Open Ownership Register: Intelligence Source Guide

Open Ownership publishes the Beneficial Ownership Data Standard and an aggregated register expressing national declarations in it. It is the only serious attempt to make who-really-owns-this comparable across borders, and it inherits every weakness of the self-declared data it normalises.

At a glance

Source	Open Ownership Register
Category	Corporate, Ownership & Legal Records › Beneficial Ownership & Legal Entities
Homepage	https://www.openownership.org/
Machine interface	`https://register.openownership.org/`
Format	`HTML`
Access	Open — no account required
Disciplines	Corporate Intelligence
Mission domains	Financial Crime, Anti-Money Laundering, Corruption & Governance

Beneficial-ownership data standard & register. — as catalogued in the platform’s own source registry.

Open Ownership is a non-governmental organisation working on beneficial ownership transparency. It produces two things that matter operationally. The first is the Beneficial Ownership Data Standard, usually abbreviated BODS: an open specification, expressed as JSON with a published schema, for describing who ultimately owns or controls a company. Its central design decision is that the unit of publication is a statement — a dated, sourced assertion by an identified publisher — rather than a row in a table. An entity statement describes a company or arrangement; a person statement describes a natural person; an ownership-or-control statement links one to the other and describes the interest, its type, its size and its period of validity. Statements carry provenance: who asserted this, on what date, from what kind of source, and which earlier statement this replaces. The second product is the Open Ownership Register, an aggregation that takes data from the small number of national registers that publish beneficial ownership openly, transforms it into the standard, and makes it searchable across jurisdictions. The organisation's emphasis has moved over time toward the standard itself and toward supporting governments implementing registers, so confirm the register's current scope, sources and status on the site before you build a pipeline that depends on it.

Corporate registers tell you the legal owner, which in any structure designed to obscure control is a nominee, a holding company or a trustee. Beneficial ownership data is supposed to tell you the natural person at the end of the chain. The analytical job Open Ownership does that nobody else does is make those declarations comparable: the same interest type, the same share representation, the same treatment of indirect ownership, and the same provenance model whether the declaration came from a European register, a resource-sector disclosure regime or a public procurement rule. Without that layer, cross-border ownership analysis is a manual reconciliation exercise repeated for every jurisdiction. With it, a chain can be traversed as a graph. The second contribution is conceptual and underrated: BODS forces the publisher to say what it does not know. A person can be declared as unknown or unspecified, with a reason; an interest can be declared as a range rather than a value; a statement can be marked as replacing an earlier one. Most ownership datasets present absence as though it were a fact. This one has a vocabulary for ignorance, and using that vocabulary properly is what separates a defensible ownership finding from a plausible-looking chart.

Who publishes it, and why that matters

Open Ownership is a small non-profit funded by philanthropic foundations and development donors, working closely with governments implementing transparency commitments. That gives it credibility with the register operators whose data it depends on and a genuine influence on how national registers are designed — several jurisdictions have adopted or aligned with the standard. It also means its priorities follow the transparency policy agenda rather than an investigator's needs: the organisation exists to improve registers, not to run a lookup service, and the aggregated register has always been a demonstration of what the standard makes possible as much as an end product in itself. Practically, this implies three things. The standard is the durable asset and is versioned, documented and stable enough to build on. The aggregated data is a convenience that depends on which governments publish openly, which is a political variable outside anyone's control. And the organisation's continuity depends on donor funding in a field where donor priorities have shifted more than once. Build against the standard; treat any specific aggregation as a source that may change scope.

Provenance is the first question to ask of any dataset and the one most often skipped. Who collects it, what their incentive is, whether they publish a methodology, and whether they correct the record when they get something wrong all bear directly on how much weight a finding drawn from it can carry.

What a record actually contains

The fields you will be working with, what each one means, and whether it is something you can pivot on. Read the meanings carefully — more analysis is wrecked by misreading a field than by failing to find one, and a field that looks like an observation is often an inference.

