Summary
- RIPE NCC made 2,638 IPv4 allocations in 2012, the year it began allocating from its final /8 under scarcity rules. In 2025 it reported 723 waiting-list allocations of IPv4 /24s. That is a 72.6% fall in the count of those labelled IPv4 allocation events, but the policy, allocation size and operational work behind the two figures differ, so the decline cannot be read as a 72.6% fall in total registry labour.
- The membership and institution moved in the opposite direction. RIPE NCC reported 8,760 members and 128 full-time equivalents at the end of 2012. For 2025 it reported 19,863 active members, 20,647 active LIR accounts and an average of 184.8 full-time equivalents. Members grew by about 126.7% between those endpoints and staffing by about 44.4%, while the service portfolio included registry accuracy, transfers, RPKI, the RIPE Database, measurement, security, governance, training and regional engagement.
- The annual fee did fall: EUR 1,800 in 2013, EUR 1,750 in 2014, EUR 1,600 in 2015 and EUR 1,400 from 2016 through 2022, with member redistributions in several years. It rose to EUR 1,550 in 2023 and EUR 1,800 for 2025 and 2026. The accurate question is therefore why the post-exhaustion reduction did not remain durable, not why management never lowered fees.
- A sensitivity test shows why transaction decline alone is limited public evidence. If narrow allocation handling represented 5%, 10% or 15% of total cost and its cost fell in exact proportion to the 72.6% volume decline, total cost would fall only 3.6%, 7.3% or 10.9%. Those are hypothetical bounds, not estimates: the actual allocation cost share and its variable component are not publicly identified on a comparable basis.
- Members need an annual workload-to-cost bridge that separates mandatory registry custody from discretionary or expandable activities, identifies fixed and variable cost, publishes comparable service units and requires explicit member authorisation when savings from declining work are retained for a broader mission.
The fee question begins with a correction
The simple story is attractive. IPv4 addresses ran out. A registry that once processed large numbers of allocations should have had less allocation work. Less work should have meant fewer staff, lower expenditure and lower annual fees. When the fee returned to EUR 1,800, the conclusion seems obvious: the institution kept the savings and expanded itself.
Part of that story is true, but the sequence is not. RIPE NCC did lower the annual fee after exhaustion. Its published charging history shows EUR 1,800 for 2013, EUR 1,750 for 2014, EUR 1,600 for 2015 and EUR 1,400 from 2016 through 2022. Members also received redistribution credits in multiple surplus years. The fee rose to EUR 1,550 in 2023 and remained there in 2024, then rose to EUR 1,800 in 2025. The current payment schedule keeps the annual fee at EUR 1,800 for 2026, alongside separate charges for some resources and sign-up.
The historical question is therefore narrower and more revealing: why did the lower-fee settlement not endure as allocation work changed? The answer cannot be read from an allocation count alone because the entity being financed changed. RIPE NCC ceased to be understood only as a queue that handed out addresses. It presented itself as custodian of an increasingly complex registry and as a provider of security, measurement, database, training, policy-support and regional-coordination services.
That may be a legitimate response to exhaustion. Scarcity can increase the importance of accurate records, transfer controls, abuse resistance and route-origin security even as primary allocations decline. A registry that distributed the last block and then disappeared would leave networks with a decaying title and routing system. The governance issue is not whether any new work was justified. It is whether members explicitly authorised the conversion of falling transaction work into a broader permanent mission, with enough cost accounting to judge each part.
Without that bridge, both sides can overclaim. Critics can treat every allocation as an equal unit of labour and infer that a 72.6% volume fall should remove 72.6% of cost. Management can list every useful activity and infer that the total budget follows. Neither claim identifies which costs are fixed, which vary with workload, which arose from member mandates, and which expanded through accumulated institutional discretion.
The workload that plainly fell was narrow but important
RIPE NCC's 2012 annual report records 2,638 IPv4 allocations, 1,289 IPv6 allocations and 2,500 Autonomous System Number assignments. It also reports 16,703 requests for Internet number resources and 9,448 allocations and assignments across resource types. That was the year RIPE NCC announced that it had started allocating IPv4 space from the final /8 under the region's last-block policy.
The 2025 annual report records 723 allocations of IPv4 /24s from the waiting list. Comparing 2,638 with 723 gives a decline of approximately 72.6% in the number of labelled IPv4 allocation events. The comparison captures a real structural change: the registry was no longer issuing ordinary IPv4 blocks from a plentiful pool.
