In the second of two articles, Jeff Alvares analyzes the competing arguments around Pix under World Trade Organization rules—a debate involving broader questions about how international trade rules need to reflect the complexity of public services in the digital economy.
As examined in Part I, Brazil’s public instant-payment system Pix, rolled out to expand banking services to those whom private financial services had previously excluded, has achieved remarkable success by vertically integrating the infrastructure and application service layers. To accomplish this integration, it resorted to a dual-barrier framework: a legal barrier against payment schemes other than Pix using the infrastructure, and an economic barrier in the form of mandatory participation and zero fees that disincentivizes the development of alternative payment schemes.
But can this design be justified under international trade rules, or did less-restrictive alternatives exist? The question recently spurred an investigation by the Office of the United States Trade Representative (USTR). Answering it requires examining the rules of international trade.
The General Agreement on Trade in Services (GATS) balances the rights of World Trade Organization (WTO) members to provide public services to its citizens against commitments to open markets and unencumbered trade. Article I:3 excludes “services supplied in the exercise of governmental authority”—services neither commercial nor competitive with private suppliers—from GATS disciplines. Article VIII addresses monopoly suppliers, requiring that they not discriminate or abuse monopoly positions in adjacent markets. Under Articles XVI and XVII, any public provider must operate on non-exclusionary and non-discriminatory terms between domestic and foreign providers.
WTO case law tests necessity and proportionality to determine if a monopolized government service complies with GATS. Exclusive provision must be necessary to achieve public-policy objectives, and no less trade-restrictive alternative can be reasonably available. Panels generally accept exclusive provision of essential infrastructure that exhibits natural monopoly characteristics, provides non-discriminatory access, and is priced to reflect costs or policy objectives. User-facing services should remain open to competition unless market failures clearly prevent private provision.
In 2012, the United States won a case against China’s UnionPay, a state-backed payment card network, arguing its monopoly over card-transaction clearing and user-facing card services violated GATS market-access and national-treatment rules. The panel agreed, ordering China to open the market.
However, the UnionPay precedent doesn’t apply neatly onto infrastructure-layer monopolies that keep applications open. Those systems function as public utilities without profit motive, preserve competitive opportunities, and exhibit natural monopoly features, making exclusivity defensible.
Vertical integration creates significantly greater WTO exposure. By mandating participation among financial institutions, providing zero-pricing transaction fees, and standardizing experience across the entire payment scheme layer—not just the infrastructure—Pix-style systems foreclose competitive opportunities that may be legally open. Operating the entire payment-scheme layer arguably crosses into commercial activity, weakening governmental-authority defenses.
Therefore, the critical question for Pix is whether controlling the front end is necessary to achieve financial inclusion objectives, or whether less restrictive alternatives exist.
The case for trade restriction: Why Pix may violate WTO rules
From a trade-law perspective, several aspects of Pix’s design raise concerns. Most fundamentally, Pix creates a two-layered barrier to competition. The legal closure of the Instant Payment System (SPI) settlement rail to other schemes eliminates the infrastructural pathway that might otherwise enable rivalry. Mandatory participation and zero pricing then impose economic barriers that would prevent competition even if infrastructure access were technically available. This combination of legal monopoly and economic foreclosure faces the highest level of scrutiny.
The infrastructure monopoly alone might be defensible on natural monopoly grounds if access were open to competing applications. But extending exclusivity to the application layer introduces cumulative restrictions that may be deemed disproportionate. Mandatory participation, zero pricing, and standardized user experience create a government-operated payment scheme that competes with (and forecloses) private alternatives not by superior efficiency but by regulatory mandate.
