Subsidiaries, Shells and the Multiplication of Membership Voice

Summary

• RIPE NCC gives each unsuspended legal member one vote and limits a single member with multiple LIR accounts to one vote. The rule blocks account-based vote buying inside one entity.
• RIPE NCC also states that related legal entities may apply for membership. Public rules do not show a general group-wide cap based on ultimate control, creating a structural route by which separate affiliates could each hold a ballot.
• Corporate separateness often reflects real geography, regulation, financing and operational responsibility. Common ownership is not proof of manipulation, and no named member should be accused without evidence.
• Election integrity requires a test for control and coordinated action, not a ban on groups. Auditors should receive proportionate ultimate-control information, publish concentration bands, aggregate affiliates for specified votes where control is established, and provide an expedited appeal.
• The objective is not to expose private ownership indiscriminately. It is to ensure that a rule advertised as one-member-one-vote does not silently become one-incorporation-one-vote when incorporation choices are controlled by the voter.

The constitutional boundary can be moved with company paper

One-member-one-vote appears to answer the power problem elegantly. A small network and a multinational each receive one ballot. A member with several LIR accounts receives no additional voice. Financial scale, address holdings and account count do not purchase formal control. The association treats its legal members as equals.

The equality depends on the boundary around a member. If a multinational holds all operations in one legal company, it receives one vote. If the same economic group uses six subsidiaries and each becomes a member, the register may contain six voters. The operational footprint may be identical. The constitutional result changes with corporate organisation.

This is not evidence that any group has manipulated an election. It is a design vulnerability. Rules based on legal personality always face the question whether several legal persons act under one ultimate direction. Company law deliberately permits groups to separate liability, licences, taxation and local duties. Association law deliberately assigns rights to each admitted member. Electoral integrity must decide when those separations should be respected and when they should be aggregated.

The answer cannot be improvised after a contested result. If common control matters, the definition, disclosure, review and remedy must exist before ballots open. Otherwise administrators gain discretion to recognise convenient entities and challenge inconvenient ones.

RIPE NCC closes the account route but not visibly the affiliate route

RIPE NCC's published membership guidance makes one safeguard explicit. A member can open more than one LIR account, but those accounts do not generate separate votes. The oldest LIR account receives the voting position. The Articles of Association assign one vote to each unsuspended member, not to each paid account.

This is a coherent anti-amplification rule. It prevents an existing member from converting sign-up fees and annual contributions into a stack of ballots merely by opening more accounts. It also confirms that the association considers legal identity, not service volume, the proper unit of equality.

The same guidance says that applications from related legal entities with the same contact person, directors, owners or applicant are processed in the same way. The statement is useful because it acknowledges affiliation. It does not say that related entities are prohibited, automatically aggregated or limited to one group vote. Processing may include scrutiny, and not every application will necessarily succeed. The public text does not supply a complete affiliation rule.

The structural gap is therefore narrow but real. Multiple accounts inside one member produce one vote. Multiple legal members under common control can, in principle, occupy several entries unless another disclosed rule applies. The equality principle moves from economic group to legal entity without explaining why that boundary is resistant to choice.

A legal entity is a good administrative unit

There are strong reasons to use legal entities. They can sign contracts, hold assets, pay fees, sue and be sued. Their authorised representatives can be verified. Corporate registers provide evidence of existence. If an election dispute arises, the institution knows which person holds the right and which jurisdiction governs its authority.

Operational units are harder. A "network" can span subsidiaries and countries. Autonomous systems do not map cleanly to companies. Brands can be shared. Business divisions appear and disappear without legal notice. Customer populations fluctuate. Ultimate investors may be passive while local directors control operations.

The legal entity is therefore not an arbitrary fiction. It is a stable point of responsibility. A subsidiary can owe separate regulatory duties, employ its own engineers and contract with its own customers. Denying it membership merely because a parent exists could exclude a genuine institutional interest.

The case for control disclosure does not reject legal personality. It asks whether formal equality should count independently controlled interests or documents capable of being multiplied by one controller. That is a familiar question in competition law, financial regulation and corporate elections, though the answers in those fields cannot simply be copied into RIR governance.

