Summary

  • AFRINIC’s chief executive was working under a three-year employment contract signed on 20 April 2015 and due to expire on 20 April 2018.
  • The Board had discussed renewal months earlier. Its 26 November 2017 record said the contract should move to a renewable five-year term, while the Remuneration Committee would conduct a performance evaluation.
  • A circular proposal on 18 April 2018 sought a one-year renewal through 20 April 2019. It received five yes votes, two no votes and one recusal, but failed because 62.5 per cent was below the cited two-thirds threshold.
  • At a special Board meeting on 19 April, convened specifically for the CEO contract, directors waived the usual 14-day notice and recorded CEO Alan Barrett’s conflict declaration and recusal from voting.
  • The Board then passed Resolution 201804.407, renewing the employment contract for three years through 20 April 2021 on the same terms and authorising the Board Chair to sign the necessary document.
  • The resolution itself said the Board would finalise the KPIs and the CEO performance evaluation in May 2018. A Remuneration Committee record dated 6 May still treated a detailed KPI report and Board-retreat discussion as an action item.
  • The sequence establishes that a fixed three-year commitment preceded completion of the measurement and evaluation work identified in the decision. It does not establish illegality, poor performance, improper motive, financial harm, an operational incident or any effect on Internet number resources.
  • The missing decision receipts are specific: the 21 March committee recommendation, final KPI set, completed evaluation, signed renewal instrument, fuller recusal and voting record, and any later review connecting measured performance to the renewed term.

The decision’s narrow but consequential hinge

The most revealing sentence in Resolution 201804.407 is not the sentence that renewed the chief executive’s employment. It is the sentence that set the timing of the work still to come. On 19 April 2018, AFRINIC’s Board committed the company to another three years on the same terms, extending the employment relationship through 20 April 2021. In the same resolution, it said that the Board would finalise the chief executive’s key performance indicators and performance evaluation in May 2018.

That order matters because the two acts answer different questions. Renewal answers whether the organisation is committed to retaining an executive, for how long and on what terms. KPIs and an evaluation answer how performance is defined, what evidence is considered and whether past delivery supports the decision. A board can lawfully possess the corporate power to hire and supervise an executive while still leaving outsiders unable to test how it exercised that power. Here, the available records describe the commitment with unusual clarity, yet leave the performance basis materially incomplete.

The finding should be kept exactly at that level. The sequence does not show that the Board had no information. The 19 April minutes recorded the Remuneration Committee’s view that the chief executive had performed despite weaknesses in some areas. Nor does the sequence show that those weaknesses were disqualifying, that a later evaluation was negative, or that the renewal terms were disadvantageous. It shows something more limited and more demonstrable: the organisation fixed the duration and continued the terms before completing the final KPI and evaluation work named in its own resolution.

This is a receipts problem because the missing material prevents a reader from reconstructing the decision chain end to end. What performance criteria existed before 19 April? What changed when they were finalised? What evidence underpinned the committee’s recommendation? What exactly did “same terms” preserve? Was the signed instrument consistent with the resolution? How was the recusal operationalised, and how did the eligible directors vote? The records reviewed here do not answer those questions. A disciplined account must neither fill the gaps with suspicion nor treat the gaps as irrelevant.

The expiry date was known, and discussion began months before it

The deadline was not obscure. The 19 April record states that the chief executive had been employed under a three-year contract signed on 20 April 2015 and expiring on 20 April 2018. The contract’s end date therefore provided a fixed point around which the Board and its Remuneration Committee could organise evaluation, negotiation and approval.

The surviving Board record shows that the matter was under discussion by 26 November 2017. At that meeting, the approaching April 2018 expiry was raised and the Remuneration Committee was expected to conduct a performance evaluation. The same record says the Board decided the chief executive’s contract should move from three years to five years renewable, with the committee to report on the evaluation. It also records a request to see the chief executive’s contract and a response that the document would be shared confidentially.

Those entries establish three important conditions without resolving the later sequence. First, the institution had several months of notice before expiry. Second, duration and evaluation were already connected in the Board’s discussion: a proposed move to five years sat alongside a promised committee report. Third, at least one participant regarded access to the underlying employment document as relevant, while the Board chose confidential rather than open sharing.

None of that proves that the delay later described by the committee was avoidable, blameworthy or strategically motivated. The records do not explain the work calendar between late November and March. They do not reveal the confidential contract’s contents. They do not show what preliminary measures the committee may have been using. But they do make it impossible to describe the 19 April decision as an entirely new matter forced upon directors without prior awareness. Renewal had been identified well in advance, and evaluation had been assigned as part of the preparatory work.

