Data for Governance: Turning Public Data into National Capacity
Why better governance depends on institutions that can see, coordinate, act and learn
Collecting more data, establishing national registers or launching digital platforms does not automatically make governments more capable. These interventions may improve access to information, reduce administrative friction and strengthen particular processes but they do not, on their own, create the institutional capacity required to govern more effectively.
Public data becomes national capacity when government can use it to develop a credible and. current view of what is happening, recognise the same people, businesses and activities across different institutions, coordinate the agencies responsible, make decisions within clear legal and budgetary authority, deliver the intended service, correct errors and learn from the. outcomes that follow.
Without this operating chain, governments may accumulate more data without becoming materially better at making decisions. They may collect more information about citizens, businesses, transactions and public programmes without becoming significantly better at identifying needs, responding to change or delivering results.
This gap becomes easier to understand when we examine how public programmes function in practice. Nigeria’s National Social Register and the National Single-Window, Kenya’s social-protection and health-insurance systems, Togo’s Novissi programme, Ghana’s management of unfinished public projects and Sierra Leone’s medicine-allocation initiative each reveal a different point at which the connection between information, institutional authority and delivery can either strengthen or begin to break.
The purpose is not to compare countries or reduce complex programmes to simple verdicts of success or failure. It is to examine what these cases reveal about the conditions under which public data improves government action, where the delivery chain remains fragile and what must change for information to become a dependable part of national capacity.
From Registration to Delivery: Where National Capacity is Tested
Nigeria’s National Social Register provides a useful starting point because it shows the difference between identifying a household as vulnerable and successfully delivering support to it.
The register is built through geographic and community-based processes used to identify poor and vulnerable households. Information is collected digitally, assessed against defined poverty indicators and used to determine potential eligibility, assessed against defined poverty indicators and used to determine potential eligibility for social interventions. Overtime, the register has grown to cover tens of millions of individuals.
But being ioncluded in the register does not mean that support has been delivered. It means that a household has entered the administrative process. Before a cash transfer can reach a household, several additional steps may have to be linked to a valid national identity record. Personal information must match across systems. Payment details must be confirmed. The household must be approved for the specific programme, funding must be released and the transaction must be completed.
The journey from identification to delivery therefore passes through several distinct stages:
Identification → Registration → Revalidation → Identity verification → Payment validation → Approval → Disbursement → Confirmation
Every stage answers a different administrative question and responsibility for completing it may sit with a different institution.
A household may appear in the register but still require revalidation. It may be revalidated but have an unresolved identity mismatch. It may pass identity verification but lack valid payment details. It may be approved bbut not yet paid. It may received the first installment without completing the full payment cycle.
This helps explain why different institutions may publish different figures for the same programme. One figure may refer to households captured in the register. Another may count beneficiaries whose identities have been verified. A third. may show households that have received. at lest one payment, while another reflects those that have received every approved installment. The figures may all be accurate while describing different. points in the delivery process.
The expansion of the targeted cash transfers to 15 million vulnerable households progressed more slowly than intended partly because records in. thye. social register has to be matched with biometric information held within the national identity system. The delay exposed a practical limitation that is easy to overlook when programme scale is discussed through headline numbers: a large register does not mean that every record is immediately useable. for delivery.
The register can help government identify households that may require support. National capacity is demonstrated by whether government can validate that need, reconcile information across information, resolve inconsistencies, release the required resources and complete the transfer.
The distinction matters because public reporting often concentrates on the stages that are easiest to count. Governments can report how many people have been identified, registered, verified or approved. For the citizen, the measure that actually matters is whether the support actually arrived.
The inclusion of a household in a register has little practical value if it does not translate into timely support. The real test is whether government can convert recognised need into actual delivery.
When More Public Data does not Lead to Better Decisions
The wider challenge is not that African governments collect too little information. In many cases, they collect a great deal of it, often from the same citizens and businesses several times over.
