ARIN is one of the clearest places to study what IPv4 transfer policy has become. It is not merely a set of administrative clauses for moving address blocks from one organisation to another. In North America, where cloud platforms, carriers, data-centre networks, hosting businesses, enterprise buyers, brokers, lawyers and financiers have treated IPv4 as a priced input for many years, transfer policy has become market architecture. It shapes which holders can sell, which buyers can settle, how much diligence is needed, how much information remains hidden, how quickly capital can move, and how much discretion a registry retains after the free pool has largely ceased to matter.
That makes ARIN more interesting than a registry in obvious institutional crisis. A failed or paralysed registry shows what happens when governance collapses. ARIN shows something subtler: how a comparatively mature registry can still carry pre-market assumptions into a post-exhaustion asset market. Its process is public. Its categories are legible. Its staff-facing procedures are more developed than those of many administrative systems. Yet the structure still joins two roles that a high-value market would normally separate. ARIN maintains the ledger that makes address claims reliable. It also acts as a gatekeeper that decides whether a transaction deserves recognition under policy tests inherited from the allocation era.
IPv4 scarcity changed the economic meaning of registry rules. When a registry had a meaningful free pool, the central policy question was distributive: who should receive scarce numbers from a common administrative supply, and on what evidence of use? After exhaustion, the more important question is transactional: when one recognised holder has agreed to move an already allocated, economically valuable resource to another party, what should the registry be allowed to test before the market receives settled recognition? Those are different questions. Treating them as the same is the source of much of the present tension. Conservation language becomes a rationale for capital control. Needs assessment becomes rationing by another name. Registry review becomes a toll on liquidity.
ARIN's own public materials are useful factual exhibits for this analysis. Its transfer guide distinguishes merger and acquisition transfers, specified recipient transfers inside the ARIN region, and inter-RIR transfers. Its Number Resource Policy Manual gives the policy grammar behind those categories. The transfer guide describes source and recipient requests, officer acknowledgement, qualification checks, documentation, fees, agreements and final registry recognition. Those materials show what the registry does. They do not, by themselves, prove that the design is economically neutral, proportionate, or optimal for the operators whose networks and customers depend on the numbers.
The institutional-economics test is more exacting. A market for a scarce operational asset needs title certainty, predictable settlement, low transaction costs, reliable information, credible dispute isolation, clean security-state transition and a public record that counterparties can trust. It also needs a registry layer that knows the difference between recording recognised control and granting permission for capital to move. If the registry preserves uniqueness, prevents double assignment, maintains accurate RDAP and Whois data, supports reverse DNS and RPKI continuity, records transfer history and isolates disputes, it increases the asset value of IPv4. If it uses the same position to judge business plans, moralise speculation, export regional policy preferences, or condition recognition on discretionary forecasts of need, it suppresses liquidity while calling the result stewardship.
That distinction runs through Lu Heng's public notes on number-resource governance, the Number Resource Society's decentralisation argument and LARUS's continuity doctrine. The common claim is not that registries are useless. It is that the registry function must be separated from registry power. Number uniqueness is real. Publication and security continuity are real. The world needs a trusted record of who is recognised for operational purposes. None of that proves that an RIR owns the economic value of the addresses, speaks for every end user in a region, or should decide whether a buyer's future plan is good enough to justify settlement. The useful principle is narrow: protect the ledger, not the gatekeeper.
ARIN's architecture should therefore be judged by what it does to certainty, liquidity, information and risk placement. The North American system has genuine strengths. It recognises specified transfers. It has a mature continuity route for mergers, acquisitions and reorganisations. It publishes transfer categories. It lists compatible inter-RIR transfer partners. It maintains a Qualified Facilitator Program and a visible fee schedule. It gives legacy holders a way to maintain core registry records even when they are not under an ARIN agreement. These features make ARIN more market-facing than a purely allocation-era registry. The same architecture, however, keeps rationing habits alive: needs assessment, wait-list restrictions, policy-compatibility filters, agreement dependence, confidential bilateral requests and broad registry discretion over recognition. The market exists. It exists inside an approval structure.
From free-pool allocation to settlement architecture
Before IPv4 exhaustion, transfer policy could be treated as secondary. The registry's main economic role was allocation. Applicants needed addresses; the registry had a pool; policy decided how to distribute that pool. Need-based allocation, utilisation thresholds and conservation language made intuitive sense in that environment because the registry was issuing unallocated administrative supply. Those rules were never perfect, but the economic stakes were different. Market purchase had not become the dominant route to meaningful IPv4 capacity.
After exhaustion, the centre of gravity moved. Address space now moves among operators, cloud platforms, hosting businesses, enterprises, brokers, asset holders and regional registries. The registry does not create the scarcity. It records and recognises changes in the control state of a scarce resource. Transfer policy therefore becomes the settlement layer of a capital market. It determines how quickly supply can meet demand, how much uncertainty a buyer must price, how much legal diligence a seller must perform, how much information an intermediary can credibly provide and whether a small operator can obtain capacity without being trapped by paperwork, delay or opaque qualification.
ARIN's materials acknowledge the market without fully surrendering allocation-era control. An organisation with unused ARIN-issued IPv4 address space or an ASN may release it to a specified recipient that qualifies under policy. A holder may also transfer unused IPv4 address space or ASNs to a qualified recipient in another RIR region if compatible reciprocal needs-based policy exists. That is not a free market in the ordinary sense. It is a recognised market whose settlement layer remains governed by registry approval. The asset can move in contract. The market only receives full practical effect when the ledger recognises the move.
