Summary

  • “Transfer tax” is an economic warning, not a legal classification. This article does not decide whether any registry charge is a tax under national law. It asks when a mandatory transaction charge begins to resemble a levy on scarcity value rather than payment for a defined service.
  • Legitimate transfer costs include identity and authority verification, dispute screening, recipient qualification where policy requires it, inter-registry coordination, secure record change, reverse-DNS and routing-security updates, audit, correction and customer support. The institution should show how its fee maps to those activities.
  • Current regional models differ sharply. ARIN charges the source a fixed non-refundable request fee and the recipient a size-banded processing fee. APNIC charges a percentage of the annual fee applicable to the transferred resources, subject to stated exceptions. LACNIC uses two transfer-size bands and an initial non-refundable payment. RIPE NCC says transfers are free of charge. AFRINIC says existing-member transfers currently carry no direct transfer fee, while new recipients and changed holdings can trigger allocation and membership charges.
  • Resource size can affect some work, but address count is not self-proving evidence of cost. One clean /16 may require less review than a contested collection of /24s. A doubling fee schedule needs a published explanation of the risk, capacity or service cost that doubles with it.
  • The strongest model separates a baseline case fee, evidenced complexity charges, onboarding cost and continuing annual service. It publishes case volumes, staff hours, external verification spending, system cost, fee revenue, surplus treatment, waivers and appeals.
  • Fees should not follow contract price, estimated asset value or market appreciation. If a fee produces revenue beyond attributable transfer and shared-registry costs, the excess should be refunded, credited, used to reduce future charges or approved as an explicit member subsidy with a defined purpose.
  • Number Resource Society can standardise cost and performance reporting and make transfer evidence portable, reducing repeated verification. It must not finance itself by taking a percentage of IPv4 value or charging for artificial permission.

The invoice can conceal a constitutional choice

A seller and buyer agree to transfer an IPv4 block. Before the recognised record changes, the relevant registry checks the parties, reviews authority, confirms that the resource is eligible, applies recipient conditions, coordinates with another registry if necessary and updates services. Those tasks cost money. No serious transfer architecture assumes that skilled staff, secure systems and audit appear without funding.

The difficult question begins after that concession. How much should the parties pay, what event triggers payment, which party bears it, and why should the amount rise? A fee schedule answers more than an accounting question. It allocates the cost of the registry and can capture part of the value made possible by scarcity.

If every transfer requires a standard set of checks, a case fee is intelligible. If a disputed legacy chain requires external documents and extended review, a complexity charge may be intelligible. If a new recipient needs an account and continuing services, an onboarding and annual service fee may be intelligible. Each has a service entity.

When the charge grows mainly because a larger aggregate is changing hands, the justification becomes harder. The institution is not necessarily doing twice the work because the prefix contains four times as many addresses. It may face greater consequence if an error affects more users, but consequence is not the same as marginal processing cost. The claimed risk must be defined and measured.

This is where a service fee can take on the economic character of a transfer levy. It is mandatory for recognised completion, attached to a change in control and scaled by a proxy for asset value. The institution collects because it controls an essential record, not because the parties purchased an optional service in a competitive market.

That does not answer the legal question of whether the charge is a tax. Legal classification depends on jurisdiction, corporate powers, contracts and public law. The governance question can be answered without pretending to give that opinion: has the registry demonstrated that the amount corresponds to verification cost, service cost and auditable spending, or is it using a chokepoint to share in scarcity value?

Cost recovery and value extraction are different revenue theories

Cost recovery starts with activities. Staff verify identity, beneficial authority and resource history. Systems receive documents, authenticate users, preserve logs, exchange messages, update registration, maintain security and support corrections. Management allocates shared overhead. The fee follows the cost of performing those tasks efficiently.

Value extraction starts with the transaction. A larger or more valuable block can bear a larger charge, so the institution asks what the deal can sustain. The fee may be expressed as a percentage of value, a resource-size ladder or a high fixed toll. Its rationale is the institution's position in the transfer, not the cost of the service.

