Summary

  • Thin trading makes every visible observation influential, but an unusual price, large holder or sharp move is not itself evidence of manipulation. Market power, urgent demand, block quality, registry path and changing transaction composition can all produce legitimate price differences.
  • The clearest suspect conduct is a false market signal: an apparently executable bid or offer placed without a genuine willingness or capacity to trade, phantom inventory circulated through several brokers, a fabricated closing, or a stale quote deliberately presented as current.
  • Related-party transactions require separate treatment. A sale between affiliates, a round trip through cooperating entities or a broker's transfer to its own inventory can be commercially real while remaining unsuitable as evidence of independent supply and demand unless the relationship and purpose are disclosed.
  • Benchmark vulnerability arises when a publisher observes only voluntary submissions, can select the assessment window, uses bids and offers when transactions are scarce, or has brokerage and inventory interests. IOSCO's benchmark and oil price-reporting principles offer a proportionate governance model, not an automatic legal classification for IPv4.
  • RIR logs are authoritative evidence of recognised registration changes, not a complete market-surveillance tape. They generally omit price, quote history, beneficial relationships, failed negotiations and whether several prefix rows belong to one commercial bargain.
  • Number Resource Society can support confidential conduct research, common quote and transaction identifiers, relationship flags, benchmark-method review and a complaints mechanism. It should neither declare every high price abusive nor use market integrity as a reason to prohibit sale, leasing, collateral analysis or other asset treatment.

Thinness magnifies information; it does not prove abuse

A market is thin when willing buyers and sellers do not meet continuously across every product class. Trades may be infrequent. Individual transactions may be large relative to visible activity. Quotes may be private. An observer may wait weeks for another comparable block to appear.

IPv4 has several sources of thinness. The total numerical supply is fixed, but only a fraction is offered at any time. Many holders continue to use their space or keep it for contingency. Large clean and transferable blocks are concentrated among organisations that received early allocations. Buyers need different quantities, registry paths and delivery dates. A /24 and a /12 serve different markets even though both can be converted to a unit price.

Thinness creates sensitivity. One distressed portfolio can pull a monthly average down. One urgent hyperscale buyer can lift a large-block clearing price. One broker with unusually broad inventory can dominate a published series. A price based on three independent transactions is easier to influence than one based on thousands.

Sensitivity is not manipulation. A holder that refuses to sell below its reservation value is bargaining. A buyer that pays more to meet a deployment deadline reveals urgent demand. A seller that accepts less for a block with reputation problems reveals quality. A broker that earns a spread for taking inventory supplies liquidity. Prices are allowed to move.

Manipulation begins with deception or an abusive mechanism designed to create a false or misleading appearance of price, demand, supply or trading. That formulation still requires evidence. Who communicated what? Was the quote presented as executable? Were the parties connected? Did a reported transaction transfer economic exposure? Did the actor have an interest tied to the published level? Was contradictory evidence concealed?

Without those questions, “manipulation” becomes a political label for any market outcome someone dislikes. A governance framework should narrow the term before expanding surveillance.

No global incidence rate can be inferred from the visible record

There is no public global file of all IPv4 bids, offers, contracts, beneficial owners, broker mandates and failed negotiations. RIR transfer logs record recognised changes under regional procedures. Broker reports reveal selected activity from their businesses. Public court and company documents disclose unusual transactions. Most direct bargains remain confidential.

That evidence can identify vulnerabilities and test particular cases. It cannot establish what percentage of IPv4 transactions are manipulated, how often false quotes occur, what share is related-party, or whether abuse is increasing. The denominator of attempted and completed commercial transactions is unavailable.

Complaint data would not solve the problem by itself. Sophisticated counterparties may settle disputes privately. Unsuccessful bidders may describe ordinary competition as abuse. A broker may identify a suspicious quote but never learn whether it traded elsewhere. A registry may see a valid authority chain without seeing the price signal that preceded it.

Published transfer rows are also not identical to transactions. One portfolio sale can create many prefix entries. A merger can change a registered organisation without a market price. A permanent transfer can follow an earlier lease. Several transfers between related companies can be internal restructuring. Counting rows as trades inflates activity and can manufacture apparent price comparables.

