Summary

  • RIPE NCC's experience shows that redistribution is operationally possible at institutional scale. Its audited 2015 accounts recorded EUR 5.014 million to be returned as invoice discounts, the 2017 surplus produced a EUR 6.092 million redistribution, the 2018 amount reached EUR 11.003 million, and the 2025 accounts recorded EUR 2.768 million for credit on 2026 invoices. These figures are separate annual results, not a cumulative member entitlement.
  • A refund changes governance because it gives unspent compulsory contributions a destination outside the institution. Without a prior return rule, conservative revenue estimates, delayed hiring, one-off sign-up income and underspending can all enlarge reserves or finance new programmes without a second decision on whether members wanted those uses.
  • The strongest rule is not "return every surplus." It is a three-way settlement made before the money is collected: fund verified operating cost, replenish a reserve within a published band, and return the residual unless members separately approve a named use. Deficits should be treated symmetrically through the same reserve rule rather than used to justify a permanently padded fee.
  • RIPE NCC's 2026 reserve policy is an important advance because it sets a 60% minimum, 80% benchmark and 100% maximum of current-year budgeted operating expense, with an obligation to publish a replenishment or reduction plan outside the band. Yet a percentage of total budget can still ratify mission growth unless the cost base itself is divided between core registry continuity and optional activity.
  • A defensible refund formula must disclose eligibility, denominator, timing, tax treatment and delivery method. A credit based on fees paid is administratively different from cash and can exclude members that closed before year-end; those choices may be reasonable, but they allocate value and should not be hidden inside billing administration.

The important feature of a refund is that the money can leave

Institutions tend to describe a refund as an accounting consequence. Revenue exceeded eligible expenditure; the annual result became clearer after audit; a credit then appeared on the next invoice. That description is accurate and incomplete. The institutional significance is that money collected under the registry's charging power has a destination other than the registry.

That outside destination changes decisions before any credit is issued. Management preparing a budget knows that an underspend cannot automatically become next year's programme. A board considering a generous contingency knows that unused protection may return to payers. Advocates of a new activity must ask members for a mandate rather than relying on a favourable outturn. Members deciding a fee can tolerate prudent uncertainty because over-collection is reversible.

Remove the return route and the incentives reverse. A fee set slightly high looks responsible. Vacancies, delayed procurement and unexpectedly strong account income create a surplus. The surplus enlarges reserves. A larger reserve makes a new initiative appear affordable. Once staff and contracts exist, the initiative becomes part of the recurring cost base. The next fee then finances the larger institution. No single step is necessarily improper, but caution has been converted into scope.

This is why a refund belongs in constitutional design rather than customer service. It determines who owns the benefit of forecasting error. If the institution keeps every favourable variance while members absorb every unfavourable one through later fees, the budget has an upward ratchet. If the residual normally returns, management can still protect continuity, but expansion requires an affirmative decision.

The distinction matters especially for a Regional Internet Registry. Its members cannot readily replace the provider of recognised registration, reverse-DNS administration, transfer records and resource certification. The invoice is formally associational, yet the exit channel is narrow. A rule that sends genuine excess back to the payer partly compensates for that weak market discipline. It does not create competition, but it prevents lock-in from becoming an unlimited claim on accumulated contributions.

RIPE NCC turned an abstract principle into an invoice credit

The RIPE NCC record supplies a useful exhibit because redistribution was not left as a theoretical right. The mechanism was changed in 2014 and first put to members for the 2015 result. A contemporaneous announcement explained the fiscal setting: an annual surplus could be returned to members to reduce the taxable result, while retaining it would generate corporate income tax and add the balance, after tax, to the Clearing House reserve.

The mechanics were concrete. Eligible members active at 31 December would receive a rebate in proportion to annual service and relevant assignment fees paid. Sign-up fees were excluded. Members joining after the first quarter received a pro-rata amount. The return would appear as a discount on 2016 invoices rather than as an immediate cash transfer.

The audited 2015 Annual Report later recorded EUR 5.014 million of member-fee redistribution and a further EUR 321,000 liability for later repayment. It identified 12,764 eligible members and an average amount of EUR 393. The figures differ from the approximate EUR 4.2 million presented before year-end because a forecast vote necessarily preceded the final audited calculation. That difference is evidence for a true-up, not evidence that the institution should wait until every number is final before deciding the rule.