The difference matters even if both models raise the same revenue in one year. Cost recovery disciplines expenditure. If automation reduces staff time, the fee should fall or support demonstrably better service. If volume rises, fixed system cost is spread more widely. Surplus receives defined treatment.

Value extraction moves in the opposite direction. If IPv4 scarcity raises prices, the institution's take can rise even when its work is unchanged. If market entities become more efficient, the fee need not fall. Revenue becomes detached from the service and creates an incentive to preserve the gate.

Resource-size charges occupy an ambiguous middle. Size may correlate with risk, recipient holdings, service category, documentation or the impact of error. It also correlates imperfectly with market value. A size ladder is not automatically rent extraction, and a flat fee is not automatically fair. The institution must publish the causal link.

The burden of proof should increase as the schedule moves away from observable work. A fixed case charge needs a cost calculation. A block-size multiplier needs a cost calculation and evidence that size drives work or risk. A percentage of contract price would need an authority and purpose far beyond registry processing. No RIR should rely on the circular argument that the fee is legitimate because its board adopted the schedule.

Membership does not erase the distinction. A member organisation can approve cross-subsidies and collective goods, but affected parties must know what they are financing and have a meaningful voice. A transfer recipient compelled to join or use the incumbent record is not in the same position as a club member freely buying an optional dinner.

Start with the service that a transfer charge can legitimately buy

The first service is source identity. The institution must determine whether the person submitting the request represents the registered holder or its lawful successor. That can require account recovery, corporate records, historical contacts and fraud screening.

The second is authority over the resource. Registration may be stale, a corporation may have dissolved, a chain of mergers may be incomplete or a court-appointed representative may act. The reviewer must separate an authentic transfer from an attempt to appropriate space associated with an inactive organisation.

The third is eligibility. Holding periods, origin restrictions, disputes and policy conditions can prevent or delay a transfer. Where recipient need remains part of policy, staff also review the receiving organisation. These tasks are policy choices layered onto record accuracy, but they still consume resources while operative.

The fourth is synchronisation. In an inter-regional transfer, source and destination institutions must agree on the entity, sequence, status and completion. A safe change avoids two recognised holders or a gap in which neither side can administer essential services.

The fifth is technical administration. Registration, RDAP and related records change. Reverse-DNS responsibility and routing-security access may need coordinated transition. Logs must preserve who changed what and when.

The sixth is remedy. Parties need correction when staff make an error, when a document is misunderstood or when systems diverge. Appeals, escalation and incident response are part of a reliable service, not luxuries to be funded outside the fee story.

The seventh is common infrastructure. Secure portals, authentication, staff training, record storage, privacy controls, external audit and continuity planning benefit all cases. A reasonable share can be allocated to transfer service, but the formula should avoid charging the same overhead again through annual membership and transaction fees without explanation.

These activities form a defensible cost base. They also reveal why address count is an incomplete driver. The number of parties, prefixes, historic entities, jurisdictions, disputes, document cycles and registry handoffs may explain work better than the number of addresses.

ARIN combines a fixed source charge with a steep recipient ladder

ARIN's fee schedule effective 1 January 2026 states that the source pays a $500 non-refundable transfer request fee for an 8.3 or 8.4 request before evaluation begins. The payment does not guarantee approval. An 8.2 merger, acquisition or reorganisation request also carries a $500 charge to the requesting organisation.

For specified-recipient IPv4 transfers, the recipient pays a separate processing fee after approval and before allocation of the resources. The current ladder begins at $187.50 for a /24 or smaller if policy permits. It rises through aggregate-size bands to $3,000 for more than a /18 up to a /16, $48,000 for more than a /10 up to a /8 and $192,000 for an aggregate larger than a /6.

The pattern is clearly tied to aggregate address size, not the sale price reported by the parties. ARIN introduced the recipient charge in 2023, saying its existing processing fees did not fully cover the risk or effort associated with the transactions and that proportionality would improve equitable cost recovery.

