Summary

  • Article 4 of the instrument archived by AFRINIC as its 2007 bylaws identified recurring funding categories and placed their review “from time to time” with the Board. It established corporate competence, not any particular charge, amount, invoice, collection or finding of necessity.
  • The power was private delegated fiscal authority. It could be compulsory in practical economic effect where reliable registry records and coordination were difficult for an operator to avoid, but it was not a tax, public levy, regulatory sanction or sovereign command.
  • Articles 3, 7, 13, 16 and 20 show that constitutional permission was only the first link in a longer chain: permitted purpose, competent decision, authenticated resolution, operational execution, accounting, audit and ex-post oversight all required their own evidence.
  • The phrase “from time to time” supplied adaptability, but no formula, review interval, notice rule, consultation duty, evidential threshold, appeal path or expiry. Each exercise therefore created a fresh obligation to show authority, cost basis, alternatives, impact, conflicts, execution, use of proceeds, exceptions and a route to reversal.
  • The historical record does not prove that any fee was reviewed, imposed, collected, waived, excessive or unlawful in 2007. It supports a governance diagnosis: recurring private funding attached to an important coordination service should be audited with the discipline applied to other practically compulsory funding mechanisms.

The constitutional hinge

The most consequential words in AFRINIC’s archived 2007 instrument are not a number. Article 4 contains none. They are the words that convert a list of revenue sources into a continuing field of Board discretion. Article 4.1 says that funding for running the company shall be realised from specified sources. Article 4.2 then says that the fees referred to in that list are subject to review “from time to time” by the Board.

That construction matters because a constitution can do two different jobs at once. It can define a category of corporate power, and it can identify the organ competent to exercise that power. It does not follow that the constitution has itself performed every later act within the category. A power to review fees is not a reviewed fee. A listed revenue source is not an invoice. A valid corporate object is not a cost model. A Board’s competence is not evidence that the Board met, considered alternatives, declared conflicts, voted, recorded dissent, fixed an effective date or caused money to be collected.

Article 4.1 was more specific than a general permission to receive income. It named membership fees; subscriptions for bulk registration services; registration fees for individual address-space assignments; maintenance fees for non-contiguous, non-ISP address space; registration fees for individual address-space transfers; registration fees for ASN allocations; and maintenance fees for ASN allocations. It then added an eighth, residual category: other sources the Board might deem appropriate from time to time.

The list constitutionalised a portfolio of possible recurring funding streams around membership, registration, maintenance, transfer and ASN activity. Yet it supplied no amount, calculation method, frequency, notice period, consultation mechanism, required evidence, publication duty, appeal route or automatic sunset. That silence is not proof of defective intent, still less of illegality. It does mean that the text alone cannot answer the questions that make a concrete charge accountable.

The analytical mistake is to compress three propositions into one. First, the company’s constitution identified possible funding sources. Second, the Board was made the recurring review point. Third, a later charge might have been designed and executed. Only the first two propositions are established by Article 4 itself. The third would require a separate record. Without that distinction, constitutional authority becomes a magic phrase that appears to authenticate any downstream conduct merely because the general power existed.

It is better understood as delegated fiscal authority inside a private company limited by guarantee. “Fiscal” describes the power’s relationship to recurring organisational revenue, budgeting and expenditure. “Delegated” describes its source and limits in the corporate instrument. Neither word turns AFRINIC into a state. The company was a private coordinator and bookkeeper of registry information. Its incorporation, Board, constitution, membership arrangements, records and consultations could organise private corporate action. They could not manufacture sovereignty.

That boundary is indispensable. AFRINIC’s records could be operationally important to networks that required dependable uniqueness coordination. The record could carry substantial practical weight. But a ledger is evidence of a coordination state, not a deed of ownership, a statute, a judgment or a sovereign grant. The institution did not create the underlying Internet by recording identifiers, did not acquire title to number resources by maintaining entries, and did not gain punitive, confiscatory, regulatory or adjudicative authority through Article 4. The fee clause financed a company; it did not crown a government.

What the surrounding articles add

Article 4 cannot be read responsibly in isolation because the instrument distributes purpose, management, decision and execution across other provisions. Those provisions do not add facts about a particular fee exercise. They map the kinds of records that such an exercise should have left.

Article 3.2 directs the company’s income and capital solely towards its objects and bars distributions to members. Article 3.3 allows reasonable remuneration for services and reimbursement of expenses. Together, the clauses establish two propositions that are easy to confuse. Operations can legitimately cost money, and income is constrained by the company’s purposes. Neither proposition proves that every expenditure is necessary, every reserve proportionate, every programme within a thin coordination core or every charge aligned with cost.

