Security, Privacy and Regulatory Governance in Digital Payments. A Systematic Review and Cross-Architecture Framework for PIX, Blockchain, and Central Bank Digital Currencies

Technical Summary: Security, Privacy, and Regulatory Governance in Digital Payments

Problem Statement
The digital payment landscape is characterized by the coexistence of three distinct paradigms: centralized instant payment systems (exemplified by Brazil's PIX), public and permissioned blockchain networks, and Central Bank Digital Currencies (CBDCs). While these systems operate concurrently, governance frameworks for security, privacy, and regulatory compliance (specifically Anti-Money Laundering and Counter-Terrorist Financing, or AML/CFT) have developed in isolation. The central problem identified is the unresolved tension between the need for transparency and traceability to satisfy regulatory obligations and the requirement for privacy protection. Current scholarship lacks a unified instrument to compare how these architectures manage the trade-off between security, privacy, and compliance, particularly regarding the "privacy paradox" in CBDCs and the structural compliance gaps in decentralized systems.

Methodology
This study employs a systematic review compliant with PRISMA 2020 guidelines, analyzing eighteen peer-reviewed studies published between 2014 and 2025. The corpus includes technical, empirical, legal, and policy-focused literature covering PIX, blockchains, and CBDCs. Sources were drawn from Scopus, Web of Science, IEEE Xplore, ACM Digital Library, RePEc, and HeinOnline.

To synthesize these findings, the authors constructed the Digital Payment Governance Framework (DPGF). This framework evaluates systems across three independent axes, each scored on a scale of 1 to 5:

1. Locus of Control (LoC): Measures the distribution of rule-making and enforcement authority (ranging from a single central authority to a diffuse protocol community).
2. Privacy Spectrum (PS): Measures the disclosure of transactional and identity data (ranging from full anonymity to full transparency).
3. Compliance Architecture (CA): Measures the point at which compliance obligations are embedded (ranging from "regulation by design" to "post hoc enforcement").

The study also appraised the quality of the included studies using the Mixed Methods Appraisal Tool for empirical/technical works and doctrinal coherence for legal/policy works.

Key Contributions and Results

• The Digital Payment Governance Framework (DPGF): The primary contribution is the DPGF, which allows for the comparative positioning of disparate payment architectures. The study identifies three distinct governance models based on the scoring of the three axes:

  • Regulatory Clarity Model (PIX): Scored 1 (LoC), 4 (PS), 1 (CA). The Central Bank of Brazil holds centralized rule-making and enforcement. Compliance is embedded in participation ("regulation by design"), requiring real-time monitoring by providers. Privacy is protected via data protection laws rather than cryptographic protocol mechanisms, resulting in high linkability for regulators but lower default privacy for users.
  • Autonomy Model (Public Blockchains): Scored 5 (LoC), 2 (PS), 5 (CA). Protocol changes require community consensus. Compliance is largely absent at the protocol layer and applied only at regulated boundaries (e.g., exchanges/VASPs) via post hoc enforcement. Pseudonymity is contingent on address linkage, and metadata often allows for re-identification.
  • Hybrid Governance Model (CBDCs & Permissioned Blockchains): Scored 2 (LoC), 2 (PS), 2 (CA). These systems feature a central authority retaining rule-making power while execution is distributed among licensed intermediaries. They aim for a balance, utilizing selective disclosure and tiered anonymity (e.g., anonymous small transactions vs. transparent large ones).

• Technical Findings:

  • Privacy-Enhancing Techniques (PETs): Zero-knowledge proofs (specifically zk-SNARKs) and homomorphic encryption are the most widely adopted techniques in the literature. However, the review notes significant scalability and metadata limitations. For instance, homomorphic encryption is computationally expensive (orders of magnitude slower than plaintext), and timestamp metadata in ring signatures remains a vulnerability for re-identification.
  • Compliance Mechanisms: "Regulation by design" (embedding compliance in the architecture) is shown to lower per-transaction compliance costs compared to post hoc enforcement but increases the cost and complexity of regulatory updates.
  • Security Threats: The study highlights specific threats, including mass surveillance risks in CBDCs (due to direct central bank access to data) and the "offline payment problem," where offline capabilities create audit windows that reduce privacy.

• Propositions: The framework generates five propositions for future research, suggesting that:

  1. Lower control scores correlate with faster regulatory response but higher surveillance.
  2. Intermediate positions on the privacy spectrum (selective anonymity) may sustain trust better than extremes.
  3. "Regulation by design" reduces transaction costs but increases regulatory change costs.
  4. Hybrid architectures may offer superior composite governance outcomes (a hypothesis for testing).
  5. Advanced PETs reduce linkability but increase latency, potentially constraining national-scale feasibility.

Significance and Claims
The paper claims that the DPGF provides a necessary instrument for comparing governance arrangements across centralized, decentralized, and hybrid architectures, moving beyond single-axis analyses of centralization or privacy. It argues that the field has converged on hybrid architectures as a design preference, though this is currently supported by advocacy rather than measured comparative outcomes.

The study emphasizes that selective disclosure embedded by design at the protocol level represents the most supported direction for next-generation payment governance. To validate this, the authors propose a set of 13 cross-system indicators (e.g., fraud rates, incident MTTR, linkability index, regulatory response time) to operationalize the propositions and measure governance effectiveness.

The authors maintain a modest stance, acknowledging that the empirical base is thin (few studies use real transaction data) and that the propositions are hypotheses for future testing rather than established results. They note that the comparison is analytical rather than a quantitative meta-analysis due to the heterogeneity of the 18-study corpus. The paper concludes that no single architecture currently resolves the three-way tension between oversight, privacy, and compliance efficiency; rather, each model offers distinct trade-offs that must be evaluated against specific contextual requirements.