Summary
- The NIR model buys local-language service, local incorporation and domestic knowledge while keeping regional policy and ledger control at APNIC.
- Seven NIRs are currently recognised; the moratorium began on 27 February 2012 and was made permanent on 26 February 2024, making the existing structure an inherited bargain rather than an expanding design.
- APNIC's 2022 hypothetical shows the representation issue sharply: one extra-large NIR would receive 64 APNIC votes, while 500 direct very-small members would receive 1,000 votes. That is not observed turnout, but it exposes the aggregation mechanism.
- Transfer and direct-membership routes can discipline service only if their cost, timing and conditions are practical; published procedure is not the same as measured portability.
A request can be local while authority remains regional
Imagine a downstream network operator approaching a National Internet Registry with an address request. This is a composite mechanism, not a real applicant and not a claimed outcome. The point is to follow the interfaces. The operator talks to a local institution, perhaps in a local language, perhaps using familiar currency, local forms and staff who understand domestic provider structures. The NIR reviews the request under rules that ultimately depend on APNIC's regional policy framework. In some circumstances, regional second-opinion or ledger procedures may become relevant.
The resulting resource record still has to fit into APNIC's regional registry system.
That journey shows the bargain. Front-line service is closer. Binding policy is still regional. The downstream organisation may have a relationship with the NIR, but not necessarily the same contract, vote or appeal route it would have as a direct APNIC member. The NIR is separately incorporated and not merely an APNIC branch office. Yet it remains an APNIC member bound by regional policies and the cooperative contract structure. The operator receives localisation and intermediation together.
The central question is whether that intermediation decentralises power or adds a layer. The answer is not one thing. It decentralises service. It can centralise governance. It may lower language and administrative cost. It may also aggregate many downstream organisations into one APNIC member position. It may make fees more intelligible locally while making total cost harder to compare. It may provide a nearby help desk while moving the decisive ledger and policy authority to a regional level.
The current APNIC description recognises seven NIRs. It also states a one-per-economy rule and, in most cases, a choice between APNIC and the local NIR, but not resources from both. The moratorium on new NIRs began on 27 February 2012 and was made permanent on 26 February 2024. Those are rule facts. They do not prove service quality, actual downstream prices, transfer completion times, appeal success or the reasons behind every policy choice.
The NIR bargain therefore has to be audited surface by surface: service, policy, fees, votes and exit. "Decentralisation" cannot answer all five.
Local service is the strongest part of the bargain
The strongest case for NIRs is practical. A regional registry serving a large and diverse region cannot assume that all operators are equally comfortable with regional meetings, English-language documentation, cross-border payment, unfamiliar legal instruments or remote staff. A local registry can reduce those costs. It can understand domestic corporate forms. It can communicate in local language. It can handle currency and invoicing in ways familiar to members. It can explain APNIC policy inside a local operator community.
That is not cosmetic. Transaction costs shape who can use governance. If a small operator must understand regional policy, pay foreign invoices, communicate across time zones and interpret unfamiliar membership rules, the operator may depend on intermediaries anyway. A well-run NIR can make the registry service more usable for organisations that would otherwise appear only indirectly in APNIC's system.
Separate incorporation also matters. NIRs are not simply APNIC branch offices. They have their own local bodies, local staff, local members and domestic responsibilities. That can create accountability closer to the downstream community. If a local operator has a service concern, the first conversation may be with an institution it knows. If language or domestic documentation is the barrier, the NIR may be the more realistic access point.
The published model also preserves regional coherence. NIRs are bound by regional APNIC policies. They do not run separate ledgers outside the regional system. The unique-registry function remains intact. That matters because localisation should not create duplicate allocation risk or competing registry truth. The bargain is meant to provide local service while preserving a coherent regional number-resource ledger.
The evidence supports the plausibility of this benefit. It does not measure the benefit across all seven NIRs. The fixed sources do not provide comparable fee schedules, service levels, refusal reasons, complaint outcomes or transfer times for every NIR. They do not show whether local-language service actually reduces delay for small operators or whether local fees offset savings from proximity. The local-service case is strong as institutional logic and unproven as a complete comparative outcome record.