Common ownership is not common control in every case

Ownership comes in degrees. A parent with all voting shares ordinarily controls a subsidiary. A minority investor may not. Joint ventures can require consent from several owners. A trust, foundation or public body can exercise influence without ordinary shares. Contractual rights, board appointment powers and financing covenants can shift practical control.

Even wholly owned subsidiaries may possess meaningful independence. Banking, telecommunications or national-security rules can require local boards and ring-fenced operations. Minority protections can constrain the parent. Insolvency duties can force directors to act for the subsidiary rather than the group. Different management teams can adopt different governance positions.

An election rule based only on a percentage ownership threshold will therefore produce errors. It can aggregate entities that decide independently and miss entities coordinated through contract or common management. The relevant concept is the power to direct the member's vote or appoint the body that does so.

That concept must be evidenced. Shared directors, registered addresses, email domains or advisers are indicators, not conclusions. A fair process would allow the member to explain why apparent affiliation does not amount to coordinated electoral control.

The word shell should be used sparingly

"Shell company" often functions as a verdict rather than a description. A company with no large staff or physical office may hold intellectual property, licences, infrastructure or financing for legitimate reasons. Network operations can be outsourced. A resource-holding entity can be designed to preserve continuity across changes in operating contractors.

Public company information is uneven across the RIPE NCC service region. Some registers disclose shareholders; others do not. Nominee directors and holding chains can obscure control without violating law. A superficial investigation risks accusing lawful members on incomplete evidence.

Governance analysis should therefore focus on functions. Does the entity have independent directors? Who approves its registry vote? Does it bear separate contractual or regulatory responsibility? Does it operate or finance a distinct network? Can the parent instruct it? These questions are more useful than attaching a pejorative label.

The title names shells because deliberately thin entities are one possible vehicle for amplification. The evidence reviewed here does not establish that named RIPE NCC members use them for that purpose. The institutional obligation is preventive: rules should work before misconduct is proven.

Vote multiplication is an incentive problem

If each separate member receives one vote and membership costs are affordable relative to the value of control, groups have an incentive to organise at the level that maximises voice. Most will not restructure solely for an RIR election. Incorporation, accounting, tax, compliance and annual fees impose real costs. Yet an existing multinational may already possess many eligible entities, making the marginal cost of additional membership lower.

The benefit varies by contest. Routine resolutions attract limited participation and may not justify coordination. A close board election, charging decision or constitutional amendment can make a handful of votes consequential. The value rises when turnout is low because each organised vote represents a larger share of ballots cast.

Institutions should design against incentives even when no abuse is observed. A bank does not wait for fraud before separating duties. An election body does not wait for duplicate ballots before maintaining a register. The existence of a profitable route can corrode trust once entities notice it.

An anti-evasion rule can reduce the incentive by making the outcome depend on control rather than the number of incorporation certificates. It must also avoid punishing legitimate growth or acquisitions. Clear prospective rules are essential.

APNIC shows how entity and scale can combine

APNIC's by-laws place membership in a person, firm, corporation or other eligible organisation and permit authorised representatives and proxies. Its tier rules assign votes according to chargeable address holdings, from one to 64. A corporate group question therefore has two dimensions: how many member entities it controls and how many votes each entity's tier carries.

The published structure does not itself prove group amplification. It does show that formal power can be distributed through legal accounts and resource holdings rather than ultimate-control groups. A group with several qualifying members could possess entitlements shaped by each member's tier.

APNIC has strong reasons to recognise separate regional businesses and resource responsibilities. The Asia-Pacific region contains national registries, public bodies and complex multinational structures. A blunt one-group-one-vote rule would fit poorly.

That complexity strengthens the case for audit rather than secrecy. Election administrators can collect affiliation data confidentially, report concentration and apply a rule only where the same controller directs several ballots. Public trust does not require publishing every ownership document.

LACNIC and AFRINIC use organisation-level gates too

LACNIC assigns membership and voting weight to qualifying organisations. Active A members receive between one and eleven votes according to address holdings, while founding members receive one. AFRINIC defines Resource Members as legal entities or persons that qualify for resources, sign the registration agreement and pay the relevant fees.

