Summary

  • Resolution 201304.171 did not create AFRINIC’s USD 100,000 Board-approval boundary. Its distinct work was to repeat that boundary inside a redesigned delegation containing three principal signatories and two conditional alternates.
  • The repetition answered one continuity question: the high-value Board exception accompanied the April 2013 framework. It did not prove the scope of “transaction,” the required signing combination, the date of bank activation, or the revocation of earlier mandates.
  • A consequential wording change remains unresolved. The 2011 sentence referred to “any external transaction,” while the 2013 sentence referred to “any transaction.” The public record establishes the difference but not whether it widened the rule, compressed the drafting, relied on a definition in the unseen delegation table, or changed nothing in practice.
  • AFRINIC’s Board was governing the treasury of a private technical bookkeeper and coordinator. That is a real corporate control, but it creates no sovereign, regulatory, police, punitive, confiscatory, prosecutorial or adjudicative power over operators, networks, routing or number resources.

The old number and the new frame

The revealing feature of Resolution 201304.171 is repetition. In April 2013, AFRINIC’s Board authorised the “above mentioned signatory persons” to act under the organisation’s authorisation policy, described in a Delegation of Authority table, subject to one expressly stated exception: any transaction in excess of USD 100,000, or the equivalent in another currency, had to be approved by the Board. Read alone, that sentence might look like the birth of a financial limit. Read in the sequence published by AFRINIC, it is plainly a restatement.

The same numerical boundary had already appeared in October 2011. The minimum baseline is enough to establish the point. Four people were then named as signatories. The Board authorised them to act under the applicable policy, except that any external transaction in excess of USD 100,000 or its foreign-currency equivalent required Board approval. Bankers were to receive the relevant documents, and a later minutes amendment inserted the threshold sentence into the documentary record. The origin and detailed economics of that earlier act belong to the 2011 episode.

The question here is different: why did an established limit have institutional work to do again in 2013?

The answer lies around Resolution 201304.171. The Board first adopted New Delegation of Authority 002-2013 in line with a new organisational structure. The sequence added the chief operating officer to the bank-account authority architecture. It then named three principal signatories: the chief executive officer, the director of finance and accounting, and the chief operating officer. Next, it provided that, in the absence of two of those three appointed signatories, two other people would also serve as alternate signatories under the delegation. Only then did the Board repeat the high-value approval exception.

A further resolution said the bankers were to be furnished with the relevant policies and resolutions.

Those surrounding acts should not be collapsed into a general history of AFRINIC’s banking arrangements. Each performed separate work. Their importance for Resolution 201304.171 is narrow but substantial: they show that the visible authority matrix had changed before the familiar ceiling was stated again. The public structure moved from a four-person slate without a published principal-versus-alternate distinction to three named principals and two people in a conditional alternate layer. The threshold did not move numerically. The human and role architecture around it did.

That is why the first institutional fact of 2013 is not USD 100,000. It is reattachment. A monetary boundary has meaning only in relation to the actors, instruments, definitions and evidence that make it operational. When those actors and instruments are redesigned, silence can create a continuity problem. Staff, directors, bankers or auditors might later ask whether a restriction written beside an earlier mandate followed the new delegation or remained attached to the old one. By repeating the exception after adopting Delegation 002-2013, the Board made the continued presence of a Board-level high-value control visible in the new sequence.

This is an inference about the work performed by the published chain, not a claim about an unpublished motive. The register does not say that directors feared confusion, anticipated a dispute, detected a weakness or intended a particular compliance outcome. It does not need to. The sequence itself shows that an old numerical rule was placed alongside a new authority design. Repetition can be institutionally meaningful even when it creates no new number.

Within the first few resolutions, then, April 2013 produced a precise but limited delta. The organisation adopted a numbered delegation aligned with a changed structure; the public register exposed three principals and two conditional alternates; and the Board again reserved transactions above the established amount to itself. One ambiguity was reduced: the exception was not left stranded in the 2011 wording.

Several others remained: who exactly fell within “above mentioned signatory persons,” what “transaction” covered, how approval connected to signatures, when the bank treated the new authority as active, and which prior mandate was formally displaced.