That explanation identifies a legitimate objective but not the complete cost bridge. “Risk or effort” contains different things. Effort can be measured in staff time, document cycles, systems and external review. Risk can mean probability of error, consequence of error, fraud exposure, liability, service disruption or institutional reputation. Each has a different relationship to block size.

The current schedule invites a direct audit question: what cost or quantified risk changes at each prefix band? If large transfers require senior review, enhanced control, longer retention, additional security steps or more complex updates, publish that. If the process is materially the same, the ladder may be charging for transaction scale rather than service cost.

There is also a double-funding question. ARIN says it recovers registry-operation costs through annual Registration Services Plan fees based on aggregate holdings. A recipient can therefore pay a transfer processing fee and move into a higher annual category. Both may be justified: one pays for the event and one for continuing service. The cost model should show that the event is not financing continuing service already covered by the annual charge.

The fixed source fee has a clearer case unit but creates its own issue. It is non-refundable even if approval fails. Some work occurs immediately, so a filing charge can be reasonable. Accountability requires average pre-decision cost, reasons for refusal and whether unused portions should be refunded when a request is closed early.

APNIC indexes the transaction to a resource-based annual fee

APNIC's current member schedule charges a transfer fee equal to 20% of the annual fee applicable to the resources being transferred, per transaction. The recipient ordinarily pays. For transfers to another RIR, the source APNIC member pays. Stated exceptions include an initial IPv4 transfer into a member account holding no IP addresses, transfers to NIR members and membership changes to or from an NIR.

The annual fee itself is calculated from holdings using a formula based on the number of address bits, a base fee and a bit factor. For 2026, the member schedule lists an AUD 1,295 base and a 1.320 bit factor, with IPv4 and IPv6 assessed separately and the larger result determining the annual fee. The schedule also applies a 50% fee discount to account holders in least developed countries under its stated criteria.

The transfer fee therefore does not take a percentage of a private contract price. It takes a percentage of a resource-based institutional price. This is more predictable than ad valorem access to confidential sale value, and it can be administered without collecting commercial terms.

It still needs a cost explanation. The annual formula is designed to allocate membership cost across holdings. Using 20% of that formula for a one-time event assumes that the annual resource metric is also an appropriate measure of transfer work. The two questions are related but not identical.

APNIC's 2026 budget materials provide useful context. They forecast transfer-fee revenue as a small share of total revenue and describe the line as unpredictable. Revenue share does not prove cost correspondence, but publishing it is better than leaving the amount invisible. A full account would place transfer-service expenditure beside the revenue.

The direction rule also affects incidence. An outbound inter-RIR case charges the APNIC source, while an ordinary case charges the recipient. The same administrative formula can therefore enter seller or buyer economics depending on path. Contracts may reallocate the cost, but the invoice still affects negotiation and settlement.

The exceptions reveal policy choices. Waiving the transfer fee for a new account that will pay a full annual fee can avoid charging twice for entry. The NIR exception reflects regional institutional structure. The LDC discount makes distributional policy explicit. Each should appear in a subsidy account showing who finances the waiver and whether the intended group benefits.

APNIC's model is auditable because the formula is public. The missing link is activity: why 20%, and how closely does the resulting amount track the incremental service and risk of a transfer?

LACNIC uses two size bands and charges before justification is complete

LACNIC's current transfer guidance states that intra-regional and inter-regional IPv4 transfers carry an administrative fee based on block size. A block at least /24 and smaller than /19 carries a total charge of $1,000. A block /19 or larger carries a total charge of $1,500.

The recipient pays an initial $200 before the justification review. If the recipient cannot justify the transfer and approval is denied, that payment is not refunded. If approval is granted, the recipient pays the remainder of the applicable total.

Where one recipient obtains blocks from several offering organisations, LACNIC treats each transfer separately and charges each one. A recipient that changes membership category can also receive a supplementary invoice for the remaining renewal period. A new recipient pays the applicable category amount in addition to the administrative fee.