Field	Type	What it means	Pivot value
`statementId`	string	Identifier for a single dated assertion, generated deterministically from the statement's content so that identical content yields the same identifier. Statements are immutable; a correction is a new statement, not an edit.	Change tracking, and linking a claim to the exact publication that made it.
`statementDate`	timestamp	The date the assertion was made or the date it describes, depending on the publisher's convention. It is not necessarily the date the ownership began, which is carried separately on the interest.	Temporal ordering of declarations; detecting how long a declaration has stood unrevised.
`recordType / statementType`	enum	Whether the statement describes an entity, a person, or a relationship of ownership or control. The three are separate objects and a complete picture requires all three.	Graph construction: entities and persons are nodes, ownership statements are edges.
`entityType`	enum	For entity statements: a registered company, another legal entity, a legal arrangement such as a trust, or an entity that is anonymous or unknown. The arrangement type is where trusts and partnerships live, and it behaves differently from a company.	Structural analysis; identifying where a chain passes through an arrangement whose disclosure rules differ.
`identifiers`	array	Identifiers for an entity with an explicit scheme, such as a national company register code or an LEI. The scheme is what makes the identifier resolvable rather than merely a number.	The join into company registers, GLEIF and sanctions data — this is the field that connects ownership data to everything else.
`personType`	enum	Whether the person is known and named, anonymous because the publisher withheld identity, or unknown because the declaration itself could not identify them. The distinction between anonymous and unknown is analytically crucial and constantly collapsed.	Gap analysis: unknown persons mark where the ownership chain actually fails, which is different from where publication was restricted.
`names`	array	Person or entity names with a type qualifier, in the script and form the source register used. Transliteration, name order and the treatment of patronymics vary by jurisdiction and are a persistent matching hazard.	Sanctions and PEP screening — with the understanding that name-only matching across scripts and cultures produces both false hits and misses at high rates.
`birthDate`	string	Date of birth, commonly published at reduced precision such as month and year only, because full dates are treated as excessive disclosure in most regimes.	Disambiguation between people with common names; the single most useful non-name attribute, and the one most often truncated.
`nationalities / addresses`	array	Declared nationality and address, usually a correspondence or service address rather than a residence. Multiple nationalities are recorded where declared and are a meaningful signal in themselves.	Jurisdictional analysis, citizenship-by-investment research, and address clustering onto corporate service providers.
`interests`	array	The nature of the interest: shareholding, voting rights, appointment of the board, other significant influence or control, or a role in an arrangement such as settlor, trustee or beneficiary. One person can hold several interests in the same entity.	Control analysis — voting rights and board appointment frequently matter more than shareholding, and are held separately in real structures.
`interest share`	string	The size of the interest, expressed as an exact figure or as a band with minimum and maximum. Most registers publish bands, because that is what their thresholds require.	Aggregation across a chain — but band arithmetic compounds uncertainty quickly and an indirect stake computed from bands is a range, not a number.
`interestLevel`	enum	Whether the interest is held directly or indirectly through intermediate entities. Indirect declarations conceal the chain, so an indirect interest is a claim about an end state without the path that produced it.	Deciding whether the chain in your graph is documented or asserted.
`startDate / endDate`	timestamp	The period over which the interest is declared to subsist. An open end date means current as far as the declaration goes, which is not the same as verified today.	Point-in-time ownership reconstruction, which is what most legal and sanctions questions actually require.
`source`	string	Where the assertion came from – an official register, a self-declaration, or a third party – with a retrieval date, alongside the publication details naming the publishing organisation, standard version and licence. This distinguishes who asserted a fact from who republished it, which matters whenever an aggregator sits between you and the register.	Evidence assessment; separating what a state register recorded from what a company said about itself.

Coverage — and what is not in it

Coverage is determined by politics, not by technology, and it is much narrower than the ambition suggests. Only a small number of jurisdictions publish beneficial ownership data openly enough to be aggregated at all, and the set has changed in both directions. The most consequential event in this field was the November 2022 judgment of the Court of Justice of the European Union invalidating the provision of the anti-money-laundering directive that gave the general public access to member states' beneficial ownership registers; several European registers closed public access immediately, and access has since been restructured around a legitimate-interest test under the European Union's subsequent anti-money-laundering package, which journalists and civil society organisations can meet but which is administered differently in each member state. The practical consequence is that a European ownership dataset assembled before that judgment describes an access regime that no longer exists. Outside Europe, open publication exists in a scattering of countries driven by extractives transparency commitments, procurement rules and open government programmes. Temporal coverage is shallow: most registers are less than a decade old, most publish current state rather than history, and reconstructing ownership as at a past date is usually impossible from the published data. Update rhythm follows each source register, which varies from continuous to annual.

Known blind spots

Absence of evidence here is not evidence of absence. These are the conditions under which Open Ownership Register will not show you something that is nevertheless real:

Declarations are self-reported and, in most jurisdictions, unverified at the point of filing. The register records what someone said, and a person willing to conceal ownership is not deterred by a form; verification regimes are only now being introduced in a few countries and are not retroactive.
Thresholds create structural invisibility. Where the reporting trigger is a quarter of shares or votes, four people holding equal stakes disclose nothing, and structuring below the threshold is trivial and undetectable from the data.
Nominee and proxy arrangements defeat the declaration entirely. A named beneficial owner who is acting for someone else appears in the data as the answer, and nothing in the record indicates otherwise.
Trusts and similar arrangements are handled inconsistently or excluded. The roles that matter — settlor, protector, beneficiary — are exactly the ones many regimes do not require to be published, so a chain passing through an arrangement usually terminates there.
Public access has been restricted rather than expanded in a major jurisdiction. Following the 2022 European court judgment, general public access to member state registers ended and was replaced by a legitimate-interest regime, so what was once bulk-accessible now requires an application in each country.
Exemptions remove significant populations. Listed companies, regulated financial institutions and, in some regimes, entities whose owners face a demonstrated risk of harm are excluded from publication, which removes both benign and interesting cases.
Data quality in the underlying registers is poor in well-documented ways: implausible dates of birth, obviously false names, entities declaring themselves as their own owners, and large volumes of filings that no system ever checked.
Historical reconstruction is generally impossible. Registers publish current state, corrections overwrite, and unless you archived the data yourself you cannot establish who was declared as owner on a past date — which is what sanctions, procurement and litigation questions actually need.
Coverage is absent in precisely the jurisdictions most used for concealment. The secrecy jurisdictions that matter most either have no register, no public access, or a register whose contents are not independently checkable.