The distinction is not semantic. In a mature land, securities or commodity market, a registry or clearing system is judged largely by whether it records valid transactions, prevents fraud, maintains the public state and settles rights predictably. In the IPv4 market, the registry still asks whether the recipient has a qualifying need. That imports a public-administration test into a private transaction. A buyer willing to pay market price for scarce IPv4 is already revealing economic need through capital commitment. The registry may still have legitimate questions about identity, authority, fraud, sanctions, dispute status, duplicate registration and security-state continuity. But when it asks whether the buyer's future use satisfies an administrative forecast, it has become more than a ledger.
ARIN's case is revealing precisely because it is orderly. Disorder can obscure structure. ARIN's procedures show the structure cleanly. Its rules are public, but publicity is not the same as title certainty. Its transfer path is real, but a recognised path can still impose economic drag. Its process is stable enough to support a market, but stability does not by itself prove neutrality. The institutional question is not whether ARIN permits transfers. It plainly does. The question is whether its design lets addresses move toward higher-valued use while preserving only those registry controls that running networks actually require.
That question becomes more urgent as IPv4 becomes a balance-sheet and continuity asset. The cost of a broken transfer is no longer just a delayed ticket. A buyer may be integrating customer infrastructure, cloud regions, VPN platforms, email systems, firewall rules, security products, telecom services or acquired networks. A seller may need certainty to close an acquisition, unwind an old estate, return capital to investors or retire an operating subsidiary. A broker may be balancing escrow, warranties, route reputation and registry timing. In that environment, the registry's settlement design is part of the asset's price.
Title certainty begins with recognition
The most valuable contribution a number registry can make to a transfer market is title certainty. In the IPv4 context, "title" must be used with care. RIRs often resist property language, and address resources are governed through registration, service agreements, policy and operational recognition rather than through ordinary land-title doctrine. Yet markets do not wait for institutions to settle vocabulary. Buyers, sellers, lessors, lenders and network operators behave as if the recognised control position over a block has economic value. They price it, finance it, route through it, litigate around it, insure around it and build customer services on it. The registry record is part of that value because it tells the world which organisation is recognised for administrative and operational purposes.
ARIN's transfer rules make recognition central. For specified recipient transfers, the source organisation must be the current registered holder of the resources. It must not be involved in a dispute about the status of those resources. It must provide a signed and notarised officer acknowledgement letter. The minimum transfer size is generally a /24. Reserved-pool addresses are not eligible. If the current registrant no longer exists, a merger, acquisition or reorganisation transfer may be required before the specified recipient transfer can proceed. Each requirement is, at one level, a title-certainty device. It tests whether the seller is the recognised source, whether corporate authority exists, whether the resource is disputed and whether the proposed transfer can be recorded without corrupting the ledger.
Those checks are defensible when they protect the record. A registry that recognises transfers from non-existent companies, forged officers or disputed blocks would destroy trust. Its job is to prevent double claims and fraudulent changes. But the same structure can become overbroad if every uncertainty is converted into institutional discretion rather than conflict metadata. A mature asset registry should distinguish invalidity, dispute, delay and administrative incompleteness. If a block is subject to a legal dispute, the ledger can record the dispute, preserve the last verified state and prevent conflicting alteration. It need not convert every unresolved question into a general power to freeze market movement beyond what the dispute requires.
ARIN's merger and acquisition route shows both the strength and the tension. The registry may process transfers where an organisation acquires assets such as customers and equipment, a network, or the organisation as a whole. It may ask for transaction instruments, merger filings, court orders, public filings, name-change documents and evidence connecting the old and new organisations. It permits redaction of financial terms and offers a standard nondisclosure agreement. This is sensible. Companies disappear, addresses remain in use, and registry records must follow economic reality. The registry is not merely blessing commerce; it is preserving continuity between operating assets and the public record.
Yet recognition remains a gate. After approval, invoices and agreements still stand between the private bargain and completed registry effect. ARIN's guide indicates that once approval, signed agreements and applicable fees are in place, resources will be transferred within a short defined period. That finality is useful, but it reveals settlement power. Until the registry updates the record, the private transaction has not achieved full market certainty. The parties may have contract rights against each other. The asset's practical title certainty still depends on registry recognition.
This dependence is the root of both value and risk. A trusted registry record makes IPv4 more valuable because counterparties can rely on a public control state. A discretionary registry gate makes IPv4 less valuable because counterparties must price review, delay and institutional interpretation. The ledger creates certainty only if the market believes it will record valid changes predictably. If recognition becomes a tool for policy judgment beyond fraud, authority, uniqueness, security continuity and dispute isolation, the ledger begins to behave like an allocator again.
ARIN is not alone in this problem. It is a structural feature of the RIR system. ARIN's region, however, makes the problem sharper because the market is more capitalised. Large cloud platforms, data-centre networks, carriers, content networks and enterprise buyers can turn a small uncertainty in recognition into a large price discount. A holder with a clean block in a region with predictable settlement can command a different market premium from a holder whose recognition path is uncertain. Title certainty is not a legal abstraction. It is the difference between an address block as bankable infrastructure and an address block as conditional administrative permission.
The best registry design therefore treats recognition as a constrained settlement duty. The source must be valid. The authority must be verified. The record must be unique. Fraud and sanctions issues must be handled. Security publication must remain coherent. Disputes must be recorded and isolated. Once those conditions are met, the presumption should favour recognition. Anything more is not mere recordkeeping. It is economic regulation.