It is not a labour-productivity series. A 2012 allocation and a 2025 waiting-list /24 were governed by different policies and size constraints. The reports do not state that they consumed equal staff time. The later registry must validate eligibility under scarcity rules, preserve a waiting list, maintain transfer and registration records and address fraud or sanctions risk. Automation may reduce handling time while scrutiny increases it. The number of addresses contained in an allocation is not a measure of labour either.
Nor did every resource transaction fall. In 2025 the registry still reported IPv6 allocations, ASN assignments, transfers and registry-maintenance work. Its services included the LIR Portal, the RIPE Database and RPKI. The correct conclusion is therefore modest: a historically central category of primary issuance fell sharply, and members should be able to see where its cost went.
That last clause is the missing account. A mature activity report should state direct staff hours, automated decisions, exception cases, median completion time, quality findings, fraud reviews and system cost for each major registry transaction. It should distinguish primary allocation from transfer, registration maintenance and certification. A count tells members what passed through the system. It does not tell them what it cost or whether the cost was controllable.
The institution grew while the allocation queue contracted
The endpoint comparison is striking. At the end of 2012, RIPE NCC reported 8,760 members and 128 full-time equivalents. In 2025 it reported 19,863 active members, 20,647 active LIR accounts and an average of 184.8 full-time equivalents. The membership count increased by about 126.7% between the disclosed endpoints. Average staffing increased by about 44.4%.
These percentages should not be forced into a productivity verdict. The 2012 membership figure and 2025 “active members” are reported under the organisation's contemporary definitions, which may not be perfectly stable. The 2012 staff figure is year-end; the 2025 figure is an annual average. A member can hold more than one LIR account. The service region, technical threat environment and regulatory exposure changed.
The comparison nevertheless invalidates the image of an unchanged registry serving the same population after one queue became shorter. RIPE NCC was serving more legal entities and accounts, and it employed more people to perform a wider set of functions. In the 2025 expense account, the organisation placed average FTE in four broad groups: 42.5 in Registry, 66.5 in Information Services, 41.9 in Community and Engagement and 33.9 in Organisational Sustainability.
Those headings make mission expansion visible, but they do not prove value. “Information Services” may contain critical measurement, security and database work; it may also contain projects whose marginal benefit members would rank below fee relief. “Community and Engagement” may support the open policy system and neglected regions; it may also grow through activities whose outcomes are difficult to test. “Organisational Sustainability” may contain necessary finance, legal, people, facilities and governance controls; it can also become an elastic home for overhead.
The member's problem is not that these categories exist. It is that their connection to the original registry obligation and to explicit authorisation is not always visible at the unit level. A budget can be approved as a whole while the activity mix changes incrementally. Each individual addition appears reasonable, but the cumulative result is a different institution funded by the same compulsory relationship between a resource holder and its regional registry.
Fee history shows adaptation, then reversal
RIPE NCC's 2025 charging consultation history provides a useful corrective to claims that fees never responded to scale. The annual charge moved down as the membership grew and surpluses accumulated.
| Charging year | Published annual fee per LIR account | Published redistribution credit |
|---|---|---|
| 2013 | EUR 1,800 | EUR 0 |
| 2014 | EUR 1,750 | EUR 0 |
| 2015 | EUR 1,600 | EUR 0 |
| 2016 | EUR 1,400 | EUR 400 |
| 2017 | EUR 1,400 | EUR 279 |
| 2018 | EUR 1,400 | EUR 359 |
| 2019 | EUR 1,400 | EUR 568 |
| 2020 | EUR 1,400 | EUR 354 |
| 2021 | EUR 1,400 | EUR 461 |
| 2022 | EUR 1,400 | EUR 614 |
| 2023 | EUR 1,550 | EUR 8 |
| 2024 | EUR 1,550 | EUR 54 |
| 2025 | EUR 1,800 | determined under the applicable result and decision |
| 2026 | EUR 1,800 | not treated here as a guaranteed credit |
The redistribution column is not a permanent fee cut. Credits depended on annual results and member decisions, and the treatment of taxes and timing could differ among members. Subtracting a redistribution from the headline fee can illustrate an effective contribution for a year, but it should not be used as the recurring price of membership. A registry needs a predictable charge rather than reliance on an uncertain refund.