The form also matters. Calling the amount a refund can suggest that the original invoice was unlawful or mistaken. The documents describe an excess contribution produced by the annual result. Calling it a dividend would suggest a return on invested capital. It was neither. It was a proportional settlement of member-funded operating revenue after the institution had met expenditure and applied its reserve and tax arrangements.

That precision should be preserved. The governance claim is not that each euro remained legally owned by the member after payment. It is that the payment instrument, service agreement and member resolution created a route by which an excess could leave. Once such a route exists, retaining the money becomes a decision requiring reasons, rather than the natural fate of every surplus.

The repeated redistributions establish feasibility, not a permanent promise

The next years demonstrate that a return mechanism can operate repeatedly without preventing registry continuity. They also show why no single historical credit should be advertised as a recurring price reduction.

For 2017, the RIPE NCC published an estimated redistribution of EUR 6,092,310. Its calculation page used EUR 23,742,825 of relevant annual fees as the denominator. A member paying a EUR 1,400 annual fee and no separate assignment fees was shown an illustrative credit of EUR 359.23. The amount was placed on 2018 invoices, and quarterly payers received it across four invoices.

For 2018, the final published amount was EUR 11,002,630 against EUR 27,142,388 of relevant fees. The page records that the figure was adjusted down by EUR 300,000 after final confirmation. A one-account member with a EUR 1,400 annual contribution and no qualifying extra assignments was shown a credit of EUR 567.51. That is roughly 40.5% of the headline annual contribution, an unusually visible demonstration of how far actual income and expenditure can diverge from the fee set in advance.

For 2019, the published redistribution was EUR 8,365,249 against EUR 33,099,600 of relevant annual fees, producing an example of EUR 353.82 on a EUR 1,400 contribution. But only half of the surplus was returned. Members decided to retain the other half in reserves.

These are not amounts to be summed and divided by today's membership. Eligibility dates, account populations, qualifying fees and institutional circumstances changed. Some members paid for multiple accounts or assignment charges. Credits were attached to subsequent invoices. The numbers prove administrative capacity and materiality. They do not prove that every later member should receive the same percentage, or that a registry can plan annual cash needs as though the largest historical credit will recur.

The governance lesson is narrower: once annual settlement is routine, cautious budgeting need not imply permanent overfunding. Members can finance uncertainty for one year without silently endowing every future year.

The 2019 split vote exposed the real choice

The October 2019 General Meeting is especially instructive because the choice was not presented as a binary contest between fiscal virtue and member self-interest. The Board proposed adding the whole surplus to reserves, adding half and returning half, or allowing the entire amount to be returned if the first two options failed.

The meeting minutes record that the full-retention option received 561 votes in favour, 1,124 against and 64 abstentions. The half-retention option then received 911 votes in favour, 738 against and 59 abstentions, and passed. The outcome was a compromise: more capital for the institution and a material credit for members.

This sequence shows why decision architecture matters. Members did not vote on an undefined proposition that the Board should "use the surplus wisely." They saw allocations with different consequences. Rejection of the first option did not produce paralysis. It opened the second. Rejection of both would have activated a known default. The money therefore could not remain in institutional limbo while management searched for a purpose.

It also shows the risk of agenda control. Sequential voting can favour the option placed first because voters may support it to avoid an uncertain fallback. In 2019 the fallback was disclosed, which reduced that uncertainty. A durable system should go further by stating the settlement order in governing documents: first determine the audited operating result; then apply the pre-authorised reserve rule; then return the residual unless a separately described use wins approval.

Separate votes are essential. A member may favour a stronger reserve but oppose a conference expansion. Another may favour a one-year security programme but oppose permanent headcount. Combining those choices into "retain the surplus" gives the institution a blank cheque. A good resolution names the amount or formula, reserve effect, programme, duration and future recurring cost. The member then decides an allocation rather than endorsing an aspiration.