Needs assessment is rationing logic after exhaustion
The most important economic residue in ARIN's transfer architecture is needs assessment. ARIN's transfer-recipient rules under NRPM 8.5 require, among other things, a minimum transfer size, qualification for an initial /24, documentation that 50 percent of requested IPv4 addresses will be used within 24 months, evidence that 50 percent of prior allocations are efficiently used and an alternative path for existing holders that can demonstrate 80 percent efficient utilisation of previous allocations. For incoming inter-RIR transfers to ARIN, recipients must demonstrate need for up to a 24-month supply of IPv4 addresses. These tests are framed as policy discipline. In market terms, they are rationing logic.
Rationing logic is not irrational in a free-pool allocation system. If a registry is giving scarce resources from an unallocated pool, it needs a method for deciding who receives them. Need assessment can prevent one applicant from taking a disproportionate share without operational justification. It reflects the old conservation doctrine: avoid waste, promote efficient use and keep addresses available for other networks. The problem is not that such logic ever existed. The problem is that it survived into the transfer era, where the resource is no longer being issued from a common pool but purchased from another holder.
In a market transaction, the buyer pays the price and bears the risk. The seller gives up the resource and receives consideration. The registry's technical interest is that the transfer not create duplicate claims, fraudulent records, invalid security data or untraceable control. The buyer's business need is primarily a matter for the buyer, its customers, investors and contracts. A company may buy IPv4 for immediate deployment, future expansion, renumbering, customer continuity, platform redundancy, reputation repair, M&A integration, cloud migration, leasing or strategic reserve. Some uses will succeed; some will not. Markets discover that through capital discipline. A registry forecast is a crude substitute.
Needs assessment also creates information asymmetry. The applicant must reveal plans to satisfy the registry. The registry sees confidential business information that counterparties and competitors do not. The buyer does not know exactly how staff will interpret its evidence, how long review will take or whether additional documentation will be requested. The seller may not know whether the buyer's qualification risk will delay closing. Brokers can help, but their value often consists of navigating approval psychology rather than improving the underlying asset. That is a sign of administrative friction.
The economic cost is not limited to delay. Needs assessment suppresses demand from buyers whose legitimate need is difficult to express in old utilisation language. Cloud platforms, content delivery networks, security businesses, VPN operators, hosting firms, AI infrastructure, adtech, telecom services and enterprise networks may use addresses in ways that do not fit historical network-growth models. A buyer may value optionality because future scarcity is costly. The registry may see optionality as insufficiently immediate. The market prices future risk; the registry asks for present justification. The result is a wedge between economic value and recognised need.
That wedge matters most for liquidity. Liquidity is not merely the presence of buyers and sellers. It is the ability to transact with predictable settlement, tolerable cost and limited uncertainty. Needs assessment reduces liquidity by narrowing the set of eligible recipients, increasing diligence costs, extending timelines and giving sellers reason to prefer buyers with easier approval profiles. A large incumbent with sophisticated staff may pass more easily than a small but fast-growing operator. That is not necessarily because the incumbent has higher economic need. It may simply have better paperwork, counsel and registry familiarity.
Heng's public argument about RIR enforcement creep captures the structural concern. When registries expand from recordkeeping into auditing, freezing, retroactive questioning and approval-based market control, scarce assets behave less like capital and more like ration coupons. The phrase is polemical, but the mechanism is conventional economics. If ownership-like control is conditional on administrative approval, the asset's liquidity and collateral value fall. If liquidity falls, price discovery weakens. If price discovery weakens, capital allocation becomes less efficient. If capital allocation becomes less efficient, smaller and newer operators pay the highest relative cost.
ARIN's defenders can answer that needs assessment protects the community from speculation and hoarding. That answer is weaker after exhaustion. Speculation is not automatically harmful in a scarce asset market. It can provide inventory, price signals and future supply. Hoarding is not solved by forcing every transaction through a bureaucratic forecast. Hoarding is reduced when idle holders can release assets to higher-valued use and are rewarded for doing so. Transferability is the mechanism. A registry that blocks or narrows transferability may increase the very underuse it claims to prevent.
The institutional alternative is to split allocation from transfer. Need-based criteria can remain for any residual free-pool distribution or special reserved pool. For already allocated resources, the transfer test should be objective: source authority, absence or recording of dispute, fraud control, accurate records, security-state transition, sanctions compliance where legally required and recipient acknowledgement of registry-service obligations. That would not abolish ARIN's role. It would refine it. The registry would protect the ledger and stop judging the buyer's business plan.
Liquidity needs information, not just permission
IPv4 transfer markets are information markets. The item being traded is technically simple: a block of globally unique addresses. The risk environment is not simple. A buyer must know whether the seller is the recognised holder, whether the block is cleanly routable, whether it appears on blocklists, whether RPKI objects exist, whether IRR data must be changed, whether reverse DNS will move, whether there are downstream users, whether abuse history will affect reputation, whether legacy status changes service access, whether the source is within a transfer restriction, whether wait-list consequences apply, whether inter-RIR compatibility exists and whether the recipient can satisfy needs assessment.
ARIN's materials partly address this complexity. The transfer guide advises source organisations to edit or delete transferring prefixes from source ROAs, review maxLength values, update or remove IRR objects, coordinate reverse DNS delegation and ensure recipients understand their responsibility for RPKI, IRR and reverse DNS after transfer. This is useful operational guidance. It shows that a transfer is not only a legal event. It is a change in the routing-adjacent and security-adjacent state of an address block. If handled badly, the buyer may receive recognised registry control but inherit broken routes, stale objects, reputation damage or customer disruption.
