ARIN is a useful institutional case because its problem is not collapse. The American Registry for Internet Numbers is a mature, documented, North American registry with public policy texts, published service pages, a visible fee schedule, an established transfer process, and a long administrative history. That is precisely why its language matters. When an institution with orderly procedures continues to describe post-exhaustion IPv4 governance through the vocabulary of conservation, the result is not merely technical language. It is a political economy with a technical accent.

Conservation was not always a suspect word. When a registry still controlled a meaningful free pool of IPv4 addresses, it needed rules for avoiding obvious waste. It had to decide how much address space a network could receive, how quickly the space should be used, how to prevent speculative stockpiling, and how to keep enough capacity available for later applicants. In that setting, conservation was a rationing rule for a finite administrative inventory. It was imperfect, but it was at least attached to a resource still being issued from a pool under registry custody.

After exhaustion, the same word does different work. It no longer sits above a deep common inventory. It sits above a market, a waiting list, legacy holdings, transfer rules, proof obligations, fee categories, contract incentives, routing-security services, and governance structures whose costs fall on firms that need stable number resources to serve customers. A word that once meant "do not waste the pool" can become a polite way of deciding who may acquire a scarce operating input, who must reveal business plans, who can wait, who must buy, who can sell, who bears compliance costs, and who gets to call a capital-allocation decision neutral stewardship.

The official record establishes the factual setting. ARIN's own welcome page describes it as a nonprofit, member-based organization established in December 1997 to manage and distribute internet number resources in Canada, the United States, and many Caribbean and North Atlantic countries. Its IPv4 addressing options page states that ARIN's free pool was depleted on 24 September 2015. Its Number Resource Policy Manual, request guidance, waiting-list page, transfer page, legacy-resource page, fee schedule, membership rules, and Policy Development Process give enough material to examine the system without speculation.

Those official pages are exhibits, not the frame. The frame is institutional economics: scarcity has turned registry discretion into a question of market design, entry cost, distributional politics, and institutional accountability. The public critique developed in Lu Heng's notes at heng.lu/all-notes is helpful because it asks the questions official vocabulary tends to soften: when is a registry protecting a ledger, and when is it protecting a gatekeeper? When does a narrow coordinating role become enforcement? When does community procedure launder a broader mandate? The same orientation appears in the Number Resource Society's public argument that scarcity and economic value make centralized registry discretion structurally risky, and in LARUS's commercial emphasis on continuity around registry-layer risk. Those are interested positions. But institutional analysis often begins by asking which interested party has named an incentive that public-interest language obscures.

The central argument is straightforward. ARIN performs a necessary coordination function. The ledger matters. Internet number resources must remain unique. Registration records must be reliable. Contact records must be maintainable. Transfers must not corrupt the record. Reverse DNS and routing-security data must be coherent. Fraud, duplicate claims, forged authority, and unresolved disputes are real operational risks. But the more ARIN's conservation vocabulary reaches beyond those ledger functions into transfer approval, needs assessment, waiting-list restrictions, return incentives, legacy-contract alignment, fee leverage, and member-shaped policy, the more a registry begins to look like a market gatekeeper. The rhetorical risk is that "conservation" becomes a clean word for deciding the distribution of scarce internet capital while denying that distribution is the question.

The word changed when the pool disappeared

ARIN's own depletion record marks the institutional break. The current IPv4 options page says the free pool was depleted on 24 September 2015 and that requests can no longer be fulfilled unless an applicant fits reserved-policy exceptions such as micro-allocations for critical infrastructure operators or a dedicated /24 route to facilitate IPv6 deployment. The remaining official choices include the waiting list, specified-recipient transfers, inter-RIR transfers, and IPv6 adoption. ARIN's archived 2015 depletion announcement said that approved requests could be fulfilled through the wait list or through the IPv4 transfer market.

That sentence ended one economic order and confirmed another. Before depletion, a qualified applicant could still imagine the registry as the place where future IPv4 capacity would come from. After depletion, ARIN could approve a request but could not create ordinary supply. Capacity would come from recycled addresses, revocations, voluntary returns, IANA distributions if any existed, reserved carve-outs, or transactions between holders and recipients. Conservation therefore moved from supply administration into demand control and transfer control.

In the allocation era, conservation had a plausible technocratic defense. A registry did not want one applicant to take more than it could plausibly use while later networks went without. Needs documentation, utilization thresholds, and anti-stockpiling rules were part of rationing a pool still under administrative custody. The registry was not merely an owner-like gate over market movement; it was stewarding a shrinking inventory.

In the post-exhaustion era, the same vocabulary becomes less innocent. If a buyer and seller agree on a transfer price, the address block is not coming out of ARIN's ordinary free pool. It is moving from one recognized holder to another. The registry still has work to do: confirm authority, prevent duplicate claims, verify that the current registrant is the recognized holder, check for disputes, update the public record, preserve reverse-DNS and routing-security continuity, and comply with lawful orders. But deciding whether the recipient has documented enough future operational use is not the same function. It is a judgment about whether capital may move.