The 2019 record does not prove that the compromise was financially optimal. Public documents cannot reconstruct every risk estimate known to the Board. It proves something more basic: surplus allocation can be made contestable, legible and capable of producing a mixed answer.

A rule must work in deficit years as well as surplus years

A refund principle that operates only when money is plentiful is not a financial constitution. It must explain deficits, otherwise the institution can return favourable variances for several years and then demand an emergency assessment at the first shock.

RIPE NCC's more recent resolutions have become explicitly symmetric. The October 2024 minutes record a decision that an excess contribution would be redistributed in 2025 while a deficit would be covered by reserves. The audited 2024 Financial Report then reported no redistribution because operations produced a deficit. A small total surplus remained after financial results, including unrealised gains and interest, but those items were outside the operating contribution used for redistribution.

That separation is important. If investment gains automatically subsidise current fees, members may become dependent on market performance. If unrealised gains are returned and markets later fall, the reserve can lose protection twice. Conversely, if every investment gain is retained while every operating surplus is also retained, accumulated capital can become detached from any stated need. The rule needs distinct treatment for operating contributions, reserve investment returns and restricted funds.

The 2025 result demonstrates the other side. The audited 2025 Financial Report recorded EUR 40.292 million of member-fee income, EUR 807,000 of sign-up fees and EUR 2.768 million of redistribution. It said the operational surplus would be returned to eligible members through credits on 2026 service-fee invoices after the October 2025 approval. After redistribution, a EUR 96,000 operating result remained because of specific tax adjustments; the broader surplus after financial results was EUR 622,000.

The contrast between 2024 and 2025 is useful. The mechanism did not manufacture a credit in a deficit year to satisfy a political promise. Nor did the following year's higher revenue disappear into reserves by default. It settled each period under a known rule.

Symmetry still has limits. A reserve can cover a temporary deficit, not a fee structurally below recurring cost. A credible constitution should require a fee review when operating deficits persist for a defined number of periods, just as persistent surpluses trigger a reduction or refund. Otherwise one side of the rule eventually consumes the other.

A reserve band is necessary, but its denominator can grow

The 2026 RIPE NCC reserve policy turns a long-running argument into measurable boundaries. It aims for a Clearing House Reserve between 60% and 100% of current-year budgeted operating expense, identifies 80% as a benchmark, and requires the organisation and Executive Board to publish a replenishment or reduction plan when the latest audited position falls below or above the band.

Using the EUR 41.125 million 2026 operating budget, the document translates those percentages into a EUR 24.7 million minimum, EUR 32.9 million benchmark and EUR 41.1 million maximum. The audited 2025 consolidated equity of EUR 33.663 million sits near the benchmark, although equity, the Clearing House balance and immediately liquid funds should not be treated as identical without reading the accounts.

This is better than an unbounded appeal to resilience. A lower trigger tells members when the institution needs repair. An upper trigger recognises that too much capital has costs. A plan obligation prevents an out-of-band result from being explained away as temporary without an identified response.

Yet the denominator creates a second-order risk. If the reserve target is a percentage of total operating expense, every expansion of the annual budget also expands the permitted reserve. Add EUR 5 million of recurring activity and the 80% benchmark rises by EUR 4 million. The institution can then argue that it needs to retain more surplus because the larger institution requires a larger buffer. Mission growth and reserve growth validate each other.

The answer is not to abandon the percentage. It is to use at least two cost bases. The core-continuity band should be tied to the cost of keeping registration, database publication, authentication, reverse DNS, resource certification, security, essential support and lawful succession running under a stress plan. A separate operating-liquidity amount can cover normal cash timing. Optional programmes should not automatically enlarge the continuity reserve; their shutdown or funding plans should be explicit.

An automatic refund trigger built only on total budget is therefore incomplete. It controls the height of the reservoir but lets the institution widen the reservoir. Members need to approve both dimensions.

Excess must be measured against an authorised cost, not any cost incurred

Accounting surplus is a necessary starting point, not a complete definition of excess. An institution can spend every euro and report no surplus even when part of the expenditure lacked a convincing member mandate. Conversely, it can report a surplus because a vital security project was delayed, in which case returning all cash may create a false economy.