People provide their names, addresses, identity documents, household details, employment records and business information to different public institutions. Health facilities submit reports on service delivery, medicines and patient activity. Schools report enrolment and attendance. Customs systems record trade transactions. Tax authorities maintain information on registered businesses. Local governments collect data on licenses, properties and public projects, while social programmes generate records on beneficiaries, approvals and payments.
Yet, even with this volume of information, governments may still lack a clear and reliable view of what is happening across their own programmes.
Which households received every payment for which they were approved?Which families have become vulnerable since the social register was last updated? Which public projects have been funded but remain incomplete?Which clinics have adequate medicines and which are experiencing shortages? Which traders have completed the requirements of one agency but remain delayed by another?
The information needed to answer these questions may already exist, but it is often spread across institutions that collect it for different purposes, update it at different intervals and define the same people, activities and outcomes in different ways. Each institution holds a portion of the picture, but no single institution necessarily sees the whole.
Official statistics remain essential because they provide a broader and more representative view of the population, economy and society. They establish national baselines, reveal structural patterns and make it possible to measure change over time. Administrative data serves a different but equally important purpose. It captures the routine activity of government through registrations, applications, payments, inspections, transactions and service delivery. These two sources are complementary, but neither is sufficient on its own.
Administrative data reflects only the people and activities visible to the institution collecting it. A household that does not appear in a social register has not necessarily escaped poverty. A business without a complete tax history is not necessarily inactive. A citizen whose identity record is incomplete has not ceased to exist simply because the system cannot recognise them.
Official surveys and statistical releases, on the other hand, may not provide the frequency or level of detail required to manage the daily operations of a public programme. They can show the scale and distribution of a problem, but they may not reveal which household missed a payment this month, which clinic ran out of medicine this week or where a particular approval process has stalled.
State capacity grows when governments can bring the credibility and broad coverage of official statistics together with the timeliness and operational detail of administrative data, then connect both to the decisions that institutions are responsible for making.
The problem, therefore, is not simply a shortage of data. It is that public information is often fragmented across the same institutional boundaries that divide authority, budgets and responsibility. Governments may know more about citizens, businesses and public programmes than ever before while still lacking a coherent view of what is happening and what must happen next.
The Fragmented System of Government
Government is often discussed as though its institutions operate within oner coherent. structure, guided by a single budget, database and chain of authority. In practice, it is a network of ministries, agencies, regulators, regional and local administrations, public enterprises, contractors, payment providers, development partners and frontline institutions. Each operating under its own mandate, leadership structure, budget cycle, technology environment and reporting obligations.
This fragmentation becomes visible in the way public programmes are delivered. A social-protection programme, for example, may depend on local structures to identify vulnerable households, a national office to manage the register, an identity authority to verify beneficiaries, a ministry to define policy, a finance authority to release funds, a payment provider to complete transfers and a separate mechanism to resolve complaints.
Each institution sees a different part of the process. The identity authority can confirm whether an individual has a valid identity record. The social register can show whether a household has been captured. The programme office can determine whether the household meets the eligibility criteria. The payment platform can confirm whether a transfer was initiated or completed. The household knows whether the support actually arrived.
No single system necessarily captures the full delivery picture. The same problem appears in the management of public infrastructure.
In 2026, Ghana’s District Assemblies Common Fund reported a backlog of unfinished projects valued at close to GH¢8 billion and committed part of its resources to completing legacy projects. Funding constraints and the abandonment of projects across successive administrations were identified as major contributors to the backlog. [6]
Behind each unfinished project are several administrative records. The budget system records what was approved. The procurement system records the tender and contract. The finance system records releases and payments. The implementing agency reports progress. Inspectors may confirm physical completion, while the local community knows whether the road, clinic, classroom or water facility is actually usable.
Unless these records are linked through a consistent project identity, government may know what was budgeted, contracted and paid without having a reliable view of what was actually completed, brought into use or maintained over time.
This is not data integration gap. It reflects the way responsibility is divided across government. The ministry that controls the budget may not oversee implementation. The agency managing the contract may depend on another institution for the release of funds. The local authority expected to operate the completed facility may have had little involvement in its design or construction. When administrations change, projects may be inherited with incomplete records, limited political ownership and no clear funding pathway for completion.