The word "conservation" therefore needs a time stamp. In 1997 it meant one thing. In 2015 it meant another. In 2026 it cannot be understood without the market. A phrase designed for the scarcity of a common pool now mediates the scarcity of an asset-like input embedded in operating companies, customer commitments, network design, route reputation, acquisition strategy, and corporate value.

ARIN matters because it keeps needs assessment alive inside that market. Its rules are more visible than many institutional arrangements elsewhere, and visibility is better than opacity. Yet clarity does not dissolve the economic choice. When a recipient must document projected use, when a waiting-list recipient cannot quickly transfer received space, when a legacy holder faces different service access depending on agreement status, and when organized participants shape the rules under which transactions clear, conservation language is doing distributional work.

The intellectual mistake is to treat exhaustion as a mere change in inventory. It is a change in institutional category. Before exhaustion, the registry had a scarce stock to allocate. After exhaustion, it regulates recognition in a world where the stock is largely held by existing organizations. That shift turns old anti-waste language into a control language for secondary allocation. The same noun now speaks with different economic force.

What ARIN's public record actually shows

ARIN's public pages are valuable because they show the architecture. The IPv4 options page states the depletion problem. The request page explains documentation. The waiting-list page explains rationing. The transfer page explains market settlement. The legacy page explains the historical boundary. The fee schedule explains direct monetary friction. The membership and PDP pages explain participation and voting structures. None of this requires a hidden record. The mechanism is visible.

The request IPv4 addresses page is especially revealing. It says ARIN still processes and approves customer IPv4 requests even though depletion has occurred. It lists reserved pools and ordinary allocation standards. Initial ISP requests may qualify for a small block in some conditions; larger requests require documentation. Examples include dynamic-pool data by service type and city or region, static customer assignments, hosted domains, internal equipment and infrastructure, customer counts, and utilization percentages. Additional requests require evidence of efficient use and reassignment information.

This is not trivial paperwork. It is a view into a firm's network, customer base, growth assumptions, internal allocation logic, and sometimes its commercial plan. For a large carrier, cloud business, or mature enterprise network, producing such material is an administrative exercise. For a small hosting company, regional ISP, data center entrant, security firm, or startup network, it can consume managerial time, technical time, and legal attention. The cost is not only a fee. It is the burden of translating a business into a registry-approved account of need.

The transfer page shows the same logic in the market context. ARIN recognizes transfers due to mergers, acquisitions, and reorganizations; transfers to specified recipients within the ARIN region; and inter-RIR transfers. But it also says transfer requests must satisfy ARIN policy to receive approval. For specified and inter-RIR transfer recipients, the guidance states that the minimum IPv4 transfer size is a /24. Recipients without allocations qualify for an initial /24. Larger initial or additional blocks require documentation that 50 percent of the requested addresses will be used within 24 months and that previous IPv4 allocations are sufficiently used. An alternative route exists for organizations demonstrating 80 percent efficient utilization of current holdings, subject to limits.

Those rules are coherent as policy. They are also the skeleton of a regulated market. A buyer with cash, customers, and a strategic reason to secure a larger block cannot rely on price as evidence of seriousness. It must fit a conservation model inherited from the allocation era. A seller with unused space cannot rely only on authority and clean record status. The transaction must satisfy a policy filter.

The waiting list adds another layer. ARIN says organizations holding more than a /20 equivalent of IPv4 space, excluding certain special-use pools, are not eligible. The maximum aggregate an organization may qualify for at any one time is a /22. An organization may have only one request on the list. Receipt of IPv4 space through the waiting list, an 8.3 specified-recipient transfer, or an 8.4 inter-RIR transfer removes the organization from the waiting list. A filled request creates a 90-day wait before additional applications unless a waiver is granted. Waiting-list space cannot be transferred for 60 months except in merger, acquisition, or reorganization transfers.

Again, the policy has a rationale: prevent large holders from draining a small recycled pool and stop queue gaming. But it also defines a distributional settlement. A /22 is rationed relief, not abundance. The 60-month hold treats the received block as operational capacity rather than freely movable capital. The 90-day rule controls repeated access. The eligibility ceiling distinguishes small or less-endowed networks from larger holders. Conservation is allocating scarcity among classes of operators.

The fee schedule shows the direct monetary layer. ARIN's 2026 fee schedule says all new customers must sign a Registration Services Agreement to receive registry services, that fees are non-refundable, and that transfer requests carry a non-refundable processing fee that does not guarantee approval. Recipient processing fees scale with aggregate transfer size. Relative to large IPv4 transactions, those fees may be modest. Relative to a small operator's administrative budget, they are part of a larger frictional cost: uncertainty, delay, disclosure, preparation, and correspondence.

The public record therefore shows a post-exhaustion regime with four components: rationed leftover access, needs-assessed market recognition, contract-mediated service access, and member-shaped policy. That is not merely a ledger. It is institutional market design.

Needs assessment is hidden central planning

Needs assessment is the decisive point. It sounds technical. It often appears as a neutral requirement to document future use. In the free-pool era it served a rationing function. In a transfer market it becomes a form of hidden central planning.