A governance-grade calculation begins with authorised cost. Each material activity needs a purpose, service population, accountable owner, period and approval route. Shared overhead should be allocated by a disclosed method rather than hidden in a broad corporate line. Capital projects should distinguish cash paid, depreciation and future commitments. Vacancies should not be treated automatically as savings if the role remains necessary and recruitable.

The settlement can then be expressed without pretending to know one universal reserve number:

returnable residual = eligible contribution income - verified eligible operating cost - approved reserve replenishment - separately approved one-off use

Every term needs boundaries. Eligible contribution income should identify whether sign-up, reactivation, transfer and sponsorship income are included. Verified cost should be reconciled to audited accounts but remain divided by activity. Reserve replenishment should follow a stress-based band. A one-off use should expire and should not create recurring expense without another decision.

The formula prevents a common manoeuvre: announcing a surplus and a new project in the same paper, then presenting retention as costless because the money has already been collected. The project may be worthwhile. It still competes with returning the residual and lowering future charges. Its proponents should show expected result, total lifecycle cost and why current payers rather than voluntary funders or future beneficiaries should finance it.

This standard does not invite members to micromanage every supplier invoice. The Board remains responsible for execution within an approved envelope. It requires that material departures from the authorised cost map cannot acquire legitimacy merely by being spent before year-end.

The formula distributes power as well as money

Once a returnable amount exists, the allocation formula decides which members count. RIPE NCC's historical formula generally followed relevant annual contributions paid. That approach has a coherent logic: those who financed more of the eligible revenue receive more of the residual. It resembles a cooperative patronage allocation more than an investor dividend.

But contribution proportionality is not neutral. A member with multiple LIR accounts may receive more than a member operating one large network. Separate assignment charges can increase the numerator. Sign-up fees have generally been excluded even when unexpectedly high sign-up income contributed to the year's favourable result. Members that close before the year-end eligibility date can receive nothing, although they paid for service during part or all of the year. A merger completed before the cutoff can alter the receiving account.

These may be acceptable administrative choices. A closed entity can be difficult to pay, and a sign-up fee may finance onboarding work rather than recurring service. The problem arises when the formula is treated as a technical detail rather than a distribution decision.

A strong rule would publish a denominator card before the vote: eligible revenue by fee type; count of paying legal members and accounts; treatment of partial-year members, closures, mergers, arrears and sanctions-related payment barriers; estimated credit at several common fee profiles; and the date at which status is fixed. After audit, the organisation would publish the actual denominator and explain changes.

Downstream incidence should also be acknowledged. An LIR may pass registry cost to customers. Returning a credit to the LIR does not guarantee a lower retail price. That does not invalidate the return, because the contractual payer remains the administrable recipient. It does mean the registry should not claim a measured end-user benefit without evidence.

The purpose of the formula is not to simulate perfect economic justice. It is to make the allocation of a scarce institutional benefit predictable and contestable. A hidden denominator turns a governance right into a billing surprise.

An invoice credit is not the same as cash

RIPE NCC's redistributions have commonly appeared as discounts on the following year's invoices. This form is efficient. The registry already has billing details, the member has a new obligation, and offsetting the two reduces transaction cost. It can also support the tax treatment described in the institution's documents.

The credit form creates consequences. It keeps the value within the continuing service relationship. A member receives the benefit only by remaining eligible and receiving another invoice. A network that leaves, merges or becomes unable to pay through ordinary banking channels may not experience the credit like a continuing member. Quarterly billing can spread the amount across several dates rather than supplying liquidity immediately.

For a routine annual true-up, these frictions may be smaller than the cost and fraud risk of thousands of cash payments. For an unusually large residual or a structural fee correction, members should be offered a clearer choice. The institution could credit the next invoice by default while permitting verified cash settlement above a materiality threshold, subject to legal and tax constraints. Unclaimed amounts should not quietly revert to unrestricted spending; they should follow a published destination.

The language on the invoice should distinguish the year of origin from the year of delivery. A 2025 excess credited in 2026 is not a reduction in the cost of 2026 operations. Treating it as such distorts fee comparisons and can make the following year look artificially expensive when the credit disappears. The gross annual charge, prior-year credit and net amount payable should be separate lines.