Write the blind spot into the product. A statement that something “was not observed in Open Ownership Register” is defensible; a statement that it “did not happen” is not, and the difference is what survives cross-examination.

Access, licensing and what you may do with it

Access model: Open — no account required

There are two distinct things to access and they should not be confused. The standard — schema, documentation, codelists and worked examples — is published openly and is the durable asset; build your data model against it regardless of where your ownership data comes from, because a growing number of national registers publish in it or map to it. The aggregated register is a web application with search and per-entity views, and its bulk data availability, source list and status have changed over the project's life, so check what is currently offered rather than assuming. For serious work, the more robust architecture is to consume national registers directly where they publish openly, transform to the standard yourself, and use the aggregation as a cross-check and for discovery. Where a European register now operates a legitimate-interest access regime, the application route is national and the criteria differ; journalists and civil society organisations generally qualify but must apply, and building that into an investigative timeline is a project management problem as much as a legal one.

Licence

The standard is published as an open specification intended for adoption and reuse, and implementing it carries no licensing burden. The data is a different matter and it is genuinely layered: each source register applies its own licence and re-use conditions to its data, the aggregator applies terms to its transformation and publication, and BODS itself carries a licence field on every statement precisely because the answer varies per statement. Read that field rather than assuming a blanket permission. Several national registers permit free re-use with attribution; others restrict bulk redistribution, commercial use, or use for direct marketing and profiling; and access obtained under a legitimate-interest regime typically comes with conditions on onward disclosure that are stricter than an open licence. If you are redistributing or building a commercial product, the licence position must be resolved per source, not per aggregate. This is tedious and it is also the difference between a defensible product and one that quietly breaches a dozen register terms at once.

Rate limits and fair use

The register is a small non-profit's web application; treat it accordingly. Do not scrape it to reconstruct a dataset — where bulk data is offered, take that instead, and where it is not, go to the source registers. For interactive use, human-cadence querying with an identifying user agent is fine. National registers have their own limits and several enforce them strictly; the ones that offer bulk downloads always prefer you use them over hitting their search endpoints, and the ones that do not offer bulk downloads usually have a reason.

Licensing changes, and it changes without warning. A dataset that was free for research this year may not be free for commercial or evidential use next year. Confirm the current terms before you build a dependency on it, and record the terms you relied on alongside the data — the licence in force at the time of collection is part of the provenance.

Collecting it

How Open Ownership Register is actually pulled, in the order you would set it up. Prefer the bulk or export interface over per-item lookups wherever one exists: it is kinder to the publisher, faster for you, and gives a reproducible snapshot rather than a series of point-in-time answers you cannot reconstruct later.

Method	Format	Cadence	Notes
Standard and codelist retrieval	JSON	per standard release	Take the schema, codelists and documentation and pin the version. The standard has evolved through significant structural revisions, and code written against one version will not consume another without work.
Aggregated register access	HTML	ad hoc	Use for discovery and cross-jurisdiction search on specific entities and people. Verify anything found here against the originating national register before it carries weight.
Direct national register collection	CSV	per register, daily to annual	The robust route. Consume the source registers that publish openly, in their native format, and transform to the standard yourself so that you control provenance and timing.
Legitimate-interest access applications	HTML	per jurisdiction, as needed	Where European registers now operate restricted access, the route is a national application. Track applications, grants and conditions as a managed process, because the conditions attach to your subsequent use.
Snapshot archiving	JSONL	per collection cycle	Store every collection as a dated snapshot. Registers overwrite corrections and publish current state only, so your own archive is the sole route to point-in-time ownership, and it cannot be created retroactively.

Ingesting it into the platform

Every step below is idempotent and cursor-based: interrupt one and it resumes from where it stopped rather than duplicating rows or losing progress. Collection is recorded per source, so a feed that quietly stops publishing shows up as a stale timestamp instead of silently thinning your coverage.

Register sources individually — Record each contributing national register in sources.php as its own source with its own licence, cadence, verification regime and access basis, rather than registering the aggregation as one source. The provenance differences between registers are larger than the similarities.
Model statements, not rows — Import through import.php preserving the statement structure: an assertion by a publisher on a date from a source, with its own identifier. Flattening statements into a current-state table at ingest destroys the provenance and change history that make this data defensible.
Keep person, entity and interest as separate objects — During ingest.php, create person and entity nodes and ownership edges distinctly in link-analysis.php, with the interest type and share band carried on the edge. Merging an interest into a node attribute makes multi-interest holdings unrepresentable.
Preserve unknown and anonymous as values — Load unknown and anonymous person statements as records rather than dropping them. They are the map of where the ownership chain fails, and a graph that silently omits them looks complete and is not.
Resolve entities against identifier schemes — Run resolve-everything.php to match entity identifiers against company registers and GLEIF using the scheme-qualified identifier rather than the name, so that ownership data joins to the rest of the corporate picture on a stable key.
Screen persons against sanctions and PEP data — Match declared beneficial owners against sanctions and politically exposed person sources, retaining match scores and the attributes used, so a name-only match is never presented with the same confidence as one supported by date of birth and nationality.
Compute chains with explicit uncertainty — Where indirect ownership is derived by multiplying share bands along a chain, carry the resulting range rather than a point estimate, and mark any chain that passes through an unknown person or an arrangement as broken rather than continuing through it.
Snapshot and version — Persist each collection as a dated version so that timeline.php can answer who was declared as beneficial owner at a past date. This is the capability the source registers do not provide and the one most legal and sanctions questions require.

