Summary

  • AFRINIC’s Board recorded Resolution 202001.547 in January 2020, adopting amended Election Processes and Guidelines. The public record does not establish the exact adoption day: 24 January is the publication date of the approval notice, while 17 January appears only as the PDF’s “Proposed Changes” date.
  • The rulebook divided responsibility among a four-person, Board-appointed NomCom, a staff-only Election Committee appointed by the CEO, three electronic-vote trustees, the CEO and the Board. That division created useful role clarity without fully separating appointment, screening, administration and review.
  • Nomination was only an entry point, not candidacy. NomCom applied criteria, could interview and check nominees, considered diversity, sought at least two candidates per seat as a Board recommendation, received nonbinding Board input and then finalised the slate.
  • Voting access depended on membership class, formalities and payment status. The text joined electronic and paper totals, confined proxies to paper voting, removed an electronic vote when a proxy was granted and imposed specific approval, signature, timing and volume rules on proxies.
  • The count rules were detailed but contained a consequential asymmetry: an initial result combined electronic and paper votes, while a tie moved to repeated paper-only voting among tied candidates and did not carry the earlier electronic votes forward.
  • The strongest case for the design is procedural competence and predictability; the strongest concern is concentrated discretion at control points. The proportionate response is better reasons, conflict disclosure, verification records and independent review—not treating AFRINIC’s private technical coordination as public authority.

January had three clocks, not one adoption date

The first discipline in reading Resolution 202001.547 is chronological. AFRINIC’s 2020 Board resolutions page places the resolution under January and says the Board, after reviewing the Election Processes and Guidelines, resolved to adopt the new guidelines as amended. That establishes the institutional act and the month. It does not supply a day, a roll call or a vote count. AFRINIC’s approval notice appeared on 24 January 2020, which establishes when that notice was published. The adopted twelve-page PDF carries the line “Proposed Changes: 2020/01/17,” which establishes a date printed on the document. None of those records permits the three dates to be collapsed into an invented exact adoption day.

This separation is more than archival caution. Governance analysis often becomes unreliable when a document date, a decision date and a publication date are treated as interchangeable. A draft may be dated before formal approval; a notice may appear after it; a resolutions index may state only a month. The sequence available here is bounded but sufficient: a document carried proposed changes dated 17 January, the Board recorded adoption in January, and an approval notice was published on 24 January. The evidence says no more about the exact decision day.

What the Board adopted was not merely a statement of electoral aspiration. The guidelines described a chain of private institutional actions: forming NomCom, opening nominations, converting eligible nominees into a candidate slate, defining eligible voters, administering electronic and paper channels, handling proxies, counting ballots, resolving ties, referring disputes and announcing results. The Board described the text as approved under article 13.2 of AFRINIC’s Bylaws, applicable to Board seats and, with necessary changes, to other elections following the same process.

It also retained the ability to provide additional guidelines under article 9.2.

That scope made the resolution an ex ante operating instrument. Before anyone could know who might win, it set out the gates through which nominations, votes and results were meant to pass. Ex ante rules do not prove that administrators later followed them. They do something narrower and indispensable: they create a benchmark against which actions can be explained. A member can ask whether the same standing test was applied to all voters. A nominee can ask whether the published criteria governed screening. An observer can ask whether a proxy met the stated conditions or whether a timetable change was genuinely minor.

Without a prior rule, those questions deteriorate into competing memories of what seemed acceptable at the time.

The institutional stakes should also be described without inflation. AFRINIC is a private technical registry, bookkeeper and coordinator. Its internal elections choose people for offices within that organisation. They do not create a sovereign electorate, enact public law, confer police power or turn membership participation into jurisdiction over the public. Yet private procedure can still matter greatly. Operators rely on registry services, and Board choices can influence budgets, staffing, operational priorities and service continuity.

The rulebook’s credibility therefore mattered as a condition of dependable private administration, not as an exercise of governmental authority.