The table still demonstrates that post-exhaustion economics were not ignored. The nominal annual fee fell by 22.2% from EUR 1,800 in 2013 to EUR 1,400 in 2016 and stayed at that level for seven years. Redistributions made the cash burden lower in several of those years. Any serious criticism must explain the later increase rather than erase the earlier reduction.
One part of that explanation is denominator risk. The consultation records 25,125 LIR accounts in 2019 and projected roughly 20,500 for 2024. The 2025 annual report gives 20,647 active LIR accounts at year-end. When a budget is spread through a flat account fee, a decline in accounts increases the charge needed from each remaining account unless cost also falls or another revenue source grows.
Another part is scope. By 2025 members were not paying only for manual allocation decisions. The organisation's total reported expenditure was EUR 38.952 million. Personnel expense alone was EUR 24.321 million, 62% of the total. Information technology, consultancy, outreach, depreciation and other operating costs supported a portfolio that did not move in direct proportion to IPv4 issuance.
The fee path is thus the combined result of scale, account counts, cost, reserves, other charges and political choice. It is not a unit price set by the number of allocations. That is precisely why a better unit account is needed.
A sensitivity range defeats false precision
Suppose, only for analysis, that the narrow handling of IPv4 allocation events represented a share of total institutional cost and that this share varied exactly with transaction volume. If the cost share were 5%, a 72.6% fall in volume would reduce total cost by about 3.6%. At a 10% share, the reduction would be 7.3%. At a 15% share, it would be 10.9%.
| Hypothetical share of total cost attributable to volume-sensitive IPv4 allocation handling | Volume decline applied | Hypothetical total-cost reduction |
|---|---|---|
| 5% | 72.6% | 3.6% |
| 10% | 72.6% | 7.3% |
| 15% | 72.6% | 10.9% |
These are not estimates of savings RIPE NCC should have achieved. The public reports do not identify a comparable 2012-to-2025 cost share for narrow IPv4 allocation handling, and much of that work may be fixed or may have shifted to waiting-list, transfer, audit and fraud controls. The calculation exposes the assumption hidden inside the simple story. A large fall in one workload produces a large institutional saving only if that workload was a large and variable share of the institution.
The opposite mistake would be to set the share near zero without evidence. Primary allocation was central to the historical registry, and the 2012 report records thousands of requests and allocations. Automation, exhaustion and policy change should have released some capacity. The institution should identify that capacity: positions removed or reassigned, system cost avoided, turnaround time changed and savings retained.
A range is more honest than a single invented amount. It tells members which missing fact drives the answer. If management can show that only 5% of cost was volume-sensitive, expectations of a dramatic total reduction weaken. If the comparable share was 20% or more and staff were simply moved into new programmes, the question becomes whether members authorised those programmes.
Sensitivity analysis should also include a fixed-cost floor. Registry custody requires secure systems, data integrity, authentication, legal capacity, audit, backup, incident response and qualified staff even if the number of new allocations approaches zero. Cost does not follow volume to zero. The relevant curve has a base, a variable slope and discretionary layers above it.
Account counts explain pressure, not legitimacy
The fall from 25,125 LIR accounts in 2019 to 20,647 at the end of 2025 changed fee arithmetic. Dividing 2025 expenditure of EUR 38.952 million by different account counts illustrates the effect:
| Account-count sensitivity | Total expenditure divided by the count |
|---|---|
| 25,125 accounts | about EUR 1,550 per account |
| 20,647 accounts | about EUR 1,887 per account |
| 20,000 accounts | about EUR 1,948 per account |
These values are not membership fees and not the cost of serving one LIR. Total expenditure supports non-LIR public services and community work. The year-end denominator is not an account-month measure. Some entities hold several accounts. Income also includes sign-up fees, resource charges, sponsorship, investments and other sources. The table is a denominator sensitivity, nothing more.
Its lesson is still powerful. The 2019 account peak could support a EUR 38.952 million expenditure envelope at an arithmetic average close to EUR 1,550. With 20,647 accounts, the same envelope exceeds the 2025 EUR 1,800 annual account fee before considering other income and charging details. A fee increase can therefore occur even when a transaction count falls, because the payer base contracts while the institution's fixed and broader costs persist.
That mechanism does not authorise the budget. It explains the pressure. Management should not be permitted to answer a legitimacy question with denominator arithmetic alone. If the membership base shrinks, members must decide whether to reduce cost, draw reserves temporarily, change the charging basis, price resource holdings differently or retain the service portfolio and pay more.
