital infrastructure to provide financial services directly to consumers, these currencies must be intermediated, and the questions of by whom and how are crucial. If CBDC are to be truly accessible, rather than replicating existing systems that lock some consumers out, they must be very carefully designed (Narula, Swartz, and Frizzo-Barker, Reference Narula, Swartz and Frizzo-Barker2023; see also Swartz and Westermeier, Reference Swartz and Westermeier2023).

The future of payment infrastructures seems increasingly stratified. Customers who are members of loyalty programs or who carry “ultrapremium” cards often have access to a differentiated experience, with a separate customer service portal, lavish treatment, and even different physical spaces within a hotel or sports arena. The flipside is true for shoppers using state benefits cards, who can only buy certain foods at the grocery store. Technology and culture scholar Nathaniel Tkacz (Reference Tkacz2019) argues that payment apps compete on the basis of offering not just payment but, perhaps more importantly, “experience” of the world. He explains that such “experience money” takes up ordinary transactions and “deliberately infuse[s]” them “with a coherent value proposition” (p. 277).

As our money becomes more plural, so too do our transactional identities. We may find ourselves using multiple monies, bouncing between different payment infrastructures, and thus oscillating in and out of a variety of transactional identities (Maurer, Reference Maurer2005, p. 13). We don’t know the shape of tomorrow’s transaction media: What is emerging is social media money: private, surveilled, and data-driven. Some new money forms are hierarchical and segmented; others are universal. It is likely that we will be asked to trust corporations with our money and our data and our ability to get paid. The segmentation of these money forms means that you could be living in a separate transactional community than the person sitting next to you, while their plurality means that we will be constantly shifting between different communities and different forms of money.

And it is worth noting that some marginalized communities are experimenting with cryptocurrency to express resistance to the imposition of colonial economics, even though most cryptocurrency projects ultimately fail (Cordes, Reference Cordes2022).

5 The “Social” Future of Payments

Venmo users with public feeds broadcast a lot of personal information without realizing it. Friends can watch each other meet cute, fall in love, and break up in the course of a few months’ transactions. See, for example, the 2017 art project by Han Thi Duc, Public By Default, which finds poignant stories in the Venmo lives of others.

But Venmo can also be used to conduct social experiments, like the one that unfolded in June 2020, as the USA convulsed with anger over the killing of George Floyd, a Black man, at the hands of White police officers. As tensions simmered, a handful of people, separately, in diverse geographic locations, began to experiment with a kind of informal reparations: peer-to-peer payments to Black people, either friends or strangers.

In Vermont, activists Moirha Smith and Jas Wheeler crowdsourced a list of Black people’s Cash App and Venmo accounts, titled “Wealth Redistribution for Black People in Vermont,” and posted it on Facebook. The accompanying “Letter to White People” noted that “one of the easiest … ways to support Black life, Black joy, Black safety, Black community is to give your money to Black people.” The list ultimately grew to over 300 names, and its organizers estimate that $65,000 was transferred in varying amounts.

Recipients reported that they felt weird about getting money from strangers, but did appreciate the funds (NPR, 2021). Meanwhile, a few Black people in other parts of the USA began to receive notifications that White friends and acquaintances had sent them small amounts of money, presumably as a form of reparation – but without any kind of organized campaign. These transfers tended to be small amounts of money, and recipients said they found them baffling and insulting (Gimlet, 2020).

Just as people are always finding new ways to communicate, we are also finding new ways to pay. Money is inherently social and new forms of sociality will necessarily be reflected in our payment systems. New money technologies, then, will offer an opportunity to make new kinds of transactional communities and also to make mistakes, forging a messy path toward an unknowable payments future.

1 Sociability in the Market

Individual participation in financial transactions has been a market feature since at least the days of the tulip mania. While in North America and Western Europe individuals have lost ground to institutional investors since the 1960s (Useem, Reference Useem1996), it is worth noting that in other major financial markets, especially Asian ones, they continue playing a significant role in terms of share of market transactions and volume. Since the late 2000s, though, we observe an increased participation of retail investors in market operations in North America and Western Europe too: Episodes such as the GameStop saga in 2021 – when groups of retail investors managed for a while to cause significant losses to some hedge funds – have brought some of this participation to public attention. Equally, periodic waves of popular enthusiasm for Bitcoin, tokens, or nonfungible tokens have contributed to this public attention, especially since of late crypto assets have gained regulatory legitimacy.

A common ground for these apparently disparate phenomena – GameStop was about the stock of a fading game retailing chain, while crypto manias are about a new and ill-defined class of assets – is the infrastructures that made them both possible. Chat forums such as Reddit, where retail traders coordinated their actions and summoned each other in real time, trading apps such as Robinhood, or crypto trading apps belong to the infrastructures that made possible this broader individual participation to financial transactions. Of course, as David Pinzur argues (this volume), we need to distinguish between ready-to-hand devices and infrastructures: Trading apps on smartphones and chatrooms belong to the former, while data centers, cloud computing, or transmission lines, as well as the algorithms calculating spreads on the GameStop stock (among many other things) would belong to the invisible background that solicits awareness only in moments of crisis. Yet, we have to notice here a few interrelated aspects: First, while communication infrastructures play a crucial role in finance (see also Coombs in this volume), social media have been seldom counted by academics as pertaining to financial infrastructures (though professional investors have recognized their significance). Second, we need to ask the question, how does social media, as part of this infrastructure, impact investor behavior? How do they (mis)align participants? What kind of social dynamics do they foster?

While financial markets are social by definition and communication has played a key role since their inception, this has been less recognized in benchmark models of financial decision-making, which have focused on individual behavior seen as striving toward utility maximization, grounded in an efficient processing of information, and risk aversion (e.g., Fama, Reference Fama1970). Forms of sociability and their consequences have been largely seen as imitative behavior and investigated as such (herding phenomena). Finance scholars have more recently recognized that social behavior in markets extends beyond imitation phenomena – hence the shift in focus toward “social finance” (Hirshleifer, Reference Hirshleifer2015) meant to emphasize a reorientation of investigations away from the presumption of individual decision-making to the effects of mediated social dynamics upon markets. While there is a decades-long body of financial research on individual investors, social media-influenced decisions are much less well understood. This opens a potentially fertile ground for dialogues between sociologists of finance and financial economists interested in social behavior.

Over the last fifteen years, social media have become more and more integrated with trading platforms, giving rise to what are called social trading platforms (STPs) (see also Tong and Preda (Reference Tong and Preda2023) for more detail). From the perspective of social research, STPs, as we have argued, add another dimension to the study of how evolving infrastructures reshape not only market institutions but also the behavior of participants.

The rise of general social media (such as Facebook) has been quickly followed by the rise of social media exclusively dedicated to traders and integrated with online trading, often built in a smartphone app (“Facebook” for traders). In the institutional realm, data providers such as Bloomberg have also integrated social messaging in their data provision. By offering much lower fees compared with traditional brokerages, coupled with a global outreach, STPs have managed to attract millions of individuals into financial transactions. Some of the largest STPs have millions of subscribers and revenues of over one billion US dollars (The Insight Partners, 2022). STPs offer platform-wide communication forums, as well as the possibility of building communication groups. Traders can exchange messages in real time – meaning as they trade and observe the market – either within distinct groups or with everyone who has an account on the platforms. At least as important, STPs use metrics for ranking the most successful traders and embed copying algorithms that allow participants to automatically copy the transactions of those traders deemed to be more skilled. Should the latter be successful, they receive a share of the profits made by those who have copied them. In this sense, STPs can be seen as integrating within broader societal trends of generating status differentials by means of commensuration and public rankings (Mennicken and Espeland, Reference Mennicken and Espeland2019).

For sociologists of finance and financial economists alike, there is very rich data to be studied from STPs, such as trading data, network data, and communication data (e.g., Tong and Preda, Reference Tong and Preda2023). These different types of data have become increasingly valuable with the rapid growth in technological innovations, such as AI and machine learning. Institutions or individuals may utilize these tools to construct trading strategies or even perform algorithmic trading. Trading data includes traders’ everyday trading records, such as daily balances, profits and losses, number of trades, trade sizes, and so on. Network data includes the structures of how traders are connected to each other as well as whether/how often they participate in the social communication features, such as online discussion forum (ODF) and one-on-one messaging. Communication data includes the discussion content on the ODF, revealing how traders perceive and frame market events, how they justify their trading decisions, and how they account for market events. As some STPs (and other trading platforms) have made this data available for social science research, it becomes possible to investigate how new communication infrastructures shape the social dynamics of markets. This chapter aims to shed light on this issue.

2 Sociability and Financial Performance

At least three streams of literature are directly relevant to the notion of “sociability” in relation to the financial performance of investors. We present three streams of literature to reflect the profound impacts of social interactions on financial decisions as well as the inherent skills and abilities of both professional and retail investors in financial markets. We aim to highlight the dynamic nature of human behavior, particularly in financial markets, in the presence of infrastructures that facilitate interactions among investors (e.g., social media). We argue that it is important to further understand the relationship between social interactions and investors’ financial performance, as well as the underlying mechanisms through which investors’ financial decisions are influenced.

The first stream of literature investigates the relationship between social interactions and investment biases, such as disposition effects (Heimer, Reference Heimer2016) and herding effects (Gemayel and Preda, Reference Gemayel and Preda2018b). We should make clear that the notion of bias, widely used in behavioral finance, does not mean “irrationality” or “prejudice” or attachment to stereotypes on the part of investors. It simply means that observed behavior does not fit the predictions of the benchmark model of individual decision-making – as such, “bias” should be understood as deviation from such predictions (it is used interchangeably with “effect” in the sense of empirically observed effects). This being said, most studies are silent on how social interactions through media impact the financial performance of individual investors (Heimer, Reference Heimer2014, Reference Heimer2016; Gemayel and Preda, Reference Gemayel and Preda2018a). Online communication represents a distinct form of social interaction. Research indicates that online chats can offer valuable information for individual investors, aiding their decision-making (Antweiler and Frank, Reference Antweiler and Frank2004; Das and Chen, Reference Das and Chen2007). A recurring theme in this body of literature is the emphasis on the significance of STPs (Gemayel and Preda, Reference Gemayel and Preda2018a, Reference Gemayel and Preda2018b) and information systems (Abuelfadl, Choi, and Abbey, Reference Abuelfadl, Choi and Abbey2016) through which individual investors make their financial decisions. Online trading platforms, including social interaction features, provide a unique avenue for researchers to explore the impact of social interactions on investor behavior and financial performance. It is important to note, however, that a majority of individual investors tend to experience financial losses on such platforms (Preda, Reference Preda2017). For instance, studies using data from investment-specific online social networks, involving 5,693 foreign exchange retail traders with around 2.2 million trades from early 2009 to December 2010, have examined the influence of social interactions on the disposition effect (investment bias). These studies have shown that after gaining access to social networks, traders tend to exhibit nearly twice the magnitude of the disposition effect. This effect refers to a trader’s tendency to sell winning stocks while holding onto losing stocks (Heimer, Reference Heimer2016). By utilizing data from the Consumer Expenditure Quarterly Interview Survey spanning from 2000Q2 to 2010Q1, Heimer (Reference Heimer2014) has demonstrated a strong association between social interactions and active portfolio management. This is more prevalent among active investors compared to passive investors. It is important to note that this study cannot establish the direction of causality in the relationship between sociability and active portfolio management, as acknowledged by the author. Furthermore, there is an implication that social interactions may increase risk-taking, which could potentially have a negative impact on the financial welfare of traders.

However, a fundamental question remains unaddressed in existing literature: whether being sociable in the market, involving more social interactions, is advantageous or disadvantageous for the financial performance of individual investors. Notably, the existing literature does not distinguish between individual investors in terms of their social characteristics. Future research should bridge this gap by examining the financial performance of individual investors in relation to their varying levels of sociability in the market.

The second strand of literature is in alignment with broader social sciences and natural sciences. It seeks to uncover the impact of social interactions on the financial performance of individual investors, from the perspective of complex human systems and social networks (Saavedra, Duch, and Uzzi, Reference Saavedra, Duch and Uzzi2011; Saavedra, Hagerty, and Uzzi, Reference Saavedra, Hagerty and Uzzi2011; Liu, Govindan, and Uzzi, Reference Liu, Govindan and Uzzi2016). This strand’s focus lies in understanding the complexity of human systems and the collective wisdom of human interactions rather than merely examining the outcomes of financial decisions (Pan, Altshuler, and Pentland, Reference Pan, Altshuler and Pentland2012; Altshuler, Pentland, and Gordon, Reference Altshuler, Pentland and Gordon2015). Research in this area highlights that the patterns and content of instant messages (IMs) sent and received by professional stock day traders in typical trading firms can be interpreted as indicators of collective wisdom among individual investors across various platforms and can potentially influence investors’ financial performance (Saavedra, Duch, and Uzzi, Reference Saavedra, Duch and Uzzi2011; Saavedra, Hagerty, and Uzzi, Reference Saavedra, Hagerty and Uzzi2011; Liu, Govindan, and Uzzi, Reference Liu, Govindan and Uzzi2016). For example, Saavedra, Hagerty, and Uzzi (Reference Saavedra, Hagerty and Uzzi2011) used a dataset consisting of 66 individual stock day traders in a typical trading firm from September 2007 to February 2009, including over 1 million trades, with 55% being profitable. Their findings indicate a positive association between synchronous trading and the probability of making a profit, and the levels of synchronous trading are closely related to the patterns of IMs. Similarly, Liu, Govindan, and Uzzi (Reference Liu, Govindan and Uzzi2016) examined a dataset from 30 professional day traders, covering around 886,000 trading records and over 1.2 million IMs from January 2007 to December 2008. Their research reveals a connection between the expressed emotions in online communications and the profitability of actual trades. Traders who exhibit minimal or excessive emotional expression tend to make relatively unprofitable trades, while those with moderate emotional expression tend to make relatively profitable trades. Pan, Altshuler, and Pentland (Reference Pan, Altshuler and Pentland2012) utilized data from the online STP eToro and provided evidence that social trades, often associated with crowd wisdom, are more likely to outperform individual trades. However, it’s important to note that social traders are not consistently optimal performers (Pan, Altshuler, and Pentland, Reference Pan, Altshuler and Pentland2012). These studies operate with a notion of collective or crowd wisdom that in part sends back to the established concept of herding, and in part attempts to identify emerging phenomena in communication processes, based on large datasets: Communication is synchronized with trading actions, while interpretive frames (for market events) emerge within communication and become objectified (more specifically, are iterated across communication sequences and cannot be attributed to a single source anymore). The results point to at least two effects of communicational infrastructures: action synchronicity and objectification of interpretive frames.

These studies suggest that social communication and interactions play a significant role in the decision-making process of individual investors, highlighting the need for a more precise behavioral model (Pan, Altshuler, and Pentland, Reference Pan, Altshuler and Pentland2012). Furthermore, Altshuler, Pentland, and Gordon (Reference Altshuler, Pentland and Gordon2015), using data from the same online STP (eToro), which involved over 3 million individual investors and more than 40 million trades spanning from 2011 to 2014, revealed an inverted U-shaped relationship between the average financial gains and the number of information sources used for decision-making. This suggests that having too little information is insufficient, while an excess of information can be harmful in terms of financial performance. As mentioned earlier, while some studies indicate an association between social interactions and financial performance, the literature does not investigate different degrees of communication in relationship to investors’ financial performance. Future research needs to address this question by taking into account different levels and degrees of communication and to develop an analytical model to explore the relationship between communicative interactions and the financial performance of individual investors.

The third strand of literature investigates the skills and abilities of investors (including both professional and retail) in relationship to the (positive) returns on investments. We distinguish here between professional and retail investors. This is because individual investors tend to exhibit different patterns of decision-making compared with professional investors (Preda, Reference Preda2017). In terms of professional investors, previous research indicates that approximately 24% of professional currency managers (drawn from a sample of thirty-four individual currency fund managers) have the potential to achieve significantly positive abnormal returns within a four-factor model in the currency market (Pojarliev and Levich, Reference Pojarliev and Levich2008). However, there is no evidence demonstrating that currency fund managers can consistently generate abnormal returns (Pojarliev and Levich, Reference Pojarliev and Levich2010). In contrast, when we consider retail investors, conventional wisdom suggests that, in the stock market, active trading individual investors tend to underperform passive trading individual investors. This underperformance is often attributed to the costs associated with a high level of trading (turnover) (Barber and Odean, Reference Barber and Odean2000). However, other studies present evidence that within the highly active individual investors there exist small subsets of individual investors that earn abnormal returns (Goetzmann and Kumar, Reference Goetzmann and Kumar2008; Dahlquist, Martinez, and Söderlind, Reference Dahlquist, Martinez and Söderlind2016). For instance, in Sweden’s Premium Pension System approximately 5.8% of active and 0.6% of highly active individual investors earn significantly higher returns, achieving average returns of 6.86% and 12.57% per year, respectively. This is in comparison to the remaining 93.5% of inactive individual investors who achieve average returns of 3.82% per year. These active investors manage their investments by reallocating money from different funds in their pension accounts (Dahlquist, Martinez, and Söderlind, Reference Dahlquist, Martinez and Söderlind2016). Moreover, there is evidence suggesting that around 2% of high-turnover and under-diversified individual investors’ portfolios perform better than their high-turnover and better-diversified counterparts in the stock market (Goetzmann and Kumar, Reference Goetzmann and Kumar2008). This demonstrates that active trading is not always hazardous to wealth, at least for some investors, although their proportion is quite small. In the context of individual currency investors, which is the focus of this chapter, certain studies employing a four-factor model (Pojarliev and Levich, Reference Pojarliev and Levich2008) indicate that individual currency investors can achieve abnormal returns even after accounting for transaction costs (Abbey and Doukas, Reference Abbey and Doukas2015).

Building upon these strands in existing literature, it becomes evident that they all place significant emphasis on communication and instant messaging within the context of STPs. Sociability, as implicitly depicted in these studies, revolves around engaging in communication with other traders through instant messaging and participating in community discussions. However, existing studies do not furnish clear-cut evidence regarding whether this sociability, understood as engaging in communication, has a positive or negative impact on financial performance. The underlying assumption is that the “wisdom of crowds” is superior to making decisions independently. But is this indeed the case? Does online communication with other traders enhance financial performance? On the one hand, one can argue that engaging in online communication enables traders to swiftly exchange information and acquire knowledge. On the other hand, however, an opposing argument can be made – that online communication distracts traders and exerts a detrimental influence on their performance. The question of whether sociability in the form of communication is ultimately advantageous or detrimental to financial performance remains a pivotal one.

3 Communication and Survivorship: A Case Study of STP

Communication alters investors’ trading behavior and decision-making process (Heimer, Reference Heimer2016; Han, Hirshleifer, and Walden, Reference Han, Hirshleifer and Walden2022; Tong and Preda, Reference Tong and Preda2023). Traders can be influenced by communication with friends and neighbors in terms of stock market participation (Hong, Kubik, and Stein, Reference Hong, Kubik and Stein2004; Guiso and Jappelli, Reference Guiso and Jappelli2005) and investing strategies (Han and Hirshleifer, Reference Han and Hirshleifer2012; Heimer, Reference Heimer2014). Empirical studies have documented that communication plays a role in retail traders’ decision to start trading in equity and foreign exchange (FX) markets (Brown et al., Reference Brown, Ivković, Smith and Weisbenner2008; Kaustia and Knüpfer, Reference Kaustia and Knüpfer2012; Changwony, Campbell, and Tabner, Reference Changwony, Campbell and Tabner2015; Chen and Roscoe, Reference Chen and Roscoe2017). Against this background, it is intuitive that traders can also be influenced by the conversations they have with other traders while trading, especially when they are discussing their ongoing trading activities and decisions. The consequences of social communication on traders can not only include the decision to participate and to adapt their trading strategies, but also the decision to continue (survive) or to cease (quit) their trading activities.

