Summary

  • AFRINIC’s 8 May call sought two consensus-selected PDP co-chairs, with one transitional one-year term and one two-year term.
  • AFRINIC-37 minutes say NomCom received two submissions from two people, but the final slate identified only Haitham El-Nakhal as eligible.
  • The minutes record a show of hands for acceptance and opposition, then say the candidate was accepted. They give no eligible-participant roll or numerical count.
  • The minutes assign the selected chair a two-year term, say a second chair could be chosen in November and record an unresolved request for an outgoing chair to provide temporary support.
  • The packet does not prove illegality or candidate wrongdoing. It shows that AFRINIC’s claim of consensus is not yet backed by the complete public control record required to audit the gate, the room and the result.

A two-seat process ended with one named appointment

AFRINIC did not begin its 2026 policy-chair exercise with one vacancy. Its 8 May call advertised two PDP co-chair positions to be filled by consensus. The exceptional transition was explicit: one successful candidate would serve one year, the other two years, restoring the staggered cycle after the organisation’s governance interruption.

The sequence that followed is now partly visible. AFRINIC’s minutes for the 24 June Public Policy Meeting say the Nomination Committee received two submissions from two unique candidates. The final slate, published on 17 June, named only one person as eligible: Haitham El-Nakhal. At the meeting, NomCom presented him for a two-year term. Participants were asked to show acceptance by raising their hands; those who disagreed were also asked to raise their hands. The minutes then say the proposal was accepted.

AFRINIC’s 30 June result announcement compresses that sequence to one line: El-Nakhal “was selected via consensus” as the new PDWG co-chair.

The appointment is a real event. So is the missing second appointment. The minutes say another co-chair could be elected at the next PPM in November. They also record a proposal that an outgoing co-chair support the incoming chair until then. One outgoing chair said he would assess whether he had time; later, a show of hands was described as consensus for an outgoing chair to provide support.

This record corrects an easy but false conclusion. The public packet is not silent about the meeting. It identifies two applications, one eligible candidate, the acceptance method, the assigned term and a planned date for revisiting the second seat. The news question is narrower and more consequential: does the record show enough to test each exercise of control that produced that outcome?

The eligibility gate decided the contest before the room met

The first decisive act did not occur when hands went up. It occurred when NomCom reduced two applications to a final slate of one.

AFRINIC’s Board-approved Election Guidelines set demanding criteria for a PDWG co-chair. A candidate must live in the service region, understand the Policy Development Process and RFC 7282, have at least three years of meaningful RPD participation, have attended two AFRINIC events in five years, be available for policy work, avoid authorship of a proposal currently under discussion and carry nominations from two Resource Members in good standing.

The PPM minutes give a shorter checklist: residence, complete documents, a nominator and seconder in good standing, AFRINIC engagement and no conflict of interest. They do not state which test the other applicant failed. The final-slate announcement simply says only one candidate was found eligible.

There are legitimate reasons not to publish every personal document in a nomination file. The absence of a name or private evidence is not proof of manipulation. Nor is El-Nakhal’s former AFRINIC Board service, or his current position at Egypt’s NTRA, evidence of misconduct or a disqualifying conflict.

Privacy, however, does not require an unreviewable decision. A public process can disclose a criterion-by-criterion decision matrix, a neutral reason code, the evidence type inspected, recusals and an appeal or correction route without exposing sensitive records. That is especially important when the gate converts a two-person submission pool into a single-candidate proceeding.

The distinction matters because a meeting cannot compare candidates it is never permitted to see. Once NomCom published one eligible name, the community’s choice was no longer between two people. It was acceptance or rejection of one person, followed by a separate question about the unfilled second position. The institution cannot use the public meeting’s openness to substitute for transparency at the earlier gate.

“Consensus” needs a participant and decision record

The Election Guidelines say eligible participants are people who form part of the African Internet community and complete registration for the PPM. Section 12.8 goes further: after registration and due diligence, AFRINIC is to publish a list of registered participants granted in-person or online access.

That list is not a decorative attendance sheet. It defines the body whose response can be described as PDWG consensus. Without it, a reader cannot determine how many eligible participants existed, how many were present, whether the online and physical rooms were treated alike or whether the people raising hands belonged to the registered decision body.

The minutes record the method but not the measurement. The floor was opened. Acceptance was requested by show of hands. Opponents were invited to raise theirs. The candidate was noted as accepted. No number is supplied for acceptance, opposition or abstention. No participant total or attendance roll appears in the fixed 20-source packet.

Consensus is not always a vote, and rough consensus is not reducible to a majority count. A hand total alone would not prove a sound deliberative judgment. But rejecting simple vote counting does not eliminate the need for evidence. A reasoned consensus record should identify the eligible group, the questions put to it, material objections, how those objections were assessed and why the facilitator concluded that the result met the announced standard.

There is also a textual tension in AFRINIC’s own rule set. Section 14.3 of the guidelines says a single unopposed candidate should be declared selected by acclamation without an election or selection exercise. The meeting minutes instead describe an acceptance-and-opposition show of hands. That may have been a sensible confirmation of community support. It may also reflect an attempt to reconcile acclamation with the PDWG’s consensus tradition. The packet does not explain which rule controlled or why this method was chosen.

That tension is not, by itself, a finding that the appointment is void. It is a reason to publish a short decision note. When an institution has written both “acclamation” and “consensus-based process” into the same framework, it should show how the two operated in the actual case.