A divided process, with appointment power still at the centre

The January design allocated different tasks to different bodies. NomCom was the principal screening and supervisory body. Each year it consisted of four people appointed by the Board: one Board member and three people selected after a call to the community. One of those three community members served as chair. The Board representative was not an observer. That person held a full vote and could report concerns about candidate qualifications to the Board, although confidential discussion of comparative merit was not to be reported.

NomCom decided by majority. If a tie remained unresolved, the chair had a casting vote. In a four-member committee, those provisions matter. Majority rule supplies an ordinary decision method; a casting vote prevents paralysis. At the same time, the mechanism gives the chair additional decisional weight at the point of deadlock. The rule is transparent on its face, which is preferable to improvising after division occurs. But transparency about a power does not eliminate the need to know when it was used, what conflict rules applied and how the committee distinguished qualification questions from confidential judgments of merit.

Staff supported NomCom but were not to participate in deliberations about candidate merits. Election logistics belonged to the Election Committee, or ECOM. ECOM was staff-only, appointed by the CEO, and its chair was likewise chosen by the CEO. The separation was intelligible: NomCom could concentrate on nominations and supervision, while staff with administrative capacity could run identity checks, ballots and other operational details. It also meant that the institution’s executive side controlled the appointment of the body performing the process, while the Board controlled appointment of the body supervising it.

Electronic voting added another small group. The CEO, the NomCom Chair and Legal Counsel served as trustees, and their signatures accompanied the electronic tally. Paper counting involved a NomCom representative, ECOM and as many as four volunteers from other RIRs. Results were announced by the NomCom Chair; the CEO was responsible for emailing the result within forty-eight hours and publishing it online at the same time. Proxy instruments required CEO approval at least twenty-four hours before voting.

This was not a simple hierarchy in which one office did everything. It was a network of dependencies. The Board appointed NomCom and could provide nonbinding input on nominations. NomCom screened candidates and supervised the election. ECOM ran the mechanics. Trustees held responsibilities around the electronic tally. The CEO appointed ECOM, approved proxies and published results. Legal Counsel participated as a trustee. Volunteers from other RIRs could join the paper count. Splitting tasks can make interference harder because more than one role must leave a record.

It can also obscure responsibility if every actor points to another link in the chain.

The central design question is therefore not whether duties were divided. They clearly were. It is whether the division created checks that were sufficiently independent and visible. Board appointment of all four NomCom members supplied a line of accountability, but it also connected the slate-forming body to the institution whose seats were being filled. A staff-only ECOM promised competence, but its appointment by the CEO made documentation of operational decisions more important. Three trustees spread custody of the electronic tally, but the sealed record contains no ceremony record or independent security examination.

The text provided an architecture. Its credibility in use would depend on records showing how the people inside that architecture exercised their powers.

Nomination opened the door; NomCom controlled passage to the slate

The rulebook drew a useful distinction that casual election language often erases: a nomination was not candidacy. A nomination brought a person into consideration. NomCom’s screening and finalisation process determined whether that person appeared on the candidate list. That distinction located a major control point before the ballot itself.

Each member could nominate one person for each open seat through authorised contacts. The text also allowed self-nominations and nominations from people who were not authorised contacts, but those routes required support from two members. The design balanced two interests. Authorised member contacts received a straightforward route to submit a name, while individuals outside that route were not categorically excluded. Requiring two member supporters gave the alternative path a threshold of institutional backing.

Thresholds of this kind can be defended as filters against frivolous submissions. They can also affect who reaches the screening stage, especially when access to member contacts is uneven. The available evidence does not establish how many people used each route, whether support was difficult to obtain, or whether anyone was excluded. It does establish the rule against which such outcomes would need to be assessed. A credible administration would preserve the nomination source, authorisation status, supporter evidence and the reason any submission failed the threshold.

NomCom then applied criteria, could interview nominees and could conduct checks. It was expected to consider diversity and to work toward at least two candidates per seat, which the Board framed as a recommendation. The difference between a binding minimum and a recommended objective is important. A binding rule would make a one-candidate slate presumptively noncompliant unless an exception existed. A recommendation sets direction but permits departure. That flexibility may be necessary if few qualified people are willing to stand; it also makes reasons for departure essential if the objective is not met.

