Summary

  • AFRINIC appointed Mike Silber chair of its Bylaws Review Committee on 2 March 2026, then named him CEO Designate on 24 August, with the executive term due to begin on 1 January 2027.
  • The committee was still reviewing submissions on 25 August. Its remit reaches rules for the CEO, directors, committees, membership termination and conflicts, while its terms route reports through the CEO or the person overseeing day-to-day management.
  • Nothing in the checked public packet proves misconduct or even an unmanaged conflict. What is missing is a dated role-transition record showing continued service, screened subjects, recusals, substitute chairing and custody of advice.

At 10:47 UTC on 25 August, AFRINIC’s Bylaws Review Committee issued a short message about two of the most consequential questions in the registry’s constitutional repair. It was considering whether to include a dispute-resolution clause. It was also weighing safeguards around the Board’s residual or discretionary power to terminate resource membership. The consultation would remain open through 30 August.

The timing matters because AFRINIC had made another announcement the day before. On 24 August, the Board and the Receiver named Mike Silber CEO Designate. He is to assume office on 1 January 2027. Six months earlier, on 2 March, the same institutional authorities had appointed Silber to the six-member Bylaws Review Committee and selected him as chair.

There is no contradiction in those dates. A designate is not yet the chief executive. A committee chair can bring exactly the legal, regulatory and institutional memory that a recovering registry needs. Nor is the committee sovereign: its terms of reference call it a non-decision-making body whose advice is nonbinding. The Board considers its recommendations, and members decide whether to approve constitutional amendments.

But those facts do not make the transition administratively self-explaining. They make a precise record more valuable.

The public question is not whether Silber may hold two labels. It is what happened to his committee authority when the future executive relationship became known. Did he remain chair? Did the Board or committee screen particular provisions? Was he absent for deliberations touching the CEO’s powers, appointment or accountability? Did another member chair those items? Who transmitted the committee’s advice, and under which version of the draft? The official documents examined for this article do not answer those questions.

That is a narrow omission with a wide governance radius.

Two appointments, one unfinished process

The 2 March appointment notice names one committee member for each of six African sub-regions. It identifies Silber as the Southern Africa representative and says the Board, with the Receiver’s consent, appointed him chair. The notice says the committee will operate under a Board mandate and engage the community. It does not set a fixed termination date for the chair.

The committee’s formal terms explain why an end date cannot be assumed from the original schedule. The body is supposed to collect, analyse and synthesise public input; report how each submission was handled; give advice on areas needing reform; and prepare a final report of amendments agreed through consultation. It dissolves only when the Board ends it after the work is complete and members have approved the amendments. A timetable can slip while the mandate remains alive.

It did slip. The April consultation timetable was explicitly indicative. It placed final analysis between 19 July and 27 August, a final report on 28 August, and an SGMM in September. The second consultation notice, published on 4 August, instead opened a further round through 21 August. It said the committee would consider the responses, make revisions, obtain an independent external legal review and only then submit recommendations to the Board.

The process was therefore plainly unfinished on 24 August. The third-consultation message removes any residual doubt: on 25 August the committee said it had begun reviewing the second-round feedback and would accept more input until 30 August. It had received a privileged legal opinion on dispute resolution and membership termination, and it was still considering options.

The CEO-designate communiqué sits directly between the second and third rounds. It describes a recruitment process led by the CEO Search Committee, records the Board’s confidence in Silber, and fixes 1 January 2027 as the date on which he will assume office. It does not mention the Bylaws Review Committee. The 25 August message is issued for the committee but does not identify its chair or record a change in chairing.

Silence is not evidence that nothing was done privately. It is evidence that a member reading the published packet cannot tell what was done.

Why “CEO Designate” changes the control problem

The cleanest counterargument is chronological. Silber does not yet control AFRINIC’s staff, budget or day-to-day administration. The appointment notice itself postpones those powers until 2027. Treating a future appointment as if it were a present executive office would be inaccurate.