A nondistribution rule answers the question of who may receive company value in one important sense. It does not answer how much the organisation needs, how shared costs should be allocated, which activities are essential, whether optional programmes should be bundled into recurring charges, or whether a surplus should trigger reduction. A nonprofit can face real costs and still have an agency problem. Indeed, when profit distribution is unavailable as an explanation, institutional expansion, budget preservation and organisational relevance can become the incentives that require scrutiny.

Article 13 then concentrates the management surface. Article 13.1 places management under the Board’s direction, subject to the Companies Act and constitution. Article 13.2 allows the Board to exercise company powers not reserved by the Act or constitution to members in general meeting. Article 13.3 includes power to determine the financial budget, set and vary expenditure ceilings, fill an auditor vacancy, and reduce or waive fees or amend payment conditions. Article 13.4 permits delegation to committees, officers, employees or other delegates on terms and within limits set by the Board.

Read with Article 4, that architecture places price review, budget formation, expenditure limits and exceptions within a closely connected control surface. Such concentration is not inherently improper. A Board ordinarily manages a company and needs to connect expected income with planned operations. The governance risk lies elsewhere: the institution proposing the scope of work can help determine the budget for that scope, review the charges that fund it, and control the exceptions. If the record is thin, expansion can justify revenue and revenue can enable further expansion in a self-reinforcing circle.

Article 13.3(viii), the waiver and payment-condition power, is especially significant beside Article 4.1(8), the residual revenue source. One is flexible on what additional funding sources may be considered; the other is flexible on relief and payment conditions. Flexibility can be humane and operationally useful. It can protect continuity and respond to unusual circumstances. Yet the combination makes scope, purpose and exception records essential. Without stated criteria and aggregate reporting, a waiver can be sensible accommodation or invisible favour, and an additional revenue source can be necessary adaptation or mandate drift.

The instrument alone does not tell us which.

Article 16 supplies the ordinary architecture by which a Board decision might be authenticated. It addresses quorum, majority voting, dissent, minutes and written resolutions. The text states that a majority of Primary Directors constitutes a quorum, with at least three at an adjourned meeting; each director has one vote; ordinary questions are decided by majority; a director present is presumed to assent unless dissent is recorded; and a written resolution signed by at least two-thirds of directors can take effect if entered in the minute book.

Those mechanics turn an abstract claim that “the Board decided” into testable questions. Which directors participated? Was quorum present? What precise proposition was put? Were conflicts disclosed? What alternatives were presented? How did directors vote? Was dissent recorded? Was a written resolution entered in the minute book? Article 16 does not prove that a fee-review record answering those questions exists. It explains why minutes and resolutions, rather than retrospective summaries, are the evidence needed to bridge constitutional competence and decision.

Article 20 sits further downstream. It sets execution rules for instruments and payment instructions, involving the chief executive, directors and authorised persons; a Board-authorised payment policy; Board choice of bankers; and multiple signatories. This is the point at which an authorised corporate choice may become operational conduct. Even then, the article proves only the designed execution layer. It does not prove an effective schedule, a valid delegation, an invoice, a receipt, an account movement or a reconciliation.

Finally, Article 7.4 supplies an ex-post accountability setting. It gives Full Members powers that include approving financial statements, receiving the auditor’s report, considering the annual report, determining general policies, approving major transactions, amending the constitution and appointing the auditor. For the narrow question here, those powers matter because they can test the Board’s financial account after decisions have been made. They do not erase the need for decision-level evidence.

An annual statement may accurately describe aggregate finances while remaining silent on why a particular charge class was chosen, how its design related to cost, whether alternatives were considered or how objections were answered.

The constitutional map therefore contains at least six distinct stations: purpose, source, competence, decision authentication, execution, and accounting with assurance. Treating them as one station invites category errors. The constitution is not the invoice; the resolution is not the receipt; the receipt is not proof of permitted use; the financial statement is not the original reasons paper; and an auditor’s report is not necessarily a review of the policy choice that produced the charge.

Compulsory in effect, private in law

Why apply the scrutiny associated with compulsory funding if the charge remains private? Because legal classification and economic experience answer different questions.

A network operator may depend in practice on accurate registry records, recognised coordination and reliable administrative continuity. Exit from that dependency may be costly, uncertain or operationally unrealistic. A recurring charge attached to that coordination can therefore feel unavoidable as a condition of maintaining ordinary network operations. The pressure can resemble compulsion because the service is not a casual consumer option and the cost of refusing it may extend beyond the invoice.

