Summary

  • On 2 March 2022, AFRINIC’s Board voted five to three for Resolution 202202.676, adding one year to the incumbent’s term in Seat 6, the Eastern Africa seat.
  • The strongest case for the decision was a real one: article 13.5 stated a three-year term, an earlier resolution had set a two-year mandate, and a supporter warned that too many simultaneous elections could endanger continuity or quorum.
  • Yet an election date is also a control surface. Moving it determines when members may replace a director and who remains able to participate in Board decisions during the added year.
  • The minutes show that the Board considered ordinary bylaw amendment and a special general meeting, but it extended the incumbent before separately commissioning the wider rule repair.
  • The record does not publish the two legal opinions, settle their disagreement, establish the resolution’s ultimate legal validity, or support claims of improper motive.
  • A defensible term repair should leave a public receipt: the defect, authority, affected seat, alternatives, conflict treatment, evidence for urgency, duration, expiry and route to member amendment or ratification.

Five votes and one more year

The decisive act came in a roll call. Five directors—Abdalla Omari, Benjamin Eshun, Habib Youssef, Silvio Almada and S. Moonesamy—voted in favour. Three—Oluwaseun Ojedeji, Adewale Adedokun and Eddy Kayihura—voted against. Eshun proposed the resolution and Youssef seconded it. With that five-to-three result, AFRINIC’s Board extended the incumbent of Seat 6, representing Eastern Africa, by one additional year.

The official number is Resolution 202202.676. AFRINIC’s Board register groups it under February, while the minutes place its adoption in the continuation held on 2 March 2022. The date presentation is untidy, but the official identifier and the operative decision are clear. The resolution recited two propositions. Article 13.5 of the Bylaws gave an elected director a three-year tenure. Resolution 202009.568, adopted on 21 September 2020, had given Seat 6 a two-year term. The Board’s answer was to add the missing year to the sitting term.

Read only as arithmetic, the decision has an attractive neatness. Three years minus two years leaves one. Add that year and the term conforms to the duration stated in article 13.5. No other seat has to be shortened. No vacancy has to be manufactured. A director elected under a published two-year arrangement remains for the length associated with elected directors in the Bylaws. The correction is easy to state and easy to administer.

But the unit being adjusted was not an abstract interval. It was a year in office. During that year, members would not have the opportunity they had expected to use to contest Seat 6. The incumbent would continue to occupy the seat and, subject to the ordinary operation of the Board, participate in decisions that arrived before the revised expiry. A calendar entry therefore changed the distribution of corporate power. That conclusion requires no allegation about anyone’s intentions. It follows from what an election date does.

The distinction matters because AFRINIC is a private member-based regional Internet registry. NRS’s account fixes that institutional character. AFRINIC keeps and coordinates number-resource records and provides services within a defined operational remit. It is not a government for a service region. Its Board does not legislate for a continent, and a Board vote does not become public law because its consequences matter to network operators. Heng Lu’s ledger-not-throne analysis supplies the essential limit: bookkeeping and coordination can be consequential without becoming sovereignty.

Resolution 202202.676 thus deserves neither grandiose language nor clerical dismissal. It was not an act of a legislature, regulator, police force, prosecutor, punisher, confiscator or public-law court. It created no such powers. It was a private corporate act about office-holding under AFRINIC’s own governance arrangements and applicable law. Yet within that private institution, the act mattered. Members exercise one of their most concrete forms of control when a seat becomes contestable. Changing that date changed when their control could be used.

That is the tension at the centre of the resolution. The Board could plausibly see itself as restoring the term promised by article 13.5. At the same time, it chose an office-holder-specific outcome while the general rule problem was unresolved. The schedule became internally neater because one person stayed longer. The constitutional question is not whether one year can be added correctly on paper. It is why the office-holder result came before the rule repair, and what safeguards should accompany that sequence.

The history immediately behind Seat 6 explains how the mismatch arose. AFRINIC’s 2021 Election Process said Robert Ford had been elected to the seat at the 2019 Kampala AGMM and resigned after one year. On that account, the seat was vacant for the remaining two years. NomCom then called nominations for a two-year mandate in order to preserve the election sequence. Resolution 202202.676 later described Resolution 202009.568 as the source of the two-year term. By early 2022, the effort to maintain one schedule had collided with the three-year language applied to an elected director.

The 2021 material is important only for that chain. It shows that the two-year mandate was presented as a way of preserving sequence after a resignation, not as a free-floating invention in March 2022. It does not answer whether the later Board possessed the power to extend the incumbent, whether a different transition was preferable, or whether members should have acted first. Those questions must be examined from the 2022 record itself.

That record begins on 26 January, when the Board was reviewing the 2022 Election Guidelines and the seats due for election. The Chair asked the CEO to seek further legal advice on a director’s tenure. A continuation planned for 2 February was postponed because of cyclonic conditions in Mauritius. The Board reconvened on 21 February and examined how many elections the 2022 AGMM might require, the tenure attached to each seat and the effect of casual vacancies that had arisen in 2021.

