II. Central banks in payments: From wholesale (reserve accounts) to retail (instant payments)

Central banks are usually mandated by their country’s laws to ensure that their payment systems are solid (sound, trustworthy, and transparent), and efficient (competitive, cheap, fast, easy, and inclusive). One of the ways in which they fulfil this mandate is by acting as regulators and operators of reserve accounts held by banking/financial institutions (wholesale payment systems).Footnote ¹⁰ A central bank uses this function to (i) intermediate interbank transactions by using its own balance sheet as the settlement vehicleFootnote ¹¹; and (ii) provide liquidity standing facilities to institutions to ensure smooth settlements through reserves (up to providing emergency liquidity as a lender of last resort).Footnote ¹² In this institutional design, the central bank provides payment services to banks upstream, and the banks themselves provide payment services to households and businesses downstream. Lo, the so-called two-tiered structure.

Central banks use the instruments above to (indirectly) provide soundness to retail payments systems. However, even if that wholesale role is executed perfectly, retail payment systems might still be inefficientFootnote ¹³ : in other words, be expensive for payers/payees, slow, and non-interoperable between different PSPsFootnote ¹⁴. As a critical example, we have households and businesses who see themselves obligated to participate in multiple PSPs. Inefficiencies in retail payment systems raise the transaction costs of goods and services, harming commerce, hindering financial inclusion of the unbanked, and ultimately burdening economic activityFootnote ¹⁵.

If a retail payment system is sound, but inefficient, then the mandate of its central bank is only half accomplished. With this gap in mind, central banks have expanded their attention from wholesale reserve accounts to also the entire payments supply chain and end users.Footnote ¹⁶

Among the options to build on top of the two-tiered structure (again, where the central bank provides payment services to banks upstream, and the banks themselves provide payment services to households and businesses downstream), the preferrable option to deliver efficiency today is instant payments.Footnote ¹⁷ They pave the way for cheaper transactions, immediate fund availability, streamlined/quality digital experience, and so on – central banks are interested in promoting them on behalf of efficiency in retail payment systems.

From a traditional viewpoint, these authorities’ classic role has always been to address market failures in finance, such as information asymmetries, elevated barriers to entry, market power, and social cost of bank failure. In other words, they have traditionally worked to regulate away the negative externalities in financial markets.

Now, central banks may serve a greater cause. To cumulate the (classic) role of mitigating negative externalities with the (novel) role of promoting positive externalities and concrete social gains. As the BIS General Manager has described: “central banks are rolling up their sleeves to be at the front of the innovation process and help shape the digital transformation in the financial system – all while continuing to deliver on their mandates. In other words, they aren’t afraid of the algorithms, but are seeking to master the art of designing them”Footnote ¹⁸. Regarding how central banks may advance instant payments, according to the Committee on Payments and Market Infrastructure, there are three (non-exclusive) fronts.

First, the central bank may act as a catalyst. This is an institutional role at its core, where the authority leverages its expertise and relationship with public/private actors to advance IP agendasFootnote ¹⁹. Examples include setting national strategies and payments roadmaps, facilitating cooperation between public authorities (i.e., ministries, competition/telecom agencies etc.), supporting interoperability between schemes, establishing sandboxes, and coordinating experimental linkages for cross-border payments, DLT, CBDC, etc.Footnote ²⁰ Since this role is less crucial to our point than the other two, we will not refer to it again.

Second, the central bank may act as a regulator. To fulfil their objectives of soundness and efficiency in payments, central banks already monitor existing/planned payment systems, assess them against these two objectives and, where necessary, induce change.Footnote ²¹ They might control entry/exit of PSPs with license requirements, supervise licensed PSPs, and determine/enforce minimum standards (i.e., obligating PSPs to settle payments in a time limit, capping card interchange fees etc.). This role is business as usual for central banks in payments, and naturally extends to IP.

Third, the central bank may act as the operator, which is our focus here. The role where central banks provide IP-related services directly to the private sector, instead of only regulating the IP initiatives originated by the private sector. They may operate the financial infrastructure of IP schemes, known as real-time gross settlement system (RTGS, which includes providing liquidity), and the database directory that links payers/payees’ aliases with their account information, enabling the transactions to be clearedFootnote ²². This infrastructure is where/how instant payments are settled between PSPs after being commanded by their client. Over the last few years, the adoption of IP in countries has closely followed the adoption of RTGS by their respective central banks, revealing how these institutions incentivise IP when they are providers of its infrastructureFootnote ²³. Additionally, central banks may operate not only the infrastructure, but the instant payment scheme itself.

