The need for better identifiers - discuss!

This may be a matter for legislation before it can work, but the registry is badly in need of a unique identifier for officers (as there is for companies) to unify all the filings relating to each human. If you search, for example, for officers named as follows:

You will get at least 91 different identifiers for that name, nearly all at the same address in E18 and all born in July 1982. It makes it very hard to see the extent of one individual’s directorships across the UK. If you make an assumption that someone born in the same year and month is the same person, then you get this record of 112 directorships past and present:

https://webb-site.com/dbpub/positions.asp?p=12865969

But it would be preferable not to make that assumption, particularly for more common names.

DOB as a differentiator
One could be more certain about identities before day-of-birth was removed from the public system, but still it wasn’t an ideal identifier as there may be more than one John Smith/ Thomas Owen/ James Stuart born on a given day and matches in a given month are much more likely.
Address as a differentiator
Addresses are no longer a good differentiator because so many people now choose their corporate service address rather than residential address, and so many companies share the same address at a corporate secretarial firm. Also, addresses can take multiple formats with typos, so comparing them by machine requires an amount of fuzzy matching.
Names as identifiers
It may seem obvious, but people often give a different variation of their name, with or without a middle name, or using nicknames such as Tom rather than Thomas, or changing their name without stating former names on directorship forms, so matching them based on name is harder, even without the fact that there are other people who share the name. People who are looking to escape a history of directorships of insolvent companies may adopt such tactics, or just wait 6 years before UKCH removes them from view.
Passports or national IDs as identifiers
Passport numbers are generally unique and not recycled, although an individual will have more than one in her lifetime, so a disclosure of the previous passport numbers on each form would be needed. Not everyone has a passport though. Many if not most countries have a national ID or social security system with unique identifiers and these are generally assigned for life (except for government-sanctioned change of identity such as witness protection).

Corporate secretarial firms generally have a “Know Your Client” (KYC) obligation to gather such forms of identification when registering directors, so they should be able to use the passport or ID number in a public registry. Indeed many companies registries such as Hong Kong use the passport or ID number as an identifier and the public search system for those registries uses that.

I throw this out for discussion. Does anyone have a better solution to avoid duplicate identities and mistaken identities?

It looks as though you have been underwhelmed by responses to this important topic which may link in with my latest post here POTENTIALLY FICTICIOUS ADDRESS and / or PERSON(S).

I have been here before, having previously spent almost a decade working with PERSONS data for a regional group of police forces. In that arena the provision of deliberately misleading information is endemic and we are seeing some traces of the same activity with Companies House data.

I like your thinking and if you would like to discuss further drop an e-mail to me - frank@statbooks.co.uk

I’d held off posting in case of response from CH. I don’t have any solutions to add here and I’m not implementing the API! I’m sure most / all users would be in favour of such improvements. For us even merely avoiding the “multiple entries for the same individual / entity” would be a great improvement. This could take the form of CH issuing a unique identification without needing to “expose” further data e.g. even passport / ID. I don’t see offhand why that would pose any legal / privacy issues.

I think that even as things stands there could be improvements in this area:

  • Better user input design (especially given electronic filing) / filtering and validation. Examples: avoid incorrect categorisation of human PSCs as corporates or corporate offices as humans, don’t allow officers to have birth dates which make them 1 (is this legal) etc.
  • Is it possible - possibly for new data only - to unify some of the data sets? Examples: at the moment officers and PSCs do not correlate, the disqualified officers is a separate data set also (yes I’ve seen the post about this being a decision), the way that changes to PSC data are recorded.
  • Beyond simple identification, data is recorded - e.g. shareholders - but not available directly via the API. So you’ve still got to look through the pictures (filings) to find data which is input in textual form.

I imagine anything further - and possibly even the above - would require legislative change. In an ideal world not only would people / entities be uniquely identified but the incoming data would requirements be much more strict. At the moment the impression is “as long as you file it, put as much or as little as you like - no-one’s going to come bang on your door”.

(I now veer completely outside the forum scope.) This almost certainly comes down to money - for people to check input or develop the systems to do so and - equally important - for all the machinery of enforcement. No point worrying about the rules if there is almost no chance of cost to your business if you ignore them. I suspect that the trend of political focus in the UK (even allowing for the current turbulence) is elsewhere and if anything tending to the opposite direction from “stronger regulation”.