APNIC Account Holders Are Not All Equal Principals

Summary

• APNIC's institutional population includes direct Members, direct Non-Member account holders, National Internet Registries that are themselves APNIC Members, and organizations served downstream by an NIR. A resource record or fee payment does not collapse these relationships into one membership class.
• Direct Members form APNIC's governing body under the by-laws. They elect the Executive Council and exercise other member powers, but their electoral votes are weighted by seven tiers: one vote for an Associate Member through 64 votes for an Extra Large Member.
• Non-Member account holders can receive direct resource services and have notice, response and Executive Council appeal rights under their agreement, yet APNIC's voting terms expressly exclude them from online voting.
• An NIR is one APNIC Member. Its own members or account holders contract locally with the NIR and do not each receive an APNIC ballot merely because their resources sit within the regional hierarchy. APNIC's published explanation says each recognized NIR has 64 votes, regardless of how many organizations it serves.
• Policy access is broader than electoral status: anyone with an interest may propose and discuss resource policy, and a show of hands is an aid to consensus rather than a vote. Even so, member-meeting confirmation and Executive Council endorsement remain separate stages, while a customer's remedy usually follows its direct contract.
• APNIC should publish a rights map and unduplicated population statistics for each relationship, together with votes used by tier, NIR representation, policy participation and appeal outcomes. Without those denominators, account holder is an operational label that conceals unequal institutional standing.

One registry system, several principals

The phrase APNIC account holder sounds like a single legal category. Operationally, it can be useful: an organization has credentials, contacts, fees, resources or services associated with an account. Institutionally, it is too broad. One account can belong to a direct APNIC Member. Another can belong to an organization that contracts directly for resource services without membership. A third organization may receive resources and account administration from a National Internet Registry, while the NIR itself is the APNIC Member. These parties occupy the same regional registry system but do not stand in the same relationship to APNIC.

That difference matters because governance claims use words such as member, community and stakeholder as though they were interchangeable. They are not. A direct Member may vote in an Executive Council election, exercise powers at a Member Meeting and make requests under the membership agreement. A Non-Member may have a direct service contract and a defined appeal without a ballot. An NIR customer may participate openly in regional policy discussion while relying on a local contract for service and remedy. The NIR may cast the APNIC votes attached to its own membership tier.

The resulting structure is not necessarily illegitimate. Layered administration can provide local language, local payment, local law and proximity to network operators. Direct non-membership can serve organizations unable or unwilling to join. Tiered voting can be defended as reflecting resource scale and financial exposure. Open policy consensus can amplify people who hold no electoral vote.

The problem begins when operational connection is presented as equal authority. Paying is evidence of a financial relationship. Registration is evidence of a record. Receiving resources is evidence of a service chain. None, by itself, identifies who can elect, amend, appeal, petition or bind whom. The governing instrument and direct counterparty must be identified first.

The corporate member is a specific legal position

The APNIC Membership Agreement does more than purchase registry services. It says APNIC Pty Ltd is an Australian non-profit proprietary limited company and that the Company established a Special Committee called APNIC under its corporate constitution. The organization signing the agreement is accepted as a member of that Special Committee. The Committee includes Members, General Meetings, an Executive Council appointed by Members and a General Secretariat staffed by the Company.

That construction explains why membership carries governance rights beyond service delivery. The agreement says the Company must establish open communication mechanisms, promote training, undertake activities required by Members, consider member requests concerning operations, ensure Executive Council consideration of referred requests, and provide rights and services under APNIC documents. The Member pays fees, keeps information accurate and complies with the agreement and applicable documents.

The membership term is one year and renews by payment under the then-current standard agreement. Failure to renew within the stated period can lead to revocation of member rights. Membership is therefore neither a permanent status attached to a prefix nor an automatic consequence of being visible in registry data. It is a renewable institutional relationship.

This matters for the word principal. A direct Member is one of the parties through which the Special Committee is governed. It is not merely the entity of a registry record. But even direct Members do not individually direct staff or own APNIC's resources. They exercise defined collective powers through the by-laws, meetings, elections and consensus procedures. Principal is therefore a useful governance metaphor only if kept bounded: direct Members are constituents with formal rights, not owners of every institutional act.