Accordingly, this article offers a conduct taxonomy and evidence design, not a prevalence estimate. Any future study must define its observed population, preserve failed and excluded cases, link multiple legs and state coverage by contributor, region and block class.

A high price is not a false price

Scarcity creates disagreement about value. A buyer compares address acquisition with leasing, carrier service, address sharing, customer delay and IPv6 transition. A seller compares cash proceeds with continued use, leasing income, strategic reserve and expected future value. The negotiated price falls between private alternatives that outsiders do not fully observe.

A high price can be rational if a buyer avoids a larger operational loss. A low price can be rational if a seller values certainty, speed or a package sale. Different prices can coexist because blocks differ in size, history, fragmentation, region, agreement status and reputation. The last completed trade does not compel the next entity to transact at the same level.

Large holders can influence price simply because their decisions change available supply. If a university releases a substantial clean block, buyers gain an alternative. If an operator postpones sale, available inventory tightens. Market impact is not proof that the holder created a false signal.

The same caution applies to concentration. A broker with many mandates may have bargaining power. A buyer with substantial recurring demand may receive discounts or move a size band. Competition policy may ask whether conduct excludes rivals or exploits dependency. Manipulation analysis asks the different question of whether the actor falsified market information or used a deceptive transaction.

Asset treatment is also separate. Calling IPv4 rights an asset for accounting, tax, collateral or commercial analysis does not cause manipulation. Refusing the word asset does not prevent false quotes. Market integrity depends on conduct, evidence and governance, not the vocabulary a registry uses for its contractual relationship.

The first red flag is a quote that was never meant to trade

An OTC quote can take several forms. An indication describes a possible level. A request response offers a price subject to diligence. A firm quote commits specified capacity for a stated period and conditions. Manipulation risk rises when an actor presents weak interest as strong, executable interest.

A seller may advertise a large block it does not control, hoping that visible supply pushes competing sellers to lower prices. A buyer may circulate an aggressive bid without funds or qualification, hoping that a holder delays another sale or that a publisher marks the market higher. A broker may duplicate the same mandate across several listings and present each as independent inventory.

The decisive evidence is intent and context. Quotes are legitimately cancelled when facts change, a counterparty fails diligence, a validity period expires, the block sells elsewhere or the registry path proves unavailable. A quote is suspicious when the party could not perform at submission, repeatedly withdraws after the other side relies, changes undisclosed conditions only after acceptance, or uses the quote solely to influence a report or valuation.

The CFTC's 2013 interpretation of spoofing supplies a bounded analogy. In markets covered by its statute, spoofing involves bidding or offering with intent to cancel before execution, while good-faith cancellation is not enough. IPv4 transfers are not thereby transformed into futures or swaps, and the CFTC rule does not govern every address negotiation. The useful principle is evidential: cancellation alone is ambiguous; contemporaneous intent, capacity and repeated pattern matter.

A practical rule should therefore prohibit misrepresenting an indication as executable, capacity that is not controlled, funding that is not available, or a quote's conditions. It should retain the original message, timestamp, quantity, validity and withdrawal reason. It should not force a seller to keep an expired quote open or punish a buyer for discovering a genuine defect.

Phantom inventory is especially potent in a fragmented market

Inventory can appear larger than it is because the same block travels through several channels. A holder gives a non-exclusive instruction to three brokers. Each broker lists the full quantity. An aggregator copies the listings. Market observers see three sellers and perhaps three price points, although only one block exists.

Duplication can be innocent. Non-exclusive representation helps a seller reach more buyers. Confidentiality can prevent brokers from learning that another mandate exists. A holder may offer different portions conditionally. The problem arises when the duplicated presentation is used to claim independent supply or pressure the market.

Common resource identifiers can address much of this risk without exposing the holder. Each listing can carry a cryptographic commitment to the prefix set and a mandate identifier validated by an independent service. Competing brokers can learn that their offers overlap without seeing confidential terms. A public display can count unique capacity rather than advertisements.

Authority must also be current. Historical registration does not prove that the person marketing a block can transfer it. Corporate dissolution, mergers, disputes, security interests, court orders and previous commitments can intervene. A broker should label whether authority is asserted, document-checked or confirmed through a current registry process.