India’s Unified Payments Interface (UPI) provides a compelling alternative model. The National Payments Corporation of India (NPCI), a public-private partnership, launched UPI in 2016, with the Reserve Bank of India regulating the system. UPI is a decentralized, open-API protocol that sits on top of the Immediate Payment Service settlement infrastructure—the equivalent of Brazil’s SPI—and allows banks and third-party developers to offer their own user-facing applications. India delineated the layers where exclusive public provision was justified (settlement, protocols, interoperability) from layers where private competition should thrive (front-end applications). NPCI provides mandatory infrastructure to all licensed providers while keeping payment-scheme rivalry.
The architecture enables universal access with competitive innovation. Numerous third-party applications—including Google Pay, PhonePe, and Paytm—operate atop the UPI API, creating an intensely competitive ecosystem of payment providers. The results rival Pix’s. UPI accounts for 85% of digital transactions, processing over 20 billion monthly transactions worth $280 billion, with 491 million individuals enrolled. Crucially, success hasn’t come at the private sector’s expense. Firms generate substantial revenue through merchant services, lending, and insurance built on UPI.
UPI demonstrates that monopolized public infrastructure can catalyze rather than stifle private competition. Companies compete over user experience, merchant tools, and complementary financial offerings while public rails ensure interoperability and universal access.
The existence of UPI raises the key WTO question: was Pix’s vertical integration necessary and proportionate to achieve financial inclusion, or would infrastructure monopoly with competitive applications have sufficed? If UPI achieves similar policy outcomes with less market foreclosure, it may be difficult to demonstrate that Pix was the least trade-restrictive measure reasonably available.
Beyond the Pix vs. UPI comparison, WTO necessity analysis would require Brazil to demonstrate why other less-restrictive approaches couldn’t achieve the same objectives. These might include mandatory interoperability requirements on private providers, price regulation of instant-payment systems, competition-law remedies against banking oligopolies, or subsidies for private instant-payment development. Each of these alternatives would preserve more competitive space than Pix’s integrated model while potentially addressing the market failures that Brazil sought to overcome.
The case for justification: How Brazil might defend Pix
Should the USTR’s probe lead to WTO litigation, Brazil could raise several defenses, beginning with the question of timing. UPI launched in 2016 but achieved scale only around 2020-2022—after Brazil had committed to Pix’s architecture. At the time Pix was designed (2018-2020), UPI’s viability at scale was not yet proven. Brazil could argue that WTO panels should assess necessity based on information reasonably available when the measure was adopted, not on ex-post evidence.
Brazil’s institutional context further differentiates its situation. The payment market wasn’t simply underserved; it was dominated by entrenched oligopolies. A few large banks controlled both infrastructure and customer relationships, with limited regulatory capacity to discipline them effectively. Brazil might argue that alternatives such as price regulation or competition enforcement weren’t realistically available given institutional constraints. This argument challenges the assumption—often implicit in WTO necessity analysis—that private-market regulation is a feasible substitute. Institutional feasibility, not just theoretical availability, may matter.
Brazil could also argue that the legal restriction of SPI to Pix was a technical necessity. Opening SPI to multiple schemes could compromise settlement integrity, introduce operational risks, and create coordination burden. A single scheme simplifies standards, dispute resolution, and operational responsibility: all essential for processing billions of transactions monthly. Whether these claims withstand scrutiny, given UPI’s success, would be central.
Similarly, mandatory participation and zero pricing may be characterized as coordination mechanisms necessary to overcome network effects and entrenched incumbents. Voluntary participation might have resulted in fragmented adoption, with dominant banks delaying implementation to protect fees. Universal participation and free pricing ensured immediate network completeness, arguably essential to achieve rapid financial inclusion.
Brazil may also argue that public payment systems have distinct data-governance implications. Private payment systems would concentrate vast amounts of data, raising surveillance-capitalism concerns. Public control over the application layer allows democratic oversight and enforceable privacy protections. In this view, infrastructure and applications cannot be cleanly separated.
