RIPE NCC and the economics of fee incidence and regressivity

Summary

• RIPE NCC fee incidence begins after the invoice is paid: the member named on the invoice is not necessarily the operator, customer, shareholder or end user that ultimately bears the economic burden.
• A per-LIR account charge can be formally equal while economically regressive, because the same euro obligation absorbs a larger share of cash flow, staff time and purchasing power for smaller networks and poorer markets.
• The RIPE NCC region is too unequal for nominal sameness to be neutral; a fee that is minor for a large Western carrier can be a material fixed cost for an access network in a lower-income or high-friction market.
• Account-based charging changes incentives around multiple LIRs, consolidation, transfer timing, legacy-resource agreements, sponsoring relationships and the choice between holding addresses directly or through intermediaries.
• IPv4-rich holders and small access networks face different incidence: the former may treat fees as a carrying cost on scarce capital, while the latter experience them as a fixed overhead on thin service revenue.
• Payment channels, exchange-rate exposure, sanctions screening, billing-contact accuracy, documentation work and suspended request processing are part of incidence because they decide who can absorb delay without losing operational options.
• A fair charging debate should replace accidental cross-subsidy with explicit incidence reporting, transparent tradeoffs and a narrow account of what costs belong in the compulsory registry relationship.

Incidence starts after the invoice is paid

Every fiscal system has two payers. One is legal and visible: the person or organisation that receives the bill. The other is economic and often hidden: the party that ultimately bears the cost after contracts, prices, wages, investment plans, payment frictions and market power have done their work. The distinction is familiar in public finance. A payroll tax may be remitted by an employer but partly borne by workers through lower wages. A sales tax may be collected by a retailer but partly borne by customers and partly by the retailer's margin. A regulated network fee may be invoiced upstream and paid downstream through higher access prices, lower investment or reduced service quality.

That is the right starting point for RIPE NCC fees. The invoice may name a member, a Local Internet Registry account, a sponsoring LIR or a legacy resource holder. The burden does not stop there. It moves through the operator's business model. It may be charged to a retail access division, included in enterprise connectivity pricing, embedded in wholesale transit, absorbed by shareholders, passed to hosting customers, deducted from a rural expansion budget, capitalised into IPv4 leasing economics or treated as the cost of keeping an address portfolio clean enough for transfer. In poorer markets it may also move through foreign-exchange costs, bank-transfer delay, documentary overhead and the lost value of staff time.

RIPE NCC's Charging Scheme 2026 is clear about the legal incidence. The model is based on an annual contribution per LIR account, with additional charges for independent and legacy Internet number resources and ASN assignments. For 2026, the annual contribution remains EUR 1,800 per LIR account, the sign-up fee is EUR 1,000, the separate charge for certain independent resource assignments is EUR 75 and the ASN assignment charge is EUR 50. The Billing Procedure 2026 adds further mechanics: invoices are issued for each LIR account, members must pay the annual contribution for all LIR accounts before transfers can take place, and payment obligations are in euros.

Those facts are only the beginning. They describe the statutory shell of the charge. Incidence asks what happens next. A large telecommunications group with millions of customers, in-house counsel and a treasury desk experiences EUR 1,800 differently from a small fixed-wireless provider, a small hosting firm, a regional ISP in a lower-income market or a network that holds a modest amount of address space as its main scarce asset. The legal invoice can be the same while the economic invoice is radically different.

This is why the fee debate should not be reduced to a narrow argument about whether a charge looks high or low in Amsterdam terms. The first question is distributional. Who can spread the cost? Who must absorb it as a fixed overhead? Who can pass it on? Who loses optionality because a small charge arrives with process, timing and compliance conditions? And who pays indirectly because the registry relationship is not a discretionary subscription but a recognised route into globally unique Internet number resources?

A flat fee is not flat in economic life

Flatness is a legal design, not an economic result. A flat charge treats the account as the unit of equality. Each account owes the same amount, subject to defined add-ons. That is simple, auditable and administratively attractive. It also has a familiar public-finance problem: the same nominal fee can be regressive when it consumes a larger share of income, margin or operating capacity for smaller and poorer payers.

Regressivity is not the same as unfairness in every case. Some flat charges are defensible because the service cost is genuinely flat, because the charge buys equal access to a common facility or because variable charging would create worse distortions. But the burden must be recognised. A road toll of EUR 10 is mathematically identical for a luxury car and a delivery van; it is economically different if one trip is discretionary and the other is needed to earn a day's revenue. The same principle applies to a registry account. The charge may be identical per account. The capacity to absorb it is not.