The term should not be exaggerated. ARIN is not setting production quotas for every network. It is not dictating every customer contract or retail price. But it is deciding when a recipient's future plans are sufficient to justify recognition of a transfer above certain thresholds. That is a planning judgment. It substitutes an administrative account of operational need for the market signal created when a buyer risks capital on a scarce block.

In ordinary markets, willingness to pay is imperfect but information-rich. A buyer that pays for IPv4 is revealing expected use, option value, customer demand, future revenue, risk tolerance, and strategic need. It may be wrong, but it bears the downside. If it buys too much, capital is tied up. If it buys too little, customers may be lost. If it buys polluted space, reputation cleanup costs follow. Price is a harsh tutor.

Needs assessment shifts part of that judgment to the registry. The buyer must show projected use in a form the institution recognizes. This can punish unusual but legitimate models. A fast-growing hosting company may need inventory for onboarding volatility. A security platform may need addresses for segmentation and reputation management. A data center operator may need capacity before a customer signs a final contract. A cloud business may need reserve capacity to meet enterprise commitments. A telecom operator may need breathing room because renumbering customers is not frictionless. These uses are not always easy to express as a 24-month utilization spreadsheet.

The conservation answer is that without needs assessment, addresses may be hoarded or speculated upon. But "hoarding" is itself a loaded word. In an exhausted market, holding inventory can be a rational way to manage uncertainty. Warehousing risk is part of how markets work. A holder may be making future liquidity possible. A broker may discover underused space. A lessor may offer short-term access to networks unable to purchase. A buyer may hold reserve capacity because demand arrives in bursts, not in policy-sized increments. Some of this behavior can be socially useful even if it offends allocation-era morality.

This does not mean every market behavior is healthy. Fraud, forged authority, undisclosed disputes, route hijacking, shell transactions, and abusive reputation laundering are real risks. But those risks call for ledger controls: authority verification, dispute marking, public record integrity, routing-security transition, reverse-DNS coordination, and fraud prevention. They do not require the registry to decide that the buyer deserves only a certain quantity because a forecast failed to satisfy a conservation formula.

The cost of needs assessment falls unevenly. A large incumbent can hire counsel, consultants, and network architects to assemble a clean file. It can wait through correspondence. It can revise projections. It can use alternative entities or inventory. A small operator may have real demand but limited staff. Its founder may be the network engineer, compliance lead, and finance lead at once. The administrative process therefore does not merely test need. It tests process capacity.

That is one of the small-operator compliance costs hidden by conservation rhetoric. Official language says the system protects efficient use. In practice it may privilege actors who can document themselves in the format the institution wants. A rule that seems fair on paper can become regressive when the burden of proof consumes a larger share of a small operator's resources.

There is a second distributional effect. Needs assessment can reduce liquidity, and reduced liquidity often hurts weaker actors. If sellers fear delay or uncertainty, fewer blocks move. If buyers must fit administrative categories, prices may be distorted. If long-term inventory cannot be recognized, small entrants may end up leasing under less transparent conditions or buying smaller, fragmented blocks. Conservation can therefore produce the opposite of its stated aim: less visible movement, more informal dependence, and higher total cost.

The hard institutional question is simple: once ARIN has no ordinary free pool to conserve, what exactly is needs assessment conserving in the transfer market? It may conserve an idea of operational virtue. It may conserve a tradition of registry discretion. It may conserve the appearance that addresses are not capital. But if the practical effect is to decide who can acquire scarce capacity, the policy is not only conserving addresses. It is conserving power.

The market is real even without property vocabulary

ARIN's official pages do not deny the transfer market. They describe it. The 2015 depletion notice used the phrase "IPv4 transfer market". The current transfer page explains specified-recipient transfers, inter-RIR transfers, pre-approval, current-registrant requirements, recipient requirements, qualified facilitators, fees, and waiting-list consequences. ARIN is not pretending bilateral transactions do not occur.

Yet the transfer market sits uneasily beside the older conservation culture. A market treats addresses as valuable transferable claims around a scarce input. The RIR tradition resists strong property vocabulary. The compromise is to speak of registration rights, operational use, policy compliance, and stewardship. That compromise may avoid legal overstatement, but it can also obscure economic reality.

The 2014 paper "A Primer on IPv4 Scarcity" is useful because it described the transition before ARIN's final depletion. It noted that IPv4 would continue to matter for a long time and that address blocks had already become goods exchanged on secondary markets. It also described the evolution from early informal registration to needs-based allocation and then depletion. The historical account matters because it shows scarcity was not a narrow administrative inconvenience. It changed the category of the resource.

Once a block is bought, sold, leased, routed, reputationally evaluated, and embedded in customer operations, it behaves like capital even if lawyers avoid the word property. A hosting company with stable IPv4 has productive capacity. A cloud network with clean blocks can onboard customers. A telecom operator with enough addresses can reduce address-sharing pressure. A seller can turn unused space into cash. A buyer can convert cash into capacity. A lessor can turn inventory into recurring revenue. A customer can treat address continuity as a service attribute.