The Board could offer input on nominations, but the text described that input as nonbinding. NomCom finalised the candidate list after receiving it. On paper, that preserved a distinction between institutional advice and the slate decision. The Board could identify a qualification issue or convey its view without formally choosing the candidates. The difficulty lies in observing influence that is expressly nonbinding. A final decision can legally or formally belong to one body while another body’s views carry practical weight because it appointed the decision-makers or controls resources around them.

The answer is not to assume that advice became command. The evidence does not support that claim. The answer is to identify the record that would let an observer test the distinction: the criteria in force before nominations opened; conflict declarations; the issues the Board characterised as qualification concerns; NomCom’s disposition of those concerns; and reasons for acceptance or rejection that do not reveal confidential personal material. Such a record would protect both sides. It would make improper pressure easier to detect while preventing every adverse screening decision from being recast as hidden Board intervention.

The public comment period added a further input channel. Comments could remain open until the day before the election, and inappropriate comments could be removed. Late availability gave members time to respond to developments, but removal authority introduced a judgment about what counted as inappropriate. Moderation can protect candidates from abuse, irrelevance or harmful disclosures. It can also affect the public information environment if standards and reasons are unclear.

A published moderation policy, a removal log that protects sensitive content and a timely review route would preserve the legitimate function while making inconsistent use more visible.

Withdrawal rules likewise anticipated timing problems. A candidate withdrawing at least seven days before electronic voting could be removed. A later withdrawal remained on the ballot, although votes cast for that person were ignored. That rule may have protected ballot stability after voting preparations had advanced. It also created a situation in which a visible choice could no longer produce a valid vote. Clear notice to voters would be crucial. The sealed evidence establishes the rule, not whether any withdrawal occurred or how any notice was delivered.

Eligibility translated membership records into voting access

Election legitimacy is shaped not only by ballot secrecy or counting but by the administrative definition of the electorate. The January guidelines made eligibility depend on membership category and standing. Good standing required completion of formalities, current membership and payment of membership and allocation fees already due. Invoices not yet due were disregarded. Resource Members in good standing and all Registered Members could vote under the rulebook’s stated terms.

This is a bookkeeping-heavy definition, as one would expect in a private membership organisation. It depends on records of membership status, contact authority, invoices, due dates and payment. None of those determinations is inherently political, but each can determine whether a person or organisation reaches the ballot. A payment posted late, an outdated contact record or disagreement over completed formalities can become an election issue when the consequence is disenfranchisement inside the organisation.

The rule’s treatment of invoices not yet due is a significant limiting detail. It prevented future obligations from being used to classify a member as not in good standing before payment was required. That made the relevant date and account state more precise. Even so, the available record does not show how disputed invoices, late reconciliations or recent status changes were treated. The appropriate inference is not that the system succeeded or failed, but that a reproducible eligibility snapshot would be important: what records were checked, at what time, under which due-date rule and through what correction process.

A corporate Resource Member was represented by its registered contact without a proxy. That placed considerable importance on the register of contacts. The difference between direct representation and proxy authority turns on who the system recognises as entitled to act. For natural persons and organisations alike, the quality of the contact record could shape practical access. A private registry can legitimately require formal representatives; it should also provide timely means to correct a record before the voting cut-off.

Candidate status did not itself produce a vote. A candidate who was not otherwise an eligible member voter could not vote merely because the person was standing. A candidate could, however, vote for organisations represented by that candidate if the person had been the registered contact for twelve months, or since the organisation became a member if membership was newer. These details separated the right to stand from the right to vote and tried to prevent a last-minute representational change from manufacturing electoral weight.

Again, the text created a benchmark rather than evidence of performance. There is no sealed turnout figure, voter list, rejected-voter ledger or correction record. It would be wrong to call the electorate inclusive or exclusive in practice, just as it would be wrong to infer confidence from the existence of detailed rules. What can be said is that the guidelines located voting entitlement in verifiable corporate records. That choice reduced some ambiguity while making the accuracy, timing and reviewability of those records a central governance concern.

