Inter-RIR transfers look technical until the moment they matter. A block recorded in one regional registry is to be recognised in another. The seller signs, the buyer pays, the records are checked, and, in the tidy version, the internet continues without noticing. That description is useful only in the way a customs form is useful: it names the paperwork while hiding the political economy behind it. Since IPv4 exhaustion, a cross-registry transfer is no longer just a clerical movement of a prefix. It is a settlement between regional ledgers that lived through scarcity differently, wrote different rules for need, developed different levels of comfort with markets, and now have to decide whether a private bargain made across their borders deserves public recognition.
APNIC sits in the middle of that settlement problem. The Asia-Pacific region is not a single address economy. It contains giant mobile markets, cloud regions, carrier groups, data-centre corridors, government networks, national registries, island economies, fast-growing broadband providers and small access networks whose requirements are modest in volume but severe in consequence. For some holders, IPv4 has become a balance-sheet asset. For others, it remains a working input without which customers cannot be served cleanly. For the registry system, it is both a scarce commodity and a public co-ordination record. Inter-RIR transfers expose the strain between those roles more sharply than domestic transfers do. Within one registry, a community can argue over its own bargain. Across registries, every rule becomes a condition imposed on people who did not necessarily vote for it, budget for it or even recognise its institutional assumptions.
That is why the subject should not be reduced to brokers, prices, escrow accounts or sanctions checks, although all of those are real. Nor is it chiefly about APNIC's internal transfer market design, which is a separate question of how a regional market discovers supply, reviews recipients and records movement within one policy community. The cross-registry question is different. It asks how far one ledger should have to accommodate another ledger's philosophy before a transfer is final. It asks when due diligence protects the integrity of the numbering system and when it becomes a regional tariff. It asks whether need tests, holding periods, legacy-resource conventions, national registry procedures and compliance screening are settlement safeguards or instruments of protection.
The practical answer matters because IPv4 transfers now perform some of the economic work once done by new allocations. They are how a late entrant obtains addresses after the free pool has gone. If cross-registry recognition is predictable, the market can be expensive without being arbitrary. If recognition is slow, opaque or philosophically overloaded, the market adds a second scarcity: not just the scarcity of addresses, but the scarcity of reliable recognition. That scarcity is distributed unevenly. It falls hardest on the buyer that lacks a specialist broker, the regional operator whose corporate documents are ordinary at home but unfamiliar abroad, the NIR-mediated network that must align local and regional records, and the small provider that cannot keep capital trapped in a conditional transaction for months.
The economics of inter-RIR transfer politics therefore begins with a simple observation. An IPv4 address block is globally routable, but registry recognition is institutionally local. The prefix may not change its technical nature when it crosses from one RIR to another. The conditions under which it becomes trusted, billable, contactable, certifiable and operationally clean do change. A transfer is the negotiated replacement of one recognition context with another. The price of that replacement is now part of the price of IPv4.
Scarcity turned a technical record into a settlement
The pre-exhaustion registry system was not innocent, but it was easier to narrate. Registries received address space, evaluated requests and made allocations or assignments according to policies that linked addresses to documented use. The central problem was conservation under growth. Mistakes and privileges accumulated, especially in early allocation history, but a network with a new requirement could still imagine an administrative route to fresh space. Scarcity has changed the institutional meaning of the registry record. Where a new allocation once expanded the ledger, a transfer now reallocates a position inside a ledger that has become economically valuable.
This is why sale language and registry language sit uneasily together. Buyers and sellers speak of purchase, price, escrow, closing, warranty and title. Registries speak of holders, contacts, authorisation, policy eligibility, demonstrated need, resource certification and database updates. Neither vocabulary is simply false. The seller really may receive a market price. The buyer really may treat the prefix as a purchased input. But the buyer does not receive the useful thing merely by paying. It receives the useful thing when the relevant registry, and the operational world that relies on that registry, recognises it as the holder.
In a domestic transfer, that recognition occurs inside one institutional frame. The registry checks the source holder, reviews the recipient, updates the public registration data, adjusts reverse DNS delegation, supports RPKI transitions where relevant and records the new service relationship. Participants may dislike the rules, but the rulebook is single. Inter-RIR transfers multiply the rulebooks. The source registry must be willing to release the block. The recipient registry must be willing to receive it. Each may have views on the seller's authority, the buyer's eligibility, the resource's transfer history, the age of the block in its current account, the kind of recipient need that counts, the legal status of the parties, and the technical state of the records. The transaction closes only when these views can be made compatible enough for both ledgers to move without contradiction.
The word "settlement" is useful because it shifts attention from ownership mythology to finality. A buyer can sign a contract and still lack a reliable position if the source registry has not released the resource or the recipient registry has not recorded it. A seller can accept funds and still face trouble if the transfer authority is disputed. A broker can match supply and demand but cannot make a prefix clean if the records show unresolved history. Settlement, in this sense, is not just payment. It is the point at which private agreement, registry recognition and operational continuity line up.
Financial analogies should not be pushed too far, but they illuminate the structure. A securities trade is not secure simply because buyer and seller agree on price; entitlement, clearing and custody have to align. A bank transfer is not complete simply because two parties want money to move; the relevant ledgers must change under valid authority. IPv4 is not money or equity. Yet its market now depends on comparable institutional finality. The registry record is the infrastructure that turns a bargain into a usable position. In cross-registry transfers, there is no single bookkeeper, so compatibility becomes the clearing mechanism.
This makes inter-RIR politics more than an administrative inconvenience. A registry rule can alter who can close, how quickly they can close and at what all-in cost. A need requirement can export one region's view of stewardship into another region's market. A document convention can favour large corporate buyers over smaller networks. A slow review can make an otherwise fair price unaffordable. A cautious fraud check can protect everyone, while an unbounded one can become a discretionary veto. The difficulty is that all these things can use the same institutional language. The task is to distinguish the safeguards that make settlement reliable from the preferences that make settlement unequal.