The registry is not the source of all this value. It is the settlement and coordination layer. Its record is essential because uniqueness depends on recognized claims, but the value comes from scarcity, routability, reputation, operational embedding, and demand. This distinction is crucial. If the registry mistakes its ledger role for ownership-like moral authority over value, it can justify too much control. If market actors ignore the ledger, they risk chaos. The right balance is neither registry supremacy nor pure laissez-faire. It is constrained record authority.

ARIN's transfer process contains many legitimate record safeguards. The current registrant must be real. Disputes matter. Officer acknowledgements matter. RSA execution matters for the recipient. Routing-security cleanup matters. Reverse-DNS coordination matters. Inter-RIR compatibility matters. These are not decorative requirements. A bad transfer can damage both parties and third parties. RPKI and IRR state can cause route rejection or misrouting. Public data can be wrong. Fraud can occur.

Needs assessment belongs to a different category from record integrity. It is not about whether the seller is real, the buyer is real, the block exists, the officer has authority, the record can be updated, or a court order constrains the transaction. It is about whether the buyer's planned use satisfies a policy conception of efficient utilization. That is where the transfer market becomes a gate.

This matters most for entrants. A large incumbent with historical holdings can conserve privately by keeping inventory. A new entrant must acquire capacity under review. The old distribution of IPv4 therefore becomes a continuing advantage. Conservation rhetoric may criticize "speculators", but it often leaves historical luck less disturbed than future entry. A firm that received large space decades ago can hold option value. A firm born after exhaustion must prove itself.

That is not ARIN's fault alone. It is the structural fact of IPv4 history. But a post-exhaustion institution should be honest about it. Conservation no longer simply protects future entrants from waste. It can also protect incumbents from a more liquid reallocation of old supply. If the market is allowed to clear under clear ledger rules, underused historical space has a stronger path to productive use. If the market is filtered through moralized need, some supply remains stuck or moves through less visible channels.

The waiting list and return incentives

ARIN's waiting list is a compromise between fairness and scarcity. It is also a study in return incentives. The list can only be filled when addresses come back into available inventory. ARIN says this typically occurs through revocations for non-payment, voluntary returns, IANA distributions, or other paths that make space available for reissue. The queue therefore depends on a trickle of returned or recovered resources rather than a deep supply.

That architecture creates a tension. The registry wants unused space to come back. But a rational holder knows that IPv4 has market value. If returning a block to ARIN produces no market consideration, while selling or leasing can produce value, the return path depends on altruism, compliance pressure, corporate cleanup, inability to transact, reputational choice, or non-payment. Conservation rhetoric may praise return to the common pool, but the economic incentives usually point elsewhere.

ARIN's rules partly acknowledge this reality. Transfers exist. Inter-RIR transfers exist. Pre-approval exists. The waiting list is not the only route. Yet the moral vocabulary around return remains powerful. The resource revocation, returns, and reinstatement page describes revocation for non-payment and returning resources to available inventory. It also explains that revoked resources may be reissued. The IPv4 options page notes that returned or revoked space can fill waiting-list requests.

The problem is that "unused" and "unneeded" are not purely technical categories. A block may be unrouted today but valuable tomorrow. It may be held as resilience capacity. It may support future customers, acquisition strategy, or migration plans. It may have market value even if the current holder does not use every address. A registry sees idle capacity; a finance team sees optionality; a network team sees safety; a buyer sees supply; a smaller market participant sees a price barrier.

The waiting-list restrictions reveal how conservation translates into distribution. Organizations with more than a /20 equivalent cannot apply. The maximum size is a /22. Receipt of space via the waiting list or relevant transfers removes the applicant from the list. Waiting-list space cannot be transferred for 60 months except in merger, acquisition, or reorganization transactions. These rules direct recycled supply toward smaller applicants and prevent quick resale. That is a distributional choice in favor of certain entrants and against immediate liquidity.

There is nothing inherently illegitimate about such a choice. The issue is whether it is named honestly. A queue that gives small applicants a chance at modest capacity is a social policy inside a registry. It should be evaluated as such: how much supply does it deliver, how long do applicants wait, how many organizations rely on it instead of buying or leasing, how often do returned blocks match requested sizes, how many applicants are removed after transfer activity, and what the 60-month hold does to business flexibility. If the queue is framed only as conservation, those economic questions are softened.

The waiting list also creates opportunity costs. A small network that waits for uncertain recycled supply may delay deployment, overuse address sharing, buy reputationally weaker space, or spend time navigating policy rather than serving customers. If it buys through transfer, it may lose waiting-list position. If it receives a small block, it may be locked into a narrow operational use case for years. Scarcity makes each path costly.

Return incentives are therefore a central test of post-exhaustion governance. A registry that wants more supply to circulate should make legitimate transfers predictable, reduce fear around record updates, clarify the treatment of legacy space, and distinguish fraud control from commercial suspicion. If holders believe visibility invites moral scrutiny or contract leverage, they may hold tighter, lease quietly, or structure around the official path. If they believe the ledger will record clean changes without turning every transaction into a trial of virtue, more supply can move.