Electronic access, paper presence and the proxy trade-off

The rulebook offered electronic and paper voting. It limited proxies to the paper channel and encouraged electronic voting over paper ballots and proxies. This was a hybrid design, but the channels were not interchangeable. Each carried different requirements, intermediaries and risks.

Electronic voting was integrated with MyAFRINIC. It used the same candidate list as paper voting, began at least seven days before the election and closed at the same time as paper voting. That period allowed eligible members to vote without being physically present at the AGMM. In a region of large distances and uneven travel costs, remote access has an obvious practical defence. The text did not, however, establish how many members used it, whether access was reliable or how the system was independently tested.

The CEO, NomCom Chair and Legal Counsel were trustees for electronic voting, and their signatures accompanied the system tally. Three roles sharing responsibility can reduce dependence on one individual. Their institutional relationships still matter: the CEO also appointed ECOM and approved proxies, the NomCom Chair led the screening and supervisory body, and Legal Counsel was an internal legal role. Trustee multiplicity is not identical to external independence.

The value of the arrangement would be strengthened by evidence of the tally ceremony, access controls, reconciliation and exception handling, none of which is available in the sealed record.

Paper voting was designed for the AGMM. ECOM validated voters and distributed secret ballots that were stamped and signed. A ballot containing personal identifying information was invalid, and multiple marks for one seat spoiled it. These rules protected secrecy and unambiguous choice. A stamp-and-signature requirement also created physical controls over which papers entered the count. It would require disciplined reconciliation: ballots issued, ballots returned, spoiled ballots, unused stock and valid votes should connect in a coherent record.

Proxy voting offered a way for an absent member to participate through an eligible voter at the meeting. The CEO had to approve the proxy at least twenty-four hours before voting. A person could carry no more than five proxies, and candidates could not carry them. A proxy could be produced through MyAFRINIC or on a member’s letterhead. For a letterhead proxy, the registered administrative contact or a director had to sign. A corporate seal was not mandatory, and natural persons did not need letterhead.

Those details reduced uncertainty about form. They also gave the CEO an approval gate close to the election. An approval requirement can prevent forged, incomplete or ineligible instruments from entering the count. It can become a point of unequal treatment if similar documents receive different outcomes or if rejection reasons arrive too late to correct. The proportionate safeguard is not to eliminate verification; it is a recorded, consistently applied test with swift notice and a review route capable of acting before the vote.

The five-proxy limit constrained concentration. It prevented one eligible voter from arriving with an unlimited block of delegated votes. The candidate prohibition separated candidacy from the role of proxy carrier. Yet proxy rules still redistributed practical influence toward people able to attend the paper vote. Their significance becomes clearer when combined with the channel-switch rule: giving a proxy removed the member’s electronic vote. That prevented double voting, but it also meant a member choosing representation committed to the paper channel and its human intermediary.

The design therefore made channel choice consequential. An electronic voter acted directly through MyAFRINIC over a period of at least seven days. A paper voter acted at the AGMM. A proxy giver delegated a paper vote, lost the electronic entitlement and depended on timely approval. The initial count later combined electronic and paper totals, but the routes into those totals were institutionally different. Credibility depended on members understanding those differences before choosing a channel and on administrators preserving a traceable account of approvals, rejections and ballot reconciliation.

The count combined channels, but a tie changed the decisive electorate

At the first count, NomCom supervised and ECOM performed the election process. Paper ballots were counted by a NomCom representative, ECOM and up to four volunteers from other RIRs. Electronic and paper totals were then added. Abstentions were recorded and announced. The NomCom Chair announced the result while separating the electronic and paper counts, and the CEO was to email results within forty-eight hours and publish them on the website simultaneously.

Several choices here support intelligibility. A combined total made both valid voting channels count toward the ordinary result. Separate disclosure of electronic and paper numbers allowed observers to see the channel composition rather than only a final aggregate. Recording abstentions preserved the difference between participation without candidate choice and nonparticipation. A defined announcement role and publication deadline reduced room for indefinite delay.