The November position also provides a meaningful comparison with what followed. Five years renewable was the direction then recorded. By 21 March 2018, according to the committee chair’s report at the later special meeting, the Remuneration Committee was recommending three years instead. On 18 April, a circular resolution proposed only one year. On 19 April, the Board adopted three years. The term length moved across three distinct positions. The reviewed material establishes that those positions existed, but it does not supply the reasoning paper that would explain the movement from five to three to one and back to three.

The ordinary meeting on 18 April did not contain this resolution

Chronology becomes especially important across 18 and 19 April. AFRINIC held an ordinary Board meeting on 18 April 2018. The minutes reviewed for that meeting covered financial statements, draft minutes for a prior Annual General Members’ Meeting and the notice for the 2018 Annual General Members’ Meeting. They did not record Resolution 201804.407.

That absence is not a claim that the ordinary meeting was defective or that the chief executive’s contract could only have been discussed there. It serves a simpler purpose: it prevents the two dates and two decision mechanisms from being collapsed into one. The renewal ultimately adopted was considered at a special Board meeting on 19 April whose agenda was the CEO contract. The preceding day matters because it also carried a different attempted action, a circular proposal for a shorter term.

Treating the events as separate preserves the vote mechanics. The 18 April circular resolution proposed renewing the contract for one year, through 20 April 2019. It was recorded as receiving five yes votes, two no votes and one recusal. The minutes calculated five of eight directors as 62.5 per cent and said this fell below the two-thirds threshold drawn from Article 19.12 of the bylaws. On that basis, the circular proposal failed.

The failure tells us what did not happen: the chief executive did not receive the proposed one-year renewal through that circular vote. It does not tell us why two directors voted no, why five preferred yes, whether the objections concerned term length, process, evidence or something else, or how each director was identified. The detailed identities behind those positions are not specified in the minute text reviewed here. Any effort to supply motives would exceed the evidence.

Nor should the circular proposal be portrayed as the same substantive decision later carried into a room and approved. It was a one-year proposal with a recorded supermajority threshold problem. The special-meeting resolution was a three-year renewal on the same terms, passed under the mechanics recorded for the 19 April session. The change in duration is not a cosmetic difference. It multiplied the period of commitment and aligned the result with the committee’s reported recommendation rather than the previous day’s shorter proposal.

The 19 April special meeting had an explicit agenda and an explicit recusal

The decisive session was not of unknown character. It was a special Board meeting on 19 April 2018, and its agenda was the CEO contract. The directors waived the 14-day notice normally required for the meeting. The minutes also recorded that CEO Alan Barrett declared a conflict of interest and would not vote.

Those facts are essential because accelerated procedure and conflict handling are two separate governance surfaces. Waiving notice addresses how quickly the Board could convene. Recusal addresses whether the person whose employment was at issue took part in the vote. The record says both occurred. It would be wrong to erase the recusal and insinuate that the chief executive voted on his own renewal. It would be equally incomplete to cite the recusal as if it resolved every question about participation, information flow or the reasons for accelerating the meeting.

The minutes provide a basic motion trail. The three-year renewal motion was proposed by HE, seconded by SM and recorded as passed unanimously among the voting directors while the chief executive was recused. This differs from the circular record, which provided an aggregate five-yes, two-no and one-recusal tally. The special-meeting text establishes unanimity among those voting but, in the material reviewed, does not reproduce a fuller individual voting schedule or a more detailed conflict-handling protocol.

The 14-day notice waiver should also be read in context rather than turned into a verdict. The prior contract was expiring on 20 April, one day after the special meeting. That proximity makes urgency visible, but the records also show that expiry had been under discussion since at least November. The existence of urgency does not prove that the waiver was improper. It does sharpen the question of process design: when a known deadline results in a last-day commitment, the minutes and supporting documents carry more of the burden of explaining how directors had enough time and information to decide.

The combination is therefore neither exculpatory nor accusatory. There was a disclosed conflict and recorded recusal. There was also a waived notice period and a materially longer term than the failed proposal from the day before. A trustworthy decision history needs both sides of that picture. Governance analysis loses accuracy if it treats every safeguard as conclusive proof of sound judgment, or every compressed step as proof of misconduct.

The committee recommendation and the evaluation were not the same document

At the special meeting, the Remuneration Committee chair traced the matter back to the November–December 2017 Lagos meeting and said delay had followed. The chair also said that on 21 March 2018 the committee recommended a three-year renewal instead of the five-year term originally decided in Lagos. The minutes recorded the committee’s view that the chief executive had performed, despite weaknesses in some areas.

