Summary

  • AFRINIC’s notice for its 25 June 2026 annual general members’ meeting proposed that the Board be authorised to appoint the external auditor for the financial year ending 31 December 2026 and determine the auditor’s remuneration. No certified result, signed final minutes, named appointee, engagement letter, fee decision or independence assessment had been found in the examined public record by 10 August.
  • Mauritius company law makes the annual meeting central to auditor appointment and distinguishes appointment from the method for setting fees. AFRINIC’s Audit Committee charter, meanwhile, places recommendation, independence checks, conflict review, fee adequacy and direct auditor access at the centre of the control chain. A delegation could therefore be defensible, but only if the legal route and the committee’s work are visible.
  • The proposal could allocate no more than a private corporate task. AFRINIC is only a private bookkeeper and technical coordinator for number resources. It has no sovereign, regulatory, police, enforcement or punitive authority, and neither a member vote nor an audit opinion can create such power.

Ten minutes, two levers

At 12:10 UTC on 25 June, according to the published agenda, AFRINIC’s Audit Committee chairperson was due to take item 6.5. Ten minutes were allotted. The proposed resolution packed two decisions into a single sentence: the Board would be authorised to appoint AFRINIC’s external auditors for the financial year ending on 31 December 2026, and to determine their remuneration.

The brevity is not, by itself, evidence of a defect. Annual meetings routinely dispatch formal business quickly when the preparatory work has already been done. Nor is it unusual for a company’s directors to negotiate with the professional firm that will audit it. The practical knowledge needed to compare firms—licensing, staffing, timetable, sector experience, scope and price—rarely sits with a large membership assembled for an annual meeting. Sensible governance often delegates execution.

But the sentence joined two unusually sensitive levers. The first is access: which firm receives the mandate to examine the financial statements and the stewardship behind them. The second is incentive: the fee on which the firm’s willingness and capacity to do the work depend. The Board was not merely being asked to procure office supplies. It was proposed as selector and pay-setter for the independent examiner of a reporting process for which the Board itself bears responsibility.

That arrangement is a familiar corporate principal-agent problem, not proof of misconduct. Every external auditor is paid by the institution it audits. Independence is therefore never produced by physical or financial distance alone. It is produced by constraints that outsiders can test: who recommended the firm, what conflicts were checked, whether management was excluded from sensitive communications, whether non-audit services were controlled, how the fee reflected the scope, and whether the auditor could report directly to an oversight committee without seeking management’s permission.

For AFRINIC, the authority chain matters as much as the commercial terms. Its notice was dated 3 June 2026. It said the annual general members’ meeting would begin at 11:00 UTC on Thursday 25 June during AFRINIC-37 at the Emara Ole Sereni Hotel in Nairobi. It was issued by order of the Board and named Executive Services Limited as company secretary. Those facts establish the announced process. They do not establish the result of item 6.5.

That distinction changes the grammar of any responsible account. Resolution VII proposed to authorise. The notice does not show that it authorised. By 10 August, the examined record contained no certified vote total or denominator, no signed final minutes for the meeting, no appointment resolution, no named firm for 2026, no engagement letter, no fee figure, no tender record and no independence assessment. A later announcement about a Governance Committee election is not the missing corporate-resolution result. The most consequential word in this story remains an unproved verb: passed.

The absence of that record does not prove that the resolution failed. Nor does it prove that no appointment occurred away from public view. It means the public cannot move from proposal to completed mandate without evidence. Where the identity of the Board itself is contested, treating an agenda as an outcome would do more than make a clerical mistake. It would allow the institution that announced a power to become the source of proof that it acquired it.

What the annual meeting was being asked to do

The Mauritius Companies Act gives the annual meeting a specific place in the audit architecture. Subject to stated exceptions, a company is to appoint an auditor at each annual meeting. That auditor holds office until the conclusion of the next annual meeting and audits the financial statements for the next accounting period. This is not merely a general instruction that the company should somehow obtain an audit. It describes an appointment associated with the meeting.