The market's information problem is larger than any checklist can solve. Sellers know more about historical use than buyers. Brokers know more about current demand than either side may fully disclose. Recipients know more about future deployment than the registry. ARIN knows more about pending request status than a counterparty in a separate confidential request. Blocklists and geolocation databases may lag. RPKI transitions may be technically clean in ARIN systems but messy for relying parties. The block's economic value is therefore shaped by public registry data, private warranties, broker reputation, operational diligence and post-transfer remediation.
ARIN's confidential request process protects sensitive information, but it also creates bilateral opacity. The guide states that ARIN cannot provide information on requests from other organisations and that parties should coordinate directly to monitor progress. That is understandable from a privacy perspective. It is also a market friction. The seller cannot fully see the buyer's approval risk. The buyer cannot fully see the seller's documentation risk. Each side may be waiting on a registry process whose details are not visible to the other. The broker becomes an interpreter of silence.
A mature transfer market therefore needs standardised representations and public-state signals. The registry need not publish private contracts, prices or business plans. It can, however, make the public ledger more informative: current holder, resource status, historical transfer record, dispute notation where appropriate, routing-security state, reverse DNS status, public eligibility category, legacy or agreement status where disclosure is permitted and operational warnings where the record itself supports them. The more objective the public state, the less the market depends on insiders who know how to read registry process.
ARIN's retired Specified Transfer Listing Service illustrates the evolution. The STLS, supported from 2010 until its retirement in June 2023, was an optional service to connect organisations seeking IPv4 space, organisations looking to transfer unused space and parties interested in facilitating transfers. Its retirement left the Qualified Facilitator Program as the more visible ARIN-sanctioned support structure. The change is telling. The market no longer needs a registry-operated classified board as much as it needs competent execution. But execution still depends on registry recognition, and recognition still depends on policy qualification. The information problem has moved from finding counterparties to managing settlement risk.
Brokerage developed because ARIN's ledger is necessary but not sufficient. A broker can source supply, screen buyers, coordinate documents, advise on reputation, negotiate escrow, structure closing and manage communications. But a broker that only forwards documents to the registry is a thin intermediary. Heng's "broker question" reframes the issue: the real question is not which intermediary can introduce a seller, but who can carry registry-layer risk when the transfer process becomes uncertain. That distinction matters even to readers who do not accept every commercial conclusion attached to LARUS or i.LEASE. Brokerage in IPv4 is less like ordinary sales intermediation and more like risk architecture around a monopoly settlement layer.
Information asymmetry therefore has two cures. The first is better market infrastructure: cleaner records, standardised diligence, clearer dispute-state publication and predictable transfer-state signalling. The second is narrower registry discretion: fewer subjective reasons for a valid transaction to fail late. ARIN has done more than many institutions to make process visible. Yet the continued presence of needs assessment and policy-compatibility filters means information asymmetry remains structurally high. Participants are not merely discovering facts about the asset. They are discovering how the registry will interpret their facts.
Inter-RIR transfers turn settlement into political economy
ARIN's inter-RIR transfer rules show how a technical registry system becomes political economy. Inter-RIR transfers under NRPM 8.4 allow organisations in the ARIN region that hold unused IPv4 address space or ASNs to transfer to a specific qualified recipient in another RIR region, and allow transfers from another RIR region to an organisation inside the ARIN region. The decisive condition is compatibility. The transfer can occur only between RIRs that share reciprocal, compatible, needs-based policy. ARIN's public transfer page lists APNIC, LACNIC and RIPE NCC as approved transfer partners, while AFRINIC is not approved. It also notes that inter-RIR transfers involving ARIN may not include IPv6 addresses.
Compatibility sounds technical. Its effect is political. A resource can move across regions only if two policy regimes recognise each other. The registry border becomes a customs border for number-resource capital. If compatibility exists, the asset may cross. If it does not, the asset remains trapped in its regional administrative system or must move through alternative structures. This is not routing necessity. BGP does not care which RIR updates the record. The constraint is institutional recognition.
Inter-RIR compatibility also creates bargaining power between registries. A region with strict needs-based rules can export that discipline by refusing to recognise transfers with a looser region. A region that wants inbound liquidity may adjust policy to meet another registry's compatibility standard. A region with no compatible outbound path imposes a discount on resources registered under it. A buyer in one region may prefer a block from a region with smoother recognition even if the technical addresses are equivalent. Policy becomes part of price.
ARIN's role is especially important because North America is both a major source and a major destination for IPv4 demand. If ARIN requires reciprocal needs-based compatibility, it helps preserve needs assessment as a global transfer norm. Even when another region might prefer a more market-led approach, access to ARIN-region transfers may require policy resemblance. That is institutional gravity. ARIN does not need to claim global authority to shape global behaviour. Its market weight gives its recognition rules extraterritorial influence.
The politics are clearest in the AFRINIC contrast. AFRINIC's governance controversies and policy disputes have generated arguments about regional control, outbound movement, institutional legitimacy and the difference between registry continuity and gatekeeper protection. ARIN's page simply states that AFRINIC is not approved for transfers. As a factual exhibit, that line is modest. Economically, it is large. It means ARIN-recognised inter-RIR liquidity does not include the African registry. The reason is policy compatibility, not packet compatibility. The address space remains technically global. Its recognised movement is institutionally regional.
This is why the "service region" should be treated as administrative metadata, not as moral title. The Internet was not born as five political economies. RIR regions were a scaling device for registration and policy administration. Once IPv4 became capital, regional service boundaries became potential capital controls. Some controls are explicit, as in rules that restrict outbound movement. Others are softer, as in compatibility requirements, needs-based certification and agreement conditions. ARIN's system is not the most restrictive, but it demonstrates the same category: capital movement depends on registry recognition.