The by-laws reserve governing functions to Members

The current By-laws of APNIC state that Members are the governing body of APNIC. They identify the Executive Council as acting on APNIC's behalf and the Secretariat as the operating administration. Membership is open to qualifying persons and organizations engaged in open-system network services, subject to dues and Executive Council authority.

The listed member functions are substantial. Members determine general policies for APNIC's entities, consider Executive Council reports and adopt decisions, examine and approve accounts where appropriate, elect the Executive Council, consider by-law amendments, review or amend Executive Council decisions by a two-thirds majority of the votes of the entire membership, and call a Special Meeting through a petition representing at least one quarter of the votes of the entire membership.

These are not ordinary customer-service benefits. They constitute the formal accountability surface of the institution. A customer that can update contacts or request an address is not thereby part of the body entitled to approve accounts or elect the council. The by-laws make the boundary explicit even when public discussion uses the wider term community.

The boundary is reinforced by office rules. Executive Council members serve in a personal capacity and must act in the interests of the APNIC membership rather than the organization with which they are associated. This prevents a simplistic principal-representative model in which a council member carries instructions from one employer. The electorate chooses individuals, but those individuals owe their role to the membership collectively.

That fiduciary-style separation is valuable. It does not solve who constitutes the membership or how votes are distributed. Those questions are answered by the agreements, tiers and account relationships that sit below the general statement that Members govern.

Direct Members do not have equal electoral weight

APNIC's Membership: Tiers and Voting rights document defines seven levels: Associate, Very Small, Small, Medium, Large, Very Large and Extra Large. The effective tier is determined by chargeable IPv4 and IPv6 holdings, assessed under the published prefix thresholds. Resources that are not charged, including some historical and experimental resources, do not count in the same way.

The vote schedule doubles at each step. Associate Members receive one vote; Very Small, two; Small, four; Medium, eight; Large, 16; Very Large, 32; and Extra Large, 64. The same document says all Members receive equal access to core services, while voting entitlements depend on tier. Equal service access and equal political weight are deliberately separated.

The design has an intelligible rationale. Organizations with larger chargeable holdings can have larger fees, broader infrastructure and more exposure to registry decisions. Weighted voting may prevent thousands of small accounts from governing an institution whose continuity depends heavily on large operators. The geometric sequence is also simple to administer.

But holdings are not the only defensible measure of institutional interest. A small operator may serve an underserved economy, provide critical local connectivity or face a decision with severe relative impact. A large corporate group may hold resources through several accounts. A resource-based tier can turn historic allocation patterns and acquisition strategy into present electoral power. The public rules do not show how concentrated votes are after accounting for entities, proxies and votes actually cast.

It is therefore inaccurate to say APNIC Members vote equally. They have equal access to listed core benefits, not equal ballots. Any report of member participation should disclose both organizations and weighted votes, together with the tier distribution and, where privacy allows, concentration bands for related entities.

Authorization adds a second gate inside each Member

Vote entitlement belongs to the Member organization, not automatically to every individual associated with its account. APNIC's Online voting terms and conditions limit online voting to current Member organizations using MyAPNIC. Eligible individuals are Corporate Contacts and other authorized contacts to whom voting rights have been assigned. Two-factor authentication is required.

The terms treat a submitted ballot as the Member's ballot and record the Member, submitting person, number of votes, time and network address for audit, while keeping the choices separate from voter identity. A Corporate Contact can distribute voting authority among contacts. Operational control of credentials and contact roles therefore mediates the organization's formal entitlement.

This layer is necessary for security and attribution. It also means that an organization can be a Member with votes on paper while failing to exercise them because contacts are stale, authority is unclear or authentication is incomplete. Conversely, one individual may control voting for several related or separately authorized organizations. Public totals should distinguish members entitled to votes, members with valid voting contacts, members that cast at least one vote and weighted votes actually used.

The same terms provide the clearest negative rule: Non-Member account holders are not eligible to vote online. They can possess an APNIC account and direct services, but the voting system recognizes the corporate membership relationship, not account existence alone.

