Nigeria’s BVAS, Voting Failures and the Risk of Vote Buying

In short: How Nigeria’s biometric accreditation affects access and vote-buying risks, and what election teams can do to protect ballot secrecy and track failures.

By Micheal Eze, Founder and Principal Researcher, Critical Frontiers, in collaboration with Risalat Consultants International

In short. Nigeria's Bimodal Voter Accreditation System (BVAS) verifies voters before they receive paper ballots. It can make impersonation harder, but it cannot prevent vote buying on its own, and failures can keep eligible people from voting. Election administrators need reliable accreditation, a secret ballot and evidence on what happens when either protection breaks down.

What BVAS does at a polling unit

Accreditation and the paper vote

The Independent National Electoral Commission (INEC) uses BVAS for accreditation. The device can verify identity using fingerprints or facial recognition. Accreditation precedes the issue of a ballot. The subsequent act of marking, casting and counting ballots takes place on paper. This distinction matters when a voter presents a Permanent Voter's Card and appears on the register but the device cannot verify them. Registration establishes one part of the check. Under the Electoral Act 2026 and applicable procedures, the official must also follow the required accreditation process.

Accreditation totals help officials assess over-voting. Section 51 of the 2026 Act addresses cases in which votes cast exceed accredited voters in a polling unit. An ordinary difference in the other direction can have a different explanation, since the number of people accredited and the number who complete a ballot are conceptually separate. Reconciliation should preserve the counts and notes needed to investigate a discrepancy. A dashboard that simply flags any difference in red will send supervisors to the wrong cases unless it specifies what each number represents.

Results publication has a different purpose

INEC's Results Viewing Portal, or IReV, makes polling-unit result forms available for public viewing. A failure to upload an image can delay public scrutiny and create uncertainty over when a result was posted. A failure to accredit a voter affects the person's ability to receive a ballot at the polling unit. Both failures can occur in the same election and require different records, remedies and communications.

In its first-hand observation of Nigeria's 2023 general elections, the EU Election Observation Mission documented delays in publication of presidential polling-unit forms. The mission also observed shortcomings in secrecy and instances of vote buying in the vicinity of polling units during the state elections. Its observations describe the units and events observed. They provide evidence of risks and specific failures, with limited ability to measure the prevalence of transactions across Nigeria.

How accreditation affects vote-buying incentives

Uncertainty changes the risks of an illegal bargain

Vote buying predates BVAS. An exchange depends on a buyer believing that an eligible person can vote and that the vote will be cast as requested. Biometric checks increase uncertainty for a buyer whenever accreditation is slow or fails. This can create pressure to defer an inducement until after accreditation or voting. It can also make a purported way of verifying a vote more valuable to the people offering rewards. These are plausible incentives. Establishing a systematic change in payment timing would require observations collected across comparable elections, including evidence about enforcement, turnout and local competition.

Secrecy changes the second part of the bargain. If a voter can mark and fold a ballot in private, a broker has a weaker basis for believing that an inducement secured a particular choice. Polling staff should therefore consider how queues, booths, party agents and observers are positioned. The observer needs a view of the procedure while voters need privacy at the moment of marking. A small room with an exposed writing surface can compromise that privacy even when every accreditation succeeds.

Keep observation distinct from inference

The original account describes the author's experience as a polling officer during a 2025 Edo senatorial bye-election. The author reports that a device continued to accredit other people while repeatedly failing to verify some individuals who appeared on the register with valid cards. The account illustrates the distinction between a unit-wide outage and a failed match affecting an individual. It cannot establish whether the cause lay in the device, stored records, capture conditions or another factor. The commission's bye-election timetable supplies the electoral context. The details of the individual encounters are the author's observations.

A similarly careful distinction applies to suspected vote buying. A queue that becomes crowded late in the day can reflect delayed polling, people arriving from work or an attempt to influence voters. An incident log should describe what an observer actually saw, the time and the action taken. Suspected offences can be referred through authorised channels. The 2026 Act addresses bribery in section 125 and secrecy in section 126. Staff can document process and preserve evidence without asking anyone to disclose their vote.

Diagnose failures and protect access

Rehearse more than a dead battery

A useful exercise gives polling teams several events to diagnose. In one scenario the device will not start. In another it runs but repeatedly fails to verify one voter. In a third the accreditation and count finish, but connectivity prevents publication of a result form. Ask the team to identify the lawful response, the person who can authorise the next action, what to tell waiting voters and which record must be retained. Include a busy queue, a voter who needs assistance and a place where technical support is some distance away.