GATS recognizes members’ rights to regulate services to meet public-policy goals. Brazil might argue that monopolized public infrastructure allowed the country to leapfrog from a cash-dominated economy to a digital leader, and that temporary departures from more competitive models should be permitted for development. Network effects reinforce this argument: because payment markets tend toward concentration, public monopoly might be deployed to overcome early-stage coordination failures, even if only temporarily. Private providers wouldn’t individually invest in infrastructure without assurances that competitors and merchants would adopt simultaneously.
Finally, Brazil might analogize Pix to a subsidy delivered through direct provision rather than funding. GATS Article XV recognizes the role of subsidies for development purposes. If direct provision achieves inclusion more efficiently than subsidizing private providers does, it should not face more stringent scrutiny.
Key analytical tensions
Several tensions complicate the assessment under WTO rules. Evidence suggests that Pix increases the use of other payment methods: payment slips, bank wires, and card acceptance all grew following Pix adoption. This complementarity paradox prompts competing interpretations. On one hand, Pix isn’t substituting for other methods but expanding the digital ecosystem, suggesting less market distortion than feared. On the other hand, if Pix alone achieves the goal of teaching people to use digital payments, controlling the application layer may not be necessary—infrastructure monopoly might suffice to enable competitive applications to capture the same benefits.
Additionally, the competitive-effects analysis depends critically on market definition. If the market is “instant payment services,” Pix is effectively dominant. If the market is “digital payments broadly,” dominance is lower—around 52%. If the market is “all payment methods including cash,” dominance is lower still—approximately 51%. WTO panels must define the relevant market to assess foreclosure, and different definitions yield different conclusions about competitive harm.
While the legal restriction of SPI to Pix is clear, the critical WTO question is whether this closure constitutes a permissible technical necessity or instead an arbitrary monopoly grant. Brazil would argue the former—that system integrity, operational reliability, and unified standards require single-scheme infrastructure. Critics would counter that UPI demonstrates multi-scheme infrastructure is technically feasible. This tension between technical claims and comparative evidence is central to the necessity analysis. Similarly, the economic barriers from mandatory participation and zero pricing could be characterized either as anticompetitive design or as necessary coordination mechanisms given oligopolistic market structure and network effects. Determining whether these features constitute legitimate tools or disproportionate restrictions will be central to WTO analysis.
However, even if a WTO panel found that Pix violates GATS obligations, the practical remedy is unclear. With 177 million users, extensive business integrations, and strong network effects, separating Pix’s joined-up infrastructure and application layers would be disruptive. WTO necessity analysis considers whether less restrictive alternatives are “reasonably available,” which includes practical feasibility. An alternative that would work well if adopted from the start but would require a prohibitively disruptive transition may not qualify as “reasonably available” at the dispute stage. This creates an asymmetry: it may be easy to adopt closed infrastructure first but hard to transition to private competition afterwards, raising questions about whether countries can lock themselves into more restrictive models by early adoption choices. A more likely outcome might be prospective opening only, where new schemes can access SPI but Pix continues operating.
Finally, the conventional analytical framework distinguishes infrastructure layers (where monopoly may be efficient) from application layers (where competition creates value). But in practice, this boundary may be less clear. User-experience standardization spans both layers. Data-governance concerns span both layers. Network effects and coordination problems span both layers. Technical integrity and operational reliability may require unified control. If the infrastructure/application distinction is less clean than assumed, the proportionality analysis becomes more complex. Brazil’s dual barriers might be reframed not as cumulative restrictions but as integrated features of a unified technical architecture that cannot be cleanly separated without compromising system objectives.
Restructuring options and challenges
A potential approach to addressing USTR’s concerns could be restructuring Pix toward UPI-style decentralized architecture. Two routes are possible. The first would transform Pix from a payment scheme into an open-API protocol through which payment-service providers (PSPs) connect to SPI, enabling the emergence of competition at the application level. This solution would require opening access to SPI and creating standardized APIs for competing schemes, replicating UPI’s core design.