The account denominator also carries distributional consequences. A flat annual charge treats one LIR account as the unit even though accounts vary in resource holdings, revenue, operational complexity and ability to pay. Large organisations may hold multiple accounts; small networks may face a larger burden relative to turnover. A shrinking denominator can increase pressure to redesign fees, but redesign should not obscure whether total cost is efficient.
Allocation exhausted; custody did not
The strongest substantive answer to the fee question is that number-resource work changes after exhaustion rather than ending. The registry becomes less like a warehouse issuing stock and more like a title, transfer and security institution.
Scarce IPv4 addresses acquire higher market value. Transfers make accurate registration and authority checks more important. Hijacking and fraud risks increase. Resource holders need reliable RDAP and database records. Networks rely on RPKI certification to express route-origin authorisations. Members still need account access, corporate updates, mergers and acquisitions support, sanctions handling and dispute resolution. IPv6 issuance continues even when IPv4 is scarce.
These functions can require more judgement per exceptional case than routine allocation did. They also require resilient technology whose fixed cost is not sensitive to the number of tickets in one year. A secure signing system, replicated registry, audit trail and incident-response capability must work continuously.
This is not a blank cheque. Custody work should have its own service units and outcome measures. Members should see cost per maintained account and resource record, percentage of records validated, authentication failures, transfer completion time, RPKI availability, certificate incidents, database data quality, security findings and recovery-test results. Where publication could aid attackers, the detail can be aggregated or independently assured.
The critical distinction is between necessary custody and chosen expansion. Maintaining authoritative records, secure access and operational continuity follows directly from holding the registry mandate. A new training programme, research platform, government-engagement initiative or conference expansion may be valuable, but it requires a separate argument. Placing both under a broad claim of “registry service” denies members the ability to accept one and challenge the other.
The cost bridge should begin with stable service families
RIPE NCC's activity plans contain substantial detail, but organisational categories and reporting structures change. A member trying to compare 2012 with 2025 can encounter renamed teams, reallocated staff and different boundaries. That makes a long-term cost bridge difficult even when each year's report is internally clear.
A stable service-family account would solve much of the problem. It should preserve the following functions across reorganisations:
- Primary allocation and assignment.
- Transfer, merger and resource-holder change.
- Registry data maintenance, RDAP and database operations.
- RPKI and route-origin certification.
- Member account, billing and support.
- Security, fraud prevention, sanctions and compliance.
- Policy secretariat and formal member governance.
- Training, community events and regional engagement.
- Measurement, research and public technical information.
- Corporate overhead, legal, people, facilities and executive governance.
For each family, the report should show direct FTE, direct technology and supplier cost, allocated overhead, major workload units, service outcomes and year-on-year change. Shared systems can be allocated by documented drivers rather than hidden. Restatements should preserve at least five years of comparable history when categories change.
The purpose is not to pretend that every minute can be attributed perfectly. Activity-based accounting has its own judgement and overhead. The purpose is to make large shifts visible. If allocation staff move to RPKI, members can see the transfer. If a measurement platform grows while registry cost falls, the trade is explicit. If corporate overhead rises because legal or security exposure changes, the cause is named.
Most importantly, the bridge should identify authority. Each material activity should cite the member decision, policy obligation, statutory duty or board mandate that supports it, along with the review date. “Continuing activity” is not a perpetual authority category. A programme whose original rationale has expired should return for decision.
Personnel is where the workload question becomes real
Personnel accounted for EUR 24.321 million, or 62% of RIPE NCC's 2025 expenditure. Any claim about workload and cost must therefore explain people.
Staff numbers need not fall in line with transaction counts. The skill mix can change. A hostmaster role can be replaced by software engineering, security, legal or data-quality work. Growth in members creates support demand. Regulatory and cyber risks require expertise that a 2012 registry did not carry at the same level. Training and community functions may be deliberately expanded.
But reassignment is a budget decision, not a natural law. When automation or exhaustion releases capacity, management has at least four choices: remove the position through attrition; improve quality or resilience in the same service; move the capacity to another authorised function; or propose a new function to members. Only the third and fourth choices preserve the cost, and both need a visible rationale.
The annual plan should therefore include a capacity bridge. Opening FTE, hires, departures, automation released, work transferred, new mandates and closing FTE should reconcile by service family. Average FTE should be accompanied by year-end counts and vacancy assumptions. Contractor substitution should be disclosed so a nominal staff reduction cannot hide the same labour in consultancy.