The second seat became a continuity arrangement without a complete mandate

The original call promised two staggered appointments. The meeting made one, assigned the longer term and moved the second chair to a future PPM.

The minutes show participants recognised the operational risk. One participant proposed appointing a temporary co-chair to support the incoming chair. The outgoing chairs were asked whether one could remain. Darwin da Costa said his professional focus had shifted and that he could not then give a definitive answer, although he would assess his availability. Later, the minutes say there was a show-of-hands consensus for an outgoing chair to support the incoming chair.

What the packet does not show is equally specific. It does not identify a named outgoing chair who accepted a formal interim appointment. It gives no written delegation, start and end date, conflict rule or account of which co-chair powers “support” includes. It does not say whether this was an appointment, an informal handover or an expression of preference contingent on later consent.

The safe conclusion is not that the second seat was abandoned. AFRINIC’s minutes say it could be filled in November. Nor can the record support a claim that an outgoing chair unlawfully stayed in office. It supports a more disciplined finding: the continuity arrangement was discussed and endorsed in general terms, while the identity, authority and duration of any resulting role remained unresolved in the bounded public packet.

That is precisely where governance documents should be most exact. A temporary bridge can be necessary. But necessity does not answer who may exercise power on the bridge.

This is not a ceremonial post

AFRINIC’s guidelines call co-chairs procedural custodians rather than policymakers. That description should not obscure the power embedded in procedure.

Co-chairs evaluate whether a policy proposal has reached rough consensus. They guide proposals through discussion, initiate and end Last Call, document determinations, report outcomes and transmit adopted policy to the Board for ratification. AFRINIC’s current PDWG materials also make clear that chairs publish minutes and decide the procedural status of proposals.

Those acts shape the operating rules for number resources on which networks depend. A chair does not allocate an address block by personal fiat, but a chair can decide whether a disputed proposal advances, returns to discussion or reaches the Board. The eligibility screen that determines who may hold that role is therefore part of the registry’s control surface, not an internal human-resources detail.

Heng Lu’s doctrine asks the correct questions at each link: who has the capacity to decide; who carries the economic downside; what prevents the agent from substituting institutional interest for the principal’s interest; and what remedy exists when power is exercised wrongly. Here the chain runs from the Board-approved rules to NomCom, from NomCom to the registered meeting body, from the room to the consensus interpreter and from the selected chair to future policy outcomes.

The people and networks affected by those outcomes bear the practical cost of error. Yet the fixed packet does not assign equivalent liability or a clear correction duty to the people who control eligibility and consensus. That asymmetry makes documentation a minimum control, not a public-relations courtesy.

The current authority dispute raises the standard, not the allegation

AFRINIC says its Board approved the Election Guidelines on 4 February 2026. AFRINIC had announced the Board after its 2025 election and later said an Interim Management Committee was created with Receiver consent, reporting jointly to the Board and Receiver. NRS disputes the legal finality of the 2025 election and calls for transaction-specific Board, delegation, Receiver and court authority records.

Those competing positions do not allow this briefing to decide that the Board, NomCom or co-chair appointment was unlawful. The packet contains no final merits judgment resolving the broader dispute. It also contains no evidence that the Receiver supervised or approved the PDWG selection.

They do explain why institutional titles cannot carry the proof alone. A rule described as Board-approved should be linked to the resolution or authority on which it relies. A NomCom eligibility decision should show the published criteria and review path. A community consensus should identify the participating body and the reasoning. A temporary support role should have a named holder, defined scope and end date.

The LARUS-hosted legal opinion, although directed principally at APNIC, makes the structural point: the appearance of elected bodies is not enough; actual corporate authority has to be traced. NRS applies a similar documentary demand to AFRINIC. These are attributed analyses and demands, not judicial findings about El-Nakhal or the June meeting.

A publishable control record is straightforward

AFRINIC can close the public gap without disclosing sensitive nomination files or pretending that consensus is a ballot.

For applicant screening, it can publish a two-row disposition table with identifiers that preserve privacy where necessary. Each row can show every criterion, pass/fail/not-applicable status, the evidence type, the decision date, conflicts and recusals, a concise reason code and the route for correction or appeal.

For the meeting, it can publish the registered-participant list promised by section 12.8, or a privacy-preserving verifiable roll if data law requires less disclosure. It can state how many eligible participants attended in person and online, how many indicated acceptance, opposition or abstention, what objections were raised and why the facilitator treated the response as consensus.

For the seats, it can state that El-Nakhal received the two-year position, identify the legal status of the remaining seat, publish the November timetable and explain whether a one-year transition term will still be used. Any outgoing-chair support should identify the consenting person, authority, functions, conflict rules and automatic expiry.

This would not predetermine whether the appointment survives every legal challenge. It would do something more basic: make the decision chain inspectable by the people in whose name consensus is claimed.

AFRINIC’s minutes already provide facts that a shorter result notice omitted. That is valuable. But “two applications, one eligible, hands raised, accepted” is a narrative of a process, not yet a complete audit record of it.

The bounded conclusion is therefore neither celebratory nor accusatory. One co-chair was appointed for two years. A second selection was deferred. The packet contains no competent finding of illegality and no basis to accuse either applicant of wrongdoing. It does show that control was exercised at three points—eligibility, participation and consensus—without the complete public evidence needed to independently test those acts.

Readers seeking a broader treatment of election legitimacy can consult the separate BTW research listed below. That longform is further reading; this briefing reports the distinct 2026 PDWG co-chair event and does not count the research article as news.

Sources