The movement of government data is ultimately constrained by how authority is divided across institutions.
When institutional mandates overlap, accountability becomes harder to trace. Where agencies compete for funding, relevance and influence, control of data can become a source of institutional leverage. In that environment, no single organisation may feel fully responsible for maintaining a complete and reliable record of the programme from approval through to delivery.
Interoperability therefore requires more than the technical exchange of data between systems. Institutions must agree on common definitions, establish which records are authoritative, determine who may update them, and assign clear responsibility for action and accountability when delivery falls short.
Without that alignment, government may hold many accurate pieces of information yet still lack a single, dependable view of the outcome it is expected to deliver.
Visibility: When Public Data Falls Behind Reality
The first requirement of national capacity is visibility. Government must be able to identify the people, places and activities that require attention, but that visibility depends on more than the volume of data available. It requires both coverage and currency.
Coverage determines how much of the population or activity a system can observe. Currency determines whether the information still reflects present conditions.
Social registers illustrate this distinction clearly. Household vulnerability is not fixed. A family may lose income, relocate, experience illness, suffer a climate-related shock, close a business or change in composition after its initial assessment. At the same time, another household may remain on a register even though its circumstances have improved.
A large register can extend the reach of a programme, but it still reflects household conditions at a particular point in time. Its usefulness depends on whether records can be updated and whether significant changes in people’s circumstances can enter the system early enough to influence eligibility and delivery.
Togo’s Novissi programme provides a useful example of how government can create visibility quickly when existing administrative systems are incomplete.
The programme was introduced during the pandemic to deliver emergency assistance through mobile money. Available identity information, mobile-phone data, survey records, geospatial information and predictive methods were combined to identify and prioritise vulnerable people. This enabled support to reach poor communities that would have been difficult to assess through conventional field-based processes during the public-health restrictions. [4]
Novissi demonstrated the value of organising available data around a clearly defined public decision. Government needed to identify people whose livelihoods had been disrupted and deliver support quickly. The programme used the information and payment channels already available to respond to that need.
The programme revealed the limitations of the underlying infrastructure.
Eligibility depended partly on whether people could be identified through voter and mobile-phone records, which made those outside these systems harder to reach. Payments through mobile channels created further access conditions because beneficiaries needed a phone and an active SIM, even though phone ownership and usage were uneven, particularly among women, younger people and rural households.
The programme demonstrated both institutional capability and structural constraint. By combining available data with existing payment channels, government was able to respond quickly, but the experience also exposed the limits of relying on temporary arrangements as the permanent basis for sustained public delivery.
Togo’s subsequent social-protection reforms have sought to address those gaps by investing in unique identification, a dynamic social registry, a multi-programme payment platform and stronger monitoring and evaluation. [4]
The significance lies in what followed the emergency response. The programme began to strengthen state capacity when its lessons, systems and institutional capabilities were carried forward into permanent public infrastructure, clearer institutional arrangements and more reliable delivery mechanisms.
Verification: Protecting Programme Integrity Without Excluding Beneficiaries
Visibility must be matched by the ability to verify that the people and organisations appearing in public records are genuine, that the information held about them is accurate and that programme benefits are reaching those for whom they were intended.
Verification is therefore central to programme integrity. It helps identify duplicate records, remove deceased beneficiaries, reduce fraud and protect limited public resources from being diverted away from eligible recipients.
The challenge begins when a programme depends on a single method of verification and offers few practical alternatives when that method fails. In those circumstances, a control introduced to strengthen the programme can become a barrier to access for legitimate beneficiaries.
Kenya’s Inua Jamii programme brings this tension into focus. In 2026, biometric proof-of-life verification was introduced for beneficiaries of cash-transfer programmes supporting older people and persons with severe disabilities. Fingerprint and facial-recognition technologies were used to confirm identity and continued eligibility as part of wider efforts to update the beneficiary register and strengthen programme accountability. [2]
Social-protection programmes cannot depend indefinitely on records that are neither regularly updated nor independently verified. Institutional capacity is tested by whether the programme can resolve cases in which the standard verification process fails.