A listing should expire. Stale inventory distorts both supply and price. The system should preserve expired entries for audit but remove them from current depth. If the seller renews, the record should show renewal rather than pretend a new independent lot appeared.

None of this requires a public list of every block for sale. Confidential duplicate detection can improve the denominator while protecting strategy. The target is false multiplicity, not private negotiation.

False bids can influence sellers even when no benchmark exists

Market manipulation is often discussed as an attempt to move a published index. In a thin OTC market, the target can be one counterparty's belief.

A large holder considering sale may ask several brokers for market colour. If a buyer or intermediary plants an unrealistically low “recent trade,” the seller can accept less. If a purported bidder submits a high offer and then disappears, the seller can reject a genuine lower bid, lose time and later return under pressure. Both strategies alter decisions through false information.

Evidence should distinguish hard comparables from stories. A completed comparable needs a date, block class, price basis and independent confirmation that economic exposure moved. Confidential terms can remain protected, but a claim such as “a /16 traded yesterday” should not enter a formal valuation solely because a market entity said it.

Broker communications need capacity labels. A house bid backed by the broker's balance sheet differs from an unnamed customer's interest. A bid contingent on buyer approval differs from available funds in escrow. A range based on the broker's judgment differs from a submitted quote. The recipient should know which one it is hearing.

Record retention protects both sides. If a seller alleges a false bid, the quote, acceptance attempt and withdrawal reason can be reviewed. If a buyer is accused merely because it walked away after diligence, the defect and stated conditions provide a defence.

The aim is not to eliminate informal conversation. Markets need exploratory communication. The aim is to stop exploratory claims from being promoted to evidence classes they have not earned.

Related-party trades may be real and still be price-ineligible

A transaction between affiliates can transfer registration, allocate operational responsibility or support financing. It can have a contract and consideration. It can be valid under RIR policy. Yet it may not reveal the price that independent parties would accept.

A parent can sell a block to a financing subsidiary and lease it back. Sister companies can move addresses during reorganisation. A broker can purchase into inventory and later sell to a customer. An investor can transfer between controlled vehicles. A large holder can transact with a strategic partner under broader commercial terms.

The danger is not the relationship itself. It is presenting the price as arm's-length market evidence without disclosure. Connected parties can choose a price for tax, accounting, financing or internal allocation reasons. A broker's principal purchase may include a liquidity discount and later resale margin. A reciprocal deal may bundle services that make the address allocation arbitrary.

FINRA's TRACE framework provides a useful comparison because transaction reporting distinguishes counterparty capacity and includes treatment for non-member affiliate principal transactions. That regime applies to eligible securities and cannot be imposed on IPv4 by analogy. The transferable lesson is that relationship status changes how a price observation should be interpreted.

An IPv4 transaction record should therefore identify, under confidentiality protections, whether parties are under common control, whether the intermediary acted as agent or principal, whether consideration includes non-address obligations and whether a linked transaction reverses or offsets the exposure. Public aggregates can exclude connected prices from an arm's-length benchmark while reporting their volume separately.

Exclusion is not condemnation. A related transaction can be lawful and operationally important. It simply answers a different price question.

Wash and circular trades create activity without new risk

A wash trade is commonly understood as a transaction that creates an appearance of trading without a meaningful change in beneficial exposure. In IPv4, the exact legal label will depend on jurisdiction and instrument. The economic pattern remains worth detecting.

Imagine a holder transferring a block to a controlled entity, then buying it back at a higher reported price. The registrations may change twice. If the parties present the second price as independent demand while retaining the same ultimate exposure, the sequence can support an inflated valuation. A group of cooperating actors could circulate blocks and publish each leg as a new market trade.

Round trips can also arise innocently. A transaction may fail after registration and be reversed. A buyer may discover a defect. A company may unwind a restructuring. A financing vehicle may return collateral after repayment. The record needs purpose and timing before drawing a conclusion.

Detection should link transactions by prefix, beneficial relationship, payment path, control rights, timing and reversal terms. A rapid return at a predetermined price is more informative than registration movement alone. So is cash that returns to its source through side agreements.

Public price series should not count every leg automatically. A reversal should be marked. A financing transfer should be classified separately. A portfolio split and recombination should retain one economic-event family. If the analyst cannot determine whether exposure changed, the observation should be excluded from an arm's-length benchmark and included in the unknown category.