The alternative is not a single global registry monopoly. That would merely concentrate discretion at a higher level. The better design is interoperability with portability: common technical standards, auditable records, objective fraud controls, recognised dispute metadata and the ability for resources to receive registry services without being trapped by a regional policy monopoly. Inter-RIR transfers should be a settlement question, not a political negotiation. If the source holder is valid, the recipient identity is valid, no superior dispute block applies, the security-state transition is clean and the destination registry can maintain accurate records, recognition should follow.
ARIN's current architecture stops short of that model. It is compatible with markets, but it is not market-neutral. It facilitates cross-border movement among approved policy peers while preserving an ideological preference for needs-based control. That makes ARIN a stabiliser for the existing RIR order and a constraint on a more liquid global IPv4 market. Whether one calls that prudence or protectionism depends on the prior view of registries. Institutional economics asks a narrower question: who bears the cost of the constraint? Usually, it is not the registry.
Legacy resources are the original title problem
No analysis of ARIN's transfer market is complete without legacy resources. Legacy IPv4 resources are the residue of the pre-RIR and early RIR era. ARIN's legacy guidance explains that early Internet number resources were allocated before formal legal agreements became standard, and that when ARIN was formed in December 1997 it took responsibility for IPv4 addresses and ASNs not already administered by the European and Asia-Pacific registries. The ARIN board decided to provide registration services for those legacy resources without requiring original holders to enter into a Registration Services Agreement or pay service fees.
Legacy resources are economically important because they expose the difference between registration service and asset control. A legacy holder not under an ARIN agreement may still maintain unique registration in Whois and RDAP, update public registry data, manage reverse DNS delegations, maintain records in ARIN Online and access DNSSEC. ARIN's guide says RPKI and IRR access require an ARIN agreement. That creates a two-tier certainty structure. The holder can preserve basic public recognition without signing, but modern routing-security and routing-registry functions are tied to agreement status.
The Legacy Registration Services Agreement tried to bridge the gap. ARIN offered the LRSA from October 2007 through December 2023 for organisations and individuals in the ARIN service region with legacy resources. Fee-cap treatment continues for active LRSAs entered into before January 2024, with ARIN's 2026 fee schedule describing a $250 annual cap for those covered legacy resources and annual $25 increases. Legacy resources brought under agreement after January 2024 fall under the regular Registration Services Plan fees. These details matter because fees, security services and agreements affect the carrying cost and operational certainty of legacy blocks.
From a market perspective, legacy status is both valuable and complicated. It can signal historical claim strength. It can also raise diligence questions. Is the holder under LRSA, RSA or no agreement? Are RPKI and IRR services available? Are contacts current? Has corporate identity changed? Is the block subject to old internal records, successor disputes or acquisition history? Will a transfer require signing a current RSA? Will legacy fee treatment survive, change or disappear? The answers affect price.
Legacy resources also test ARIN's institutional restraint. If a resource holder can maintain core registration without an agreement, the registry implicitly recognises that its ledger function is not identical to a contract-based service monopoly. It must preserve uniqueness and accuracy because the Internet needs the record to be true. That is the ledger role. When additional services require agreement, the registry moves into the service-operator role. Both roles may be legitimate. They should not be confused.
The danger is agreement leverage. If modern operational safety increasingly depends on RPKI, IRR, automated updates and advanced registry services, legacy holders face pressure to enter the contractual framework even if their historical position was outside it. That may be commercially reasonable in many cases. It may also shift bargaining power. A holder that needs routing-security features for customer trust may accept terms it would otherwise resist. The registry can say the holder remains free. The market may say the freedom is increasingly costly.
This is the same structural issue that appears across RIRs. The registry database begins as a technical coordination tool. It becomes part of the asset's value. The institution operating it then gains leverage over asset holders. In low-value conditions, the distinction barely matters. In a high-value IPv4 market, it becomes central. Legacy certainty is not only a backward-looking legal problem. It is a forward-looking test of whether the registry layer can provide service without converting service dependence into control.
ARIN has handled this more transparently than many institutions. Its legacy pages are explicit about available services, agreement requirements and fee treatment. That transparency improves certainty. But the deeper market question remains: should access to modern registry-security functions be tied to acceptance of a mutable private agreement, or should core security publication be treated as part of ledger continuity? A live network's security status is not merely a premium convenience. It is part of the public reliance environment around the number. As RPKI becomes more central to routing decisions, this question will become harder to avoid.
Brokers, facilitators and contract risk
ARIN's Qualified Facilitator Program is a practical admission that transfer markets need intermediaries. The programme is optional. ARIN says qualified facilitators assist organisations seeking help in acquiring or transferring IPv4 address space or ASNs, and that using one may streamline the process with Registration Services. Facilitators must satisfy requirements including legal registration and good standing within the ARIN region, coverage under specified RSA versions, operation within the ARIN service region, sanctions-screening constraints, designated representatives, transfer qualification, liability insurance, background checks, indemnification, customer references, annual confirmation and fees. The annual facilitator fee is listed at $10,000 in the 2026 fee schedule.
This creates a semi-official intermediary class. ARIN does not require parties to use a facilitator, and it states that transfer negotiations and financial terms are matters for the parties. Yet by vetting facilitators, ARIN creates a reputational channel. The facilitator does not replace registry approval. It helps the market navigate the approval layer. In economic terms, the facilitator is a transaction-cost response to settlement risk.
That settlement risk has several layers. The source must be eligible and authorised. The recipient must qualify. The parties must open the right requests. Documentation must satisfy ARIN. Fees must be paid. Agreements must be signed. Routing-security and registry records must transition cleanly. The private contract must allocate failure risk if registry recognition does not occur. A buyer and seller can agree on price, but the transaction still depends on an external recognition process.