This model is less steep than a multi-band ladder, but it still uses resource size and transaction count as cost drivers. The /19 boundary creates a discontinuity. A transaction just above the threshold pays more even if the parties, documentation and review are otherwise identical. The institution should identify the additional work or exposure represented by that boundary.

Per-source charging can reflect real case work. Separate sellers require separate authority checks and resource histories. But it can also penalise a buyer forced to assemble fragmented supply. If one seller transfers four prefixes under one authority record, is that one case or four? If four sellers each transfer one /24, the work is plainly greater. The published rule should align the fee unit with the work unit.

The non-refundable initial payment is easier to defend if it covers an actual intake and justification review. LACNIC should publish the average cost of rejected or withdrawn cases, the point at which the amount is earned and any refund when the institution does not begin substantive work.

Membership adjustment again needs separation from transaction service. Receiving more addresses can place an organisation in a different continuing service category. That does not by itself justify a larger event fee. A transparent invoice should identify intake review, completed transfer, new membership and prorated annual-service effects as separate items.

LACNIC's statement that it does not intervene in commercial operations is important. The fee should be consistent with that boundary. Charging for verification and record change supports non-intervention. Charging according to the value the parties exchange would contradict it.

RIPE NCC shows that transfer processing need not carry a transaction toll

RIPE NCC's current transfer guidance says that transfers within its service region are free of charge. The 2026 charging scheme instead places the main burden on an annual contribution of EUR 1,800 per LIR account, plus specified charges for independent resources and ASNs, and a sign-up fee for new membership.

This does not mean transfer processing is costless. Staff, systems and controls are funded through the broader service model. Transfer users are cross-subsidised by annual contributions to the extent that their cases consume more than their share; quiet members may pay part of active users' costs.

The model proves a narrower point: a per-transfer fee is not technically necessary for a registry to process transfers. Costs can be pooled. Whether pooling is fair depends on member preferences, case concentration and transparency.

RIPE NCC's annual model contains a governance mechanism relevant to surplus. Members vote each year on the treatment of excess paid fees or a shortage through redistribution. That does not answer every cost-allocation question, but it gives members a visible decision over the gap between contributions and expenditure.

The absence of a transaction charge also changes market behaviour. Parties do not face a registry fee cliff based on prefix size or number of transfers. Fragmentation still creates broker, legal and operational cost, but the registry invoice does not add another marginal toll.

There can be hidden distributional effects. Members processing many transfers may benefit more from pooled funding. End users may depend on sponsoring arrangements. New membership and annual contributions can still be entry costs. “Free transfer” should therefore mean no separate transfer charge, not no institutional cost.

RIPE NCC has publicly consulted on alternative charging models, including proposed transfer charges, without making those proposals the 2026 scheme. That history is instructive. A fee proposal should be debated as a distributional choice, compared with annual funding and either adopted or rejected transparently. A consultation is not authority to invoice.

The RIPE model sets a useful benchmark for the other regions: if one mature registry processes transfers without a direct toll, a size-based charge elsewhere needs evidence specific to its service and funding structure, not a claim that such charging is inherent to registry operation.

AFRINIC separates existing-member transfers from entry and category cost

AFRINIC's published fees page says that transfers between two existing resource members currently carry no fee for IP or ASN transfer. The recipient's category is recalculated after the transfer, and a new annual fee may apply at the next renewal. All involved accounts must be in good standing.

Where an existing member transfers to a new organisation, the recipient applies for membership and resources. AFRINIC states that an allocation fee for the approved size and a membership fee for the recipient category apply. Its transfer guidance similarly says applicable fees must be paid and distinguishes existing and new recipients.

This creates a hybrid model. The event may be free between established members, while entry and increased holdings create size-related costs. The design can be defended if allocation fees cover the onboarding and review of a new recipient and annual categories cover continuing service.