Older beneficiaries may have fingerprints that are difficult to capture, while persons with disabilities may face physical barriers in reaching enrolment centres. For those in rural areas, long travel distances, weak connectivity and limited administrative support can make verification even more difficult. Differences in names, dates of birth and other personal details across government systems may create further mismatches that beneficiaries struggle to correct.
In these circumstances, a technically robust control can still result in the exclusion of legitimate beneficiaries.
The real question is whether the programme provides workable alternatives when the standard verification process fails. Assisted enrolment, alternative verification methods, accessible record correction, human review and a clear appeals process are necessary to protect programme integrity without allowing technical or administrative difficulties to become automatic grounds for exclusion.
Effective verification requires more than identifying inconsistencies in a record. It must also provide a reliable process for investigating them, correcting errors and reaching a fair, accountable decision.
The same issue is evident in Nigeria’s cash-transfer programme. Linking beneficiaries to national identity records can strengthen traceability, but it also means that delivery depends on the social register, identity system and payment process working together effectively.
A household may be genuinely vulnerable and correctly identified through community-based processes, yet still remain outside the payment system because the intended recipient has an incomplete identity record or because personal details differ across institutions.
Where identity, eligibility and payment systems are developed separately, the points at which they intersect can become points of failure. Verification contributes to state capacity only when those connections are supported by clear institutional responsibility, accessible processes for correcting errors and practical pathways through which eligible beneficiaries can complete verification and receive the support intended for them.
Decision Quality: When Automation Scales the Wrong Assumptions
Even after government has identified and verified a citizen, it must still interpret the available data and decide how that data should influence access, contributions or benefits.
Public programmes increasingly use formulas, statistical models and automated rules to estimate income, assess vulnerability, determine payment obligations and allocate support. These approaches can improve consistency and allow decisions to be processed at a scale that would be difficult to manage manually.
Yet automation does not eliminate judgement from public administration. It relocates that judgement into the choice of variables, the assumptions applied, the thresholds established and the outcomes the model is designed to produce.
Get Seghe Nwamaka Momodu’s stories in your inbox
Join Medium for free to get updates from this writer.
Kenya’s Social Health Insurance Regulations provide for means testing of households whose income is not derived from salaried employment. The assessment may use indicators such as housing conditions, access to basic services and household composition to estimate income and determine contributions. [3]
This responds to a genuine administrative challenge. Where much of the population earns income informally, irregularly or outside formal payroll systems, payslips and employment records cannot provide a complete measure of a household’s ability to pay.
The difficulty is that household characteristics do not always provide an accurate picture of current income or financial circumstances.
A family may own its home and still have very little disposable income. Indicators such as electricity access or the quality of a roof cannot fully capture the effects of debt, illness, unstable earnings or the number of dependants relying on the household’s income. Financial circumstances may also change more quickly than the data and assumptions used to estimate them.
A 2026 investigation into the means-testing system reported that it often overstated the incomes of poorer households while understating those of wealthier ones. It also raised concerns about the transparency of the model and the difficulty citizens faced in understanding or challenging the contributions assigned to them. [3]
The case draws an important distinction between administrative efficiency and decision quality. Administrative efficiency is concerned with processing cases quickly and applying rules consistently. Decision quality depends on whether the resulting classifications are accurate enough, explainable, responsive to changing circumstances and open to review.
A process can therefore become faster and more consistent without becoming more reliable.
An incorrect manual assessment may affect a single case. When the same flawed assumption is embedded in an automated model and applied across a large population, the error is reproduced systematically and at scale.
The use of models in government is not the problem. The issue is whether they are governed with the same seriousness as other instruments of public administration. The underlying data must be sufficiently representative, performance must be tested across different household and income patterns, and decisions affecting access to essential services must be explainable. Citizens must also have practical routes to correct inaccurate information, understand how a decision was reached and request a review.