This is the ledger-versus-gatekeeper distinction in practical form. A ledger encourages returns and transfers by making rights and changes legible. A gatekeeper encourages defensive behavior by making recognition contingent on institutional approval beyond objective record integrity. Conservation rhetoric often favors the gatekeeper because it imagines hoarding as the enemy. In a scarce market, the enemy is also opacity.

Legacy resources expose the institutional boundary

Legacy resources are the strongest evidence that the ledger can be separated from the full gatekeeping bundle. ARIN's legacy-resource page explains that early IP address space was allocated before formal legal agreements, that ARIN inherited administration of many resources not already under RIPE NCC or APNIC, and that these resources are commonly called legacy number resources. It also says ARIN's Board decided at formation to provide registration services for legacy resources without requiring original holders to sign a Registration Services Agreement or pay service fees.

That historical fact is inconvenient for any theory that the full contract stack is required for registry continuity. ARIN can maintain unique registration, Whois/RDAP data, reverse DNS, DNSSEC access, and some record functions for legacy holders not under an ARIN agreement. It withholds other services, including RPKI and IRR access, unless the holder is under an ARIN agreement. In other words, the core ledger and the expanded service relationship are divisible.

This does not mean ARIN is wrong to encourage agreements. Modern services have costs and reliance risks. RPKI affects routing security. IRR data affects routing policy. The registry is entitled to define terms for services it offers. But the legacy boundary clarifies the institutional reality: uniqueness does not inherently require every holder to accept the same contractual and policy bundle. Some core record functions can continue without full alignment.

The legacy fee structure adds another incentive. ARIN says the legacy fee cap expired on 31 December 2023, but organizations with an active LRSA entered into before 1 January 2024 continue to have limited fees for legacy resources covered before that date, while legacy resources brought under agreement after that date are covered under regular Registration Services Plan fees. This creates categories within the same registry: legacy resources not under agreement, legacy resources under older LRSA treatment, legacy resources brought under agreement after the cutoff, and ordinary resources under current terms.

Those categories affect cost, service access, transfer preparation, and due diligence. A buyer assessing a legacy block may need to understand the history of the holder, corporate authority, record accuracy, agreement status, service access, routing-security state, and fee treatment. A seller may need to clean up decades-old records. The registry sits at the point where history becomes market friction.

The deeper distributional issue is that ARIN's region contains a large share of early internet history. Early North American institutions received significant space under assumptions that no longer apply. Later networks, including small operators and entrants outside the old allocation era, face scarcity as an entry cost. Conservation rhetoric can make this history appear to be a neutral inheritance rather than a market advantage.

The correct lesson is not to punish legacy holders. Punishment would likely reduce transparency and liquidity. The better lesson is to preserve the ledger, encourage clear records, make voluntary contract alignment attractive without coercion, and make transfers predictable. If old supply is to move toward higher-value use, holders must believe the official path is safer than avoidance. A gatekeeper posture discourages that confidence.

Lu Heng's note "The Registry Continuity Fallacy -- Protect the Ledger, Not the Gatekeeper" captures the point in broader terms. Continuity of registry services matters; continuity of every institutional power claim is a different question. In ARIN's legacy context, the distinction is visible. The essential ledger can persist across different contractual statuses. That should make the institution more modest about which parts of its authority are truly necessary.

Member power and the politics of community

ARIN's membership page says membership is not required to obtain direct resources, participate in PDP discussions, submit suggestions, or join public consultations. That matters. It limits the simplest criticism that only members can speak. At the same time, the page states that General Members in good standing are eligible to vote in ARIN elections through a designated voting contact, and that the General Members mailing list is available only to General Members, Trustees, and key staff for governance discussion.

This distinction matters because policy language often uses "community" as if it were a single public. In practice, participation is layered. There are resource holders, service customers, General Members, voting contacts, staff, Advisory Council members, Board members, meeting participants, mailing-list participants, brokers, large incumbents, small operators, counsel, consultants, civil-society voices, governments, and ordinary affected users who never enter the room. These groups do not have equal time, knowledge, incentives, or voting power.

ARIN's Policy Development Process says internet number resource policy must satisfy principles of fair and impartial administration, technical soundness, and support by the internet community. It defines technical soundness through conservation, aggregation, and registration. It says community support need not be unanimous and may be shown by the active part of the community discussion so long as support substantially exceeds opposition. That is a plausible operating rule for a technical policy body. It is not the same as democratic representation of every affected interest.

In institutional economics, participation is never free. The cost of joining a policy debate is lower for organizations with dedicated staff. A large network can send people to meetings, track policy lists, model how a proposal affects holdings, and vote through membership structures. A small operator may know a policy is harmful only after it becomes a compliance burden. A customer of an ISP may be affected by address scarcity but have no practical voice. The "community" is therefore more organized than representative.

This is not a claim of bad faith. It is a claim about incentives. The people with the strongest stake or most available expertise participate more. That can improve technical quality. It can also entrench insider language. Terms such as stewardship, operational need, conservation, and community support acquire authority because they are repeated inside the process. Outsiders then encounter them as if they were neutral principles rather than contested choices.