Unknown is a legitimate result. Forcing every RIR row into “sale” or “manipulation” creates the very false precision the framework is meant to prevent.

Selective submission can move a benchmark without any fake trade

A voluntary price publisher sees what contributors choose to provide. A broker can submit its strongest sales and omit discounts. A buyer can share low bids and hide higher completed prices. A holder can disclose a premium small-block trade while withholding a discounted portfolio. Every disclosed event may be authentic, yet the resulting picture is biased.

IOSCO's 2013 Principles for Financial Benchmarks identify voluntary or selective submissions, varied submitter composition, discretion, conflicts and false or misleading data as benchmark vulnerabilities. The principles call for governance, quality of methodology, accountability and controls proportionate to the benchmark's risks.

The 2012 IOSCO Principles for Oil Price Reporting Agencies are especially relevant as a market-design comparison. Physical oil assessments can rely on completed transactions, bids, offers and other information, particularly when transaction data are sparse. IOSCO requires documented methodology, priority for concluded transactions where appropriate, explanations for exclusions, audit trails, assessor supervision and conflict controls.

IPv4 is not oil, and a broker's monthly report is not automatically a regulated financial benchmark. The structural similarities are narrower: voluntary reporting, thin cells, heterogeneous products, human judgment and contributors with positions affected by the published level.

A robust IPv4 series should therefore define contributor obligations. If a entity chooses to contribute a category, must it submit all qualifying transactions or a declared sample? How are cancelled and corrected events handled? Can the publisher verify consideration? Are related transactions tagged? What happens when the contributor's records contradict the submission?

Coverage metrics should reveal selective risk. Publish the number of contributors, contribution concentration, eligible events, included events and exclusions. If the eligible denominator is unknown, say so. A series that cannot measure its own selection should not be used as a contractual settlement price.

Assessment windows can be engineered

Every benchmark chooses time. A daily close, monthly average, rolling median and last transaction can produce different results from the same observations.

In a thin market, a entity can time a small transaction near the end of a window to influence a mark. A publisher can choose a cutoff that includes a desired trade and excludes a contrary one. A broker can report contract date for one transaction and registry completion date for another. The series appears systematic while the dates are not comparable.

The method should choose the economically relevant event in advance. Contract date measures when parties fixed price. Completion date measures when recognised delivery occurred. Quote date measures current interest. A market report can publish all three, but it should not switch among them to smooth or accentuate a trend.

Sparse windows need explicit rules. If no qualifying transaction occurred, the publisher can carry forward the previous value, use a model, widen an uncertainty band or publish no assessment. It should not silently transform an indication into a trade. A carried value must show its age.

Outliers require the same discipline. An unusual price can reflect a large block, distressed seller, reputation impairment or bundled service. Exclusion may be justified, but the reason must follow a published rule and remain in the audit trail. Otherwise “outlier” becomes a method for removing inconvenient evidence.

Corrections should be visible. If a price was entered in the wrong currency or a portfolio was double-counted, the revised series should identify the change, date and effect. Quiet historical rewriting allows the benchmark administrator to improve apparent performance after the fact.

Composition can imitate manipulation

Suppose the average unit price falls sharply in one month. That could mean every block class became cheaper. It could also mean the month contained more /16-and-larger transactions, which often trade at a different unit price from /22-/24 blocks. Without composition, the headline invites an accusation that sellers depressed the market.

Registry path can have the same effect. An inter-RIR transaction with more uncertainty may price differently from a straightforward intra-regional transfer. Legacy status, agreement obligations, fragmentation, reputation and continued use can alter value. A mixed average moves when the mix moves.

Price publishers should report stratified observations before a global average. Block-size bands are the minimum. Region path, transfer type and material quality flags should follow where sample size permits. A hedonic model can estimate like-for-like movement, but its variables, coefficients and missing-data treatment must be disclosed.

Small cells create confidentiality and reliability problems. Publishing an exact mean for one transaction identifies the transaction and gives it benchmark authority. The correct response may be a wider band, delayed publication or no value. “Limited public evidence independent observations” is more informative than a precise number built from one trade.