Why a repeated limit can be a redesigned control

A threshold is often described as if it were a solitary number. In practice, it is a junction. It joins a defined class of activity to an approver, a measurement method, an execution route and an audit trail. “Above USD 100,000 requires Board approval” can be clear as a headline while being incomplete as an operating control. To apply it consistently, an institution also needs to know whose proposed commitments are covered, which transactions count together, when foreign currency is converted, what evidence constitutes approval, who may sign after approval, and when banks recognise a changed mandate.

The April 2013 restatement matters because one of those connected elements—the participant architecture—had visibly changed. Two years earlier, the public sequence listed four individuals together. In 2013, the chief executive officer and director of finance and accounting remained among the principals, the chief operating officer entered the principal set, and the Board representative in Mauritius and the director of human resources and administration appeared in an adjacent conditional-alternate provision. This was not merely a change of names.

The register introduced a difference of status: principal authority on one side, alternate service under a stated absence condition on the other.

When a delegation changes in that way, there are at least two ways an approval ceiling can fail to travel clearly. First, the new instrument may name new actors without stating whether the old exception constrains them. Second, the exception may be repeated but disconnected from the mechanics that show how it applies. Resolution 201304.171 addressed the first risk in public view. It did not fully address the second.

The control value of repetition can be seen through a simple counterfactual. Suppose the Board had adopted Delegation 002-2013, named the new principal set, created an alternate layer and stopped. A later reader would know the human architecture changed but would have to infer whether the 2011 high-value exception survived. The inference might be reasonable, especially if the unseen delegation incorporated it. Yet the published chain would leave room for a claim that the previous restriction belonged to a previous authorisation policy or slate. Repeating the threshold avoids that particular uncertainty.

Now consider a more complete version. The Board could repeat the threshold while also identifying the superseded instrument, stating the effective date, defining the covered transaction class, specifying categorical signing combinations and recording when the bank accepted the change. That would turn continuity from a textual inference into a versioned control trail. The April 2013 register does not supply those links. It therefore sits between silence and a complete transition record: better than leaving the old ceiling unmentioned, but less than a publicly testable implementation chain.

This middle position is analytically important. Governance analysis often swings between two unhelpful extremes. One treats any repeated rule as empty boilerplate because the words appeared before. The other treats a new resolution number as proof of a newly created power. Neither fits the evidence. Resolution 201304.171 performed distinct work by carrying the exception into the redesigned sequence, but it did not originate the amount or prove that every operative term stayed the same.

The resolution also illustrates why continuity should be assessed by links rather than labels. The title “Delegation of Authority 002-2013” signals a version. The word “new” signals change. The surrounding resolutions identify roles and people. The threshold sentence identifies a Board approval exception. The banker-notification instruction signals an intended implementation bridge. But these components do not automatically prove one another. Adoption is not activation. Notification instruction is not proof of bank receipt. Board approval is not execution. A named alternate is not necessarily an active signatory at every moment.

A repeated number is not a repeated scope.

The institutional achievement was therefore modest and real. The Board left a visible marker that high-value approval remained reserved at Board level even as the authority matrix changed. This made it harder to read the organisational redesign as an implicit removal of the ceiling. The institutional deficiency was equally specific. The public record did not expose enough of the version, scope and activation chain to let an outside reader determine precisely how that continuity operated.

The missing word that cannot be wished away

The most consequential textual delta between the two published formulations is one word. In 2011, the exception applied to “any external transaction” in excess of USD 100,000. In 2013, it applied to “any transaction” above the same amount. The disappearance of “external” is observable. Its intended and operational effect is not.

There is an obvious temptation to translate the difference directly into scope. On ordinary reading, “any transaction” can appear broader than “any external transaction.” It might seem to include internal transfers or other activity that an external-only rule would exclude. But that conclusion outruns the public evidence. The archived register does not define either phrase. It does not reproduce Delegation 002-2013 or its table. It does not show whether “transaction” was a defined term, whether “external” had been considered redundant, or whether an internal classification elsewhere preserved the earlier perimeter.