The terminology requires caution. A transferred block is not newly produced inventory, yet the new recipient pays an “allocation” fee pegged to size. The institution should explain which transfer activities that charge finances and why an existing member receiving the same block may not incur it. If the difference is account creation and contract onboarding, those are better cost drivers than address count alone.

AFRINIC's public pages have been updated at different times and include language tied to earlier transfer conditions. Current implementation, especially for broader inter-regional arrangements, should be verified case by case before a party relies on a quoted amount. That uncertainty is itself a fee-accountability problem: parties should be able to calculate the institutional cost before signing a deal.

The existing-member waiver shows that no universal transaction levy is assumed. The category recalculation recognises continuing service cost. The new-recipient charge recognises onboarding. Those separations are useful if supported by current, consolidated publication and audited expenditure.

AFRINIC also demonstrates why comparisons must include more than a line called “transfer fee.” A region can report zero direct charge while imposing a size-based entry fee and a higher annual category. Another can charge both an event fee and annual service. The buyer needs the full institutional price over a defined period.

Five regimes reveal choices, not natural laws

The regional comparison contains at least five cost-allocation philosophies. ARIN combines a fixed source filing fee, a resource-size recipient ladder and holdings-based annual service. APNIC applies a percentage of a resource-based annual formula to each transfer. LACNIC uses two size bands, a non-refundable intake payment and category effects. RIPE NCC pools transfer cost into annual and related service charges. AFRINIC waives direct charges in some existing-member cases but applies onboarding, allocation and category fees in others.

No technical property of IPv4 uniqueness dictates one of these models. The same essential record change is financed differently because institutions made governance choices about incidence, membership and revenue.

The variation defeats two weak arguments. The first is that every transfer must pay a size-based charge because larger blocks inherently cost more to process. RIPE's model shows that direct charging is optional. The second is that zero direct charge is necessarily fair. Pooled funding can shift cost to members that rarely transfer.

The relevant comparison is total and functional. For a defined case, calculate source filing, recipient processing, membership entry, prorated annual change, sponsoring cost, external verification and any charge at the counterpart RIR. Then identify which service each item funds.

Inter-regional transfers complicate the picture. Source and destination institutions can each impose charges under their own rules. A party can face an outbound source fee, an inbound recipient fee, annual category changes and private facilitation. A cost-recovery claim should account for duplicated work and shared evidence.

Different fee units also alter behaviour. Per-case charges encourage consolidation. Per-source charges penalise fragmented acquisition. Size bands create thresholds. Holdings formulas affect whether resources sit in one account or several. Free transfers reduce direct friction but can increase the burden on annual members.

The goal is not uniform prices. Labour, currency, institutional design and member preferences differ. The goal is a common proof: cost driver, service entity, expenditure, revenue, surplus treatment and review.

Address count is a poor substitute for measured complexity

Consider one clean /16 held by an active company with current contacts, a clear agreement and no dispute. Now consider sixty-four /24s assembled from several dissolved entities, with stale records, different routing histories and inter-regional destinations. The second case can involve fewer addresses yet much more work.

Prefix count may matter because each entity needs validation and update. Counterparty count matters because authority is established separately. Historic chain matters because old records can require reconstruction. Inter-regional handoffs matter because two institutions coordinate. Dispute and court status matter because ordinary processing may need to pause.

Aggregate size may still affect consequence. An erroneous /8 record can disrupt more networks and create a larger correction event than an erroneous /24. Large transfers may warrant dual control, senior approval, enhanced monitoring or additional insurance. Those are testable costs.

The fee model should therefore use size only where the institution can describe the control triggered by size. A published matrix might show baseline intake, per-counterparty verification, per-prefix technical change, inter-registry coordination and enhanced-control tiers. Parties could see why an amount changes.

Complexity charging carries a discretion risk. Staff could classify a case as difficult and increase the bill after filing. The remedy is objective criteria, advance estimates, caps, reasons and appeal. A simple size ladder is predictable; the challenge is preserving predictability while making cost attribution more truthful.