Brokers and facilitators make money because this process is not self-executing. They reduce search costs, screen participants, explain documentation, coordinate timing and sometimes manage escrow or warranties. But their role also shows the market's institutional weakness. In a fully mature asset market, intermediaries compete mainly on price discovery, financing, execution speed, custody and distribution. In IPv4, they also compete on familiarity with registry rules and staff expectations. That is a sign that the clearing layer remains discretionary enough to require translators.
The contract risk is not only between buyer and seller. It is also between operator and registry. Public discussions of ARIN's Registration Services Agreement note that recognised resources sit inside a policy-bound and contract-bound framework, with service obligations, fee obligations, changed policy exposure, possible enforcement paths and limited registry-side liability. The exact legal effect of any clause is a matter for counsel and fact-specific circumstances. The market signal is simpler: a buyer purchasing expensive IPv4 blocks must price not only the block, but the ongoing institutional relationship attached to recognised control.
This is where LARUS's first-party leasing argument enters the wider debate. LARUS presents direct holding as potentially riskier than leasing from a specialist holder because direct purchase places the operating company inside the registry-facing layer: payment, audit, policy, compliance, termination and revocation machinery. A buyer may think it has reduced risk by putting its own name in the registry database. The counterargument is that it has placed registry risk directly on the balance sheet of the company running the live network. Leasing from a first-party pool, in this view, separates production use from upstream registry exposure.
That claim is commercial as well as doctrinal, and should be read as such. But the risk-placement question is real. If the registry contract is thin relative to the operational value of the block, direct holding may not be as safe as it appears. If a broker only matches buyer and seller, it does not solve that mismatch. If a facilitator only helps a request pass, it does not guarantee continuity after recognition. The market therefore needs to distinguish transaction brokerage from continuity architecture.
ARIN's facilitator model improves customer experience, but it does not remove the structural issue. The registry still controls recognition. The facilitator is vetted by the same institution whose process it navigates. That may increase trust for ordinary users; it may also consolidate the registry's influence over the intermediary ecosystem. A facilitator who depends on ARIN qualification has incentives to align with registry process. That is not inherently corrupt. It is an institutional fact. In a market where the registry is both ledger and gatekeeper, intermediaries tend to become gatekeeper-facing.
Sophisticated buyers should separate three questions. Who can find supply? Who can close the transfer? Who can carry continuity risk after closing? The first is brokerage. The second is process execution. The third is structural risk placement. ARIN's architecture has made the second question increasingly professional. The market's next stage will be judged by whether the third question becomes explicit.
Capital controls can exist without a ministry
Capital control normally evokes states: exchange restrictions, currency controls, outbound investment limits, approvals, quotas and licensing. IPv4 has no ministry of capital movement. Yet transfer policy can create analogous effects. When a scarce, priced, transferable asset can move only if an administrative institution recognises the transfer, and when recognition depends on need, region, compatibility, agreement, wait-list status and documentary approval, the market has a form of capital control. It is private, technical and policy-based, but its economic function is familiar.
ARIN's version is moderate compared with more restrictive regimes. It does not impose a general outbound embargo on ARIN-issued IPv4. It recognises inter-RIR transfers with compatible partners. It has a published process and defined categories. It allows private price negotiation. It does not claim to set the sale price. But capital controls do not require price-setting. They require a gate over movement. ARIN has that gate.
The gate appears in several places. A source that transfers IPv4 resources to another party cannot apply for IPv4 space from the ARIN wait list for 36 months. A source must not have received a transfer or allocation from ARIN within the prior 12 months, subject to ownership-control exceptions. A recipient on the wait list may be removed and barred from reapplying for 90 days after applicable transfer activity. Inter-RIR transfers require compatible reciprocal needs-based policy. Recipient qualification includes utilisation tests. Transfer fees and recipient processing fees must be paid. Qualified facilitators must satisfy region, sanctions and insurance requirements. None of these rules is a customs wall. Together, they shape the flow of IPv4 capital.
Some of the rules have legitimate anti-abuse functions. A wait-list recipient should not be able to obtain scarce space cheaply from ARIN and immediately arbitrage it into the market. A source that recently received resources should not exploit policy for quick resale. Sanctions compliance is a legal reality for US-region actors. Fraud and dispute checks are essential. The problem is not that every friction is irrational. The problem is that the architecture rarely separates anti-abuse controls from broader market suppression with enough precision.
A stronger policy design would ask what each control prevents. If a rule prevents double registration, fraud, sanctions breach, immediate free-pool arbitrage or unresolved dispute, it has a narrow justification. If it prevents a buyer from acquiring resources because the registry is unconvinced by the buyer's business forecast, it is a rationing control. If it prevents movement because another registry's policy ideology differs, it is a regional compatibility control. If it burdens a small buyer more than a large incumbent because documentation costs are fixed, it is a scale-biased control. Each may be defensible only if its cost is acknowledged.
RIRs often speak of such controls as community protection. Institutional economics is less impressed by moral vocabulary. A policy can be well-intentioned and still misallocate capital. A rule can sound fair and still favour incumbents. A registry can say it is preventing speculation while actually reducing supply. A compatibility requirement can be framed as reciprocity while functioning as protectionism. The question is not whether the policy has a virtuous name. The question is whether the policy improves the functioning of the market and the continuity of live networks relative to its cost.
ARIN's transfer architecture is a compromise between allocation-era legitimacy and market-era reality. It permits capital movement but disciplines it through the old language of need. It accepts private counterparties but requires public-registry recognition. It allows cross-region movement but only among policy peers. It supports intermediaries but qualifies them through registry standards. It preserves legacy records but ties advanced services to agreements. This is not a free capital market. It is a managed market.