We should take a step back to differentiate the IP infrastructure and the IP scheme, since conceptual overlaps may happen between them in a scenario where the central bank operates both. We may adopt Brazil’s Pix as a practical guide since the Central Bank of Brazil operates both the financial infrastructure and the scheme. In short, Pix is the IP scheme launched in 2020 by the Central Bank to great success: over 70% of adoption by individuals in only 2 years, dramatically lower costs, relevant financial inclusion, and numerous international awards.Footnote ²⁴ In fact, the significantly lower financial costs of Pix adoption when compared to debit/credit cards, especially for individuals and small businesses, led the Brazilian IP scheme to surpass Brazilian card schemes in number of transactions.Footnote ²⁵

Brazilian PSPs participate in Pix’s infrastructure in the following way (to avoid any confusion, “Instant Payments System” in the following flowchart means the RTGS, and not the scheme) (Fig. 1).Footnote ²⁶

Figure 1. Brazil’s Pix flowchart of participants (Source: Central Bank of Brazil, 2023).

Per the image above, Pix’s infrastructure (RTGS and database) is operated by the Central Bank of Brazil and provided to the PSPs (for a low fee), which in turn provide IP services to Brazilian clients. These clients order payments, and the PSP settles these payments (instantly) between themselves inside Pix’s infrastructure. For reference, this database infrastructure is already populated by more than 700 million aliases from Pix payers and payees.Footnote ²⁷ These steps above only describe how the central bank operates the IP infrastructure, and not yet the IP scheme. Beyond them, the Central Bank of Brazil is additionally responsible for Pix’s entire rulebook, admission of PSPs into the scheme, branding/communication strategies etc. In practice, the authority owns the scheme.

As per the BIS publication on Pix, “the [Central Bank of Brazil] plays two roles in Pix: it operates the system, and it sets the overall rulebook. As a system operator, the [Central Bank of Brazil] fully developed the infrastructure and operates the platform as a public good. As rulebook owner, the [Central Bank of Brazil] sets the rules and technical specifications (i.e., APIs) in line with its legal mandate for retail payments”Footnote ²⁸.

This model might seem familiar, considering that the role of operator central bank in IP schemes is analogous to the role of card networks in card schemes. Similarly to how Visa and Mastercard “own” their card schemes, the Central Bank of Brazil is the “owner” of Pix. The similarities are very clear: (i) they make and enforce the rules, each in their own way; (ii) they are not PSPs nor front-end for consumers, but control the participation of the actual PSPs in their schemes; (iii) they have a firm hand on all matters of branding/communication strategy of the scheme.

III. The existing criticisms against the operator central bank

Many authors and authorities have highlighted the benefits of central banks operating instant payment (IP) schemes. This paper innovates by putting this operator role through a stress test. The following paragraphs outline the most common criticisms and concerns raised by academia and the market. Later in the paper, we address each of these points in detail.

First, it is common for many state intervention to be called “unwarranted intervention” in the free market. This is not different in central bank-led IP schemes. In short, critics state that (i) there is no clear market failure that justifies the state intervention of a central bank that provides instant payments; and (ii) there is no concrete evidence that market forces cannot spontaneously address the demand for instant payments.Footnote ²⁹ These critics state that central banks should passively wait for (and maybe assist) private actors to develop instant payments infrastructures and schemes, instead of developing it themselves. Bureaucracy, as these critics understand it, should wait for the free market to solve a problem – and if they do not solve it, then maybe there was never any problem in the first place. This rationale is not entirely dissimilar from the decades-old economic theory of regulation, which was sponsored by neoclassic economists (and business interests) to ultimately provide the rhetorical support for the deregulation policies across the twentieth century.Footnote ³⁰ In a similar way, advocates today argue broadly that a public authority (central bank) should not intervene in a free market (payment sector).