That distinction is important. NIRs may be useful without every NIR being equally useful. Local service is a reason to keep the model under serious review, not a reason to stop measuring it.
Binding policy remains with the regional ledger
The second surface is policy authority. The NIR may be local, but it is not a sovereign allocation regime. APNIC's operational policy for NIRs, current transfer procedures and model relationship agreement keep the NIR inside the regional policy and ledger architecture. That is the point of the model. Local service should not fragment the registry.
For downstream organisations, this creates a layered authority structure. The local registry may process the request, explain requirements and maintain the direct relationship. But the rules are regional. Future APNIC policies can bind the NIR relationship. Operational procedures can require second opinion, database coordination, service standards or transfer steps. Termination and dispute surfaces sit in the contract chain.
This structure is efficient if everyone understands which decision belongs to which layer. It is frustrating if the local office appears to control a decision that is actually constrained by APNIC policy, or if APNIC appears to control a service problem that is actually local. A downstream operator needs to know whether a refusal, delay or condition came from APNIC policy, NIR interpretation, local contract terms, missing documentation or ledger procedure.
The model agreement is useful because it defines cooperation and obligations. It is limited because a model agreement is not proof that every executed agreement is identical, current or effective in practice. The public record used here does not provide executed current agreements for all seven NIRs and does not compare clauses across them. Without that comparison, the public cannot know how much variation sits beneath the common model.
This is where decentralised service and centralised governance can coexist. APNIC can retain policy authority while NIRs deliver service. That may be the right design for a unique regional ledger. It still leaves the downstream operator navigating two institutions. The practical question is whether the two layers produce clearer service or more friction.
Fees can become harder to see
The third surface is cost. Direct APNIC membership has a fee relationship with APNIC. NIR intermediation can introduce a different fee path: the NIR pays APNIC as a member, and downstream organisations may pay the NIR under local arrangements. Local currency and local invoicing can be helpful. They can also make total cost harder to compare.
The evidence available here does not include comparable downstream fee schedules for all seven NIRs. It therefore cannot say which NIRs are cheaper or more expensive for similarly situated organisations. It cannot say whether local service saves more than it costs. It cannot say whether fees discourage direct membership, NIR membership or transfer. It cannot say whether small operators face higher effective burdens through one path.
APNIC's 2022 explanation of NIR structure, annual fees and voting entitlement offers one important hypothetical. It compares one extra-large NIR receiving 64 APNIC votes with 500 direct very-small members receiving 1,000 votes, alongside materially different aggregate APNIC fees. The example is not observed turnout. It does not include local NIR fees. It does not prove actual representation or actual price burden. It does reveal the mechanism: aggregation can reduce APNIC voting entitlement compared with many direct memberships, while fee incidence is distributed through a different structure.
That mechanism is the heart of the bargain. The NIR can make service easier for many downstream organisations. But those organisations do not necessarily appear as direct APNIC voters. Their local fees may not be visible in APNIC's aggregate comparison. Their service conditions may be set partly by local policy or contract. The APNIC fee paid by the NIR is only one part of the downstream cost story.
A serious fee audit would compare total annual cost by member size and resource profile under direct APNIC membership and under each NIR. It would include local NIR fees, APNIC fees, transfer fees, administrative documentation cost, payment friction, and cost of using complaint or appeal routes. Without that audit, the fee bargain remains asserted rather than measured.
The cautious conclusion is that NIRs can lower transaction costs through local service, but the public record here does not prove that they lower total cost for every class of downstream organisation.
Voting aggregation is a governance cost
The fourth surface is voting. This is where the NIR bargain becomes most visibly asymmetric.