Registered sources and their last-collected state are listed in sources.php, and the scheduled chain that keeps them current is in automation.php.

How it is wrong, and how to tell

Every dataset is wrong in characteristic ways. Knowing which ways is the difference between using a source and being used by one, and it is the part of source evaluation most often skipped because it is the part that takes work.

Assess this source in two layers, because they have opposite characteristics. The standard is high quality: it is well specified, documented, versioned, developed openly with implementers, and it models uncertainty and provenance better than any competing approach. If you build your ownership data model on it you will not regret it. The data is only as good as the register it came from, and most registers rest on unverified self-declaration. The documented failure modes in the largest open registers include implausible dates of birth, obviously fictitious names, circular declarations and very large volumes of filings that no human or system reviewed. Verification regimes — identity checking of filers, cross-referencing against other government data — are being introduced in a few jurisdictions and are neither universal nor retroactive. The aggregation layer adds transformation risk on top: mapping a national schema into the standard involves judgement calls, and a mapping error propagates silently. The correct posture is to treat any beneficial ownership declaration as a lead with a named source and a date, to verify anything that matters against the originating register, and to treat convergence between a declaration and independent evidence as the actual finding rather than the declaration alone.

Characteristic false positives

The declared beneficial owner is a nominee. The data shows a named natural person with an apparently complete chain, and the record contains nothing that would distinguish a real owner from someone holding on another's behalf.
Threshold structuring produces a legitimately empty record. An entity with no declared beneficial owner may be evenly held below the reporting trigger, which looks identical to non-compliance and to an exemption, and the data cannot tell you which.
Name-only matching against sanctions and PEP lists produces both false positives and false negatives at high rates, particularly across scripts and naming conventions, and truncated dates of birth remove the attribute that would resolve most of them.
Indirect interest computed from bands presented as a number. Multiplying two twenty-five-to-fifty-per-cent bands yields a range from six to twenty-five per cent, and reporting the midpoint as an ownership percentage is a fabrication with arithmetic attached.
Anonymous confused with unknown. A person withheld from publication for protection reasons and a person the declaration could not identify are entirely different findings, and collapsing them produces either a false gap or a false certainty.
Stale declarations read as current. An open-ended interest with no end date means nobody has filed a change, not that anyone has confirmed the position still holds, and in registers without periodic confirmation that can persist for years.
Transformation artefacts from schema mapping. Where an aggregator maps a national schema into the standard, interest types and share representations that do not map cleanly are approximated, and the approximation is invisible downstream.
Entity resolution merging distinct companies. Similar names across jurisdictions, and the same name reused after dissolution, cause ownership chains to be joined at a node that represents two different companies — which produces a network that is both wrong and compelling.

None of these make the source unusable. They make it a source that requires corroboration before an assertion built on it goes into a product, which is true of every source and admitted by few.

Ageing

Beneficial ownership data ages badly and gives almost no signal that it has. Ownership changes on transaction dates that nobody is obliged to publish promptly; filing deadlines in most regimes are measured in weeks and compliance is imperfect; and a register that requires only event-driven updates will carry a declaration indefinitely until someone files a change. An interest with no end date is therefore an assertion that nothing has been reported, not that anything has been confirmed. Person attributes age too: names change, nationalities are acquired, addresses move, and service addresses at corporate providers persist long after the relationship ends. The most consequential ageing is at the regime level — access rules, verification requirements and publication scope have all changed materially in recent years, so a methodology built on what a register offered three years ago may not be executable today. A stale usage looks like a due diligence report asserting current ownership from a declaration of unknown vintage, or a chain traversed through an entity that has since been dissolved. Detect it by displaying the statement date and source retrieval date beside every ownership claim, and by re-collecting at the point of use rather than relying on your store.

What this source feeds

A source is only worth what it lets you conclude. These are the disciplines that collect through it, the mission domains it serves and the data points it yields — every one is a tag, so you can follow any thread from here into the rest of the library.

Collected by these intelligence disciplines

Corporate Intelligence

Serves these mission domains

Yields these data points

How each sector uses Open Ownership Register

The same dataset is worked very differently depending on who you are, what authority you hold, and what you are ultimately producing. A military analyst is supporting a commander’s decision; a journalist is meeting a publication standard; an NGO caseworker is protecting a person. The records are shared — the constraints, thresholds and outputs are not.

🎖 Military and defence

The application is procurement and supply chain assurance: establishing whether a supplier, sub-tier vendor or research partner is ultimately controlled by an entity or person of concern. Beneficial ownership declarations are the only public data that even attempts to answer that, and the standard's interest types matter here because control through voting rights or board appointment is more relevant to a foreign ownership assessment than a shareholding percentage. The limits are severe and should be stated plainly in any assessment: coverage is confined to jurisdictions that publish openly, declarations are largely unverified, thresholds make minority control invisible, and the jurisdictions most likely to be used for concealment are absent. Use it as one input to a foreign ownership, control or influence assessment, and never as the determination. Where a chain terminates in an unknown or an arrangement, that termination is itself a finding worth escalating.