The second approach would keep Pix as a payment scheme while allowing directfront-end competition to develop. This path would also require opening access to SPI settlement, in addition to eliminating Pix’s mandatory participation and allowing banks and nonbank PSPs to choose whether to support Pix, competing schemes, or both. To preserve universality, the Central Bank could require all PSPs to offer free, interoperable basic functionality alongside premium services, ensuring that financial inclusion gains are maintained even as competition potentially increases.
However, these restructuring paths face significant challenges, as noted earlier. Transitioning to a more competitive model could cause disruption. Businesses would need to modify systems, users would face changed interfaces, and network effects might initially decline. PSPs would face the burden of supporting multiple systems simultaneously during transition.
Opening SPI to competing schemes isn’t merely a policy choice and could require substantial technical architecture changes. Multi-scheme settlement may require more complex protocols, dispute resolution mechanisms, and operational coordination than single-scheme systems. These technical challenges could partially justify Brazil’s current architecture as necessary for reliability.
Private firms that previously competed against Pix might be reluctant to build on infrastructure they perceive as initially unfairly advantaged, creating adoption challenges. Foreign payment schemes might demand compensation for lost market opportunities during Pix’s exclusive period.
The transition should be managed so as to retain Pix’s social benefits—rapid financial inclusion, cost reduction, universal access. Maintaining interoperability is key. Fragmenting into competing schemes could reduce network benefits and slow further inclusion. The integrated model may have enabled faster adoption precisely because of the coordination it provided.
Broader implications: The future of digital infrastructure
The USTR’s Pix investigation represents a critical test for how trade policy adapts to digital markets where traditional public-private boundaries blur. Current GATS rules drafted in the 1990s provide insufficient guidance for digital-infrastructure markets. GATS assumes clear public-private distinctions that become faint when governments provide exclusive infrastructure enabling competitive private applications. Modern trade rules should recognize legitimate public infrastructure monopolies while requiring open access for private competitors in adjacent markets.
The Pix case particularly highlights the challenge of dual barriers—when legal closure of infrastructure combines with economic foreclosure through regulatory design. Trade rules need more explicit frameworks for analyzing these cumulative effects and determining when such combinations constitute disproportionate restrictions versus integrated technical architectures.
The case raises questions about whether developing countries should have greater flexibility to use public digital infrastructure for rapid economic transformation, even if it reduces competition on certain markets. The WTO framework’s balance between market access and development needs may require updating for digital services. Particularly relevant is whether countries should be judged on alternatives available at policy adoption time versus dispute time, and whether transition costs from integrated to competitive models should factor into the necessity analysis.
Pix achieves financial inclusion and enhances consumer welfare through greatly reduced transaction costs. Yet long-term net welfare effects remain uncertain: has its success created innovation barriers or laid the groundwork for broader financial services competition? Evidence of complementarity with other payment and financial services suggests more nuanced effects than simple foreclosure.
As central banks explore digital currencies (CBDCs) and the U.S. retrenches, the Pix precedent will influence how new money forms, particularly central bank-issued ones, interact with existing markets, raising similar questions about the boundaries between public policy and competitive distortion. If Pix faces WTO challenges, CBDCs may face even greater scrutiny, especially if they involve infrastructure- and application-layer control via restrictive participation and pricing rules.
Beyond payments, the Pix case illuminates broader questions about when governments should provide digital platform infrastructure and when they should regulate private platforms. The infrastructure/application distinction may provide a useful framework for other digital services, though the Pix controversy indicates this boundary may be less clear in practice than in theory.
Author Disclosure: Jeff Alvares is senior counsel at the Central Bank of Brazil. The views expressed are his own and do not represent those of the Central Bank of Brazil or the Brazilian government.
Articles represent the opinions of their writers, not necessarily those of the University of Chicago, the Booth School of Business, or its faculty.
Subscribe here for ProMarket’s weekly newsletter, Special Interest, to stay up to date on ProMarket’s coverage of the political economy and other content from the Stigler Center.