APNIC's 2022 hypothetical gives the cleanest illustration. One extra-large NIR would receive 64 APNIC votes. If the 500 downstream organisations were direct very-small APNIC members, they would receive 1,000 votes. This comparison must be handled precisely. It is a hypothetical. It is not observed election turnout. It does not show how the downstream organisations would vote. It does not include local NIR governance or local fee structure. It does not prove that 500 direct members would participate.
Even with those limits, the example matters because it shows how aggregation changes formal voting entitlement. Many downstream organisations can be represented in APNIC through one NIR member position. That may be administratively efficient. It also means that service decentralisation can coexist with regional voting centralisation.
The difference between voice and vote matters. APNIC policy discussions may be open to nonmembers. Downstream organisations may be able to participate in policy discussion, attend meetings or comment through community channels. That is voice. Voting entitlement inside APNIC's member structure is different. Contract rights are different again. A downstream organisation that receives service from an NIR may have local membership rights, APNIC policy voice and no direct APNIC member vote. Those categories should not be merged.
The governance cost is not that every downstream organisation must vote directly. Direct voting by all downstream entities may be impractical or undesirable. The cost is that aggregation can make the regional institution appear broader in service reach than in direct control rights. A downstream operator may depend on APNIC's ledger and policies, pay through an NIR, receive local service, and still not hold the same APNIC vote as a direct member.
This does not make the NIR model illegitimate. It makes the model a trade-off. Local service and voting aggregation should be presented together. If APNIC or an NIR argues that the model improves access, it should also explain the formal voting effect and how local governance compensates for it.
The missing evidence is actual election participation attributable to direct members, NIRs and downstream communities. Without that, the 2022 example remains a structural warning rather than a turnout finding.
Exit is the test of the bargain
The fifth surface is exit. A layered system can be disciplined if organisations can move when service is poor, fees are high or representation is inadequate. APNIC's current materials indicate that membership changes and multiple transfer scenarios exist, including APNIC-to-NIR, NIR-to-NIR and inter-RIR transfer settings. That is important. Procedure existence means the model is not necessarily a closed trap.
But procedure existence is not measured portability. A transfer route disciplines service only if it is timely, affordable, predictable and not burdened by conditions that make exit unrealistic. A direct-membership option disciplines an NIR only if downstream organisations know it exists, can qualify for it, can afford it and can move resources without unacceptable service risk. Inter-RIR transfer options discipline the system only if they are relevant to the operator's actual resource and business position.
The current policy allows one NIR per economy and in most cases a choice between APNIC and the local NIR, but not resources from both. That choice can simplify ledger management. It can also make exit consequential. If an organisation cannot use both channels, switching is not merely adding a service provider. It may require changing its registry relationship, contracts, fees, documentation, and resource management path.
The moratorium matters here. Seven NIRs are currently recognised. The moratorium began on 27 February 2012 and was made permanent on 26 February 2024. That means the model is not open to new NIR entrants as an ordinary competitive response. Existing NIRs are part of an inherited structure. The rule may be justified. The fixed evidence here does not provide the reasons and vote records for making the moratorium permanent. It therefore cannot assess the rationale. It can say that the permanent moratorium raises the importance of measuring the performance of the seven existing NIRs.
Exit is the practical test because it turns rights into discipline. If a downstream organisation can leave a poor local arrangement and become a direct APNIC member without excessive cost, the NIR has an incentive to serve well. If transfers are slow, expensive or uncertain, the NIR layer becomes harder to discipline.
The evidence needed is concrete: transfer completion times, rejection reasons, fees, documentation burdens, service interruptions, complaint outcomes and the number of organisations that moved from NIR to direct APNIC membership or from direct membership to NIR service. The public record used here does not supply those denominators.
One NIR per economy changes the competitive logic
The one-per-economy rule should be treated as a design choice with consequences, not as a neutral administrative detail. If one NIR is recognised for an economy, the local intermediary role is not normally contestable inside that economy. That can simplify coordination. It prevents overlapping local registries from creating confusion over policy explanation, database responsibility and member service. It can also concentrate local gatekeeping power.
