Summary

  • Draft Article 13.5 would restore staggered AFRINIC Board elections by assigning three-year terms to the three candidates with the highest valid-vote totals, two-year terms to the next three and one-year terms to the following two in 2028.
  • Draft Articles 13.3 and 13.4 separately preserve six regional seats and two competency-based non-regional seats. Article 13.5 does not say whether it ranks every candidate or only the eight seat winners.
  • A valid total can identify the winner inside one seat without being comparable to a total from another seat. Ballot completion, abstention, invalid votes, an uncontested race and the size of the candidate field can change totals even when the eligible-member list is shared.
  • The cleanest transition assigns duration to each seat before nominations or voting. If AFRINIC keeps a candidate-based ranking, it needs a predeclared metric and a public allocation receipt that separates seat victory from term length.

Eight boxes on the ballot, one second consequence

Imagine an election screen with eight boxes. Six are labelled for geographic seats. Two are for non-regional seats based on competence. A member may be choosing a different person in every box because every box fills a different office.

Now add a second calculation after the screen closes. The winner in each box is not only elected. All the relevant vote totals are placed in one league table. The top three people receive three years in office, the next three receive two, and the last two receive one.

That second calculation is the unexplained part of AFRINIC's 4 August draft Constitution.

The proposal has a defensible purpose. The 2025 election reconstituted all eight elected Board seats at once. The operative 2020 Constitution was built around sequential elections and ordinary three-year terms. If eight directors begin together and all receive the same duration, they can also leave together. A one-time set of unequal terms is a straightforward way to recreate a stagger.

Draft Article 13.5 states that purpose expressly. As an exceptional transition for the 2028 elections, the three candidates receiving the highest number of valid votes would each serve three years. The next three would serve two. The following two would serve one. Thereafter, elected directors would return to three-year terms and elections would be arranged so that, as nearly as practicable, one-third retire each year.

The arithmetic is clear. The comparison set is not.

The draft defines seats before it ranks candidates

Draft Article 13.3 does not describe a single eight-member at-large contest. It describes eight elected seats. Seats 1 through 6 each represent a named sub-region: Northern, Western, Indian Ocean, Central, Southern and Eastern Africa. Seats 7 and 8 are non-regional and are filled on competence rather than regional representation. The CEO occupies a ninth, ex-officio seat.

That structure carries forward a durable feature of the operative Constitution. The 2020 text gives one directorship to each of six regions, creates two distinct non-regional seats and joins them to a sequential cycle. Every elected director ordinarily receives three years. The sequence is attached to offices, not to a retrospective popularity table.

AFRINIC's official September 2025 result notice also records the election as eight seat outcomes. It names one winner for Seat 1, another for Seat 2 and so on through Seat 8. It does not publish candidate totals in that notice.

An indexed capture of AFRINIC's official 2025 Election Guidelines makes the historical mechanics more explicit. Each eligible Resource Member was allocated one vote per open seat. The candidate with the most votes won for each seat. The result was supposed to include total votes for every candidate. That direct URL returned 404 at this article's cutoff, so it is evidence of the 2025 implementation, not a promise about 2028.

The 2028 Election Guidelines do not yet exist in the reviewed record. They may use a different ballot, electorate or certification method. The point is not to project 2025 mechanically into the future. It is to identify the join the proposed Constitution must define if it wants one set of seat elections to produce a second, cross-seat allocation of time.

Candidate can mean two different populations

Article 13.5 says the three candidates with the highest valid-vote totals receive three-year terms. It does not say the three elected candidates, the three seat winners or the three directors-elect.

One reading is that every candidate across all eight contests enters a common ranking. That reading collides with the seat architecture. A person can receive a high total yet lose the particular seat sought because another candidate in that seat receives more. A different person can correctly win another seat with a lower total. Article 13.3 still requires one director for each regional seat and two directors for the non-regional seats. A global list of all candidates cannot by itself decide those offices.

The more plausible reading is therefore a two-stage process: determine the winner of each seat, then rank the eight winners by their raw valid-vote totals and use that rank to allocate duration. But the draft does not say so. A constitutional transition should not depend on silently inserting winning before candidates.

Nor does the tie clause resolve the non-tie question. Article 13.5(c) sensibly requires a fair, transparent and objective mechanism where equal totals affect term allocation, and forbids allocation until the tie is resolved. That deals with equality at a boundary. It does not define which totals enter the ranking, why they are comparable or what happens when their denominators differ without being equal.

The same electorate is not the same denominator

Suppose the 2028 rules give every eligible member one choice in every seat, as the 2025 rules did. That creates a common opportunity to vote. It does not necessarily create a common number of votes cast.

A member may choose a candidate in Seat 1 and abstain in Seat 2. One seat may have three competitive candidates while another has one eligible candidate. A ballot rule may permit a none option. A mark can be invalid in one contest without invalidating choices in seven others. A correction can change one seat result. Even if every seat is open for exactly the same hours and every voter comes from the same frozen register, the valid-vote total can still vary by seat.

That variation is not automatically suspicious. It may be an accurate record of voter choice. The problem begins only when a total created to select one office is reused as a measure of relative mandate across other offices.