That boundary should be kept. Yet governance conflicts are not limited to present command. They also concern incentives, expectations and the reasonable perception that future authority may shape present judgment. The constitutional draft illustrates the point without proving any individual failure.

The draft proposed amended Constitution defines the chief executive as the company’s highest-ranking employee, responsible for day-to-day affairs and accountable to the Board. Its proposed Article 17 addresses CEO appointment, eligibility, powers, remuneration authority and acting arrangements. Elsewhere, the draft treats the CEO as an ex-officio director where the context requires. Those are not decorative clauses. They distribute operating authority.

The committee chair helped lead the process that produced and revised those provisions. He is also the named future occupant of the office they describe. That combination need not disqualify him. It does create identifiable subjects for a transition decision.

One subject is CEO eligibility. Proposed Article 17.3 says the Board should determine, reasonably and in good faith, whether an actual conflict or another circumstance is likely materially to impair independent and effective performance. Another is the reporting path. The committee’s terms say reports to the Board go through the CEO or through the person or body designated to oversee day-to-day management. If the committee survives into January, its former or continuing chair could become part of the channel through which its work reaches the Board.

A third subject is conflict policy itself. Proposed Article 19.13 would require an organisation-wide policy covering identification, disclosure, assessment, management and, where appropriate, mitigation of actual, potential and perceived conflicts. Its listed scope includes committee members as well as directors, officers and employees. The proposal is not yet an adopted rule, and it does not establish that Silber has any actual conflict. But it supplies the committee’s own vocabulary for describing the transition: actual, potential and perceived.

The right response to that vocabulary is process, not insinuation.

Independence statements are principles, not event records

AFRINIC has already recognised that confidence in the review depends on more than technical drafting. On 19 May, after criticism from an outside group, the committee issued a statement on independence and integrity. It said members were conscious of duties to act independently, impartially and in the community’s best interests. It rejected claims that one organisation controlled the exercise.

That statement deserves to be credited. It also shows the difference between an assurance and a control record.

An assurance says what standard members accept. A record says how the standard was applied to a later event. The August appointment was such an event because it changed one member’s future relationship with the institution. The relevant record need not reveal deliberative confidences. It could say that the new role was disclosed on a given date; that the Board or committee assessed specified subject areas; that Silber continued on non-affected work; that another named member chaired certain agenda items; and that any abstention was recorded in the minutes.

Without those fields, readers face an unhelpful binary. They can either trust a general statement or infer a problem from silence. A well-designed register creates a third option: verify the procedure while withholding privileged advice and personal detail.

This distinction is especially important because the committee relied on confidential legal advice. The 25 August message says the opinion cannot be disclosed because it is legally privileged. That may be entirely proper. Privilege protects the substance of counsel’s advice; it does not automatically require opacity about who chaired a meeting, who participated, which subject was screened or whether a recusal occurred. AFRINIC can preserve the first boundary while documenting the second.

The committee’s advisory status reduces one risk, not all of them

Another sensible defence is that the committee cannot enact its own draft. The Board and members retain authority. Advice is nonbinding. That substantially limits the chair’s formal power.

It does not eliminate agenda power, information power or drafting power. A chair can shape meeting order, frame alternatives, decide when a text is mature enough to circulate, and become the natural public interpreter of a compromise. A drafting committee can also narrow the menu that later reaches decision-makers. None of those functions is improper. They are simply forms of influence that a role-transition record should make legible when the chair is preparing to enter the executive office.

The risk is therefore not that one person can secretly amend the Constitution by fiat. The system contains multiple veto points: committee agreement, legal review, Board consideration and a members’ vote. The risk is that the provenance of recommendations becomes difficult to reconstruct. If a disputed clause later affects executive authority, membership termination or the route for resolving disputes, stakeholders may ask whether the future CEO helped chair its treatment. A record produced now is cheaper and more credible than a retrospective explanation produced during a dispute.

This is a recurring institutional lesson. Separation of final decision rights is necessary, but it is not the same as separation of upstream influence. Good governance records the latter because downstream votes rarely disclose how options were assembled.