This is the first decisive answer to the title. Account holders are unequal before any candidate is considered. Some organizations have no APNIC electoral entitlement. Among Members, the number of votes varies sixty-fourfold. Within a Member, only authorized individuals can activate those votes.

Non-Member does not mean unprotected customer

The APNIC Non-Member Resource Services Agreement establishes a direct relationship with organizations that receive resource assignment, registration or database services without joining APNIC. The agreement begins when APNIC receives the non-member service fee and continues until termination. The organization pays maintenance charges, supplies accurate information and complies with applicable documents.

The absence of membership removes the governance layer, not every procedural protection. If APNIC reasonably believes the organization breached the agreement or another applicable document, it must send written notice describing the believed breach, required corrective action, a reasonable response period and intended consequence. The organization may deny the breach, show that it has cured it or identify exceptional circumstances. If APNIC proceeds to revocation, the organization may appeal to the Executive Council, which must consider the appeal within 30 days and withdraw the notice if the appeal is justified.

These rights are consequential. A Non-Member can challenge an adverse service decision through a direct APNIC contractual path. It does not need an electoral vote to possess a remedy. This is why voice and remedy must not be collapsed. Corporate membership can supply voting power without deciding every case-level dispute, while a service contract can supply a case-level appeal without political power.

The agreement also imposes hard consequences. A revocation notice can require immediate cessation of specified delegated resources, and APNIC may seek court restraint subject to judicial discretion. An appeal considered within 30 days may not preserve operations unless interim protection is available through another route. The published form proves the presence of a remedy, not its practical adequacy in every urgent case.

No complete public statistics reviewed here show how many Non-Member appeals are filed, how quickly they are decided, how often notices are withdrawn or how service is preserved during review. Their governance status should not be described as voiceless, but neither should a contractual appeal be presented as equivalent to membership.

An NIR is one Member, not a bag of direct ballots

The APNIC and NIR Member Relationship Agreement creates a distinct bridge. APNIC recognizes the NIR organization as both an APNIC Member and the National Internet Registry for a named country or economy. The agreement describes NIRs as serving organizations locally and adapting procedures and services to cultural differences while remaining consistent with regional and global resource policies.

The legal member is the NIR. The agreement does not say that every ISP, end user or local account holder served by that NIR becomes an APNIC Member. Instead, the NIR must enter a formal membership agreement or suitable contract with its own members or account holders and require compliance with policies consistent with APNIC's. The NIR must take reasonable steps to enforce those terms.

This creates two contractual levels. APNIC provides rights, services and resources to the NIR under the regional agreement. The NIR provides local rights, services and resources under its customer or member agreements. An organization at the lower level may be recorded in resource administration and subject to APNIC-derived policy without being a direct party to the APNIC-NIR agreement.

The agreement's termination provisions confirm the separation. If the NIR relationship ends, APNIC has a right to reinstate members of the NIR as APNIC Members only if those organizations enter an APNIC Membership Agreement. Direct status is not automatic. A new contract is the bridge.

The document also says, to the extent local law permits, the NIR should guarantee that local registries, providers and end users can choose between APNIC and the NIR as the registry from which they receive resources. This is an important protection in principle. Its practical availability, cost and legal constraints can differ by economy, and the reviewed materials do not provide a current comparative measure.

NIR aggregation changes the relationship between people served and votes

APNIC's official 2022 explanation, The NIR structure: Annual fees and voting entitlement, says the regional NIR structure began in 1996. At the time of the article, APNIC recognized seven NIRs: CNNIC, IDNIC, IRINN, JPNIC, KRNIC, TWNIC and VNNIC. Each NIR was one APNIC Member in the Extra Large tier with 64 votes.

The article offers a hypothetical comparison. If one NIR served 500 organizations, each with a /22, the NIR would have 64 APNIC votes. If the same 500 organizations were each direct Very Small Members, the aggregate entitlement would be 1,000 votes because each would receive two. The associated fee comparison also changes dramatically because NIR holdings are aggregated and a multiplier is applied.