The RIPE NCC account is not an ordinary association subscription. It is attached to a registration relationship for IP addresses, ASNs, reverse DNS, RPKI, transfers, database accuracy, billing status and continuing administrative standing. The RIPE NCC service list describes a registry function that assigns and allocates Internet number resources, maintains contractual information, processes transfers, reviews registry data and provides certification and database services. Members can dislike the price, complain, vote or reorganise. They cannot replace the recognised registry relationship with a cheaper competitor in the same region.

That lack of substitutability is what makes the flat charge more like a compulsory infrastructure levy than a market price. In a normal market, a small firm facing a high subscription can choose another supplier, cut the service, self-provide or negotiate a different package. A network needing recognised registry standing cannot treat the relationship that casually. It may reduce the number of accounts, use a sponsor, delay a transfer, consolidate holdings, pass costs to customers or exit a business line. But the underlying relationship is not optional in the way a conference ticket or software licence is optional.

The flat charge therefore has a double character. It is administratively easy, and that matters. It reduces gaming, simplifies billing and avoids turning every annual fee into a dispute about size, income, address count, use, country, revenue or hardship. At the same time, it loads a fixed cost onto unequal firms. Once the charge is fixed, the incidence depends on scale. The larger the denominator - customers, revenue, gross margin, address portfolio value, enterprise accounts, wholesale contracts - the smaller the burden. The smaller the denominator, the more the same invoice behaves like a tax on entry, survival and administrative capacity.

There is also a difference between average institutional cost and marginal member burden. A registry may reasonably say that a simple account fee is the least costly way to fund common services. That may be true from the institution's accounting perspective. Yet the member's burden is not the average cost of running the institution; it is the marginal pressure the bill places on the member's next decision. For a large incumbent, the margin may be whether to allocate a tiny overhead line to network operations or corporate administration. For a small entrant, the margin may be whether to buy another upstream connection, keep a spare router, pay a consultant for RPKI configuration or delay direct membership for another year. The same accounting line therefore lands on different business margins.

The RIPE NCC region is too unequal for nominal equality to be neutral

The RIPE NCC service region is not a single-income economy. The organisation states that it serves more than 75 countries across Europe, the Middle East and parts of Central Asia, with more than 20,000 organisations acting as LIRs in their own countries. That region contains global financial centres, wealthy small states, large mature broadband markets, oil economies, post-Soviet transition economies, lower-income states, conflict-affected territories, sanction-exposed jurisdictions and markets where a few thousand euros can be a material operational decision.

This matters because the fee is denominated in euros and because the registry's administrative centre of gravity sits in a high-cost European environment. A euro-based charge is simple for the institution and for many members. It is also a distributional choice. Operators earning revenue in weaker currencies, working through fragile banking systems or serving customers with lower purchasing power face a different real burden. For them, the annual fee is not just EUR 1,800. It is EUR 1,800 plus exchange-rate exposure, bank fees, treasury inconvenience, documentary work, compliance checks and the possibility that payment delay affects operational requests.

Nominal equality can look attractive in a region this diverse because any alternative appears politically difficult. A country-income adjustment would be controversial. A revenue-based charge would require disclosure, verification and rules for multinational operators. An address-count charge could penalise efficient address use or distort IPv4 holding decisions. A service-use charge could make essential registry functions feel transactional. A means-tested relief scheme could be gamed and would require the registry to become a judge of members' hardship. These are real objections. They do not make the burden disappear.

The public-finance point is that equality of rule and equality of burden are different concepts. A fee can be non-discriminatory in the legal sense and regressive in economic effect. In a homogeneous region, that gap might be tolerable. In the RIPE NCC region, it is too large to ignore. The same account charge crosses Iceland, Germany, the United Kingdom, the Netherlands and Switzerland, but also markets where revenue per user, bank reliability, legal capacity and staff availability look very different. It crosses high-margin enterprise networks and small access networks that may be trying to build redundancy, customer support and compliance capacity on modest cash flow.

That is where the poverty penalty enters. The penalty is not only that poorer members have less money. It is that they often need to spend more administrative effort per euro of fee paid. The same email invoice can trigger more internal work. The same transfer prerequisite can consume more scarce management time. The same payment deadline can matter more when international banking is slower or when currency conversion is volatile. The same registry request delay can matter more when a small network has fewer spare addresses, fewer engineers and less legal support. Poverty is not merely low income. It is lower resilience against fixed institutional friction.