However, the relationship between survivorship in trading and social communication is unexplored in the literature. The investigation of the survival of traders has a distinct value for understanding the dynamics of a trader’s lifetime decision-making processes, which is different from the decision to participate (at the beginning of a trading life) and to choose their trading strategies (in the middle of a trading life). It is the decision to quit trading (at the end of a trading life) which finally concludes the story of a trader’s trading life. This decision constitutes an important aspect of the characteristics of a trader’s trading life.

It is not fully clear to the academic community what traders talk about and how the various aspects of their trading activities are influenced by the content of the conversations they have while making their trading decisions (let alone examining the impact of social communication on traders’ behavior). However, in the setting we explore in this chapter (data from a STP) we can observe what traders talk about while trading and how their behavior is subsequently altered by such social communication. We observe that traders are keen to talk about the future in the ODF. For example, “Today is looking very sketchy, I’m going to hold a long aud/jpy averaged about 77.90 and call it a week,” “What do yu [sic] think the EURUSD pair is going to do in the next 5 hours?,” and “Maybe MyFXtrade will have a real-time graph of these numbers in the future we can use.”

Intuitively, these discussions anchor traders’ expectations regarding the future. Traders should therefore be more curious to check out their expectations in the future and more likely to stick around to see what happens, compared to instances where they do not have any expectations at all. Consequently, traders should have the incentives to continue to stay (survive) in the market (as opposed to exiting the market) after having such conversations regarding the future of the market. Therefore, we examine whether social communication impacts the survival of traders.

Such an investigation is especially relevant since, as we have argued, technological evolutions have led to integrating communication with real-time trading. This integration changes the way transactions are organized, in the sense that it becomes possible to obtain real-time information about how fellow traders make decisions, swap opinions, and interpret market information jointly. Evidence shows that communication on social media can predict prices in equity markets and FX market movements (Ozturk and Ciftci, Reference Ozturk and Ciftci2014; Reed, Reference Reed2016; Lachanski and Pav, Reference Lachanski and Pav2017). FX markets are of particular interest because entry barriers are usually lower compared with the stock market, attracting a broader spectrum of investors of different financial means. Crypto asset markets are another domain of interest here, but studies of crypto traders are still in an incipient stage. Recently, studies have developed theoretical models in order to describe information transmission in the market through network communication, capturing the implications for asset prices (Ozsoylev, Reference Ozsoylev2004; Han and Yang, Reference Han and Yang2013; Han, Hirshleifer, and Walden, Reference Han, Hirshleifer and Walden2022).

4 Communication and Survivorship: Possible Explanations

We encounter several converging explanatory approaches. One, coming from the sociology of finance, is that market participants (i.e., professional traders) use face-to-face communication or online messaging to coordinate with each other, build joint expectations based on what they observe while trading, and reciprocal obligations (e.g., Knorr Cetina and Bruegger, Reference Knorr Cetina and Bruegger2002; MacKenzie, Reference MacKenzie2009; Laube, Reference Laube2016). This explanation is grounded in studies of institutional trading floors and trading rooms, studies that do not examine massive online communication.

Another explanatory approach is that capitalist organizations generate fictional projections of the future as a means of coping with uncertainty (Beckert, Reference Beckert2016). However, such projections are generated at an organizational level, including various tools (e.g., business plans). It is unclear how they impact survival at organizational or individual level (if at all).

A third approach is provided by the anticipatory discourse theory which has been advanced in applied linguistics and psychology studies (Kinsbourne and Jordan, Reference Kinsbourne and Jordan2009; Streeck and Jordan, Reference Streeck and Jordan2009; Saint-Georges, Reference Saint-Georges and Chapelle2012; Poli, Reference Poli2019). Specifically, the anticipatory discourse theory suggests that “futurity is an inevitable component of text, talk, and more largely of social life, because human action has an intrinsically forward-looking nature” (Saint-Georges, Reference Saint-Georges and Chapelle2012). The “forward-looking nature” embedded in human communication takes two forms in the discourse processes, namely projection and anticipation (Kinsbourne and Jordan, Reference Kinsbourne and Jordan2009). Streeck and Jordan (Reference Streeck and Jordan2009) suggest that the forward-looking nature “consistently emerges in any discussion of interaction” (p. 93).

These insights reveal an important theoretical implication on the dynamics of human behavior subsequent to communication. That is “the very fabric of interaction and communication seems to be imbued with forward-looking anticipatory structures that facilitate ongoing, fluid interactions in a dynamic social environment” (Streeck and Jordan, Reference Streeck and Jordan2009, p. 95). Applied to the case discussed here, it would mean that communicational infrastructures present in markets embed such anticipatory affordances – they provide participants with opportunities to project the future repeatedly – and such anticipations ground actions in the market.

This theoretical implication is not exclusive to finance. We find that, in the context of STPs, these insights are evidenced by the discussion contents of the ODF. When reading through the content of the ODF, one significant feature is that traders are keen to talk about events in the future, share their predictions about the future, and discuss trading strategies based upon their perception of different states of the market in the future.

Given the discussed forward-looking nature of online discussions, we would expect that social communication increases the survival of traders. This is because traders, based upon the online discussions, may change their future expectations about the market or the platform, alter their perception of their own trading skills, and try out new trading strategies. These influences can be eventually translated into an increased survival probability of traders in the short term or a prolonged trading period in the long term. Therefore, we would expect that social communication increases the survivability of traders on a STP.

5 Sociability and the Wisdom of Crowds

This section aims to shed light on the effect of social media on the wisdom of crowds, and among different types of crowds, most of which are affected by communication. At least two strands of literature are directly relevant to the issue. The first one is the influence of social media on human behavior and the second one is the wisdom of crowds. As the literature shows that social media has broad influences on human behavior, we have sufficient grounds to expect that social media plays a (positive or negative) role in the decision-making process of individual investors. However, the wisdom-of-crowds literature focuses more on when and why crowds make better decisions. It remains unclear whether this wisdom can be influenced by social media and whether it is influenced differently according to different types of crowds. For instance, we can expect that social media accelerates crowd formation and/or polarization of opinions, and that there are differences in this respect between media and other communicational infrastructures.

6 Conclusion

We have examined here two interrelated issues: the integration of communication infrastructures into trading platforms and the impact of social media on trading behavior. This chapter has primarily focused on three aspects: the potential impact of communication on trading decisions and associated outcomes (i.e., financial performance); decisions to quit trading (i.e., survivorship); and the wisdom of crowds (i.e., group decisions). We discuss existing literature on each of these aspects and highlight potential areas for future research.

We have formulated two arguments: The first, theoretical, is that communication infrastructures, long seen as essential in finance, need to include social media. These play a key role not only in the realm of individual traders – which we have discussed here – but also in that of institutional traders. As we have mentioned in the opening, institutional data providers have integrated social messaging into their offerings, while, to the best of our knowledge, we have limited evidence on the impact of social media on the behavior of institutional traders. We know that social media data is intensely used in devising trading strategies, including algorithmic ones. Especially as communicational infrastructures evolve rapidly under the impact of AI and machine learning, it is imperative to examine closer both their evolution and their impact in finance and beyond.

The second argument we have made here concerns the impact of social media on trading behavior. Evidence points to the fact that social media usage increases imitative behavior and conformism (perhaps not surprisingly), but also that financial performance, at least in the realm of individual traders, is not positively impacted by social media usage (except for a tiny minority). This raises, among others, regulatory issues with regard to the integration of social media with trading platforms, even more so as these media incessantly evolve and as market infrastructures are regulated.

1 Introduction

Through digitalization the reach of formal finance can be expanded to previously underserved territories and populations, thereby enhancing the capacity of financial infrastructures to increase monetary flows. This transformation is observable in developing countries, where various actors collaborate to integrate informal economic activities into financial circuits, trying to connect them to the financialized capitalist system and adapt financial infrastructures. This chapter focuses on the role of philanthrocapitalist actors in this process, specifically examining the efforts of the Mastercard Foundation to advance the digitalization of financial infrastructures around the African agribusiness sector.

Our approach uses both critical development studies and science and technology studies (STS) to investigate infrastructuring processes. Grounded in the insights of Star and Ruhleder (Reference Star and Ruhleder1996), we perceive networking as a pivotal element of infrastructural change, encompassing interconnected social, organizational, and technical dimensions. We study the infrastructural transformation process by analysing the Foundation’s practices in agriculture and digital finance through discourse and programme analysis, as well as by examining its network of partnerships and fundings using social network analysis (SNA).

In keeping with Carse’s historical observation that infrastructure originally pertained to the organizational groundwork that preceded the construction of physical artefacts (Carse, Reference Carse, Harvey, Jensen and Morita2016), we posit that the Foundation contributes to financial infrastructural changes by assembling organizations, technologies, and capital through its networks, thereby creating platforms capable of (re)directing African resource flows into formal finance circuits. Previous research has shown that the Foundation helps to connect digital financial infrastructures to the financialized capitalist economy and that the firm Mastercard is frequently involved in those circuits (Langevin, Brunet-Bélanger, and Lefèvre, Reference Langevin, Brunet-Bélanger, Lefèvre, Chiapello, Engels and Gresse2023). This is not to suggest that these efforts to alter financial infrastructures are guaranteed to succeed; in fact, they encounter various challenges, gaps within existing infrastructures, and complexities in connecting them to peripheral economic elements such as rural finance. Nevertheless, we observe continued digitalization of agricultural finance in which the Foundation is actively involved. Our goal is to understand the project and shed light on the consequences of this process in terms of wealth circulation. We ask: How might the evolution of relationships reshape the credit landscape, impacting accessibility, costs, and organizational channels? We thus aim to contribute to social and policy debates about the power of (digital) financial infrastructures and their agency (Bernards and Campbell-Verduyn, Reference Bernards and Campbell-Verduyn2019; de Goede, Reference de Goede2021; Pinzur, Reference Pinzur2021; Campbell-Verduyn and Hütten, Reference Campbell-Verduyn and Hütten2023).

We begin in Section 2 with our conceptual framework, followed by our methodology. Section 3 delves into the Foundation’s approach and analyses its project related to the emerging digital financial infrastructure in the African agribusiness sector. We then explore the various networks, organizations, and actors involved, and finally, we conclude by reflecting on the connections between this evolving infrastructure and global power structures in the financialized capitalist economy.

2 Philanthrocapitalism and Digital Financial Infrastructures

In a conception akin to an agencement, as described by Callon (Reference Callon2021), Pinzur (this volume) emphasizes, like Bernards and Campbell-Verduyn (Reference Bernards and Campbell-Verduyn2019), that infrastructures do not inherently exist but instead take shape or materialize through labor that often remains unseen. We specifically argue that the contributions of organizations associated with philanthrocapitalism (or strategic philanthropy) to the emergence and transformation of market infrastructures are overlooked and remain largely imperceptible to analysts in critical development studies and STS. Nevertheless, entities like the Mastercard Foundation are significant actors in the realm of development and the capitalist economy. It is crucial to explore how the practices of these mega-foundations can potentially alter the circulation of wealth as they contribute to transformations such as the creation of new platforms, the digitalization of financial infrastructures, their recombination with other elements, and even the construction of new infrastructures.

Created by Mastercard International, a global payment technology firm that plays a major role in financialized capitalism, the Mastercard Foundation is one of the largest philanthropic institutions in the world in terms of capitalization. It is driven by its mission to advance education and financial inclusion as a catalyst for inclusive growth in developing countries. This raison d’être inherently places the Foundation within the process of financial infrastructures transformation and its digitalization. The critical literature on financial inclusion highlights the pivotal role of digital financial technologies in extending financialization at the margins (e.g., Langevin, Reference Langevin2019; Natile, Reference Natile2020; Bernards, Reference Bernards2022). These mechanisms are vital in establishing financial circuits that link peripheral agricultural markets to the dominant circuits of financialized capitalism. To make payment and credit transactions viable and profitable, operational digital financial infrastructures are essential.

By looking at the prevailing constellation of actors involved in the Foundation’s financial inclusion project and targeting marginal spaces in the global political economy through a neo-colonial analytical lens, we seek to identify the power relations being played out. To do so, we draw upon a broad concept of infrastructures conceived as a background operation and define the financial ‘infrastructuring process’ as an agency process of emerging new capabilities born at the confluence of innovative practical configurations to link, in this case, segments on the periphery of formal economic and financial circuits. This perspective is essential for incorporating infrastructural agency because, as de Goede (Reference de Goede2021, p. 353) contends regarding global payment infrastructures, their inherently political nature holds the potential to ‘reinscribe power relations and reroute money flows’.

The Mastercard Foundation’s existence is intricately linked to global power structures. As part of its 2006 IPO (initial public offering), the firm Mastercard International provided the Foundation with its capital from the firm’s own shares. With a substantial endowment of over $39 billion (Canada Revenue Agency, 2022), the operational capacities of the Foundation are thus partly built on the profitability of the firm Mastercard, a dominant financial capitalism corporation. What Mastercard International does in global capitalism is provide technological financial services to states, consumers, and enterprises to make their financial transactions fluid and secure. Simply put, the business case for the firm is that the higher the volume of transactions through their financial infrastructures, the more user fees are collected. Our previous work revealed that the Foundation’s practices reinforce Mastercard’s global organizational power by helping establish new market infrastructures, particularly in sub-Saharan Africa (Langevin, Brunet-Bélanger, and Lefèvre, Reference Langevin, Brunet-Bélanger, Lefèvre, Chiapello, Engels and Gresse2023). In this chapter, we narrow our focus to agribusiness, a strategically vital area for development institutions, for-profit entities, public entities, and philanthrocapitalist organizations. Globally, the Foundation ranks among the top ten private foundations that invest in agriculture (OECD, 2018, p. 61), with a significant portion of its investment portfolio directed to this sector (OECD, 2023).

The infrastructures discussed in this chapter go beyond being mere physical conduits; they involve farmers who are integrated into the global political economy, albeit through various pathways and socio-technical relations. On a relational level, we aim to differentiate what impacts the potential for inclusion and wealth capture in this infrastructuring process. We explore how new infrastructures emerge and interconnect with existing ones and consider the adaptable nature of the ‘installed base’ (Star, Reference Star1999, p. 382). Socio-technical relations may in fact evolve ‘often subtly – as alternative bundles of sociotechnical relations arise and interact with sociotechnical relations that already exist’ (Campbell-Verduyn and Hütten, Reference Campbell-Verduyn and Hütten2023, p. 461). We put forward the following question: What new relationships between actors and technologies alter the landscape and influence who can access credit, at what cost, and through which organizational channels?

3 Method, Empirical Materials, and Analytical Grid

In Chapter 1 of this volume, the editors Westermeier, Campbell-Verduyn, and Brandl ask contributing authors to adopt an ‘infrastructural gaze’ for their analyses. Although our primary focus is a major player in development finance, we employ a form of ‘infrastructural gazing on finance’ that delves into micro-level empirical data to explore how and where infrastructures undergo change. Our particular interest lies in understanding the actors involved in the Mastercard Foundation’s network and the impact of digitalization on core infrastructure functions like payment and credit in the African rural economy. We examine the Foundation’s role in Africa’s agricultural sector in a relational manner, encompassing relationships with technologies, public and private entities, institutions, investment flows, and other resources within the Foundation’s networks of fundings and partnerships.

Using a database of the Foundation’s discourse on its website from its inception until 2019, we extracted and analysed references related to agricultural finance and the Foundation’s programmes in the African agricultural sector. In addition to examining brief statements such as news items and blog posts, we conducted an in-depth analysis of key documents, particularly those involving the Foundation and its partners in agricultural finance, such as the Fund for Rural Prosperity (FRP) and the Rural and Agricultural Finance Learning Lab (RAFLL). These documents include action plans and assessments regarding agricultural finance and digitalization of the sector. Our objective was to understand the project undertaken by the Foundation in this domain, with a focus on the role of technology in constructing new digital infrastructures or transforming existing ones, along with the (re)integration of certain components.

We combined this method with SNA to examine how actors who play particular roles and mobilize capital are organized and brought into circulation by the Foundation. Our network dataset covers 2,206 fundings granted to 575 organizations from 2007 to 2021. We have also identified 68 partnerships spanning from 2008 to 2022 involving 166 partner organizations. SNA helps us understand the composition and functioning of networks associated with the Mastercard Foundation. It focuses on the structural patterns in relational data, where units (e.g., individuals and organizations) are interconnected within these networks, which is significant for both the unit and the network. This method aligns well with the concept of infrastructure as a relational entity and enables observation of how the Foundation’s practices stimulate network development in various sectors, like agriculture. It sheds light on the process of infrastructural transformation, including its (re)connection to the ‘installed base’. This approach helps us understand the Foundation’s relational capacity in digitalizing and formalizing informal sectors and reshaping markets, without assuming hierarchical control over the entire network or its components.

Our analytical framework, inspired by the ‘infrastructural gaze’, focuses on infrastructure as transformative socio-technical relations. In the discursive aspect, we scrutinize how the Mastercard Foundation defines and characterizes the infrastructuring project by observing their framing of current issues in the African financial sector for agriculture, their action plan, recommendations, the organizational constellation they aim to engage, the mobilizing and transformative roles digital technologies should play in the process, and implications for development circuits and flows. We also examine the diverse networks involved in this infrastructuring project, detailing the organizations participating in the Rural and Agricultural Finance (RAF) initiative and RAFLL, as well as those funded by the Mastercard Foundation, probing their functions and agency in this infrastructural process and potential relationships with power structures and macro-circuits of capital.

4 Agribusiness, Capital, and Technology: A Narrative of Necessity and Innovation

This section examines the Mastercard Foundation’s discourse on its website and in selected reports related to the RAF initiative and RAFLL. The goal is to capture the overall narrative regarding the transformation of financial infrastructure for agribusiness and understand the project promoted by the Foundation and selected partners. This process has been ongoing for about a decade, and the discursive materials, including programme details and sector diagnostics, provide insights into the Foundation’s intentions, practices, and, in some cases, results.

5 Network of Partnerships

The Foundation’s practices induce the development of new networks around the agricultural sector in Africa, overlapping public and private components. We now turn our attention to the relational changes at the infrastructural level by unpacking the components and connections in these networks. Thus, networking conceived as the core infrastructural change ‘becomes a primary analytic phenomenon’ (Star and Ruhleder, Reference Star and Ruhleder1996, p. 113). Figure 26.1² maps the partnership network surrounding the Foundation since its launch. The mapping exposes partnership clusters in specific sectors with different kinds of actors and organizations, such as in education, the public domain, private consultancy firms, think tanks, NGOs, and private corporations. We noticed a myriad of key actors in the field of financial inclusion, including the Bill & Melinda Gates Foundation, USAID (United States Agency for International Development), Women World Banking, Grameen, CGAP (Consultative Group to Assist the Poorest), GSMA (Global System for Mobile Communications Association) Foundation, SEEP (Small Enterprise Evaluation Project) Network, UNCDF (United Nations Capital Development Fund), and Accion. Notably, the Foundation’s initiatives connect influential players on the African continent, tied to colonial power structures, to foster the digital financial services market for the poor, evident in partnerships with Equity Bank, Equity Group Foundation, and the Kenya Commercial Bank (KCB) (details in Langevin et al., in press).

Figure 26.1 Mastercard Foundation’s Partnership Network, 2007–2021. For a higher-resolution, color version visit: www.cambridge.org/Westermeier

Source: Authors’ elaboration.

Figure 26.2 Legend. For a higher-resolution, color version visit: www.cambridge.org/Westermeier

Source: Authors’ elaboration.

Within this constellation of actors, we focus now more specifically on the RAF initiative (Figure 26.3³), which, as noted earlier, involved the Foundation supporting a portfolio of private organizations funded through the FRP. The project was carried out from 2015 to 2019 in collaboration with the RAFLL, an organization created by the Foundation dedicated to learning, best practices, dissemination and mobilization. This trio (RAF, FRP, and RAFLL) collaborates synergistically to disseminate practices across the financial and agricultural sectors, uniting private sector businesses and NGOs (Table 26.1). The budget for this partnership is approximately $148 million.

Figure 26.3 Partnership. For a higher-resolution, color version visit: www.cambridge.org/Westermeier

Source: Authors’ elaboration.