APNIC's awkward geography of scarcity
APNIC is a hard case because the Asia-Pacific region did not become address-scarce in a tidy sequence. Some networks received space early, when the internet was smaller and the administrative culture was more generous. Others expanded after conservation had become strict. Some economies built large consumer internet markets before IPv6 deployment could remove the operational need for IPv4. Others still have small networks whose aggregate demand is not spectacular but whose local dependence is intense. The region contains both buyers with professional transaction teams and operators for whom an inter-RIR transfer may be the first expensive institutional transaction they have ever attempted.
This diversity makes APNIC a destination, a source and a translator. Blocks may move into the region for broadband expansion, hosting, mobile services, cloud edge deployment, enterprise connectivity, payments infrastructure or public-sector networks. Blocks may also leave the region when holders monetise unused or underused space. Some parties deal directly with APNIC. Others sit in national internet registry environments, where local records, local language, local company law and APNIC's regional recognition have to work together. Foreign registries and sellers may understand the direct APNIC path better than the NIR-mediated path; domestic operators may understand the local path better than the foreign one. The result is not a single market interface but a layered recognition system.
That layering matters economically. A resource held directly under APNIC may follow one sequence. A resource held through an NIR-linked environment may require local confirmation before regional and foreign recognition can align. A recipient that is ordinary in its home setting may need to show the same facts in a form legible to a foreign registry: incorporation status, signatory authority, corporate continuity, operational plans, beneficial control or network use. A seller in another region may be accustomed to its own transfer conventions and may hesitate when faced with APNIC-region structures it has not seen before. None of these frictions is necessarily irrational. Together they become path dependence.
Path dependence is not a minor administrative inconvenience. It changes price and access. If a buyer's route to recognition is familiar, predictable and quick, the seller can treat the buyer as a credible counterparty. If the route is unfamiliar, the seller may demand a higher price, insist on more protective escrow terms, refuse to tie up the block, or simply choose another buyer. The buyer then faces a thinner market. This is how institutional variety becomes market segmentation. The same /20 is not economically identical in every corridor if the probability and timing of recognition differ.
APNIC has to manage that reality without pretending that the region can be simplified. NIRs are not a defect to be wished away. They can lower local participation costs and provide service in languages and legal environments that a regional body alone might not handle well. But APNIC's transfer politics must make the NIR dimension intelligible to outsiders. If a local registry's confirmation answers a question of authority, that should be clear. If APNIC must also review a matter, the sequence should be clear. If a foreign registry needs assurance that the local and regional records match, APNIC should be able to supply that bridge. Otherwise, useful local institutions get priced as uncertainty.
The Asia-Pacific scarcity story also gives the subject a political edge. Late entrants in the region are often asked to buy from an address distribution created by earlier internet history. They must pay for resources that others received administratively under a different era's assumptions. That may be unavoidable after exhaustion, but it makes additional procedural burdens harder to justify. A registry cannot undo early allocation history. It can decide whether today's recognition rules compound its inequities.
Recognition finality is the product being bought
The buyer in an inter-RIR transfer is not really buying a string of numbers. It is buying a recognised and operationally usable claim. Recognition finality is the moment when the new holder can rely on the record without fearing that another valid claimant, an unfinished source process, a registry disagreement or a broken technical dependency will disturb the position. In practice, finality means more than a line in a database. It means that the recorded holder is correct, the contacts are reachable, reverse DNS delegation can be controlled, RPKI status can be made consistent with intended routing, and upstream networks have no special reason to treat the prefix as suspect.
This is why doubt has a price. Sellers prefer buyers who can close without policy surprises. Buyers prefer blocks with clean history, a clear holder and a predictable source registry. Brokers prefer corridors where previous transactions have established timing and evidentiary patterns. Escrow providers build payment release around registry milestones because neither side wants the transaction to move ahead of recognition. If doubt rises, the spread widens. A seller may demand compensation for time and risk. A buyer may discount the block, require indemnities, seek staged release or avoid the corridor. The market does not wait for a formal policy paper to price institutional uncertainty. It prices it through behaviour.
Finality is also asymmetric. A global cloud platform can divide purchases across sellers and regions, keep specialised counsel, maintain relationships with brokers and absorb delay as a treasury problem. A small access provider may need a modest block to connect customers in a particular quarter. For the large buyer, a slow review is an inconvenience. For the smaller buyer, it may be the difference between accepting new customers and turning them away. A rule that appears neutral in policy text can therefore produce an unequal economic result. The cost of doubt follows bargaining power.
APNIC's cross-registry posture should treat finality as a policy objective in its own right. The question is not whether the registry can ask for more evidence. It always can. The question is whether the evidence requested materially improves settlement integrity. Authority checks are essential. Chain-of-recognition questions can be essential, particularly where legacy space, dissolved companies, acquisitions, receiverships or disputed representatives are involved. Fraud screening is essential. Legal prohibitions must be respected. But an indefinite process does not become fair merely because it is attached to a legitimate category of risk. A registry has to show the connection between the doubt it creates and the integrity it protects.
Clear closing conditions are therefore not customer-service niceties. They are economic infrastructure. Parties should know which documents are required, which eligibility tests apply, how long review normally takes, which events pause the clock, how source and recipient registries communicate, and what can be done if one side rejects or stalls. Without such conditions, the market still functions, but it adds an uncertainty premium. That premium is not just a private transaction cost. It affects who can enter the network economy after exhaustion.
Recognition finality also disciplines the role of registries. A registry is not a mere clerk, because a wrong update can legitimise a disputed transfer, create duplicate claims or break operational trust. But it is not a sovereign allocator of moral desert either, because a valid transfer can be obstructed by excessive discretion. The institution's proper role is to make a safe settlement final. The more it can say in advance what safe means, the less room there is for politics to hide in process.
