openownership / data-standard

The Beneficial Ownership Data Standard (BODS) is an open standard providing a specification for modelling and publishing information on the beneficial ownership and control of corporate vehicles
http://standard.openownership.org
Other
63 stars 13 forks source link

Feature: Representing missing information #389

Open kd-ods opened 2 years ago

kd-ods commented 2 years ago

[This ticket helps track progress towards developing a particular feature in BODS where changes or revisions to the standard may be required. It should be placed on the BODS Feature Tracker, under the relevant status column.

_See Feature development in BODS in the Handbook._

The title of this GitHub ticket should be 'Feature: XXXXX' where XXXXX is the feature name below. The information in this first post on the thread should be updated as necessary so that it holds up-to-date information. Comments on this ticket can be used to help track high-level work towards this feature or to refine this set of information.]

Feature name: Representing missing information

Feature background

What user needs are met by introducing or developing this feature in BODS? [Summarise these needs. Link to user stories, reports, blogs and other evidence where possible. Or add user stories here directly.]

BODS represents more than just company structures. It represents declarations (statements) about company structures. Therefore the schema needs to facilitate the representation of various situations where there are known gaps in the information declared and published.

Here are 10 examples of such missing information:

Example 1 The declaring company is exempt from disclosing beneficial owners on any of a given set of grounds. E.g:

Example 2 Declaring company declares that a significant shareholder (direct or indirect) is a public, listed company (details undisclosed).

Example 3 Declaring company declares that nobody meets the definition of a reportable beneficial owner.

Example 4 Declaring company declares that information about potential beneficial owners has not been provided by a single upchain entity (info not retrievable). Other beneficial owners have been identified.

Example 5 Declaring company declares that information about potential beneficial owners has not been provided by a single upchain entity (because potential beneficial owner(s) unco-operative). Other beneficial owners have been identified.

Example 6 Declaring company declares that information about potential beneficial owners cannot be retrieved, details of all legal owners are supplied instead.

Example 7 A beneficial owner is granted anonymity on stated grounds.

Example 8 A company is required to make a declaration but has not done so.

Example 9 The chain of intermediaries between a declaring company and its beneficial owner is declared, but the particular interests connecting the intermediaries are unknown (and not required to be disclosed).

Example 10 A declaring person is exempt from declaring the companies they beneficially own on the grounds that they don’t have greater than an x% stake or control over a company with a turnover > $n.

Typology Looking at these examples, a typology suggests itself. When representing declarations of beneficial ownership it needs to be clear to users of the data:

Additional requirement [added 08-02-2022]

It will often be the case that some information about owners and controlling interests is required to be disclosed in a declaration system, even when no people meet the definition of 'beneficial owner'. So however missing information within a declaration is represented in BODS, disclosed information must be able to be represented too. See the point about Liberia below.

What impact would not meeting these needs have?

If the data standard does not offer a way of representing why information is missing, then publishers may either: not represent it at all, or represent it in non-standardised ways. In either case, this will make it hard for people to correctly interpret data (especially when aggregated from different sources).

How important is it to meet the above needs?

These needs must be met in order to cover the full scope of disclosure possibilities.

How urgent is it to meet the above needs?

The standard should not reach version 1 without incorporating this feature.

Are there any obvious problems, dependencies or challenges that any proposal to develop this feature would need to address?

We already have (as of BODS 0.2) various codelist options and modeling options that go some way to addressing the above needs. However they are not complete, not well documented, and not consistent. There are - therefore - numerous issues that point to these deficiencies:

We would expect these to be resolved with any proposal to address this feature in the standard.

Feature work tracking

[Link to proposals, bugs and issues in the repository to help track work on this feature]

StephenAbbott commented 2 years ago

@kd-ods The situation in Liberia (quoted below, found on p18-19 of Open Ownership's Beneficial ownership in law: Definitions and thresholds briefing) looks like a new example which is slightly different to examples 3 and 6 above:

Finally, governments should apply reporting requirements for firms where no person falls within the threshold or definition of BO. It is possible for no individual to meet the criteria for disclosure outlined within a legal definition of BO, even if the definition is robust and the thresholds are low. In such cases, it is recommended that countries require the disclosure of the name of a natural person in a senior role with managerial responsibility for the company in question. Many countries – including Argentina, the Republic of Korea, and EU member states under AMLD5 – require the submission of names of senior managing officials of the relevant entity instead (e.g. the directors, CEO or board members). The EITI, in its work on BO disclosures for mining, oil, and gas firms, has held up the example of Liberia as a case study in good practice. In that country, where no natural persons meet the definition of BO, the reporting entity must declare the identity of those individuals with the five largest stakes in the firm.