The regional dimension also changes the meaning of solidarity. A single fee across a broad service region can be defended as a way to keep the registry common rather than fragmented into national price schedules. But if that common fee finances activity whose benefits are concentrated among larger, richer or more internationally connected members, the solidarity claim weakens. The question is not whether lower-income markets should receive charity. It is whether they should be asked to finance a regional institutional package whose optional benefits they use less and whose fixed costs they feel more sharply. A fair common region needs common infrastructure. It does not automatically need common financing for every surrounding activity.

Account structure turns fees into incentives

An account-based charging model does not merely collect money; it shapes behaviour. When the unit of charge is the LIR account, members have an incentive to ask how many accounts they should hold, whether multiple accounts still make sense, whether legacy resources should be held directly or through sponsorship, whether address portfolios should be consolidated, and whether a transfer should occur before or after a billing date. These are not abuses of the model. They are predictable responses to the unit of assessment.

The Billing Procedure 2026 says invoices are issued for each LIR account and that existing members are invoiced for the full year for each account they hold on 1 January 2026. It also says that if a member closes one or more LIR accounts during 2026, the contribution for all accounts must still be paid in full unless a valid closure request was submitted before the end of 2025. It adds that members must pay the annual contribution for all LIR accounts before a transfer can take place. These rules make sense from the institution's perspective: they prevent free-riding, avoid mid-year administrative complexity and protect collectability. They also affect timing, consolidation and liquidity.

For a large member, the account question may be a portfolio-management issue. Multiple LIR accounts may reflect acquisitions, historical structure, internal divisions, address-management convenience or transfer strategy. Paying another EUR 1,800 can be a small carrying cost compared with the operational convenience or option value of keeping the account. For a smaller member, the same account may be a larger budget line. Maintaining a second LIR account may be hard to justify unless it carries clear address, transfer or organisational value.

The fee therefore creates an incentive to economise on accounts. That can be efficient if it cleans up unnecessary structures. It can be harmful if it pushes members to consolidate in ways that reduce transparency, merge operationally distinct networks, increase dependence on sponsors or make future transfers less flexible. It can also encourage timing behaviour around year-end account status, because the billing date matters. Again, the point is not that members are gaming the system. The point is that fiscal rules define margins, and margins drive behaviour.

The same logic applies to sponsoring relationships. A small end user with independent resources may not become a member directly. It may rely on a sponsoring LIR and ultimately bear the charge through the sponsor's price. The legal fee may sit at the sponsoring relationship, but economic incidence can move to the end user as a line item, a higher management fee or reduced service responsiveness. If the sponsor market is competitive, pass-through may be limited. If switching sponsors is cumbersome or if the end user values continuity, the sponsor may recover more of the charge. The registry fee becomes part of a small private market in administrative standing.

IPv4-rich holders and small access networks face different burdens

The same registry charge can attach to very different assets. For an IPv4-rich holder, the annual account fee can resemble a carrying cost on scarce digital capital. For a small access network with limited addresses and thin retail margins, it can resemble a fixed overhead on connectivity service. The distinction is central to incidence because it changes who can absorb the cost and what behaviour the fee encourages.

An address-rich holder may hold legacy IPv4 space, acquired space, transferred space or historically accumulated allocations. Even when the legal nature of the holding is not the same as ordinary property, the economic characteristics of IPv4 scarcity are evident. Addresses can support customers, hosting, leasing, mergers, financing discussions, network continuity and transfer value. Against that denominator, EUR 1,800 may be modest. It is part of the cost of keeping the recognition relationship in good order. The fee may be capitalised into the expected value of the address portfolio, much as maintenance, custody or title-cleanliness costs affect other scarce assets.

A small access network sees another arithmetic. Its scarce asset is not only address space; it is local trust, field crews, backhaul contracts, customer support, cash flow and the ability to keep churn under control. If it serves a lower-income market, average revenue per user may be low. If it serves remote customers, operating costs may be high. If it has little bargaining power with upstream providers, supplier costs may be sticky. The registry fee then competes with router replacement, tower maintenance, customer-installation subsidies, security work and staff training. A charge that is small relative to an address portfolio can be large relative to a small access network's discretionary budget.

This difference matters for policy because "resource holder" is not a homogeneous category. A fee that looks neutral across LIR accounts may be light for an address-rich holder and heavy for an access network. Conversely, a pure address-count fee might relieve small access networks with little space but burden holders whose space is valuable even if their current operating revenue is small. There is no frictionless answer. But the tradeoff must be explicit. Otherwise the charging debate treats unlike cases as though they were morally and economically identical.