These rules are region-specific, but they share the same general dependence on legal organisation. Corporate groups can contain several organisations. Whether those organisations are admitted, how holdings are attributed and whether votes are coordinated become election-integrity questions.

Comparative evidence should not be overstated. The absence of a prominently published ultimate-control rule on a reviewed page does not prove that staff perform no related-party checks. Application procedures, fraud controls or governing law may supply additional safeguards. The finding is about public auditability: an outsider cannot reconstruct how group voice is measured from constitutional rights alone.

Because legitimacy claims are public, the governing principle should also be public. Confidential implementation details can protect security and privacy, but members must know whether affiliation changes their vote.

Three models for treating affiliates

The first model is strict legal-entity equality. Every admitted entity votes independently, regardless of ownership. This is simple and respects corporate separateness. It is also most vulnerable to deliberate multiplication and can give differently organised groups unequal power.

The second is full group aggregation. All entities under common ultimate control receive one collective vote. This blocks multiplication but can erase distinct regulated businesses and create difficult control disputes. It may also concentrate the vote at a parent outside the service region or operational context.

The third is conditional aggregation. Separate affiliates vote independently by default, but their votes are combined for specified matters when one person can direct them or when they coordinate under a common voting instruction. Independent subsidiaries can rebut aggregation with evidence of separate governance.

Conditional aggregation is more accurate and more demanding. It requires definitions, declarations, audit and appeal. The burden may be justified for board elections and constitutional changes while unnecessary for informal consultations. Institutional design should target the decisions where multiplied formal power can alter control.

Ultimate-control disclosure is the missing input

Election administrators cannot apply any group rule without knowing who controls members. Legal names and registration numbers identify immediate entities. They do not reliably reveal the natural person, public body, foundation or listed parent at the top of a chain.

Other regulatory fields have developed beneficial-ownership concepts, including the Financial Action Task Force's guidance on accurate and current information about legal persons. Those standards address financial crime, not association elections, and should not be imported wholesale. They demonstrate that collecting control information is administratively possible.

An RIR disclosure should be narrower. It needs the person or body able to direct the member's governance decisions, affiliated RIR members, relevant ownership percentages and board appointment rights. It does not need every passive investor or full wealth history.

The declaration should be renewed before each election and when control changes. A risk-based auditor could request documents from complex or inconsistent entries. Members should certify accuracy through an authorised officer. Deliberate concealment should carry proportionate consequences after notice and review.

Privacy is not a reason for institutional blindness

Ownership information can be sensitive. Publishing the home address or identity documents of a small company's owner would create security risks. Some jurisdictions restrict disclosure. Listed companies have dispersed investors. Public bodies and cooperatives do not fit a natural-person template.

The choice is not between universal publication and collecting nothing. A trusted independent auditor can receive detailed records. Public reports can name a corporate group where already public, or use anonymised concentration bands. The election body can state how many member votes were affiliated without exposing protected personal data.

Data retention should be limited. Documents should be encrypted, access logged and deletion scheduled after legal challenge periods. Members should know the purpose and lawful basis. Information collected for elections should not drift into unrelated service decisions.

Privacy-protecting audit is stronger than relying on rumours assembled from inconsistent public registers. It gives members a defined process and a chance to correct errors.

Coordinated action can matter without common ownership

Two members can coordinate votes without being affiliates. Trade associations, consultants, candidate campaigns and informal coalitions exist in every election. Coordination is not inherently wrong; collective advocacy is part of democratic participation.

The distinction is between persuasion and hidden control. A campaign that publicly asks independent members for support is ordinary politics. A consultant holding several credentials and casting ballots without separate instructions presents a different risk. Contractual arrangements that require a member to follow another entity's vote may resemble control even without ownership.

An affiliation rule should therefore include coordinated direction but avoid criminalising shared views. Evidence might include common authorised contacts, identical instructions issued by one controller, voting agreements or inability of subsidiary boards to decide independently. Identical ballots alone are not evidence; many independent voters choose the same candidate.

Investigations should begin only on objective indicators and follow published procedure. Fishing expeditions would chill legitimate association.

Candidates and nominators create another concentration point

Group influence can operate before ballots. Several affiliated members may nominate the same candidate, satisfy support thresholds, occupy speaking slots or provide campaign resources. If only votes are aggregated, the group can still shape the ballot and agenda disproportionately.

