Lessons from building a prototype single-search tool for beneficial ownership registers

Photo by Lisa Zoe on Unsplash.

Beneficial ownership (BO) data is useful only once it is in the hands of those who need it. However, our recent research with users of BO data found that lack of access to information is still one of the biggest barriers to its effective use.

Our work developing the Open Ownership Register showed that there is demand for accessing BO information from multiple countries. Yet, only a few countries publish the bulk BO data required to power such tools.

Many countries are making it possible for users to search and explore BO data at a national level. But if you’re conducting an investigation into a company or person, you might not know where to look. Therefore, knowing whether a register holds specific information and which register holds that specific information could save time and resources in investigations.

A little while ago, we had a simple idea – what if there was a place where you could search multiple countries’ BO registers at once? Would that be possible? Would that be useful? We spent some time over the last few months creating and testing a tool to start answering these questions.

We wanted to learn:

• To what extent would users of international BO information find a simple signposting tool useful?
• What advantages (or disadvantages) might a tool like this have over the information sources they are using already?
• Are existing BO registers set up to enable access in this way?
• What can this tell us about effective implementation of BO reforms?

This blog post describes the tool we built in light of these questions, and what we learned in the process.

Initial thinking

To start thinking about what this might look like, we developed a product vision board to consider our target audience and their needs. We then discussed what kind of product might be able to serve these needs.

This exercise helped us make a few key decisions:

• The tool would be aimed at anyone investigating specific companies or individuals from a range of different countries, including law enforcement agencies, financial intelligence units, bank staff, and journalists.
• We would start off just by signposting users to where they could find more information – our hypothesis was that the quantity of sources was more important for these users than providing BO information directly.
• Search by company and by person would be important to investigators, so we would attempt to do both.
• We would only access official registers rather than unofficial data (e.g. Panama Papers leaks). Our priority as an organisation is to provide advice to governments, so we wanted to prioritise learning lessons about official government sources.

Identifying sources

We came up with a longlist of 33 potential sources of information – BO registers – for the tool based on the Open Ownership map. These potential sources either provided some level of public search functionality or were present in commercial tools like Moody’s Orbis database.

We then needed to figure out which of these would be accessible. Our criteria were that the source:

• provided an application programming interface (API) to search companies and/or persons (technically, we may have been able to search some registers either by using browser imitation, or downloading and searching bulk information, but this would have required significantly more developer time and/or storage costs); [1]
• did not require user registration to search (although registering to access an API would have been fine).

This left us with a list of nine accessible BO registers we could search by company name or entity ID number, and four we could search by a person’s name.

Note: this list is not exhaustive – there may be other registers that we could have connected to, but these were the ones we identified in the time available.

Source	Company search	Person search	Comment
Bulgaria	✔️		Searches of the Bulgarian company register need to be in the Cyrillic alphabet, so the tool converts the Latin alphabet to Cyrillic
Czechia	✔️
Denmark	✔️	✔️	Technically, it is possible to search by beneficial owner, but we were often blocked by their server
Estonia	✔️	✔️
GLEIF	✔️
Latvia	✔️
Nigeria	✔️	✔️
Poland	✔️		Requires exact date of birth in order to search by person name
Slovakia	✔️	✔️
United Kingdom	✔️		Provides BO information in bulk data download but it is not possible to search by beneficial owner via the API

The tool

With these sources, we designed a simple tool. It has just a search bar and radio button which allows you to choose whether to search for a company name or a person’s name.

After inputting the name of a company and selecting "Company search", the tool will return the list of sources and number of matches found in the source.

For example, searching the company name “Aurubis” returns:

It then links to each separate BO register, where the user can choose to go and recreate the search to try to access the actual BO information if they so choose.

Similarly, the user can instead choose to search for a person’s name – doing so returns a similar but shorter table of registers, just those that allow a search by a person’s name:

Users are then able to go and recreate the search on the relevant registers if they get a match.

If you want to try it yourself, the tool is available here – you will need to follow the install instructions in the README file.

Testing with users

After creating the tool, we wanted to get it in front of some real-life users of BO data in order to see if it would be something they found useful.

We identified a number of types of users that might find this kind of information helpful, including:

• financial intelligence units;
• banks;
• journalists and civil society.

We conducted a total of six interviews with relevant users. We asked each interviewee:

• How do you currently access international BO data, and what do you use it for?

Following a demonstration of the tool, we subsequently asked:

• Do you think you would find the tool useful in its current form?
• What additional features might it need in order to be useful for you?

What we learned

We spent 11 days of developer time, plus another 20 days on design and testing. Despite the many limitations, we learned some interesting things:

• People found it useful – in principle

All users who tested the tool expressed that there was value in a simple signposting tool. As one bank employee said:

“Even if I can see that Italy is listed and the register is closed and I still have to file a request to Italy, then that would still be useful… If I go to open sources, I don’t know what I’m looking for, I just search and search and search.”