The discussion was not a single-minded march to extension. The CEO told the Board that it had an internal legal opinion and an opinion from C&A Law and said both provided for six elections. The minutes do not reproduce either opinion. One director questioned whether the unsigned C&A document was admissible. The CEO responded that the email provenance could be traced and that C&A had confirmed the document as final. The public record available here contains neither an independent authentication nor the text needed to compare the reasoning line by line.

The minutes also record disagreement over substance. Participants disputed whether the opinions actually differed and whether the Bylaws needed amendment. That disagreement is itself an important fact, but it cannot be converted into a public holding about which side was legally right. Readers can see that advice was sought, that two opinions were said to exist and that their relationship was contested. They cannot see the full premises, qualifications or conclusions. The opacity places a hard limit on responsible judgment.

The CEO’s Board paper proposed six elections. It included reduced transitional terms for Seats 4, 5 and 6 and contemplated a later amendment to article 13.14. It also identified a Special General Members’ Meeting before the AGMM as an alternative. A motion followed: six elections, normal three-year terms for three seats, a one-year term for Seat 4 and two-year terms for Seats 5 and 6. The vote divided four to four, so the motion failed.

After that failure, the minutes say the Board agreed that the Bylaws could be amended if the normal procedure, including community consultation, was followed. The Board also agreed in principle to consider a Special General Members’ Meeting. Those passages matter because they show that rule-first routes were not imaginary solutions invented after the event. Amendment and member involvement were on the table during the decision sequence.

When the meeting continued on 2 March, however, the Board first resolved Seat 6’s immediate status by adding a year. It then separately adopted Resolution 202202.677, asking the Governance Committee to undertake a Bylaws-amendment exercise and recommend workable solutions by the end of April. It also separately adopted Resolution 202202.678, concerning amended Election Process and Guidelines. The narrow sequence is the revealing part: the office-holder outcome was settled before the wider inconsistency was sent onward for repair.

This does not prove that the Board acted with a hidden purpose. It shows the mechanism the Board selected. Faced with a collision between a prior two-year mandate and a three-year tenure provision, it could have pursued rule amendment, a member meeting, another transitional allocation or some temporary bridge. It chose to preserve the incumbent for an additional year. Only after making that choice did it commission the broader bylaw work through a separate resolution.

The strongest benign account must be stated without sarcasm. Article 13.5 said that an elected director’s three-year term expires at the AGMM on or around the third anniversary of appointment. The earlier two-year treatment did not fit that formulation. A supporter could reasonably prefer to honour the stated three-year duration instead of preserving a schedule produced by a casual vacancy. Extending by exactly one year was simple. It did not shorten other directors’ tenures. It could also reduce the operational strain of turning over too many positions together.

Eshun’s recorded argument sharpened the continuity concern. The minutes attribute to him the view that the legal opinions were not the same and the concern that simultaneous elections involving a majority of Board members could produce a quorum problem after the AGMM. For a body that must continue taking decisions, quorum is not decorative. A transition design that could leave a Board unable to function deserves serious attention.

LARUS’s operator-facing governance analysis helps explain the broader principle: predictable institutional controls matter to organisations that depend on registry services, even when a particular governance act cannot be tied to a specific outage or loss.

The available record does not test the forecast. Ojedeji is recorded as replying that a similar configuration involving elections for a majority of seats had occurred before without a subsequent quorum problem. Neither assertion becomes proven merely because it appears in minutes. The supporter identified a risk; the dissenter challenged it with a past comparison. A careful analysis can recognise the legitimacy of continuity planning while still asking what evidence supported this particular intervention.

Continuity also does not decide who should bear the transition cost. If the problem is too many elections at once, there may be several ways to stagger terms. Some approaches shorten future terms, some delay contests, some alter rules prospectively, and some use a tightly limited bridge. Each distributes power differently. “Protect continuity” is therefore a goal, not a complete decision rule. It cannot by itself tell members why the sitting holder of Seat 6 should receive the added year.

The vote exposed how contested that choice was. Ojedeji’s objections are recorded in unusually forceful terms: he called the extension a “coup”, said it was made “arbitrarily”, and objected that it proceeded “without recourse to members”. Those are his attributed words. They are not this article’s findings, not a court’s conclusions and not proof that the resolution was unlawful. Their analytical importance is narrower: a director identified the member-power consequence at the moment the Board acted.

Calling the schedule a control surface gives that concern a precise form. Election rules determine when office-holders must face replacement, which candidates may compete at a given meeting, how many seats members decide together and which directors remain inside the room for later votes. A one-year extension shifts each of those conditions for the affected seat. It may be justified, but it is never merely the correction of a neutral spreadsheet cell.

Heng Lu’s analysis of AFRINIC governance places term extensions inside a larger problem of member control. The relevant insight is not that every extension is capture or that every incumbent is suspect. It is that the timing of accountability can detach practical power from the members who are supposed to exercise it. When the body controlling the immediate calendar also benefits institutionally from preserving continuity, a good process makes the authority, alternatives and limits unusually visible.