Election rules should identify which rights require affiliate treatment. Candidate nomination thresholds, petitions, special-meeting requests and constitutional proposals may all depend on member counts. A group-wide cap can be appropriate where the right is designed to show breadth of support.

This does not mean affiliates may not support a candidate. Their support should be counted transparently as coming from one control group when breadth is the purpose of the threshold. For ordinary debate, each member can retain its voice.

The design principle is functional. If a rule asks "how many independent members support this step?", control matters. If it asks "who wishes to speak?", affiliation may matter only for time allocation and conflict disclosure.

Acquisition creates a timing problem

Corporate control can change during an election. A member may be acquired after the register closes. Two previously independent members can become affiliates. A group can divest a subsidiary but retain contractual influence. Election rules need a record date and a duty to report material changes.

Automatic retrospective invalidation would create uncertainty. A better approach is to define control at the voter-register cutoff, require updates through ballot close and give the election body power to suspend or combine entitlements where a change is verified. Completed elections should be disturbed only if the undisclosed change could have affected the result.

Mergers also create transitional fairness concerns. Newly affiliated operating companies may need time to combine membership and records. A temporary exception can preserve service while preventing duplicate governance power. Exceptions should be published in aggregate and expire automatically.

Clear timing rules stop control analysis from becoming a weapon deployed selectively against a winning side.

Independent administration is essential

The board should not decide secretly whether its supporters count separately. Ultimate-control declarations, aggregation decisions and appeals should be handled by an election body with protected tenure, a defined budget and no candidate conflicts.

The body needs access to membership and payment records, corporate documents and authorised-contact history. It should publish guidance before nominations open. Decisions affecting eligibility should give reasons, identify evidence and allow rapid review by an independent panel or arbitrator.

Because ownership structures can be complex, the standard should be civil and administrative rather than accusatory. The question is whether votes should be aggregated, not whether a member committed a crime. Sanctions for intentional false statements require a higher evidentiary threshold than a provisional aggregation decision.

An auditable chain of custody matters. Members must know who saw their documents, when a status changed and how ballots were treated. Trust in the rule will depend as much on administration as on its text.

The remedy must preserve both election and service rights

An affiliation dispute concerns governance, not entitlement to registry service. An institution should not threaten resource records or service continuity merely because two members' votes are combined. Conflating the domains would make members afraid to disclose ownership and give election officials excessive leverage.

The ordinary remedy is electoral: combine votes, reject a duplicate nomination, correct the register or require a fresh ballot if the number was outcome-determinative. Service agreements, resource records and customer relationships should continue unless a separate legal ground supports action.

The member should have an expedited appeal capable of resolving the issue before voting closes. If time is limited public evidence, ballots can be held provisionally under secure custody. Public reporting should disclose the number of aggregated or disputed entitlements without exposing confidential details.

Separating remedies is a central rule of administrative fairness. A defect in one institutional role should not become punishment in another.

How to measure concentration without publishing a target list

Annual election reports can publish useful aggregates: total eligible member entities; number belonging to disclosed control groups; largest affiliated group size; share of ballots linked to the ten largest groups; number of members claiming operational independence despite common ownership; and number of disputes or appeals.

Weighted systems should publish both member entities and voting units. Reports should distinguish potential entitlements from ballots actually cast. A group with ten eligible subsidiaries but one active voter presents a different risk from a group casting all ten.

Historical series would show whether concentration is rising. Sudden increases near an election could trigger audit. Stable, low affiliated voting might demonstrate that the theoretical vulnerability is not material. Either result would improve confidence.

Privacy safeguards can suppress small categories and avoid naming groups unless their structure is already public or disclosure is necessary to explain a decision. Institutional audit does not require public spectacle.

The strongest objections can be answered

The first objection is complexity. Control analysis costs money and delays elections. The answer is proportionality: simple declarations for most members, deeper review only on risk indicators, and aggregation limited to consequential votes.

The second is that groups pay separate fees and hold separate obligations. Payment supports service but does not necessarily justify multiplied constitutional power. Separate obligations can support an independence rebuttal when operations and governance are genuinely distinct.

