When a state transfers its public services to a digital platform, it enters into a contract whose clauses it has not always read. International IDEA published a report in 2026 examining whether these infrastructures serve citizens or the states deploying them. The answer depends less on technology than on governance, a domain where governance vulnerabilities are identified without their prevalence being quantified.
The essentials
- Public digital infrastructures—identity, payment, data exchange—organize access to rights as much as they concentrate exclusionary power.
- International IDEA identifies three fundamental building blocks (digital identity, payments, data exchange) and recommends for each independent supervisory authorities with real resources, effective remedies, and preventive competition rules.
- Without open and contestable architecture, states risk double dependence: on private suppliers who hold the code, and on automated exclusion logics that nobody monitors.
- Models exist—certification of identity systems, independent audit committees, portability obligations—but remain rare outside a handful of pioneering countries.
- Current architectural choices can have lasting consequences for freedoms and the capacity for public action.
Three building blocks, the same capture risk
The basic idea is appealing. Digitize administration, reduce queues, distribute social benefits without bureaucratic intermediaries, identify each citizen by a unique number. India did it with Aadhaar, Estonia has been doing it for twenty years, dozens of countries are taking it up. The efficiency is real. So are the risks.
The report covers the three common functions of DPI: digital identity verification, digital payments, and secure data exchange. Each raises the same structural problem. The moment a state delegates one of these building blocks to a private supplier without retaining technical and legal control, it runs the risk of transferring part of its operational sovereignty. The supplier can become a de facto arbiter in access to social aid, health services, and official documents.
Supplier lock-in is a known phenomenon in the private sector. In the public sector, its consequences are of a different nature. A company can change accounting software at the cost of time and money. A state wanting to change the system on which the identities of forty million citizens rest is playing a different game. Migration takes years, and during those years, the initial supplier holds the keys.
Asia today concentrates the most ambitious deployments, and therefore the most revealing trade-offs. The Philippines and Indonesia have recently developed large-scale digital identity components, but India’s Aadhaar and Bangladesh’s NID programs date back to 2010. Some have opted for open architectures with public standards. Others have signed with private consortiums whose technical specifications remain proprietary. The difference in governance between these two trajectories already sketches two very different futures.
Automated exclusion, the real political risk
A poorly governed digital infrastructure does not merely create technical dependence. It can exclude. And automated exclusion has a particular feature: it is invisible until you are a victim of it.
When an algorithm decides that a file is incomplete, that an identity number does not match, that a transaction is suspicious, the decision can occur without explanation, sometimes without accessible recourse. The official who once examined a file at an open counter could be challenged, questioned, persuaded. The automated system responds with an error code.
International IDEA recommends that complaint and recourse mechanisms, supported by administrative and judicial oversight, be accessible to all. Nuance matters. In many countries that have provided online recourse forms, the success rate of these recourses remains unknown due to lack of public statistics. Formal access to the right to contest does not guarantee the actual capacity to do so.
This question is far from theoretical. Populations have been deprived of social benefits due to errors in identity databases uncorrected through official channels. Some migrants have had their rights suspended due to identity document formats not recognized by certain systems. These cases, documented in several countries in South and Southeast Asia, illustrate what International IDEA calls the “exclusionary power” of public digital infrastructures, a power as real as the access power they are supposed to organize.
The report identifies a fundamental asymmetry here. The benefits of digitization—speed, cost reduction, coverage expansion—are visible and politically valuable. The costs of exclusion are diffuse, poorly measured, and primarily affect populations least able to assert their rights. This mechanism echoes other dynamics of public resource allocation, where perception biases distort political priorities to the detriment of the most vulnerable.
Common points in models that work
Counter-examples exist, and they deserve close examination rather than citation as totems.
Estonia is the most documented case. Its digital identity system relies on open standards, a multi-level control architecture, and legislation guaranteeing each citizen access to data the state holds on them, as well as traceability of every consultation. An official who consults a file without legitimate reason leaves a trace. This principle of data transparency reverses the usual logic: the state is transparent to the citizen, not just the citizen to the state.
The other distinctive element of the Estonian model is its interoperability architecture, a technical layer called X-Road that allows different administrative systems to communicate without centralizing all data in a single database. Data decentralization reduces both the risks of massive breach and the temptation toward comprehensive surveillance.
These architectural choices were not inevitable. They resulted from political decisions made in the early 2000s, when Estonia was building its digital state on the ruins of Soviet administration. Urgency could have pushed toward the fastest solution. Estonian decision-makers chose the most robust one.
India offers another, more ambiguous case study. Aadhaar enrolled over one billion two hundred million people and genuinely extended service access to populations that had no identity documents. Direct transfers to beneficiaries reduced certain forms of corruption. But the system also led to exclusions—people unable to access food rations due to limitations of biometric sensors with damaged fingerprints. And Aadhaar governance long lacked an independent supervisory authority with real powers.
The Estonia-India comparison designates neither winner nor loser. It isolates the variables that make the difference: open or proprietary standards, independent control or not, effective or formal recourse, decentralized data or central database.