Tax effects belong in the same disclosure. A decision to retain a surplus may create tax cost under a particular arrangement. That cost is relevant but not dispositive. Members should see the net amount available under each option: return, retain after tax, or finance an approved use. Otherwise tax efficiency can become a rhetorical substitute for deciding whether the institution needs the money.

Delivery design may sound mundane. In a captive service relationship, it determines whether the right is portable, timely and intelligible. That makes it governance.

Cooperative practice supplies a principle, not a legal label

Regional Internet Registries are not all cooperatives, and legal forms differ. Cooperative practice is still useful because it confronts the same question: when users finance a common service, who controls the capital and the residual?

The International Cooperative Alliance's third principle says members contribute equitably to, and democratically control, cooperative capital. Members may allocate surpluses to reserves, development of the cooperative, benefits in proportion to transactions, or other member-approved activities. The principle does not require emptying the treasury. It requires that retained capital and other uses derive from member choice.

United States cooperative tax guidance adds a more exact feature. The IRS instructions for Form 1120-C describe patronage dividends as amounts paid under a pre-existing obligation, based on the quantity or value of business done with or for a patron and determined by net earnings from that business. The legal and tax details cannot be transplanted into a Dutch association or another RIR. The institutional insight can: an ex-ante obligation is stronger than annual generosity.

If a board decides after seeing the surplus whether members deserve a return, the credit remains a concession. If governing documents say that eligible residuals follow a stated formula unless members approve another use, the return becomes a property of the charging system. Managers can budget around it; members can price the likely range; auditors can test compliance.

Cooperative practice also warns against equal distributions detached from use. One member, one vote may be suitable for governance while patronage remains proportional to transactions. A registry can similarly use equal or capped member voting while returning over-collected contributions according to fees paid. Political equality and financial incidence need not use the same denominator.

The comparison should not be romanticised. Cooperatives can hoard capital, obscure allocated reserves and delay redemption. A registry refund constitution should borrow the discipline of pre-existing obligation and member allocation without assuming that a cooperative label guarantees either.

Utility true-ups show why forecasting error should be reversible

Regulated utilities provide a second comparison. They invest in essential infrastructure with large fixed costs, uncertain demand and customers who cannot easily choose another network. Price controls often set allowed revenue in advance and reconcile over- or under-recovery later.

The analogy is not legal equivalence. An RIR membership association is not a state-regulated water or electricity monopoly. Yet the economic problem is recognisable: a provider must collect enough to maintain a continuous shared service before exact demand and cost are known.

Ofwat explains that it sets the revenue water companies may recover over a five-year control and requires charging differences to reflect cost differences. Its reconciliation materials describe symmetric adjustments for earlier over- or under-recovery. The United Kingdom Competition Commission has likewise required customer refunds for over-recovery in a network price-control period, as recorded in the Northern Ireland Electricity determination.

The useful principle is not external regulation. It is reversible forecasting. A fee can be prudently set before the period without granting the institution every favourable error. If actual eligible revenue exceeds allowed cost, future charges fall or money returns. If revenue falls short for legitimate reasons, a reserve or later adjustment can restore balance. Both directions use the same cost boundary.

This is more disciplined than demanding perfect budgets. Forecasts will be wrong. Membership counts change, sanctions affect collection, projects slip and security costs emerge. A system that punishes every underspend encourages rushed procurement; a system that lets every underspend be retained encourages padding. True-up design permits caution without rewarding inaccuracy.

There is a warning too. Regulated price controls can become technically impenetrable and dominated by specialists. An RIR should not reproduce a utility regulator inside its annual meeting. Its rule can be shorter: a small number of cost pools, audited actuals, a reserve band, a variance threshold and an automatic destination. Complexity should match the institution, not provide shelter from scrutiny.

The board needs discretion inside the year, not ownership after it

No registry can run safely if every unexpected expense requires a member referendum. Security incidents, legal obligations, supplier failures and staffing emergencies demand executive action. A refund rule should strengthen, not disable, that capacity.