Critics also say that, if an operator central bank provides an IP scheme to the public, the institution becomes a “competitor” of private PSPs that also wish to provide their IP schemes. Since the central bank is also the supervisor of the PSPs, holding relevant regulatory sway over them, any competition between one and the other would be “unfair competition.” In this theory, an operator central bank would have the power to favour its own scheme in detriment of private schemes. The result would be that the central bank would pressure PSPs into embarking in its own scheme, instead of developing their own schemes or embarking in other schemes (generating what some might call “artificial” network effects). In short, critics are concerned that an operator central bank would leverage its supervisor role to coerce PSPs.Footnote ³¹

This “unfair competition” argument indirectly raises another argument – if a central bank operates IP schemes, while still carrying all its legacy responsibilities, would this situation result in conflict of interest inside that central bank? This criticism considers that the objectives/incentives of a regulator role are naturally different from those of an operator role, but both roles coexist inside a central bank that leads an IP scheme – and this coexistence is more conspicuous when the same department oversees the two roles.Footnote ³²

For example, we might consider a central bank that has the institutional power to enact rules/obligations that are objectively better in the payment system it supervises (i.e., minimum security requirements for payment schemes, maximum permitted time to settle, technical security standards, ISO formatting), and also operates an IP scheme in this system. However, that same central bank would not have the operational/technological capacity to comply with these intended rules/obligations, should they be enacted. Would the authority still enact these rules, only to become non-compliant itself? Or would the central bank’s regulator role hesitate in favour of its operator role?Footnote ³³

IV. Efficiencies vs. criticisms of central bank-owned IP schemes

Now that we have presented the current criticisms from market and academia against the operator central bank, we will present the potential advantages of the model (already made implicit on Chapter II), and confront the criticisms from Chapter III. We find that it is desirable for central banks to bear the role of operators of IP infrastructure and IP schemes – especially because of the efficiencies arising from this role alone and from its synergies with the role of regulator. We identify four of these crucial efficiencies.

First, lower entry costs and production costs for PSPs. The operator central bank is not a profit-maximising institution. As per the BIS General Manager, “central banks are designed to ‘make money’ only in the literal sense”, and their “bottom line” is exclusively the public good.Footnote ³⁴ Since profit is unneeded, the (economically neutral) central banks charge PSPs participating in the IP schemes only the bare cost needed by the authority to keep the infrastructure functional – without any price mark-up, like what already happens in reserve accounts and other services provided by central banks (recoup human/technology expenditure). As a result, we have (i) low entry costs for (small and large alike) entrant PSPs to participate in central banks’ IP schemes and compete with incumbent PSPs for consumers (business-stealing effects)Footnote ³⁵; and (ii) low production costs incurred by PSPs to offer IP to consumers, increasing the likelihood of lower prices for households and businesses (provided there is a competitive market).

Second, efficient infrastructure. The operator central bank’s IP infrastructure (RTGS and database) results in a single-rail scheme that (i) maximises economies of scale and internalises many externalities particular to payment systems (i.e., payers and payees paying for redundant, non-interoperable systems)Footnote ³⁶; (ii) enables provision of liquidity to PSPs by the central bank; (iii) allows the central bank to create synergies with its other regulatory systems/databases, leveraging action against money laundering/fraud and other goals from its regulator roleFootnote ³⁷; and (iv) allows the central bank to use decades of expertise/critical mass acquired from running reserve accounts systems, putting them to good use in the IP infrastructure.

Third, PSPs may focus on products and not rails. The operator central bank allows the PSPs to develop what the private sector excels at developing: product. While the authority manages the IP scheme’s operation and IP infrastructure’s rails, the PSPs can focus on developing better payment products (cheaper, better, and more innovative) to compete against other PSPs for clientsFootnote ³⁸. When Mastercard openly criticised the Fed’s intention to provide RTGS back in 2018, its letter encouraged the authority to drop this more interventive role and, instead, “let private industry do what it does best – innovate and compete to meet market demand.”Footnote ³⁹ That is an excellent reason for the operator central bank to provide an IP scheme – so that the PSPs can compete on top of a centralised system instead of competing with the development of redundant rails.

Fourth, leveraging the operator role with powers from the regulator role. The operator central bank can combine its powers from its supervisory function to increase its IP scheme’s efficiency and adoption rate. The best examples are:

• (i) obligating the largest PSPs in the respective financial system to participate in the central bank’s IP scheme and IP infrastructure. By bringing all their clients to the IP scheme, the PSPs’ presence generates instant/exponential network effects. This requirement was fundamental for the success of Brazil’s Pix (where all PSPs with more than 500,000 accounts were obligated to join)Footnote ⁴⁰. The absence of any such requirement in the U.S.’ FedNow (a possible result of pressure from American incumbents) might explain its lacklustre adoption by American PSPs.Footnote ⁴¹ We note that this strategy is not exclusive to IP schemes, and has already been implemented by central banks in wire schemesFootnote ⁴²;