This passage matters because it shows that the Board was not acting without any articulated recommendation. It had a committee position on duration and a general assessment of performance. Yet a recommendation to renew is not identical to a completed performance evaluation, and the Board’s own language preserves that distinction. The minutes said the performance evaluation would be completed and presented at the forthcoming face-to-face meeting in Dakar. Resolution 201804.407 then said the Board would finalise both the KPIs and the CEO performance evaluation in May.

The reviewed evidence does not include the committee’s 21 March recommendation paper. It therefore cannot answer whether the recommendation set out measurable objectives, compared delivery against previously agreed standards, discussed the weaknesses mentioned in the minutes, evaluated alternative term lengths, or addressed the continuation of existing conditions. The fact that a recommendation existed is established by the special-meeting account. Its detailed analytical basis is not.

This distinction prevents two opposite errors. One error would be to say there was no performance consideration at all. That is contradicted by the committee chair’s reported view. The other would be to treat the general statement that the chief executive had performed as equivalent to the final KPI set and completed evaluation that the resolution itself placed in May. That would erase the sequencing gap with a document the record does not provide.

“Despite weaknesses in some areas” also requires restraint. The phrase cannot support a finding of poor overall performance. It does not identify the areas, their severity, whether they were corrected, or how they affected the committee’s recommendation. It may signal that the assessment was not unqualified, but it is not a licence to invent deficiencies. The proper question is documentary: how were those acknowledged weaknesses represented in the final evaluation, and how did the final KPIs respond to them? The answer is unknown from the material reviewed.

What Resolution 201804.407 actually committed

Resolution 201804.407 did three concrete things relevant here. It renewed the chief executive’s employment contract for another three years, through 20 April 2021. It kept the same terms of employment. And it authorised the Board Chair to sign the document necessary to give effect to the renewal. The motion was recorded as passed unanimously by the voting directors, with the chief executive recused.

Each element carries a different documentary consequence. The three-year duration defines the period of commitment. “Same terms” incorporates an existing set of conditions by reference, but the full content of those conditions is absent from the reviewed record. Signature authority identifies who could execute the necessary instrument, but the signed renewal document itself is also absent from the material considered. A complete chain would allow a reader to compare the resolution, the pre-existing contract and the executed renewal instrument. Here, only the high-level instruction is visible.

It would be improper to infer salary, benefits, termination rights or any other remuneration provision from the words “same terms.” The phrase indicates continuity, not the details of what continued. There is no evidentiary basis here for a claim of increased pay, reduced protections, unusual benefits, hidden penalties or financial loss. Nor is there enough to judge whether the unchanged conditions were favourable or unfavourable to AFRINIC. The gap is one of visibility, not a discovered defect in the terms.

The fixed duration is clearer. Three years is not a provisional bridge to a later evaluation. The resolution did not say that renewal would become effective only if the May work was satisfactory. It committed the organisation through 20 April 2021 while stating that the KPIs and performance evaluation would be finalised the following month. Nothing in the reviewed record establishes a condition subsequent, reopening clause or later confirmation vote. It would be equally wrong to assert that no such mechanism existed in the unseen contract. The honest conclusion is that the visible resolution does not supply one.

The signature provision is important for a further reason. Governance does not end when a motion passes. The decision must be translated into an instrument that matches what directors authorised. Without the signed document, one cannot verify execution date, exact wording or consistency with “same terms.” This is not evidence that the Chair exceeded authority. It is a reason to seek the instrument that would prove faithful implementation.

May remained the month of unfinished measurement work

The resolution’s forward reference to May was not merely abstract. A Remuneration Committee record dated 6 May 2018 still listed a detailed KPI report and discussion at the Board retreat as an action item. It also said that the chief executive’s contract and salary had not yet been discussed in that committee meeting.

This later entry reinforces the chronology without answering every question. It supports the finding that KPI work remained active after the three-year renewal. It does not show that the Board failed to complete the work later in May. It does not reveal the final KPI language, the completed evaluation, its date, or whether either was eventually adopted and used. It also does not establish anything about salary. The statement that contract and salary had not yet been discussed in that particular meeting cannot be converted into a claim that salary was never discussed anywhere or that the renewal changed remuneration.