Legacy treatment intensifies the point. The 2026 charging scheme says the fee for legacy resource holders that enter a direct agreement with RIPE NCC is identical to the annual fee per LIR account, with no sign-up fee for direct legacy agreements and no sign-up fee for legacy holders that become members without requesting additional resources. That is administratively simple and arguably fair as a service relationship. Yet incidence differs between a large legacy holder that treats the fee as address-title hygiene and a small historical holder that sees it as a cost of maintaining recognisable standing around a resource that may be valuable but not liquid without further legal, registry and commercial work.

The policy question is therefore not whether IPv4-rich holders should pay more or less as a slogan. It is whether the chosen unit of charge accurately matches cost, benefits, risk and ability to pay. Account-based charging scores well on simplicity. It scores less well on distributional precision. A serious fee debate should acknowledge both scores rather than pretending that one metric settles the matter.

The poverty penalty hides inside administrative sameness

Poorer operators rarely pay only the headline fee. They pay in smaller margins, weaker currencies, thinner administrative teams, slower banking, fewer lawyers, less familiarity with procedures and a higher cost of mistakes. A large member can assign billing, legal, compliance and registry work to specialised staff. A small member may assign all of it to the founder, the network engineer or the finance officer who also handles customer support and supplier negotiations. The euro amount is visible; the opportunity cost is not.

Administrative sameness can therefore conceal unequal burden. The same billing contact requirement is routine for a large operator and risky for a small one if staff turnover is high. The same 30-day payment window is routine for a member with treasury operations and more serious for a member that depends on international wires, local banking holidays or manual approvals. The same requirement to pay all LIR accounts before a transfer can proceed is a predictable rule for a well-capitalised group and a liquidity trap for a smaller firm trying to complete a transfer to raise cash, restructure resources or satisfy a customer.

The poverty penalty also changes the cost of attention. Public-finance debates often treat money as the scarce variable. In small networks, attention is equally scarce. Reading a billing procedure, checking whether a resource is charged as an independent assignment, reconciling invoices, planning year-end account closure, confirming payment references and ensuring that a request is not blocked can consume managerial time that would otherwise be spent on service quality or sales. That time has a higher marginal value when the firm is small.

This is where registry fees differ from many ordinary supplier costs. A small operator can often negotiate with a vendor, delay a purchase, choose a cheaper router, defer travel or switch accountants. It cannot easily ignore the registry standing attached to number resources. That dependence gives administrative friction a quasi-tax character. It is imposed by the need to remain properly recognised in the system rather than by a purely optional purchase.

There is a temptation to dismiss this as sentimentality: every institution has forms, invoices and deadlines. But incidence analysis is not sentiment. It is the discipline of following costs to their final bearer. If poorer and smaller members bear a higher non-monetary burden per euro of invoice, then the economic charge is larger than the accounting charge. A fair model may still choose a simple flat fee. But it should not pretend the flat fee is the whole cost.

Payment friction is part of fee incidence

Payment mechanics are often treated as housekeeping. They are not. Payment friction can decide who bears the burden of the fee and how large the burden becomes. The RIPE NCC billing procedure requires payment obligations in euros, says any exchange difference remains payable or receivable, asks members to include the LIR number and invoice number as a reference, prefers one transaction, does not accept cheques, and warns members to account for the time required for international bank transfers. It also says invoices need to be paid within 30 days and that no new or ongoing requests will be processed if payment is not received within 60 days.

For large members, these are ordinary controls. For smaller and cross-border members, each is an incidence channel. A euro obligation means currency risk. A payment reference requirement means operational risk: a misallocated payment can create delay even when the member has paid. International transfer timing means liquidity risk. Online payment variation by country means unequal access to convenient rails. The 60-day request-processing consequence means the payment system can become an operational bottleneck, not merely a finance issue.

The same is true of sanctions and high-risk-country complications. The RIPE NCC's 2026 Activity Plan and Budget notes conservative income assumptions partly because of consolidation of LIRs and income it is not able to collect from certain high-risk countries and members awaiting clearance from potential sanctions matches. The exact treatment of particular cases belongs to compliance rules, not to fee theory. But the incidence point is plain: collection risk and compliance delay do not affect all members equally. They cluster around geography, banking access, state risk, documentation and political exposure.

Payment friction also affects transfer timing. If annual contributions for all LIR accounts must be paid before resources can be transferred, a member with multiple accounts, a disputed invoice, a delayed international wire or sanctions-clearance uncertainty may face a more expensive transfer process. The explicit fee is unchanged. The economic cost rises through time, uncertainty and foregone optionality. In thin markets, timing can be the difference between completing a transaction and losing it.

The easiest reform is not necessarily to reduce the nominal fee. It may be to reduce friction around payment certainty: clearer status dashboards, faster allocation of payments, more robust local payment options, better pre-transfer billing checks, clearer communications for high-risk jurisdictions and lower probability that a small clerical issue turns into an operational hold. These improvements would not solve regressivity. They would reduce a hidden component of it.