Section 47 provides for the accreditation process and addresses failure of the deployed device. Its provision on cancellation and another election is subject to the conditions in the Act, including the potential effect on the overall result. That provision concerns the failure of the device in a unit. A single unsuccessful match calls for the applicable polling instructions and escalation procedure. Training should make the boundary clear so an official can respond consistently during a stressful encounter.

Record the outcome without exposing a voter

An incident form can record the polling unit, time, type of failure, whether the problem affected one person or the whole unit, what the official tried, and how the case ended. Each incident needs an identifier that allows a supervisor to follow it without placing biometric material or identifying details into a public spreadsheet. If the same encounter produces several attempts, the reporting definition should prevent those attempts from being counted as several people. The public needs enough information to distinguish an initial failure from an unresolved one.

An illustrative supervisor receives ten reports from three units. Six refer to repeated attempts for a single person who was subsequently accredited. Two concern an outage that ended when a replacement arrived. The remaining two describe people whose attempts were unresolved. The immediate response should be guided by the polling instructions. After the election, these cases suggest different follow-up work: verify counting rules, examine replacement times and review individual-match failures. Aggregating the ten reports into one rate would conceal those distinctions.

Preserve secrecy through operations and assurance

Inspect the physical layout

Before opening, a presiding officer can walk the route a voter will take, from joining the queue to leaving the polling unit. Look at the marking surface from the queue, from the observer's permitted position and from the exit. Confirm where an eligible person who needs assistance can request it and how staff will preserve secrecy in accordance with the rules. At busy units, retest the layout once queues form. The arrangements that worked in an empty room may break down when people bunch around the entrance.

Supervisors can compare incident patterns across locations with similar conditions, using the comparison to decide where additional guidance or observation would help. Observer briefs should include a common description of what counts as interference, a secrecy breach and a technical failure. Consistent categories make reports easier to interpret. They also help prevent a documented defect at one station from turning into an unsupported claim about all stations.

Publish indicators people can interpret

Useful public measures include opening delays, units reporting a device-wide outage, restoration time and unresolved individual failures where reliable counts exist. Explain the reporting period, denominator and coverage. If people left queues without reporting a problem, the incident data cannot capture their reasons. Pair publication statistics with a separate account of accreditation incidents. An election authority can make these data available with clear definitions while protecting the identity of affected voters and the integrity of any investigation.

What public agencies and donor-funded projects can do

Fund resilience and access alongside equipment

Election-support projects can budget for field rehearsals, accessible voter information, technical support, transport for replacement devices and the documentation needed to close incidents. Design the specification around realistic field conditions. Require a supplier to demonstrate recovery from an outage, the handover of a replacement and the reconciliation of logs. Test the support arrangement with staff who will actually be deployed. The responsible election authority should own the legal procedures, approve the data collected and set retention requirements.

An institutional capacity assessment can help a support team identify who controls training, equipment replacement and complaint escalation. A digital readiness assessment can help it examine support coverage, connectivity and data handling. Translate each gap into a named action and an acceptance test before the equipment and staff are sent to the field.

Give independent assurance a clear remit

A ministry or donor financing technical assistance can ask for aggregate evidence that polling teams followed procedures and that staff closed incidents with documented outcomes. An independent observer can compare reported performance with a sample of visits, while an authorised reviewer can examine a secure sample of records. Keep those functions independent of candidate selection, electoral decisions and access to individual vote choices.

After the election, arrange a review with administrators, accessibility representatives and observers. Examine how long voters waited, whether replacements reached remote units, how secrecy worked in constrained rooms and which questions the data could answer. Assign each agreed correction an owner and a date before the next deployment. A better test script, clearer signage or a reliable escalation contact is a concrete improvement that can be checked at the next rehearsal.

Preguntas frecuentes

Does BVAS cast or count votes?

BVAS accredits voters. Voters cast paper ballots, which polling officials count under the applicable procedures.

Is IReV the same system as voter accreditation?

IReV publishes polling-unit result forms for public viewing. Accreditation is the earlier identity check before a person receives a ballot.

Does biometric accreditation stop vote buying?

It can alter the risks of an illicit exchange. Ballot secrecy, observation and enforcement address pressures that identity verification alone cannot remove.

What should an election authority report about failures?

Publish defined, aggregate measures distinguishing outages, repeat attempts, successful recoveries and unresolved encounters where reliable counts exist.