Whilst it is important to note that these people are not necessarily beneficial owners, for law enforcement purposes it is preferable to have the name of somebody who has real responsibility for the company rather than having no name at all (or the names of formation agents). Where this is the case, it should be recorded in the published data that these individuals have been disclosed because there are no others that qualify as beneficial owners under the country’s legal definition.

Should we add example 11 with text like "Declaring company declares that information about potential beneficial owners cannot be retrieved, details of largest stakeholders are supplied instead"?

kd-ods commented 2 years ago

Thanks, @StephenAbbott - the Liberia regulations are a good test case. It will often be the case that some information about owners and controlling interests is required to be disclosed in a declaration system, even when no people meet the definition of 'beneficial owner'. So I think this aspect of the Liberian system does fall under example 3 above.

But it does highlight an additional requirement of any proposal to address this feature: that however missing information within a declaration is represented in BODS, disclosed information must be able to be represented too. I'll add that to the body of this feature ticket above.

cosmin-marginean commented 2 years ago

I'm worried about the subtleties in the many reasons why information might be missing and trying to unify this in a coherent model.

My main concerns are around situations like Example 3

Declaring company declares that nobody meets the definition of a reportable beneficial owner.

From a consumer perspective, it has to be clear that the company meets certain criteria stating that no BOs should be listed. This is logically (and legally) simply an absence of BOs.

This is in friction with notions from Examples 4/5 where the information does exist, but it's not available.

In other words, when representing this data, it has to be clear when there is nothing there (and the reason why). It also has to be clear when some entity has an interest here but it can't be identified.

The modelling for the former has to refer to the entity company that's (legitimately or not) missing this data, which seems to require a new type of statement/metadata (that isn't ownerhsip-or-control).

The latter, however needs to allow a consumer to reason about a number of "?" entities that exist because there is ownership-or-control. Worst, this relationship can have a cardinality greater than one. Which then raises the issue of representing these anonymous ownership-or-control statements with some IDs that are unique (at least in context).

And this is where the confusion between anonymous and unknown entities is relevant. The fact that the entity has been identified but we have no data for it is, from a consumer perspective, logically equivalent to "entity hasn't been identified" because we simply don't know. Although we could agree that the reason for why the info about the entity is missing is different.

In short, we feel that there is a need to split the domain further in two overarching concepts: a) missing data and b) reasons why data is missing. Because while a) can/should be represented with statements (types), b) would be best handled as metadata. What would seem impractical is an ever-expanding set of subclasses for EntityType to cover situations that essentially mean "there is something, but we don't know what it is". Because this would limit adding future reasons that mean "we don't know" without a new schema.

Apologies for the long comment and if it misses any recent schema developments. I'm wearing mostly the consumer hat on this.

kd-ods commented 2 years ago

@cosmin-marginean - thanks for those reflections.

In short, we feel that there is a need to split the domain further in two overarching concepts: a) missing data and b) reasons why data is missing. Because while a) can/should be represented with statements (types), b) would be best handled as metadata.

I agree. Do you think the current pattern we use in BODS adequately splits (a) from (b)? In particular, we have used the unspecificedEntityDetails property (and its equivalent in Person statements) to provide (b), the reasons why data is missing.

(My current thinking is that we will need to break out the current UnspecifiedReason codelist into separate codelists (see #240) and do a fair bit of work to tighten up the codes. Additionally, we'll need to extend this pattern to cover unknown types of interest too. I've done some work on a proposal along these lines; I hope to be able to share it in a couple of months.)

cosmin-marginean commented 2 years ago

I agree. Do you think the current pattern we use in BODS adequately splits (a) from (b)? In particular, we have used the unspecificedEntityDetails property (and its equivalent in Person statements) to provide (b), the reasons why data is missing.

I believe it does specifically for when we know that there is an interest we just don't know whose interest it is. The more important subtlety is when we have "no-beneficial-owners" (Example 3). What we don't want there is an anonymous/unknown entity because that means there would be some BO, when clearly there isn't. So this would ideally be handled at the level of OwnershipOrControlStatement.

This is an example why I think this work probably needs to be split up into smaller chunks so we can reason about them independently. At a high level these being: "we don't know the entity" vs "there isn't an entity" vs "there is an entity but we don't know the interest" vs "there is an entity and an interest but we don't know what type of interest", etc.

It could be over-ambitious to try to handle this in a single exercise. Even better, once these use cases are identified and materialised it would probably be easier to materialise some additional abstractions. Just a thought.