Outcome measures must accompany staffing. More security staff should correspond to control coverage, response readiness and reduced exposure, not merely a headcount. More community staff should correspond to sustained member and operator participation, not only events held. More information-services staff should correspond to reliability, use and decisions enabled. A public-interest institution cannot always monetise outcomes, but it can state what changed.
Training and engagement require their own mandate
The 2012 report already showed that RIPE NCC was more than an allocator. It trained about 2,000 people, supported regional meetings and operated information services. Mission expansion did not begin suddenly after exhaustion. The organisation had long argued that technical coordination required skills, measurements and community infrastructure.
That history weakens any claim that members were surprised to fund all non-allocation activity. It does not settle how large those activities should become. A longstanding programme can still outlive its best format or grow beyond its mandate.
Training provides a useful example. Better-trained operators can improve routing security and registry accuracy. The benefit may reach members that never attend. Yet course counts and entity satisfaction do not reveal whether practice changed. Members need cost per completion, target audience, organisation type, geographic distribution, demonstrated skill gain and follow-up evidence. Free training is not free to the membership; it is a deliberate transfer from fees to capability building.
Engagement has a similar structure. Government relationships, regional events and policy outreach may protect the registry's operating environment. They can also become difficult to bound because success is described as access, awareness or influence. Each programme should identify the concrete risk or coordination gap it addresses, the constituency served, the time horizon and the decision that follows if evidence remains weak.
Separating these activities does not diminish them. It gives them a defensible mandate. A member can support regional training while asking for lower conference cost, or support measurement while questioning generic public relations. Aggregate approval suppresses those distinctions.
Surplus redistribution is not a substitute for cost discipline
RIPE NCC redistributed substantial per-account amounts in several years, including EUR 614 for 2022 according to the charging history. Returning surplus can be evidence that management did not seek to retain every euro. It can also reveal forecasting and structural questions.
A redistribution occurs after income and expenditure have produced a result. Members learn the effective burden after the operating decision has largely been made. A lower planned fee or smaller budget changes the authority before expenditure occurs. The two mechanisms are not equivalent.
There are legitimate reasons to prefer a stable fee with a later credit. Membership counts and new LIR openings can be uncertain. Unexpected risk may require reserves. A registry should not cut charges so far that it must raise them abruptly after a shock. Yet repeated large redistributions can indicate that the institution is collecting above its demonstrated need or using members as an interest-free buffer.
The policy should define a target accuracy band. If recurring operating results exceed it, the next plan should explain whether the cause was account growth, vacancy, delayed projects, investment result or conservative budgeting. Structural surplus should lower the planned charge or fund an explicitly authorised one-off investment, not silently enlarge the baseline.
Conversely, the disappearance of large credits as the account base contracts should not be portrayed as an unexplained loss. The annual bridge should show the effect of denominator change separately from cost growth. Members can then see whether they are paying more because there are fewer accounts, because activities cost more, because the mission expanded, or because all three occurred.
Fee design cannot repair an opaque cost base
Changing from a flat fee to a category or resource-based charge can improve distribution. It cannot determine whether EUR 38.952 million is the right total. A perfectly progressive charging scheme can still finance an inefficient institution; a simple flat charge can finance a disciplined one.
Cost scrutiny should therefore precede or accompany fee redesign. The registry should first identify the mandatory custody floor, authorised shared benefits, discretionary portfolio and reserve requirement. It can then decide how those costs should be distributed among accounts, resource holders and other beneficiaries.
The beneficiary principle has limits. Charging directly for every public database query or training benefit could undermine open coordination. Some goods are deliberately shared because exclusion would harm the Internet. The appropriate response is to state the subsidy and its mandate, not to pretend every cost is attributable to the LIR account that receives an invoice.
Resource-based charges also need care. Address holdings are not a clean proxy for current ability to pay, operating complexity or service use. Legacy allocations, IPv6 strategy and transfers can distort the relationship. A charging redesign should publish winners and losers by anonymised cohort, transition effects and the share of total cost that remains common.
The discipline lies in keeping the two decisions visible: how much the institution may spend, and who pays which share. Combining them allows a debate about distribution to displace the question of size.