A hybrid can work. Charge a modest baseline for standard cases. Include a defined quantity of review. Add published charges for exceptional external verification or unusually numerous counterparties, subject to consent and a cap. Fund common systems through annual service. Waive or credit duplicated checks where portable evidence is accepted.

The institution should not bill inefficiency. Repeated requests caused by unclear instructions, staff turnover or system error are not legitimate complexity costs. Time records must distinguish applicant-caused incompleteness from institutional rework.

Most importantly, estimated market value should never substitute for complexity. The registry does not create the buyer's business opportunity or the seller's scarcity gain. It maintains the record that makes the change legible.

A mandatory charge needs an activity-based account

The account begins with case volume. Publish completed, withdrawn, refused, pending, inter-regional, legacy, disputed and corporate-change cases by size and prefix-count band. Preserve confidentiality with minimum cell sizes.

Then publish labour. Show median and percentile staff hours for intake, source review, recipient review, coordination, technical update, remedy and appeal. Separate applicant-response days from staff time. Do not convert elapsed delay into labour cost.

Publish direct external spending. Identity services, document verification, legal review, secure signing, audit and specialist support should appear by category. A party should know when an unusual external cost is passed through.

Allocate systems carefully. Transfer portals, authentication, logging and automation have fixed and shared components. State the depreciation period, allocation key and share already funded by annual fees. Capital improvement should not be charged repeatedly after recovery.

Publish overhead. Management, premises, finance, insurance and governance can be allocated using a declared rule. “Overhead” should not become a container for unrelated expansion.

Place revenue beside cost. Show source fees, recipient fees, onboarding amounts, annual category increments attributable to transferred holdings, waivers, refunds, bad debt and currency effects. A transaction-fee line without the related expenditure cannot prove recovery.

Explain surplus. If revenue exceeds attributable cost, the institution can lower the next schedule, issue credits, fund a reserve capped at a stated level or ask members to approve a named cross-subsidy. The decision and beneficiary should be public.

Independent assurance should test allocation, not merely arithmetic. An auditor can confirm that reported expenditure occurred yet miss whether transfer users were charged for unrelated activities. The audit question is whether the cost driver is reasonable and consistently applied.

This account can be concise. It need not expose salaries by person or commercial cases. It must be detailed enough for a member or affected non-member to reproduce the fee logic.

Non-refundable does not mean unearned

Several regimes require payment before all substantive review is complete. An intake fee protects the institution from spending on speculative or incomplete requests and then chasing payment. That is a legitimate concern.

The word “non-refundable” should not end the analysis. A fee can be contractually non-refundable and still exceed the work performed in a case that closes immediately. Accountability asks when the institution earns each portion.

A staged model is clearer. A small filing amount covers account validation and initial triage. A source-review amount becomes due when that work begins. A recipient-review amount follows if required. Completion and technical-change cost is charged when the record is ready to move. Exceptional external cost requires notice.

If a case is refused because the applicant supplied false documents or ignored clear requests, retaining incurred amounts is defensible. If the institution discovers that the resource is ineligible through information it already held, charging the entire completion fee is harder to justify. If the institution cancels for its own operational reason, a refund should follow.

Published refusal reasons allow testing. A large share of early rejections may justify a separate intake service. It may also reveal poor guidance that the institution should fix rather than monetise.

Appeals should suspend disputed incremental charges where practicable. A party should receive an itemised reason and a route to challenge classification without losing the transaction. The appeal body should not depend financially on upholding the charge.

Refund policy also affects market access. A small operator may avoid filing if it can lose a meaningful amount before learning whether it qualifies. Pre-assessment tools and binding estimates can reduce that barrier.

The rule is simple: mandatory payment should vest as defined service is performed. Non-refundable status should protect real work, not transform uncertainty into revenue.