Several explanations remain possible. The Board may have intended a broader rule. The drafting may have been compressed without substantive change. The new delegation may have supplied definitions that made the adjective unnecessary. The published register may have summarised a fuller instrument. Or the wording may simply have varied without anyone addressing its consequence. The available record cannot choose among these possibilities. Calling the change a deliberate expansion would convert a textual fact into an unsupported statement of intent. Calling it a harmless omission would do the same in the opposite direction.

This is not pedantry. A threshold is only as stable as its perimeter. If “external transaction” excludes some category that “transaction” includes, the identical number can govern a materially different set of acts. If the phrases are functionally equivalent under definitions in the unseen table, the number and perimeter may both have continued. Without those definitions, numerical continuity and scope continuity must be kept separate.

The problem becomes clearer with categories that the register does not address. It does not say whether an inter-account transfer is a transaction, whether a recurring commitment is measured by each payment or its total contractual value, whether related invoices are aggregated, or whether an emergency instruction follows a different path. It does not say which foreign-exchange rate applies, when the equivalent amount is measured or how rounding works. These are not allegations that the organisation lacked rules. They are examples of why the bare phrase cannot establish its own operational reach.

Nor should the missing adjective be used to manufacture a story of evasion or enforcement failure. Nothing in the published material identifies a transaction tested under the resolution. There is no disclosed approval, refusal, split payment, breach or loss. There is no basis to claim that anyone exploited the wording, misunderstood it or acted improperly. The responsible finding is narrower: a potentially meaningful scope word changed, and the public materials needed to determine the effect are absent.

The wording delta also shows why restatement is not identical to replication. The amount, the “in excess of” construction, the foreign-currency limb and the Board approver remained materially aligned with the earlier sentence. The transaction adjective did not. The 2013 act carried forward the visible core of the boundary while changing at least one visible component. It therefore supplies evidence of continuity at the level of amount and approver, but only uncertainty at the level of perimeter.

A mature control record would surface this distinction without publishing sensitive banking details. It could state that the 2013 delegation continued, narrowed or expanded the predecessor’s transaction definition, or that the terminology changed without intended effect. It could identify the controlling definition by version. That small addition would let staff, directors, auditors and members distinguish a conscious scope decision from a drafting variation. The present register does not.

Who were the “above mentioned signatory persons”?

Resolution 201304.171 authorised the “above mentioned signatory persons” to act under the policy described in the delegation table. The phrase seems straightforward until it is placed in sequence. Immediately before it, the Board had published one resolution naming three principals and another providing for two alternate signatories when a stated absence condition was met. The plural referent could point only to the principal signatories, or it could encompass the principals together with alternates when the alternates were conditionally active. The wording does not unambiguously decide.

That uncertainty should remain open. The alternate provision is not a general grant that the two people were always equivalent to principals. It says they would “also” serve as alternate signatories in the absence of two of the three appointed signatories. Yet the subsequent phrase is broad enough that a reader may naturally look back across both preceding resolutions. Without the delegation table, the grammatical sequence cannot be converted safely into an operational matrix.

The difference matters because participant class is one link in the control chain. If Resolution 201304.171 addressed only the principals, the authority of an activated alternate would need to be established through the alternate clause and the underlying table. If it addressed all five people subject to their respective conditions, the exception might travel through both layers. Either design could be coherent. The public wording does not prove which one governed.

It would be equally unsafe to infer signing combinations. Naming three principals does not show that any one could act alone, that two were always required, or that a particular pair was mandatory. Saying that alternates served in the absence of two principals does not show whether one or both alternates had to join an instruction, whether another principal was required, or how the absence condition was evidenced. Resolution 201304.171 concerns authorisation under a policy and a Board approval exception. It does not publish the combination table.

This is a useful reminder that a public register can disclose the existence of a control without disclosing its operative topology. The register tells us there were principals, conditional alternates, an authorisation policy and a high-value exception. It does not tell us the complete route from initiation to Board approval to bank execution. That missing topology is not proof of poor control. It is a limit on what can be concluded publicly.

The remedy is categorical, not personal or security-sensitive. A version notice could say that ordinary instructions require a defined combination of principal roles, that alternates become eligible only under a defined condition, and that high-value instructions require a cited Board approval before the applicable combination may execute them. It need not reveal specimen signatures, credentials, bank-account identifiers or authentication methods. It need not describe any actual transaction. The goal is to show how classes of authority connect.