🕵 National intelligence

For financial intelligence and sanctions work, this is the open-source layer that connects a corporate entity to a natural person, which is where designation, asset tracing and ownership-test analysis all lead. The standard's provenance model is what makes it usable in an assessment: every statement says who asserted it, when and from what source, so you can characterise the evidence rather than inheriting it. The ownership tests in most sanctions regimes turn on control and on aggregated indirect holdings, which is exactly where share bands and undeclared chains defeat you, so the honest output is frequently a range and a set of unresolved nodes rather than a determination. Two structural points deserve to be in any assessment using this data: declarations are unverified in most jurisdictions, and European public access was curtailed by the 2022 court judgment and reconstituted on a legitimate-interest basis, which changes what can be collected and how.

👮 Law enforcement

For money laundering, corruption and asset recovery investigations, beneficial ownership registers are a starting point and a source of leads, not evidence. The declaration is a third-party assertion, frequently unverified, and where a case depends on ownership you will obtain certified extracts, banking records and the filing history through legal process. What the data does well is direct that process: it tells you which entity to serve, which jurisdiction to request assistance from, and which natural person to place at the centre of a request. False or misleading declarations are themselves offences in a growing number of jurisdictions, which makes a discrepancy between a filing and the true position a chargeable matter rather than merely an analytical inconvenience. Where access is restricted under a legitimate-interest regime, law enforcement generally has a separate and broader access route through the register operator; use it rather than the public interface.

🔍 Private investigation and corporate security

This is the free source that answers the question clients most often ask and most often misunderstand. It will tell you what was declared; it will not tell you who owns anything. Communicate that distinction explicitly in writing, because a due diligence report that presents a declaration as a finding creates a liability when the nominee behind it surfaces. Practically, the highest-value uses are identifying where a chain terminates in an unknown person, spotting service-address clustering that indicates a corporate service provider, and establishing the identifiers needed to order documents from the primary registers. Note the access change in Europe: general public access ended following the 2022 court judgment, and a commercial investigative firm may or may not meet the legitimate-interest test in a given member state, which is a matter to resolve before promising a client coverage you cannot deliver.

📰 Journalism and OSINT media

Beneficial ownership data has driven a great deal of accountability journalism, and the standard's provenance model is well suited to the evidentiary discipline good reporting requires — you can say precisely who declared what, when, and to whom. The most productive stories are usually about the failures: entities with no declared owner, implausible declarations, chains terminating in secrecy jurisdictions, and companies winning public contracts whose ownership is undeclared. Two things to carry into the newsroom. The declaration is not proof; corroborate before you publish an ownership claim about a named individual, because the person named may be a nominee or the filing may simply be wrong. And the access position in Europe changed in 2022: journalists can generally meet the legitimate-interest test, but it must be applied for, member state by member state, and that lead time belongs in the reporting plan rather than at the end of it.

🌍 NGO, humanitarian and human rights

Anti-corruption, tax justice and extractives transparency organisations are both the users and the constituency for this work, and the standard exists largely because of their advocacy. The data supports procurement monitoring, resource contract analysis, and the identification of politically exposed persons behind licence holders and government suppliers. It is also evidence about itself: the proportion of entities with no declared owner, with unknown persons, or with declarations that fail plausibility checks is a directly quantified measure of a register's effectiveness, and is more persuasive to a reform audience than any argument from principle. For casework involving individuals, apply the same care as with any personal data about people who may face retaliation, and note that some regimes permit protection-based exemptions from publication for exactly that reason — an anonymous person statement may be protecting someone rather than hiding them.

🎓 University and research

For research in corporate governance, illicit finance, tax and political economy, this is the most usable public beneficial ownership resource, and the standard makes cross-jurisdiction comparison tractable in a way that raw register dumps do not. The research designs that work best exploit the data's own error structure: measuring declaration quality, threshold effects, compliance rates and the effect of verification regimes tells you about the policy instrument, which is a better-identified question than trying to measure hidden ownership from data designed to be evadable. Model the selection explicitly — only some jurisdictions publish, publication regimes changed materially in 2022, and coverage correlates with the transparency commitments that are themselves the treatment in many designs. Ethics review is warranted: the data concerns identifiable natural persons, including people in jurisdictions where being named as an owner carries real risk.

Playbook: working Open Ownership Register end to end

A repeatable sequence from first pull to finished product. Each phase states what you are trying to establish, not merely what to click — the objective is a defensible chain of reasoning, not a completed checklist.

Phase 1 — Decide what standard of proof the question requires

Establish at the outset whether you need a lead, an assessment or evidence. Beneficial ownership declarations are unverified assertions in most jurisdictions and are excellent leads, adequate for assessment with caveats, and insufficient as evidence. Setting this before collection prevents the familiar failure of a compelling chart that cannot support the conclusion drawn from it.

Phase 2 — Establish the access regime for each jurisdiction

Determine, per country, whether the register is publicly accessible, restricted to legitimate interest, or closed. In Europe this changed materially after the 2022 court judgment. Where an application is required, start it immediately — the lead time is measured in weeks and it will otherwise become the critical path.

Phase 3 — Pin the standard version and model statements natively

Build against a specific published version of the standard, preserving statements as immutable dated assertions rather than flattening to current state. The structural revisions between versions are significant, and code that assumes a shape will break silently on data published under another.