This is the softer ARIN version of mandate laundering. A private, member-based, region-serving institution passes decisions through public-sounding procedures and technical vocabulary until policy outcomes appear to possess legitimacy greater than the representation behind them. Lu Heng's "Mandate Laundering: From RIR Fantasy to Transition Architecture" uses a sharper version of the argument: a narrow coordinating role can be wrapped in procedure, regional rhetoric, consensus language, and institutional myth until private discretion emerges as quasi-public mandate. ARIN's version is less theatrical than crisis cases, but the mechanism is recognizable.

The conservation word is ideal for this mechanism. No one wants waste. No one wants hoarding. No one wants routing chaos. No one wants fraud. A policy that claims to conserve resources and support operational need begins with a moral advantage. Critics can be made to look self-interested, speculative, or anti-community. Yet the policy may still shift costs from incumbents to entrants, from the registry to operators, from large firms to small firms, or from current holders to future buyers.

Member power must therefore be evaluated through outcomes, not only formal openness. Do policies reduce uncertainty for small operators? Do they improve transfer predictability? Do they avoid turning registry records into leverage? Do they distinguish fraud control from commercial suspicion? Do they publish enough aggregate data for outsiders to understand who benefits? Do they limit discretionary review to objective criteria? These are accountability questions, not public-relations questions.

Compliance cost is not a side issue

One of the easiest ways to misread ARIN is to focus on large transfers and overlook the operator at the margin. A small hosting company, local ISP, enterprise network, SaaS platform, managed-security firm, or data center entrant does not experience conservation rhetoric as an abstract doctrine. It experiences it as forms, forecasts, fees, correspondence, uncertainty, and the need to disclose business logic to a registry whose approval affects growth.

ARIN's documentation examples are operationally detailed. An applicant may need to show dynamic pools by service type and region, static assignments by customer, hosted domains by IP address, internal infrastructure by hostname or device, reassignment data, utilization percentages, and future plans. For a mature large network, much of this data may already be structured. For a smaller network, the process can require cleanup, mapping, counsel, and repeated correspondence.

This cost is regressive because it scales poorly. A large operator can spread compliance staff across many requests and transactions. A small operator may incur nearly the same procedural burden for a much smaller block. A transfer processing fee may be small relative to a large transaction but more meaningful for a small one. Recipient processing fees scale by size, but fixed attention cost does not. Waiting-list delay may be survivable for an incumbent with reserve capacity and painful for an entrant with immediate customers.

Compliance cost also changes competitive timing. If a small operator needs a /24 to serve customers, the difference between two days, two weeks, and two months matters. A large incumbent can absorb delay with inventory. A new entrant may lose the customer. If a buyer needs pre-approval before negotiating, the market may move while the file is reviewed. If a block's history requires cleanup, the smaller buyer may lack experience. Conservation may promise fairness, but delay and uncertainty can make fairness expensive.

There is also a confidentiality cost. Business plans, customer lists, utilization data, and internal architecture are sensitive. ARIN has reasons to ask for documentation under current policy. Applicants have reasons to be cautious. A small operator may not have formal confidentiality procedures or legal capacity to decide what to share. This creates a chilling effect: the easiest path may be to ask for less, grow slower, buy in smaller fragments, or use less transparent arrangements.

The burden is not only ARIN's. IPv4 scarcity itself imposes cost. But policy can worsen or reduce that cost. A lean ledger model would ask for what is necessary to maintain the registry and confirm objective transfer criteria. A thick conservation model asks applicants to prove virtue in a format that favors administrative maturity. The difference matters for competition.

Small-operator costs also challenge the idea that conservation protects the weak. If conservation means rationed waiting-list access to a /22 or smaller, it may help some small networks. If conservation also means paperwork intensity, transfer uncertainty, and limited liquidity, it may hurt others. The net effect is empirical, not rhetorical. It should be measured.

The Number Resource Society's claim that scarcity turns centralization into structural risk is strongest at this small-operator level. Centralization is not only a constitutional concern. It is a cost center. Every additional gate between a network and usable capacity increases the value of knowing the rules, knowing the people, having staff, surviving delay, and interpreting policy culture. That is a tax on entry.

Registry accountability when the ledger becomes valuable

ARIN accountability cannot be assessed only by asking whether it publishes documents. Publication is necessary, not sufficient. The question is whether the institution's power is matched by responsibility, reviewability, proportionality, and narrowness.

IPv4 scarcity changes the scale of harm. A delayed transfer can affect a sale. A rejected transfer can strand capital. A stopped service can disrupt records. A dispute over authority can impair financing. A routing-security transition error can break reachability. A documentation burden can delay growth. A policy change can alter the expected value of holdings. These are not minor clerical issues.

ARIN's resource revocation page describes service stoppage and revocation for non-payment under the RSA. At a stated delinquency point, services stop and records can be removed from public directory services. Later revocation can terminate the agreement and return resources to available inventory for reissue. Reinstatement may be possible if resources have not been reissued, subject to process. These rules are understandable for fee compliance. They also show how registry status can affect a valuable operational input.