The need for visibility increases when one sitting director receives the direct benefit of time. The public record supplied for this decision does not show that the incumbent designed, requested or voted for the extension. It does not contain a beneficiary-recusal analysis. That absence should not be filled with speculation. It should instead shape the procedural question: did the institution isolate the affected office-holder, disclose how any conflict was managed and demonstrate that the remedy was selected independently? A published answer would protect both members and the beneficiary from unsupported inference.

The same discipline applies to the legal opinions. Their existence cannot carry more weight than their disclosed content permits. The minutes tell readers that an internal opinion and a C&A opinion were before the Board, that the CEO said both contemplated six elections, that authenticity and difference were disputed, and that the texts were not reproduced. That supports a conclusion about uncertainty and disclosure, not one about ultimate legality. There is no basis here to declare Resolution 202202.676 lawful, unlawful, void, fraudulent or corrupt.

Nor is there evidence of later judicial review of this resolution. BTW’s account of a later court-supervised election plan belongs to another period and another institutional mechanism. It can clarify the boundary by contrast, but it cannot be used to project later court involvement backward onto the 2022 extension. Likewise, BTW’s comparison of AFRINIC’s Bylaws before and after crisis owns the larger version history. The Seat 6 decision should not be inflated into a substitute history of AFRINIC’s constitution.

Keeping those boundaries intact makes the criticism stronger, not weaker. The case does not depend on proving bad faith or invalidity. It depends on the observable order of operations. The Board had an unresolved schedule conflict. It had considered ordinary amendment, consultation and a special meeting. A transitional motion had failed. It then awarded the sitting term another year and commissioned the general rule repair separately. The control question arises from that sequence alone.

A rule-first approach would have reversed the order. The institution could have published the exact collision between the two-year mandate and article 13.5, stated which version and resolution created each side, set out the available transition designs and explained the evidence for any quorum risk. It could have used the normal amendment procedure or convened a Special General Members’ Meeting before making the election date durable. Members would then have decided the governing rule before the rule produced a personal result.

If time truly made that sequence impracticable, a minimum reversible bridge was available as a governance design. Such a bridge would protect only what was necessary to maintain a functioning transition. It would have a short duration, an automatic end, a fixed review date and timely member ratification. It would disclose the affected seat and office-holder class, isolate the beneficiary from the decision, and state what happens if members reject the proposed amendment. The point is not to demand paralysis. It is to make emergency power visibly temporary and subordinate.

This is a design standard, not an opinion about what Mauritius law compelled. Private companies need ways to handle contradictions and unforeseen vacancies. A rigid insistence that no action occur until every interpretive dispute is settled could itself damage continuity. But institutional agility is safest when its effects can be undone. A bridge that lapses unless members affirm it is easier to defend than a completed transfer of an extra year whose wider rule justification is examined afterward.

The authority-liability principle makes the same point from another direction. The more a private coordinator’s decision changes practical control, the more clearly responsibility should follow the power used. Who decided? On what disclosed authority? Who benefited? What evidence made delay necessary? When could members review the act? How would an error be corrected? These questions keep a registry within an honest corporate frame. They do not pretend that AFRINIC possesses sovereign authority, and they do not ask it to punish anyone.

For operators, the lesson is not that this extension caused technical failure. The sealed public record supports no claim that it produced an outage, loss, service breakdown or later institutional crisis. The relevant connection is more modest. Predictable governance reduces uncertainty around the organisations that maintain essential coordination functions. When election rules are adjusted through a personal term result, participants need a reliable account of how and why that happened. Confidence rests on inspectable controls, not on metaphors of territorial rule.

The watchpoint is therefore the next schedule anomaly, not a retroactive accusation. Will AFRINIC identify the rule conflict before deciding who stays? Will it publish enough of the authority path to let members distinguish legal constraint from managerial preference? Will continuity risks be supported rather than asserted? Will an affected office-holder be isolated? Will a temporary adjustment terminate automatically unless the membership approves a durable rule? Those are practical tests of institutional learning.

Resolution 202202.676 answers one narrow question conclusively: the Board selected an additional year for the incumbent of Seat 6 by a five-to-three vote. It does not conclusively answer the legal questions participants debated. The absence of the opinions, an independent power-to-extend analysis, a member resolution, a recusal account, a full options comparison and a later judicial ruling prevents that. Any account that claims more certainty than the available record is substituting inference for evidence.

The resolution nevertheless permits a firm institutional judgment. The schedule was power because it fixed the next moment of member choice. By resolving the discrepancy through incumbency before members repaired the rule, the Board made a contested interpretation effective in the person holding office. Even on the benign continuity case, that sequence required stronger procedural limits than the public record displays.

The repair need not have been elaborate. Publish the collision. Separate the beneficiary. Explain the opinions’ provenance and disagreement without disclosing protected advice. Compare the options. Demonstrate the minimum continuity need. Set a short expiry. Put the rule to members. Record what finally happened. The required restraint is thin enough for an operating institution and strong enough to preserve member control.

The one-year extension should therefore be remembered neither as proof of villainy nor as calendar housekeeping. It is a compact example of how governance power moves through dates. A private registry’s authority remains narrow, but narrow authority must still be accountable where it affects who governs. The Board made the term match three years. It did so by keeping one incumbent in office while the rule behind the mismatch waited its turn.