Phase 4 — Collect from source registers, not only the aggregation

Where a register publishes openly, take it directly and transform it yourself. This gives you control of timing, provenance and licence position, and it removes a transformation layer in which mapping errors are invisible. Use the aggregation for discovery and for jurisdictions you cannot collect yourself.

Phase 5 — Snapshot before you analyse

Archive every collection as a dated version. Registers publish current state and overwrite corrections, so point-in-time ownership — which is what sanctions tests, procurement rules and litigation actually ask about — exists only if you built it. It cannot be recovered later.

Phase 6 — Resolve entities on scheme-qualified identifiers

Join companies using the register scheme and identifier, not the name. Cross-jurisdiction name matching merges distinct companies and splits single ones, and an ownership graph with a wrongly merged node is worse than no graph because it is persuasive.

Phase 7 — Traverse the chain and mark every failure honestly

Walk from the entity toward natural persons, and where the chain reaches an unknown person, an anonymous person, an arrangement or a jurisdiction with no register, stop and label the node as a termination with its reason. Chains that quietly continue through gaps are the standard way ownership analysis goes wrong.

Phase 8 — Carry uncertainty through the arithmetic

Where indirect holdings are computed from share bands, propagate ranges rather than point estimates and present the result as a range. If your conclusion requires the midpoint, your conclusion is not supported by the data and needs another source.

Phase 9 — Screen persons with more than a name

Match declared owners against sanctions and PEP data using date of birth, nationality and address alongside the name, and record which attributes supported each match. Name-only screening on this data produces both false hits and misses at rates that will embarrass you in either direction.

Phase 10 — Test declarations against independent evidence

Check the declared position against company filings, land and vessel registers, procurement awards, court records and press reporting. The finding is the convergence or the divergence; a declaration standing alone is a claim, and a divergence between a filing and reality is frequently an offence in its own right.

Phase 11 — Escalate the gaps as findings

Entities with no declared owner, unknown persons, chains terminating in secrecy jurisdictions and implausible declarations are results, not failures. In procurement, sanctions and public contracting contexts they are precisely what the enquiry should surface, and reporting them is more useful than a partial chain presented as complete.

Phase 12 — Record with provenance and access basis intact

File the analysis in cases.php with each statement's publisher, source type, statement date and retrieval date preserved, and with the legal basis on which restricted-access data was obtained. Ownership findings get litigated, and a record that cannot say where a claim came from cannot defend it.

The platform ships this as a step-checked workflow in playbooks.php, so progress is recorded against a case rather than held in someone’s head.

What to pair it with

No single source carries a finding. These are the datasets that corroborate, extend or contradict this one — and a source that contradicts is worth more than one that agrees, because it is the only thing that will tell you when you are wrong.

Source	Relationship	What it adds
OpenCorporates	prerequisite	Company register data across jurisdictions, including officers and filings. The legal-ownership and corporate-existence layer that beneficial ownership declarations sit on top of.
GLEIF LEI	corroborates	Global entity identifiers with the local register code and company number, plus accounting consolidation relationships that can be tested against declared ownership.
OpenSanctions	extends	Consolidated sanctions, PEP and watchlist data with entity resolution, the standard screening layer for declared beneficial owners.
ICIJ Offshore Leaks Database	extends	Leaked offshore entity, officer and intermediary records covering the structures that beneficial ownership registers were created in response to and still largely do not reach.
OCCRP Aleph	extends	Cross-source search over registers, leaks and documents, useful for testing whether a declared owner appears in investigative material that the register knows nothing about.
Companies House	prerequisite	One of the largest openly published beneficial ownership regimes, and the reference case for both what open publication makes possible and how unverified self-declaration fails.
Financial Action Task Force	prerequisite	Sets the international standards on beneficial ownership transparency that national registers implement, and the source for what a compliant regime is supposed to require.
Extractive Industries Transparency Initiative	extends	Requires beneficial ownership disclosure for companies holding extractive licences, and is the reason several countries outside Europe publish this data at all.
Beneficial Ownership Data Standard	prerequisite	The specification itself, with schema, codelists and worked examples. Build your data model here regardless of which registers you consume.

Legal, ethical and operational constraints

This data is about identifiable natural persons and their financial interests, which places it squarely inside data protection law in most jurisdictions and inside heightened scrutiny in Europe. The 2022 Court of Justice judgment invalidating general public access to member state beneficial ownership registers turned on exactly this: the court held that indiscriminate public access was a disproportionate interference with privacy and data protection rights. Access has since been reconstructed around a legitimate-interest test, and if you are processing European beneficial ownership data you need a lawful basis, a purpose that would satisfy that test, and compliance with whatever conditions were attached to your access grant — including limits on onward disclosure. Beyond data protection, three constraints recur. Register licences differ per source and frequently restrict bulk redistribution and profiling. Publishing an allegation that a named individual is the true owner of an entity, where the register says otherwise, is a defamation exposure that requires evidence rather than inference. And in some jurisdictions individuals may be exempted from publication because disclosure would expose them to harm; re-identifying such a person from other sources may be lawful and is rarely defensible.