Public procurement as a revealer of actual dependence
The question of digital sovereignty is not reducible to technical architecture. It also passes through contracts. And public contracts in the digital sector have a structural particularity: they are often signed on short cycles but create long dependencies.
A state purchasing a digital identity system from a private supplier generally signs a contract for five to seven years. But after three years, the data is in the system, agents are trained on the interface, administrative procedures are rewritten around the supplier’s specifications. Changing becomes more costly than staying. The initial market, even if put out to competitive bidding, can produce de facto concentration.
International IDEA recommends competition rules and active prevention of lock-in. It supports the use of open assets, standards, and specifications as well as prevention of vendor lock-in. These requirements exist in best practices for digital public procurement, but their application remains uneven. As analysis of public purchasing shows, administrations tend to optimize for immediate cost rather than total cost of ownership, which includes future dependence.
The problem is aggravated by market concentration. A restricted number of companies supply a significant share of public digital infrastructures across several regions. This concentration creates two distinct risks. The first is economic: an oligopoly of suppliers can impose its prices and conditions on contract renewals. The second is geopolitical: critical infrastructures of sovereign states depend on companies whose headquarters, jurisdiction, and legal obligations are elsewhere.
Two trajectories for the next ten years
The question posed by International IDEA is not whether states will digitize their administrations. That decision has been made everywhere. The question is under what architecture.
The report proposes principles and exploration questions, and the signals to distinguish them are already visible today.
The report notably proposes open standards, well-resourced independent control institutions, accessible recourse, and mechanisms preventing vendor lock-in. This trajectory requires higher initial investments in time, legal expertise, and technical capacity of the administration. It produces more robust systems in the long term, less exposed to capture and arbitrary exclusion.
The second trajectory presents risks related to power concentration and weak control institutions. Infrastructures deployed rapidly, often with help from private suppliers providing turnkey solutions; the absence of independent governance or recourse must be established case by case. This trajectory is politically attractive in the short term: it produces visible results quickly, it reduces apparent administrative complexity. Its costs appear later, when lock-in is complete, when exclusions accumulate without recourse, when changing suppliers becomes politically impossible.
The signals that allow distinguishing these two trajectories are concrete. The number of successful appeals against automated administration decisions, an indicator that very few countries publish today. The degree of interoperability of identity systems in regional agreements—the European Union is advancing on this with the European digital identity wallet, Southeast Asia much more slowly. Supplier concentration in digital public markets, an indicator that national competition authorities could follow but generally do not.
The report rather emphasizes evaluations, prior verifications, and independent, transparent, and continuous audits. Models exist in other sectors—certification of medical devices, aircraft audits. Applying similar logic to public digital infrastructures would require defining auditable compliance criteria: that data can be migrated, that automated decisions can be contested, that code can be inspected. Few governments have taken this step.
Interoperability between national systems is the other structural lever. States adopting common standards create conditions for a more competitive supplier market. Public and shared specifications can reduce vendor lock-in risk, without eliminating it alone. Interoperability is also a condition for regional cooperation on rights: a refugee whose status is recognized in one country should be able to assert their rights in another without starting over. These ambitions remain largely programmatic, but a few bilateral agreements in Southeast Asia show it is feasible.
Necessary powers of control authorities
International IDEA’s central recommendation on control authorities deserves development, because it touches on a point where intention and reality often diverge.
Many countries have created personal data protection authorities. Some have limited powers to oversee public digital infrastructures as a whole, particularly regarding system governance, decision algorithm bias, and actual accessibility of recourse.
The report envisages independent control bodies and continuous evaluations or audits. This is a preventive posture, closer to what banking regulators do than what data protection authorities do in most countries.
This ambition presupposes human resources that few national administrations have built up. Auditing a large-scale digital identity system requires expertise in cryptography, system architecture, administrative law, and social sciences to assess exclusion biases. These multidisciplinary teams are expensive and in direct competition with the private sector for recruitment. This is precisely why International IDEA insists on “real resources”: an underfunded control authority is a façade authority.
The concentration of technical expertise in a few major metropolitan areas makes this challenge even more acute for middle-income countries. AI and digital systems expertise concentrates geographically in a handful of hubs, meaning that national administrations of less well-resourced countries depend even more on external suppliers and have even less capacity to audit them.
Governing architectures before they govern in their place
The question posed by International IDEA’s report may seem technical. It is fundamentally political.
Current DPI choices can have lasting consequences for relations between citizens, state, and digital systems. These choices are theoretically reversible; they are very difficult to undo in practice, because migration costs increase over time and reinforce dependence.
The good news is that models combining efficiency and democratic governance exist. They are not reserved for wealthy countries: some West African and Latin American countries have adopted open architectures with accessible recourse, sometimes starting from a blank slate rather than renovating inherited systems. The diffusion of DPI models also depends on institutional capacities, sociopolitical context, initiative maturity, and trust.
Decision-makers—elected officials, senior civil servants, and regulators—must integrate governance into system design, not add it afterward. A public digital infrastructure lacking an independent control mechanism is an incomplete infrastructure. The report recommends user-centered design and accessible complaint and remedy mechanisms.