The danger of managed markets is not immediate collapse. It is slow underperformance. Supply stays idle because holders dislike transfer friction. Demand shifts to leasing because purchase settlement is burdensome. Buyers discount blocks from regions with uncertain recognition. Smaller operators avoid the market or rely on intermediaries with expensive margins. Large incumbents absorb compliance and keep accumulating. Prices become less informative because eligibility, not only scarcity, shapes transactions. The system remains orderly, but order hides deadweight loss.
Small operators pay the dependency tax
Transfer-market architecture is never neutral across firm size. Large cloud platforms, telecom groups, content networks and well-funded enterprises can hire counsel, retain brokers, prepare utilisation documentation, manage ARIN Online accounts, coordinate RPKI transitions, pay transfer fees, absorb delay and negotiate sophisticated warranties. A regional ISP, hosting company or small infrastructure operator may face the same procedural requirements with far less administrative capacity. The fixed cost of registry navigation becomes a dependency tax.
ARIN's fee schedule shows part of the scale issue. Annual Registration Services Plan fees rise by aggregate holdings, with a 2026 3X-Small category at $275 for /24 or smaller, then larger categories up through 5X-Large. Transfer fees include a $500 non-refundable transfer processing fee for resource-transfer requests and recipient processing fees scaled by transferred IPv4 aggregate size. The absolute amounts may be modest compared with market prices for large blocks, but fixed fees and process costs matter more at small scale. A /24 buyer experiences each document request, broker fee and delay as a larger share of total transaction value than a /16 buyer.
Needs assessment compounds the problem. A small operator may have real demand but less formal forecasting capacity. It may serve customers whose future address use is uncertain. It may need IPv4 for redundancy, customer retention, email reputation or gradual migration rather than a neat 24-month deployment schedule. It may not have a dedicated registry team. The registry asks for evidence; the operator translates business reality into policy grammar. The translation cost favours those who already speak the language.
Wait-list and transfer interactions also matter. ARIN's waiting list is a residual scarcity mechanism. Public ARIN-region coverage recorded a July 2025 waiting-list distribution of 83 IPv4 blocks. Such distributions are important to small operators because open-market purchase can be expensive. But wait-list restrictions can limit strategic flexibility. A small operator must decide whether to wait for scarce registry-distributed space, pursue a market transfer, lease capacity or renumber around scarcity. Each option has policy consequences. Large operators can run parallel strategies. Small operators often cannot.
Intermediary dependence follows. A broker or facilitator can be valuable, but it adds cost and another information layer. The small buyer may not know whether the intermediary is truly protecting its interests or mainly optimising for closing. It may not know whether the block's reputation problems are manageable. It may not know whether the seller's documentation is robust. It may not understand how RPKI or IRR records will change. It may rely on the broker, the registry and the seller at the same time, while having the least leverage among them.
This is why the ledger-versus-gatekeeper distinction is not abstract. A thin ledger helps small operators because objective records reduce the need for insider navigation. A thick gatekeeper hurts them because every subjective control requires expertise, time and institutional familiarity. Large operators can manage thick governance. Small operators experience it as dependency. A registry that claims to protect the community can end up protecting those most capable of navigating it.
NRS's public materials use sharper language, arguing that businesses should own the fundamental elements of their IP business and that number-resource governance should move toward decentralisation as scarcity and economic value reshape the registry system. The broader economic point is that operators need autonomy over the identifiers on which their customers depend. If the smallest operators must rely on a chain of registry approval, broker interpretation and policy compliance to obtain or move addresses, the transfer market has not democratised access. It has replaced allocation scarcity with process scarcity.
ARIN could reduce the dependency tax by making transfer eligibility more objective, publishing clearer pre-transfer state signals, limiting needs assessment for market transfers, standardising small-block pathways and separating anti-abuse controls from business-plan review. It could also treat routing-security continuity as a standard transition obligation rather than a premium complexity. Such reforms would not weaken the ledger. They would make the ledger more usable for the operators least able to carry administrative overhead.
The registry as ledger and the registry as gatekeeper
The central choice in ARIN's transfer-market architecture is whether the registry is primarily a ledger or a gatekeeper. A ledger records the current recognised state of number resources. It protects uniqueness, preserves history, publishes accurate data, supports security assertions, records transfers, notes disputes and prevents fraudulent changes. A gatekeeper decides whether a transaction or business model deserves recognition according to broader institutional views. The same organisation can perform both roles, but the market consequences differ.
As a ledger, ARIN is valuable. Without a trusted registry, IPv4 transfers would be more expensive, fraud-prone and operationally risky. Buyers would struggle to verify control. Sellers would face more suspicion. Operators would lack a common reference point for Whois, RDAP, reverse DNS, RPKI and related records. Courts and counterparties would lack a practical control state. The ledger increases asset value because it makes claims legible.
As a gatekeeper, ARIN creates hold-up risk. Hold-up occurs when one party invests in an asset or relationship and another party controls a bottleneck after the investment is sunk. IPv4 buyers and operators build networks around addresses. The registry controls recognition, security services and policy compliance after that reliance exists. If discretion is broad and liability is limited, the operator carries more downside than the institution controlling the bottleneck. That is the structural asymmetry at the core of Heng's notes on registry power detaching from liability.