The paper count included people from multiple roles, including potential volunteers from other RIRs. That could provide additional witnesses and technical familiarity. “Up to four” did not guarantee that four would participate, and affiliation with another RIR did not by itself create an independent assurance function. The instrument identifies who could be present; it does not provide a count record, observer report or reconciliation. It would be incorrect to call the election secure or independently verified merely because the rule allowed several counters.

The tie rule created the sharpest channel asymmetry in the document. If candidates tied, repeated voting among the tied candidates took place using paper only. The earlier electronic votes were not carried forward. This was not just another count of the same electorate. It changed the decisive participation mechanism from a hybrid electorate—including members who voted remotely—to those able to cast a paper ballot or be validly represented in the meeting-based channel.

No misconduct is required for that shift to matter. Imagine two candidates tied after remote and meeting votes were combined. The subsequent choice would be made without the electronic ballots already cast. Members who had used the encouraged electronic channel might not be physically present and, having already acted, might not have arranged a proxy. The paper-only rerun could therefore assign the final choice to a differently composed participating group. The evidence does not show that a tie occurred, so this is a design consequence, not a claim about a realised result.

There are benign explanations for the rule. A tie needed to be broken promptly at the AGMM; reopening an electronic system might have created delay or technical complexity; the meeting provided an available group of eligible voters. Those are practical considerations. But the rule’s consequence should have been explicit to members before voting, and a more channel-consistent alternative was conceivable. A tie process could preserve the earlier electorate’s participation, reopen both channels under a defined timetable or use another precommitted method that did not make physical presence decisive only after a tie emerged.

Disputes were assigned to Bylaws article 10.2. Cross-reference to an existing provision is better than silence because it identifies a formal route. The available material does not establish how independent, timely or remedially effective that route was in relation to an election. A dispute mechanism is not credible solely because it has an article number. Its practical value depends on standing, deadlines, access to relevant records, conflict handling, reasoned decisions and remedies capable of addressing an error before it becomes irreversible.

The result-publication provision also needs precise treatment. The rule says the CEO was to email the result within forty-eight hours and publish it online simultaneously. It does not prove that this happened in 2020. Nor does the printed schedule prove that electronic voting opened on 29 May or that an election occurred on 11 June. The schedule contemplated NomCom work from 22 January, electronic voting from 29 May and election on 11 June, while also saying it should be followed as far as practical and that minor deviations did not require Board approval.

Those were planned and discretionary parameters, not observed events in the available record.

The strongest defence of the January architecture

The rulebook deserves to be tested against its best justification, not against a caricature. A four-person NomCom could be small enough to deliberate efficiently. Board appointment of its members could create a clear accountability line rather than leaving formation to an undefined crowd. Selecting three members after a community call introduced people beyond the sitting Board, and requiring one of those three to chair limited the Board representative’s formal control of the committee.

The Board representative’s full vote can be defended as a way to connect NomCom to the legal and operational responsibilities of the organisation. The restriction on reporting confidential merit deliberations recognised that nominees need some protection. Allowing qualification concerns to be reported could help the Board identify issues relevant to corporate office without giving it formal control of the final slate. Nonbinding Board input, on this view, balanced institutional knowledge with NomCom’s final decision.

Staff support and a staff-only ECOM can likewise be defended on competence. Elections require contact verification, deadlines, system administration, ballot production and communications. Assigning those tasks to people familiar with AFRINIC’s records may be more reliable than building a temporary external administration each year. Prohibiting staff from participating in merit deliberations drew a boundary between operational support and candidate selection.

The hybrid voting design also had a coherent rationale. Electronic voting expanded access beyond the meeting room. Paper ballots retained an in-person option. Proxies allowed some absent members to use the paper channel through representation. The same candidate list applied to both voting methods, and their totals were joined. Limits on proxy volume and a ban on candidate proxy carriers constrained obvious concentrations and conflicts.