The third is discrimination against multinationals. A neutral control rule applies equally to domestic conglomerates, public bodies and private groups. It addresses direction, not nationality or size.

The fourth is evasion through nominees and trusts. No rule eliminates concealment, but officer certification, document access and sanctions raise its cost. The fifth is that small groups may be over-aggregated. A reasoned appeal and evidence of independent decision-making protect them better than unreviewed staff discretion.

What cannot be claimed from the present record

The reviewed public material does not provide a list of RIPE NCC member affiliates or the share of votes under common control. It does not show that a past election was changed by subsidiaries. It does not establish that staff ignore related-party applications. No named company should be described as a shell or electoral manipulator on this evidence.

What the record does establish is the architecture. The vote is allocated per unsuspended member. Multiple LIR accounts within one member do not multiply votes. Related legal entities may apply. Public constitutional texts do not supply a complete ultimate-control denominator.

That architecture creates a testable risk. Responsible reporting should ask for the missing denominator rather than fill it with accusation. Institutions can answer by publishing their rule and aggregate results. Silence leaves members to speculate.

The burden of proof differs by action. Reforming the disclosure rule requires evidence of vulnerability and proportional benefit, which exists. Annulment or sanction would require proof of actual control, concealment and material effect, which has not been presented here.

The register sees legal edges, not economic centres

The membership register is built from legal persons because legal persons can sign agreements, pay fees, appoint contacts and receive notices. That is a reasonable administrative choice. It creates a stable surface for billing and accountability. But a legal edge is not always an economic centre. A subsidiary may be capitalised by a parent, instructed by a group treasury function, staffed by a shared operations team and represented by the same policy adviser as other group companies. The register can show several members while the actual decision centre is one boardroom.

This distinction matters most where the association claims equality. One-member-one-vote is not merely a counting rule; it is a legitimacy story about independent principals. If several entities vote under common direction, the formal count overstates the breadth of judgment. If a single group has one vote because it wisely consolidated all resources in one entity, the count may understate the affected networks. Corporate architecture becomes a governance variable.

An institution cannot eliminate this problem by pretending that corporate law has answered it. Corporate law allows groups to organise for tax, liability, financing, licensing, acquisition and local-compliance reasons. It does not guarantee that each incorporated entity is an independent political voice. A registry association should respect corporate form for service responsibility while adding its own test for electoral independence.

The test should be modest. It should ask who ultimately controls the voting entity, whether affiliates also hold membership, whether instructions are independently approved, and whether a common representative exercises authority across entities. These facts can be collected confidentially and reported in aggregate. The purpose is to measure concentration, not to expose every commercial structure.

Multiplication can happen without bad faith

The word shell suggests deceit, but many multiplication effects arise without fraud. A multinational may create local subsidiaries because local law requires them. A network group may acquire small operators and leave their legal memberships intact during integration. A hosting company may separate infrastructure assets from retail operations. An investment vehicle may own several providers that still have distinct brands, staff and customers. Each entity may have a legitimate service relationship with the registry.

The governance issue is not moral blame. It is whether the association's voting rule still measures what it says it measures. If the rule assumes independent member interests, then common control should be visible. If the rule deliberately permits group voting through separate entities, then members should know that the electorate is partly a function of corporate structuring. Either answer is better than silence.

Bad faith becomes relevant only when an entity is created or retained primarily to multiply votes, conceal control or evade a concentration rule. Proving that requires evidence of timing, control, purpose and material effect. A fair system should not infer manipulation from affiliation alone. It should create disclosure duties that allow manipulation to be tested if alleged.

This is why aggregate reporting is useful. A public statement that, for example, a certain percentage of voting members belong to declared control groups would not accuse anyone. It would help members decide whether the association needs a related-party rule, campaign-disclosure rule or affiliation cap. It would also reduce suspicion by replacing rumour with measured facts.

A related-party rule must protect real subsidiaries

Some subsidiaries are genuinely independent in governance terms. They may have local management, local customers, separate licences, separate risk, separate capital and separate operational teams. Aggregating them automatically could silence real constituencies. A rigid anti-affiliate rule may therefore be as unfair as complete blindness.