The compatibility problem that policy language hides
Compatibility sounds neutral. It suggests that two systems either fit or do not fit. In inter-RIR transfers, however, compatibility is often the name given to a negotiated political choice. Two registries can have different internal policies and still settle transfers if they agree on the facts that must be verified. They can also appear compatible while one side quietly exports more of its policy than the other. The key question is not whether policies look the same. It is whether the differences create a real settlement risk or merely offend a regional preference.
The core compatibility questions are practical. Is the source holder authorised to transfer the resource? Is the block eligible to move, or is it subject to a local restriction, a recent-transfer bar, a dispute, a court order or a service condition that prevents release? Is the recipient a valid organisation or network party that can be recorded and contacted? Does the recipient meet whatever recipient eligibility test the receiving registry lawfully applies? Will the transfer create conflicting records, broken reverse DNS, invalid certification, stale contact data or confusion over who should be contacted for abuse and operations? These questions go to settlement integrity.
Other questions are less central. Does the buyer's business model satisfy another region's theory of proper address use? Does a foreign registry approve of the price? Does a source community dislike the idea that addresses might leave its region? Does a recipient's local corporate form look unfamiliar even though it proves the required fact? Does a historical allocation principle deserve to follow the block forever? These may be important to communities, but they are not automatically compatibility requirements. They become legitimate cross-border conditions only when they can be tied to a concrete settlement risk.
Policy language can blur that distinction. "Stewardship" can mean accurate records and accountable use. It can also mean reluctance to let a market reallocate resources away from incumbents. "Need" can mean evidence that the recipient has a real operational requirement. It can also mean a demand that a foreign buyer conform to a documentary culture designed for domestic allocations. "Due diligence" can mean checking authority and fraud. It can also mean a rolling search for reasons not to recognise a politically uncomfortable transfer. The words do not settle the question. Proportionality does.
For APNIC, compatibility should be framed as a set of settlement tests rather than a moral referendum on markets. A transfer into or out of the region should be acceptable when the source holder is legitimate, the recipient can be recorded, local transfer restrictions are not being evaded, the relevant ledgers can move without contradiction, and operational dependencies can be preserved. That does not require APNIC to ignore other registries' rules. It requires APNIC to identify which rules are necessary for shared finality and which are external costs imposed by another policy community.
This distinction matters for accountability. If APNIC-region buyers face burdens because a source registry imposes a need test, APNIC should not describe the entire burden as generic compatibility. If APNIC itself requires a step, it should say why. If an NIR confirmation is necessary, the role of that confirmation should be visible. If a foreign registry has the practical veto, the buyer should know where the veto sits. Hidden responsibility is a subsidy to discretion. Visible responsibility is the first condition of reform.
Need tests and constitutional export
The need test is the most politically charged instrument in inter-RIR transfers because it carries the old allocation ethic into the market era. Its defenders see it as the bridge between stewardship and transfer. If addresses were originally allocated for use, not speculation, a transfer should not become pure asset accumulation. Its critics see it as a relic that imposes paperwork on buyers already disciplined by price and by the operational cost of holding scarce IPv4. Both positions contain a truth. The conflict begins when one region's answer becomes a condition for another region's participants.
In a domestic setting, a need test is a community's choice about its own registry. In a cross-registry setting, it can become constitutional export. A buyer in the Asia-Pacific region may need to satisfy a foreign conception of need because the source registry will not release a block without it. An APNIC-region seller may find that a foreign recipient's registry applies its own recipient standard, shaping the sale even though the source community chose a different balance. The policy no longer governs only the community that adopted it. It travels with recognition.
This export can be legitimate when it prevents sham transactions, hoarding through shell structures or evasion of explicit transfer rules. A recipient that cannot explain any operational use may present a genuine risk if the policy community has decided that transfers must connect to deployment. But need tests become suspect when they function as a moral veto over another region's scarcity economy. A small operator's demand may be real but hard to express in the format expected by a foreign registry. A network may need IPv4 for transition, customer compatibility or commercial reachability even while deploying IPv6. A provider in a fast-growing market may have operational evidence that is informal, local or customer-driven rather than wrapped in the forecasts and procurement records expected by larger firms.
The distributional asymmetry is obvious. Incumbents rarely have to keep proving why they should retain historical holdings. Late entrants must prove why they deserve to buy. That asymmetry may have been easier to defend when a registry allocated from a free pool and needed to ration administrative grants. It is harder to defend in a paid secondary market where the buyer is not taking free addresses from a community pool but acquiring resources from a willing holder. A strict need test can recreate the old hierarchy: early recipients enjoy optionality; later entrants must submit to suspicion.
APNIC should not answer this by turning against all need tests. Nor should it accept them as universal virtue. The proper question is what the test is doing in the transfer. If it confirms that the recipient is a real network party with accountable use, it may support settlement confidence. If it forces the recipient to satisfy a foreign allocation philosophy unrelated to fraud, uniqueness, authority or continuity, it is a political burden. The test should be judged by transfer relevance, not by institutional nostalgia.
Need-test asymmetry also affects price. A seller choosing between two buyers will prefer the one with the easier recognition route unless the riskier buyer pays more. A buyer facing uncertain need review will discount the transaction or demand more protection. A broker will steer clients toward corridors with predictable evidence requirements. The result is not a clean moral market in which needy users win. It is a market in which documentary sophistication can substitute for operational need. That is a bad outcome for a region like APNIC, where many real demands arise from growth, transition and customer dependency rather than from the neat corporate planning documents that international transaction markets prefer.