Other details reduced the need for improvisation. Nomination routes and supporter requirements were stated. Good standing had a defined payment rule. Candidate withdrawal had a timing consequence. Public comments had a closing point. Electronic voting had a minimum opening period. Paper ballots had secrecy and validity rules. Counters and trustees were named by role. Ties, disputes, announcement and publication were addressed. Abstentions were to be recorded rather than disappearing from the reported participation picture.

From this perspective, Resolution 202001.547 represented a serious attempt to turn a complex recurring event into a repeatable set of steps. Private institutions need administrative discretion; no text anticipates every contact-record error, candidate issue or operational interruption. Requiring the Board to approve every minor timetable adjustment could produce paralysis. Allowing a chair to break a committee deadlock could prevent failure to form a slate. Allowing inappropriate comments to be removed could protect useful discussion. Allowing the CEO to verify proxies could protect the vote from defective instruments.

That defence is substantial. It explains why eliminating all discretion would be neither practical nor necessarily fair. It also clarifies the proper standard of criticism. The question is not whether judgment existed, but whether each judgment was bounded by advance criteria, attributed to a responsible role, recorded with reasons and open to timely correction. The January text did part of that work by naming roles and rules. It left the quality of records and review mechanisms as the decisive second half.

Where discretion accumulated, narrower safeguards were available

The main structural risk was not one dramatic clause. It was the accumulation of connected powers. The Board appointed every NomCom member, placed one fully voting representative on the committee, could receive reported qualification concerns, could give nonbinding input on nominations and retained power to issue additional guidelines. NomCom controlled the candidate list and supervised the election. Its chair could cast a deciding vote after a deadlock and served as an electronic-vote trustee. The CEO appointed the entire staff-only ECOM, approved proxies, served as another trustee and published the result.

Each power has a plausible administrative explanation. Together, they create dependence. A nominee might encounter criteria applied by a Board-appointed committee after Board input. A voter using a proxy depended on approval by an executive who also appointed the election administrators. An observer of the electronic tally depended on three trustees drawn from roles already embedded in the process. The text distributed tasks but did not fully distribute institutional allegiance.

The right response is narrower safeguards, not assertions of bad faith unsupported by evidence. NomCom appointment notices could disclose selection criteria and relevant conflicts. Qualification standards could be fixed before nominations. The Board’s nonbinding input could be logged by subject and disposition without exposing private details. NomCom could publish aggregate reasons for slate decisions and disclose use of the chair’s casting vote. These measures would preserve the Board’s appointment power and NomCom’s confidentiality while making their boundaries testable.

Comment moderation could use a clear definition of inappropriate content, notice to the commenter and a rapid reconsideration route. Proxy approval could use a checklist tied to the published requirements, timestamped receipts, reasons for rejection and correction time where feasible. ECOM could preserve a voter-validation ledger, ballot issue and return reconciliation, and an exception register. Trustee actions could be documented through a signed record of system opening, closing, tally extraction and any irregular event.

Timetable flexibility also needed definition. The rulebook said the schedule should be followed as far as practical and minor deviations did not require Board approval. That avoided burdening the Board with operational adjustments. “Minor” is nevertheless a classification with distributional consequences. Moving a deadline can affect who completes a nomination, resolves a payment issue, secures a proxy or comments on a candidate. A deviation notice should state the original deadline, the change, the reason, the affected rights and why the change was considered minor.

The tie procedure called for a substantive design correction. If electronic participation was sufficiently legitimate to count in the ordinary result—and sufficiently preferable for the Board to encourage—it should not disappear solely because the aggregate produced a tie. A channel-consistent tie rule would maintain the composition of the participating electorate or reopen equivalent access. At minimum, members should know in advance that choosing electronic voting might leave them outside a paper-only decisive round.

Dispute review is the backstop for all these points. Because the available source only assigns disputes to Bylaws article 10.2, no conclusion can be drawn about how a particular dispute would have been handled. A credible private process would need decision-makers without a material conflict, access to the relevant institutional records, a chance for affected parties to be heard, written reasons and a remedy available soon enough to matter. Review after an irreversible transition may document an error without correcting it.