Cross-subsidy can be legitimate only when it is named

Registries provide collective goods beyond individual transfer cases: secure public records, routing-security infrastructure, policy development, training, research and continuity. Members may decide that transaction users should help fund some of them.

That is a political and membership choice, not pure cost recovery. It should be described as a levy for a named collective purpose, with a budget, beneficiary, duration and review. Hiding it inside “processing” prevents affected parties from assessing the trade.

A cross-subsidy may support small networks, least developed economies, security improvements or emergency continuity. Those can be valuable. The institution must show why transfer parties are the appropriate funding base rather than all members, service users or donors.

Incidence matters. The party invoiced may pass the amount through the contract. A source fee can reduce seller proceeds. A recipient fee can raise acquisition cost. A facilitator may bundle it. Ultimately, customers of the network may bear some cost. The institution should not assume that charging a large transfer means charging a wealthy speculator.

Size-based waivers and discounts also need funding disclosure. APNIC's LDC discount is explicit; the next question is who absorbs the reduction. A new-recipient waiver can avoid double payment, but the annual account should show the exchange.

Cross-subsidies should expire unless renewed. A security project has a delivery date. A reserve has a target. An emergency fund has a cap. Permanent percentage charges tend to outlive the reason offered for them.

The prohibition should be equally explicit: no fee should rise with private contract price or estimated IPv4 appreciation merely to capture the parties' gain. The registry's contribution is reliable coordination, not creation of scarcity value.

Named subsidy is accountable. Unnamed surplus is rent.

Fee design changes market structure

A high fixed fee weighs more heavily on small blocks and small operators. A steep size ladder weighs more heavily on large consolidations. A per-source charge makes fragmented supply expensive. A holdings-based annual fee changes the cost of retaining addresses after completion.

These effects can alter deal structure. Parties may combine transfers to avoid repeated filing fees, split them to remain below a band, use affiliates, choose a different regional path, lease instead of transfer or delay record changes. Not every response is abusive. Many are ordinary reactions to price.

Fee cliffs are especially likely to distort. A small change in aggregate size can move a case into a higher category even though service work barely changes. A smoother formula reduces cliffs but may make the charge look more like a value-linked percentage. The solution is not mathematical elegance alone; it is evidence about cost.

The burden can affect record quality. If updating the recognised holder is expensive, parties may keep contractual arrangements outside the top-level record, use leasing or postpone cleanup after corporate change. The registry then loses accuracy while collecting a fee intended to support it.

Inter-regional fees can fragment the global market. A source-region charge and destination-region charge may make one path more expensive than another. If the difference reflects actual coordination, it is a service cost. If it reflects cumulative scarcity tolls, it becomes a private trade barrier.

The institution should test behaviour after fee changes. Report case size, fragmentation, withdrawals, pre-filing inquiries, unregistered-change corrections and path substitution. Do not infer evasion from a drop in filings without examining prices and policy changes.

An impact assessment should precede large increases. Model standard cases by size and entity type, include annual consequences and show the share of total institutional cost recovered. After implementation, compare actual revenue and workload with the model.

A fee is not neutral because it appears in a schedule rather than a policy manual. It changes which transactions are worth recording.

A defensible fee schedule has ten controls

First, define the service entity. State whether the charge covers intake, source review, recipient review, inter-registry coordination, technical change, remedy or continuing service.

Second, define the cost driver. Use cases, parties, prefixes, document cycles, external checks, enhanced controls or another observable measure. Explain any use of aggregate address size.

Third, separate event cost from annual service. A recipient should see why both apply and which ongoing functions the annual category buys.

Fourth, publish activity and revenue annually. Volumes, hours, direct spending, allocated overhead, waivers, refunds and surplus should use stable definitions.

Fifth, cap exceptional charges. Parties need a binding estimate before they commit funds, with consent required for unusual external work.

Sixth, stage non-refundable amounts. Payment becomes earned as defined work occurs. Institution-caused cancellation and unperformed service produce a refund or credit.