Member authorisation must attach to retained savings
When a workload falls, a public-interest organisation can use the released capacity well. It may automate another weak service, improve security, validate old records or extend access to members that were underserved. The benefit can exceed a fee reduction.
The principal should decide. Management should present the released amount or capacity, the options and the expected return. Members should be able to choose among fee relief, reserve strengthening, service quality and new activity. If the amount cannot be measured precisely, management can present a range with assumptions, just as this article does.
The decision should occur when capacity is released, not years later when the new programme has staff, contracts and a constituency. Once a cost becomes recurring, stopping it is institutionally difficult. Early authorisation protects both the membership and the staff who would otherwise build work on an uncertain mandate.
Authorisation should also expire. A three-year investment in RPKI adoption may be justified by a measurable gap; after three years, the programme should show adoption and reliability outcomes before its temporary capacity becomes permanent. A training intervention should return with evidence of reach and skill use. A measurement service should show users, decisions and operational dependence.
This is not anti-innovation. It is the difference between experimentation and entitlement. A registry can reserve a bounded innovation portfolio while requiring successful experiments to compete for permanent funding.
A practical annual workload-to-fee statement
The annual statement should start with last year's expenditure, not next year's aspirations. It would reconcile:
- cost removed through automation, procurement, project completion and vacancies;
- cost added through membership or transaction growth;
- cost added through security, legal, regulatory or continuity requirements;
- capacity transferred among service families;
- temporary initiatives entering or leaving the baseline;
- inflation and contract effects;
- account-count change and other income;
- reserve contribution or draw; and
- the resulting fee requirement under alternative charging bases.
Each line should be supported by a service unit and authority. The statement should include a sensitivity band for membership and account counts rather than a single confident forecast. If the budget depends on 21,000 accounts, members should see what happens at 20,000 and 19,000. If investment income or sponsorship is uncertain, core custody should not depend on the optimistic case.
The statement should then show the fee options. A lower-cost plan would state which activities or service levels change. A baseline plan would state the approved mission. An expansion plan would identify the additional outcome and sunset. Members could make a real choice rather than vote on an aggregate whose internal substitutions remain invisible.
Independent assurance need not certify management's policy judgement. It should verify that cost categories reconcile, workload definitions are consistent, major reallocations are disclosed and sensitivity arithmetic is correct. The board remains responsible for recommendation; members remain responsible for mandate.
The right conclusion is neither “fees must track allocations” nor “scope must grow”
IPv4 exhaustion did not abolish the need for a registry. It changed the registry from an issuing institution toward a custody, transfer, security and coordination institution. That change can justify substantial fixed cost and specialist capacity.
It did not automatically authorise every programme that could be connected to Internet coordination. A useful training course, measurement tool or government meeting is not necessarily the best use of compulsory member income. The burden remains on the institution to show mandate, cost and outcome.
The fee history demonstrates both adaptation and institutional persistence. RIPE NCC lowered the annual charge and redistributed surplus during years of account growth. Later, as accounts fell and the broader portfolio cost remained, the charge returned to EUR 1,800. That sequence is more informative than a claim that fees ignored workload.
What remains absent is a durable economic contract for the post-allocation era. Members need to know the cost of authoritative custody, the cost of shared public-interest services, the cost of chosen engagement and the capacity released by automation or declining transactions. They need the right to decide whether savings reduce fees, strengthen continuity or finance a new mandate.
Until that contract exists, management can answer every falling-workload question with a list of new responsibilities, while critics can answer every budget with a declining allocation chart. A workload-to-cost bridge would force both accounts into the same frame. It would show that falling work does not mechanically produce lower fees—and that a broader mission does not become legitimate merely because the old queue became shorter.
What a durable lower-fee test would require
A serious lower-fee proposal should not begin by selecting a politically attractive number. It should begin with the custody floor. The registry would identify the people, systems, suppliers, assurance and reserve capacity required to keep records authoritative, credentials secure, RPKI available, transfers controlled and recovery tested. That floor should be costed under a normal case and a stress case, with no discretionary growth hidden inside it.
The second layer would contain member-authorised shared services whose value is not captured by transaction charges: policy support, public registry access, essential measurements, documented security coordination and narrowly defined operator capability. Each should have a service outcome and review date. The third layer would contain choices that can genuinely expand or contract—events, new research, experimental tools, campaigns and temporary engagement programmes.