APNIC's annual address analysis uses IPv4.Global data to show long-term price patterns and notes variation by block size and market period. It is valuable secondary analysis of a broker dataset, not proof that every global transaction followed the same path. The data's source and coverage remain part of the interpretation.

Composition analysis protects legitimate volatility from being mislabelled as abuse. It also makes manipulation harder by preventing a entity from moving the headline merely by choosing which size class to submit.

Broker conflicts require capacity and remuneration disclosure

A broker can represent a seller, represent a buyer, act for both with consent, or trade as principal. Each model can be legitimate. Each creates different incentives.

An exclusive seller agent seeks the best acceptable result for the holder but may prefer a quick closing that earns commission. A buyer agent seeks suitable supply and may be paid according to savings or quantity. A dual agent holds both reservation prices. A principal dealer benefits if it buys lower and sells higher. A benchmark publisher may hold inventory whose valuation changes with the published series.

The market should not infer duty from the word broker. The engagement should state capacity, client, remuneration, exclusivity, principal interest, referral payments and whether the intermediary contributes data to a price report. If capacity changes during a transaction, the change needs consent.

Price reports should separate customer-to-customer agency events from principal purchases and resales. Counting both legs of a riskless principal sequence as independent demand doubles activity. Reporting only the higher customer resale can hide the seller's net result. Both observations may be useful if clearly linked and labelled.

Conflict management requires organisational controls where one firm combines brokerage, inventory, valuation and publishing. Data inclusion decisions should not be made by the trader whose book benefits. Method staff need independent reporting lines, documented overrides and review. Personal and corporate holdings relevant to the benchmark should be disclosed under a proportionate policy.

External audit should test process rather than promise the “correct” price. Did the publisher include all qualifying internal transactions? Were exclusions documented? Were related parties tagged? Did the method change after seeing the result? Those questions are answerable even when fair value remains a range.

Registry logs cannot reveal intent

RIR records are essential evidence of recognised registration. RIPE NCC transfer statistics identify blocks, parties, transfer type and processing date. ARIN publishes transfer files and procedures. These records can reveal repeated movement, rapid reversals and concentration patterns worth further inquiry.

They generally cannot establish manipulative intent. A transfer row does not show the quote that induced the bargain, the beneficial relationship, full consideration, broker capacity, side agreements or whether economic exposure changed. The registered organisations may be affiliates whose relationship is not apparent from names. One economic transaction may produce many rows.

Route data add another view but not a verdict. A block that changes origin near a transfer can indicate operational handover. Continued routing by the seller can indicate a leaseback, transition or delayed use. A new origin can be a customer or service provider. BGP does not identify purchase price or beneficial control.

RDAP, RPKI and IRR histories can corroborate timing and authority. Court and corporate records can reveal relationships. Payment and escrow records can show consideration. Communications can establish quote intent. A conduct finding requires the evidence chain, not one technical signal.

Registries should cooperate with lawful investigations and preserve accurate histories. They should not become price police merely because they operate the most authoritative public record. Expanding RIR review to commercial purpose could delay legitimate transfers without supplying the missing evidence of deception.

The right architecture links registry events to confidential market evidence under due process. It does not ask a registration desk to infer motive from a prefix.

Surveillance needs one event identity across many records

Manipulation detection fails when the same economic event appears under different names. A broker records a mandate. An escrow provider records a case. Two RIRs record prefix changes. A price publisher records a sale. Route collectors observe announcements. Without a common link, analysts either miss the sequence or count it several times.

A transaction family identifier can connect those records while preserving access controls. The identifier should be created when parties reach a binding commercial confirmation, with earlier quote and mandate identifiers linked to it. Every prefix leg, correction, reversal and operational handover retains the family link.

Party identifiers should distinguish legal entity, registry organisation and beneficial-control group. Public release can pseudonymise them. The surveillance layer needs relationship flags so it can separate independent, affiliate, principal-inventory and unknown cases.

Price evidence needs a money trail at the appropriate level. An escrow confirmation can verify that stated consideration moved without publishing bank details. If a bundled transaction allocates value to addresses, the method should identify who made the allocation and whether an independent valuation supported it.

Quote evidence needs message integrity. Timestamp, quantity, price basis, conditions, validity and withdrawal should be retained. A hash can prove that the preserved communication is unchanged, but it cannot prove that the quote was genuine. Capacity and conduct still require review.