The fee provision is phrased differently. The company at the meeting may fix the auditor’s fees and expenses, or determine the manner in which they are to be fixed. When directors appoint an auditor, directors fix the remuneration; when the Registrar appoints, the Registrar fixes it. The distinction matters. It gives members an express choice about a fee-setting method, which can include delegated negotiation. It does not automatically answer the separate question of whether a prospective resolution may leave the auditor’s identity entirely to the Board when the annual-meeting appointment rule would otherwise apply.

Resolution VII therefore sat across two statutory ideas that should not be blurred. Delegating the method of fixing remuneration has an evident textual foothold. Delegating the identity decision requires the corporate record to explain which appointment route was being used and why it fit the circumstances. Perhaps the members were being asked to determine an appointment method. Perhaps an incumbent’s status affected what remained to be done. Perhaps another exception applied. The available record does not settle those possibilities, and an article cannot settle a point the documents leave open.

Two further provisions make assumptions particularly hazardous. First, the Act provides for automatic reappointment of an incumbent auditor at an annual meeting unless one of a specified set of conditions applies. Those conditions include disqualification, replacement by company resolution, a qualifying decision by a small private company not to appoint, or notice from the auditor that it does not wish to be reappointed. Without the incumbent’s identity, eligibility and willingness, it is impossible to know whether the main issue was a fresh choice, a reappointment, a replacement or simply the commercial terms of continuing work.

Second, a qualified incumbent facing removal or replacement has procedural protection. The Act provides for at least 28 days’ written notice and a reasonable opportunity to make representations to shareholders. The record examined does not show whether there was such an incumbent or whether replacement was intended. That is precisely why the question belongs on the disclosure list. Silence cannot be converted either into an accusation that protections were breached or into reassurance that they were irrelevant.

The Act also imposes qualification and disqualification rules. These include restrictions connected to directors and employees, their partners or employees, a liquidator or receiver, specified residence and indebtedness conditions, and conflicts involving related companies. The Financial Reporting Act separately requires a person offering audit services for remuneration to hold the applicable licence, subject to statutory exceptions. A name and price would therefore be only the beginning of a proper appointment record.

The firm’s legal eligibility and professional standing have to be shown, as do relationships capable of compromising independence.

This is the statutory tension at the heart of Resolution VII. The members may have had a rational reason to let a smaller body complete procurement. Company law permits choices about how remuneration is determined and provides Board routes in particular circumstances. But the members’ approval, even if eventually proved, would not make the statutory categories interchangeable. The right record would identify the category, the legal basis and the work done under it.

A casual vacancy is a route, not a label

AFRINIC’s bylaws allow directors to fill a casual vacancy in the office of auditor. The Companies Act does the same. It also provides a backstop: if no auditor is appointed or reappointed at an annual meeting, or if a casual vacancy remains unfilled for one month, the Registrar may appoint one. These routes show that the law anticipates appointment failure and unexpected vacancies. They do not show that either condition existed in 2026.

AFRINIC’s own history illustrates the difference. In December 2019, Board minutes recorded the resignation of PwC, identified the resulting casual vacancy, cited section 195(2) and the bylaws, and then recorded the appointment of BDO for the year ending 31 December 2019 together with the determination of remuneration. Whatever judgment one makes about that decision, the documentary logic was legible: resignation, vacancy, legal power, appointment and fee.

No equivalent chain has been established for the 2026 proposal. Resolution VII was framed prospectively for an accounting year. The available material does not establish an auditor’s resignation or another event creating a casual vacancy. Invoking the Board’s express vacancy-filling power without first proving a vacancy would put the answer before the fact that activates it.

There is another tempting but misleading comparison. In October 2025, a Board-receiver communiqué said that the receiver had appointed Forvis Mazars, described as licensed by the Financial Reporting Council, to conduct audits for 2022, 2023 and 2024 in exceptional circumstances without an annual members’ meeting. That was a backward-looking response to an accumulated gap, undertaken by a court-created transitional actor. Resolution VII concerned prospective authority over the 2026 audit. The exceptional appointment neither supplies a permanent source of Board authority nor retrospectively settles the legal basis for the later proposal.

These comparisons are useful only because they expose the need to name the route. An appointment by members, an automatic reappointment, a Board appointment to fill a casual vacancy, an exceptional act by a receiver and a Registrar appointment are not different descriptions of the same undifferentiated power. They have different triggers and different decision-makers. Good governance begins by resisting the urge to call them all “appointment” and move on.