The better design is rebuttable aggregation. Related entities would be presumed affiliated for electoral-concentration reporting. They could rebut aggregation for a particular vote by showing separate management authority, distinct operational exposure and independent approval of voting instructions. The evidence need not be public in detail; an election auditor or independent reviewer could assess it.

For board elections and constitutional votes, the association could publish two numbers: legal-member votes and control-adjusted support. The legal result would remain the one required by the Articles unless a specific rule says otherwise. The control-adjusted number would inform legitimacy and future reform. Over time, members could decide whether the gap is large enough to justify a formal threshold.

This approach avoids the common false choice between counting every company blindly and treating corporate groups as illegitimate. It recognises that independence is factual. It can vary by group, country, business line and decision type. The governance system should be able to see those differences before they become a crisis.

Campaign finance is the overlooked companion issue

Vote multiplication is not the only way a group can dominate an association. Campaign time, candidate recruitment, travel support, mailing-list activity and professional advice can matter as much as the formal ballot. A corporate group with several experienced policy staff can influence agenda and candidate perception even if it has only one vote. A network of related entities can amplify messages without formally coordinating.

This does not make campaigning improper. Members should advocate. Candidates should build support. The problem is undisclosed scale. If a candidate is materially supported by a group of related members, or if a campaign manager represents several voters, the electorate should know the category of support. The same principle already appears in many non-profit and cooperative settings: disclose conflicts, sponsorship and material assistance so voters can evaluate independence.

Registry associations can adopt a light rule. Candidate statements could disclose material support from member groups, advisers or organisations with multiple membership relationships. Campaign communications could identify when they are sent on behalf of more than one member. Election auditors could review complaints about coordinated undisclosed support. Sanctions should focus on disclosure failure, not on ordinary endorsement.

The purpose is to preserve trust in the equality story. A one-member-one-vote election can tolerate persuasion. It cannot easily tolerate the hidden conversion of money, control or professional campaign capacity into unmeasured influence while still describing the outcome as the simple will of independent members.

One member should mean one independently directed interest

The ideal behind one-member-one-vote is not worship of incorporation. It is equality among independent principals. Legal personality is a practical first approximation because it identifies responsibility. Ultimate control is the necessary second check when the approximation can be chosen by the voter.

A defensible reform would define control narrowly, collect confidential declarations, report concentration, aggregate affiliates for elections and breadth-of-support thresholds, preserve independently governed subsidiaries through rebuttal, separate electoral remedies from registry service, and provide rapid independent appeal.

The reform should not attempt to redesign every corporate group. It should ensure that the association's own promise is true. One member with many accounts already receives one vote because accounts are not independent principals. Several entities directed by one centre pose the same constitutional question in a different legal form.

Future institutions, including Number Resource Society if it develops a membership or operator franchise, should build control disclosure into the electorate from the outset. A new label will not prevent old corporate incentives. Legitimacy depends on verifiable independence, not on counting certificates.

Subsidiaries can be real operators. Holding companies can bear real risk. Thin entities can serve legitimate purposes. None of those facts removes the need to know who directs the vote. The fair rule is neither suspicion nor blindness. It is disclosed control, reasoned classification and an electorate that counts independently governed interests once.

The same discipline should apply when the risk seems small. A low-stakes election can set habits that matter during a crisis. If the association never learns to ask who directs multiple votes, it will be poorly placed to answer the question when a contested fee, suspension, certification rule or court-sensitive decision makes the answer urgent. Control disclosure is cheapest when nobody is accused and most members treat it as routine hygiene. Waiting until a disputed result produces suspicion makes every later request look punitive.

The practical standard is therefore preventative. Collect the minimum facts, protect commercially sensitive details, publish aggregate concentration and give members a narrow right to challenge material concealment. That does not weaken one-member-one-vote. It makes the phrase accurate enough to bear the legitimacy weight placed on it.

It also gives honest corporate groups a defence. When a group can show separate management, distinct exposure and independent instruction, the record should say so. Transparency should protect independence where it exists, not merely expose concentration where it does not.

That is the procedural bargain: the association does not presume bad faith, and the group does not ask the association to pretend that incorporation alone proves independent judgment. It is a small bargain, but without it the count is too easy to game and too easy to doubt.