Cross-subsidy should be explicit, not accidental

Every common charging scheme contains cross-subsidy. The issue is not whether cross-subsidy exists. It is whether it is intentional, legible and defensible. In a flat account model, members whose service cost is lower than the average subsidise members whose service cost is higher. Members that rarely use training, events, complex support or policy infrastructure may help pay for members that use more of those services. Larger members may subsidise smaller members in some dimensions because the fee is not proportional to scale. Smaller members may subsidise larger members in other dimensions because a fixed fee lets scale-rich networks spread the same charge across a much larger revenue base.

That ambiguity is the problem. A cross-subsidy that is explicitly justified as funding the common uniqueness layer, the registry database, RPKI, reverse DNS, security, service resilience and basic member support can be debated. A cross-subsidy that accidentally bundles events, outreach, measurement, training, public-policy activity, regional engagement and institutional overhead into a single compulsory charge is harder to evaluate. Members are left arguing over the total fee rather than the incidence of each function.

RIPE NCC's 2026 budget places the total planned income around EUR 41.140 million and costs around EUR 41.125 million, with activity areas including the Registry, Information Services, External Engagement and Community, and Organisational Sustainability. Some of those activities are core to the registry relationship. Others may be valuable public goods for the region. Others are institution-maintenance costs. The incidence question is not whether they are useful. It is whether they should be financed through the same compulsory account charge, and if so who bears the burden.

A useful public-finance distinction is between benefit taxation and ability-to-pay taxation. Benefit taxation charges users according to the benefit or cost they impose. Ability-to-pay taxation charges stronger payers more because they can bear it. A flat LIR fee does neither precisely. It is not clearly cost-based, because the cost of serving members differs. It is not clearly ability-based, because large and small members pay the same per account. It is a hybrid of administrative simplicity, common-cost recovery and political compromise.

Hybrids can be legitimate. But they should not be hidden. If large operators are being asked to subsidise the regional public-good functions of the registry, say so. If small operators are being asked to help fund services they rarely use because those services strengthen the overall registry ecosystem, say so. If address-rich holders are receiving a valuable recognition layer at a low account-level price, say so. If low-income-market members are being asked to pay the same euro fee because any relief scheme would be too complex or too vulnerable to manipulation, say so. Cross-subsidy is not automatically wrong. Accidental cross-subsidy is weak governance because nobody can tell whether it matches the institution's purpose.

The same principle applies to reserves, but only as a separate question. A reserve can protect continuity, smooth shocks and prevent sudden special assessments. It can also soften budget pressure if members cannot see which activities are truly necessary and which are carried forward because the common fee can finance them. Reserve adequacy is not the same as incidence. A registry can have prudent reserves and still impose a regressive charge. It can have thin reserves and still distribute current fees unfairly. Treating reserves as the whole fee debate blurs the sharper question: which members and downstream users bear the current cost of the model, and are they the ones most able to carry it?

Fee design can encourage consolidation, fragmentation or pass-through

Fees are not neutral once firms can reorganise around them. An account-based fee can encourage consolidation when multiple accounts no longer justify their cost. It can discourage fragmentation by making each additional LIR account expensive enough to require a business reason. It can also encourage pass-through, because members may treat the annual fee and additional resource charges as costs to be recovered from customers, sponsored resource holders or internal divisions.

Consolidation can be healthy. It can remove dormant accounts, reduce administrative overhead, simplify data and lower the number of invoices. But consolidation also has side effects. If smaller operators prefer to avoid direct accounts and rely on sponsors, registry standing becomes more intermediated. If acquired networks are folded into larger account structures, their distinct operational needs may become less visible. If multiple regional networks consolidate address administration inside one corporate function, local managers may lose direct understanding of registry obligations. Administrative efficiency can trade off against operational transparency.

Fragmentation can also be rational. A group may maintain separate LIR accounts to preserve address-management boundaries, isolate acquisition histories, support transfer optionality, separate regulated businesses, manage risk or keep clean records for future transactions. The annual fee then becomes the price of optionality. That price is easier for large groups to pay than for small ones. If the fee is too low, unnecessary fragmentation may persist. If it is too high, useful boundaries may disappear. The correct answer depends on evidence about how accounts are used, not on a general preference for fewer or more accounts.