• (ii) capping fees that PSPs may charge from clients to make payments in the IP scheme. While this might initially sound like exaggerated government intervention, it is important to remember that the PSPs are accessing the IP scheme and infrastructure while paying cost price to the central bank. The central bank’s strategy when capping fees is to emulate a state of perfect competition and complete transfer of the low-cost IP infrastructure to the end user. The Central Bank of Brazil has achieved this by obligating PSPs to offer Pix transactions free of charge to natural persons and small businesses (i.e., those who might be financially included by cheap digital payments), but allowing the PSPs to freely charge non-small businesses. The Reserve Bank of India also adopted this approach by establishing a “zero MDR” rule for consumers and merchants in UPIFootnote ⁴³;

• (iii) standardising ISO messaging formatsFootnote ⁴⁴ and APIs to ensure an open-architecture/interoperable IP scheme. From a technology point of view, this is the gateway for the private sector to develop the most diverse/creative integrations (i.e., aggregation of financial and non-financial services like messaging and content), eventually reaching the point of smartphone “super-apps”Footnote ⁴⁵;

• (iv) establishing the governance of the IP scheme with joint participation/representativity from all segments of the payment sector, like incumbents and entrants, banks, and non-banks (i.e., Brazil’s advisory committee “Forum Pix,” and the umbrella of Indian institutions under National Payments Corporation of India forming UPI’s governance). This leverages the central bank’s relationship with the entire sector to ensure that the IP scheme will not be dominated by one private interest or another.Footnote ⁴⁶

In short, central banks can pair up their roles as operators and regulators to maximise the efficiency of their IP schemes.

Now that the practical advantages of central banks acting as operators have been clarified, especially when combined with their traditional regulatory role, it is possible to address the criticisms explored in Chapter III. Critics argue that there is no market failure justifying such intervention and that central banks create unfair competition against the private sector.

V. Beyond the traditional criticisms: Additional concerns that might impact operator central banks

When central banks become operators, they gain access to a more versatile toolkit to fulfil their mandate. Specifically, the mandate of ensuring that retail payment systems are not only sound, but also efficient. A non-operator central bank does not have access to this toolkit. This toolkit is needed to address inefficiencies and market failures which are not addressed by pure competition in the market. To correct these market failures is no more than the goal of regulation. In fact, the only novelty here is that the operator central bank is not correcting market failures from an exogenous role (regulator), but from an endogenous role (agent). Footnote ⁵⁹

While we have addressed the most common (and somewhat partial) criticisms against the operator central bank and its IP infrastructure and schemes, there are more legitimate concerns to keep in mind. First, there are opportunities to improve the operator central bank IP model.

We suggest simple governance solutions to address some of the criticisms above. There is no apparent reason for the operator role and the supervisor role to be carried out by the same personnel and same department inside a central bank. As such, to prevent any undue use of data from one role by the other, a central bank may distribute the roles to different departments, and establish a functional Chinese wall between them. As benchmark, the Reserve Bank of Australia has a clear division between the commercial department and the other departments.Footnote ⁶⁰

Besides governance, a minimum legal mandate accommodating the operator central bank may deter the allegations of acting outside of Congress’ intent. As argued in Chapter II, the central bank’s mandate to ensure efficiency and financial inclusion in the payment sector justifies its dual role as both regulator and operator.Footnote ⁶¹ Still, if a concrete law is put in place, that would further prevent disputes over the boundaries of operator central banks.

Those issues, however, pose no major difficulties to the discussion on the central bank´s operator role. There are far more complicated issues that should be under discussion but are still overlooked due to the attention dispensed to the criticisms addressed earlier.

First, there is data privacy. The volume of personal information generated by daily use of retail IP by individuals and businesses is colossal and being steadily stored by the public authority. A Chinese wall inside the State is needed to address the valid concern of privacy and surveillance powers, considering the plethora of data that may be collected and hoarded by an owner of a payment scheme, and eventually used for nefarious purposes by a public authority without the proper governance.Footnote ⁶² If a central bank owns an IP scheme that generates/stores half a country’s payment data, we cannot rely on traditional governance models for this new reality.