Operational security

Querying a beneficial ownership register is a targeted act that names the company or person you are interested in, and those queries are logged. Where access is granted under a legitimate-interest regime, your identity, your organisation and your stated purpose are on file with a public authority, and in some jurisdictions that record is discoverable or shareable. For sensitive corporate investigations, assume the subject may learn of the enquiry — some registers notify, some leak, and professional intermediaries talk. Mitigations are the ordinary ones: prefer bulk collection over targeted queries where bulk is available, separate research infrastructure from your organisation's main network, and stage enquiries so that the most sensitive lookups happen last, after the target's attention is least likely to be raised. Downstream, remember that an ownership graph you publish reveals your methodology and your scope as clearly as it reveals the structure, and that named individuals in it may face consequences from third parties that your analysis never contemplated.

Two rules that hold regardless of jurisdiction. Collection that is lawful is not automatically proportionate, and a dataset assembled for one purpose does not carry consent for another. Where the records concern identifiable people, the question is not only whether you may hold the data but whether holding it serves the purpose you are accountable for.

Is it earning its place?

Sources accumulate. Feeds get added during an incident and are never reviewed again, and a decade later the pipeline is carrying dead weight that nobody dares remove. These are the measures that show whether Open Ownership Register is contributing anything, and they are worth baselining now so the answer is available later.

Share of entities in your working set with any declared beneficial owner, per jurisdiction, which measures the register's reach rather than the ownership picture and is the honest denominator for everything else.
Proportion of ownership chains that terminate in a named natural person, versus those terminating in an unknown person, an arrangement or a jurisdiction with no register — the direct measure of how much of your analysis is actually resolved.
Plausibility failure rate in declarations you process: impossible dates of birth, circular ownership, obviously fictitious names, as a running indicator of the quality of the underlying register.
Verification rate: the proportion of load-bearing declarations you confirmed against an independent source, which should be complete for anything appearing in a client or legal deliverable.
Age distribution of the statements in your store, measured from statement date rather than collection date, since an old declaration and a fresh collection look identical if you only track the latter.
Number of dated snapshots retained per source register, which determines whether you can answer a point-in-time ownership question at all.
Screening precision on declared persons, tracked separately for name-only matches and for matches supported by date of birth or nationality, because the two have completely different error profiles.

Beware of volume. Indicator counts rise easily and say almost nothing. Unique contribution — findings this source produced that no other source in your stack would have — is the measure that matters, and it is usually far lower than anyone expects.

Tradecraft notes

The distinctions that separate a competent analyst from a fast one:

A declaration is not a finding. In most jurisdictions nobody checked it, and the correct output of an ownership search is 'X was declared as beneficial owner on date D by publisher P', which is a materially weaker and more honest claim.
The gaps are the product. Unknown persons, absent declarations, chains ending in arrangements and threshold-structured holdings are what an ownership investigation is looking for, and a workflow that treats them as missing data throws away the finding.
Distinguish anonymous from unknown every time. One is a publisher withholding a name, often for protection; the other is a chain that failed. Collapsing them produces either a false gap or false confidence, and the vocabulary exists precisely to prevent that.
Control is not shareholding. Voting rights, board appointment rights and other significant influence are separately declarable and are how real control is usually held; an analysis that ranks by percentage held will miss the person who actually decides.
Propagate ranges through indirect chains. Multiplying bands produces a band, and reporting a midpoint as a percentage is a fabricated precision that will not survive scrutiny.
Snapshot before you need to. Registers publish current state and overwrite; point-in-time ownership is the question everyone eventually asks and it can only be answered from archives you started keeping earlier.
Join on scheme-qualified identifiers, never on names. A wrongly merged company node in an ownership graph is more damaging than a missing one, because it is persuasive and invisible.
Re-check the access regime before every project. This field's rules have changed sharply and asymmetrically since 2022, and a methodology that worked last year may now require an application in each member state.
Treat the standard as the durable asset and any particular aggregation as contingent. Registers open and close for political reasons; a data model built on the standard survives that, and one built on a specific aggregator's schema does not.

Questions analysts actually ask

Does this tell me who really owns a company?

It tells you who was declared as owner, by whom, and when. In most jurisdictions that declaration is unverified, thresholds hide minority holdings, and nominee arrangements defeat it entirely. Treat it as a sourced lead, corroborate before relying on it, and describe it in those terms in any deliverable.

Why can I no longer access European beneficial ownership registers freely?

The Court of Justice of the European Union held in November 2022 that general public access under the anti-money-laundering directive was a disproportionate interference with privacy rights, and several registers closed public access. Access has since been rebuilt around a legitimate-interest test administered nationally, which journalists and civil society organisations can generally meet but must apply for.

What does it mean when an entity has no declared beneficial owner?

Any of several very different things: the ownership is structured below the reporting threshold, the entity is exempt, no owner could be identified, or nobody filed. The data usually cannot distinguish them, and that ambiguity is itself worth reporting rather than resolving by assumption.

Is this data verified?

Mostly not. Registers have historically accepted filings without checking identity or substance, which is why they contain implausible dates of birth and obviously false names. Identity verification requirements are being introduced in a few jurisdictions, are recent, and are not retroactive to existing filings.

Should I use the aggregated register or the national ones?

National registers where you can, for control over provenance, timing and licence, and the aggregation for discovery and for jurisdictions you cannot collect directly. Always verify a finding against the originating register before it carries weight.