Pass-through is the least visible channel. A hosting provider may recover registry costs through server pricing. A retail ISP may add a small amount to monthly service plans. A sponsor may charge end users for independent-resource management. A transfer intermediary may build billing standing and payment-timing risk into transaction costs. A corporate network may allocate fees to business units. In each case the legal invoice is only a waypoint. The final burden may land on customers, smaller counterparties or internal projects with less bargaining power.

This is why a charging scheme can produce industrial-organisation effects without intending to. It can alter the relative advantage of large integrated operators versus smaller direct members. It can change the economics of sponsorship. It can shift bargaining power between resource holders and service providers. It can influence whether IPv4 is held, sold, leased, consolidated or kept dormant. And it can affect whether new entrants enter directly, enter through intermediaries or avoid the market altogether.

Members price uncertainty as well as euros

The economic cost of a registry fee includes uncertainty around the fee. Members do not simply ask what they owe today. They ask what the model implies for next year, whether additional resource charges will grow, whether the account unit will remain stable, whether service scope will expand, whether compliance costs will rise, and whether the charging scheme can be forecast with enough confidence to support contracts and investment. A predictable fee is less costly than an unpredictable fee of the same amount.

Uncertainty matters more when the registry relationship is irreplaceable. A software vendor can change prices and lose customers. A registry can change charges through member-approved procedures, but members cannot simply buy equivalent regional recognition elsewhere. That makes fee predictability part of the economic contract, even if the legal contract says the scheme can be changed. A small operator deciding whether to become a direct member, sponsor resources, buy IPv4, lease addresses or rely on upstream arrangements must forecast registry charges. If it cannot, it discounts the value of direct standing.

The 2026 scheme is stable relative to 2025 in its headline fees. That stability is valuable. But the future debate is not only about the next nominal amount. It is about the basis of charging. Per-account, per-resource, per-ASN, address-weighted, category-weighted, income-sensitive, service-use and hybrid models distribute uncertainty differently. A model that tracks address holdings may make small access networks happy but worry address-rich holders. A model that tracks accounts may be simple but burden poorer small members. A model that expands per-resource charges may affect end users and sponsoring markets. A model that uses revenue or geography may raise verification, privacy and political problems.

Uncertainty itself is regressive when the capacity to manage it is unequal. Large members can run scenarios, lobby, vote, read consultation material, forecast budget lines and absorb mistakes. Small members may only discover the practical effect when the invoice arrives or when a transfer is held. This is why incidence reporting matters. Members need to see not only the expected institutional revenue but the distribution of burden across account sizes, resource profiles, regions and member types.

In public finance, a tax authority that changes rates without distributional analysis invites distrust. A registry charging scheme should be held to a similar standard, not because it is a state, but because its relationship to members is compulsory in economic substance. A small number of clear incidence tables would lower the uncertainty premium. They would not settle the politics. They would make the politics honest.

The customer eventually meets the registry bill

No member-funded registry charge stops at the member boundary. Some of it is absorbed by the member's owners or donors. Some is paid through lower wages, lower investment or lower surplus. Some is passed to customers. The ability to pass through depends on competition, demand elasticity, customer type and whether the charge can be itemised. But over time, upstream fixed costs tend to appear in downstream prices or service quality.

For a large broadband provider, the annual LIR fee is too small to be visible on a retail bill. It disappears into overhead. For a small network, it may not be visible either, but it can still affect choices at the margin. A small provider deciding whether to extend service to a village, discount installations, upgrade a router, buy spare addresses, maintain a second upstream or hire a support engineer works inside a tight budget. The registry charge may not determine the decision alone. It adds to the fixed-cost stack that makes marginal expansion harder.

The same is true in hosting and enterprise services. Address costs, registry fees, abuse-contact work, RPKI management, transfer due diligence, reverse-DNS maintenance and customer documentation all become part of the cost of offering stable IP-based services. Customers rarely know which portion reflects the registry layer. They see a server price, a managed-service fee, a setup charge or a requirement to pay for address management. Incidence has occurred even if no one names it.

This does not mean that every fee cut would become a customer price cut. In competitive markets, some savings would likely pass through. In less competitive or highly differentiated markets, some would remain as margin. In constrained networks, some would fund investment. In public or nonprofit networks, some might preserve service. The distribution would vary. But the burden still travels. A registry fee is not sealed inside the member organisation.

This is why the "every end user pays" claim should be treated carefully. It is too broad if it implies a direct, measurable line from the RIPE NCC invoice to each household. It is accurate in the more important economic sense: essential upstream overhead becomes part of the cost base of connectivity and digital service. The more fixed and unavoidable that overhead is, the more it matters for weaker operators and lower-income customers.