The event identity is a coordination tool, not a global public ledger. Sensitive documents remain with authorised custodians. Investigators or auditors can obtain them under agreed rules. The public receives enough aggregated evidence to judge the benchmark without learning every party's strategy.

A manipulation screen should start narrow

A useful surveillance screen can rank cases without declaring guilt.

The first group concerns quote integrity: repeated high or low quotes withdrawn immediately after acceptance; quotes exceeding verified authority or funding; identical inventory represented as independent; and large changes in undisclosed conditions.

The second concerns relationship and circularity: transactions between common-control parties; rapid return of the same prefix; prices far from independent contemporaneous trades combined with side agreements; and multiple linked legs that leave ultimate exposure unchanged.

The third concerns benchmark interaction: a trade or quote near an assessment cutoff; a contributor with contracts, collateral marks or inventory tied to the level; selective submission after seeing other data; and unexplained overrides by assessment staff.

The fourth concerns record inconsistency: price reports that do not reconcile to confirmed quantity; several observations mapped to one transaction; registry transfer type inconsistent with claimed market sale; or a correction that removes an adverse point without published reason.

Every flag needs an innocent hypothesis. Urgent acquisition, block defects, restructuring, financing, failed delivery, currency error and data duplication can explain anomalies. The screen should request evidence and preserve response, not publish an accusation from an algorithm.

Thresholds should be calibrated on actual coverage. A “20 percent deviation” rule has no meaning if the reference price mixes unlike blocks or rests on one observation. Statistical rarity is not intent. The more heterogeneous and thin the class, the more the review must rely on facts.

Remedies should repair information before restricting exchange

When a quote was mislabelled, correct its status and remove it from executable depth. When a transaction was duplicated, link the records and revise the series. When a related-party price entered an arm's-length benchmark, reclassify it and publish the correction. When a contributor withheld qualifying trades contrary to the method, suspend its submissions and audit the affected period.

Serious intentional deception may justify contractual termination, loss of accreditation, civil claims or referral to a competent authority under applicable law. The response depends on jurisdiction and the legal character of the conduct. A private standards body should not invent criminal jurisdiction.

Due process matters. The subject should receive the evidence, methodology and opportunity to respond. Decision-makers should be independent of the transaction. Findings should distinguish proven facts, inference and unresolved uncertainty. Appeals should exist for accreditation and benchmark decisions.

The remedy should not hide historical prices. Corrected data should preserve the original publication and explain the change. Researchers need to know what entities saw at the time. Silent deletion creates a cleaner chart and a weaker audit trail.

Nor should a disputed transaction be erased from the registry merely because its price was misleading. If authority transferred validly, the registration event and the market-conduct issue are separate. Commercial or legal remedies can address deception without corrupting authoritative resource records.

Prohibiting asset treatment would target the wrong mechanism

Some critics respond to market abuse by denying that IPv4 rights should be sold, leased, valued or used in finance. That response confuses the existence of value with the integrity of price formation.

Scarcity exists because IPv4 has a finite address space and deployment demand persists. Transfer policies recognise that address capacity moves between organisations. Companies incur costs to acquire, lease, renumber and protect use. Courts, auditors, tax authorities and lenders may characterise the resulting rights differently, but economic value does not disappear when a registry rejects property language.

Manipulation can occur in markets for property, licences, services, securities and commodities. It can also occur in allocation systems through false need, collusion or strategic applications. Banning one label does not remove incentives to deceive.

A ban on sale can push transactions into mergers, leases, service contracts or offshore arrangements where price evidence is worse. A ban on publication can strengthen intermediaries who already possess private comparables. A ban on holding surplus can force distressed disposal and make the market easier for large buyers to influence.

The proportionate response is conduct-focused. Require truthful quote labels, authority evidence, relationship classification, benchmark governance and correction. Preserve legitimate confidentiality and strategic choice. Let legal systems determine property, tax and accounting consequences within their jurisdictions.

Market integrity improves when the rules identify the false signal. It deteriorates when they treat every valuable transaction as suspect.

Legitimate conduct can resemble the warning signs

Any surveillance design that ignores false positives will eventually deter the liquidity it is meant to protect. Several ordinary practices can look suspicious when only the public record is visible.