KPMG East Africa is the organization which receives the most funding from the Foundation within this partnership, receiving $74 million. The professional services firm manages the FRP. The FRP’s office and phone numbers are the same as KPMG’s offices in Kenya, thus we deduce that the funding is intended for the management of this programme. The Fund’s original aims included improving access to finance for over one million farmers, supporting the commercialization of innovative products and services, and fostering agricultural finance markets. In our analysis of the Foundation’s employment relationships network since its inception, which includes board members and senior and middle management, we identified five connections linking both entities, including current and former COOs of the Foundation.⁴ These connections are significant, as the Big 4 global professional service firms like KPMG ‘exercise power in the global economy by commanding transnational infrastructures of expertise that provide stability and order to globalization, and which form a critical resource that other actors – namely corporations and regulators – depend on to act’ (Christensen, Reference Christensen2022, p. 1). AgDevCo, a British not-for-profit investment firm specializing in African agribusiness, partnered with the Foundation to enhance productivity and market access for nearly 500,000 smallholder farmers. The KCB collaborated with the Foundation to improve financial service access for up to two million farmers, providing financial and non-financial services through the mobile solution, KCB MobiGrow, with a commitment to make $200 million in credit available (Mastercard Foundation and KCB Group, 2016).

It is worth noting that the private sector partners in this network hold influential positions in the financialized capitalist political economy. KPMG, a major accounting and consulting firm, is part of the elite Big 4, AgDevCo focuses on the African agricultural sector from the United Kingdom, and KCB, known for its connections to neo-colonialism, extends its influence beyond Kenya across the African continent (Bernards, Reference Bernards2022; Langevin et al., in press).

Among NGO partners, ICCO Cooperation (International Cocoa Cooperation) in the Netherlands received over $28 million to offer affordable agricultural financial products to 200,000+ farmers. Kenyan NGO AGRA, founded by the Bill & Melinda Gates Foundation and the Rockefeller Foundation, collaborated with the Foundation to expand financial services access to 700,000 smallholder farmers, emphasizing agribusiness as a primary financing method and supporting agricultural market information systems and payment innovations. One Acre Fund, a Kenyan NGO, provides East African farmers with asset-based financing and agricultural training. US-based NGO Kiva Microfunds, focusing on inclusive finance, established Kiva Labs to provide 700,000 clients access to new financial products.

Global Development Incubator, an American NGO, served as the learning partner for the RAF initiative, overseeing the RAFLL. It shared best practices, improved stakeholder understanding, and developed the monitoring, evaluation, and learning framework for the entire RAF portfolio. The role of this American NGO is an indicator of the significance of the organizational capacity-building and the continuous learning process within infrastructuring seen as a process. Organizational work is a vital aspect of an infrastructural perspective. Carse’s genealogy of the concept notes that, in the early twentieth century, infrastructure ‘referred primarily to the organizational work required before railroad tracks could be laid: either establishing a roadbed of substrate material (literally beneath the tracks) or other work functionally prior to laying tracks like building bridges, embankments, and tunnels’ (Carse, Reference Carse, Harvey, Jensen and Morita2016, p. 28). The Mastercard Foundation’s network-driven transformations empower various actors to document and make visible this learning, accessible through websites like those of the Foundation, RAFLL, and affiliated entities such as Dalberg and ISF, two development consulting firms.

Root Capital, a US-based NGO, engages in social investment in marginalized rural communities. Collaborating with the Foundation on the ‘Expanding the Frontier of Rural Agricultural Finance in West Africa’ initiative, with a budget of approximately $7 million, the partnership aimed to enhance the livelihoods of over 200,000 smallholder farmers. It sought to connect them to local markets, offering access to training and financial services. This collaboration reflects the process of integrating agricultural economies into formal marketplace structures, aligning with Brooks’ (Reference Brooks2021, p. 1) concept of inserting new market subjects ‘into value chains and wider circuits of capital and data’. This aligns with Mann’s (Reference Mann2018, p. 28) observation that digital data in Africa is designed to ‘become a source of power in economic governance’.

From this analysis, we deduce that the majority of these actors are based in the Global North, where innovations are conceived before being implemented in Africa with the support of the Foundation’s capital, expertise, and programmes. These innovations are frequently crafted to integrate new consumers into digital finance platforms assembled in Africa, often through neo-colonial corporate telecommunication, digital, and data infrastructures, enrolling diverse populations previously excluded from formal financial relations under colonial regimes (Langley and Leyshon, Reference Langley and Leyshon2022, p. 401).

6 Conclusion

Our analysis, guided by the infrastructural gaze, has aimed to understand how the infrastructure-(re)building process is linked to existing power dynamics and how it integrates with established circuits, such as those of the Foundation and other key channels in financialized capitalism and the neoliberal development agenda. Discourse analysis reveals an underlying agenda to generate more wealth through inclusive growth, aspiring to establish a commercial agricultural sector fully supported by formal finance and digital technologies. Our SNA indicates that actors involved in (re)building infrastructures and integrating resource flows into channels, whether through activating digital and financial literacy or constructing platforms leveraging massive data of African farmers, all share the characteristic of being anchored in the formal economy and finance.

This study sheds light on the ongoing involvement of philanthrocapitalist organizations in the financial sector in Africa, providing insights into the way financial infrastructures in the developing world are interconnected with global power structures (Bernards and Campbell-Verduyn, Reference Bernards and Campbell-Verduyn2019). Echoing concerns expressed in critical infrastructure literature, we acknowledge how infrastructural technologies ‘inscribe specific ways of doing things [and] sediment historical power relations and core-periphery relations’ (de Goede, Reference de Goede2021, p. 354). This prompts us to question the implications for the future of agribusiness in Africa, considering that the transformed financial infrastructure, through the integration of the Foundation’s networks, assigns pivotal functions to neo-colonial financial institutions like KCB. The financial infrastructure underlying agribusiness in Africa today significantly mirrors the power dynamics of the past. The project that seeks to ‘bank the unbanked’ is being made possible by the transformations of financial and development infrastructures that redistribute financial access while being anchored in powerful existing systems that can capture wealth.

1 Introduction

In 2009, the Indian government began work on introducing a twelve-digit biometric identity card known as Aadhaar, which means “foundation” in Hindi. By 2024, almost all adults in India had an Aadhaar ID, which has facilitated direct identification and provision of government payments to beneficiaries. In 2016, around the time there were one billion Aadhaar IDs, a Unified Payments Interface (UPI) was overlaid on Aadhaar to facilitate payments among banks. The third layer of the financial infrastructure, “Data Empowerment and Protection Architecture,” is meant to secure and verify digital identities necessary for any transaction. The three-layered financial infrastructure has enabled almost seamless commercial transactions among consumers, vendors, and banks, from payments to a taxi driver to a utility provider. In 2023, identity verification moved increasingly toward cloud-based platforms such as the Aadhaar-backed DigiLocker, made possible through the Ministry of Electronics and Information Technology (MEITy). The three infrastructural layers of identity, payments, and verification have made “India Stack” possible: The set of application programming interfaces (API) that, in February 2024, generated 8.6 billion transactions monthly, worth $170 billion (Indiastack.org, 2024). India’s vibrant startup culture, which largely rests upon the ease of making payments, would not be possible without India Stack.

Artifacts have politics (Winner, Reference Winner1980). India Stack is no exception. The term stack implies a physical layering, in this case of APIs above the identity and payment layers. Understood in an infrastructural sense, the layering results from the actions of multiple agents, including the state, business, and society. This chapter discusses the political economy facets of the relationship between the Indian state and the unlikely rollout of a mega-scale financial architecture such as India Stack. How did a state historically inadequate at providing public goods at scale roll out a postindustrial project of mega-proportions in record time? What are the distributional outcomes and the social meanings that arise from such an undertaking?

The short answer lies in the three interrelated c’s of state legitimacy in India: calculation, coercion, and creativity. The interrelationships show that the current “materiality” of India Stack rests upon the historical continuations of – or (in some cases) departures from – collective understandings in society and business about the role of the state. The longer explanation focuses on the three c’s in more detail: (1) The financial architecture parallels the Indian state’s historical habits of authority or postcolonial calculations that favored large-scale projects; (2) the calculations now encompass the imagination and creativity of groups such as India’s skilled and entrepreneurial IT talent; and (3) state legitimacy in India purportedly mobilizes market provision for social inclusion (e.g., in the name of “financial inclusion”), but like artifacts, markets also represent politics. While a supposedly inclusive financial infrastructure is made available to over a billion people, a set of coercive and exclusionary adjacent policies have raised surveillance and human rights concerns regarding minority religions and civil society groups, and have led to an overall decline in India’s democracy.

The India Stack project, described later in this chapter, strengthens state authority through bureaucratic and technical creativity. State authority is also enabled through a set of historically contextual cultural and economic calculations – and coercions. The financial infrastructure reflects the prerogatives of state meaning-making or legitimacy through what anthropologist James Scott terms an “administrative ordering of nature and society” (Scott, Reference Scott1998, p. 4). The India Stack project also belies claims about the weakening of the postcolonial state that cannot make territories and people “legible” or provide them with material comfort (p. 2). This underestimates the idea that the state is weak, as well as the state calculations about demographics and coercion enabled through digital networks that rest upon the identity of citizens. It is often argued that postcolonial states such as India turn to authoritarian populism as, in their view, neoliberal economic policies fail to encourage social inclusion and employment (Chacko, Reference Chacko2018). This chapter shows state calculations about demographics and coercion are enabled alongside creative provision of goods and services through the India Stack architecture.

The administrative ordering of Indian society through the Aadhaar identity card thus sits alongside the state’s historic monopoly on violence (Tilly, Reference Tilly, Evans, Rueschemeyer and Skocpol1985). Arjun Appadurai notes that recognizing “calculative action as a central feature of twenty-first century economics” is important for understanding how financial markets (or capitalism in general) function (Appadurai, Reference Appadurai2012, p. 14). This calculative tradition can be traced to the classical political economies of Adam Smith, Karl Marx, or Max Weber. Understanding the social meanings and distributional impact of India Stack entails situating its utilitarian calculations within the political and cultural anthropology of everyday life in India. As Chapter 1 of this volume notes, an “infrastructural gaze helps piece together and pierce through the complexities of finance” but “big-picture concerns of power, authority, legitimacy” in this chapter are revealed through the three interrelated aspects of calculation, creativity, and coercion.

From the introduction of railroads and telegraph in colonial India to fintech in the postcolonial twenty-first century, the constitution of authority and governance in India has been based on technology.

2 Calculations

Infrastructures are material artifacts with social and political meaning (Biejker, Hughes, and Pinch, Reference Biejker, Hughes and Pinch2012; Bernards and Campbell-Verduyn, Reference Bernards and Campbell-Verduyn2019). The techno-imaginary found in policy is indicative of political and social possibilities for the present and the future (Jasanoff and Kim, Reference Jasanoff and Kim2013). Rodima-Taylor and Campbell-Verduyn (Reference Rodima-Taylor and Campbell-Verduyn2023, p. 17) note “how collectively held visions (imaginaries) materialize into existing sociotechnical relations (infrastructures).” Imaginaries and infrastructures, therefore, reveal calculations from states and providers about the intended beneficiaries and the collectively held meanings by the providers and the users of the infrastructure. In India’s case, the calculations of the Indian state and bureaucracy have been crucial to the techno-imaginary informing infrastructures, the latest iteration being India Stack.

A historic example in India is that of railways. More was spent on raiways than on any other forms of infrastructure in India during the colonial era. The railways safeguarded British governance and investments. In postcolonial India, railways became a key means of transport and an indispensable part of India’s indigenously focused import-substitution industrialization strategy. From 1924, during the colonial era, until 2016 the Indian railways budget was presented to parliament as separate from the union (or federal) budget because of railwayss’ importance to the country. It was not a coincidence that the railwayss’ and union budget were combined in 2017, just as India turned its attention to postindustrial infrastructures.

While railways and telegraphs were considered important to the British Empire’s governance of India, the telephone, which is now at the core of India Stack, was viewed as a luxury in postcolonial India and deemed unimportant for development (Singh, Reference Singh1999; see also Handel, this volume, for an analogous understanding of the telegraph). Instead, postcolonial imaginary emphasized the radio, and later television, for developmental communication. Programming was colored by secular and inclusive constitutional norms (regardless of the reality of Indian society). The field of development communication coevolved with such communication infrastructures in India and other parts of the developing world (Gudykunst and Mody, Reference Gudykunst and Mody2002). Radio and TV broadcasting companies, along with a limited number of telecommunication providers, were either entirely state-owned or were public goods.

Three types of calculations from the Indian state, starting in the 1980s, pushed India toward thinking about new types of information-driven infrastructures that would later form the basis for Aadhaar and India Stack. First, India had a bureaucratic and engineering skillset that would assist with these efforts. The calculations from the Indian state rely on “sarkar” or a vast bureaucracy and elected officials with huge influence over people’s daily lives. At the top of the hierarchy is a set of officials from the Indian Administrative Service (IAS) who constitute an elite meritocracy within the state. In 2023, there were about 5,000 IAS bureaucrats in the country, with around 180 IAS selected yearly from over a half a million people who take the competitive civil service exams every year. Nothing in postcolonial or current India moves forward without the consent or vision of IAS officials. Therefore, the rollout of India Stack involved several IAS officials who either aided (or thwarted) the Aadhaar and other efforts.¹ India has also developed important science and technology education institutions in the postcolonial era – the foremost being the Indian Institutes of Technology – that provide graduates with the skillsets to make an IT-led infrastructure feasible. Interestingly, nearly two-thirds of those selected from the highly competitive Indian Civil Services exams are from engineering backgrounds (Economic Times, 2023). Natural science and engineering fields historically made it easier for examinees to score high on the civil service exam and get selected.

Second, India began prioritizing telecommunication in its development strategies in the 1980s and information technologies starting in the 1990s, and moved away from the state provision of telecommunications. While the rollout of telecommunications until 2000 was slow, that of mobile telephony after 1999 through private firms grew exponentially. By the end of the first decade of cell phone provision there were over 600 million cell phone subscribers, and, by 2020, over a billion. Beginning in 1991, India also veered away from its Import Substitution Industrialization (ISI) strategy to allow private actors to provide goods and services that had traditionally been seen as public goods. The state had provided landline telephones, but the private sector was allowed into the mobile telephony market. One of the main failures of the ISI strategy was an overly regulative state known as the license-quota-permit-Raj. As India liberalized, there were corruption scandals, some of the biggest being around spectrum allocation for mobile provision. Nevertheless, parallel developments provided some regulatory oversight through the Telecommunication Regulatory Authority of India and Telecom Disputes Settlement and Appellate Tribunal, which together sought to create impartial rules and ensure interoperability among providers.

Third, the colonial and postcolonial Indian state, in contrast to most in the developing world, had and has a high capacity for collecting statistics and calculating demographics. In doing so, the state has employed “a cadastral lens” that simplifies and makes visible some characteristics of populations over others (Scott, Reference Scott1998). The ten-year census survey was a prime example. In 2002, the central government proposed a National Population Registry (NPR) that would serve as a precedent for and, later, a competitor to the Aadhaar program. There was a security dimension in maintaining a population register, as proposed by Prime Minister L. K. Advani, leader of the Hindu-aligned Bharatiya Janata Party (BJP) that currently governs India under Narendra Modi. A controversial Citizenship Amendment Act (CAA) in 2019 extended Indian citizenship to persecuted religious minority refugees from neighboring countries, except if they were Muslim. CAA set off fears and protests that even Indian Muslims would face discrimination in proving their citizenship.

The historical precedent in the three calculations presented is apparent in the dirigisme and the skillsets within the state, and was fostered through public institutions and the state’s imagination about its demographics. There were other connections, such as those reflecting the ISI strategy. The shift to information technologies featured Indigenous solutions. Narayana Murthy, CEO and founder of Infosys, one of India’s biggest IT firms and one of the four firms in the country with a market capitalization above $100 billion, mentioned the 500-line switch from government’s Center for Development of Telematics in the 1980s as a foundational vision that would assist India’s adoption of IT and spread rural telephony (Murthy and Murthy, Reference Murthy and Murthy2009, p. 114). The market liberalization strategy did not immediately dismantle state ownership. In fact, initially it endowed state officials with tremendous power to decide between corporate winners and losers, and it resulted in many well-known corruption scandals.

The introduction of Aadhaar in 2009 followed the three aforementioned calculations, and linked them with finance and biometrics. The Unique Identification Authority of India (UIDAI) was created as a government body to assist with Aadhaar efforts. Nandan Nilekani, one of the Infosys founders, was brought in from the private sector to lead UIDAI. Nilkeni’s appointment is representative of collaboration with the private sector and reflects the growing economic and political importance of information technologies. The first Aadhaar card was issued in September 2010 and, by 2024, there were 1.31 billion Aadhaar cards in the country. Nilekani relied on a team of industry and highly competent bureaucrats, including several IAS officers, to plan the Aadhaar strategy (Ramnath and Assisi, Reference Ramnath and Assisi2018). Aadhaar was initially conceived as a voluntary ID card to facilitate the provision of public goods, especially income transfers and government subsidies, which were enjoyed by nearly half of the Indian population. By one estimate, before Aadhaar was launched only 15% of government subsidies reached the intended beneficiaries because intermediaries would siphon off some of the funds (Raghavan, Jain, and Varma, Reference Raghavan, Jain and Varma2019). An Aadhaar registry would facilitate payments directly to the beneficiaries.

3 Creativity

Aadhaar was initially planned as a voluntary scheme, but it increasingly took on compulsory dimensions for Indian citizens and became a central and inescapable component of proof of identity in India. Aadhaar identity is now a prerequisite for verification for almost all private and public services. Aadhaar is also the foundation of the financial infrastructure in India that links banks with public and private providers of goods and services. Although conceived by the Congress-led central government in India, the current BJP-led government embraced it in 2014 when it came to power. Therefore, the India Stack project also reveals continuities rather than breaks in the Indian state’s historical steerage of its financial infrastructure.

The term creativity is employed in this section for the financial affordances or possibilities enabled through the India Stack platform described later in the chapter (Earl and Kimport, Reference Earl and Kimport2011). The creativity of India Stack lies in affordances that were unavailable earlier. This includes payments and receipts, access to public records, registrations, and – most importantly – the foundational layer for the startup culture in India that has benefited from the API architecture that rests upon Aadhaar and the UPI payments system.

Aadhaar from its inception was conceived as an infrastructure that would enable other platform or layers to be stacked upon it. Bernards and Campbell-Verduyn (Reference Bernards and Campbell-Verduyn2019) point out five characteristics of infrastructures, which include their facilitation, openness, and centrality to a variety of activities, their durability over time, and their relative obscurity operating in the background. The uniqueness lies in tracing the scale and the interconnections that allowed the platform to grow and take on several functions. Transfer payments employing the Aadhaar platform began to be made in 2014 through a central government program called PMJDY (Prime Minister’s Jan Dhan Yojana), which entailed opening bank accounts through the biometric identification inherent in Aadhaar. By 2018, there were 310 million bank accounts associated with this program (Ramnath and Assisi, Reference Ramnath and Assisi2018, p. 128) and in 2022 there were 460 million (Financial Times, 2022).

The next step came in 2016 with the launch of the UPI enabled through the National Payments Corporation of India, a nonprofit jointly owned by major banks, which made all banking transactions interoperable. The unique feature of UPI was that it enabled transactions across various payment systems such as Paytm (launched in 2010), Google Pay, and Amazon Pay. After its entry in 2017, Google Pay both boosted and accounted for over half the transactions through UPI. The government launched its own interface for interoperable transactions, called Bharat Interface for Money or BHIM, which enabled transactions among users; this was ostensibly also to deflect the critique that all major existing applications were commercial or foreign owned.