The better standard is modest. Need evidence, where required, should be specific, proportionate and open to functional equivalents. It should not punish a buyer because its local business records do not resemble those of a large North American or European network. It should not convert IPv6 ambition into evidence against current IPv4 requirement. It should not let a registry protect historical allocation privilege by making latecomers prove an impossible purity. In a scarce market, a need test is not an ornament. It is a rationing device. It should be treated with that seriousness.
Reciprocity is not sameness
Inter-RIR transfer policy often speaks in the language of reciprocity. The word can be clarifying if it means mutual recognition of roles: the source registry verifies source-side facts, the recipient registry verifies recipient-side facts, and both co-operate so the ledger moves once and only once. It becomes misleading if it means that every registry must adopt the strictest rule in the chain or pretend that regional policy philosophies are interchangeable. Reciprocity is not sameness. It is disciplined restraint.
Disciplined reciprocity begins with a division of labour. The source registry is best placed to confirm the source holder's status, transfer authority, local restrictions, prior-transfer timing and any known dispute. The recipient registry is best placed to confirm the recipient's eligibility, service relationship, contactability and post-transfer operational setup. Both sides have an interest in fraud control and in preventing double recognition. Neither side should use the transfer to relitigate every historical policy choice that produced the block. The purpose of co-operation is settlement finality, not policy conversion.
The problem is that power is rarely symmetrical. A registry with large available supply, strict release conditions or a cautious institutional culture can shape the terms of cross-border trade beyond its formal region. A recipient registry that wants its members to access that supply may have to live with exported conditions. Conversely, a registry that makes inbound recognition especially easy can pressure others by attracting demand, even if the source side remains wary. Inter-RIR compatibility is therefore a market in institutional leverage as well as a policy arrangement.
APNIC's interest is to avoid both extremes. If it is too permissive, it risks becoming a venue for transactions structured to escape controls elsewhere, damaging confidence in APNIC records and inviting other registries to tighten recognition. If it is too deferential, it lets other regions set the effective price of entry for Asia-Pacific networks. The middle position is not compromise for its own sake. It is a principled claim that shared settlement requires verification, not conversion.
Rule export should therefore be analysed case by case. What risk does the exported rule address? Is that risk present in the cross-registry transaction? Can it be addressed by a less burdensome or more locally appropriate form of evidence? Does the rule make a difference to authority, uniqueness, eligibility, continuity, legal prohibition or fraud? Or does it simply preserve a regional preference about how scarce addresses ought to circulate? If the rule passes the settlement test, APNIC should accommodate it. If it fails, APNIC should identify it as a cost imposed by the other regime rather than absorbing it into neutral terminology.
This distinction is not academic. Once a condition becomes routine, it shapes behaviour without further debate. Brokers advise clients away from difficult corridors. Sellers learn which buyers are slow. Buyers self-select out of transactions they might otherwise need. The rule then acquires the authority of habit. APNIC should be alert to that soft law. The most important inter-RIR politics may occur not in public policy meetings but in the accumulated expectations of brokers, lawyers, sellers and registry staff.
Reciprocity also requires historical humility. No registry's allocation history is a perfect expression of justice. Early internet timing, language, institutional proximity, public-sector involvement and commercial maturity all influenced who received space and who did not. A registry that imposes strict conditions on cross-border buyers while leaving inherited holdings untouched may be protecting integrity. It may also be protecting a windfall. The difference is whether the present rule is necessary for present settlement or convenient for incumbent comfort.
The NIR inheritance
The Asia-Pacific national internet registry model is one of APNIC's defining institutional features. In transfer debates it is sometimes treated as a procedural complication. That is too narrow. NIRs are part of the region's political economy. They make registry service local, mediate language and legal differences, and embed number-resource administration in national internet development paths. They can lower participation costs for domestic networks. They can also make cross-border transfers harder to price because recognition may depend on a chain of local, regional and foreign confirmations.
The difficulty is not that NIRs are irrational. It is that outsiders may not know how to value their confirmations. A local registry may have the best view of a member's standing, authorised contacts and domestic corporate documentation. APNIC may have the regional relationship and the inter-RIR interface. A foreign registry may need to know which confirmation answers which risk. If the roles are not explicit, the market sees ambiguity. Ambiguity becomes delay. Delay becomes price.
Consider a buyer whose ordinary corporate evidence is in a local language and a local legal form. Domestically, the evidence may be routine. In a cross-registry transaction, it may need translation, notarisation, a board resolution, proof of signatory authority, an explanation of corporate succession or a mapping between local membership records and APNIC's regional records. Each request may have a reason. Together they form a fixed cost that does not shrink much when the prefix is small. A buyer seeking a modest block can therefore face a per-address process cost far above that of a large buyer.
The same problem appears for sellers. A seller outside the region may not understand whether an APNIC-region recipient is direct, NIR-mediated or moving between local and regional service states. A seller may worry that funds will be held while local confirmation is sought. A broker may not know how long the combined sequence usually takes. The rational market response is to demand more certainty, more money or a different buyer. The NIR inheritance then affects liquidity even when no one objects to the recipient's actual use.
APNIC can reduce this penalty without weakening NIRs. It can provide clearer guidance on how NIR confirmations fit into inter-RIR transfers, what documents are normally expected, which local equivalents are acceptable, and how foreign registries should understand regional records. It can co-ordinate with NIRs so that members do not receive inconsistent instructions. It can distinguish missing evidence from unfamiliar evidence. It can tell parties where in the sequence they are. These are not glamorous reforms, but in a scarce market they matter more than slogans.
The larger principle is that local institutional diversity should not become a hidden exclusion rule. A transfer system that works smoothly only for direct members with internationally familiar corporate structures is not neutral. It privileges one organisational style. APNIC's challenge is to make the region's institutional pluralism legible enough that it does not carry an unnecessary risk premium. That is a governance task and an economic development task at the same time.