Second, and somewhat related, is cybersecurity resilience. Banks worldwide are still grappling with understanding cyber risk and how to quantify and mitigate it. This challenge is heightened by the interconnectedness of today’s financial system, the heavy reliance on third-party tech providers, and other consequences of the digital age.Footnote ⁶³

If a central bank operates IP infrastructure and schemes, the cyber risk is further magnified, becoming even more of a concern for society at large. How technically secure are the central bank’s data storages (full of individuals and businesses’ info) against hacker attacks? How much of the authority’s budget is being dedicated to a functional cybersecurity scheme? How effective are their firewalls? What criteria are they adopting when selecting contractors for improving all that protection? Are the operator central banks aware of these discussions?

Taking a concrete case, Brazil’s Pix has faced some data leaks over the last few years due to cyber vulnerabilities in the participating PSPs.Footnote ⁶⁴ The weak link has never been the Brazilian Central Bank itself; the PSPs were penalised, and the leaked data was relatively minor and non-transactional. Still, the leaks ratify the essentiality of increasing the sophistication of cybersecurity in the financial system.

On the same security note, IP schemes greatly facilitate digital fraud, made easier by the fast and irreversible nature of the system.Footnote ⁶⁵ There are already efforts to address the problem by regulators (i.e., KYC, “mule” accounts) and by technology (i.e., behavioural analytics, biometrics, geographical “safe zones” for transactions, etc). Even now, as the Basel Committee enforces the necessity to act on financial fraud prevention,Footnote ⁶⁶ that should be a larger concern for operator central banks that enable millions of retail payments daily through IP infrastructure and schemes.

VI. Conclusion

The brave/digital new world has enabled very advanced technology and high consumer expectation for digital services. In the financial sector, this is best represented by instant payments – to allow individuals and businesses to transfer money with speed, low cost, and availability around the clock. A welcome evolution against classic payment instruments, such as cash (costly to handle and easy to launder), cheques and wire transfers (slow and inconvenient), and cards (costly, especially for small businesses).

Yet, while the efficiencies of instant payments are evident, the private sector was remarkably slow to develop them. As such, central banks have decided to lead the way, building IP infrastructure and schemes with a dual role of regulator and operator. They leverage their institutional positions as economically neutral actors to take the reins and develop IP infrastructure and schemes with low entry cost, rulebook enforcement, and (preferably) mandatory participation by relevant PSPs. In a way, the central banks play in IP a structurally similar role to Mastercard and Visa’s roles in cards – owners of a payment scheme.

These IP schemes owned/operated by central banks have had resounding success, especially in Brazil (inclusion of 70 million persons) and India (over 300 million active users). Both schemes are led by their countries’ central banks, involved obligatory participation by the largest PSPs, and mandatory low/zero cost for individuals and small businesses (thus emulating a “perfect competition” scenario where the low cost of the PSP is passed on to the consumer).

By now, many reports have praised the practical advantages of IP infrastructure/schemes owned by central banks. This paper innovates by putting these practical advantages through the stress test, responding to market-oriented rhetoric levied by interested parties against operator central banks (with the corroborating data).

First, to address the claim of “needless government intervention without a clear market failure to justify it.” While there may be potential gains from IP (i.e., larger customer bases, improved loyalty, rails for innovative products), these gains depend on a large enough number of PSPs developing/participating in IP, and they may choose to favour the guaranteed revenue of less efficient payment instruments. This is a market failure of collective action, thus justifying regulatory guidance. As clear evidence to the problem, no fully private IP scheme has been as successful as Brazil’s Pix and India’s UPI, and even these were only developed when it became clear that PSPs were not gaining meaningful ground on providing instant payments.

Second, we need to address the claim that central banks owning IP schemes are a “source of unfair competition and conflict of interest against private agents.” We argue that central banks do not compete as innovators, but rather lead the way as coordinators for innovation. At most, central banks use the operator role to develop IP schemes that maximise efficiency and network effects – with fundamental assistance from the regulator role – and enables full competition between the PSPs on top of the central bank’s rails (with more innovative IP products for the end users).

While major scheme owners are challenged by the Brazilian Central Bank’s IP scheme (i.e., card networks and closed-loop schemes), depository institutions and payment initiators rejoice as they welcome new clients aboard. More individuals and businesses embark into the system, enjoying an easier and cheaper mechanism to transfer money, and consumer welfare flourishes.

For future research, it is essential to understand how the creation of new functions for instant payments aids in social inclusion and competition with other payment methods. Additionally, examining the potential impact on international business growth between countries from implementing cross-border instant payment systems is necessary.