The committee is the hinge, not the shield

AFRINIC’s Audit Committee charter offers a plausible mechanism for managing the Board’s dual role. Approved by the Board on 6 October 2021 as Resolution 202110.634, it establishes a standing committee to assist with oversight of financial-statement integrity, legal and regulatory compliance, the independence of internal audit, the performance of internal and independent audit, and the quality of operations.

The charter is careful about institutional role. The committee does not perform management functions. It chiefly makes recommendations to the Board. Within that boundary, however, it is assigned the work most relevant to Resolution VII: recommending the appointment and compensation of the external auditor; overseeing the auditor; pre-approving audit work and permitted non-audit services; and receiving direct reporting from the auditor.

It also requires an annual review of external-auditor performance and recommendations on appointment, reappointment or termination. It directs attention to independence and conflicts. It requires assurance that the fee is appropriate and sufficient to permit an effective audit. The committee should meet the external auditor at least annually without management present, review the engagement letter, terms, scope and approach, and ensure that the engagement aligns with the applicable Board approval.

The current committee page names Dr Fiona Makokha Asonga as chair, with Benjamin Mark Roberts, Professor Abdelaziz Hilali and Dewole David Ajao as members. The annual-meeting agenda assigned the ten-minute presentation to the chairperson. This identifies the people and the announced presenter. It does not demonstrate that the charter’s procedures were applied to a particular 2026 appointment. A committee page is evidence of structure; it is not a substitute for a recommendation, an independence declaration or minutes showing a decision.

That is why the committee should be understood as a hinge rather than a shield. If it works as designed, it separates evaluation from executive preference. It can assemble candidate criteria, examine licences, assess the incumbent, compare scope and capacity, inspect commercial and personal relationships, control non-audit work, judge whether the fee buys a credible audit, and preserve a channel through which the auditor can speak without management in the room. Its recommendation then gives the Board a reasoned basis for action.

If none of that work is visible, invoking the committee’s existence does not establish independence. The charter cannot operate itself. Nor can the committee confer on the Board powers the members did not validly grant or powers inconsistent with the governing statute. It can make a lawful mandate better controlled; it cannot cure the absence of a mandate.

The same point applies to conflicts among directors. AFRINIC’s bylaws require a director to disclose a potential or actual conflict and to recuse from voting; a conflicted director may also be excluded from discussion. This is a practical protection where a prospective auditor or its affiliates have prior relationships with Board members, management or connected companies. Yet the useful evidence is not a generic assurance that conflicts are taken seriously.

It is a register showing that interests were considered, minutes recording recusals where necessary, and enough disclosure to establish the control without publishing protected audit material.

Independence has a price as well as a name

The fee power deserves separate scrutiny because a poorly designed remuneration decision can compromise an audit without any explicit instruction to the auditor. A price can be too high, suggesting weak discipline or undisclosed additional work. It can also be too low, leaving insufficient time, senior attention or specialist capacity for an effective examination. Independence is damaged if the auditor becomes economically dependent on ancillary work, fears losing a valuable relationship, or must repeatedly negotiate scope with the managers whose reporting is under review.

AFRINIC’s charter addresses the central point: the committee is to assure itself that the fee is appropriate and permits an effective audit. That calls for more than publishing a final number. The basis should show the expected scope, risk, timetable, staffing and treatment of expenses. It should explain whether changes require further approval. If non-audit services are allowed, their nature and pre-approval should be recorded so that members can judge whether the auditor’s commercial interests could weaken scepticism.

The Board’s power to determine remuneration, if validly acquired, would thus remain bounded by purpose. It would be authority to secure an effective independent audit, not discretion to reward a congenial firm or starve a demanding one. There is no evidence in the examined record that either abuse occurred. The point is institutional: without a fee rationale and variation controls, outsiders cannot distinguish prudent negotiation from influence, even if the directors acted properly.