Due diligence, or the price of being believed
Due diligence is where the registry's public function meets the transaction market's suspicion. IPv4 scarcity attracts forged authority documents, compromised accounts, shell companies, disputed legacy claims, insolvency disputes, hijacked prefixes and attempts to sell resources by people who do not control them. Cross-registry movement amplifies these risks because documents come from different legal systems, languages and historical service relationships. A registry that fails to check authority harms not only the immediate buyer or seller. It weakens trust in the ledger that the market depends on.
Serious verification is therefore indispensable. The source holder must be the party entitled to transfer. The representative must have authority. Corporate continuity must be understood where names have changed, mergers occurred or the original holder no longer exists. The block should not be under unresolved dispute or legal restriction. The recipient must be an identifiable party that can be recorded, contacted and held to registry service obligations. The technical record should be capable of moving without producing duplicate claims or operational gaps.
Yet due diligence can also become the respectable language of institutional doubt. The registry can always ask one more question. It can always say that a document is unfamiliar, a forecast is incomplete, a signature needs further validation or a corporate chain requires more proof. Sometimes that caution is justified. Sometimes it reflects a deeper discomfort with cross-border trade, a desire to prevent outflow, or a habit of favouring documentation styles associated with larger buyers. A process does not become fair because it uses the vocabulary of risk.
The economic issue is the price of being believed. Large firms can buy belief. They arrive with counsel, translated records, corporate secretaries, audited accounts, recognisable transaction forms and brokers who know what registry staff expect. Smaller networks often arrive with real demand but less polished evidence. They may have local-language documents, informal customer demand, short planning horizons and limited experience with escrow. If due diligence is not calibrated, it rewards institutional presentation rather than substantive legitimacy.
Calibration means several things. Requests should be tied to a specific risk: authority, identity, eligibility, dispute, continuity, legal prohibition or technical dependency. Deficiency notices should state what fact remains unproved. Functional equivalents should be accepted where they prove the fact, even if they are not the preferred foreign format. Review should be staged so parties can identify fatal problems early rather than after money and time have been committed. Complex cases should receive scrutiny, but ordinary cases should not be trapped in a bespoke investigation simply because they cross a regional boundary.
The distinction between verification and obstruction is not always obvious in a single case. It becomes visible in patterns. Are similar documents accepted from large buyers but questioned from small ones? Do some corridors produce repeated delay without clear explanation? Are certain local corporate forms treated as inherently suspect? Are legacy-resource transfers examined for real chain issues or for unease about market sale? Does a registry ask questions that can be answered, or questions that keep expanding? APNIC's legitimacy in inter-RIR transfers depends on paying attention to these patterns.
Brokers, escrow and the market for institutional certainty
Brokers and escrow providers are often treated as proof that IPv4 has been commercialised. They are that, but they are also evidence that registry finality is uncertain enough to require private risk infrastructure. A broker in an inter-RIR transaction does not merely find supply. It estimates whether the block can actually move. It reads policy corridors, seller history, transfer waiting periods, documentation burdens, recipient eligibility, NIR involvement, possible compliance concerns, RPKI implications, reverse DNS handover and the likelihood that both registries will accept the transaction within a useful time.
That knowledge is valuable because the public process is not always transparent enough for ordinary participants. A competent broker can tell a buyer which sellers are likely to close, which blocks have problematic history, which registries will ask for which evidence, and how escrow milestones should be framed. The broker's fee therefore buys search, negotiation and institutional interpretation. In opaque corridors, the interpretive component grows. Private expertise becomes a toll charged on public uncertainty.
Escrow plays a related role. The seller does not want to release a valuable resource before payment is secure. The buyer does not want funds released before recognition is effective. Escrow turns registry milestones into payment milestones. But escrow cannot remove policy risk; it can only hold money while the risk is resolved. If review takes longer than expected, the buyer loses time and may lose business. If recognition fails, funds may return but the deployment plan does not rewind. If the seller can find a faster buyer, it may avoid the uncertain one from the start.
This infrastructure is useful and should not be romanticised away. A valuable scarce resource needs professional intermediation. But the size of the broker and escrow role tells us something about the state of public settlement. The more opaque the recognition rules, the more buyers must buy expertise. The more uncertain the timing, the more escrow and contract drafting must compensate. The more fragmented the corridors, the more market access depends on knowing the right intermediaries. What looks like a private market service may be the capitalised cost of institutional ambiguity.
APNIC should care because this cost is not neutral. Large buyers can purchase expertise repeatedly and learn from each transaction. Small buyers may rely on a seller's broker, a general lawyer or internal staff making a first attempt. They may not know which questions to ask until the process is already expensive. If the registry's public guidance is thin, market knowledge becomes proprietary. That is not a healthy settlement regime for a resource that remains embedded in a public co-ordination system.
Reducing ambiguity would not eliminate brokers or escrow. It would improve their function. Brokers would compete more on sourcing, negotiation and service rather than on decoding hidden procedural expectations. Escrow would remain a prudent protection rather than a long holding pen for regulatory uncertainty. The market would still price scarcity. It would price less of APNIC's or another registry's avoidable obscurity.
What liquidity pricing really captures
IPv4 price quotes usually compress too much into a single number. A prefix length, a reported transaction range or an informal broker estimate may suggest a market price, but the buyer's actual cost is the price of usable, recognised, routable addresses by the time they are needed. That cost includes the amount paid to the seller, broker fees, escrow fees, legal review, internal staff time, translations, financing cost, delay, failed-deal risk, technical transition and future reputation risk if the block's history is messy. Inter-RIR transfer rules enter almost every part of that calculation.
The first channel is time. A predictable 30-day review and an unpredictable multi-month review are not economically equivalent even if the registry fee is the same. A buyer may need addresses for customer activation, a migration, a data-centre opening or a contract deadline. Delay can force interim leasing, carrier-grade NAT expansion, renumbering, temporary upstream arrangements or delayed revenue. The seller also bears time risk and may demand compensation or prefer another buyer. Liquidity is therefore not just the number of willing sellers. It is the number of sellers whose blocks can be recognised on a schedule that buyers can use.