That form of disclosure would also reduce dependence on the ambiguous phrase “above mentioned.” Cross-references by resolution number, role class and delegation version are more durable than positional language. A reader should not have to decide whether “above” means the immediately preceding slate, two preceding resolutions, or a table not reproduced on the page. Where the stakes include corporate treasury, precision in referents is itself a control property.

Approval is not execution

The most important conceptual separation in Resolution 201304.171 is between the Board’s approval threshold and the signatories’ execution authority. The sentence reserves approval for a class of high-value transactions. It does not state how an approved instruction is signed, how many signatures are needed, which roles must combine, or whether an alternate is eligible. A Board approval and a bank mandate answer different questions.

The approval rule asks whether the transaction may proceed at all without collective Board assent. The execution rule asks who can issue or authenticate the instruction after all required approvals exist. In a well-designed process, these controls can reinforce each other, but neither substitutes for the other. A bank may require a particular combination of authorised signatories while the company separately requires Board approval above a value. Conversely, the company may possess a Board approval while still needing the correct signatory combination to execute it.

Resolution 201304.171 proves the first layer only in broad terms. Transactions in excess of USD 100,000, or the equivalent in another currency, had to receive Board approval. Because the trigger is “in excess of,” exactly USD 100,000 is not captured by this sentence. Other rules might apply at that exact amount, but they are not established here. Precision matters: replacing “above” with “at or above” would change the line.

The foreign-currency limb preserves the economic idea of the boundary while opening practical questions. A transaction denominated in another currency must be compared with USD 100,000, but the published record gives no conversion source, valuation time or rounding method. It does not say whether the amount is assessed when a contract is approved, when an invoice is issued, when an instruction is prepared or when a bank processes it. Again, these omissions are not proof that no operative method existed. They show that the public sentence cannot answer the question.

Aggregation raises a similar problem. The resolution does not define whether related payments are tested separately or together, whether recurring commitments are measured per instalment or over their term, or how amendments to an existing commitment are treated. These questions belong to the scope and measurement layer. Their absence cannot be used to infer splitting, circumvention or misconduct. No such event is established. But an audit-ready system would address them somewhere in the controlling delegation or supporting procedure.

The evidence of Board approval is another distinct link. The register does not state whether approval must appear in a resolution, minutes, a written consent or another record. It does not describe quorum, conflict treatment, emergency approval or post-payment ratification. Nor does it identify any particular decision made under the rule. To connect approval to execution, a system would normally preserve a reference that signatories, finance personnel, auditors and perhaps the bank can test without seeing more transaction detail than they need.

The distinction between approval and execution protects against over-reading the named people. A principal signatory was not thereby given power to approve a transaction that the Board had reserved to itself. An alternate’s conditional availability did not itself prove that the alternate could approve or execute alone. The Board’s collective reservation did not itself instruct the bank on the exact signature combination. The public sequence contains related controls, not one undifferentiated grant.

This distinction also keeps institutional power in its proper domain. The Board was setting the conditions under which its private organisation’s treasury could be committed and its banking instructions executed. It was not legislating for networks or conferring public-law authority. The fact that an internal corporate approval could affect an important technical coordinator does not transform the approval into regulation of the Internet.

The version bridge that the public record does not complete

Delegation of Authority 002-2013 is a numbered instrument, and the register calls it new. Those features imply a versioned control environment. Yet the public chain does not identify the precise predecessor delegation, quote an express supersession clause, state the effective date or show when earlier personal mandates were revoked. It records adoption, roles, a threshold and an instruction to furnish bankers with documents. The transition from decision to operational replacement remains unverified.

The difference between these stages is practical. Board adoption establishes that the corporate decision was made. An effective date establishes when the organisation intended the new rule to govern. Delivery places the documents before the bank. Acceptance or activation establishes when the bank’s systems recognised the new mandate. Revocation confirmation closes the old path. If any of these dates differ, the transition period may require explicit handling.