The integration of Aadhaar with UPI and BHIM fed into the creation of India Stack and associated digital security measures that have further boosted the fintech industry and associated startup culture in India. Almost all apps in India are part of India Stack and operate through the underlying UPI. Recent work showing the overlap of often-state-enabled and widely accessible infrastructures with ICT-driven platforms is useful here (Plantin et al., Reference Plantin, Lagoze, Edwards and Sandvig2018; Westermeier, Reference Westermeier2020): The infrastructural layer comprising Aadhaar and UPI overlaps with the set of open interfaces or the platform known as India Stack. India Stack fits the conceptualization of platform capitalism (Sell, Reference Sell2022).

India’s experience has attracted global attention. The digital ID and fintech interfaces being developed globally now regularly cite the Aadhaar and India Stack platforms, and Indian writers – politicians, business leaders, and journalists – point out at length that the experience has attracted attention from international organizations such as the World Bank and the International Monetary Fund, global platforms such as Microsoft and Google, and multiple national governments (Nilekani and Shah, Reference Nilekani and Shah2016; Ramnath and Assisi, Reference Ramnath and Assisi2018; Kant, Reference Kant2019). While the creativity of India Stack is often praised in the West and India, the underlying coercions are largely overlooked.

Section 4 describes India’s struggles with data laws but, in the meantime, the technical solutions being developed have led to the MEITy developing the electronic eKYC (Know Your Customer) and DigiLocker as cloud-based services for consent-based digital identity verification. Aadhaar founder Nandan Nilekani has described the platform as follows: “the Aadhar program has now conclusively proven that we need not look only to the Silicon Valley of the world for cutting-edge innovation in technology” (Nilekani and Shah, Reference Nilekani and Shah2016, p. 46).

4 Coercion

The legitimacy of any infrastructure rests upon the way that state, commerce, and society are co-constituted. Sections 2 and 3 have described the calculations and creativity of state and commerce. This section turns to society and its collective understandings and participation in India Stack. Economic sociology rejects the view that states and markets are in opposition to each other or that the increasing role of one decreases that of the other (Block and Evans, Reference Block, Evans, Smelser and Swedberg2005). They are, in fact, co-constituted, and market rules are often embedded in social understandings of the roles economic actors perform (Granovetter, Reference Granovetter1985). A social understanding that accepts the authority of state and business would be one where the state, through its “cadastral lens,” (p. 47) produces a “prostrate civil society” (Scott, Reference Scott1998, p. 5). In a hierarchical postcolonial society such as India, the temptation to produce a prostrate civil society with authority and coercion, the basis of state legitimacy, is always present. Meritocratic bureaucrats used to enforcing top-down rules or technocrats designing engineering utopias may have scant regard for involving civil society, or fail to understand how social norms and civil society function in a democratic polity. Unfortunately, state coercion has increasingly replaced active and explicit societal consent as the financial infrastructure has evolved in India.

There are two ways to examine the role of society in India Stack. One looks at the direct involvement of societal actors in deliberating and formulating the rules that went into the formation of India Stack. However, infrastructures also reflect their political environments and societal hierarchies (Winner, Reference Winner1978). A polity’s overall vision and rules informing societal engagement with the state are, therefore, equally important. Both are discussed in what follows.

The Congress Party-led government, in power until 2014, had a difficult time passing legislation that would make Aadhaar a legal instrument. At issue was its status as a voluntary instrument when introduced, to it increasingly becoming a compulsory undertaking. Eventually the legislation was passed as a “money bill” in the Indian Parliament in 2016. Money bills do not require the consent of the upper house, and so are easier to pass. A civil society challenge to the money bill did not succeed in the Indian Supreme Court. The plaintiff in both Indian Supreme Court cases was a retired judge from the Karnataka High Court and the cases are known as the Puttuaswamy cases after him.

The societal contestation over Aadhaar and India Stack at first glance might point at anything other than a prostate civil society. For example, a successful civil society challenge from the Internet Freedom Foundation in 2016 defeated moves that would have compromised net neutrality. Another challenge resulted from Aadhaar’s data collection and breaches, and fears regarding the surveillance capacity of the state or other actors with access to Aadhaar data. One of the most fundamental revisions to the legal framework came from a 2017 Supreme Court case that enshrined privacy as a fundamental right in India and enjoined the central government to come up with a data privacy framework. In August 2023, the Indian Parliament passed the Digital Personal Data Protection Act (DPDPA). It was critiqued for providing enormous powers to the central government and introducing the language of “deemed consent” or implicit consent of individuals rather than explicit consent (Sabharwal, Reference Sabharwal2023). In early 2024, the Editors Guild of India, a national journalists’ organization, filed a petition with the government, noting that “deemed consent” could also be used against journalists reporting any story about individuals.

The preceding narrative points to legal recourse for Indian civil society and journalists, but this is in the context of legal charges being filed against journalists by the BJP-led central and state governments for reporting unfavorably about Aadhaar and of widespread attacks on civil society organizations in India. The rollouts of Aadhaar and India Stack have been top-down processes with hardly any direct consultation with Indian society. Therefore, most accounts of civil society engagement are either in the form of challenges in courts to the infrastructure or to the perceived legitimacy of these state maneuvers. An analysis from Mahrenbach and Pfeffer (Reference Mahrenbach and Pfeffer2023) showed the declining legitimacy of Aadhaar in the first 10 years of its operation from an analysis of nearly 250,000 tweets.

This analysis tracks a period when the Aadhaar efforts were new and during the time when the Supreme Court cases took place. Critiques of Aadhar are hard to find on Twitter now, especially from media and civil society leaders, and press freedom has declined in India. The restrictions on civil society groups and nongovernmental organizations continue to increase.

Most broad indicators about democracy or democratic deliberation in India present a bleak picture. India’s rank in the World Press Freedom Index at its inception was 80 in 2002, declining to 122 in 2010, and in May 2023 stood at 161 out of 180 countries. In 2000, the V-Dem scores for Indian democracy were as follows: 0.66 for deliberative democracy, 0.97 for electoral democracy, 0.58 for liberal democracy, and 0.48 for participatory democracy (see Figure 27.1). In 2022, these score had declined substantially from being at par with prosperous liberal democracies to among the lowest worldwide: Deliberative democracy was 0.29, electoral democracy 0.4, liberal democracy 0.31, and participatory democracy 0.26 (Coppedge et al., Reference Coppedge, Gerring, Knutsen, Lindberg, Teorell, Altman, Bernhard and Cornell2023).

Figure 27.1 V-Dem scores for India, 1972–2022.

Source: Author’s elaboration based on data in Coppedge et al. (Reference Coppedge, Gerring, Knutsen, Lindberg, Teorell, Altman, Bernhard and Cornell2023).

Civil society involvement has been marginal in the evolution of Aadhaar and India Stack. The Indian government can claim electoral mandate and global attention as constitutive of societal consent. As shown, India continues to decline on almost all measures of democracy. Society is envisioned mostly as a recipient or a consumer in India Stack calculations rather than as a deliberative entity – questioning, contesting, or engaging with Aadhaar and India Stack policies. With the decline of deliberation, Scott (1998, p. 4) is right in calling attention to a “prostrate civil society” whose primary interaction with the infrastructure is to partake of its core functions that have been internalized as collective understandings.

5 Conclusion

The three c’s discussed in this chapter each reveal alternative facets of India Stack across different realms of state authority in India. The state’s infrastructural gaze has been central to the endeavor of the fintech infrastructure that offers both continuities and departures from the way the Indian state has functioned historically. With the liberalization of Indian markets in the 1990s, the envisaged pluralistic political-economy realm in liberal thought, in which firms would check state power, has instead evolved into the coproduction of state legitimacy and coercion through an identity-based infrastructure that the state controls. Large-scale projects often consolidate state control and coercion (Acemoglu and Robinson, Reference Acemoglu and Robinson2012).

From the railroads to artificial intelligence (AI), large infrastructures are part of the legacy of state techno-imaginaries in India. The state authority sector in India is understaffed: There are only 3.2 million federal government employees for a population of 1.4 billion people (Sinha, Reference Sinha2023). State authority is better understood as unrivaled and far-reaching, especially through the actions of its competitively selected bureaucracy – the IAS – two-thirds of which is now drawn from engineering schools. Nevertheless, these vast infrastructural calculations have included Indigenous capacity-building, whether the locomotives or engines for the railroads, or the rise of Indian startups and unicorns for the AI infrastructure.

The Indian fintech infrastructure, enabled in its everyday usage through the India Stack platform, is unique and is already looked up to as a development model by several developed and developing countries and international organizations. In the mid-2010s, India seemed to be behind in fintech and in mobile money infrastructures (Singh and Flyverbom, Reference Singh and Flyverbom2016), and it seemed that the extant banking sector and government regulations had thwarted mobile money efforts. Instead of allowing mobile payments through SIM cards, as the pioneering Kenyan government did for M-Pesa, the Indian government demanded that bank accounts be linked to these payments. In hindsight, the rollout of Aadhaar and the UPI, and the facilitation of government payments and subsidies, have led to an exponential increase in bank accounts and, subsequently, mobile money in India. The identity and payment layers have also now led to India Stack being central to India’s vibrant startup sector with regard to information technologies and AI. Despite the continuities, the departures also demonstrate a creative state. The twenty-first-century Indian state works closely with business; the twentieth-century postcolonial state kept it at arm’s length.

There has always been a beneficial payoff for the Indian state in large infrastructures. For the British Empire, technologies of telegraph and the railroads were those of colonial extraction and of maintenance of the empire. Postcolonial India arguably needed these same forms of carriage and communication to extend governance over a vast territory and people. The railroads budget had symbolic importance and railroads were part of the popular imagination of being Indian. The fintech architecture, endorsed by all political parties in power for the last fifteen years, extends and expands the control of the Indian state. The license-quota-permit-Raj is gone. Businesses now grow in a liberalized marketplace and the classic British or postcolonial Raj or governance is history. In its place is the identity-surveillance-silencing Raj of an increasingly autocratic Indian state. The fintech architecture both represents and replicates that control.

Do AI-driven fintech architectures or all platform architectures in AI drive toward autocratic control? At the core of AI is data that identifies people (in this case Aadhaar), which is activated through algorithms for service provision – be this social media or fintech applications. However, the indicators correlating AI infrastructures with autocracy point to a mixed picture (Singh et al., Reference Singh, Shehu, Dua and Wesson2025).

First, there are dissimilarities across AI infrastructures themselves. Not all states with fintech and AI offer a similar architecture with inadequate data controls. The European Union, with its 2016 General Data Protection Regulation and 2023 AI Act, is a strong exception, but even countries such as Kenya and Colombia have strict regulations on data. Second, the debate on platforms and their entanglement with the state needs further research and evidence before grand claims can be warranted. For every grand claim, there is contestation. India’s fintech architecture works within a controlling state, but also with continuities of history and enormous creativity. India’s precipitous decline in V-Dem and press freedom scores is also not unique. The 2023 V-Dem report on the state of democracy worldwide is ominous and points to a decline in democracy in the Global North and South (Papada et al., Reference Papada, Altman, Angiolillo, Gastaldi, Köhler, Lundstedt, Natsika and Nord2023). The democracy levels in 2022 worldwide were those of 1986, with the most precipitous declines in Asia-Pacific, followed by Latin America and Central Asia. V-Dem makes special note of the decline in India. The declines in democracy seem correlated with per capita incomes more than how deeply AI fintech infrastructures have permeated a country. Our own policy research also shows that there are varieties of AI infrastructures with varying values across states, and it would be inaccurate to correlate all AI platforms with autocracy (Singh et al., Reference Singh, Shehu, Dua and Wesson2025).

This chapter has provided evidence that large infrastructures in the past and fintech infrastructures in the present are both steered by, and contribute to, state authority and control in India. There is not enough evidence to state definitively that fintech infrastructure and platform economies always lead to democratic decline, but the negative evidence cannot be dismissed either, including the centralization and manipulation of large national or even global data sets to service a few firms or states. Similarly, the thesis that neoliberalism has created an underclass in India that must be controlled through authoritarianism needs questioning: The India Stack architecture recently, and neoliberalism since 1991, have created affordances and a secular rise in overall incomes. This chapter locates the rising authoritarianism instead in historical social factors and political calculations.

Acknowledgments

Thanks are due to the editors of this volume for detailed feedback. This research is supported by a $1.389 million grant from the Minerva Research Initiative (2022–2025) and a Fulbright-Nehru Professional Excellence Award September 2022–January 2023.

1 Introduction

Are you who you say you are? The infrastructures for financial inclusion exist primarily to answer this one question. Identification is a prerequisite for inclusion into a financial system that grants credit and processes transactions that are essential for day-to-day activities. But identification is complex; financial inclusion would certainly not be a global challenge if it could be done easily. The evidence for this is that, despite the growing power of finance and the agility of digital technology, over one billion individuals, mostly from the poorest households, remain disconnected from their own national financial systems (World Bank, 2021).

Advances in digital technology have made it easier for many individuals to access the financial system to make payments, to borrow, and to lend. The scale of these transactions has meant that many large financial institutions have gained immense influence and global power since the late 1990s. Nevertheless, financial access has been elusive for the most marginalized individuals. The Global Findex Database offers compelling statistics to highlight this point (World Bank, 2021). The four largest developing economies with unbanked individuals are: China (224 million), India (191 million), Pakistan (99 million), and Indonesia (97 million). In each of these countries, over half of the unbanked are women. And globally, about one-third of the unbanked are young people; between the ages of fifteen and twenty-four.

Financial inclusion has thus not been pervasive enough to close the gap in access to banking services between what the World Bank categorizes as high-income and low-income countries (Demirguc-Kunt, 2018). Data from the Global Findex Database shows that, on average, in high-income countries 90% of individuals have bank accounts, relative to under 30% in low-income countries (Demirguc-Kunt, 2018).

And even when and where financial access is available, its ostensible benefits have been challenged (see Duvendack et al., Reference Duvendack, Palmer-Jones, Copestake, Hooper, Loke and Rao2011). Critical scholars have repeatedly highlighted how new digital strategies and technologies, commonly known as fintech, have strengthened the role of global finance in poor countries. These perspectives build on earlier critiques of strategies such as microcredit and microfinance, which note how these practices are counterproductive and reproduce inequalities (e.g., Roy, Reference Roy2010; Soederberg, Reference Soederberg2014; Mader, Reference Mader2015).

A more recent set of studies explores the phenomenon of ‘digital financialization’, which shows how the monetization of data widens inequalities because the most marginalized groups are compelled to relinquish the most privacy to access the financial system (Jain and Gabor, Reference Jain and Gabor2020). These concerns have become more meaningful as fintech infrastructures in many countries have become foundational to governance – specifically, e-governance – through the provision of digital identity. As examples from India and Pakistan – of Aadhaar and NADRA (National Database and Registration Authority) respectively – illustrate, financial access has become one part of a narrative that also includes digital welfare. For these two countries, the melding of financial access and digital welfare entailed the absorption of policies from global development discourses, particularly to assemble collaborations between private and public organizations to expand into what has been described as the ‘informal economy’; an undocumented space which is regularly problematized by development practitioners for low tax contributions, weak productivity, limited access to finance, and social protections (see Alter Chen, Reference Alter Chen2005).

As I show later in this chapter, perhaps the most striking aspect about efforts to incorporate this informal economy into the financial mainstream has been an emphasis on payment systems. These have expanded as new financial infrastructures, seeking to advance financial inclusion, have driven a shift from traditional over-the-counter methods to transactions that are digital, mobile phone-based, and biometrically verified. These have resulted in a system designed not just for the poor and unbanked, but also the non-poor, who have access to the mainstream banking system. This reflects an emergent consensus driven by the specific features of the financial infrastructure, including actors, objects, and processes (see also Swartz, this volume).

My focus is on how new payments systems serve as infrastructures for financial inclusion on the condition that they are backed by digital identification technologies. I show how these infrastructures for financial inclusion are a product of global and local shifts in development strategies. I use examples from India and Pakistan to show how these infrastructures have their roots in the KYC or know-your-customer requirement that policymakers sought to address through biometric databases. Thus, a core argument of this chapter is that tools initially intended to enhance financial access eventually became foundational to the broader issue of social policy, which in developing countries is increasingly organized through the concept of a digital welfare state (see Alston, Reference Alston2019). As such, I observe how the repurposing of technology has consequences that are often unplanned, highlighting a tension between chance outcomes and the purported linearity of techno-determinism.

The remainder of the chapter is organized as follows: Section 2 is a discussion on the international context of KYC requirements, showing how digital financial infrastructures see these obligations as the core challenge for financial access. Section 3 offers the specific local contexts of biometric databases in India and Pakistan respectively. Section 4 examines how initiatives to use biometric data to overcome KYC requirements have tended to rely on partnerships between multilateral agencies, governments, private financial institutions, and philanthropic foundations. Section 5 notes how the infrastructures for financial inclusion have elicited state support as exemplified by the respective instances of active support from the central banks in India and Pakistan. Section 6 concludes.

2 KYC: An International Agenda

The financial inclusion movement has repeatedly highlighted how KYC requirements are an impediment to financial access because poor people often lack identification documents (for an overview see Jafri, Reference Jafri2023). This problem underpins the case for such individuals to be given a digital identity as part of various development strategies.

As such, one of the key promises of fintech has been that digital identification will enhance development interventions. Fintech firms are increasingly collaborating with financial institutions, development organizations, platform technology companies, and philanthropic foundations to enhance access to the financial system through digital identity databases. Not only has a business model of data mining and data monetizing proven immensely successful but, as Gabor and Brooks (Reference Gabor and Brooks2017) observe, development practice has been shaped by behaviouralism based on big-data analysis and accumulation.

Digital identification has thus become central to digital financial transactions. Most of these transactions are what is known as G2P or government-to-person payments. The digitization of these transactions is a key facet of e-governance and underpins what has been describe as a digital welfare state, characterized by the increasing uptake of digital data and technologies in welfare design, partnerships, administrative processes, and service provision (Alston, Reference Alston2019; Van Zoonen, Reference Van Zoonen2020). The World Bank is a vocal supporter of this approach, which is seen as a comprehensive strategy to improve financial inclusion, women’s economic empowerment, and government fiscal savings (World Bank, 2023). Aside from forming the infrastructure for governance and welfare in poor countries, digital financial transactions are a favoured strategy of the World Bank and its affiliated institutions (e.g., CGAP, 2023; ID4D, 2023) because they are cost-effective, growth friendly, business friendly, and inclusive.

There is also an important narrative around how digital transactions restrict corruption, terror finance, human and drug trafficking, and tax avoidance and evasion (see Campbell-Verduyn, Rodima-Taylor, and Hütten, Reference Campbell-Verduyn, Rodima-Taylor and Hütten2021). This has become persuasive since the turn of the century, immediately prior to which, in the 1980s, growing concerns about money laundering and the war on drugs – and eventually also nuclear proliferation – created the impetus for what became guidelines to address AML/CFT, that is, anti-money laundering and countering the financing of terrorism (Amicelle, Reference Amicelle2017)

These guidelines are centred on KYC processes to establish a customer’s identity and identify risk factors for fraud and other financial crimes. The origins of KYC are in the USA Patriot Act, officially known as the Uniting and Strengthening America by Providing Appropriate Tools Required to Intercept and Obstruct Terrorism Act 2001. This legislation was enacted by President George W. Bush on 26 October 2001, shortly after the attacks in New York on 11 September. A quick and forceful response to the terrorist events, this legislation had – and continues to have – a profound impact on financial institutions worldwide.

The main components of KYC emerge from the AML rules of the Patriot Act. These include the customer identification programme (CIP), customer due diligence (CDD), and enhanced due diligence (EDD). The Act mandated the implementation of CIPs for new customers in specific financial institutions. Additionally, it specified EDD requirements for correspondent banking and private banking customers, particularly those who were non-US individuals. These measures expanded on existing AML legislation in the United States and imposed new obligations regarding CDD procedures for private banking and correspondent bank accounts involving non-US individuals. The Patriot Act thus encompassed various special measures, promoted cooperative AML efforts, prohibited unlicensed money transmitters, and established significant extraterritorial powers. These requirements significantly impacted the landscape of AML and CFT, both domestically and internationally (see FinCen, 2023).