The second channel is rejection risk. A transaction that may fail requires protective terms. The buyer discounts the bid. The seller resists because the block is tied up while review proceeds. Escrow reduces payment risk but not opportunity cost. A failed transfer can expose a buyer's demand to the market, disrupt internal planning and waste scarce managerial attention. Corridors with higher rejection risk therefore clear at different effective prices, even when headline address scarcity is the same.
The third channel is documentation burden. Authority checks, need evidence, resource-history review, NIR confirmations, corporate succession proof and compliance screening all require effort. Some of that effort is the legitimate cost of a safe market. Some is the friction of institutional mismatch. Fixed costs are especially important. The same categories of proof may apply to a modest transaction and a large one. The large buyer spreads the cost over more addresses and more transactions. The small buyer cannot. Fixed process costs are regressive.
The fourth channel is segmentation. A so-called global IPv4 market is actually a set of recognition corridors. Some corridors are well travelled, familiar to brokers and predictable in review. Others are unusual, culturally harder, document-heavy or exposed to conflicting policy assumptions. Blocks in easy corridors are more liquid. Blocks in difficult corridors require discounts or specialised buyers. Buyers with strong documentation and broker access reach more supply. Buyers without it face a thinner market. Scarcity is global in routing terms but regional in settlement terms.
This is where the subject differs from a simple capital-control story. Liquidity friction is an effect, not the whole frame. The deeper issue is why the friction exists, who can justify it, who benefits from it and whether it protects ledger integrity. A rule that prevents fraud may reduce liquidity but increase trust. A rule that traps supply to protect incumbents may also reduce liquidity, but for a very different reason. Both show up in price. Price alone cannot tell us which one we are seeing.
APNIC cannot make IPv4 cheap. The remaining scarcity is real, and private holders will price it. What APNIC can do is reduce the portion of the price that comes from avoidable uncertainty in APNIC-related settlement. It can make criteria clear, publish the role of NIR steps, co-ordinate with other RIRs, distinguish essential checks from inherited preferences and give parties a reasoned basis for decisions. That would not abolish scarcity rents. It would reduce institutional rents.
The small operator's tax
The most unfair burden in inter-RIR transfers is not the existence of cost. A valuable resource moving across ledgers should be checked. The unfairness lies in the way fixed and uncertain costs fall. Smaller and later entrants pay a higher effective price because the transaction system rewards institutional capacity. This is not a sentimental point about small firms. It is a structural point about market access after exhaustion.
Late entrants already pay for history. They did not receive large allocations in the era of abundance. They enter a market where IPv4 must be bought, leased, conserved or substituted away from, even though much of the internet still requires IPv4 reachability. IPv6 deployment is necessary and should be encouraged, but it does not instantly remove the need to serve customers, partners, legacy systems and enterprise contracts through IPv4. A late entrant's demand may be operationally prudent rather than backward-looking.
Small operators then face the fixed-cost problem. A due-diligence checklist does not shrink neatly with prefix size. The buyer may still need authority evidence, recipient eligibility, need support, escrow, legal review, account setup, reverse DNS planning and RPKI changes. If the block is modest, the per-address cost rises. If the buyer has never done an inter-RIR transfer, the learning cost rises. If its domestic records are unfamiliar to the source registry, the documentation cost rises again. The market may advertise a per-address price, but the small buyer pays a per-transaction tax.
Bargaining power adds another layer. Sellers want certainty and speed. A small buyer with a complicated recognition route may have to pay more, accept a less favourable contract or lose the block to a larger buyer. Brokers may rationally spend more attention on larger transactions. Escrow and legal costs may be less negotiable. Internal staff may be diverted from operations. The buyer's need may be more urgent than the large buyer's, yet its transaction posture looks weaker.
This matters for the Asia-Pacific region because modest address acquisitions can have meaningful local effects. A small ISP may need public IPv4 to serve new customers while it expands IPv6. A regional hosting provider may need addresses to support customers that cannot yet operate cleanly without them. A business connectivity provider may need enough space to avoid brittle NAT arrangements. If inter-RIR process makes modest acquisitions uneconomic, policy has silently favoured scale.
The remedy is not a weak registry. Fraud does not become acceptable because the buyer is small. The remedy is proportional process. Evidence requirements should be intelligible in advance. Review should identify the actual missing fact rather than issue vague objections. Local-document equivalents should be accepted where they prove the point. Timeframes should be predictable enough for small buyers to plan. Escalation should not require expensive advocacy. The goal is not to subsidise small operators with other people's addresses. It is to avoid making public recognition a luxury good.
Need tests illustrate the problem. A small buyer may have immediate operational need but poor documentary performance. It may not have long forecasts, formal customer letters or elaborate utilisation models. Its evidence may live in network growth, customer tickets, local contracts and operational constraints. A large buyer may be better at producing polished forecasts even when its purchase is more strategic. If registry evaluation prizes format over substance, the transfer system rewards the wrong capability.
APNIC's policy language should therefore remember the late entrant. The region's internet is not complete. A transfer regime that works only for incumbents, large buyers and well-served jurisdictions mistakes historical advantage for institutional competence. Predictable recognition is most valuable precisely where market power is weakest.
Anti-fraud verification and protectionism
The line between verification and protectionism is the central ethical line in inter-RIR transfers. Verification protects the shared ledger from falsehood. Protectionism uses the ledger's authority to keep resources, bargaining power or policy influence inside a preferred region. Both can be described in respectable language. The difference lies in proportionality, evidence and relevance to transfer integrity.