Fee options could then be tested against the layers. A reduced charge might preserve the custody floor and core shared services while postponing selected initiatives. A baseline charge might maintain the present portfolio but require measurable productivity commitments. A higher charge might finance a time-limited security or renewal investment whose end date is fixed before recruitment and contracting begin.
The member-count sensitivity must remain visible. If the proposed fee assumes 20,700 accounts, the plan should show the result at 20,000, 19,000 and 18,000, with a rule for whether a shortfall triggers reserve use, cost reduction or a mid-cycle decision. Management should also disclose the difference between accounts, unique legal members and annual paying account-months. Selecting the largest denominator to make unit cost appear low would be as misleading as selecting the smallest to dramatise burden.
The same discipline applies to workload. Allocation counts, transfer cases, registry corrections, support contacts and security exceptions should be forecast as ranges. Productivity assumptions should identify the share of cost that actually varies. A ten-percent fall in tickets does not save ten percent of a team when minimum coverage and specialist knowledge must remain. Conversely, automated handling should not be credited as efficiency if the released staff and supplier cost simply enter an unapproved programme.
Finally, the option paper should state the decision threshold. If the lower-fee plan meets continuity and service standards, management should need evidence to reject it. If it creates an unacceptable failure or legal exposure, that risk should be specific enough for members to evaluate. “Complexity” is not a cost category, and “community benefit” is not an outcome by itself.
Such a test would make a lower fee possible without turning thrift into neglect. It would also make a higher fee defensible when the additional amount buys an authorised and measurable capability. The governing question is not whether EUR 1,400 or EUR 1,800 feels reasonable in isolation. It is which layers the members have chosen to fund, under which workload and denominator assumptions, with what right to reclaim savings when the assumptions change.
The bridge should distinguish price change from institutional growth
A nominal fee series cannot by itself show whether members received a real reduction. General price levels, salaries, technology contracts and the currencies in which suppliers invoice all change over time. Applying an inflation index may help, but choosing the index is a judgement: a European consumer-price measure does not necessarily match the registry's international labour, cloud, travel and security cost basket.
The annual statement should therefore publish both nominal and purchasing-power views, with the selected index and limitations. It should also show the organisation's own cost drivers separately. If personnel cost rose because specialist salaries outpaced the index, that is a labour-market explanation. If the average grade or staffing mix changed, that is an institutional choice. Combining both under “inflation” would conceal growth.
Currency effects need the same treatment. RIPE NCC reports in euros, while some suppliers or activities may be exposed to other currencies. A favourable exchange movement can lower a contract without any management efficiency; an adverse movement can increase cost without a larger service. The budget bridge should separate rate, volume, scope and currency effects whenever each is material.
This distinction matters to the post-exhaustion claim. Holding the annual fee at EUR 1,800 in 2013 and 2026 does not mean the economic burden or service bundle was identical. The later euro has different purchasing power, the payer denominator changed, separate resource charges changed and the institution performs different work. Conversely, saying the fee is “unchanged in real terms” would not prove that every added activity was authorised.
A useful reconciliation would begin with the prior year's recurring cost. It would add or subtract general price effects, supplier-specific rate changes, workload volume, productivity, account growth or decline, new legal or security obligations, and member-approved scope. The resulting requirement would then be divided under the chosen charging model. Each bridge item would carry a source and confidence range.
That structure prevents two rhetorical shortcuts. Management could not attribute an expanded portfolio to unavoidable inflation. Critics could not describe every nominal increase as institutional growth. Members would see which part of the fee preserves existing service, which part reflects the denominator, and which part purchases a larger mandate. A durable lower fee then becomes an evidence-based option rather than a slogan tied to one declining transaction count.
Sources and method
The 2012 membership, staffing, request, allocation, assignment, support and training figures come from the RIPE NCC Annual Report 2012, read alongside the RIPE NCC Activity Plan and Budget 2012 and the organisation's last-/8 announcement. The 2025 membership, LIR-account, staffing, activity and expenditure figures come from the RIPE NCC Annual Report 2025 and Financial Report 2025.
The annual-fee and redistribution series is reproduced from RIPE NCC's Charging Scheme 2025 consultation history, supplemented by the official 2013, 2014, 2015, 2016 and 2017 charging documents and the current 2026 payment page. The account-count and cost-share scenarios are sensitivity calculations, not forecasts, invoices or audited unit costs. No cross-year comparison assumes that the member definition, staffing measure, allocation policy or activity classification remained perfectly constant.