What a role-transition record should contain

AFRINIC does not need a theatrical disclosure. It needs a small, versioned control object tied to dates and subjects.

First, record both appointments. The committee-chair appointment date is 2 March 2026. The CEO-designate appointment date is 24 August. The executive assumption date is 1 January 2027. These three dates prevent the record from collapsing future and present authority.

Second, record the service decision. State whether Silber continued as chair, continued only as an ordinary member, stepped aside for defined subjects, or left the committee. If the arrangement changes later, preserve the previous version rather than silently replacing it.

Third, name the assessor and the test. Was the decision made by the Board, the Receiver, the committee without the affected member, outside counsel or some combination? Did it examine actual, potential and perceived conflicts, or use another standard? A conclusion without an owner is hard to audit.

Fourth, map the decision to subjects rather than personalities. The obvious list includes CEO appointment and eligibility, CEO powers, ex-officio directorship, executive remuneration, acting-CEO arrangements, conflict policy, Board residual powers, membership termination, dispute resolution and the channel for submitting committee reports. Screening by subject allows useful expertise to remain available where no conflict is identified.

Fifth, record meeting consequences: attendance, chairing, participation, abstention and who took the chair. A recusal that exists only as a general promise cannot later establish which text was handled without the affected participant.

Sixth, attach document provenance. Each relevant decision should identify the draft version, agenda item, minute or public report to which it applied. That is the bridge between governance assurance and the wording members will eventually vote on.

Finally, close the record. Note when the committee dissolves, when the chair’s role ends, who holds the archive and how recommendations were transmitted. If the body continues past 1 January, the reporting path requires an explicit answer: the incoming CEO should not be left to receive a report whose preparation he chaired without a documented separation or approval arrangement.

The case for keeping Silber involved

A fair analysis must address the positive case, because removing every person who later accepts institutional office can also damage governance.

Silber’s published biography includes regulatory, legal, Internet-governance and board experience. AFRINIC’s communiqué says the Board selected him after a rigorous process to help restore stability, governance and trust. Those are directly relevant qualifications for constitutional work. A chair who understands the history of the dispute may expose drafting consequences that an outsider misses. Continuity can reduce handover errors just as the process approaches completion.

There is also a sequencing argument. His executive mandate starts months later. The committee is advisory. Legal review and member approval remain. A categorical demand that he depart on 24 August could treat perception as guilt and sacrifice expertise without measuring the actual subject-level risk.

That is why a public record is the proportionate remedy. It avoids both extremes. It does not pretend that a future CEO appointment changes nothing, and it does not presume that the appointment contaminates everything. It separates matters where continued participation can be defended from matters where independent chairing would protect the result.

The more valuable Silber’s expertise is, the stronger this solution becomes. A documented arrangement allows AFRINIC to retain it without asking the community to accept an invisible boundary.

The public test is reproducibility

Members do not need access to privileged legal advice to evaluate the review. They need to reproduce the chain of institutional responsibility.

For each material clause, they should be able to move from a consultation submission to the committee’s treatment, the relevant draft, the legal-review status, the Board’s disposition and the resolution placed before members. Where a participant’s institutional role changes midstream, the chain should also show the transition decision. The committee’s own terms already promise a report explaining how input was handled and an impact assessment for each amendment. Adding role-transition metadata is consistent with that architecture.

The absence is most visible at the seam between documents. The March appointment notice says Silber chairs. The August CEO communiqué says he will take office in January. The next day’s committee message says the review continues. Each document can be accurate on its own while the combined record leaves the operating question unresolved.

AFRINIC can close that seam with one page. It can date the disclosure, state the service arrangement, name the approving authority, list screened subjects, identify any alternate chair and link the decision to draft versions. If there was no recusal because the assessor found no conflict, say that too. Transparency does not require a predetermined substantive outcome; it requires a reviewable one.

The committee is trying to write a Constitution that will survive future disagreements. Its handling of this transition is an early test of the same proposition. Institutional trust is not built by insisting that expertise and independence never collide. It is built by showing, in a form that can be checked later, how the boundary was drawn.

Sources