These proposals do not transform AFRINIC into a public body. They are ordinary disciplines for a private organisation making consequential internal choices. Their purpose is proportional: allow necessary judgment, but make materially different treatment visible. The more concentrated the appointment and operational relationships, the more important that visibility becomes.

Why procedure matters without converting dependency into jurisdiction

AFRINIC’s registry services sit in the practical environment on which network operators depend. Board decisions can affect budgets, staffing, service priorities and continuity. Operator-side discussion from LARUS has emphasised the importance of address continuity in a scarce IPv4 environment. That practical dependence explains why an internal election can attract attention beyond the people casting ballots.

Dependency, however, is not sovereignty. NRS’s institutional account and Heng Lu’s critiques of attendance-based mandate and institutional overreach supply a useful boundary: technical coordination, membership and community participation do not create public jurisdiction. AFRINIC keeps records and coordinates registry functions through private institutional arrangements. An election can decide who occupies a Board seat under those arrangements. It cannot give the resulting officeholder legislative, police, prosecutorial, confiscatory or public-law adjudicative power.

This distinction protects the analysis from two opposite errors. One error trivialises procedure because AFRINIC is not a government. Private institutions can control services, contracts and operational resources on which others rely; opaque internal decisions can impose real costs and uncertainty. The other error treats technical importance as a mandate to govern the public. That move confuses reliance with consent and coordination with authority.

The economic mechanism is indirect but real. If operators believe internal control is unpredictable, they may face greater uncertainty about service continuity and administrative treatment. If election roles, deadlines and records are clear, uncertainty can be reduced. No available source quantifies such an effect, and no economic outcome can be attributed to Resolution 202001.547. The responsible claim is limited: predictable office-selection mechanics can support confidence in a private institution whose operational choices matter, while weak controls can raise questions about internal accountability.

This framing also keeps the resolution in its proper editorial lane. BTW Media’s broader examination of AFRINIC election legitimacy addresses a larger field, while its account of Resolution 201404.196 concerns an earlier ex ante rule story. The January 2020 instrument should be judged for its own architecture: what it specified before voting, which roles it connected and where it retained judgment. It should not be used to infer later controversies, actual results or the validity of other years’ processes.

The public value of examining the text is therefore diagnostic rather than jurisdictional. It shows where a private technical institution located control. Members and operators can ask whether those locations were accompanied by reasons and records. They need not pretend that an internal election is a constitutional act in order to insist that it be coherent, even-handed and reviewable.

The August exception confirms the baseline, not its performance

The January rulebook was built around a pre-pandemic setting that included paper voting at the AGMM. On 4 August 2020, the Board approved a distinct virtual-only exceptional process under COVID-19 conditions. That later instrument shows that the January architecture became a baseline from which specific arrangements were changed when physical-meeting assumptions became impracticable.

That is the limit of the comparison. The August act does not prove that the January text failed, that its planned schedule was implemented or that any voting channel performed in a particular way. It does not convert the January rulebook into a virtual-election story. Its relevance is methodological: an emergency change can be understood only if the ordinary starting point is clear.

The baseline identified rights and dependencies that a later instrument might alter. Paper voting was tied to the AGMM. Proxies existed only through that paper channel. Electronic and paper totals were meant to be combined. A tie moved to repeated paper-only ballots. Once a physical meeting became impracticable, those clauses could not simply be assumed to operate unchanged. A separate instrument was necessary to state what replaced them.

Baseline clarity is one of the quiet strengths of ex ante rules. It makes exceptions legible. Without a January text, an August administrator might simply announce a new method without identifying which prior rights or safeguards had moved. With the text, members could in principle compare channel access, custody, counting and tie handling. The available evidence does not provide that full comparison here, and this article does not attempt it. It recognises only that the later exceptional process confirms the January instrument’s status as the ordinary pre-pandemic reference point.

Five counterfactuals reveal what the text controlled

Counterfactuals are useful because they test mechanisms without inventing outcomes. The first concerns nomination criteria. Suppose the criteria, supporter requirements and two-candidate objective are fixed and published before nominations open. Nominees can plan, members know what support means, and later observers can distinguish a screening judgment from a rule introduced during the contest. If the same matters are clarified only after names arrive, even a substantively reasonable decision is harder to defend.