The numbers are an illustration, not a count of any actual NIR electorate. They nevertheless expose the institutional mechanism. Aggregation can turn hundreds of local service relationships into one APNIC membership and one 64-vote entitlement. The downstream organizations may influence how the NIR uses that entitlement through local governance, but the strength and form of that influence depend on each NIR's constitution, membership rules and national context.

This is neither straightforward vote suppression nor straightforward representation. The NIR may provide local accountability that direct APNIC membership cannot: language, local meetings, domestic legal recourse and knowledge of national infrastructure. It may also concentrate regional voting authority in an intermediary whose customers do not have a direct say over the APNIC ballot.

APNIC-level transparency should therefore show NIR votes separately and encourage each NIR to publish how its APNIC voting position is authorized. A claim that NIR customers are represented is testable only if the chain from customer voice to NIR decision is visible.

Policy proposal access is broader than membership

APNIC's Policy Development Process prevents a simple conclusion that non-members lack policy voice. It says policies are developed by the Internet community through bottom-up consultation and consensus. Anyone with an interest in the management and use of Internet number resources in the Asia Pacific may join the mailing list, attend the Open Policy Meeting physically or remotely, discuss proposals and take part in decision-making.

Anyone can also become a proposal author. A formal proposal is submitted to the Policy SIG Chairs, discussed on the mailing list and presented at an Open Policy Meeting. The Chairs assess whether general agreement exists. They may ask for a show of hands expressing strong support, support, neutrality, opposition or strong opposition, but the document expressly says this is not a vote. It is a way to gauge opinion.

This openness matters for an NIR customer or Non-Member. Such an organization does not need an Executive Council election ballot to identify a resource-policy problem, propose text, build support or entity. In principle, expertise and reasons can matter more than account status. A small downstream operator can persuade a room containing much larger Members.

Formal openness is not the same as equal capacity. English-language drafting, time zones, travel, employer support, mailing-list confidence and sustained attendance can shape whose reasons are heard. An organization may be entitled to speak but unable to maintain the months of participation needed to carry a complex proposal. APNIC's process values reduce legal exclusion; they do not establish empirical equality.

Policy authorship and electoral membership should therefore be reported separately. A broad, diverse proposer base can offset some representational limits of tiered elections. It cannot be assumed without data on who proposes, entities, speaks and remains through final consensus.

Consensus and voting perform different constitutional jobs

APNIC governance contains at least two decision languages. Executive Council elections allocate tier-weighted votes to Member organizations. Resource policy uses a consensus method in an open forum, followed by continued consensus at the AGM or APNIC Member Meeting, a public comment period and Executive Council endorsement.

The distinction protects the broader community. At the Open Policy Meeting, the holder of 64 electoral votes does not formally cast 64 hands. Chairs consider reasons, objections and the degree of agreement. Non-members can participate. The process can stop a proposal that has numerical enthusiasm but unresolved major objections.

Membership still matters at a later stage. The formal process requires consensus at the AGM or Member Meeting after Open Policy Meeting consensus. If there is no consensus at either forum, the proposal cannot proceed on that cycle. After the final comment period, the Executive Council endorses the proposal before implementation. Electoral authority, member confirmation and open community deliberation therefore intersect without becoming identical.

This layered design can be a strength. It prevents resource policy from being decided solely by weighted ballots while retaining a role for the body formally responsible for APNIC. It can also obscure accountability if a decision is described only as community consensus. Readers need to know which forum reached consensus, who could participate, who assessed it, whether objections remained and what the Member Meeting and Executive Council did.

The thesis is therefore more precise than saying account holders have unequal proposal rights. Formal proposal access is broadly equal: anyone may propose. The path after proposal is not institutionally flat. Members occupy a confirmation forum, Chairs judge consensus, and the elected Executive Council supplies final endorsement. Equal entry does not mean equal power at every stage.

Remedies follow privity, not regional visibility

When a direct Member disputes an APNIC action, its membership agreement supplies notice, response and Executive Council appeal terms. When a direct Non-Member disputes an action, the non-member agreement supplies a similar but separate path. When APNIC acts against an NIR, the NIR relationship agreement governs notice, at least 45 days in specified circumstances, response, revocation and an appeal to the Executive Council within 30 days.