The accountability problem is not that a registry should never enforce invoices. Any service body has terms. The problem is that the consequence of registry action may far exceed the fee relationship. A small unpaid invoice can coexist with an address block whose operational value is vastly larger. An operator's customers may suffer even if the dispute is administrative. The public record may change in ways that affect routing, reputation, and commercial trust. That is why registry power should be narrow, predictable, and subject to meaningful review.

Lu Heng's "Why Registries Must Never Become Enforcers" frames the issue in categorical terms: a registry should maintain records and execute transfers according to defined procedures, while punishment should belong to proper legal channels. That statement may be too absolute for some administrative fee scenarios; any continuing service has contract terms. But as a principle for scarce infrastructure it is right. The registry's most dangerous temptation is to use the address book as leverage.

The ledger-versus-gatekeeper distinction supplies a practical test. A ledger answers objective questions: who is the recognized holder, what records are attached, what authority documents exist, what dispute notices exist, what security assertions exist, what transfer satisfies defined criteria, and what court orders constrain action? A gatekeeper answers broader questions: does the institution approve of the holder's business model, planned use, customer geography, timing, market behavior, or political posture? The former is necessary. The latter is where mandate expands.

ARIN often operates closer to the ledger side than weaker institutions. It has published criteria, documented processes, and a mature administrative culture. But needs assessment, waiting-list restrictions, transfer consequences, fee leverage, and service access show that gatekeeping is still built into the model. The danger is less arbitrary drama than normalized scope creep. A calm gatekeeper is still a gatekeeper.

Accountability should therefore be designed around the economic reality of IPv4. The registry should publish clear aggregate transfer outcomes and processing statistics, explain rejection categories without exposing confidential business data, allow independent review for high-impact adverse decisions, limit discretionary inquiry to necessary record and policy criteria, and avoid treating conservation as a universal answer to every market concern. It should also make explicit when a rule is distributional. If the policy is meant to favor smaller operators, say so and measure whether it works. If it is meant to deter speculation, say what behavior is being deterred and what liquidity cost is accepted. If it is meant to protect the waiting list, report the trade-off.

The worst accountability posture is to treat every policy as technical and every criticism as ignorance of the model. Scarcity has made the model economic. Economic policy requires economic accountability.

Conservation as mandate laundering

Mandate laundering is useful because it describes a mechanism rather than an insult. A narrow administrative function enters a process of rhetorical expansion. It is wrapped in community, stewardship, conservation, technical soundness, regional service, policy consensus, and institutional continuity. It exits as a broader mandate to decide questions that look more like economic governance than record maintenance.

ARIN is not the most extreme example of this phenomenon, but it is a sophisticated example. Its conservation language comes through official policy, public process, and long institutional habit. Its region contains large, early, valuable holdings. Its transfer market is mature. Its members include sophisticated operators. Its documents are clear enough to make the distributional choices visible. If mandate laundering exists here, it does not require chaos. It can occur inside order.

The mechanism is sequential. A real technical requirement is stated: number resources must be unique and records must be accurate. A reasonable allocation-era principle is attached: scarce addresses should be conserved and issued for documented operational use. The principle survives the disappearance of ordinary free-pool supply. The same principle is then applied to transfers, waiting-list behavior, agreement incentives, and service access. The institution says it is still performing technical stewardship, even though it is now shaping market liquidity and distribution. The mandate has expanded, but the language remains technical.

This laundering works because each step has a plausible defense. Uniqueness is necessary. Record accuracy is necessary. Conservation once had a direct pool-management purpose. Transfers need safeguards. Waiting-list gaming should be deterred. Fraud should be prevented. RPKI should be reliable. Fees must be paid. None of these statements is false. The problem is aggregation. Together they can make a registry look entitled to supervise the address economy.

The key sign of mandate laundering is refusal to name trade-offs. If a policy reduces liquidity to deter speculation, that is a trade-off. If it favors small waiting-list applicants over larger holders, that is a trade-off. If it requires business-plan disclosure to approve a transfer, that is a trade-off. If it withholds certain services from legacy holders outside agreements, that is a trade-off. If member voting affects rules for non-member affected parties, that is a trade-off. Technical language should not make those choices disappear.

This is why official narratives should not be treated as the conclusion. An institution has an incentive to describe its own power in the most public-spirited terms available. It will say stewardship, not leverage. It will say conservation, not distribution. It will say community, not organized subset. It will say policy, not market design. Those words may be partly true. They are also politically useful.

The Lu Heng notes are valuable here not because they are neutral, but because they attack the vocabulary. The ledger-not-gatekeeper framing asks what continuity truly requires. The anti-enforcement note asks whether an address book should be used as punishment. The mandate-laundering note asks how a private coordination function acquires public-sounding authority. These are the right questions for ARIN because ARIN's calm institutional form can make expansion harder to see.

The question is not whether ARIN is acting maliciously. The question is whether its conservation inheritance now gives it too broad a role in allocating economic value after IPv4 exhaustion. A system can be professional and still overextended. It can be transparent and still distributional. It can be well run and still need sharper limits.