Fraud risks are not imaginary. Scarce IPv4 creates incentives for forged letters, compromised registry credentials, false claims by former employees, shell-company structures, disputed legacy holdings, insolvency manoeuvres and attempts to sell prefixes that have been informally routed but not legitimately controlled. Cross-border transactions add language, legal and corporate-form complexity. A registry that waves through doubtful transfers would damage buyers, sellers and the wider network. Strong anti-fraud controls are a public good.
Protectionism appears when the control is not tied to a specific risk or cannot be satisfied by reasonable evidence. A registry may be reluctant to let address space leave its region because local members feel scarce. It may prefer buyers that resemble its own historical members. It may treat foreign corporate documents as suspicious by default. It may turn a need test into a barrier that protects inherited holdings. It may prolong review without explaining what fact remains unresolved. In each case, the formal claim may be caution, but the economic effect is to discourage movement.
APNIC has reasons to oppose both weak verification and disguised protectionism. Weak verification would expose APNIC-region networks to tainted blocks, disputed claims and operational instability. Protectionism would make APNIC-region demand pay tribute to another region's scarcity politics. The registry's role is to defend the ledger without letting the ledger become a fortress.
The practical test should be direct. What risk is being addressed? What evidence would satisfy the concern? Is the requested evidence proportionate to the transaction and the risk signals? Is the same standard applied across buyer types and regions? Does the requirement protect against false recognition, duplicate claims, legal prohibition, technical breakage or clear policy evasion? Or does it merely make a transfer less attractive because someone dislikes its market consequence? If the answer is the latter, the requirement should be named for what it is.
Transparency is important because protectionism thrives in ambiguity. A denied or stalled party should be able to tell whether the issue is source authority, recipient eligibility, need evidence, timing, dispute, compliance, NIR confirmation or operational continuity. Some details may be sensitive, but the category of concern should not be hidden. A registry that cannot state the curable defect risks turning discretion into policy without public accountability.
This is also in the interest of honest stewardship. The public function of a registry is stronger when it can explain why it says no. A refusal tied to forged authority, unresolved litigation or broken continuity is defensible. A refusal grounded in discomfort with cross-border price movement is not. The difference should not be left to inference.
Operational continuity is political economy
Transfer debates often dwell on eligibility, but operational continuity is where recognition becomes visible to the network. The buyer needs more than a changed holder name. It needs reverse DNS delegation under the right control, accurate contact records, a coherent RPKI state, updated routing-related records where they are used, service-account access and a clean handoff of responsibility. If these pieces move badly, the transfer may be formally recorded but operationally fragile.
RPKI makes the point clearly. A prefix's certification state can influence whether route announcements are treated as valid, invalid or unknown by networks that use origin validation. A transfer that changes the registry context may require certificate and ROA changes so the intended origin is not contradicted. Poor sequencing can create a period of invalidity or uncertainty. This is not merely a technical nuisance; it affects the buyer's ability to use what it has purchased and the seller's ability to exit cleanly.
Reverse DNS is similar. Customers, abuse desks, mail systems and operational tools still rely on delegation and contact coherence. If reverse DNS is stale, if contacts point to the former holder, or if abuse records are not refreshed, the new holder inherits confusion. That confusion can damage reputation, delay deployment and make the block look riskier to counterparties. Again, registry-controlled dependencies are part of economic value.
Inter-RIR transfers complicate continuity because the recognition environment changes. The recipient may use different registry tools, authentication methods, certificate arrangements or contact formats. The source block may have routing history tied to another autonomous system or another region. Third-party databases may lag. Upstreams may ask for proof. Customers may be moved in stages. A commercially closed transfer can still require careful operational settlement before it is truly useful.
This is why APNIC should treat continuity as part of the transfer regime rather than as an afterthought. Parties should know when reverse DNS can change, how RPKI will be handled, what happens to existing certificates or ROAs, which contact records need updating, whether route objects or related records must be refreshed, and how NIR-mediated records move. The point is not for APNIC to run the buyer's network. It is for APNIC-controlled pieces of the handoff to be predictable.
Operational continuity also provides a principled boundary for registry authority. Requirements that prevent duplicate certification, stale contacts, broken reverse DNS or inconsistent public records belong near the centre of compatibility. Requirements that delay a transaction because the registry disapproves of the price, the buyer's business model or another region's policy culture do not. The network cares whether the prefix is legitimately and safely usable. Registry politics should be disciplined by that operational fact.
For APNIC-region operators, this is a practical issue. Customers do not care that a block crossed a regional registry boundary. They care whether services work. A transfer regime that preserves technical continuity while reducing ideological friction is not only fairer; it is more faithful to the internet's dependency structure. The ledger exists so networks can co-ordinate around accurate, usable facts. Continuity is the test of whether it has done so.
Compliance as boundary condition, not core explanation
Cross-border transfers inevitably encounter legal and compliance constraints. Parties may need to consider sanctions, prohibited counterparties, beneficial ownership, court orders, insolvency restrictions, anti-money-laundering expectations around payments and evidence of authority. Registries cannot pretend those constraints do not exist. A transfer involving opaque ownership or a legally restricted party can create serious risk. Compliance is part of the environment in which inter-RIR settlement occurs.
But compliance is not the theory of the system. Most inter-RIR transfer politics are not sanctions cases. They concern compatibility, need tests, documentation, NIR sequencing, waiting periods, legacy-resource treatment, due diligence standards, operational continuity and the distribution of scarcity costs. Treating the whole field as a compliance problem narrows the analysis and expands discretion. It can allow policy preference to borrow the moral force of legal risk.
The distinction matters for APNIC. Where a legal prohibition exists, it must be respected. Where ownership or control is unclear, additional evidence is justified. Where payment structure suggests evasion or fraud, caution is necessary. Where a block's history is disputed, recognition should not be rushed. But where the real issue is that a foreign registry dislikes the recipient's market context, or that a buyer's need does not fit the strictest documentary template, compliance language should not do work that belongs to policy debate.