A downstream organisation may still have a choice between APNIC and the local NIR in most cases, but the rule that it cannot obtain resources from both channels changes the meaning of that choice. It is not like choosing two support vendors. It is choosing a registry relationship. The operator must weigh language, currency, local support, APNIC voting rights, local fees, direct regional contact, transfer cost and future administrative flexibility. The more resource management is embedded in a given channel, the more costly switching becomes.
That is why the NIR bargain cannot be judged only by whether a choice exists on paper. Choice disciplines power when it is usable. If a small operator can compare the local NIR and direct APNIC membership, understand both fee structures, forecast transfer cost and move without service disruption, the choice is meaningful. If comparison is difficult, if local fees are opaque, if transfer timing is uncertain, or if direct membership is administratively burdensome, the choice becomes formal rather than practical.
The one-per-economy rule also affects representation. If many downstream organisations use the local NIR, their relationship to APNIC is aggregated through one member. Local governance may be strong, weak or variable. The fixed evidence here does not show downstream participation in each NIR's own governance. It does not show whether local members can influence how the NIR votes, how it interprets policy, or how it raises issues with APNIC. Without that evidence, APNIC cannot treat NIR service reach as equivalent to downstream APNIC control.
The rule may still be justified. Multiple NIRs in one economy could fragment the service, confuse applicants and create disputes over local responsibility. A single recognised NIR can build expertise and stable relationships. The point is not that one-per-economy is wrong. The point is that exclusivity increases the need for measurable accountability.
The permanent moratorium freezes the inherited bargain
The moratorium history changes the analysis from design theory to institutional inheritance. Seven NIRs are currently recognised. The moratorium began on 27 February 2012 and was made permanent on 26 February 2024. That means the NIR system is no longer an open path for new national registries to emerge under ordinary conditions. The existing NIRs are not simply examples in a growing model. They are the recognised set under a frozen design.
A moratorium can be sensible. APNIC may have concluded that new NIRs would add complexity, reduce consistency, raise costs or create governance risk. The fixed evidence here does not provide the reasons and vote records for the 2024 permanence decision, so it cannot judge the rationale. It can identify the consequence: when entry is closed, performance monitoring matters more.
If no new NIR can ordinarily be recognised, then poor service by an existing NIR cannot be answered by local institutional competition. The disciplinary tools become APNIC contract enforcement, local governance, direct APNIC membership, transfer options, complaint routes and ultimately termination or policy change. Each tool has a cost. Each requires evidence.
The permanent moratorium also affects economies without NIRs. Their operators may use APNIC directly or through other arrangements, but they cannot expect a new NIR as the normal local-service solution. That may be acceptable if direct APNIC service and remote support are strong. It is less acceptable if language, fee or time-zone barriers persist. A frozen NIR map should therefore be evaluated together with direct-service quality in non-NIR economies.
For the seven existing NIR economies, permanence can create stability. It can reassure local communities that their registry channel will not be disrupted. It can also entrench old bargains that were made under earlier market conditions. The Internet markets in those economies may have changed. Downstream member expectations may have changed. Resource scarcity, IPv4 transfers, IPv6 deployment, abuse contact expectations and corporate compliance burdens may have changed. A moratorium made permanent in 2024 should be supported by contemporary performance evidence, not only historical familiarity.
The archival test is therefore not optional. The public needs the record explaining why the moratorium became permanent, what alternatives were considered, what evidence was reviewed, who supported or opposed the change, and how APNIC will monitor the seven-NIR structure after closing ordinary entry. Without that record, permanence looks administratively tidy but democratically thin.
Procedure text is not service evidence
APNIC's operational policy for NIRs, updated on 20 July 2023, is important because it gives structure to requests, second-opinion mechanisms, database duties, service-level expectations and membership-transfer procedures. It also acknowledges disproportionate administrative complexity. That is a serious institutional signal. The policy recognises that the NIR layer requires coordination, not merely goodwill.