The public register does not say that the stages differed, and no implementation failure may be inferred. It simply does not provide the dates or acknowledgements. Resolution 201304.172’s instruction to furnish bankers with the relevant policies and resolutions is an intended bridge, not proof that the bridge was crossed. A direction to deliver is not a receipt. A receipt is not necessarily activation. Activation of a new mandate is not necessarily revocation of every earlier personal authority.

This missing version bridge is especially relevant because the participant architecture changed. The chief executive officer and finance director remained in the principal set, while the chief operating officer entered it and two people from the earlier slate appeared in a conditional alternate layer. Continuity of some people can make a transition appear seamless, but role continuity and mandate continuity are not identical. A person can remain named while the conditions, combinations or scope of authority change.

The public record also does not assign Resolution 201304.171 to a precise day. It places the act in April 2013. A later resolution approved minutes for Board meetings held on 24 April and 28 May, but the register does not map this threshold resolution to one exact meeting date. The safe date is therefore April 2013. Inventing a day would create false precision at the very point where version timing matters.

A complete bridge would be compact. It would identify Delegation 002-2013, its predecessor, the adoption and effective dates, the role categories affected, the status of the threshold language, the date of bank delivery and activation, and confirmation that superseded mandates were revoked. It could preserve operational security by omitting account numbers, credentials, specimen signatures, verification procedures and transaction details. Such a notice would allow a reader to trace status without exposing the machinery an attacker might exploit.

Version control matters because rules live over time. A Board may make a clear decision at one moment, yet ambiguity can arise later if implementers cannot tell which version applied on a given date. A current-state list of names is not enough. A historical register should show the path by which one configuration replaced another. Resolution 201304.171 is a visible continuity marker inside that path, but not a complete history of activation.

The strongest benign reading

The strongest benign account deserves to be stated in full. The Board may have been doing careful, conservative corporate housekeeping. AFRINIC had a new organisational structure. It adopted a numbered delegation to reflect that structure, added the chief operating officer to the banking framework, named three principals, preserved operational resilience through two conditional alternates, repeated the familiar USD 100,000 Board exception and instructed that the relevant documents be provided to the bankers. On this reading, the sequence is concise because the full authority table was maintained privately where it belonged.

That private table may have resolved every ambiguity visible on the public page. It may have defined “transaction,” explained why “external” was no longer necessary, specified the principal and alternate combinations, set the foreign-exchange method, identified the predecessor, chosen an effective date and required bank confirmation. The public register may have been designed as a summary of corporate decisions rather than an operating manual. Security and privacy would give the organisation good reasons not to disclose account details, specimen signatures, authentication steps or live transaction information.

The omission of “external” may therefore have been no more than drafting compression. If the table defined the relevant category or applied the threshold only within a clearly bounded part of the delegation, repeating the adjective in the resolution summary may have added nothing. Similarly, “above mentioned signatory persons” may have been perfectly clear to directors and the bank because the complete document identified the covered class. Public readers may be seeing only the top layer of a sound private system.

Nothing in the published record disproves that account. There is no evidence of a failed control, a disputed instruction, an unauthorised transaction or a neglected revocation. There is no basis for faulting any named person. Concision is not itself a governance failure, and confidentiality can be necessary in banking administration.

But plausibility is not verification. The same evidence that permits the benign reading does not prove it. A public summary can reasonably withhold sensitive mechanics while still stating the categorical facts needed to establish continuity: which version governed, whether the transaction perimeter changed, which role classes were covered, what category of combination applied, when the mandate became active and whether the prior version was revoked. The gap is not that the public cannot see credentials. It is that the public cannot see status.

The narrow answer is therefore not maximal disclosure. It is bounded disclosure. AFRINIC could show the skeleton of the transition while keeping the nervous system private. A dated version note could confirm that Delegation 002-2013 superseded a named predecessor, that the high-value exception applied to a defined transaction class, that principal and alternate roles followed stated categorical conditions, that the bank activated the mandate on a stated date and that older mandates were revoked. An approval-reference convention could connect high-value decisions to execution without revealing their commercial substance.

This remedy respects both sides of the governance problem. Members and operators have a legitimate interest in whether an institution they depend on maintains traceable authority. The institution has a legitimate interest in not publishing attack-enabling financial information. Bounded transparency serves both. It makes control architecture testable without turning the register into a security manual or a transaction ledger.