Without these safeguards, automation may improve the speed and consistency of administration while reducing the fairness and legitimacy of the decisions it produces.
Coordination: Building a Shared View of the Same Transaction
National capacity also depends on coordination. Public data has limited value when the institutions responsible for a service cannot recognise the same citizen, business, project or transaction, or do not share a consistent understanding of its status.
Nigeria’s National Single Window illustrates this challenge because international trade requires several public and private institutions to act on the same transaction.
An importer or exporter may need to engage with customs authorities, port operators, regulatory agencies, shipping companies, banks and other service providers. Each institution may require permits, certificates, declarations, licenses or payment confirmations, often through separate systems. The result is a process in which the same information may be requested repeatedly, verified more than once or interpreted differently by each agency.
The National Single Window entered its first operational phase in March 2026 with a limited scope involving one port and one shipping line. Implementation is expected to expand gradually as additional institutions and processes are brought onto the platform. [5]
The phased approach reflects the complexity of coordinating institutions with different mandates, procedures, systems and regulatory responsibilities.
A single window is not simply a portal through which traders submit documents. It requires participating institutions to agree on common data requirements, recognise information that has already been verified by another agency, redesign approval processes and clarify which institution has decision-making authority at each stage.
A shared platform does not, by itself, create a shared operating model.
If every agency retains its existing procedures, requests the same information and repeats the same checks after receiving documents through the platform, the process has merely been moved online. The interface may be different, but the underlying delays and duplication remain.
Meaningful integration occurs when the same business identifier, shipment record, payment confirmation or regulatory approval is recognised consistently across all participating institutions. It also requires legal recognition of exchanged information, clear service standards, agreed responsibilities and a practical mechanism for resolving disagreements between agencies.
The National Single Window shows why interoperability is not simply a technical exercise. APIs can move data between systems, but they cannot make institutions trust one another’s records, accept decisions made elsewhere or relinquish unnecessary control over parts of the process. Those outcomes depend on institutional agreement, legal authority and the redesign of how government works.
Execution: Can the Public Data be converted into Delivery?
Accurate and connected data has adds little value to national capacity unless the government can translate the data into delivery.
The data may show that a health facility is running short of medicine, a public project has stalled, a household has not received an approved payment or a trade transaction remains delayed. Acting on that information may still depend on the release of funds, procurement approval, staff deployment, legal authority, logistics or cooperation from another institution.
Data can guide delivery, but it cannot create the institutional capabilities required to execute it.
Sierra Leone’s medicine-allocation initiative illustrates how public data can strengthen execution when it is connected to a clearly defined operational decision.
The programme addressed a practical question: how should limited supplies of essential medicines be distributed among health facilities facing different levels of demand and operating with uneven data quality?
Rather than attempting to integrate every available health dataset or automate the wider health system, the initiative used machine learning and optimisation to support an allocation decision that government was already responsible for making. It was implemented with the Ministry of Health and the National Medical Supplies Agency as a decision-support tool within an existing operational process.
Nationwide deployment was associated with an estimated 19 per cent increase in the consumption of allocated products in participating districts. The initiative was subsequently expanded to support services reaching an estimated two million women and children under five. [7]
The value of the programme does not lie in the suggestion that an algorithm resolved the wider challenges facing Sierra Leone’s health system. It lies in the discipline of designing the intervention around a specific public decision, placing it within the institutions responsible for acting on that decision and assessing it against a practical service-delivery outcome.
The limits of the intervention remain equally important. A stronger allocation recommendation does not ensure that medicines will reach the facilities where they are needed. Delivery still depends on procurement, warehousing, transport, inventory management and frontline implementation.
The model can improve how limited supplies are allocated, but effective delivery still depends on the wider operating system. Public data strengthens state capacity only when institutions can act on the decisions it supports. Strong analysis must therefore be matched by clear authority, adequate funding, reliable logistics, effective coordination and the discipline to carry decisions through to delivery.