The NIR customer's position is different. Its immediate counterparty is normally the NIR under the local agreement required by the APNIC-NIR contract. If the local registry denies, suspends or changes a service, the customer's remedy must first be found in that local agreement, the NIR's rules and applicable law. The regional APNIC agreement can require the NIR to maintain compliant arrangements, but it does not automatically grant every downstream customer the NIR's appeal as if each were a direct party.

This is the remedy gap concealed by a shared registry vocabulary. Two organizations can hold similarly sized address space under the same regional policy. The direct APNIC Member may appeal to APNIC under its agreement. The direct Non-Member may also appeal under its own form. The NIR customer may need to pursue a local review whose independence, timing and interim relief differ.

No conclusion follows that local remedies are weaker. Some may be faster, available in a familiar language and enforceable in a nearby court. Others may be less developed or less transparent. The evidence needed is comparative: NIR customer agreements, notice periods, independent review, interim protection, published outcomes and escalation routes.

APNIC can improve clarity without assuming local jurisdiction. It can maintain a public matrix identifying the direct contractual party for each relationship and linking to each NIR's current customer terms and complaint route. An account dashboard should tell an organization which agreement governs it rather than leaving status to be inferred from billing language.

Payment is not a franchise fee

All of these relationships involve money, but payment has different institutional meanings. A direct Member pays dues as a condition of effective membership and renewal. A Non-Member pays service and maintenance fees under a contract that expressly preserves non-member status. An NIR pays APNIC as one Member based on aggregated holdings and its fee rules, while its customers may pay the NIR locally.

It is tempting to invoke a simple principle: if an organization pays, it should vote. That principle overlooks the choice to purchase services without membership and the intermediary structure of an NIR. It also overlooks tier weighting, under which payment does not purchase one equal ballot but accompanies a resource-based vote schedule.

The better accountability rule is no unrepresented burden without a visible channel. A payer should know whether it is a Member, who represents it if not, how it can influence fees and service standards, and where it can challenge an adverse decision. The channel can be direct election, open consultation, local NIR governance or contractual review, but it should be explicit and usable.

Fee consultation is especially sensitive. A direct Member can connect fee concerns to Executive Council elections and member meetings. A Non-Member can communicate as a customer but lacks that ballot. An NIR customer may face local charges shaped by both APNIC's upstream structure and NIR decisions. Publishing the components, decision-maker and consultation record allows each constituency to identify responsibility.

Payment establishes stake. It does not establish identical institutional status. APNIC's legitimacy depends less on pretending otherwise than on showing why each relationship exists and how each affected organization can respond.

Registration is not membership either

RFC 7020 describes the Internet Numbers Registry System as a hierarchy. IANA coordinates global pools; Regional Internet Registries serve regions; local or national registries and providers can distribute resources further. Registration aims to preserve uniqueness and maintain information at the relevant level. The architecture anticipates layers rather than one direct contract for every operational user.

That hierarchy explains why an organization can appear in registry data without becoming a corporate member of the upstream institution. A downstream assignment can be recorded under a provider. An NIR can maintain local customer information while coordinating regional records. A resource may be visible to APNIC because regional uniqueness requires it, not because the operational user signed APNIC's Membership Agreement.

Confusing record visibility with membership produces two errors. The first overstates rights: the recorded organization is assumed to possess an APNIC vote or direct appeal that its contract does not grant. The second overstates APNIC's direct responsibility: every downstream service dispute is assumed to be between APNIC and the user, even where the NIR or provider made the immediate decision.

Hierarchy does not remove upstream responsibility. APNIC sets regional rules, recognizes NIRs, contracts for compliance and maintains services on which downstream parties rely. It should monitor whether intermediary arrangements meet published standards and whether continuity exists if an NIR relationship ends. But accountability should track the actual chain rather than leap over it.

An accurate public account needs three separate fields: where the resource is registered, which organization provides the service, and which organization holds governance membership. Those answers can coincide for a direct Member. They can diverge for a Non-Member or NIR customer.