The clean division is temporal. Before the year, members approve the charging method, material activity envelope, reserve policy and residual rule. During the year, the Board moves money within defined limits, draws contingency where triggers are met and reports material departures. After the year, independent accounts establish the result and the residual rule settles what remains.

This gives the Board operating discretion without granting it a perpetual option over unspent money. Emergency spending remains possible. If the emergency did not occur, the contingency does not become a free fund. If a new need emerges late, the Board may propose a one-off retention with reasons and a sunset. Members can approve it instead of receiving the credit.

Thresholds prevent trivial decisions from consuming the association. Small residuals can be carried into the next fee calculation if the per-account amount would cost more to administer than it returns. The threshold itself should be fixed and periodically reviewed. A percentage of eligible income is better than an amount management can outgrow through inflation.

Board proposals should include a counterfactual. If EUR 3 million is retained, how many months of core continuity does it add? What fee reduction or average credit is forgone? What recurring cost follows? What happens if the proposal fails? Without those questions, members hear only the benefits of spending and none of the price of keeping the money.

Auditors can verify the arithmetic and conformity with the adopted rule. They should not decide the mission. An independent finance committee can test cost allocation and stress assumptions. It should report to members, not turn technical expertise into another source of unreviewable discretion.

The objective is a board strong enough to act and constrained enough to settle. Those are complements, not opposites.

Mission expansion should face a separate invoice

The most important effect of an automatic return is on activities that have not yet become ordinary. Training, research, grants, public-policy engagement, measurement, meetings and regional outreach may generate real collective value. They are not identical to maintaining authoritative registration, reverse DNS, resource certification and secure account control.

When all activities share one contribution and every residual stays inside, the distinction weakens. A new programme can begin with temporary underspend elsewhere, demonstrate a constituency, hire staff and then enter the base budget. Members are later told that cutting it would harm service or community expectations. The financing sequence has decided the mission before the membership confronts the recurring cost.

A refund constitution forces a cleaner choice. Core and already authorised shared services remain in the base contribution. A material mission addition receives a separate resolution stating cost, duration, target population, evidence of benefit and exit conditions. It may be financed by a temporary assessment, a specified share of the residual or voluntary contributions. Approval converts the money into an authorised use; rejection sends it back under the default rule.

Separate financing also improves evaluation. If a EUR 1 million programme is buried in a EUR 40 million total, its result is difficult to connect to its cost. A named envelope can report outputs and outcomes. At expiry, members can renew, redesign or stop it. The institution cannot argue that general solvency proves programme value.

This does not mean only directly transactional services are legitimate. Number-resource coordination creates public and regional benefits. Training can improve registry data and routing security. Research can reveal operational risk. Engagement can bring underrepresented networks into decisions. The point is that diffuse benefit requires more explanation, not automatic exclusion.

The separate-invoice principle is constitutional rather than literal. The institution need not issue multiple payment documents. It needs separate authority. A single invoice can show core contribution, approved temporary levy, resource-specific charge and prior-year credit as distinct lines. The member can then see which obligation preserves the ledger and which choice expands the institution around it.

Bad refund rules can create austerity, gaming and inequity

Refunds are not harmless by definition. An automatic rule tied to the wrong measure can damage the service it is meant to discipline.

First, management may defer necessary maintenance to produce a politically popular credit. Boards seeking re-election can advertise a return while technical debt grows. The response is a multi-year asset and security plan with condition indicators, not permission to retain every variance.

Second, departments may rush to spend near year-end because unused budget returns. This is a familiar use-it-or-lose-it effect. Carry-forward should be allowed for approved multi-year projects, with named milestones and a cap. Unexplained carry-forward should enter the residual.

Third, members with the most accounts or chargeable resources may receive most of the return even where smaller operators bear greater affordability pressure. The formula should follow the contribution basis unless members deliberately choose a different allocation, but the distribution by member size should be published. A progressive fee design belongs in the charging decision, not in an opaque refund correction.

Fourth, a credit can favour incumbents over departing members. Eligibility based on year-end status is easy to administer but weakens the exit right. A better design would preserve claims for members that paid eligible fees and left in good standing, at least above a threshold and for a reasonable claim period.