Correction and Learning: Can Government Detect, Correct and Learn from Error?
No public dataset remains completely accurate indefinitely, as people relocate, household structures change, beneficiaries die, bank accounts become inactive and personal details are recorded incorrectly.
Information may differ across identity, payment and programme systems. Models may classify households inaccurately, frontline officials may make mistakes and connected platforms may fail to update one another as intended.
The credibility of data-driven government therefore depends not only on the quality of the original decision, but also on whether errors can be identified, challenged and corrected without imposing an unreasonable burden on the citizen.
A household should be able to update its circumstances without repeating the entire registration process. A beneficiary should not have to visit several institutions to resolve the same identity mismatch. An incorrect eligibility decision should be open to timely review, and a decision produced by a model should be capable of being reconsidered by an official with the authority to change it. Where an error is corrected in one system, that correction should also reach the other systems that rely on the same information.
Grievance mechanisms, audit trails, record-correction procedures, appeals and human review are therefore not administrative extras. They form part of the institutional machinery needed to preserve the accuracy, fairness and credibility of public decisions, while also creating the conditions for institutional learning.
Public programmes often measure activity more consistently than outcomes. Governments may know how many people were registered, how many applications were approved, how much funding was released or how many facilities were constructed. They may know far less about whether household vulnerability declined, access to healthcare improved, trade processes became faster or completed infrastructure remained functional.
Data supports institutional learning when evidence from implementation changes what government does next. This requires connecting programme inputs and activities to delivery and outcomes, then using the results to revise policies, processes, models and the allocation of resources.
It also requires institutions to acknowledge when an intervention is not working as intended, which is usually less attractive than announcing its launch. Yet the ability to identify failure, correct course and apply the lesson to future decisions is one of the clearest ways in which public data strengthens state capacity.
The Political Economy of Public Data
It would be convenient to assume that public-data fragmentation persists only because governments lack technology, funding or technical expertise.
In practice, better information is not always institutionally convenient.
It may reveal that public resources have been concentrated in politically favoured areas, that an expensive programme has produced limited results or that reported delivery differs from what citizens actually received. It may expose duplicate beneficiaries, procurement delays, unfinished projects, weak revenue collection or persistent gaps between approved spending and completed work.
Better data can strengthen government, but it can also shift influence within it.
An institution that controls a valuable dataset may regard that control as part of its mandate, relevance or bargaining power. A ministry may resist common standards because greater transparency would expose weaknesses in its records. Programme managers may prefer to report registrations, approvals and funds released because information on delivery and outcomes invites more difficult questions about performance.
Not every fragmented system is simply awaiting a technical solution. In some cases, fragmentation protects an existing political or institutional interest.
Public-data reform therefore cannot be treated as a neutral technology exercise. It changes who can access information, who can challenge a decision, which institution is responsible for acting and who becomes accountable when delivery falls short.
Effective reform must address these incentives directly. Institutions need clear mandates for sharing and maintaining data, defined responsibility for its quality and use, and credible consequences when information is withheld, manipulated or allowed to deteriorate. Leadership is equally important because technical standards will achieve little where institutions have no incentive to follow them.
The challenge is therefore not only to build systems that allow government to see more clearly. It is to create the authority, incentives and accountability required to act on what those systems reveal.
Building Institutional Capacity from Digital Platforms
The solution is not to consolidate every government dataset into a single national platform and assume that institutional coherence will follow, because centralising inaccurate, outdated or poorly governed data merely concentrates the same weaknesses at a larger scale; a more effective approach begins with the public decision that needs to improve.
In social protection, the decision may be how to identify and support households whose circumstances require intervention; in trade, how to reduce repeated checks and shorten transaction times; in public investment, how to link approved funding, procurement, payment, physical delivery and the eventual use of the asset; and in healthcare, how to direct scarce medicines to the facilities where they are most needed. Once the decision is clearly defined, government can determine the information, institutions, authority and operational steps required to support it.