Can I compute an indirect ownership percentage through a chain?

You can compute a range. Most registers publish share bands, and multiplying bands produces wide ranges quickly. If your conclusion depends on a specific percentage — as sanctions ownership tests often do — you need filings or transaction records, not declarations.

Why does the standard have a category for an unknown person?

Because ownership data is frequently incomplete and the alternative is silence that reads as absence. Being able to state that a person exists but could not be identified, with a reason, is the difference between a chain that visibly fails and a chart that looks complete and is not.

Can I redistribute this data or build a product on it?

It depends per source. Each contributing register applies its own licence, some restrict bulk redistribution or commercial use, and data obtained under a legitimate-interest regime usually carries conditions on onward disclosure. The licence field on each statement exists because there is no single answer.

How do I handle a trust in an ownership chain?

Carefully, and usually as a termination. Arrangements are handled inconsistently across regimes, the roles that matter — settlor, protector, beneficiary — are frequently not published, and a chain reaching a trust generally stops there in the public data. Record it as an unresolved node rather than continuing through it.

Standards, formats and interoperability

What this source speaks natively, and what it has to be translated into before a partner can consume it. Work that arrives in a recognised format is easier to defend, easier to hand over and easier to automate against:

The Beneficial Ownership Data Standard is the specification itself: a JSON schema with codelists for interest types, person types, entity types and source types, published openly and versioned, with structural changes between major versions that consuming code must account for.
FATF Recommendations 24 and 25 set the international expectations for beneficial ownership of legal persons and arrangements, and are the policy source that national registers implement with varying fidelity.
The European Union anti-money-laundering framework, as reshaped by the 2022 Court of Justice judgment and the subsequent legislative package, governs access in member states and replaced general public access with a legitimate-interest regime.
The Extractive Industries Transparency Initiative standard requires beneficial ownership disclosure from companies holding extractive licences, and accounts for much of the non-European open publication.
Open Contracting and public procurement disclosure regimes increasingly require beneficial ownership of bidders, which is a growing and under-exploited source of the same data in a different context.
Scheme-qualified organisation identifiers, including national register codes and LEIs, are what allow entity statements to resolve to real companies; identifiers without schemes are not usable for joining.
The platform exports ownership graphs and entity records in CSV, JSON, JSONL, STIX 2.1 and MISP, retaining statement provenance so that a downstream consumer can see who asserted each edge and when.
Nothing in the ownership graph is model-generated: every node and edge originates in a published statement, and Copilot summaries describe the statements rather than inferring relationships between them.

References

Primary documentation and authoritative references for this source. Publishers revise and retire material, so treat the retrieval date as part of the citation and re-check before relying on any of it in a formal product.

Open Ownership — Open Ownership. The organisation: guidance on implementing beneficial ownership registers, research on data quality and policy, and the current status of its data products.
Beneficial Ownership Data Standard — Open Ownership. The specification, schema, codelists and worked examples. The durable asset here, and the right foundation for any ownership data model.
Open Ownership Register — Open Ownership. The aggregated cross-jurisdiction register and search interface. Use for discovery, then verify against the originating national register.
Financial Action Task Force — FATF. The international standards on beneficial ownership transparency that national regimes implement, and the evaluations that assess how well they do.
Companies House — United Kingdom Government. One of the largest open beneficial ownership regimes, and the best-documented case study in both the value of publication and the failure of unverified self-declaration.
OpenCorporates — OpenCorporates. Cross-jurisdiction company register data providing the legal-ownership and corporate-existence layer beneath beneficial ownership declarations.
OpenSanctions — OpenSanctions. Consolidated sanctions and politically exposed person data with entity resolution, the standard screening layer for declared owners.
GLEIF — Global Legal Entity Identifier Foundation. Global entity identifiers and accounting consolidation relationships, useful both as a join key and as a structure to test declarations against.
ICIJ Offshore Leaks Database — International Consortium of Investigative Journalists. Offshore entity and intermediary records covering the structures that registers were created to expose and still largely miss.
OCCRP Aleph — Organized Crime and Corruption Reporting Project. Cross-source search across registers, leaks and documents for testing declared owners against investigative material.
Extractive Industries Transparency Initiative — EITI. The disclosure requirement responsible for much of the beneficial ownership publication outside Europe, particularly in resource-producing states.
Court of Justice of the European Union — CJEU. The court whose November 2022 judgment ended general public access to member state beneficial ownership registers and reshaped access across Europe.

Link integrity: every reference above was verified with a live request when this page was generated. Where a publisher had moved or withdrawn a document, the link was repointed at a preserved copy in the Internet Archive and marked as archived. Anything with no reachable copy anywhere had its link removed rather than left to rot — the source is still credited, it simply cannot be linked.

Put it into practice

The Quantus Intel threat intelligence platform operationalises this source: it ingests beneficial ownership as dated, sourced statements rather than flattened rows, keeps unknown and anonymous persons in the graph as explicit terminations, propagates share bands as ranges through indirect chains, resolves entities on scheme-qualified identifiers into the wider corporate picture, and retains dated snapshots so that point-in-time ownership can actually be answered.. Browse the full source catalogue, or follow any tag above into the rest of the library.