The 6 May action item is especially useful because it prevents the later work from being treated as a drafting formality already complete on 19 April. A detailed KPI report and Board-retreat discussion were still to occur. Yet the action item’s existence cannot disclose the content or significance of the future discussion. It preserves the same bounded conclusion: the decision had already locked in the renewed term, and measurement work continued afterward.

Whether that order was a conscious governance choice or a product of deadline pressure is not established. The committee chair referred to delay after the Lagos discussion. The contract expiry created an immediate need by 20 April. Directors may have believed that continuity warranted renewal while measurement details could follow. They may have relied on performance information not reproduced in the public record. Those are possible explanations, not proven facts. The institution could clarify them by releasing contemporaneous material rather than asking readers to infer intent.

The corporate power existed; the public-authority story does not

AFRINIC’s 2016 bylaws placed the company’s business and affairs under Board direction and empowered the Board to appoint a chief executive on terms and conditions it determined. They also described the chief executive as managing day-to-day business and reporting to the Board. That is the relevant authority frame for Resolution 201804.407: an ordinary private-company board exercising employment and supervision powers within its corporate structure.

Heng Lu’s controlling doctrine draws a necessary line around that corporate competence. In his account, a regional Internet registry is a narrow private technical bookkeeper and coordinator for number-resource records. Practical reliance on its registry function can make its decisions consequential, but dependency does not transform the institution into a sovereign, regulator, police force, prosecutor, punitive body, confiscating authority or court. Registration is not ownership, and repeated institutional conduct is not proof of public authority.

Applied here, the doctrine produces two conclusions that must remain together. First, the Board’s power to employ and supervise a chief executive should not be denied merely because AFRINIC is a private registry. The bylaws provide a corporate basis for that decision. Second, the importance of registry operations does not elevate a CEO-renewal resolution into an exercise of governmental power over the Internet or number resources. Scrutinising the decision’s documentation is not the same as claiming the Board needed sovereign authority to make it.

That boundary also protects the analysis from inflated causal claims. Resolution 201804.407 is not evidence of an outage, a routing effect, resource loss, an allocation decision or harm to Internet operations. The reviewed material establishes none of those outcomes. The significance lies in institutional accountability: an operationally relied-upon private coordinator should be able to show the chain by which its own leadership commitments were authorised, informed, executed and later assessed.

Operational centrality raises the value of precision because users may have few practical substitutes for registry coordination. It does not change the legal character of the employment act. The remedy is therefore not to invent public powers or public-law offences. It is to demand a narrow, testable chain: governing instrument, meeting authority, notice treatment, conflict declaration, eligible vote, performance inputs, signature authority, executed instrument, monitoring and a route to correction.

What the sequence proves, and what it leaves open

The proven sequence can be stated without adjectives. A contract signed on 20 April 2015 was due to expire on 20 April 2018. By November 2017, the Board had discussed the expiry, indicated a five-year renewable direction and assigned a performance evaluation to the Remuneration Committee. The committee later recommended three years on 21 March 2018. A one-year circular proposal failed on 18 April after five yes votes, two no votes and one recusal did not reach the cited two-thirds threshold.

At a special meeting on 19 April, with notice waived and the chief executive recused, the voting directors unanimously adopted a three-year same-terms renewal and authorised the Chair to sign. The same resolution put final KPI and evaluation work in May. A 6 May committee record still carried a detailed KPI report and retreat discussion as an action item.

From that sequence, one conclusion follows securely: commitment preceded finalisation of the performance framework and evaluation identified in the resolution. The conclusion remains true whether the eventual KPIs were demanding or lenient, whether the final evaluation was positive or mixed, and whether directors believed renewal was obviously warranted. Those later possibilities do not alter the recorded order.

Several stronger claims do not follow. The order does not prove the renewal was unlawful or invalid. It does not prove the Board lacked power under the bylaws. It does not prove the chief executive was unsuitable or performed badly. It does not prove an improper motive, self-interested conduct by the recused executive, a breach of duty, financial loss or operational harm. It does not prove the later evaluation was never completed. It does not prove an effect on routing or Internet resources. Each of those claims would require evidence absent here.

The distinction between a procedural vulnerability and a legal conclusion is central. A board may decide that continuity, market conditions, institutional knowledge or other considerations justify renewal before every metric is complete. None of those considerations is established in the present record, so none can be credited as the actual reason. But their abstract possibility demonstrates why sequence alone cannot decide legality or wisdom. What sequence can do is identify the explanation and documentation that accountable governance should provide.