That practical pressure does not turn the charge into taxation. A tax rests on public-law authority. Article 4 was a clause in the constitution of a private company. The Board could not create a public levy by resolution, legislate for a population, punish an operator as a sovereign, confiscate an asset, determine title or adjudicate ownership. Nor could a mailing list, an election, a consultation or the regional importance of the registry convert corporate procedure into popular sovereignty.

The distinction protects both sides of the analysis. If the fee is loosely called a tax, critics risk granting the institution the very governmental status they should deny. If it is described as merely optional because it is private, defenders risk ignoring the operational dependence that gives the funding power its real economic force. “Private recurring charge with compulsory practical effect” is less dramatic and more accurate. It preserves the legal boundary while explaining why a high standard of evidence is warranted.

The audit analogy follows from dependence, not from a claim of sovereign character. Where ordinary exit is constrained, the fee setter should be able to show the same kinds of things demanded of any institution controlling an important recurring funding mechanism: the exact source of authority, the class of charge, the funded purpose, the cost base, the decision method, the impact, the collection route, the disposition of proceeds, the treatment of exceptions, independent assurance and genuine reconsideration.

This is also where the bookkeeper doctrine has practical force. A bookkeeper earns trust by maintaining an accurate, dependable and intelligible record. The bookkeeper becomes unsafe when the importance of the record is treated as proof of a larger mandate. Registry necessity is a reason to protect the thin coordination core. It is not a licence to treat the ledger as authorship of the reality it records. Funding power should therefore track the costs of coordination rather than the institution’s appetite to enlarge the meaning of coordination.

“From time to time” is not a blank cheque

Article 4.2’s phrase “from time to time” can be defended as sensible constitutional drafting. A permanent constitution should not freeze an operating company’s prices forever. Conditions change. Staffing needs, security requirements and other operating inputs can change. A Board needs room to adapt recurring income, and a power to reduce or waive payments can be necessary for continuity.

But continuing discretion creates continuing accountability. The absence of a fixed interval does not make review less evidential; it makes the timing and reasons for each review more important. The phrase cannot identify when a review should occur, what information should trigger it, who should prepare the analysis, which alternatives should be considered, or when a charge should fall rather than rise. Those questions must be answered outside the clause, in the record of each exercise.

The word “review” should describe a testable event, not a euphemism for adjustment. A genuine review should be capable of producing several outcomes: an increase, no change, a reduction, a redesigned waiver rule, consolidation with another charge, or retirement. If the decision architecture can produce only persistence or expansion, then review has hardened into ratification. The constitutional wording does not require that asymmetry. It is a governance choice.

Every exercise should therefore begin with instrument and version discipline. The decision maker must identify the exact constitution relied upon, its effective status, its amendment history and the precise clause authorising the charge class. That requirement is unusually important here because AFRINIC’s archive labels the document “Bylaws 2007,” while the instrument’s internal adoption and signature fields retain unresolved brackets. One line refers to adoption on an unspecified day in 2006; another date field ends in 2007.

The archive supports the careful description “the instrument archived as the 2007 bylaws.” It does not establish a precise adoption or effective date.

Next comes classification. The Board should identify the triggering service, assessed unit, duration and fit within one of Article 4.1’s enumerated categories. Reliance on the residual eighth source demands still more detail. “Such other sources as the Board may deem appropriate” is open-textured, but not infinite. It remains corporate authority bounded by the company’s objects and private legal form. A reasoned record should show why the new source serves those objects, why it is necessary, how broadly it operates, which costs it funds and how it will be reviewed or retired.

Purpose must then be separated from necessity. Article 3’s income constraint can show that money ought to serve company objects. It cannot demonstrate the operational need for each funded activity. A useful fee paper would distinguish the costs of accurate records, uniqueness coordination and dependable operations from optional projects. It would identify incremental and shared costs, the allocation method, reserves, excluded costs and the reason recurring income is needed. Otherwise, a thin essential function can become the carrier for an unbounded institutional scope.

Decision integrity is the next bridge. A complete file would include the Board paper, authority analysis, material alternatives, conflicts, quorum, votes, dissent, final resolution and minute-book entry. It would explain the treatment of impact across operator size, resource use and operational dependence. That analysis is not a contest over corporate franchise; it is a proportionality test. Fixed administrative and funding burdens can affect small and large operators differently even when the legal form of the charge is uniform.

Notice and response should be evidence, not theatre. The record should show what was proposed, when it was disclosed, which material objections were received and how the final decision addressed them. Consultation can improve information. It cannot supply missing constitutional authority, prove representative consent or make the resulting choice correct merely because a comment window existed.