An engagement letter completes the architecture. It should state the audit’s scope and approach, the reporting line, access to records, timetable and escalation rights. It should protect direct access to the Audit Committee and private sessions without management. It should be consistent with the appointment actually authorised. Until the engagement is documented, an external-auditor label tells members little about the practical freedom to investigate and report.

The firm’s professional independence is necessary, but it cannot bear the weight of the whole institution. An auditor audits financial statements within an engagement and applicable standards. An audit opinion does not decide whether a disputed Board was validly elected, whether every contract was authorised, whether each price was reasonable, whether every actor is liable, or whether the registry possesses powers it does not have. Treating the auditor as a constitutional certifier would compromise understanding even if the financial audit were exemplary.

The authority chain cannot certify itself

AFRINIC’s public Board page identifies the directors whom the institution currently holds out as its Board and sets out AFRINIC’s account of their powers. It proves the institution’s stated position. It does not prove that the Board’s legitimacy is uncontested. Research into the repeated 2025 election process treats later Board acts as carrying a governance risk premium while the electoral authority chain remains disputed. That is context for Resolution VII, not a judicial finding about the vote or the meeting.

The challenge was live close to the meeting. On 24 June, NRS published a member-action page contesting the purported Board’s authority to call the 25 June meeting and asking for objections to be recorded. Its express request to vote against resolutions concerned Resolutions III through VI. The checked page does not establish an NRS instruction on Resolution VII, and none should be invented. The relevant fact is narrower: affected members were disputing the authority of the corporate organ that issued the notice.

That dispute creates an additional burden of documentation. If a Board with an accepted mandate appoints an auditor under a clearly valid member instruction, an incomplete public record may look like ordinary opacity. If the Board’s mandate is disputed, the same omission affects the authority chain itself. The body announcing the meeting, counting or reporting the vote, selecting the auditor and setting the fee may appear to validate itself through a firm it has chosen.

The remedy is not to assume invalidity. Competent courts and the Registrar have statutory or adjudicative roles where corporate authority, appointment failure or validity needs a binding determination. Members and advocates can assemble evidence, object and demand records, but they do not substitute for those public-law institutions. A careful analysis therefore holds two propositions together: Board authority was materially contested, and the record examined contains no final determination resolving that contest for Resolution VII.

This is also where participation must be kept in its proper place. Attendance and voting can evidence an internal corporate decision within the lawful scope of the electorate. They are not a source of sovereignty. A vote does not transform a private membership body into a public government, and participation does not cure a decision that exceeded the company’s legal power. If Resolution VII was carried, the vote could at most authorise a bounded corporate appointment method. It could not consecrate the Board or expand the institution’s authority over the wider internet economy.

The best case for delegation

There is a strong, practical defence of the proposal. Members do not necessarily serve independence by naming an audit firm before due diligence is complete. The Audit Committee may possess better information about licence status, capacity, sector knowledge, incumbent performance, audit scope, conflicts and price. A premature public choice could narrow competition, weaken negotiation or endorse a firm before independence checks have concluded.

Resolution VII was explicit about the two powers and confined them to the financial year ending 31 December 2026. The Act allows the company at the meeting to determine the manner in which fees are fixed. The committee charter supplies a detailed oversight framework. Corporate practice inevitably places the audited company near the appointment and payment decision; the answer is controlled proximity, not the fiction that an auditor can be selected and paid by nobody.

On that view, the members would approve the method, the committee would do the specialist work, and the Board would execute a time-limited mandate. The arrangement could improve independence by delaying final selection until candidates and conflicts had been properly tested. It could also protect price negotiation while ensuring that the scope, rather than the cheapest bid, drives the fee.

This contrary case is more persuasive than the claim that delegation is inherently improper. But it establishes the plausibility of an architecture, not its use. It does not show that Resolution VII passed. It does not identify the statutory analysis for deferring the identity decision. It does not establish the status of an incumbent, the treatment of replacement rights, the candidate list, the committee recommendation, recusals, the fee basis or the engagement terms. Precisely because a well-controlled delegation is possible, the absence of its control record is consequential.

The Board could dispel much of the concern without disclosing confidential audit strategy. A certified result could show the vote denominator, votes cast, abstentions and outcome. Signed minutes could record objections and the exact mandate. A short legal note could identify whether the arrangement relied on annual-meeting appointment, a member-determined method, automatic reappointment or another statutory route. The auditor’s identity, effective date and applicable licence could be stated. A committee report could describe the candidate criteria, independence review and recommendation without revealing protected material.