But procedure text is not outcome evidence. A second-opinion rule can be well designed and rarely used. It can be used quickly or slowly. It can correct local inconsistency or merely add delay. A service-level expectation can discipline performance if measured and enforced. It can become decorative if no results are published. A transfer procedure can make portability real if completion is timely and affordable. It can be a formal path that few organisations use because it is too costly or uncertain.
The missing evidence is case-level and comparative. How many requests required second opinion by year and by NIR? How long did they take? How many were approved, modified, returned or refused? How often did APNIC intervene in database or service issues? How many downstream organisations changed membership path? How many complaints were filed? How many were resolved in favour of the downstream organisation? How often were remedies used by small operators rather than only larger entities?
Without these denominators, public analysis can only say that APNIC has procedures. It cannot say the procedures work. It cannot compare seven NIRs. It cannot identify whether one local model produces better service than another. It cannot show whether direct APNIC membership is a credible alternative or a theoretical one.
This distinction is especially important because NIR users may face two layers of explanation. A local registry may point to APNIC policy. APNIC may point to local implementation. A downstream operator may not know which layer caused the problem. Published case data would reduce that ambiguity. It would show whether friction arises from regional policy, local interpretation, documentation burden, fee structure or transfer mechanics.
The policy text creates the right questions. The outcome record must answer them.
Local incorporation solves proximity, not regional accountability
The fact that NIRs are separately incorporated should be given full weight. Local incorporation can create a domestic counterparty. It can allow contracts, employment, bank accounts, local-language service, domestic dispute routes and local accountability. It can give operators an institution that understands their market better than a regional secretariat can. That is a genuine decentralising feature.
But local incorporation does not automatically answer APNIC-level governance. A downstream organisation may have rights against the NIR under local arrangements. It may have policy voice in APNIC forums. It may not have a direct APNIC member vote. It may not have a direct contract with APNIC. It may not control how the NIR uses its APNIC voting entitlement. Local accountability and regional accountability are different surfaces.
Local incorporation can also make comparison harder. Each NIR may have different local legal obligations, membership categories, fee designs, service practices and governance structures. That variation can be good because it reflects local conditions. It can also conceal differences in cost and rights. The public record used here does not provide executed current agreements and local governance documents for all seven NIRs in a comparable form.
A serious rights comparison would ask: What rights does a downstream organisation have inside the NIR? Can it vote locally? Can it appeal service decisions? Can it inspect fees? Can it influence the NIR's APNIC vote? Can it leave for direct APNIC membership without losing service continuity? Can it transfer resources to another path? Are local rights stronger or weaker than direct APNIC membership rights for similar organisations?
Those questions do not accuse NIRs of failure. They acknowledge that a local institution can be closer and still less transparent at the regional layer. The NIR bargain works best when local accountability and regional accountability reinforce each other. It is weakest when each layer assumes the other provides the remedy.
The public record is sufficient to describe the architecture. It is not sufficient to rank the seven NIRs by rights, cost or service. That is the unresolved part, and it is precisely where operator confidence should be tested.
Open policy voice is not the same as a member vote
APNIC's policy environment can remain open while the NIR structure still changes formal control. This is a common source of confusion. A downstream organisation may be able to read policy discussions, attend a meeting, submit a comment or persuade others. That is meaningful. It can affect rules and should not be dismissed. But open voice is not the same as APNIC membership voting entitlement, contract rights, fee accountability or a direct remedy.
The distinction matters because NIR defenders can point to open participation while critics point to vote aggregation. Both can be describing real features. A downstream operator may have policy voice but not direct APNIC votes. It may have local NIR membership rights but not a direct APNIC contract. It may be affected by APNIC policy while relying on the NIR to implement it. It may be able to complain locally but not easily move if local service is weak.
The 2022 voting hypothetical is therefore not answered by saying that policy discussion is open. The hypothetical concerns formal APNIC votes: 64 for one extra-large NIR versus 1,000 for 500 direct very-small members. Open discussion may soften the problem if downstream organisations actually participate and influence decisions. It does not erase the difference in formal voting entitlement.