Once the decision has been defined, government can determine what information is required, which institution produces it, which record should be treated as authoritative, how frequently it must be updated and who is responsible for verifying it. It must also establish which institution has the legal, administrative and budgetary authority to act, what must happen after the decision is made, how errors can be challenged and corrected, and which outcome will indicate whether the intervention has worked.
This decision-led approach does not require government to integrate every available dataset before it can improve public outcomes. It requires the information, institutions and authority relevant to a specific decision to be connected through a clear operating model that carries the decision through to delivery. Once that model is working and responsibilities are clearly assigned, it can be extended to other programmes and decisions.
The principle is straightforward: Start with the public decision, not the database.
Technology still matters. Governments need authoritative registers, shared identifiers, common definitions, secure data exchange, audit trails, reliable analytics and appropriate access controls so that information can move safely and consistently across institutions.
But these systems will only endure when they are supported by sustained operational funding, capable public servants, clear institutional ownership, realistic maintenance arrangements and accessible processes through which citizens can correct inaccurate information.
The real measure of success is not whether government can launch a platform, but whether the system remains accurate, useful and operational after the consultant departs, the donor programme ends, the minister changes, the technology contract expires and the original project team moves on.
At that point, public data becomes part of the nation’s enduring institutional capacity rather than another platform that performs well at launch and declines once attention, funding and ownership move elsewhere.
The Core Point
A social register contributes to national capacity when it helps support reach the right household at the right time. An identity system contributes to state capacity when it verifies people accurately and allows administrative errors to be corrected, while a means-testing model does so when its classifications are reliable, explainable and subject to review.
A trade platform contributes to state capacity when it eliminates duplicated processes across institutions rather than simply presenting them through a shared interface. A project-monitoring system does this when it links public expenditure to physical completion, functionality and continued use, while a medicine-allocation model does so when improved recommendations translate into medicines being available where they are most needed.
The value of public data is not determined by the volume government collects, the number of platforms it launches or the speed at which it automates decisions. It is determined by what government becomes reliably able to see, decide and deliver.
Public data becomes state capacity when it supports a dependable chain from observation to action. Government must be able to identify a problem, trust the information describing it, coordinate the institutions involved, make a decision within clear authority, carry that decision through to delivery, correct errors and use the outcome to improve what happens next.
Without that chain, digital programmes may generate more records, shorten processing times and produce more sophisticated dashboards while leaving the citizen’s experience of government largely unchanged.
Source notes for the case studies
Nigeria’s National Social Register and cash transfers: NASSCO describes the register’s community-based targeting, digital enumeration, proxy means testing and use in cash-transfer programmes. The World Bank reports that the planned expansion to 15 million households was slowed by integration with biometric identity data [1]
Kenya’s Inua Jamii biometric verification: Kenya’s official news agency reported the 2026 launch and subsequent expansion of fingerprint and facial proof-of-life verification for older-person and disability cash-transfer beneficiaries. [2]
Kenya’s social-health means testing: Kenya’s regulations specify that means testing may use housing characteristics, access to basic services and household composition. A 2026 investigation reported systematic inaccuracies and limited transparency in the resulting income estimates [3]
Togo’s Novissi and successor social-protection architecture: The World Bank documents Novissi’s use of mobile money, geospatial data, surveys and predictive models, as well as the later development of unique identification, a dynamic social registry and a multi-programme payment platform [4]
Nigeria’s National Single Window: The first phase went live in March 2026 with limited initial participation, while official agency communication confirmed the platform’s use for import-certificate processing [5]
Ghana’s unfinished public projects: The District Assemblies Common Fund reported nearly GH¢8 billion in unfinished projects and committed 20 per cent of its resources to completing legacy infrastructure. Ghana’s Ministry of Finance had also allocated part of the Fund towards abandoned legacy projects [6]
Sierra Leone’s medicine-allocation programme: The peer-reviewed study reports a staggered nationwide deployment, an estimated 19 per cent increase in consumption of allocated products in treated districts and expansion to services covering about two million women and young children [7]