NIR choice is valuable only if it is practical

The NIR agreement requires the NIR, so far as local law permits, to guarantee local registries, providers and end users freedom to choose between APNIC and the NIR for resources. This clause recognizes that an intermediary should not automatically become a compulsory governance gate. It offers an answer to the representation concern: an organization that wants direct APNIC membership may choose it.

The strength of that answer depends on facts not published in one regional table. Can an organization in every NIR economy contract directly with APNIC for the resource type it needs? Are fees, currency, tax, documentation and language manageable? Does domestic regulation require use of the NIR? Can an existing NIR customer transfer its relationship without operational disruption? Does direct membership change the handling of existing resources?

The agreement qualifies the obligation by local law, appropriately acknowledging national constraints. That same qualification means formal choice cannot be assumed to be uniform. APNIC and NIRs should publish economy-specific guidance that identifies legal restrictions and practical steps without portraying one route as superior.

Choice also has collective consequences. If only large, internationally connected organizations can move direct while smaller networks remain downstream, APNIC's direct electorate may become skewed even though every organization has a nominal option. If direct membership is easy and commonly used across organization sizes, that evidence would reduce the concern.

Current unduplicated numbers are therefore essential: direct Members located in NIR economies, NIR customers by broad size band, organizations changing relationship, declined direct applications and stated legal constraints. Without those figures, the choice clause is an important protection whose practical reach remains unknown.

Corporate groups can magnify or mask inequality

APNIC's tier schedule applies to Member accounts and holdings, while the by-laws contain rules concerning associations with organizations and corporate groups in Executive Council composition. The public voting materials reviewed here do not provide a complete annual map of beneficially related Member accounts and their combined entitlements.

This matters in both directions. One corporate group could hold several memberships and aggregate votes across them. Another large operation might be represented through one NIR membership and share its 64 votes with hundreds of unrelated local customers. Counting Member organizations alone would treat these arrangements as comparable when their underlying constituencies are very different.

Privacy and commercial sensitivity limit disclosure. APNIC need not publish confidential ownership records or individual ballots. It can report concentration in bands: the share of entitled and cast votes associated with the largest disclosed corporate groups, the number of Members linked to multiple accounts after internal review, and the distribution of votes by tier and NIR status. Methodology and uncertainty should be published.

The same care applies to proxies and authorized contacts. A contact exercising votes for several Members can be legitimate, especially in a group or service organization. Aggregate concentration shows whether the practice is exceptional or structurally important. It does not establish coordinated voting, which would require separate evidence.

Without affiliation data, assertions of capture are speculative. So are assertions that weighted votes are broadly dispersed. Institutional confidence should rest on measured concentration, not assumptions based on account count.

A rights matrix is more useful than the account-holder label

APNIC could make the structure understandable through a public matrix with rows for direct Member, direct Non-Member, NIR Member and NIR customer. The columns should identify:

Relationship	Direct counterparty	APNIC election vote	Policy proposal access	Principal case-level appeal	Member governance powers
Direct APNIC Member	APNIC	1-64 votes by tier, subject to current status and authorization	Open	APNIC agreement and Executive Council route	Yes, under the by-laws
Direct APNIC Non-Member	APNIC	No online vote	Open	Non-Member agreement and Executive Council route	No
NIR Member	APNIC	64 votes under the published Extra Large classification described in 2022	Open	NIR agreement and Executive Council route	Yes, as the NIR organization
NIR customer or member	NIR	No automatic APNIC vote from customer status	Open	Local NIR terms and applicable law	Depends on NIR governance, not automatic APNIC membership

The table is a conceptual map, not a substitute for current agreements. APNIC should maintain the authoritative version, date it and link each cell to governing text. Where a local NIR gives its customers voting rights over NIR decisions, that should be described separately rather than treated as an APNIC ballot.

An account dashboard could display the same facts privately: relationship type, governing agreement version, member tier, vote entitlement, authorized contacts, upstream or downstream registry, appeal route and policy-participation links. A customer should not need corporate-law expertise to learn whether it is a Member.