A seller may withdraw a quote after a buyer requests a different registry path, extends the delivery date or reveals a use that changes contractual risk. The visible sequence is offer and cancellation; the private sequence is a changed bargain. A buyer may bid through two brokers to ensure coverage, producing apparently duplicated demand even though it intends only one acquisition.

A block can move to an affiliate and return after a financing ends. Registration changes and circularity are present, but the transaction may have allocated collateral and control exactly as documented. A broker can buy as principal and resell quickly because it had a buyer ready; the rapid round trip supplies immediacy rather than fake volume if capacity and both prices are disclosed appropriately.

A transaction near a benchmark cutoff may occur because parties faced a quarter-end budget or court deadline. A price outside the recent range may reflect a clean, large aggregation or an impaired reputation history. A publisher can exclude it under a pre-existing quality rule without suppressing inconvenient evidence.

These explanations do not make review unnecessary. They define the evidence needed to close it. The reviewer should inspect quote conditions, mandate overlap, beneficial control, financing documents, payment path, capacity, benchmark method and contemporaneous reasons. If the explanation fits records created before the challenge, the anomaly should be resolved without stigma.

Surveillance outputs should therefore use stages such as data inconsistency, inquiry opened, explained, method correction and substantiated breach. A statistical flag is not a public finding. Aggregate reporting should include the number of flags cleared and the main innocent explanations so users can judge the screen's precision.

Protecting legitimate conduct is not leniency. It makes enforcement more credible by reserving strong conclusions for evidence of a false signal rather than market behaviour that merely looks unusual from outside.

Information barriers protect both the benchmark and the client

Where one firm brokers transactions and publishes prices, separation must operate in both directions. Traders should not learn confidential submissions from competitors through the assessment team. Assessors should not receive pressure to include a house bid that improves the firm's inventory mark. Sales staff should not promise clients a particular published level in exchange for business.

Access controls, separate reporting lines and logged communications can create that barrier. The benchmark team may receive validated transaction fields without client identity where identity is unnecessary. A compliance reviewer can test related-party status and completeness. Overrides require two-person approval and a recorded method reason.

Compensation matters. If an assessor's bonus depends directly on brokerage revenue or inventory value, formal separation can fail. Remuneration should reflect method compliance, timeliness and correction quality. Senior management should certify that commercial targets did not determine the level.

Client confidentiality remains important. A contributor should know who can see its data, how long records are retained, when an authority or auditor may obtain them and what aggregates can be published. Data gathered for integrity should not become a sales lead or negotiating advantage.

Whistleblowing and complaints complete the barrier. Staff and contributors need a route outside the commercial chain to report fabricated submissions, pressure or selective exclusion. The oversight body should publish how complaints are classified and resolved while protecting identities.

These controls are proportionate to the combined business model. A publisher that uses only public records faces different conflicts from one that brokers, holds inventory and values collateral. Governance should describe the actual conflict rather than repeat a generic independence statement.

Benchmark governance can be proportionate

Not every market newsletter needs the controls of a systemically important interest-rate benchmark. The intensity should follow use.

A broker commentary used only as general market colour can disclose its source, conflicts and limits. A valuation service used in financial statements needs documented data selection, comparable adjustments, review and retention. An index written into loans, earn-outs or investment contracts needs stronger independence, fallback rules, complaints, audit and contingency for sparse data.

IOSCO's benchmark principles support this proportionality. They emphasise administrator responsibility, oversight, conflicts, control framework, data sufficiency, methodology transparency, change procedures, complaints, audit and cooperation. They do not promise that every benchmark has one mathematically correct design.

An IPv4 benchmark statement should identify purpose and prohibited uses. It should say whether the level measures gross sale price, net seller proceeds, buyer cost or an assessed mid-market value. It should define size bands, regions, dates, fees, quality criteria and related-party treatment. It should publish fallback logic when transaction count is limited public evidence.

Oversight members should represent different interests and include independent methodology expertise. Contributors and traders can advise but should not control inclusion of their own data. Minutes should record conflicts and overrides without exposing transactions.

Annual assurance can test compliance with the published method. A more frequent review is warranted after material corrections or unusual concentration. The benchmark should publish a coverage report, not just an assurance badge.