A related event is the October 2001 meeting of the G10 central banks and other supervisory authorities on terror financing, where it was noted that – to prevent abuse of the financial system – it was necessary to develop and implement effective KYC and CDD procedures (Bantekas, Reference Bantekas2003). These events also underlie the transformation of the organization known as FATF or the Financial Action Task Force on Money Laundering. Formed in 1989 by the G7 with the goal to combat money laundering from drugs crimes, until 2002 this was simply an intergovernmental body entrusted with the development and promotion of relevant domestic and international policies. However, at an extraordinary plenary held in late 2002, it broadened its mandate to encompass terrorist financing. More recently, FATF has formally acknowledged that financial exclusion is a money-laundering and terrorist-financing risk: this is detailed in its open ended mandate (FATF, 2013).

The FATF mandates that every nation implements effective CDD procedures, which involve authenticating the identity of individuals engaged in financial transactions. In its ‘Recommendation 10’ on CDD, the FATF provides a set of guidelines that encompass various obligations, such as ‘identifying the customer and validating their identity through trustworthy, unbiased source documents, data, or information’ (FATF, 2013).

Because FATF rules necessitate the use of considerable documentation and rigid procedures to verify the identity of clients, they limit financial access. World Bank data shows that tedious documentation requirements are the reason – reported by 17% of adults – for not having a formal bank account (see Allen et al., Reference Allen, Demirguc-Kunt, Klapper and Peria2012). To address this, banking regulators in some countries have sought to offer simplified accounts. These ‘no-frills’ accounts are suited to FATF guidelines because they limit transaction frequency and quantities; they thus allow the issuing bank to take a risk-based or ‘proportional’ approach to KYC, which assumes that the risk from low-value transactions is inconsistent with the need for onerous documentation (FATF, 2013).

FATF guidelines thus have a tyrannical influence on financial infrastructure. Infrastructural actors, particularly banks and financial institutions in the Global South, have little choice but to comply with this international regime. The geopolitical salience of this, especially Western-centric conceptualizations of identity documentation, is explained by de Goede and Westermeier (this volume).¹

3 The Global Project for Digital Identity: Features and Critiques

The demands of the FATF constrain financial inclusion and the response from global institutions has caused the infrastructural gaze of finance to rest on identity verification. As a result, the agenda for digital financial access is now a globalized one, and a product of initiatives guided by two respective institutions: Consultative Group to Assist the Poor (CGAP) and Identification for Development (ID4D).² Both these institutions are led by the World Bank, using a partnership model that involves philanthropic and development organizations. This model is based on a consensus drawn from three assumptions about the benefits of financial inclusion: (1) financial inclusion initiatives are a part of the commercial and retail banking system; (2) financial inclusion advances social policy, specifically through digital welfare; and (3) financial inclusion counters money laundering and terror finance. An overarching theme is that the lack of digital identification impedes not only financial inclusion but development more broadly.

The CGAP and the ID4D initiative have an overlapping history. CGAP was launched in the 1990s as ‘a multi-donor effort to broaden and deepen the success of the work done by pioneer institutions’ in microfinance (CGAP, 1998). Eventually, as the financial access agenda expanded, the CGAP established itself as the leading think tank for knowledge, particularly ‘best practice’ on financial inclusion; Roy (Reference Roy2010) and Mader (Reference Mader2015) offer critical commentaries on this process.

This success of the CGAP drove the creation of another World Bank project: the ID4D initiative, which frames digital identification technologies as having transformative potential for poor countries. This initiative acknowledges that individuals who lack birth registration and official forms of identification are typically the most vulnerable people in the poorest countries (ID4D, 2023). The World Bank’s access to global knowledge and expertise, financial instruments, and private sector networks are salient features of an approach which seeks to establish digital identification systems for the delivery of basic services to the poor.

At the core of this strategy is the ID4D Multi-Donor Trust Fund, which was established in 2016 and is supported by several organizations including the Bill & Melinda Gates Foundation, Omidyar Network, and the Australian Government (ID4D, 2020). Digital identification technology has thus gained a reputation as the leading edge of technology for development and builds on earlier narratives around the transformative potential of access and connectivity. These perspectives are fuelled by copious examples of development interventions based on digital technology.

4 Digital Identity Projects

Scholars of the Global South have expressed heavy scepticism about the intrusive nature of digital financial inclusion and the practice of alternative data capture to expand financial markets (Aitken, Reference Aitken2017; Gabor and Brooks, Reference Gabor and Brooks2017; Bernards, Reference Bernards2019). These concerns are amplified as these practices have become centred on digital data, including biometrics and locational data. By relying on the hardware and software of personal mobile phones, fintech is utilized through ‘platforms’. Platforms enable payments to be made electronically for various services and goods. As Roitman (this volume) shows, platforms have had great success in advancing alternative modes of banking: this is reflected in the success of mobile money in several – African and South Asian – countries, and in the profitability of many various online platforms, including for e-commerce, food delivery, taxis, and so on. As such, digital finance is now at the cutting edge of development interventions centred on technology.

Because of these shifts – in technology and in development strategy – digital finance has augmented the need for digital identification. But, digital finance can only be used by those who are financially included.

The notion that financial access is elusive for those who lack official identification documents is a recurrent theme in financial inclusion scholarship. For many years the fix for this was to offer alternative products to enhance financial access (see Collins et al., Reference Collins, Morduch, Rutherford and Ruthven2009). Financial access itself has undergone a series of conceptual shifts, as microcredit gave way to microfinance, and microfinance gave way to the more nebulous terminology of financial inclusion and inclusive finance, which includes digital finance. A detailed discussion of these transformations takes place within Natile’s (Reference Natile2020) study of mobile money in Kenya.

These shifts have drawn attention to strong overlaps between those who are financially excluded or unbanked and those who lack identification documents. The World Bank reports that 1.7 billion people are without financial access (World Bank, 2021). Furthermore, an estimated 1.5 billion persons globally have no form of identification; mostly in the Global South, and often migrants and refugees (ID4D, 2016). Identification projects, which have sought to issue documents to prove citizenship and entitlements to public goods and services, precede digital identity databases. But digital finance requires that identification data should be digital. As such, in contemporary development practice, identification documents are seen as complementary to digital finance, and projects to increase access to identification documents are attached to initiatives for financial access.

As a development tool, digital finance has two primary utilities: (1) as mobile money and (2) for G2P payments. Mobile money does not, in theory, require digital identification to operate, but there is increasing regulatory pressure to link mobile money with digital identification, as FATF recommendations have become embedded in domestic banking regulations.

The other use of digital finance in development strategies is for G2P payments. These include social transfers – including conditional cash transfers – as well as wage and pension payments. The advantages of digitizing G2P payments are covered in CGAP (2009): they include improvement in financial inclusion by connecting recipients to branchless banking channels, but also reductions in government costs by streamlining transactions, and decreases in leakages through theft, fraud, and corruption. Recent analyses by organizations such as the World Bank (2018) and the UN (2020) find that ‘the identity gap’ (Beduschi, Reference Beduschi2019, p. 2) sizeably impedes access to basic healthcare, education services, and social safety nets. To address this, development interventions led by international organizations have been actively assisting states in expanding digital identity. These efforts have been targeted at domestic as well as refugee populations with the support of organizations such as the World Bank, the Asian Development Bank, and the United Nations High Commissioner for Refugees (see Beduschi, Reference Beduschi2019).

5 Discussion: Financial Infrastructures, Inclusion, and Access

Explicit and active support from the respective central banks in India and Pakistan has been central to the rollout of affordable payment systems. Payments, in contrast to loans and deposits, have gained prominence relative to credit and other financial products since around 2010. This is partially an outcome of the Global Financial Crisis 2007–2009, after which it became clear that microfinance is no less crisis-prone than other sources of finance (see Di Bella, Reference Di Bella2011; Wagner and Winkler, Reference Wagner and Winkler2013). The relatively recent importance attached to payments and payments system has been a notable outcome of the discursive shift from microcredit to microfinance to financial inclusion (see Mader and Sabrow, Reference Mader and Sabrow2015). Financial inclusion encompasses an extensive array of services, particularly mobile money, remittances, and even government payments; these go beyond the traditional realms of microfinance, such as credit, savings, and insurance.

In taking a proactive approach to advancing financial inclusion, both central banks – the Reserve Bank of India and the State Bank of Pakistan – have benefited from high rates of cash usage and low rates of payment card usage. As D’Silva et al. (Reference D’Silva, Filková, Packer and Tiwari2019) note, existing frameworks could expand from the modernization of financial infrastructures. For the payment cards – debit and credit – that were in usage, the domestic banks used systems that were not interoperable; furthermore, users could not integrate their payments across systems. For example, sending money from an account at one bank to an account at another bank could take up to two calendar days, if sent after a cut-off time, and even longer over weekends or holidays.

Perhaps the most interesting feature of these new payment infrastructures is that the turn to digital biometrically verified transactions brings financial inclusion to the mainstream, rather than constraining it to specific, purposively designed microfinance institutions. This emphasis on payment systems is a departure from approaches centred on poverty finance (Bernards, Reference Bernards2022), or a bifurcated system defined by mainstream commercial banks based on a traditional bank intermediation model on the one hand, and inclusive finance based on a disintermediated, or shadow banking model, on the other (Jafri, Reference Jafri2023). Not only are the new payment systems designed to be used by consumers through mobile phones, they are also the same whether used through a commercial/retail bank or microfinance bank. In making this possible, these central banks – and others which promote digital payment systems – acknowledge that financial access can be improved not only for the poor but for more affluent users as well, especially by linking the two. There are concerns, however, about inequities arising from unevenness in digital literacy and also from gendered patterns of mobile phone ownership (GSMA, 2021).

6 Conclusion

In this chapter I have explored how an infrastructural gaze on financial inclusion has caused national payment and identification systems to be enlisted in a global project that foregrounds the role of data in development. This project has the support of global institutions and sets out to create financial infrastructures that respond to socio-economic and political realities.

The need for these financial infrastructures is motivated in part by an international security agenda with KYC requirements at its core. These obligations are the core challenge for financial access. In India and Pakistan, biometric databases have been incorporated into financial infrastructures to ease bank access while complying with rules to limit money laundering and terror finance. As such, initiatives to use biometric data to overcome KYC requirements have tended to rely on partnerships between multilateral agencies, governments, private financial institutions, and philanthropic foundations.

Through these partnerships the respective central banks of India and Pakistan have created new payment infrastructures to advance digital biometrically verified transactions. The success of these infrastructures is that they have pushed financial inclusion into the mainstream of banking rather than limiting it to specialized microfinance institutions. This represents a departure from poverty-finance approaches or a bifurcated system. As such, central banks recognize that improving financial access benefits not only the poor but also more affluent users, particularly when the two become infrastructurally connected. However, concerns remain regarding disparities in digital literacy and gendered patterns of mobile phone ownership, which can contribute to inequities in access and surveillance.

The macro and micro-level aspects of this infrastructural gaze illuminate wider questions of power given the way in which financial infrastructures are not only designed and deployed, but also imposed. On the macro-level, the role of the FATF and the governance of financial transactions cannot be understated. While there is a rich and growing literature on the detriments of this for many countries in the Global South, it has tended to focus on the geopolitical constraints that come from the FATF regime. An infrastructural gaze is useful for highlighting how the strategies to address these constraints – often framed in the language of development – have shaped the wider financial system. Similarly, there are micro-level implications too, particularly the tension between the public and private actors. As the examples in this chapter show, shifts in the wider financial system – whereby financial access becomes a product of public infrastructure – have been driven by the support of private and philanthropic institutions, which have endorsed a model in which user costs are either low or non-existent. How viable this is depends on how the role of financial infrastructures is perceived in the future.

1 The Digital Yuan: A History of Multiple Layers and Scales

In January 2016, during an international financial conference, officials of the People’s Bank of China (PBoC) publicly announced that the Bank was working on developing a digital version of the renminbi (RMB) (Diyi Caijing Ribao 第一财经日报, 2016).¹ As representatives of the Bank explained, work on the project had already begun in 2014, when the Bank set up its own digital currency task force in response to the multiplication of privately issued cryptocurrencies at the time. The task force was given the mission to examine the possible benefits accruing from the development of a digital currency backed by the Chinese state. Since this initial statement in 2016, the Chinese digital currency, even though still in its pilot phase at the end of 2023, has elicited a wealth of journalistic as well as scholarly commentary. Thus, in the specialized financial press, the digital yuan (e-CNY) has dominated discussions surrounding central bank-issued digital currencies (see Figure 29.1). Furthermore, several books have been published that are targeted toward a nonexpert readership and devote much space to discussions of the Chinese digital currency (Turrin, Reference Turrin2021; Aglietta, Bai, and Macaire, 2022; Chorzempa, Reference Chorzempa2022).

Figure 29.1 Financial Times articles mentioning CBDCs and the digital yuan, 2016–2023.

Source: Author’s elaboration based on data obtained from Europresse on December 1, 2023; frequencies include the Financial Times’ paper and online editions as well as blog entries. Search string: “digital yuan,” “dcep,” “dc/ep,” “digital RMB,” “digital Renminbi,” “Central Bank Digital Currency,” “Central Bank Digital Currencies,” “CBDC,” “CBDCs,” “Central Bank Cryptocurrency,” “Central Bank Cryptocurrencies.”

Arguably, this heightened degree of attention to the seemingly dull topic of payment systems reform (cf. Swartz and Westermeier, Reference Swartz and Westermeier2023) – of which the introduction of digital currencies is a part – stems from the weaponization of international financial infrastructures since the turn of the millennium. Thus, over the 2000s and 2010s, the payment infrastructures underpinning the global economy, which in less turbulent times operated quietly in the background, have come to the center of attention in policymaking and journalistic circles, due to their repeated instrumentalization for the enforcement and perpetuation of Western and especially American sanctions policies (de Goede, Reference de Goede2012; de Goede and Westermeier, Reference de Goede and Westermeier2022; Mallard and Sun, Reference Mallard and Sun2022; Nölke, Reference Nölke, Braun and Koddenbrock2022). Hence, payment systems reform has suddenly come to be perceived as an issue with clear geopolitical implications and thus as a topic which is relevant to a much broader public beyond financial industry professionals.

To some extent, the current geopolitical context has also affected the scholarly debate surrounding the digital yuan. Indeed, research in this field has at times strong strategic and conjectural undertones. This is apparent first of all in the fact that the current social scientific debate revolves crucially around the issue of the probable future consequences of the digitalization of the Chinese currency (Gruin, Reference Gruin2021; Huang and Mayer, Reference Huang and Mayer2022; Deng, Reference Deng2023; Peruffo, Cunha, and Haines, Reference Peruffo, Cunha and Haines2023). As a side effect, there is so far no single study about the currency’s historical origins, besides the People’s Bank’s own official account. Second, research about the digital yuan has tended to analyze the digital currency from the point of view of Western countries, questioning in particular the currency’s possible implications for the future of dollar hegemony (Zhang, Cui, and Campbell-Verduyn, Reference Zhang, Cui and Campbell-Verduyn2023). By contrast, the repercussions of the digital yuan within the People’s Republic of China (PRC) and its immediate geographical vicinity have received much less critical scrutiny. The present chapter aims to fill this gap in the literature by providing a somewhat longer history of currency digitalization in the PRC, of which the work on the digital yuan represents only the current endpoint. In this endeavor, this chapter draws on a variety of public statements by the People’s Bank as well as a selection of newspaper articles on financial reform in Chinese media.²

Following Westermeier, Campbell-Verduyn, and Brandl (this volume), we adopt the infrastructural gaze and bring the history of the digital yuan into a broader conversation with scholarship concerned with large technical systems in different societal and historical contexts. There are two benefits gained from using “infrastructure” as a sensitizing concept in the study of currency digitalization in the PRC. First of all, existing scholarship concerned with the history of large technical systems such as the Chinese digital payments infrastructure provides us with a set of assumptions concerning the mechanisms through which infrastructural change occurs. Specifically, this strand of scholarship suggests that new large technical systems are rarely ever built from scratch. Rather, new infrastructures often depend in crucial ways on existing technical systems. Thus, infrastructural change frequently operates through a mechanism which is sometimes called layering. Layering refers to the construction of new sociotechnical entities on top of already existing ones and to the reassembling of existing sociotechnical entities in novel ways. We will show that layering adequately describes several aspects of the development of a state-backed currency in the PRC.

The second advantage of exploring currency digitalization in the PRC through the infrastructural gaze is that it directs our attention toward the interplay between sociotechnical systems across multiple geographical scales (Bernards and Campbell-Verduyn, Reference Bernards and Campbell-Verduyn2019, pp. 779–780; Zhang, Cui, and Campbell-Verduyn, Reference Zhang, Cui and Campbell-Verduyn2023). Thus, at the level of the PRC, the digital yuan amounts to a network of several distinct sociotechnical entities, while at the level of the global economy the digital yuan is itself a node in an emerging network linking several payment spaces to one another. As we will show, this means that the objective socioeconomic function of the e-CNY depends, crucially, on the evolution of the structure of this emerging transnational network in which the Chinese digital payment infrastructure is embedded.

2 A History of Multiple Layers: The Digital Yuan and End of Finance as We Know It?

As a rule, large-scale technical systems evolve continuously. This change rarely ever occurs as a sudden wholesale replacement of one sociotechnical system by another (Bernards and Campbell-Verduyn, Reference Bernards and Campbell-Verduyn2019, p. 778). Rather, as infrastructure scholars Edwards, Jackson, Knobl, and Bowker have pointed out, more often than not new infrastructures are constructed on top of existing ones, so that “many infrastructures […] are themselves deeply embedded within and dependent on other infrastructures” (Jackson et al., Reference Jackson, Edwards, Bowker and Knobel2007). When infrastructural change occurs through such a cumulative process, with one type of technology being added upon another, infrastructure scholars refer to this process as layering (e.g., Reilley and Scheytt, Reference Reilley, Kornberger, Bowker, Elyachar, Mennicken, Miller, Nucho and Pollock2019). By focusing on the layered character of many sociotechnical systems, scholars can develop some critical distance with respect to assertions from social actors making overly bold claims regarding the “disruptive” socioeconomic impact of certain new technologies (Bernards and Campbell-Verduyn, Reference Bernards and Campbell-Verduyn2019, pp. 779–780). Indeed, while the notion disruption emphasizes a fundamental discontinuity between new technologies and past social practices and technological systems, the concept of layering frames technological change as a path-dependent, incremental and cumulative process (Star, Reference Star1999, pp. 381–382; Anand, Gupta, and Appel, Reference Anand, Appel, Gupta, Anand, Gupta and Appel2018, p. 12). Layering is an apt metaphor to describe how the history of currency digitalization in the PRC unfolded and, by using it, we can show how a policy proposal that had the inherent potential to completely overhaul the financial system as it exists gradually morphed into an additional layer on top of the current institutional arrangement (Larue, Fontan, and Sandberg, Reference Larue, Fontan and Sandberg2020, p. 120; Ortiz, Reference Ortiz, Fourcade and Fassin2021; Mozorov, Reference Mozorov2022).

Since the late 1990s, currency digitalization has been a topic of discussion off and on within the Chinese central bank as well as in the broader ecosystem of economic experts involved in the elaboration of financial policy in the PRC. On several occasions, the issue arose in reaction to the development of new products and services by private sector entities which facilitated payment in online retail transactions. Thus, calls for RMB digitalization often went hand in hand with statements underlining the necessity to enlarge the Bank’s purview in the quickly growing online economy, where it might otherwise be marginalized. Around the turn of the millennium, two economists of the People’s Bank, Xie Ping and Yong Lin, published a research paper raising the question of the potential costs and benefits of a publicly issued digital currency (Xie Ping 谢平 and Yin Long 尹龙, 2001). According to the two authors, the question of currency digitalization had become urgent in the wake of the steady increase in commercial transactions taking place on online platforms and due to the fact that the rise of this new internet economy had incentivized private companies to issue their own form of unregulated electronic money, in this case, prepaid cards. These two related changes could potentially weaken the efficiency of the monetary policy instruments in the Bank’s arsenal. Thus, the fundamental issue at stake in the discussion concerned the future role of the People’s Bank in an economic environment which was undergoing structural transformation. The two authors remained skeptical as to whether the central bank should attempt to provide a digital version of its currency for the sprawling internet economy or whether it should simply accept the gradual transformation of its role in a changing economic environment. Ultimately, they argued that an intervention might indeed hinder private sector innovation and would cause regulators to settle for a technologically mediocre solution.