Seventh, provide independent appeal. Classification, duplicated charges, eligibility-stage billing and refusal cost should be reviewable quickly.

Eighth, disclose cross-subsidy. Name the beneficiary, amount, duration, authority and renewal process. Do not hide it inside processing.

Ninth, prohibit value linkage. Contract price, broker estimate, market index and unrealised appreciation are not service-cost drivers.

Tenth, schedule expiry and recalibration. Automation, volume, inflation, security requirements and policy scope change. A formula should not become permanent merely because billing systems can apply it.

These controls do not require identical regional charges. They require every institution to explain its own amount in a common language of service, cost and mandate.

NRS should lower duplicated verification, not add another percentage

Number Resource Society can reduce transfer cost by making evidence portable. Verified organisation identity, authority, beneficial control, resource history, dispute state and prior checks can be expressed in signed, time-limited attestations. A destination service need not reconstruct every fact from the beginning.

Portable evidence does not mean blind acceptance. Institutions can verify signatures, freshness and scope, and can request additional evidence when a documented risk exists. The savings should appear in lower baseline fees or credits for accepted proof.

NRS can publish a comparative transfer-cost account using common definitions: baseline case, counterparty, prefix, inter-service handoff, enhanced review, completion, correction and appeal. Operators could compare price and performance without disclosing contract values.

It can also provide an open fee estimator. Parties enter service path, number of counterparties, prefix count and required controls, then receive a binding institutional estimate. Changes are versioned and explainable.

NRS must not charge a percentage of IPv4 contract price, resource appraisal or market index. It should not create certificates whose only purpose is to sell permission that the shared record could verify directly. Its own fees should pass the same activity-based test.

Competition and portability provide discipline. If another qualified service can verify the same evidence and maintain a compatible record, a provider cannot rely indefinitely on captivity. Open formats, correction rights and export make cost comparison real.

The positive model is deliberately narrow: fund accurate records, secure change, audit and remedy. Leave the scarcity value with the parties that bear the commercial risk. Reduce repeated work instead of charging every institution's full cost along the path.

NRS succeeds when a transfer becomes cheaper because proof travels. It fails if it becomes a sixth collector standing between agreement and recognition.

The institution may recover cost; it may not inherit the market's upside

IPv4 transfer services are not free to provide. Fraud, stale corporate records, contested authority and inter-regional coordination make serious verification necessary. Secure systems, skilled staff and remedies deserve stable funding.

The five regional models show that the funding method is a choice. One institution charges a source and a size-banded recipient. Another takes a percentage of a resource-based annual formula. Another uses two size bands. One pools transfer cost into annual contributions. Another distinguishes established-member transfers from new-recipient entry and category change.

That diversity should end the fiction that any particular fee follows naturally from the number of addresses. It should also end the fiction that a zero line means no one pays. Every model has incidence. Accountability requires seeing it.

The boundary is auditable. A legitimate charge maps to identity checks, authority review, eligibility work, synchronisation, technical update, correction, shared systems and reasonable overhead. The institution publishes cost, revenue and surplus. Parties receive an estimate, reasons, staged refunds and appeal.

A scarcity levy follows a different logic. It rises because the block is larger, the market is richer or the institution can hold completion. It funds unspecified activity, produces an unexplained surplus or preserves a gate whose cost should have fallen. Its invoice may say “processing,” but its incidence says rent.

No legal label is needed to reject that model. A private registry can have contractual power to invoice without having a democratic mandate to participate in every scarcity gain. Membership approval is stronger when affected parties can examine the cost account; it is weaker when the fee base includes recipients that cannot practically choose another recognised record.

The standard should be demanding and simple: charge for work performed, continuing service delivered and expenditure that can be audited. Name any subsidy. Return or credit excess. Never calculate the institution's share from the private value transferred.

The market price belongs to the bargain between holder and operator. The registry's fee belongs to the service of making the record true. Confusing those two turns a bookkeeper into a revenue authority without ever answering who authorised the levy.

Sources