A treasury rule, not a claim to rule the Internet

Resolution 201304.171 concerns AFRINIC’s own corporate transactions. AFRINIC is a private technical bookkeeper and coordinator. Its Board can govern the company’s treasury, adopt delegations, appoint signatories, create alternate arrangements and reserve large commitments for collective approval. Those are genuine corporate powers with consequences for the organisation and the members who rely on its operational competence.

They are not sovereign powers. AFRINIC is not a government, regulator, police force, punishment authority, confiscator, prosecutor or court. A bank mandate does not confer jurisdiction over operators. A Board resolution does not create public-law control of routing, networks or number resources. A private organisation can maintain an important registry and coordinate technical records without becoming the ruler of the systems those records describe.

Keeping this boundary clear improves rather than diminishes the analysis of the financial control. If the threshold is inflated into evidence of general authority, its actual institutional purpose becomes harder to see. The rule was designed to constrain how AFRINIC’s own authorised people could commit or execute company transactions. Its legitimacy rests on private corporate governance, not on a claim that the organisation may command the Internet community.

The distinction also clarifies why traceability matters. Members and operators may depend on AFRINIC’s continuity as a coordinator. Weakness in corporate administration could have operational consequences even though the institution has no sovereign mandate. That dependency justifies interest in versioned controls, clear roles and bounded audit evidence. It does not justify converting operational importance into coercive authority.

The bookkeeper principle is especially useful here. A bookkeeper’s ledger should reflect operational reality: who is authorised, under which version, for what class of action and subject to which approval. When the operational matrix changes, the record should change with it. Repeating the threshold was consistent with that discipline because it attached the high-value exception to the new sequence. The unresolved scope and activation links show where the published ledger stops short of proving the whole operational state.

There is no contradiction between saying the corporate control was real and saying the institutional authority was narrow. The Board’s reservation above USD 100,000 could be binding within AFRINIC while having no power to punish an operator, seize number resources, adjudicate a dispute or regulate a network. Precision about domain is a condition of institutional legitimacy. It prevents an internal safeguard from being mistaken for an external mandate.

What the April 2013 sequence establishes—and what it does not

The April sequence establishes five things with confidence. First, the Board adopted Delegation of Authority 002-2013 in connection with a changed organisational structure. Second, the public signatory architecture contained three principals and a separate conditional-alternate layer of two people. Third, Resolution 201304.171 authorised the “above mentioned signatory persons” under the delegation framework. Fourth, the Board again reserved transactions in excess of USD 100,000 or equivalent for its own approval. Fifth, the bankers were to be furnished with relevant policies and resolutions.

It also establishes a precise comparison with the minimum 2011 baseline. The amount, the above-threshold construction, the foreign-currency equivalent and the Board approver were carried forward. The visible participant design changed. The word “external” disappeared. Those are the durable public facts.

The sequence does not establish the Board’s motive for repeating the rule. It does not prove whether omission of “external” altered scope. It does not resolve whether “above mentioned signatory persons” meant only the three principals or included conditionally active alternates. It does not reveal the signing combination, the lower-tier authority framework or separation among initiation, approval and execution. It does not establish an exact April day, an effective date, bank receipt, bank activation or revocation of predecessor mandates.

Most importantly, it does not establish any transaction or incident. No payment, commitment, approval, refusal, split, breach, loss, fraud, corruption, negligence, illegality or personal fault is identified. A control gap in the public record is not evidence that the underlying control failed. Uncertainty about implementation is not evidence of wrongdoing.

The resulting judgment is deliberately narrow. Resolution 201304.171 is evidence of a control being carried through organisational redesign. By saying the threshold again, the Board reduced the risk that readers would treat the new delegation as detached from the high-value exception. By changing the surrounding matrix and one word of scope without publishing the table or version bridge, it left other questions open.

That combination is what gives the resolution its institutional significance. It is neither a duplicate with nothing to teach nor a new origin story. It is a transition marker. It shows the Board understood, at least at the level of the published sequence, that an approval boundary belongs beside the authority it constrains. It also shows why a number alone cannot complete a control: the perimeter, participant class, execution path and activation history must be traceable too.