In keeping with this skeptical line of argument, over the following years the People’s Bank maintained a relatively lenient stance toward private sector initiative in the realm of digital payments. This was evident with respect to both cryptocurrencies, and bitcoin in particular, and third-party payment providers, such as Alipay and WeChat Pay. Thus, during the early 2010s, thanks to a comparatively tolerant policy toward cryptocurrencies, China grew into the single most important global hub for bitcoin mining and trading. In December 2013, Chinese regulators made a first step toward a tightening of regulations by defining bitcoin as a digital commodity, as opposed to a currency, since it lacked legal tender status (PBoC, 2013). Nevertheless, instead of immediately shutting down bitcoin mining and trading activities on Chinese soil, authorities at first contented themselves with tightening regulations whenever the value of the cryptocurrency suddenly appreciated and the market seemed to overheat (Campbell-Verduyn, 2018, pp. 99–103). Similarly, when several technology companies, most notably Alibaba, developed their own payment services, Chinese regulators did not treat them as financial service providers at first, leaving the sector more or less unregulated (Liu, Reference Liu2021; Wang, Reference Wang2021; Chorzempa, Reference Chorzempa2022).

Toward the middle of the 2010s, against the backdrop of a general shift in regulation of digital finance and cryptocurrencies in the PRC (Wang, Reference Wang2021; Zhang, Cui, and Campbell-Verduyn, Reference Zhang, Cui and Campbell-Verduyn2023), certain reform-minded policy experts began to advocate for the development of a new nationwide state-operated payment infrastructure which could support the digital economy. Xie Ping, the coauthor of the 2001 research paper, provided an important impetus to the renewal of these discussions after having played a leading role in the reform of public shareholding (Wang, Reference Wang2015, p. 616) and having occupied a central position in the largest Chinese sovereign wealth fund. In the early 2010s, Xie coauthored a series of widely cited research articles about the digitalization of financial services in China.³ The steady digitalization of financial services had fundamental implications for the entire economy, leading to the emergence of a new regime of financial intermediation for which they coined the term “internet finance” (互联网金融) (Xie, Zou, and Liu, Reference Xie, Zou and Liu2015). Foreign observers rendered the unwieldy “internet finance” simply as “fintech” (Schueffel, Reference Schueffel2016). For Xie and his coauthors, however, “internet finance” meant something far more consequential than the mere application of new technological solutions to old problems. Rather, its emergence would amount to nothing less than a complete overhaul of the existing financial system and the consolidation of a new regime of financial intermediation which would ultimately replace the financial industry in its entirety.

Xie and his coauthors framed the rise of internet finance as a process of financial “democratization” and “disintermediation,” since the then-current “elitist” financial system conferred undue power to a small number of professionals who made far-reaching choices on behalf of their customers. In the decades to come, this industry would gradually be replaced by a single state-controlled online platform on which individual citizens could make financial decisions on their own. In their view, all financial transactions, retail as well as wholesale payments and securities transfers, would be realized via this state-operated system without the interference of any further intermediary. Indeed, as they explained, in the internet finance system “individuals and organizations (moral persons) alike will have accounts at the central bank’s payment center.” In this institutional configuration, the “system of bank accounts at commercial banking institutions would cease to exist” (谢平 Xie Ping and 邹传伟 Zou Chuanwei, 2012, p. 13).⁴

The term “internet finance,” as well as the idea that it was necessary to construct new public payment channels in the age of the internet economy, quickly caught on. Thus, between 2014 and 2015, then-Premier Li Keqiang repeatedly endorsed the “healthy development of internet finance” as a priority for public policy (Gruin and Knaack, Reference Gruin and Knaack2020, p. 380). However, the underlying societal project to do away altogether with the private financial sector did not gain much traction among political elites. Indeed, when the People’s Bank made its first public announcement about the digital currency in 2016, in its broad outline, the project seemed like a realization of Xie’s original proposal: It was a digital retail payment system operating on the ledger of the central bank. However, in contradistinction to Xie’s favored scenario, representatives of the People’s Bank took great care to explain that the new system would not threaten the private banking industry. Thus, in one of the earliest statements about the concrete technical design of the new digital currency in 2016, the Bank’s Vice-Governor Fan Yifei underlined that the digital yuan could in theory be developed single-handedly by the central bank, but that the Bank preferred nevertheless to cooperate with existing financial institutions since this would make implementation easier (范一飞 Fan Yifei, 2016). The first technical blueprint of the payment system was published in 2018 (姚前 Yao Qian, 2018) and confirmed this model in which private corporations and the central bank would work hand in hand, a configuration which came to be known as the two-tier model (二元体系). In this model, private sector entities would function as intermediaries between the central bank and households, purchasing the digital currency from the central bank and then distributing it, not unlike cash.

Not only did the two-tier model not undermine the fundamental role of the private financial industry in the distribution of money, it also did not threaten the economic power of the country’s two corporate giants in the field of digital payments, Alipay and WeChat Pay. Regarding the latter two, Xie Ping had once advocated a more radical proposal, arguing that the development of a public infrastructure for digital payments should be used as an opportunity to break the duopoly in the field of digital payments (谢平 Xie Ping, 2020). However, as it turned out, both companies quickly came to be associated with the digital yuan project. Thus, both Alipay and WeChat Pay developed functions to allow users of their applications to make payments via the digital yuan platform operated by the People’s Bank. Furthermore, both companies contribute to the technological development of the new infrastructure, with Alipay said to contribute its integrated development environment as well as its distributed database system, OceanBase (Liu, Reference Liu2021). Hence, whereas some of the early rhetoric surrounding the digital yuan framed currency digitalization as a means to build a new financial mechanism at the heart of the Chinese political economy, the project actually transformed into a supplementary layer in the existing “ecology of money infrastructures” (cf. Rella, Reference Rella2020).

3 The Varying Scope of the Digital Yuan and the Transnational “Race” to Develop a CBDC

In the history of the modern nation-state, infrastructures have been instrumental in logistically unifying territory (Mann, Reference Mann1984; Maier, Reference Maier2012; van Laak, Reference Van Laak2018). Yet, as Nick Bernards and Malcolm Campbell-Verduyn (2019) have underlined, the geography of infrastructures rarely overlaps with the spatial patterning of political authority. Infrastructures rarely end at the territorial boundaries of the nation-state because part of their function is to create linkages allowing for all kinds of circulation across political frontiers. Thus, even seemingly “national” infrastructures are frequently embedded in a transnational web of technical systems. The characteristics of this international network, which serves as the broader context of the infrastructure, determines the utility, function, and reach of the latter in crucial ways. The same holds true for the digital yuan, whose potential sociopolitical impact has been fundamentally altered by the rise of a transnational central bank digital currency (CBDC) movement in the latter half of the 2010s.

Following the People’s Bank’s first public statement about RMB digitalization in 2016, the project received a great deal of attention in the international central banking community and beyond. Among foreign observers, one of the key questions concerning the new Chinese payment infrastructure was how the digital yuan would affect the global balance of geopolitical and geoeconomic power. Especially in the United States, experts and policymakers were quick to interpret the People’s Bank’s work on the digital currency as a possible threat to dollar hegemony and to frame currency digitalization as a race between the major global economies, which China was then leading due to its early engagement with digital payments (Chorzempa, Reference Chorzempa2021; Himes et al., Reference Himes, Lynch, Dean, Ocasio-Cortez, Auchincloss, Barr, Sessions, Williams, Hill, Zeldin, Davidson, Gonzalez, Gottheimer, Torres, Foster, Emmer, Waters and McHerny2021). Thus, work on currency digitalization in the PRC came gradually to be perceived as a sign of the country’s economic superiority, rather than as a solution to a series of idiosyncratic problems which were specific to the Chinese payments sector.

An important aspect of this process was the rebranding of the digital yuan as a “central bank digital currency” or “CBDC.” In the second half of the 2010s,⁵ and especially in the context of the COVID-19 pandemic, the issue of CBDCs rose to prominence in the world of central banking and financial policy (Ortiz, Reference Ortiz, Fourcade and Fassin2021; Kuehnlenz, Orsi, and Kaltenbrunner, Reference Kuehnlenz, Orsi and Kaltenbrunner2023; Swartz and Westermeier, Reference Swartz and Westermeier2023). Several international organizations, such as the Bank for International Settlements (BIS) and the International Monetary Fund (IMF), actively contributed to the promotion of the topic by subsuming a variety of dispersed initiatives undertaken by central banks across the globe under a common label, categorizing them first as “central bank cryptocurrencies” (Bech and Garratt, Reference Bech and Garratt2017) and then later as “central bank digital currencies” (BIS Committee on Payments and Market Infrastructures, 2018). The BIS in particular conducted surveys amongst central bankers about their stance toward CBDCs (Barontini and Holden, Reference Barontini and Holden2019; Boar, Holden, and Wadsworth, Reference Boar, Holden and Wadsworth2020; Auer et al., Reference Auer, Boar, Cornelli, Frost, Holden and Wehrli2021; Kosse and Mattei, Reference Kosse and Mattei2022, Reference Kosse and Mattei2023), classified existing projects to show commonalities and differences (Auer et al., Reference Auer, Boar, Cornelli, Frost, Holden and Wehrli2021), drafted scenarios for future implementation (BIS, 2021), and even actively participated in certain experimental projects via its Innovation Hub (BIS Innovation Hub, 2021). Think tanks such as the American Atlantic Council as well as journalistic outlets dedicated to blockchain and cryptocurrencies helped to draw attention to the issue of payment systems reform amongst a greater variety of actors beyond the limited community of central bankers.

While in the late 2010s, it seemed obvious to many commentators that China was leading in the race to develop a CBDC, being “the first major economy” to do so (Kumar and Rosenbach, Reference Kumar and Rosenbach2020), there was no clear-cut consensus as to what the Chinese CBDC precisely was. In a widely remarked article first published in 2019, IMF economists Tobias Adrian and Tommaso Mancini-Griffoli (Adrian and Mancini-Griffoli, Reference Adrian and Mancini-Griffoli2019, Reference Adrian and Mancini-Griffoli2021) briefly discussed the Chinese model for developing a CBDC. For the authors, the Chinese model did not, however, refer to the retail payment system which the People’s Bank had been working on since 2014. Instead, the authors focused on several regulatory changes concerning the third-party payment sector in the PRC. Their brief discussion of the Chinese model of currency digitalization described indeed what both authors termed a “synthetic CBDC.” Such a synthetic CBDC had come into being in the PRC after two related regulatory changes affecting the business model of the two main companies in the Chinese digital payment sector, Alipay and WeChat Pay.

First of all, in 2017, the central bank set up a clearinghouse, NetsUnion, which was in charge of handling all transactions made via third-party payment applications (Xing, Hei, and Pu, Reference Xing, Hei and Pu2018, p. 1221; Ba, Reference Ba2022, pp. 211–220). The reform was intended to increase the monitoring capacity of the central bank over payments realized via the firm’s applications. Before 2017, third-party payment providers such as the two technology giants had usually opened multiple accounts within the existing banking network, between which they moved funds following the transactions made by users. For officials at the People’s Bank, this meant that payments made with the applications appeared only as bulk transfers between two different bank accounts belonging to the same payment corporation. When the People’s Bank asked the companies for access to more detailed data, in some documented cases they refused (Yang and Liu, Reference Yang and Liu2019; Yu, Reference Yu2021, Reference Yu2022). After the establishment of the new clearinghouse, individual transactions made via NetsUnion applications became visible to the central bank.

In June 2018, the legal framework for the two giant payment operators changed in another decisive way, as the People’s Bank obliged both Alipay and WeChat Pay to deposit 100% of their customers’ funds in their custody into accounts at the central bank, instead of reinvesting them (Carstens, 2018, p. 9). In the PRC, neither of those measures had initially been conducted under the banner of the “digital yuan.” Yet, in conjunction, these two legal changes had led to a situation in which payments made via WeChat Pay and Alipay came to bear characteristics similar to those described in blueprints for CBDCs: Retail payments with the two applications could be monitored by the central bank and they were realized with liabilities entirely backed by reserves. This institutional setup was the “Chinese model” of constructing a CBDC, according to the IMF’s researchers (Adrian and Mancini-Griffoli, Reference Adrian and Mancini-Griffoli2019, Reference Adrian and Mancini-Griffoli2021). The “Chinese CBDC,” thus defined, neither necessitated the use of any new technologies in the realm of payments, nor commanded any considerable amount of public investment in the construction of a novel infrastructure. Rather, for the creation of a synthetic CBDC in the PRC, it had been sufficient to impose tighter regulations on formally private entities already operating in the field of digital payments.

Within the Chinese politico-administrative elite, the label “CBDC” was not uncontested. While certain Chinese governmental agencies willingly took up the idea that China was leading in the race to develop a CBDC (State Council of the PRC, 2019), in 2020, former Governor of the People’s Bank, Zhou Xiaochuan, explicitly refuted the notion that the digital yuan was a CBDC (Zhou, Reference Zhou2020). Zhou argued that the key difference between the digital yuan and an ideal-typical CBDC lay in their respective liability structures: Whereas CBDCs were ultimately the liability of the central bank, the digital yuan would be a liability of the private financial sector backed by some amount of central bank reserves. This view was ultimately contradicted upon the June 2021 publication of a white paper on the digital yuan in which the People’s Bank stated that the digital yuan was actually its own liability and that the private sector would merely be charged with distributing it. Furthermore, the white paper explicitly labeled the digital yuan as a “retail CBDC” (Working Group on E-CNY Research and Development of the People’s Bank of China, 2021, p. 4). The digital yuan had thus become a CBDC and the People’s Bank the leading institution in the transnational race toward monetary modernity.

The rise of a transnational CBDC movement had thus led to an accumulation of symbolic capital by the People’s Bank. The second, and arguably more important, consequence of the multiplication of CBDC initiatives across the globe was the fact that it created a new window of opportunity for Chinese policymakers to reconsider the purpose the Chinese digital currency might serve. Most notably, it opened up new possibilities for the digital yuan to contribute to the broader restructuring of international financial infrastructures and thereby to strengthen the RMB’s role in the global economy. From early on, officials at the People’s Bank had repeatedly refused the notion that the digital yuan was intended to alter the function of the currency in the international monetary architecture. Thus, Mu Changchun, head of the central bank’s digital currency institute stated that there was no direct causal link between the international use of the Chinese currency and its material form (Mu, Reference Mu2020). According to Mu, international use depended first and foremost on economic fundamentals and then other structural variables of the economic system, a position which has found multiple echoes in the scholarly community (Prasad, Reference Prasad2021; Eichengreen, Reference Eichengreen2022; Peruffo, Cunha, and Haines, Reference Peruffo, Cunha and Haines2023). Similarly, former People’s Bank Governor Zhou Xiaochuan reiterated this point, stating during a public conference that the initial intention of the project had never been to influence international wholesale payments (Xu Wei, 2022). Rather, the digital yuan had been conceived with “shopping in mind” and the cross-border use of the digital yuan would probably remain limited to small retail transactions.

However, against the background of payment system reform in other countries, the digital yuan could become a tool to experiment with new ways of linking the Chinese payment system to systems operating in other jurisdictions. In early 2021, the People’s Bank publicized its participation in a joint collaborative “multiple CBDC” project conducted in partnership with the Bank of Thailand, the Bank of the United Arab Emirates, and the Monetary Authority of Hong Kong: Project mBridge. The explicit aim of the collaboration was to establish a shared platform for interbank settlements in all four jurisdictions based on a commonly operated blockchain. In the first joint report about the advancement of the project, Mu Changchun argued that mBridge was yet another threshold in the brief history of the digital yuan. In the same context, he stated once more that the currency was first and foremost a payment instrument for domestic retail transactions (BIS Innovation Hub, 2021, p. 13). With Project mBridge, the digital yuan had, however, acquired an undeniable transnational dimension.

4 Conclusion

This chapter has provided a brief history of currency digitalization in the PRC since the turn of the millennium by drawing on a variety of publicly accessible sources including newspaper and journal articles. In this endeavor, we have borrowed insights about the dynamics and mechanisms of infrastructural change from a selection of works by scholars concerned with the development of large technical systems in other societal and historical contexts. Application of this “infrastructural gaze” has helped us to formulate two distinct arguments. First of all, in a way similar to many other large societal infrastructures, during its early years the construction of the digital yuan was at times presented as a means to completely overhaul existing sociotechnical systems. Yet, when work on the project began in earnest over the second half of the 2010s, a process of layering unfolded, whereby the digital yuan came to rely on already existing sociotechnical systems. Second, not unlike other sociotechnical systems, the digital yuan is embedded in a transnational network of financial infrastructures which crosses national boundaries. The nature and structure of this transnational network determines in crucial ways the ultimate sociopolitical function that the digital yuan plays and will play in the international political economy.

Financial infrastructure is of decisive importance in efforts to enhance financial efficiency and promote economic growth (Amable and Chatelain, Reference Amable and Chatelain2001; Bossone, Mahajan, and Zahir, Reference Bossone, Mahajan and Zahir2003). This chapter explores how the development of China’s financial infrastructure also affects financial inclusion. For a long time China had lagged behind in terms of conventional financial inclusion, but China’s digital financial inclusion (DFI) is rapidly developing (Bei, Reference Bei2017; Jiao, Reference Jiao2019).¹ One of the most important catalysts has been the development of digital financial infrastructure, particularly digital credit infrastructure and digital payment infrastructure. This chapter first examines the concepts of financial infrastructure and financial inclusion and their Chinese contexts. It then traces the historical evolution of financial inclusion in China since 1949 and lastly dissects the significant role of digital financial infrastructure in promoting Chinese DFI since around 2013.

1 Financial Inclusion and Financial Infrastructure in the Chinese Context

2 Evolution of Financial Inclusion in China: Effects of Conventional Financial Infrastructures

The evolution of China’s financial inclusion can be divided into five stages, and the five features of financial inclusion (accessibility, affordability, commercial sustainability, responsibility, and opportunity equality) can be used to evaluate and compare these stages. In addition to detailing the development of China’s financial inclusion, this section will discuss the role of conventional financial infrastructures, specifically credit infrastructure and payment infrastructure, during these stages. Credit infrastructure includes laws, management, institutions, standards, and markets related to credit reporting. By providing third-party evaluation, credit infrastructure can reduce information asymmetry and transaction costs, thereby indirectly improving the level of financial inclusion (Huang, Reference Huang2018). As the cornerstone of any financial system, payment infrastructure generally encompasses the accounts, instruments, systems, organizations, and regulations that support payments. Payment infrastructure has a direct effect on financial inclusion by, for example, expanding coverage of financial services and improving settlement efficiency. This section reviews the evolution of financial inclusion in China and examines the impacts of financial infrastructure on financial inclusion.

3 Digital Financial Infrastructure since 2013: A New Stage of China’s Financial Inclusion

The normative and commercial objectives of financial inclusion are hard to realize simultaneously through the traditional means of providing financial services, largely because of the Pareto principle mentioned earlier in this chapter. For instance, the flaws of microfinance in improving financial inclusion become prominent, and skepticism emerges (as discussed by Shakya and Rankin, Reference Shakya and Rankin2008; Duvendack et al., Reference Duvendack, Palmer-Jones, Copestake, Hooper, Loke and Rao2011; Bateman, Reference Bateman2014). In the period since 2013, China’s financial inclusion has greatly improved (Guo et al., Reference Guo, Wang, Wang, Kong, Zhang and Cheng2020), from the country being far behind the international pioneers to gradually becoming a global leader of DFI. This achievement has been mainly driven by the advances in digital financial infrastructures, including digital credit infrastructure and digital payment infrastructure.