Precise categories protect both integrity and access. A party should be able to tell whether a delay arises from legal prohibition, ownership uncertainty, source authority, recipient eligibility, need evidence, operational continuity or regional policy mismatch. The remedy differs in each case. A prohibited-party problem may be fatal. A missing authority document may be curable. A reverse DNS sequencing issue may be operational. A need-test disagreement may be political. Lumping them together under compliance makes the process harder to challenge and easier to over-price.
This is why the subject should remain distinct from a sanctions-centred account. Sanctions and legal restrictions are real cross-border constraints, but they are not the main engine of APNIC's inter-RIR transfer politics. The deeper mechanism is institutional compatibility after exhaustion: how ledgers with different scarcity histories decide which private reallocations become publicly recognised. Compliance is one border of that mechanism, not its centre.
How APNIC can reduce cross-registry rents
APNIC cannot abolish scarcity and should not promise frictionless transfers. The registry's job is not to make every private bargain effective. It is to make legitimate transfers safe, predictable and final without using public recognition to shelter incumbents from scarcity. That still leaves a substantial reform agenda, most of it practical rather than ideological.
First, criteria should be legible before contract. Parties should be able to see the categories that determine whether a transfer can be recognised: source authority, recipient eligibility, resource eligibility, transfer-history restrictions, need evidence where applicable, NIR steps, compliance concerns and technical-continuity requirements. The guidance should include acceptable evidence and functional equivalents for different legal systems. It should distinguish a missing fact from an unfamiliar format.
Second, time should be treated as a cost. Complex cases need review, but ordinary cases should not disappear into open-ended correspondence. APNIC and its counterpart registries should identify expected review periods, the events that pause them and the escalation route when another institution has not responded. Time limits need not force approval. They force the registry to recognise that delay is borne by users, not absorbed by abstraction.
Third, decisions should be reasoned. A rejected or stalled transfer should identify the category of deficiency and, where possible, the cure. Some information may be confidential, but the parties should know whether the problem is authority, eligibility, need, timing, dispute, compliance, NIR confirmation or continuity. Reason-giving is not an indulgence. It is how discretion becomes accountable.
Fourth, the audit trail should be strong. Inter-RIR transfers depend on trust between registries and between registries and market participants. Records should show who confirmed source authority, when release was approved or declined, what recipient checks were completed, which operational updates were made and why any exception was granted. An audit trail protects the registry as well as the parties. It also makes later disputes easier to resolve.
Fifth, operational continuity should be integrated into the standard transfer process. Reverse DNS, contacts, RPKI, service access and relevant routing records should be addressed as ordinary settlement elements rather than left to ad hoc repair. Where NIRs are involved, the handoff between local and regional records should be explicit. A buyer should not have to discover after closing that the useful parts of recognition are still scattered.
Sixth, proportionality should guide evidence. The same risk categories may apply to large and small transactions, but the burden should not be mechanically identical when lower-cost evidence proves the same fact. Proportionality is not weaker fraud control. It is a way to avoid making fixed process costs the enemy of modest legitimate demand.
Seventh, APNIC should name external burdens. If a source registry's rule creates a cost for APNIC-region buyers, APNIC can respect the practical requirement while making clear where it comes from. That transparency helps buyers plan, helps the community debate rule export and prevents every cross-registry cost from being misdescribed as neutral compatibility.
These reforms would not make APNIC a market booster at the expense of stewardship. They would define stewardship in a way suited to the exhaustion era. Accurate records, accountable holders, fraud control and operational continuity remain central. What changes is the refusal to let those goods be used as cover for regional privilege. The aim is not laissez-faire. It is a disciplined settlement regime.
The next settlement question
IPv4 scarcity is the immediate setting, but the institutional lesson is broader. The internet will face other scarce, permissioned or trust-dependent resources. Some will involve identifiers. Some will involve routing security, reputation, access to shared co-ordination systems or the credibility of public records. In each case the same temptation will appear: either treat the ledger keeper as if it owns the value, or treat private exchange as if the ledger were a minor administrative detail. Both temptations are wrong.
The ledger is indispensable because it creates reliable recognition. Without it, buyers could not know what they were acquiring, sellers could not exit cleanly, networks could not find responsible contacts and routing security would be harder to interpret. But the ledger's purpose is not to preserve the power of the ledger keeper or to freeze historical distributions. It exists so the network can co-ordinate around accurate facts. Private exchange is also indispensable because exhaustion made reallocation necessary. But exchange cannot override the need for clean authority, unique recognition and operational continuity.
The settlement test is therefore practical. Does the rule protect the ledger from falsehood, duplication, fraud, dispute, legal prohibition or operational breakage? If so, it belongs near the centre of compatibility. Does the rule require another region's buyers or sellers to carry a symbolic burden unrelated to transfer integrity? If so, it deserves challenge. Does the process provide clear criteria, time expectations, reasoned decisions, an audit trail and continuity planning? If not, the market will add a risk premium, and weaker participants will pay first.
For APNIC, the strategic position should be clear. It should defend rigorous verification while resisting arbitrary obstruction. It should co-operate with other RIRs without accepting the export of regional privilege as technical necessity. It should make NIR-mediated paths legible rather than allowing local diversity to become a penalty. It should treat small and late entrants not as exceptions to the market, but as the users for whom predictable recognition matters most. It should remember that compatibility is not the same as conformity.
Inter-RIR transfer politics will not disappear. IPv4 scarcity is too valuable, regional histories are too different and registry communities are too invested in their own settlements. But the politics can be made more honest. The question is not whether APNIC participates in a market; it already does, because exhaustion made transfer recognition part of registry life. The question is whether APNIC helps build a compatibility regime that recognises the market without surrendering the ledger, and protects the ledger without turning the bookkeeper into a gatekeeper.