There is also a counterargument based on the committee’s reported judgment. If the committee had already concluded that the chief executive performed despite some weaknesses, perhaps the May evaluation only formalised an assessment sufficiently made. The difficulty is that the resolution itself did not describe the later work as mere formatting. It said KPIs and performance evaluation would be finalised in May, and the 6 May record still anticipated a detailed KPI report and retreat discussion. The counterargument remains plausible, but the missing recommendation and evaluation prevent it from being tested.

A second counterargument is that the contract’s imminent expiry forced the Board to preserve continuity. The timeline supports the existence of deadline pressure. Yet the November record also shows advance notice of the expiry and an assigned evaluation process. Urgency may explain the special meeting and notice waiver, but it does not explain why the preparatory work was unfinished after months of lead time. Again, the answer may be benign. The contemporaneous records needed to establish it are not present.

A decision-receipt chain that can be tested

The right request is concrete rather than speculative. First is the Remuneration Committee’s 21 March 2018 recommendation. Its release would show the evidence considered, the rationale for three years, the treatment of acknowledged weaknesses and the reason for departing from the five-year direction recorded in Lagos. If the paper contains legitimately confidential personal information, an appropriately limited disclosure could still preserve its reasoning and criteria.

Second is the final KPI set. Because the resolution expressly deferred finalisation, dates and version history matter. A KPI document created or approved after 19 April cannot retroactively become a precondition of the renewal, though it may show how the renewed term was subsequently governed. The useful questions are when the indicators were drafted, who approved them, how they differed from any earlier measures and whether they were tied to review points during the three-year term.

Third is the completed performance evaluation. It would establish what period was assessed, which inputs were used, how the committee addressed strengths and weaknesses, and when the Board received the result. The reviewed material does not contain the evaluation or its completion date. Absence from the material is not evidence that it never existed; it is the reason the later supervisory chain cannot now be verified.

Fourth is the signed renewal instrument. That document would allow comparison between the Board’s instruction and what was executed. It could confirm the effective dates, continuity of terms and any provisions relevant to later performance management. No conclusion about salary, benefits or termination can be drawn until the actual text is available.

Fifth is the fuller conflict and vote record. The minutes already establish that Alan Barrett declared a conflict and did not vote. Any attendance record, conflict register entry or individual voting schedule could show how recusal was implemented and which eligible directors supported the final motion. Such material would supplement, not negate, the safeguards already recorded.

Sixth is any later review that connected actual performance to the renewed term. The value of final KPIs depends on their use. If they were approved after commitment, did the Board monitor against them, document variances and preserve a route to corrective action? The present record cannot answer. A later review could show whether deferred measurement became meaningful oversight or remained an unattached document.

This chain is falsifiable. If the dated committee paper contains a developed evaluation and explains why the later work was only formal completion, that would narrow concern. If the signed instrument makes renewal conditional on a later Board step, that would alter the apparent firmness of the 19 April commitment.

If dated KPI records show substantive approval before 19 April despite the resolution’s wording, the chronology would need closer interpretation. Conversely, if final metrics were created only after the contract became unconditional and were never used in review, the sequencing weakness would become more pronounced. The point is to identify evidence that can change the assessment.

The boundary with adjacent AFRINIC stories

Resolution 201804.407 sits near several other governance subjects, but proximity is not identity. Earlier coverage addresses the 2014 CEO search and its charter. Separate coverage addresses the general design of the Remuneration Committee. Another article addresses confidential-evidence access around a March 2018 resolution. Those matters help define the institutional setting, but they do not supply the missing reasoning for the April contract renewal.

This account therefore does not revisit the original search or judge the broader committee charter. It does not absorb the March 2018 investigation, resignation, accounts or confidential-evidence disputes into the renewal decision. It does not use later leadership events to infer what directors knew or intended on 19 April. The subject is the 2017–2018 renewal sequence and only that sequence.

Maintaining that boundary is more than editorial tidiness. Governance failures are easily exaggerated when separate episodes are stacked until context becomes accusation. They are also easily minimised when each episode is isolated so completely that recurring documentation questions disappear. The disciplined middle course is to name adjacent records, leave their theses intact and require Resolution 201804.407 to stand on its own decision receipts.

On its own terms, the resolution presents a sharp institutional test. The Board had corporate authority. The conflict was declared and the chief executive recused. The voting directors approved the motion. Yet the commitment was for three years on unchanged conditions, and the performance measures and evaluation named in that same instrument were still to be finalised. The missing documents do not prove wrongdoing. They prevent the organisation from demonstrating, through the visible chain, how performance assessment informed commitment rather than followed it.

Sources