Execution requires its own controls: an authenticated schedule, effective date, authorised sign-off, billing rules, receipts, bank controls and reconciliations. Exceptions require criteria, delegated limits, reasons, duration and aggregate effect. Use of proceeds requires linkage between the charge, approved budget, actual spending, programme attribution, reserves and explained variances. Independent assurance requires accounting records, statements, auditor access, findings and remediation. None of those records can be replaced by the general existence of Article 4.

Finally, recurring review must include reversal. A decision should state the next review date or measurable trigger, and the evidence that would support reduction, redesign or retirement. Surplus, reserve accumulation, reduced core cost, narrowed service scope, collection materially above forecast, or an unresolved discrepancy between revenue and use may all warrant reconsideration. The object is not automatic reduction. It is to ensure that a temporary need does not become a permanent entitlement merely because the organisation has arranged its commitments around the income.

The strongest defence, taken seriously

The strongest institutional defence is straightforward. A nonprofit registry cannot provide stable coordination without recurring funding. It faces changing operational needs and cannot responsibly encode a permanent price table in its constitution. A Board is the normal corporate organ for budgets and pricing. Flexibility to adjust fees and make exceptions can protect service continuity. The same instrument places financial statements, annual reports and auditors within a broader accountability architecture.

A contemporary message also indicates that AFRINIC exposed proposed fee structures and constitutional revision to comment before anticipated Board action.

That defence deserves more than dismissal. Stable record keeping costs money. The nondistribution rule does not make staff, systems, security or administration free. Constitutional flexibility may be more responsible than a rigid formula that cannot respond to real conditions. Waivers may prevent a general rule from producing unreasonable disruption in exceptional cases. Consultation may reveal information the Board lacks.

The answer is not to deny discretion but to discipline it. Adaptability increases the value of a traceable record because outsiders cannot infer the basis of a variable choice from the constitution alone. The more authority the Board has over fee design, budget, expenditure limits and exceptions, the more important it is to distinguish core coordination costs from discretionary growth, identify conflicts, preserve alternatives, authenticate the resolution, control implementation and reconcile proceeds to purpose.

In that sense, accountability is not hostile to continuity. It protects continuity from institutional overreach and from suspicion. A company able to show why a charge exists, which cost it funds, how it was decided, who executed it, where the money went and when it will be reconsidered is better equipped to defend legitimate recurring income. Evidential opacity may feel convenient at the moment of decision, but it weakens the institution’s position when trust later depends on reconstructing what happened.

The narrow 2007 process trace

A message forwarded to AFRINIC’s resource-policy mailing list on 3 September 2007, attributed to the chief executive and dated the previous day, offers one glimpse of contemporary practice. It says the Board discussed fee structures for 2008 and 2009 alongside revision of the constitution. It also says that a proposed fee schedule and proposed constitution were posted for a 15-day comment period before final Board action.

The message matters because it shows that AFRINIC publicly presented future fee structure and constitutional revision as linked Board matters and invited time-limited comment. It is evidence of a described proposal and a described participatory step. It is not evidence of the outcome.

The message does not identify which constitution governed the contemplated decision. It does not prove that the proposed constitution became the archived 2007 instrument, that final action occurred, that comments changed the proposal, or that any resulting charge was necessary, valid or collected. The available record does not include the proposed schedule, proposed constitution, submissions, response to comments or final Board decision. Nor does it prove whether consultation was constitutionally required, voluntarily offered or required by another instrument.

This restraint is more than historical caution. Official texts prove what an institution wrote, proposed, recorded or presented. They do not validate a claim simply because the institution made it. An open process is evidence of openness at a defined stage. It is not proof of authority, representative consent, substantive correctness, legal compliance or faithful implementation. Institutional self-description should begin an inquiry, not end it.

What can be concluded

Article 4 placed a continuing private funding power at the Board. The surrounding constitution shows how that authority related to corporate purpose, management, Board procedure, execution and later financial oversight. The result was a recognisable governance architecture, but not a self-proving one.

The evidence does not support a finding that a particular fee was excessive, discriminatory, unlawful, punitive, confiscatory or unrelated to cost. It does not establish any fee amount or even that a listed charge was actually imposed. It does not prove a specific Article 4.2 review, a resolution, an invoice, a waiver or a reconciliation. Those absences prevent a verdict on an actual exercise.

They do not prevent an institutional conclusion. A recurring charge attached to practically important registry coordination can impose compulsory economic pressure while remaining a private charge. That combination makes the power fiscally consequential and legally bounded. It should be audited through the complete chain from clause to correction.

The central question for any exercise is therefore not whether AFRINIC could cite Article 4. It is whether the company could show how a particular charge travelled from the correct version of the constitution, through a competent and documented decision, into controlled execution, permitted use, independent assurance and symmetrical reconsideration. Authority begins at the clause. Accountability begins when the trail continues beyond it.