The customer incidence channel is also why fee debates should distinguish between large-member annoyance and small-market harm. A large member may complain because it dislikes cross-subsidy or institutional scope. A small member may complain because the fee affects operating choices. A downstream customer may never complain because the cost is invisible. Public-finance analysis exists precisely to reveal those invisible burdens.

Charging schemes are policy instruments even when they look administrative

Institutions often prefer to describe charging schemes as administrative finance. That is understandable. Fees pay bills, budgets need revenue and members need predictable invoices. But every charging scheme is also a policy instrument. It defines the unit of membership, the price of account optionality, the cost of holding resources directly, the relative attractiveness of sponsorship, the burden on small members and the implied cross-subsidy among services.

The policy content can be seen in small details. Charging per LIR account makes the account the core fiscal unit. Charging EUR 75 for independent resource assignments and EUR 50 for ASN assignments attaches a small annual cost to certain resource relationships. Charging direct legacy resource holders the same annual service fee as an LIR account treats direct standing as a service relationship of comparable institutional weight. Requiring payment before transfers links fiscal compliance to resource liquidity. Denying pro rata refunds for mid-year closure protects budget certainty while affecting exit timing. Requiring euro payment places currency risk outside the registry.

None of these choices is irrational. Each can be defended. But they are not neutral. They distribute costs, risks and bargaining power. The question is whether members and the public can see the distribution clearly enough to judge it. A policy instrument that pretends to be mere administration will be judged on narrow accounting grounds. A policy instrument that admits its distributional effects can be judged on fairness, efficiency and resilience.

There is a further industrial-organisation point. The RIPE NCC registry relationship exists in a market where IPv4 scarcity, address transfers, hosting demand, cloud growth, compliance costs and network consolidation already shape incentives. A fee model placed into that market affects behaviour at the margin. It may help keep the registry funded and stable. It may also raise entry barriers for the smallest operators, increase the attractiveness of intermediated resource holding, or make large address portfolios cheaper to carry relative to small operating networks.

The right standard is not zero distortion. No fee model achieves that. The right standard is conscious distortion: knowing what the scheme encourages and deciding whether those incentives match the registry's narrow public function. If the goal is accurate records, secure routing support, predictable transfers and broad member participation, the scheme should be tested against those outcomes rather than against revenue sufficiency alone.

Better incidence reporting would change the debate

The RIPE NCC already publishes charging schemes, budgets, activity plans and annual reports. The missing layer is incidence reporting: a distributional account of who bears the fee under different member profiles. This would not require exposing commercially sensitive data. It would require grouping members and resources in ways that make the burden visible.

A useful incidence report would show the distribution of annual charges per member, per LIR account, per independent resource count, per ASN count and per legacy-resource relationship. It would separate members with one account from members with many. It would show how many members pay only the base fee, how many pay material additional charges, and how those charges vary by country group or income band without naming individual members. It would estimate the effect of proposed charging options on small single-account members, multi-account groups, address-rich holders, sponsors, legacy direct agreements and resource-light organisations.

It would also include payment-friction indicators. How many members pay late? How many payments require manual allocation? How many members are affected by currency, banking or sanctions-clearance issues? How many requests are delayed because payment is not received within the relevant window? How many transfer processes are blocked by unpaid accounts? These figures would turn anecdote into measurement.

Most important, incidence reporting would separate institutional cost recovery from distributional choice. Members could see how much revenue the registry needs, how much each model raises and who bears the marginal change. The debate would move from "the fee is too high" or "the fee is equal" to a more serious set of questions: equal relative to what, regressive against which denominator, and justified by which common service?

The report should also model uncertainty. If a proposed scheme is expected to increase charges for certain resource profiles over three years, members should see that path. If account consolidation is expected, the revenue and incidence effects should be shown. If high-risk-country collections or sanctions screening reduce collectability, the impact on other members should be made visible in aggregate form. Hidden shortfalls are themselves an incidence issue because they are eventually covered by reserves, future fees or reduced service.

One practical table would do more than pages of abstract argument. It could compare, for example, a single-account small ISP, a multi-account national carrier, a hosting firm with many sponsored resources, a legacy holder with a direct agreement, a resource-light enterprise member and an address-rich transfer-market participant. For each profile, the table could show the legal fee, likely administrative time, payment-risk exposure, transfer-timing effect and plausible pass-through channel. The figures would be illustrative rather than determinative. Their purpose would be to make the burden visible enough that members cannot confuse formal sameness with economic neutrality.

This would not make the charging vote easy. It would make it more disciplined. Members would still disagree about whether simplicity, progressivity, cost causation, resource-based charging or low administrative overhead should dominate. But they would disagree with a shared map of burden rather than with competing intuitions.

