Blog / July 8, 2016

A first look at the UK beneficial ownership data

As we wrote on this blog last week, the UK government has just taken the bold step of publishing the world’s first fully open register of beneficial ownership. This is the first dataset of its kind, and the information promises to tell us who really owns UK companies, whilst making it harder for criminals to hide dirty money. As we explained here, we need this information so that businesses know who they are dealing with, civil and environmental rights can be protected and so that law enforcement and civil society are able to detect tax evasion and money laundering.

This is a true open government data initiative and credit should go to the folks at the Department for Business, Innovation and Skills (BIS) and Companies House for their commitment to open standards and bravely going where no other government has gone before. There are bulk files and a data feed (API) for those wanting to perform analysis and integrate the information into their own applications, and a web interface for everyday users just wanting to explore the information (see below).

Companies House BO data web interface screenshot

The Companies House web interface displaying beneficial ownership data

When it comes to the actual data, we had a quick look at what Companies House published, and what we found was promising, but not without wrinkles.

Preliminary analysis of the first batch

From our analysis of the most recent data (published on the morning of 8th July 2016) there were over 28,076 beneficial owners on record (referred to as “persons of significant control”) and over 22,595 UK registered companies had filed information on at least one beneficial owner.

The release does appear to contain genuinely new information. While we didn’t have the benefit of being able to use shareholder data to compare it to (Companies House is yet to publish this as open data although it is publicly available), we compared the named beneficial owners of specific companies with the listed beneficial owners. Over 300 beneficial owners appeared not to have already been listed as directors of those companies. This is exciting as it shows some companies are filing genuinely new information not contained within their previous filings.

The fact we could make that comparison between directors and beneficial owners is worth celebrating. It demonstrates why going to the trouble of making this available via a data feed (API) rather than just web interface is important. We were able to automate thousands of queries to Companies House servers about directors and compare the results with the listed beneficial owners automatically. This use of computational techniques like this is at the heart of much of the most exciting investigative journalism today, but these techniques need open data in order to be effective.

The wrinkles

Our initial analysis of the first batch of data did reveal some imperfections in the quality of the data. Take for instance the summary of nationalities. As the chart below shows there were multiple versions of “Uk” and “British” pointing to the inadequacy of a free-text field for this kind of entry.

Similar data quality issues were observed in other key fields such as date of birth, where some beneficial owners appeared to be less than a month old (!), and in the in the type of entity field where many people misfiled themselves as “Legal  person” rather than “Individual person”.

Although it’s far too early to tell what the level of compliance will be and how truthfully companies will be reporting their beneficial owners, it’s interesting to note that 14%  of companies appear not have reported a single beneficial owner.

Reasons varied as the table below shows, but the vast majority of companies that did not file a beneficial owner declared that none existed. This is probably because no one individual met any of the criteria for being a beneficial owner e.g. owning 25% of shares, or exercising significant control over the company. Time will tell if this is a route some companies use to avoid disclosing.

The way forward

We hope that Companies House will address some of the problems with the data that have surfaced. Many appear to be the result of people not understanding what they were meant to file, but some perhaps are flags for design choices that should be re-visited.

Companies House and the Department for Business, Innovation and Skills (BIS) have a unique opportunity here. There is serious interest in this new data amongst economists, business, campaigners and journalists. The many eyes of this crowd could be harnessed to improve the quality of this data if routes are built through which individuals can easily point to flaws in the information. Companies House have already taken some steps to directly engage developers in this data release, such as creating a forum for discussion and support, but enabling this kind of data quality crowdsourcing would be a real innovation.

It will take a full year before we have a full set of beneficial ownership filings for all UK companies. In the coming months, and as the number of filings grows, we will be exploring this data in order to understand corporate ownership in the UK and for investigative purposes. Stay tuned for information on events and hack days we will be running.

Our friends at OpenCorporates have also published their take on the early data releases which you can read here.