Raw totals, share of valid votes and share of the eligible electorate answer different questions. Raw totals reward the number of affirmative marks. Valid-vote share measures performance among the choices counted in one contest. Eligible-electorate share also includes absence and abstention in its denominator. A contest with one candidate creates a different signal from a contest with five. No conversion is neutral merely because it produces a decimal.

The institution must choose the meaning before it knows who benefits.

The strongest defence of the formula

The strongest case for Article 13.5 should be admitted. AFRINIC needs a bridge from simultaneous reconstitution to orderly rotation. Giving the longer terms to candidates with more votes can look more member-driven than letting the Board choose which of its own directors leave first. The rule is easy to explain. It also fixes the eventual 3/3/2 shape without requiring three further years of synchronized turnover.

Unequal first terms are not a penalty. They are a transition device. A one-year director still occupies a real Board seat with the same vote while in office. A three-year director does not gain three votes or a larger jurisdiction. Duration and voting power are separate fields.

This defence makes the comparison rule more important. If member votes are to justify the unequal duration, the public must be able to reconstruct what each vote authorized. A member choosing the Western Africa director may not have understood that participation in that box would also help determine whether an Eastern Africa or non-regional winner serves one, two or three years. The ballot and nomination notice must disclose the second consequence before the choice is made.

The most valuable feature of the draft is that it puts the basic formula in the Constitution rather than leaving the entire decision to Board discretion. Its weakness is that the formula stops one step before its input becomes well-defined.

Allocate time to seats before votes select people

The cleanest repair is a seat-first transition schedule.

Before nominations open, the adopted Constitution or an expressly authorized public procedure should assign three-year, two-year and one-year transitional durations to Seats 1 through 8. The schedule could follow a member-approved sequence. It could use an auditable draw among seats. It could use another neutral rule whose rationale is published before any candidate decides where to stand. The important point is temporal: the duration belongs to the office before the officeholder is known.

This separates two authorizations. The transition rule allocates time among seats. Members then select a person for each seat. Candidate support no longer becomes a hidden exchange rate between different contests.

Preassignment has costs. Some regions may object that their seat received the short first term. A public draw can distribute that burden without pretending it reflects voter preference. A fixed sequence can favour predictability but needs an explained principle. Members may prefer the current vote-ranked formula precisely because it lets electoral support influence duration.

If AFRINIC keeps that formula, it should not hide the choice behind highest valid votes. The final text or Election Guidelines should state:

  1. whether the ranking includes all candidates or only certified seat winners;
  2. the electorate and ballot-completion rule for each seat;
  3. how abstentions, none, invalid votes and uncontested seats are treated;
  4. whether the metric is raw total, valid-vote share, eligible-electorate share or something else;
  5. when corrections freeze the totals used for duration;
  6. how a seat-result challenge affects the cross-seat ranking;
  7. the tie mechanism and its public record; and
  8. when each assigned term begins and expires.

Those details must be public before candidacy and voting, not invented during certification.

A result needs two certificates

The 2028 public record should show two logically separate outcomes.

The first is the seat certificate: seat identifier, seat class, candidate list, eligible electorate, ballots completed, abstentions, invalids, candidate totals, corrections, winner and certification. That record proves who was elected to a particular office.

The second is the duration certificate: the adopted Article 13.5 version, ranking universe, comparison metric, eight inputs, ordering, ties, resolution method, allocated duration, start and expiry dates, and the later seat-replacement map. That record proves why an elected person received one, two or three years.

Both certificates can protect ballot secrecy. They need institutional counts and rules, not voter identities. Both need stable identifiers so a corrected seat total automatically points to the affected duration ranking rather than leaving two inconsistent results online.

The receipt should also state whether duration was assigned to the seat before voting or derived from candidate totals afterward. That single field prevents later readers from confusing a feature of the office with a reward for the person.

This is the ledger principle in practical form. A public claim that the transition was fair is not the evidence. The evidence is the reproducible chain from adopted rule to contest, result, comparison and expiry date.

What the record does not establish

Article 13.5 remains proposed text. The second consultation closed on 21 August. A 25 August BRC message kept a third round open through 30 August at 23:59 UTC and said the Committee was still reviewing feedback. The message discussed two other legal-design questions and did not attach a new complete authenticated draft. The 4 August PDF is therefore the latest complete proposal identified for this analysis, not proof of the final recommendation.

No reviewed source shows Board approval, an SGMM resolution, an adopted replacement Constitution or a 2028 Election Guideline. No 2028 candidate, ballot, vote, tie, winner or term exists in the evidence set. Nothing here proves manipulation, discrimination, unlawfulness or an actual injury.

The finding is narrower and useful precisely because the decision remains reversible. Eight offices are defined separately. One clause would rank valid-vote totals to allocate time across them. The public text has not yet defined the comparison.

Fixing that sentence before members vote on the Constitution is governance. Trying to infer its meaning after members vote in eight seats would be litigation over arithmetic.

Sources

  1. AFRINIC — Second Community Consultation on the Draft Amended AFRINIC Constitution
  2. AFRINIC — Draft Proposed Amended Constitution, 4 August 2026
  3. AFRINIC — Constitution 2020 / Bylaws
  4. AFRINIC — Election Guidelines 2025
  5. AFRINIC Announce — Board Election Results 2025
  6. AFRINIC Bylaws Review Committee — Third Round of Community Consultation
  7. AFRINIC — Bylaws Review Committee Terms of Reference