The second concerns institutional influence. Suppose the Board appoints NomCom and provides nonbinding input, exactly as the rule permits, but appointment reasons, conflicts and the treatment of qualification concerns remain opaque. NomCom’s formal finality remains intact. Yet observers cannot tell whether advice remained advice. Publishing bounded records would not remove Board participation; it would make the claimed separation credible.

The third concerns the tied ballot. Suppose electronic and paper votes produce a tie. Under the January text, earlier electronic votes fall away and a paper-only rerun decides among the tied candidates. The decisive participating group may differ even if every administrator acts honestly and every ballot is counted correctly. The counterfactual isolates a design effect: changing channels can change participation weights. A rule preserving equivalent access would address the issue without alleging misconduct.

The fourth concerns administrative exceptions. Suppose a proxy appears defective, a public comment appears inappropriate or a deadline must move slightly. Administrators need room to act. If they use preannounced tests, record reasons and provide timely review, discretion remains functional while inconsistent treatment becomes easier to identify. If decisions leave no reasons or arrive after correction is impossible, the same discretion becomes harder to distinguish from arbitrariness.

The fifth concerns COVID-19. Suppose a rulebook assumes paper voting at a physical AGMM and a public-health emergency makes that setting impracticable. Administrators cannot preserve the original process merely by insisting that it still exists on paper. They need a separate, clearly authorised instrument explaining how access, custody, counting and disputes will work. The August resolution demonstrates that a distinct exception was adopted; it does not establish the quality or outcome of either process.

Together, the counterfactuals show what Resolution 202001.547 achieved. It fixed enough of the operating environment to make deviations and dependencies identifiable. It also left enough judgment that the documentary practices surrounding the rules would determine how persuasive the institution’s explanations could be. A good rulebook does not abolish judgment. It makes judgment visible at the moment it affects rights inside the organisation.

A baseline is valuable only if text and outcome remain separate

The January record supports a clear conclusion, but only a bounded one. Resolution 202001.547 adopted a detailed private election architecture. It identified nominators and supporters, gave NomCom screening and slate authority, defined good standing, created electronic and paper channels, constrained proxies, assigned counting roles, addressed ties and disputes, and set an announcement responsibility. Those provisions reduced the number of questions that administrators would otherwise have answered ad hoc.

The same text preserved meaningful discretion. Board appointment and input sat close to NomCom’s final slate decision. The CEO appointed ECOM, approved proxies and published results. Comment removal and minor timetable deviation depended on standards not fully elaborated in the extract. The tie rule changed the decisive channel. An additional-guidelines power remained with the Board. None of these features proves manipulation. Each identifies a place where consistent criteria, conflict disclosure, reasons and timely review would carry particular weight.

What the record cannot support is equally important. It cannot establish the exact January adoption day, the Board’s vote count, the effect of public consultation, actual compliance with the schedule, turnout, candidate diversity, system security, ballot reconciliation, proxy rejections, comment removals, disputes, member confidence or economic consequences. It cannot establish that the process was fair, inclusive, successful, unsuccessful or lawful under every applicable national and corporate rule. Those questions require evidence of implementation that is not present here.

The resolution’s lasting analytical value is thus as a map. It shows the path the institution said an election should follow before the contest unfolded. It shows where multiple roles could check one another and where their relationships could create dependence. It shows that electronic access and paper procedure were joined at the first count but separated in a tie. It shows how a later emergency instrument could be recognised as a departure from a pre-pandemic baseline.

For a private technical bookkeeper and coordinator, that is the right scale of judgment. AFRINIC did not acquire public authority by writing election rules, and its members did not become a sovereign people by voting under them. The organisation did, however, make consequential internal office selection more legible. The credibility of that choice rests not on grand language but on a modest institutional promise: rules known in advance, judgment constrained by reasons, records capable of verification and correction available while it can still make a difference.