VladimirAlexiev commented 2 years ago

Is it relevant to describe here the GLEI level 2 "Reporting Exceptions" model?

kd-ods commented 2 years ago

Thanks @VladimirAlexiev . Yes, how GLEIF level 2 handles exceptions to reporting should certainly inform our thinking in this area.

StephenAbbott commented 2 years ago

Noting here that the European Union's draft sixth anti-money laundering directive contains the following:

Where no person is identified as beneficial owner pursuant to Article 45(2) and (3) of Regulation [please insert reference – proposal for Anti-Money Laundering Regulation - COM/2021/420 final], the central register shall include:

(a)a statement accompanied by a justification, that there is no beneficial owner or that the beneficial owner(s) could not be identified and verified;

(b)the details of the natural person(s) who hold the position of senior managing official(s) in the corporate or legal entity equivalent to the information required under Article 44(1), point (a), of Regulation [please insert reference – proposal for Anti-Money Laundering Regulation - COM/2021/420 final].

Clause A reinforces the need to capture information on why BO data cannot be collected or is missing.

StephenAbbott commented 2 years ago

Open Ownership has recently emphasised the importance of capturing missing information in our new technical guidance 'Building an auditable record of beneficial ownership' - see https://www.openownership.org/en/publications/building-an-auditable-record-of-beneficial-ownership/feature-four-visible-information-gaps/

StephenAbbott commented 1 year ago

As part of the The Money Laundering and Terrorist Financing (Amendment) (No. 2) Regulations 2022, the UK will be introducing new "material discrepancy" reporting rules for those obliged to check the beneficial ownership details submitted by overseas entities to the UK's new Register of Overseas Entities. Some of these regulations come into force in April 2023.

The creation of the Register of Overseas Entities was originally legislated for under the Economic Crime (Transparency and Enforcement) Act 2022.

Regulation 9 amends regulation 30A of the MLRs to extend the obligation on relevant persons to report discrepancies to the registrar of companies between information they hold on the beneficial ownership of a customer and information on the register. The obligation to report is an ongoing one in line with the duty in the MLRs to undertake customer due diligence and ongoing monitoring. The provision also limits the duty to report to ‘material’ discrepancies, not just any discrepancy. Regulations 16 and 17 are consequential changes to primary legislation to update the law to align with new regulation 30A and give the registrar clear powers to deal with a material discrepancy in the register.

These regulations tighten rules for how people carrying out due diligence tests/verification tests on overseas entities have to report beneficial ownership discrepancies to Companies House. What is meant by a 'material discrepancy' is set out below:

SCHEDULE 3AZA Material Discrepancies A material discrepancy in this Schedule may arise, as the case may be, in relation to information about a beneficial owner within the meaning of regulation 3 of these Regulations (including about a person of significant control within the meaning of Part 21A of the Companies Act 2006(36)) and in relation to information about a registrable beneficial owner within the meaning of Part 3 of Schedule 1 to the Economic Crime (Transparency and Enforcement) Act 2022.

  1. A material discrepancy in this Schedule is one which satisfies the condition in paragraph 2, including one which is in a form listed in paragraph 3.

  2. The condition in this paragraph is that the discrepancy, by its nature, and having regard to all the circumstances, may reasonably be considered—

  • (a)to be linked to money laundering or terrorist financing; or
  • (b)to conceal details of the business of the customer.
  1. Discrepancies listed in this paragraph are in the form of—
  • (a)a difference in name;
  • (b)an incorrect entry for nature of control;
  • (c)an incorrect entry for date of birth;
  • (d)an incorrect entry for nationality;
  • (e)an incorrect entry for correspondence address;
  • (f)a missing entry for a person of significant control or a registrable beneficial owner;
  • (g)an incorrect entry for the date the individual became a registrable person.”

This example shows the importance for publishers and declarants in being able to be very clear about laying out the reasons that certain types of beneficial ownership information may be missing from declarations.

kd-ods commented 8 months ago

Just linking to some thinking I did years ago about this issue, in case it's useful in the future.: https://docs.google.com/document/d/1j-b-6IaNuSpDqAFjzkwyCP3n-2fS0ydMiLzsmYp8sdY/edit?pli=1

StephenAbbott commented 3 months ago

https://github.com/openownership/data-standard/issues/389#issuecomment-1170859971

Is it relevant to describe here the GLEI level 2 "Reporting Exceptions" model? @VladimirAlexiev

Open Ownership and Open Data Services have just updated our mapping of GLEIF data to BODS to better capture the lifecycle of LEI data changes.

If you read the details in our blog post, you can see the challenges we've experienced handling GLEIF reporting exceptions.