Clarity would also improve APNIC's public claims. Instead of saying account holders participated, reports could say direct Member organizations voted, open community entities discussed policy, or NIR customers responded through a named consultation. Each statement would identify the relevant principal.

The missing denominators prevent a complete legitimacy test

The published instruments establish rules with considerable precision. They do not provide all the population data needed to judge representation. A complete annual report would include unduplicated counts of direct Members by tier, direct Non-Members, NIR Members and distinct NIR customer organizations. It would explain whether one organization appears in more than one class and how duplicates are handled.

For elections, APNIC should report organizations entitled to vote, weighted votes entitled, organizations voting, weighted votes used, turnout by tier, NIR votes used, proxy use and privacy-preserving affiliation concentration. Candidate totals alone cannot show whether a result drew broad organizational support or concentrated weighted support.

For policy, useful measures include proposal authors, mailing-list contributors, Open Policy Meeting speakers, remote entities, expressed objections and sustained participation by relationship class, economy, gender and organization type where voluntary and safe. Consensus should never become a quota calculation, but participation data can reveal whose absence the Chairs must address.

For remedies, APNIC and participating NIRs should publish notices, cures, appeals, decision times, withdrawals, interim measures and restoration times in aggregate. Contract rights become more credible when their use can be evaluated.

None of these denominators was complete in the materials reviewed for this article. Current totals should therefore not be invented from old annual reports, account labels or the 500-organization hypothetical. The evidence supports a structural conclusion, not a current headcount of each constituency.

A defensible layered model needs reciprocal duties

APNIC's architecture can be defended if each layer supplies a real accountability path. Direct Members receive tiered electoral rights and corporate powers; in return, vote concentration and member decisions should be transparent. Non-Members receive a direct service contract and appeal; in return, APNIC should show that the route is timely and meaningful. NIRs receive recognition, services and votes; in return, they should demonstrate local accountability and compliant customer remedies. NIR customers receive local service and open regional policy access; in return, they need clear terms and a practical escalation route.

The model fails when rights disappear between layers. A customer should not be told to influence the NIR if the NIR offers no governance channel. An NIR should not be treated as representative merely because it aggregates resources. A Non-Member appeal should not be described as equivalent to a vote. A direct Member's 64 votes should not be described as community consensus in a policy forum that uses a different method.

Reciprocity also requires APNIC to consult beyond its electorate. The Executive Council is elected by Members, but its decisions can affect non-members and downstream users. Dedicated Non-Member and NIR-customer consultation, with published response analysis, can widen the evidence available to council members without altering the corporate franchise.

Equal treatment does not require identical contracts. It requires differences to be relevant, visible and accompanied by proportionate protection. Local registry service can justify an intermediary. Resource scale can inform fees. Neither should become an unexplained reason why one affected organization cannot find who hears it.

Evidence boundaries and watchpoints

The strongest public evidence is textual. The by-laws identify Members as the governing body. The tier document assigns votes from one to 64. Voting terms exclude Non-Members. The NIR agreement makes the NIR the direct APNIC Member and requires local contracts. The policy document opens proposal and discussion to anyone with an interest. The agreements provide distinct appeal routes for direct counterparties.

Several practical facts remain unknown: the current number of unduplicated organizations in each relationship; votes actually used by tier and affiliation; the number of NIR customers; the terms and remedies of every local NIR; the practical availability of direct membership in each economy; and the demographic distribution of policy influence. The article does not infer them.

Future watchpoints include revisions to the tier schedule, new NIR agreements, publication of election turnout by class, stronger disclosure of corporate-group concentration, economy-level choice guidance and comparable appeal statistics. Policy meetings should also be watched for whether open access produces broad participation rather than merely formal permission.

The central conclusion is stable even without those totals. APNIC's account population is not a single demos. It is a chain of direct members, direct customers, institutional intermediaries and downstream users. Some can propose but not vote. Some can appeal but not elect. Some hold weighted votes on behalf of an organization that serves many others. Institutional legitimacy begins by naming those differences, then proving that every layer has an accountable route to the decisions it must live with.