Nor should formal voting be treated as the only form of control. If downstream NIR communities have strong local governance, transparent fee setting, regular consultation, and clear ways to instruct or discipline the NIR's APNIC role, aggregation may be legitimate. If local governance is weak or opaque, aggregation becomes more serious. The missing evidence is therefore local as well as regional.
The proper test is layered. At the APNIC layer, measure voting entitlement and actual participation. At the NIR layer, measure downstream membership rights, local elections or consultations, fee transparency and complaint routes. At the portability layer, measure whether dissatisfied organisations can leave. Only all three layers together can show whether open policy voice compensates for aggregated formal votes.
This layered view also prevents a false choice between APNIC and the NIRs. The issue is not whether regional policy should abolish local service or whether local registries should escape regional discipline. The issue is whether the two-layer design can prove that each layer checks the other. Local service should make APNIC more reachable. Regional policy should prevent local fragmentation. Portability should discipline both. If any one of those checks is missing, the bargain tilts toward gatekeeping.
A compact rights matrix
The downstream organisation has local service access through the NIR, possible policy voice through APNIC community channels, local contract or membership rights depending on the NIR arrangement, and transfer or direct-membership options depending on policy and eligibility. It may not have a direct APNIC vote if it is served through the NIR rather than direct membership.
The NIR has APNIC membership, contractual obligations to APNIC, local incorporation, downstream service relationships, and a voting entitlement inside APNIC calculated as an APNIC member. It must follow regional policy and ledger procedures. It may have local accountability to its own members, but that accountability is not the same as direct APNIC voting by downstream organisations.
APNIC has regional policy authority, ledger control, contractual authority over the NIR relationship, recognition criteria, operational policy, transfer procedures and member governance. It does not operate the NIR as a branch office. It also does not directly govern every downstream operator's local service experience unless the relevant policy or contract surface reaches that issue.
This matrix shows why the NIR bargain cannot be judged by one label. It is decentralised at the service counter. It is regionalised at the policy and ledger layer. It is aggregated at the voting layer. It is uncertain at the total-fee and portability layers unless outcome data are published.
The finding: useful intermediation, unresolved discipline
The ranked finding is as follows.
First, the strongest case for NIRs is service access. Local language, local currency, domestic knowledge and separate local incorporation can materially lower transaction costs for operators that would otherwise struggle with direct regional membership. NIRs are not APNIC branch offices, and the local-service benefit should be taken seriously.
Second, the model is structurally asymmetric. Regional policy and ledger authority remain with APNIC; downstream organisations may be mediated through the NIR; formal APNIC votes can be aggregated; and total cost cannot be assessed from APNIC fee figures alone. The 2022 hypothetical of 64 APNIC votes for one extra-large NIR versus 1,000 votes for 500 direct very-small members shows the voting mechanism, not observed behaviour.
Third, the public evidence is weakest on discipline. Direct-membership and transfer routes exist, but their practical cost, timing and success rates are not shown across all seven NIRs. The permanent moratorium makes measurement more important because new NIR entry is no longer the ordinary answer to poor service.
The portability remedy is to publish and enforce a clear NIR exit standard: maximum transfer time, complete fee disclosure, plain-language eligibility, no unnecessary service interruption, and a written reason for any refusal or condition. A right to leave must be observable, not merely described.
The seven-NIR outcome measurement should report, for each NIR and each year, downstream members, fee schedules, average request time, second-opinion cases, refusals, appeals, complaints, transfers into and out of the NIR path, direct-membership switches, APNIC voting entitlement, actual election participation and policy participation. The data should distinguish local NIR rights from APNIC rights.
The archival test is a public executed-record file: current agreements for all seven NIRs, clause-by-clause comparison, local fee tables, service-level reports, transfer records, complaint outcomes and the documented reasons and vote records behind the 2024 permanent moratorium. That file could show that the bargain is well disciplined. Without it, the model remains plausible and useful, but not fully resolved for operators who must rely on it.