What a ledger-first ARIN would emphasize

A ledger-first ARIN would not abolish policy. It would narrow the role of policy around post-exhaustion market movement. It would distinguish the tasks essential to global coordination from those that preserve allocation-era authority.

The essential tasks are clear. The registry must preserve uniqueness. It must maintain public registration records. It must support accurate contact data. It must record authorized transfers. It must prevent duplicate claims and forged authority. It must mark disputes where appropriate. It must maintain reverse DNS and routing-security data within clear service terms. It must comply with lawful orders. It must publish policy and process information. It must keep services resilient and secure.

The questionable tasks are broader. Should the registry decide how much inventory a market buyer may acquire beyond objective anti-fraud and record rules? Should it treat 24-month utilization forecasts as superior to capital risk? Should it constrain liquidity because it fears speculation? Should it use service access to move legacy holders toward contracts? Should it let conservation language do the work of competition policy without admitting that competition policy is at stake?

A ledger-first model would not answer all these questions with a reflexive no. It would answer them with a burden of proof. Any gatekeeping beyond record integrity should be justified by a specific, measurable harm; limited in time and scope; reviewable; and reported in economic terms. If a rule reduces transfer liquidity, ARIN should say so. If the rule protects small waiting-list applicants, ARIN should measure whether those applicants actually get meaningful capacity. If a needs-assessment threshold blocks transfers, the institution should publish aggregate categories showing why. If a service restriction encourages contract alignment, the trade-off should be explicit.

Such a model would also reduce small-operator burden where possible. It would standardize documentation, offer clearer safe harbors, publish anonymized examples of approvals and denials, separate confidential business forecasting from record verification, and avoid repeated information demands where a transfer is small or the record risk is low. It would treat speed as an accountability metric, not merely an internal service target.

For legacy resources, ledger-first means record clarity without coercive moral pressure. Encourage agreements by offering useful services, not by implying that holders outside the full bundle are less legitimate. Make transfer cleanup predictable. Publish the practical differences between agreement statuses in plain commercial terms. Respect the historical boundary while reducing uncertainty for buyers and sellers.

For membership and PDP, ledger-first means humility about "community". Keep participation open, but do not pretend organized participants equal all affected parties. When policy affects market liquidity, seek evidence from small operators, buyers, sellers, lessors, customer-facing networks, and non-member affected firms. Treat absence from a mailing list as a signal of participation cost, not consent.

For accountability, ledger-first means independent review for high-impact adverse decisions. A resource holder or recipient should not have to accept that a policy interpretation affecting large value is merely staff discretion. Review should be timely, technically competent, and capable of correcting overreach. The institution's legitimacy improves when decisions can be challenged without implying disloyalty to the community.

The broader point is cultural. Conservation should become a narrow term again. It should apply to actual remaining pools and clearly defined anti-waste rules, not to every institutional preference about the movement of scarce address value. Once IPv4 is a market input, conservation must share the stage with liquidity, competition, entry costs, reliance, transparency, and proportionality.

The ARIN lesson

ARIN's conservation rhetoric is not empty. It comes from real history. IPv4 is finite. Waste was real. Classful allocation left scars. CIDR, private addressing, NAT, IPv6 planning, needs-based allocation, and registry coordination all emerged from a real technical and economic constraint. A world without uniqueness and records would be worse.

But a word can outlive its original job. After exhaustion, conservation is no longer just the discipline of a shrinking pool. It is a political economy. It helps decide whether old holders keep option value, whether new entrants buy under review, whether small networks wait, whether returned space becomes rationed supply, whether buyers must reveal plans, whether legacy holders sign agreements, whether members shape rules for non-members, and whether the registry is seen as a ledger or a gatekeeper.

That is why ARIN is a better test than a failed institution. If the problem appears only in crisis, defenders can blame personalities. ARIN shows the structural version: calm procedures, public documents, clear rules, and still a conservation vocabulary that masks distributional choices. The risk is not always dramatic abuse. It is the normalization of economic governance under administrative language.

The right conclusion is not that ARIN should vanish, nor that markets should operate without records. The conclusion is that registry legitimacy now depends on restraint. Protect uniqueness. Protect accuracy. Protect security metadata. Protect lawful transfer. Protect customers from avoidable continuity shocks. But do not pretend every control over address movement is technical conservation. Do not launder mandate through community vocabulary. Do not make small operators pay hidden compliance taxes in the name of fairness. Do not deny the asset-like reality of IPv4 while exercising power over its movement.

ARIN's post-exhaustion challenge is therefore conceptual as much as operational. It must decide whether it is primarily a ledger for a scarce-resource economy or a gatekeeper preserving allocation-era authority after the pool is gone. The ledger role is essential. The gatekeeper role is the problem. Conservation rhetoric makes the two sound the same. They are not.

If ARIN can separate them, it can become a model for a mature registry in a scarce-address world: precise, limited, accountable, transparent, and market-aware without becoming a broker or a sovereign. If it cannot, conservation will remain the polite word by which an inherited institution decides who may control scarce internet capital while insisting that it is only keeping the books.