The ARIN system contains both tendencies. The source-holder requirement, officer acknowledgement, dispute check and routing-security transition are ledger-protecting. Needs assessment, reciprocal policy compatibility and agreement leverage are gatekeeper-like. Qualified facilitators sit between the two: they help customers deal with the ledger, but they are qualified by the gatekeeper. Legacy services also sit between the two: basic records reflect ledger duty; advanced services tied to agreement status reflect service and control.
The policy debate should therefore move away from slogans. "Community stewardship" is too vague. "Property rights" is too legally loaded. "Market freedom" is too broad. "Registry stability" is too self-serving if it ignores live-network continuity. The useful question is functional: what does the running Internet require from this rule? Does the rule protect uniqueness? Does it preserve accurate records? Does it prevent fraud? Does it maintain security-state coherence? Does it isolate disputes without destroying operations? Does it improve transfer reliability? If the answer is yes, the rule belongs in a thin registry architecture. If the answer is no, the rule should be treated as economic regulation and justified with evidence that benefits exceed costs.
Needs assessment after market purchase fails that test more often than its defenders admit. So does policy compatibility when it exports a regional ideology rather than solving a technical settlement problem. So does any agreement mechanism that makes security continuity depend on acceptance of mutable terms unrelated to the core ledger. By contrast, source authentication, signed officer acknowledgement, dispute notation, RPKI transition hygiene and public transfer reports pass the test easily. They protect the record.
This is not an argument for chaos. It is an argument for discipline. A registry with a narrower role can be stronger because its legitimacy is easier to defend. It does not need to speak for a region, forecast business need, police speculation or shape capital movement. It needs to maintain the record better than anyone else, publish the state reliably, prevent corrupt changes and make settlement predictable. The more valuable IPv4 becomes, the more important that discipline becomes.
What to watch
The practical watchpoints follow from the architecture. The first is whether ARIN keeps needs assessment central to market transfers or gradually narrows it toward objective settlement checks. If recipient qualification remains a business-plan review, liquidity will continue to depend on administrative interpretation. If it becomes a lighter identity, authority, fraud, dispute and security-state test, the market will become more predictable.
The second is whether ARIN's inter-RIR compatibility model remains a policy border. If compatibility continues to mean reciprocal needs-based ideology, ARIN will keep exporting allocation-era discipline into the global transfer market. If compatibility becomes more technical and record-focused, cross-region portability will improve and regional discounts should narrow.
The third is legacy-resource treatment. The market will watch whether modern security services remain tied to agreement status in a way that pressures legacy holders into the contractual framework, or whether ARIN treats security publication as part of the core public reliance layer. This question will grow as RPKI becomes more central to routing practice.
The fourth is intermediary accountability. Qualified facilitators can reduce friction, but they can also turn registry navigation into a semi-official profession. The market needs to know whether facilitators improve execution for small users or mainly professionalise dependence on the gate. It also needs clearer distinctions between brokers that match supply, advisers that close recognition and structures that carry continuity risk after closing.
The fifth is public-state transparency. A better transfer market does not require publication of private prices or confidential business plans. It does require clearer public signals about resource status, transfer history, dispute posture, routing-security state, reverse DNS continuity and eligibility-relevant constraints. If the ledger becomes more informative, insiders lose some advantage and smaller operators gain bargaining power.
Conclusion: orderly, but not yet neutral
ARIN's transfer-market architecture is more orderly than many alternatives. It recognises specified transfers, provides inter-RIR pathways with approved partners, maintains legacy-resource guidance, publishes fees, supports qualified facilitators and gives operators a known route from private bargain to registry recognition. These are real institutional assets. They help explain why the ARIN region remains central to the global IPv4 market.
Yet order should not be mistaken for neutrality. ARIN's architecture still embeds allocation-era judgments in a post-exhaustion market. Needs assessment asks the registry to evaluate demand that market price already reveals. Reciprocal compatibility turns regional policy into a border condition for capital movement. Agreement dependence can transform service access into leverage. Confidential requests and documentation review increase information asymmetry. Small operators face fixed process costs that large incumbents can absorb. Brokers and facilitators become necessary because settlement remains too dependent on institutional navigation.
The institutional economics are clear. ARIN increases title certainty when it acts as an accurate, auditable and predictable ledger. It reduces liquidity when it acts as a discretionary gatekeeper. It reduces information asymmetry when it publishes objective resource-state signals. It increases information asymmetry when counterparties must infer request risk, staff interpretation or needs-assessment outcomes. It supports market development when it standardises execution. It constrains market development when it carries rationing logic into private transfers.
The best reform principle is not anti-registry. It is pro-ledger. ARIN should protect uniqueness, record control, publish accurate data, maintain RDAP and Whois reliability, support reverse DNS and RPKI continuity, prevent fraud, record disputes without unnecessary operational damage and settle valid transfers quickly. It should be cautious about judging business models, future need, regional economic morality or acceptable capital movement beyond narrow legal and technical constraints. The more valuable IPv4 becomes, the less credible broad discretionary control becomes.
ARIN is a test case because it is mature enough to show the structure without the distraction of collapse. Its transfer rules reveal the unresolved bargain at the heart of the RIR system: a registry born as a coordinator now sits above assets that markets treat as capital. The old legitimacy came from stewardship of scarcity. The new legitimacy will come from disciplined recognition of reality. The Internet needs a ledger. It does not need a gatekeeper that mistakes the ledger for its own authority.
The future IPv4 market will reward institutions that understand the difference. Title certainty, liquidity, low information asymmetry, cross-region portability, legacy-resource clarity, broker accountability and small-operator access all point in the same direction. Protect the record. Protect live-network continuity. Let capital move where it is most useful. Keep the registry powerful enough to make the ledger trustworthy, and constrained enough that the ledger does not become a throne.