Fifth, the prospect of a credit can distract members from the gross budget. A EUR 300 return after a EUR 400 fee increase is not proof of economy. Members should compare gross contribution, eligible cost, reserve movement and credit over several years.

Finally, refunds can become theatre. A board can set an inflated fee, return a fraction and claim generosity. Repeated material credits should trigger a prospective fee reduction or a revised forecasting method. The return mechanism is a safety valve, not a substitute for setting the valve correctly.

These risks argue for a better rule, not for abandoning return. A system with no release mechanism contains all the incentive to over-collect and none of the visible correction.

A practical refund constitution

A member-funded registry could adopt a compact settlement article with nine parts.

One: define eligible income. State which annual, resource, transaction and joining charges enter the true-up, and explain excluded revenue.

Two: define authorised cost. Publish core continuity, shared member service, optional programme and restricted project pools, with a stable overhead allocation.

Three: set the reserve corridor. Tie the core portion to tested continuity scenarios and specify lower, benchmark and upper actions. Do not let optional spending automatically enlarge the core target.

Four: settle symmetrically. Use reserves for bounded eligible deficits and return eligible residuals above the target. Persistent variance in either direction triggers a prospective fee review.

Five: make return the default. Retention outside the adopted corridor requires a separate member resolution naming amount, purpose, duration, recurring consequence and unused-balance treatment.

Six: publish the denominator. Show eligible payers, fees, status date, closures, mergers, arrears, partial years and common-account examples before and after audit.

Seven: protect exit. Permit credits or claims for members that paid during the year and departed in good standing, subject to proportionate identity and payment controls.

Eight: separate delivery from origin. Show the gross next-year charge and prior-year credit separately, with cash alternatives for material amounts where practicable.

Nine: audit compliance. Reconcile the settlement to audited accounts, publish all formula inputs in reusable form, and require an independent opinion on whether the rule was followed.

The constitution should also contain an anti-avoidance clause. Renaming operating income, moving cost to a controlled affiliate, accelerating a discretionary commitment or classifying an ordinary programme as a reserve purpose should not defeat the settlement. Material judgement calls should be listed with their effect on the residual.

No formula eliminates politics. It improves the entity of politics. Members argue about a visible reserve band, a named project and an explicit forgone credit instead of debating whether an undefined surplus is "safe" to keep.

The evidence supports a right to a rule, not a claim to every euro.

The public record permits strong conclusions and imposes firm limits.

It establishes that RIPE NCC has repeatedly calculated and delivered substantial member redistributions; that members have chosen full, partial and conditional returns; that recent resolutions treat operating surplus and deficit symmetrically; that audited 2025 accounts recorded a EUR 2.768 million redistribution; and that the 2026 reserve policy now contains a 60%-to-100% corridor with action outside it.

It does not establish that every annual accounting surplus is excess capital. Financial returns, restricted balances, tax adjustments, unpaid invoices, lease commitments and operational contributions have different meanings. It does not establish that a particular reserve percentage is correct for every RIR. It does not show the complete cost of cyber recovery, litigation, banking disruption or service succession. It cannot prove how much of a credit reaches downstream network customers.

The evidence also does not make RIPE NCC a cooperative or regulated utility. Those comparisons contribute design principles: member allocation of residuals, pre-existing obligation, allowed-cost discipline and symmetric reconciliation. Legal implementation remains specific to the association, service agreement, tax ruling and member resolutions.

The defensible claim is therefore constitutional. Members who finance an indispensable common service should have a prior rule for what happens when collections exceed authorised cost and prudent protection. That rule should be capable of returning money, capable of retaining it for a named purpose, and incapable of converting silence into permanent scope.

A refund will not by itself make a registry accountable. It will not correct a weak election, improve a disputed policy or make exit portable. It does one important thing: it denies the institution automatic ownership of its own forecasting error.

That is enough to change behaviour. A fee becomes a bounded contribution rather than an open-ended transfer. A reserve becomes an instrument with upper as well as lower conditions. A new activity must compete with an observable alternative. A board can act during the year and still settle with members after it.

The most credible refund may be the one rarely paid because fees and authorised costs converge. Its presence is still essential. The release valve disciplines pressure even when it remains closed.

Sources