3.2 Digital Payment Infrastructure: Explicit Impact on Inclusion

Digital payment infrastructure has had a more direct impact on transforming China’s financial inclusion in DFI. As discussed, the digital development of payment infrastructure is the result of the application of a series of digital technologies in the payment field, including cloud computing, bio-identification, QR (quick response) codes, and blockchain. In China, two digital payment infrastructures have the potentially greatest impacts on financial inclusion: third-party mobile payments and digital RMB. As the latter is still in the pilot stage, we will focus mainly on the former.

Third-party mobile payments mainly refer to the mobile payment services offered by nonbank payment companies, especially internet payment companies. In China, Alipay and Tencent Finance together account for more than 90% of the third-party payment industry (Yang, Reference Yang2019). Since 2013, mobile payments have experienced dramatic development in China. As shown in Figure 30.1, the total volume of third-party mobile payments increased more than 200 times, from 1.2 trillion yuan in 2013 to 249.2 trillion yuan in 2020.

Figure 30.1 Third-party mobile payment transactions in China (1 trillion yuan).

Source: Authors’ elaboration based on data from iResearch (2021).

The explosive growth of third-party mobile payments has revolutionized Chinese financial inclusion in terms of the five core features of financial inclusion mentioned: namely, accessibility, affordability, commercial sustainability, responsibility, and opportunity equality.

First, third-party mobile payments have greatly enhanced the accessibility of financial services. Although traditional payment infrastructures such as bank outlets, ATM, and POS (point-of-sale) terminals still play an important role, use of these traditional infrastructures has been shrinking. By the end of 2021, China had 947,800 ATM terminals, a decrease of 66,000 from the previous year (PBOC, 2022b). In contrast, by the end of June 2022, the total number of cell phone subscribers in China reached 1.67 billion, a net increase of 25.52 million over 2021, twice as many as in 2010 (Ministry of Industry and Information Technology, 2010, 2022). Besides, as of June 2022, the number of internet users in China was 1.05 billion, with the internet penetration rate reaching 74.4% and the proportion of mobile internet access reaching 99.6% (CNNIC, 2022). According to PricewaterhouseCoopers, China’s mobile payment penetration rate reached 86%, ranking first in the world in 2019 (PwC, 2019). Third-party mobile payments have made financial services significantly more accessible.

Second, third-party mobile payments and the digital technologies behind them have also enhanced the affordability of financial services. One of the key factors for affordability lies in reducing the operating costs of financial service providers. On one hand, the improved accessibility of mobile payments also reduces the cost of many financial services. In the field of rural credit, the annual interest rate of microfinance institutions and informal credit was relatively high, reaching 18% or 19%, or even 20%, due to traffic restrictions and other factors (Bei and Li, Reference Li2017). In contrast, the annual interest rates of Alipay’s Jiebei and Tencent Finance’s Weilidai are usually 14.6% and 10.95%, respectively (Wu, Reference Wu2019). On the other hand, digital technologies can directly lower the cost of financial services. For example, based on cloud-computing technology, Alipay’s single payment cost is 0.02 yuan, which is ten times lower than traditional payment methods (Bei, Reference Bei2017).

Third, third-party mobile payments have more potential for business sustainability. Due to the aforementioned improvements in accessibility and affordability, these mobile payment agencies can achieve boliduoxiao (a sales strategy to increase total revenue by selling goods at a low profit per unit price). In 2021, nonbank payment institutions processed 827.3 billion online transactions worth 294.56 trillion yuan (PBOC, 2021), including 249.2 trillion yuan in third-party mobile payments (see Figure 30.1). The profits of the third-party payment platforms are mainly from sinking funds, handling fees, advertising revenue, and other value-added service income (Guo, Reference Guo2021). Taking Alipay as an example, after the launch of Yu’E Bao in 2013, it absorbed 400 billion yuan of investment in less than a year (Lu, Reference Lu2018). The revenue of Jiebei was 3.8 billion yuan, 11.2 billion yuan, and 11.8 billion yuan in 2016, 2017, and 2018, respectively, and the net profit was 1.8 billion yuan, 6.1 billion yuan, and 3.5 billion yuan, respectively (Zhongtai Securities, 2020).

Fourth, the third-party mobile payment agencies and the regulatory authorities have jointly safeguarded the responsibility of financial services. From the perspective of third-party mobile payment agencies, digital technology and companies’ self-regulatory mechanisms are two important aspects to protect the interests of customers. Fintech companies establish risk control and prevention systems through technologies such as the Internet, big data, and cloud computing, as well as through internal regulatory institutions, thus reducing the risk of financial services (Peng and Wu, Reference Peng and Wu2018). For example, Ant Group has established a consumer rights protection and management system based on architectural guarantees, internal control mechanisms, and platform construction (Ant Group, 2021). From the perspective of regulatory authorities, the PBOC and other government agencies have played critical roles. One of the representative events was the establishment of a common clearing platform for nonbank and third-party payment institutions, the NetsUnion Clearing Corporation, in August 2017. Another regulatory action was the Shanghai Stock Exchange’s suspension of Ant Group from listing on the Science and Technology Innovation Board in November 2020.

Fifth, third-party mobile payments have technically promoted opportunity equality. Opportunity equality emphasizes breaking the Pareto principle and building solid financial systems and a more equal society. As mentioned, third-party mobile payments create more opportunities by improving accessibility and affordability. They can also contribute to increasing equality between users and regions. For users, as the Ant Group claims, “we strive to enable all consumers and small businesses to have equal access to financial and other services through technology” (Ant Group, 2022). In particular, vulnerable groups can enjoy the same financial services as others, as long as they or their families have a cell phone. For regions, third-party mobile payments largely eliminate geographical and spatial limitations, enabling remote areas such as small mountain villages with less developed transportation links but good internet connections to have better access to financial services.

To conclude, the extent of financial inclusion is determined by many factors, among which financial infrastructure is most decisive. Digital financial infrastructures have transformed the conventional ways of conducting financial interactions and activities and made financial services more accessible. Meanwhile, financial infrastructures are not purely neutral and value-free at either the domestic or the international level. Within society, digitalization of the financial industry brings new redistribution effects, leading to new winners and losers. In the international system, the redistribution effects also exist among states, which has resulted in new forms of international power competition, such as the digital RMB versus other central bank digital currencies. The importance of financial infrastructures in domestic societies and international systems, as emphasized in this book, needs further exploration.

1 Introduction: Patterns behind Market Prices, an Old Obsession

John Maynard Keynes, the renowned economist, was also an outstanding mathematician – long before it became a ‘norm’ in the former discipline. Between the 1920s and 1930s, he amassed a mountain of data on securities trading in capital markets and set himself the task of identifying patterns among them. He believed there was a hidden law governing the movement of stock prices, a sentiment shared by many economists grounded in the classical tradition. However, to his disappointment, he found that the movements were almost entirely random. The behaviour of markets could not be predicted precisely. This only apparently prosaic discovery led to one of the greatest revolutions in twentieth-century economics; non-probabilistic uncertainty was elevated to the centre of economic analysis (Skidelsky, Reference Skidelsky2005). The simple yet powerful idea is that, since the future cannot be anticipated, uncertainty conditions economic agents to hold money instead of making spending decisions. Uncertainty becomes the primary reason for fluctuations in investment and liquidity preference, in a game driven by agents’ expectations.¹

Nearly 100 years later, the penetration of artificial intelligence (AI) into financial markets continues to echo this same issue. Despite abundant critiques coming from non-conventional economics and social sciences, the hidden patterns of markets and their mathematical regularities continue to be obsessively sought, often without success. Increasingly sophisticated pricing and forecasting models were developed only to, despite their advances, prove limited and fail again, as the tragic cases of Long-Term Capital Management (LTCM) in 1998 (Quiggin, Reference Quiggin2010, pp. 56–58) and the great financial crisis of 2008 (Tooze, Reference Tooze2019) didactically demonstrate.²

But what explains this relentless fight against contingency, this inglorious quest for control, at any cost, over unpredictability in the markets? Why would the verification of non-probabilistic uncertainty not have been enough to appease those who would continue this search? And why was the systemic coordination of markets with the aid of the state, something that achieved relative success during some decades of the twentieth century, also insufficient? This chapter contextualizes these questions within our digital age, exploring AI as a financial infrastructure. The objective is to underscore the persistent connections it has with a volatile, crisis-prone system, challenging the commonly espoused notion of AI as a stabilizing force in finance.

Since the 2000s, the data-centric digital circuits in global finance have increasingly steered financial markets towards a comprehensive algorithmic mode of management (Campbell-Verduyn, Goguen, and Porter, Reference Campbell-Verduyn, Goguen and Porter2017). Enthusiastically, AI is said to be the new frontier of this process. Following the meso-level ‘infrastructural gaze’ proposed by Westermeier, Campbell-Verduyn, and Brandl (this volume),³ this chapter interrogates the penetration of AI in capital markets as a combination of both continuity and change in finance. In doing so, it sheds light on the complexity arising from the interconnections of its micro and macro dimensions. Building on the Science and Technology Studies (STS) definition of infrastructures – as evolving socio-technical systems that combine human and non-human elements (Bernards and Campbell-Verduyn, Reference Bernards and Campbell-Verduyn2019) – and on the political economy of digitalized finance (Paraná, Reference Paraná2019) – I reflect on AI as an ‘emergent’ financial infrastructure.

The increasingly infrastructural dimension of AI in capital markets comes from, first, an evolution from algorithmic trading and algorithmic financial governance, and secondly, from the rise of AI as a ‘general-purpose technology’ within the financial domain. Following this rationale, I draw on insights from Danielsson, Macrae, and Uthemann (Reference Danielsson, Macrae and Uthemann2022) and Borch (Reference Borch2022) regarding AI and systemic risks, hypothesizing about the consequences of the ‘infrastructuralization’ of AI in capital markets, considering the micro–macro tension typical of capital accumulation, inequality, and crisis dynamics.

By ‘infrastructuralization’, that is, ‘becoming infrastructural’, I understand the remixing or re-establishing of socio-technical relations that position a combination of processes and artefacts as crucial for economic reproduction. Importantly, infrastructures constantly (re-)emerge, particularly during stress events (Scholz, Schauer, and Latzenhofer, Reference Scholz, Schauer and Latzenhofer2022). In depicting financial AI as ‘emergent’, I want to highlight that we should not perceive it as entirely ‘new’. Instead, it is linked to a longer history of developments in big data, algorithms, and even earlier forms of digitalization, all indicating this continuous ‘re-emergence’. By focusing on the sedimentation of somewhat but not entirely different socio-technical relations, an infrastructure perspective enables us to comprehend AI within a broader and more nuanced historical context.

Pursuing this argument, the chapter first delves into an overview of AI’s current applications in financial markets. At this point, it problematizes AI as a possible emergent financial infrastructure. It then proceeds to provide historical background around the dynamics and underlying principles of digitalized finance (Paraná, Reference Paraná2019), elaborating on the relations of finance and technology more broadly, and on the uptake of models, computation, and algorithms in the financial domain in particular. The chapter concludes by drawing the consequences (unpredictability, operational inefficiency, complexity, further concentration) and (systemic) risks arising from such an emergence of AI as a ‘new’ financial infrastructure, particularly those related to biases in data and data commodification, lack of explanation of underlying models, algorithmic collusion, and network effects. It does so by exploring the micro–macro tension, and the fallacy of composition, in financial AI.

2 AI as a Financial Infrastructure?

In addressing the infrastructural aspect of financial AI, it is necessary to understand that the effects it produces – be they good or bad – do not occur solely through the development and dissemination of certain technical artefacts. Here, we can start by unpacking the concept of financial infrastructure, brought forth by the Social Studies of Finance (SSF) literature in dialogue with that of STS and extensively developed throughout this volume.

Infrastructure can be a slippery and somewhat controversial concept (Silvast and Virtanen, in press). For the most part, infrastructures are viewed as social-technical relations and networks that enable or disable functions/actions (such as banking, payments, insurance, and trading), or, as stated, ‘systems through which basic but crucial enabling functions are carried out, but that tend to be taken for granted and assumed’ (Bernards and Campbell-Verduyn, Reference Bernards and Campbell-Verduyn2019, p. 776). This means that infrastructures include not only physical technologies but also protocols, standards, and ideas (Edwards, Reference Edwards, Misa, Brey and Feenberg2003), encompassing both ‘hardware’ and ‘software’. Infrastructures bridge different scales of action, from micro to macro. This relational understanding of infrastructure focuses attention on multiple forms of agency in key choices and possibilities in both stable and unstable times. Building from Edwards (Reference Edwards, Misa, Brey and Feenberg2003), Star (Reference Star1999), and Hanseth and Monteiro (Reference Hanseth and Monteiro1998), as compiled by Bernards and Campbell-Verduyn (Reference Bernards and Campbell-Verduyn2019) and Silvast and Virtanen (in press), STS-inspired definitions of infrastructure point to the following features: facilitation (infrastructures have a supporting function), openness (they are shared by large communities and tend to be open to new entrants), durability (they tend to persist, evolving and extending, over time), centrality (they establish how core functions are deployed), and some level of obscurity (as they are constituted of ecologies of networks operating in the background, they can seem like black boxes for users).

AI surely exhibits some of these characteristics when it comes to financial markets – such as facilitation, openness, and obscurity – while, in a sense, it is still to be seen whether it is going to fully meet other strategic requirements, such as durability and centrality. However, if we take AI (mostly machine learning(ML)) as a second-generation development of automated trading, as argued by Borch (Reference Borch2021),⁴ it is on the verge of achieving such a point of no return and coercive adoption on the part of various finance actors. Be this the case or not, the intellectual exercise I propose here, following the insights of Danielsson, Macrae, and Uthemann (Reference Danielsson, Macrae and Uthemann2022), suggests that, once and if AI becomes pervasive to the point of becoming such an infrastructure for capital markets, as indications suggest, the outcomes should be the ones I will delineate. For that, my infrastructural gaze, while maintaining an eye on the details underpinning macro-level issues, focuses on emergent systemic concerns.

AI engines have been increasingly implemented in markets in recent years. This implementation takes place through different techniques and in different application sectors. The objective, in all cases, is to respond to the financial market factors popularized through the acronym VUCA – volatility, uncertainty, complexity, and ambiguity. The expansion of production and access to financial data of all kinds, coming from different sources and locations, the increase in computational-processing capacity, the expansion of the reach and speed of connectivity, the expressive profits of funds and financial companies based on AI, and intense competition in this information-sensitive sector have led AI to appear as a general-purpose pervasive technology in the functioning of markets,⁵ with different levels and forms of application. In Srnicek’s (Reference Srnicek2019) terms, ‘AI can become a utility, like electricity, like gas.’ AI, in sum, is being pitched as a potential ‘general condition of production’ (Dyer-Witheford, Kjosen, and Steinhoff, Reference Dyer-Witheford, Kjosen and Steinhoff2019, p. 31) for financial activities; in other words, a financial infrastructure.

Among the most used AI techniques are expert systems, genetic algorithms, fuzzy logic, neural networks, and ML (Milana and Ashta, Reference Milana and Ashta2021). These techniques can be used in combination depending on the application and its objectives. Among these, the most used is ML; and while deep learning has garnered considerable attention in recent years, the predominant AI technology within this context remains supervised learning (Li et al., Reference Li, Sigov, Ratkin, Ivanov and Li2023). Financial forecasting directed to financial trading is the main subfield of AI used in finance. Although there is a growing body of literature on the promises of generative AI in finance (Ali and Aysan, Reference Ali and Aysan2023; Cao and Zhai, Reference Cao and Zhai2023; Dowling and Lucey, Reference Dowling and Lucey2023; Lopez-Lira and Tang, Reference Lopez-Lira and Tang2023), its impact is still hard to discern and difficult to evaluate at this point.⁶ By all means, financial AI is mostly ML techniques applied in different domains.

Among the most relevant practical applications are credit scoring and rating, risk management, forecasting, fraud detection, accounting, financial advisory services, and financial trading (including news trading). These applications span across different agents, such as banks, financial institutions, fintech, funds, and trader dealing in various financial instruments.

Importantly, from an infrastructural perspective, the so-called intelligent layer of AI is just the apparently superficial layer (Pasquinelli and Joler, Reference Pasquinelli and Joler2021). Such a layer is, nonetheless, of enormous importance. Proprietary models are kept under lock and key by financial market agents. However, these sophisticated models could do little or nothing without global high-speed connectivity, high computational-processing capacity, and an immeasurable amount of data produced, received, and catalogued from all parts. This is what ML, the still-dominant model for applying AI in financial markets, fundamentally requires: data/information, models/software, and hard/fast processing. Once strategically combined, these elements are the true game-changers for the dominant AI application technique in the markets.

In this regard, it is necessary to highlight that such applications are overdetermined by their materialities (Crawford, Reference Crawford2021; Steinhoff, Reference Steinhoff2021). They mobilize a large number of natural resources, human labour, and a whole complex of technical, institutional, political, and geopolitical factors that, as conditions of possibility, tend to go unnoticed (Dyer-Witheford, Kjosen, and Steinhoff, Reference Dyer-Witheford, Kjosen and Steinhoff2019; Pasquale, Reference Pasquale2020). Thus, in a sort of fractal, thinking about the infrastructures behind AI is critical to thinking about AI as infrastructure. This involves understanding why certain forms of automation were developed to the detriment of others and what social mobilization underpins such huge efforts in financial markets. However, this is only possible if we manage to bring to the analysis the different spatialities and temporalities expressed in the varied scales within which these processes are constructed (Corpataux and Crevoisier, Reference Corpataux and Crevoisier2016). Scalability – both geographical scales and scales of data and information processes within it – is fundamental to understanding not only what an AI system is, how it is produced, and how it works, but also to understanding the potential aggregate effects of its widespread application in a given sector. Scale is about power (Crawford, Reference Crawford2021): it is a product of power relations and entails and frames power dynamics.

That considered, stating that AI is being established as a financial infrastructure leads to new regulatory, institutional, and political disputes, along with emergent problems, tensions, and risks (Borch, Reference Borch2022). Examining it from a meso-level infrastructural perspective enables us to focus on the emergent scale involving interactions and material aspects connecting individual agents to the broader financial dynamics and their collective relational outcomes. Importantly, this lens pushes us to see ‘emergence’ over a far longer period – way before generative AI and ChatGPT, for example – as I discuss in the Section 3.

3 The Computerization of Markets and Its Consequences

Although the widespread adoption of AI in financial markets may seem to indicate a new financial infrastructure combining big data, deep computation, and ML models, these ‘new’ infrastructures are built upon and broadly reproduce ‘older’ ones – models, computation, and automation in general (Muniesa, Reference Muniesa2003, Reference Muniesa2007; Pardo-Guerra, Reference Pardo-Guerra2010; MacKenzie, Reference MacKenzie2015). More than supporting the STS accounts of infrastructures as never ‘settled’ and always emergent (Star, Reference Star1999; Edwards et al., Reference Edwards, Bowker, Jackson and Williams2009), this complex combination of continuity and change merits an investigation into the underlying logic this dynamic entails and, particularly, the potential risks arising from its growing complexity. This section contemplates the deeper, longer links of financial AI with the ‘installed bases’ of what existed previously.

As a departure point, it is worth noting that from the 1960s onwards, with the continuous advancement of computation, the search for anticipation, control, and parameterization within financial markets that was mentioned in Section 1 has become increasingly reliant on technological development – as also occurred in other dimensions of social life.