Fairness requires transparent tradeoffs, not sentimental exemptions

A fair fee model is not the one that produces the most sympathetic story. It is the one that states its tradeoffs honestly and aligns them with the registry's purpose. In the RIPE NCC case, the purpose is not to run regional welfare policy, redistribute income across countries or subsidise every small operator. Nor is it to ignore poverty, scale and entry barriers because the invoice is formally equal. The purpose is to maintain a trustworthy registry relationship for number resources while distributing compulsory costs in a way that is efficient, predictable and defensible.

That means sentimental exemptions are a poor substitute for design. A hardship waiver may help a few members, but it can be arbitrary and administratively costly. A country-income discount may look fair, but it can misclassify multinational firms and punish members in richer countries that serve poor customers. A revenue-based charge may track ability to pay, but it creates reporting burdens and may be intrusive. An address-based charge may track scarcity value, but it can discourage transparent address holding or punish historically allocated networks. A pure flat fee is simple, but it is regressive. Every option has a cost.

The first principle should be narrowness. The compulsory part of the fee should be tied as closely as possible to the functions members cannot reasonably avoid: uniqueness, registry accuracy, core database operations, secure resource certification, reverse-DNS continuity, essential member support, billing, transfer administration and institutional resilience necessary for those tasks. Activities beyond that narrow layer may still be valuable. The fiscal question is whether they belong inside the compulsory account charge or should rely more on sponsorship, voluntary funding, event fees, targeted service fees or explicit member-approved cross-subsidy.

The second principle should be predictability. Members can tolerate imperfect fairness more easily than unstable rules. Sudden changes create risk premiums and distort behaviour. Any move toward resource-based, account-based, hybrid or relief-oriented charging should include transition periods, caps, clear examples and forward modelling. The poorer and smaller the member, the more damaging surprise becomes.

The third principle should be visibility. A charging scheme should state who is expected to pay more, who is expected to pay less and why. If address-rich holders are asked to contribute more because they hold scarce, registry-recognised capital, the rationale should be explicit. If small members are protected because direct registry standing supports competition and regional resilience, the subsidy should be explicit. If all members pay the same because simplicity and low administrative cost are judged more important than progressivity, that too should be explicit. Fairness is not achieved by denying tradeoffs. It is achieved by making them reviewable.

The question before the next charging vote

The next charging debate should begin from incidence, not from the invoice. The invoice asks what RIPE NCC needs to collect and from which legal account. Incidence asks who ultimately bears the burden, how that burden changes behaviour and whether the distribution matches the institution's role. That second question is harder, but it is the one that matters in a region of extreme economic diversity.

The case for the current account-based model is serious. It is simple. It is predictable. It avoids complex verification of revenue, country income, address value or service consumption. It reduces administrative costs and makes the fee legible. It also prevents the registry from becoming a financial assessor of its members. Those are not small virtues. In infrastructure finance, simplicity often protects trust.

The case against the model is also serious. A fixed euro charge is regressive across unequal operators and markets. It can burden smaller and poorer networks more heavily than its nominal amount suggests. It can make direct membership less attractive for marginal operators. It can shape account consolidation, sponsorship, transfer timing and pass-through. It can let address-rich holders carry valuable scarcity at a low account-level cost while small service networks experience the same fee as a material overhead. It can conceal cross-subsidy behind administrative sameness.

The correct conclusion is not that RIPE NCC must abandon flat charging immediately. It is that the charging conversation should no longer treat formal equality as the end of the analysis. Members should ask for incidence tables, payment-friction metrics, account-profile modelling and a clear separation between compulsory registry functions and broader institutional activities. They should ask whether the fee model supports direct participation by small operators in lower-income markets or quietly pushes them toward intermediaries. They should ask whether address-rich holders, legacy direct relationships and sponsored resources are carrying the right share of the common cost. They should ask whether the customer ultimately pays through less visible channels.

Above all, they should ask what kind of registry economy the charging scheme is building. A good scheme should fund a stable registry without turning the account relationship into a hidden entry barrier. It should protect common services without disguising discretionary cross-subsidy. It should be simple where simplicity lowers cost and more precise where imprecision becomes regressive. It should recognise that a fee attached to an irreplaceable registry relationship is not just an invoice. It is a fiscal instrument placed inside the operating economics of the Internet.

That is the real question before the next vote: not whether EUR 1,800 is large or small in the abstract, but whether the burden created by the charging model falls where a serious public-finance analysis would expect it to fall. If the answer is no, then the debate is not about generosity. It is about incidence, incentives and the cost of calling unequal burdens equal.