Number Resource Society can support integrity if it accepts scrutiny

NRS states that it supports resource holders, promotes participation in Internet policy and advocates control of IP business assets. It also publishes material supporting sale and lease as monetisation strategies and describes market transparency and professional management as important.

That orientation can motivate useful work. NRS could convene a market-conduct code defining indication, executable quote, completed transaction, affiliate, principal trade, reversal and benchmark-eligible observation. It could make those definitions available to members and non-members.

It could sponsor confidential duplicate detection for listings, transaction-family identifiers and a voluntary related-party flag. It could fund independent research into quote failure, completion and benchmark coverage. Every result should report the cohort and avoid global extrapolation.

It could maintain a complaints mechanism for participating brokers and publishers, with independent review and published aggregate outcomes. Accreditation could require method compliance, record retention, conflict disclosure and corrections. It should not certify that a price is fair or that an address right has one universal legal status.

NRS's advocacy creates a reason for external safeguards. A board committee dominated by sellers or monetisation providers should not decide whether their prices enter a benchmark. Independent buyer, network-operator, statistical and public-interest voices are necessary. Funding and commercial relationships should be disclosed.

The society should also reject overreach. It should not call every registry restriction manipulation, every low bid confiscation or every high price proof of property. A credible institution applies the same evidence standard to conduct that helps and harms its preferred policy position.

A pilot should test the framework before making accusations

A twelve-month pilot could begin with a defined panel of brokers, direct holders, buyers, escrow providers and price publishers. Participation would be voluntary, but contributors would agree to submit all qualifying events in selected categories rather than only favourable examples.

The pilot would record indications, firm quotes, acceptance attempts, withdrawals, agreements, registry milestones, completion, reversals and operational handover. It would assign transaction-family and relationship classifications. Sensitive identities would remain with an independent administrator.

Published results would show unique inventory, independent quote count, quote-to-agreement conversion, agreement-to-completion, related and principal-trade shares within the panel, correction counts and benchmark coverage. These would be panel statistics, not global rates. Cases lacking relationship or price evidence would remain unknown.

The oversight group would review flagged patterns under a documented procedure. It would not publish party names or manipulation findings during the exploratory phase. Its first task would be testing whether the fields distinguish legitimate cancellation, restructuring and inventory trading from deceptive patterns.

Benchmark publishers could run their existing series beside an experimental method. Differences would be decomposed into block mix, date basis, related-party treatment, duplicate removal and judgment. The comparison would reveal which choices move the number without declaring one series fraudulent.

At the end, an independent evaluator would assess data quality, entity burden, confidentiality incidents, false positives and whether the framework improved price confidence. Rules would change before accreditation or public case reporting begins.

Integrity rests on knowing what an observation means

A thin IPv4 market will never look like a continuously traded large-cap stock. Supply is episodic, products are heterogeneous, and authoritative delivery crosses institutional boundaries. That makes single observations powerful and evidence labels essential.

The market should be sceptical of executable quotes that cannot be accepted, inventory that appears in several places, trades that return exposure to the same controller and benchmarks built from selective submissions. It should scrutinise publishers that broker, own and value the same inventory without separation. It should preserve messages, relationships, event families and corrections.

It should be equally sceptical of accusations based only on price. A premium can pay for urgency or quality. A discount can reflect size or distress. A large holder can change supply without deception. A broker can earn a spread for real risk. A related transaction can be operationally valid while excluded from an independent-price series.

RIR logs provide authoritative registration evidence and a durable historical spine. They do not contain the commercial facts needed to decide manipulation. IOSCO and FINRA comparisons show how transaction classification, methodology, conflicts and audit can improve other OTC or assessed markets, but their legal regimes do not automatically govern addresses.

NRS can help build the missing evidence layer if it accepts bounded authority and independent scrutiny. The objective is not to suppress value. It is to ensure that published value rests on transactions and quotes that mean what they claim.

The most defensible market policy is therefore precise: protect asset treatment and legitimate negotiation; expose false market signals; separate connected prices; govern benchmarks; preserve unknown denominators; and leave legal sanctions to competent authorities. Thinness calls for better information. It does not justify prohibition.

Sources