Under capitalism, technological development must respond to competition in the market. This is especially valid for financial industries, defined by the fluidity of its main ‘commodity’ – capital itself in its financial and monetary form, and, more particularly, ‘fictitious capital’ (Mollo, Dourado, and Paraná, Reference Mollo, Dourado and Paraná2022). Such competition is not only intense but, above all, fast and, more than that, ultra-sensitive to information. That is why financial markets, in many cases, not only anticipate but prefigure the development and implementation of a special type of technology, information and communication technologies (ICTs). As finance attempts to anticipate futures, it lays the ground for presents. Therefore, anticipating other fields of application by a few decades, global-reaching ICTs spread and quickly imposed themselves as the basic infrastructure for the functioning of financial markets from the 1970s onwards. And with that, a world of new–old promises opened: communication, machines, the rapid exchange of information, and their automatic processing could finally help us find, through the discovery of hidden patterns in price fluctuations, not only the path to the most profitable investment strategies but, in general, greater control, transparency, predictability, rationality, and efficiency in the markets. It is as if Keynes did not solve the problem of non-probabilistic uncertainty in markets simply because he had no access to computers, big data, and advances in modern mathematics. For some, it would only be a matter of time until these became available. The old Western techno-fetichism was updated, now under the banner of high-tech finance.

Along these lines, the prevailing perspective among major market players is a techno-solutionist approach to adopting these technologies (Campbell-Verduyn and Lenglet, Reference Campbell-Verduyn and Lenglet2023). This vision regards technology as neutral and external, almost magical, without acknowledging its co-construction within various social dimensions. This is a stance that is challenged by extensive scholarship within STS and SSF (see Pinzur, this volume). Thus, confirming the thesis of co-evolution and co-determination between finance, technology, and society, an evolved, ‘updated’ global financial market emerged, accelerating from the 1980s onwards. This would be an electronic market, globally interconnected, with negotiations in real time and increasingly mediated algorithmically. In particular, algorithmic trading and high-frequency trading (MacKenzie et al., Reference MacKenzie, Beunza, Millo and Pardo-Guerra2012; MacKenzie, Reference MacKenzie2015) emerge within what I defined as ‘digitalized finance’, the new globally integrated socio-technical management system for the valorization of financial capital (Paraná, Reference Paraná2019). This occurs through the deployment of cutting-edge automated technologies, which accelerate the compression of space-time flows to obtain short-term liquid financial gains on a global scale.

Finance and technology have long been interconnected (Pérez, Reference Pérez2002; de Goede, Reference de Goede2005; Ajji, Reference Ajji2020). Technology is not merely an external infrastructure but a fundamental component that plays a crucial role in market dynamics (Knorr Cetina and Bruegger, Reference Knorr Cetina and Bruegger2002a, Reference Knorr Cetina and Bruegger2002b; Pardo-Guerra, Reference Pardo-Guerra, Orton-Johnson and Prior2013). In this regard, a relevant fact that has largely gone unnoticed is that the implementation of mathematical models and trading algorithms in financial markets anticipates by a few decades their adoption in other fields of application that have become more popular, such as search engines and social media. It is possible to say that algorithmic governance (Zetzsche et al., Reference Zetzsche, Buckley, Arner and Barberis2018) goes from Wall Street to Silicon Valley, and not the other way around, as is commonly thought. More particularly, as Joque (Reference Joque2022) indicated, the Bayesian and inductivist logic of neural networks make them particularly fit for the financial speculation drives. This anticipation demonstrates another important aspect: if finance operates in an increasingly technologized way, contemporary technological development is carried out more and more by mimicking the modus operandi of finance itself and its form of governance – leveraged and speculative, short-term, fast, individualizing, deregulating (Srnicek, Reference Srnicek2016). Therefore, it is not by chance that financial companies increasingly resemble technology companies, and technology companies increasingly resemble financial institutions. The rise in prominence of the financial dimension of economies and the extensive digitalization of social life are mutually attractive processes in contemporary capitalism.

At this point, particularly from the mid-2000s, we can already observe the consolidation of some developments in these dynamics, which have been unfolding for several decades. Alongside the expansion of economic concentration at different levels and the additional difficulties in regulating the markets that these processes entail, the operational complexity of the markets is increasing. In my previous work (Paraná, Reference Paraná2019), based on a comparative analysis of American and Brazilian financial markets, I identified what I have come to define as the ‘spiral of complexity of digitalized finance’. This spiral (Figure 31.1) is a feedback process that fuels the growing complexity of financial markets, to the point that even many financial operators – let alone the regulators – do not exactly know what is happening in the markets. ‘Flash crashes’ are a good example of how hard it can be to explain these phenomena. In my understanding, AI amplifies this process. The spiral illustrates how completely rational individual decisions can produce ‘irrational’ social outcomes. There are three basic steps to it.

Figure 31.1 The spiral of complexity of digitalized finance.

Source: Paraná, Reference Paraná2019.

First, in highly digitalized finance, the search for unexploited financial gains incentivizes leading players to invent, deploy, and refine the means needed to overcome the technological or regulatory obstacles that stand in the way of speculation. Secondly, the wide adoption of these new technologies leads to the emergence of new institutional configurations, new modes of action, and new operating dynamics; as a result, markets are reconfigured, partly in response to the political and social conflicts that erupt. Finally, the ensuing emergence of a new institutional and technological environment encourages the development and implementation of even newer technical solutions. This cycle of growing complexity comes to entangle investors, regulators, tech companies, and other (often unwilling) participants. New systemic risks and instabilities proliferate as a result. This chapter explores it in the case of financial AI. Although the risks arising from pervasive AI adoption are new, I argue that the infrastructuralization of AI in finance, as defined earlier in this chapter, occurs along the same lines. This process underpins the spreading of AI through financial markets.

One of the main impacts that digital technologies (AI included) have on the market is to shorten space-time flows, that is, they make things faster and nearer.⁷ This increases the number and volume of trade operations and transactions. The digital systems behind those trades and transactions are immensely complex; they operate at immense speed and have tremendous learning and adaptive abilities. They make the job of regulating markets far more challenging (Pardo-Guerra, Reference Pardo-Guerra, Knorr Cetina and Preda2012; Arnoldi, Reference Arnoldi2016); there is too much systemic opacity and uncertainty involved. As regulators struggle to keep up, further concentration and centralization of capital within and between markets tend to occur. Those with more advanced technologies tend to enjoy higher competitive advantages and higher profits, leading therefore to centralization.

And here emerges another tension. As the global financial system became, in a sense, more unstable, opaque, and unpredictable – not least because of high-frequency trading and other innovations – it somehow needed to ground its operations in something that promised certainty, transparency, and predictability. This is where we find the new frontiers of this process: blockchain and, above all, AI. However, the same actors who seek stability and transparency through AI are the ones generating instability and opaqueness.

In some sense, AI’s promises represent the actualization of an old paradox. While financiers need to amass information to tame uncertainty, they themselves want to live in a low-information environment, avoiding the scrutiny of consumers, citizens, governments, and the media, at least as far as their own profit-making strategies are concerned. So, businesses always want almost full transparency from everyone but themselves. Their competitors – as well as governments and technocrats – have similar goals, producing, as a combined result, uncertain, risky, and rather opaque scenarios.

The struggle between these two tendencies – their particular combination in different contexts – is the key to understanding information governance in financial markets. It reflects the fundamental capitalist contradiction, that is, that production is a collective social process that, however, is privately conducted and controlled (Marx, Reference Marx1992 [1867]). Since the underlying contradiction can never be properly resolved, all that these disputes can do is somehow precariously accommodate these self-cancelling tendencies.

Therefore, such a ‘spiral of complexity’ entails not only a rather uncontrollable dynamic but also a highly confusing one: few experts and operators can navigate it well (Lange, Reference Lange2016; Lange, Lenglet, and Seyfert, Reference Lange, Lenglet and Seyfert2019). On the other hand, the growing opacity of the markets becomes an element of control, leaving the management of strategic information in the hands of a closed elite of financial actors. Governments and regulators, but also small investors, those acting on the edges, become increasingly dependent on large funds, brokers, and financial institutions to manage their investments. This is usually done with very little transparency. And all of this, again, occurs in the name of efficiency, stability, and predictability. Opacity is the ‘other’, the truth of the quest for financial transparency. It is in this scenario that AI makes its way as an emergent financial infrastructure.

4 The Composition Fallacy in Financial AI

After this necessary historical contextualization, we can now return to the question, presented in the Section 1 of why the obsession with controlling contingency and unpredictability, the search for patterns, and the pursuit of access to as much data and information as possible have not only persisted unabated but also intensified. Its answer lies in the individualized, albeit generalized, struggle not only for profit but for maximum profit as the foundation of survival against the competitive dynamics in financial markets. This finding may seem rather prosaic, but it helps to reveal another central problem for the elaboration of a reflection on the advance of AI in the markets: the problem of the ‘fallacy of composition’, particularly visible from a meso-level infrastructural gaze. The fallacy of composition indicates that the search for control to obtain growing profit generates, along with more profit, a greater lack of control. The insertion of AI in this process both reflects and amplifies it.

The principle of the fallacy of composition, extensively explored in economics,⁸ deals with the non-neutrality of aggregation and aggregates. In simple terms, the principle states that what is valid for the parts may not necessarily be valid for the whole. In short, the whole tends to be not only quantitatively but also qualitatively distinct from the parts. It is a logical dilemma related to the classic epistemological problem of inductivism versus deductivism, between empiricism and rationalism, in the philosophy of knowledge. In any case, the principle of the fallacy of composition helps us understand the critical phenomena of the digital economy. It explains, in part, the occurrence of the so-called network and platform effects, the sector’s tendency towards the ‘winner takes all’ model, and even the logic of ‘too big to fail’, which tends to encompass some of the main players in the market.

This leads to my core argument that the penetration of AI in capital markets has markedly different effects in the micro- and macroeconomic, or micro- and macro-social dimensions, which the meso focus of infrastructures helps us to see better. In the micro dimension, it is possible to concede that the advancement of AI actually delivers on some of its promises for some agents, particularly the largest and best-positioned ones: greater profit, transparency, predictability, and efficiency. It is also possible to concede that, at the same level, AI contributes to accounting, supervision, and risk management in particular contexts. It is as if some agents – and the most notorious examples are the great hedge fund Renaissance Technologies of the legendary James Simmons and Robert Mercer and, more recently, BlackRock – had, with the help of powerful computers, big data, and AI, finally managed to solve the problem of numerical signals hidden in the markets to guide the best investment strategies.

However, this result is not verified in the topology of the aggregates – the one regulation and systemic risk assessment typically tend to focus on. Again, scale matters. At the macro level, the sign seems to be reversing, and we observe an increase in risk, unpredictability, and perhaps operational inefficiency in the markets – an effect that, as mentioned before, is accompanied by increasing concentration and inequality.⁹ In Section 5 I will unpack this argument, discussing some of the risks and problems of implementing AI in financial markets. This argument, it is worth noting, can only be built from an infrastructural perspective, in the expanded and systemic sense of the term – for both financial markets and AI.

5 Risks, Limits, and Challenges to Financial AI

These problems stem from the structural limitations of AI that are already widely known: the problem of biases in the data (Crawford, Reference Crawford2017; Eubanks, Reference Eubanks2018), sometimes amplified by these systems, and the lack of explanation of their underlying models. To these, we can add the excessive trust deposited in artefacts and technical systems as exogenous and neutral, supposedly free from human errors and emotions, free from the so-called human factor. These problems are interrelated in a complex way in the markets; that is, ‘certain systemic conditions in markets can allow individual firms’ high-reliability practices to exacerbate market instability, rather than reduce it’ (Min and Borch, Reference Min and Borch2022, p. 277).

Conditioning these limitations, we find the so-called problem of causality (Pearl and Mackenzie, Reference Pearl and Mackenzie2018) – the fact that these systems ‘learn’ and operate based on data patterns and correlations, not on causality. These are systems that cannot give causal explanations for their actions, despite finding strong correlation and accuracy rates. This relates to explainability (Bracke et al., Reference Bracke, Datta, Jung and Sen2019), but goes beyond that. The absence of causal articulation removes some limitations but, above all, removes some advantages of human thought based on human language. Unlike machines, humans reason by combining emotions, rules, moral precepts, philosophical assumptions, flexible adaptive objectives, and a good dose of errors to elaborate theories and causal explanations for their actions. This gives us not only the most probable explanations obtained from statistical correlations but also explanations that may be highly improbable and still correct, such as the gravitation theory or the relativity theory, for example (Chomsky, Roberts, and Watumull, Reference Chomsky, Roberts and Watumull2023). Given these limits, the interactional logic of markets (MacKenzie, Reference MacKenzie2019) poses a great challenge to financial AI. As Min and Borch (Reference Min and Borch2022, p. 277) demonstrate, ‘automated markets are characterized by tight coupling and complex interactions, which render them prone to large-scale technological accidents’.

Hypothesizing the role of AI as a central banker, Danielsson and colleagues (2022) present a compelling argument for why AI can potentially undermine financial stability by ‘creating new tail risks and amplifying existing ones due to procyclicality, unknowable unknowns, the need for trust, and optimization against the system’ (p. 1), hence ‘increasing systemic risk’.¹⁰ While they focus on financial supervision (an imagined ‘Bank of England Bot’), their rationale can be extended to the wider penetration of AI in other financial domains, in our terms, its broader dissemination as an emergent financial infrastructure. Their sharp focus on hidden logic, complex dynamics, and aggregation problems demarcates a productive contrast to the sometimes circular and repetitive micro perspectives dominant in technical literature.

Expanding upon their insights, the first challenge confronting financial AI emanates from the reciprocal influence wherein economic agents responding to AI concurrently impact the system. Recognizing that the utility of AI hinges on the structural alignment with the task at hand, its optimal application occurs when addressing problems involving a solitary agent with fixed objectives and rules within a relatively predetermined space of action. Deviating from this idealized model introduces heightened complexities. Economic agents’ decisions are substantially contingent upon the environmental and infrastructural framework within which they operate. Consequently, alterations in this milieu, instigated by the actions of AI, prompt agents to adjust the parameters governing their decisions. The behavioural responses inferred by the AI engine from historical data are contextual and may falter if the engine attempts to exploit them for control purposes. This nuanced interplay extends its ramifications to elements concerning the interrelationship between market agents and political systems. In essence, deploying a machine with predetermined objectives in a highly intricate environment yields unexpected behaviours.¹¹ As noted by Ashta and Herrmann (Reference Ashta and Herrmann2021), there exists a ‘technology-oriented’ risk associated with ‘overfitting patterns’, potentially leading to the misinterpretation of data by algorithms within the dynamic and ever-evolving ecosystem of financial markets. In alignment with this perspective, Borch (Reference Borch2022, p. 9) contends that

The second challenge of AI as financial infrastructure pertains to data access for market regulation and supervision. In digital businesses, although a significant part of data circulation is relatively open, facilitating various activities, data is increasingly treated as a commodity and private property (Rotta and Paraná, Reference Rotta and Paraná2022; d’Alva and Paraná, Reference d’Alva and Paraná2024). Notably, in the digital economy, while ML and AI algorithms are mostly free and open source, the financial data used for their training are typically closed and proprietary (Bholat and Susskind, Reference Bholat and Susskind2021). Although the financial system may appear to be the ideal use case for AI given its generation of seemingly infinite amounts of data, challenges such as measurement problems, data silos, and hidden interconnections limit the information that can be gleaned. Furthermore, agents might intentionally obscure the explainability of proprietary models to maintain a competitive edge. It is anticipated that many disputes between regulators and market agents will arise from these challenges.¹²

Another significant risk associated with AI penetration in finance concerns the dynamic of trust established in expert systems, whose functioning is not likely to be questioned until a serious failure occurs. When we observe AI performing well in low-level functions, it gives the green light to at least experimental adoption in higher-level functions. Cost savings on expensive human domain knowledge will provide additional incentives to adopt AI for trading and financial decision-making. While the issue of trust is present in the current set-up, crucial differences between human decision-makers and AI make the problem particularly pernicious. Determining how an AI reasons is more challenging than assessing a human decision-maker, and holding AI accountable presents additional complexities. Additionally, because we do not know how AI would react to the unknown-unknowns of economic dynamics – statistically irrelevant yet important causes of instability – the question of trust becomes increasingly pertinent as AI encroaches on meso-/macro-like problems. Major stress events often arise from interconnections between seemingly disparate parts of the system, fuelled by political linkages that only manifest themselves once the stress is underway. Crises, by definition, are unexpected and non-statistically relevant events – black swans – that generally occur outside the universe of consideration of models looking for patterns in data.

Additionally, the widespread adoption of AI and the use of standardized models by numerous agents may lead to strategy convergence. It can unexpectedly heighten the interconnectedness of financial markets, increasing correlations among previously unrelated variables. These correlations may strengthen network effects, causing unforeseen shifts in market dynamics. As asserted by Danielsson, Macrae, and Uthemann (Reference Danielsson, Macrae and Uthemann2022), AI is poised to exacerbate the inherent pro-cyclicality within the financial system. Essentially, AI is anticipated to magnify vicious economic cycles, where micro-level rationality leads to macro-level irrationality. AI’s capacity to identify and adhere to superior risk management processes may result in a homogeneous set of techniques vulnerable to the same unknown-unknowns. Convergence in risk perception and management strategies across various AI applications, both in public and private sectors, could diminish the diversity that typically mitigates stress causalities. The risk assessment homogenization from self-reinforcing intelligent machine strategies may compromise the system’s resilience. Moreover, the heightened performance of AI during periods of stability may foster increased trust, potentially encouraging additional risk-taking behaviour. In this manner, AI manifests as pro-cyclical, echoing Minsky’s (Reference Minsky2008, p. xii) proposition that ‘stability is destabilizing’.

A distinct limitation of financial AI lies in its role in risk management, as extensively discussed by Danielsson, Macrae, and Uthemann (Reference Danielsson, Macrae and Uthemann2022) and Danielsson and Uthemann (Reference Danielsson and Uthemann2023). While the theoretical understanding of fragility, fire sales, runs, and negative feedback loops is well established, their specific manifestations depend on contextual factors, such as the current financial market structure and political environment. Despite training on comprehensive datasets that include detailed observations of past crises, AI may not identify all vulnerabilities due to its inability to reason causally. Unlike humans, AI relies on correlation rather than causation. Human regulators possess the capability to experimentally investigate and determine causal relationships, allowing them to respond proactively or reactively. While human regulators cannot predict unknown-unknowns, they possess historical, contextual, and institutional knowledge, enabling them to respond effectively. This is a notable contrast to AI, which lacks this capacity.

A final challenge involves the potential for increased coordination among economic agents in a system marked by widespread AI use, particularly when algorithms learn to cooperate. Research by Calvano and colleagues (2020) indicates that independent reinforcement-learning algorithms can sustain collusive equilibria in pricing games, maintaining prices above competitive levels. Unlike human actors, AI can facilitate tacit collusion without explicit instruction, potentially exacerbating market manipulation and raising legal and practical concerns for regulators. Paradoxically, transparency, often considered a positive attribute, may inadvertently foster coordination, potentially undermining financial stability, especially in scenarios like bank runs. These situations highlight instances where AI-induced behavioural adjustments can result in detrimental feedback loops.

6 Final Remarks

Trust in human decision-making comes from a shared understanding of values and the environment. AI’s combination of built-in values, programmed objectives, and its understanding of the environment may not be intelligible to humans. While hypothetical scenarios can be executed through an AI engine to observe its decisions, soliciting a clear explanation remains challenging.

In summary, as the infrastructural gazing at the meso-level highlights, a tension exists between macro- and micro-financial problems, significantly influencing the utility and implementation of financial AI. Despite its fast penetration, effective operation in this environment necessitates capabilities beyond the current scope of AI, including an understanding of causality, global reasoning rather than local, and the identification of potential threats before they result in adverse outcomes (Borch, Reference Borch2022; Danielsson, Macrae, and Uthemann, Reference Danielsson, Macrae and Uthemann2022).

AI, emerging as a financial infrastructure, introduces specific threats to the integrity of the entire financial system. Challenges stemming from the infrastructuralization of AI must be tackled by financial actors and regulators. A thorough understanding of these threats can be attained by adopting a perspective that considers the macro/meso/micro connections inherent in infrastructures. Addressing these concerns as systemic issues requires an integrated approach that encompasses not only the technical aspects but also the interrelated social, political, and economic dimensions.