The remuneration disclosure need not destroy negotiation. It could state the final fee or, before selection, a member-approved band or formula, together with the scope and variation rules. The interests register could identify recusals. A concise engagement summary could confirm direct committee access, private sessions, unrestricted records, escalation rights and the treatment of non-audit services. Each item would answer a different link in the chain; none would require outsiders to trust the mere existence of a committee or the prestige of an accounting firm.

Keeping the institution the right size

Audit governance can easily acquire a grandeur it does not deserve. AFRINIC maintains records for number resources and coordinates uniqueness and related technical services. It is only a private bookkeeper and technical coordinator. It has no sovereign, regulatory, police, enforcement or punitive authority. Its database function is an operational service to network reality, not a throne from which to discipline operators.

That doctrine is not a rhetorical qualification attached to an otherwise governmental institution. It defines the proper interpretation of Resolution VII. Members are principals in a private company. The Board and Audit Committee are corporate agents. The external auditor is a licensed assurance provider. Their acts can improve or weaken a technical organisation’s accountability, but they cannot make law for a continent, police private networks, punish registrants or convert record-keeping control into economic force.

The private-law character of regional internet registry bodies reinforces the boundary. The company’s constitution is auditable corporate architecture, not continental sovereignty. AFRINIC’s published bylaws provide rules for calling meetings, giving members enough notice to form a reasoned judgment, filling an actual auditor vacancy and handling director conflicts. Compliance with those rules matters because private institutions need reliable internal government. It does not transform internal government into public authority.

The distinction protects both members and the organisation. When AFRINIC is described as a quasi-state, every corporate ritual invites exaggerated claims: an election becomes a sovereign mandate, a resolution becomes regulation, and an audit becomes a seal on institutional supremacy. When AFRINIC is kept at its actual scale, the questions become more exact and more answerable. Did the members validly approve a method? Did the proper organ act? Was the auditor eligible and independent? Was the fee sufficient for the work? Were conflicts disclosed? Was the record published?

Those questions are demanding precisely because they are modest. They do not ask an auditor to settle a constitutional struggle or an annual meeting to grant public power. They ask a private organisation to prove the chain by which it purchased assurance from a professional firm.

What can—and cannot—be concluded

The public evidence supports four conclusions. First, AFRINIC announced a meeting and placed a precise proposal on its agenda. Second, the proposal would have allocated both external-auditor selection and fee determination for 2026 to the current/purported Board. Third, company law and AFRINIC’s own charter provide a framework against which that arrangement can be tested: meeting appointment, fee method, incumbent status, qualification, committee recommendation, independence, conflict controls and direct reporting. Fourth, the record examined through 10 August does not prove the outcome or the execution of the mandate.

It does not support a conclusion that the resolution passed or failed. It does not show that a particular firm was appointed or automatically reappointed, or that a particular fee was fixed. It does not show that a casual vacancy existed. It does not prove that a director, committee member, auditor, receiver, company secretary or vendor acted fraudulently, corruptly, collusively or unlawfully. And it does not establish an NRS voting position on Resolution VII.

The unresolved issues are not decorative footnotes. They are the evidence needed to distinguish three very different states: a sensible delegation executed with strong controls; a valid instruction carried out opaquely; or a mandate never validly acquired. Until the result, route, recommendation, conflicts, fee and engagement are known, no one can responsibly choose among them.

An effective audit could still emerge. The members may have approved the resolution; the committee may have completed careful due diligence; the Board may have selected an eligible firm on defensible terms; and the engagement may protect independent reporting. None of those possibilities is contradicted by the available evidence. None is proved by it either.

That symmetrical discipline is the heart of institutional analysis. Suspicion should not be promoted into a finding, and procedure should not be promoted into legitimacy. Resolution VII deserves scrutiny not because delegation is alien to corporate practice, but because two powers over the auditor were concentrated in an organ whose authority was contested. The answer is a record that can be audited independently of the Board’s own description of itself.