However, nobody followed up to ask for access following the demonstration, which would have been a much stronger signal of value.

• Connecting the tool to more sources would be more useful than retrieving the information about beneficial owners and companies

Four of the six people interviewed suggested that adding more sources within the tool would make it more useful, over and above providing information about the beneficial owners of companies or companies owned by the people searched.

However, some users did ask to see the data within the tool itself, particularly if it could either visualise BO networks or enable the search of other sources for information retrieved by the initial search. For example, if an initial search for a company name would retrieve the name of a beneficial owner, the user would be able to directly conduct a second search for that beneficial owner.

• A simple technical solution

It took 11 total developer days to set up 9 sources and a simple user interface. Of course, there are bugs and issues that could be ironed out with more time, but this is not a difficult technical problem. When the sources are accessible – even if poorly documented – incorporating information from a new data source is measured in days, not weeks or months.

• Access to sources is very limited – even amongst registers that provide some level of public access

Not many sources were accessible via this method, even amongst those that were ostensibly public. For example, while as a non-governmental organisation we were able to gain legitimate interest access to BO information via an API from the French national register, the terms of use were not clear enough as to whether we were allowed to use the information in the manner necessary to make this tool work without our users also having to apply for their own legitimate interest access – an undue burden for simple testing. Other notable sources we were unable to include (for different reasons) were Armenia and Indonesia.

For a signposting tool, the ability to search by person is arguably more useful than by company. A company is only registered in one jurisdiction, and if you are investigating a company, knowing where it is registered is likely information you already have. However, a person might own a company anywhere in the world, and being signposted to where this is will be considerably more useful.

Only four of the nine registers allowed users to search for the companies owned by a specific person, meaning this user need is often not met. This is perhaps not surprising, given that searching by the name of a person constitutes a higher degree of infringement on privacy.

• (Some) users are most interested in information from jurisdictions in their region

Which countries users are interested in depends on two main things – where the users are based, and what they do. The bank employees interviewed were mostly interested in nearby countries, where their clients were most likely to have a presence. Those interested in investigating sanctions were more interested in where individuals had exposure, meaning that having access to more sources was preferred. Those interested in financial investigation were more interested in following chains into secrecy jurisdictions.

How does the tool compare to existing ways users access beneficial ownership data?

Users listed several advantages that a tool like this might have over other ways of accessing BO data from multiple countries, including:

• Knowing where to look: While information might be either publicly available on registers or via an internet search, or accessible via professional networks (e.g. Egmont Group for FIUs), it can be difficult to know where to look or who to ask.
• Cost: Commercial offerings come at a significant cost, which is prohibitive to many. This is true not only for journalists and civil society; a bank also mentioned restrictions in its ability to use commercial tools to conduct due diligence.
• Up-to-date information: Both commercial and transparency offerings (e.g. the Organized Crime and Corruption Reporting Project’s Aleph database) can be months (or even years) out of date depending on when the data was last updated, whereas this tool searches the most up-to-date information directly from official sources.
• Navigating unfamiliar registers: Some users found value in the tool just in being able to identify where official registers were, and being able to translate the information within them when necessary.
• Provenance: Users also liked that this tool was clear about where information came from, as commercial offerings are sometimes not explicit about whether information is reported or inferred from other data.

Users were divided on the trustworthiness of the information – some found the fact that it was official information comforting, while others were worried that they didn’t know how well a country verified the BO information it collected, and thus how much it could be trusted. This latter point was particularly true for one user who was responsible for identifying sanctions evaders:

“Sanctions violation is considered a strict liability issue in the US and the EU. If you use data, even from an official register, and it is wrong, you would be held accountable for trusting that data.”

Whether or not users in financial services are able to rely on data from a central register for customer due diligence and know your customer checks – and who is liable when the information is wrong and money laundering does occur – is a persistent issue. This is likely compounded when a user would need to factor in an assessment of the quality of the information in registers from jurisdictions they are less familiar with.

Where do we go from here?

This is not a tool that we want to build out further. We just wanted to see what was possible, and what we could learn. To that end, we think this experiment could influence our future work in a number of ways.

Designing access regimes that balance privacy and usability

This experiment gave us a lot of insight into the various ways BO registers allow users to access data via APIs, with some being much more user friendly and useful than others. We should produce advice and guidance for considerations when designing and building APIs for accessing BO information.

This could include a recommendation for governments to build APIs that allow users to search to see whether or not a register holds data on a company or person of interest, without receiving other information about that company or person. While for individuals this may still constitute personally identifiable information, it will likely have a lower infringement on privacy than the information itself. Analysts from FIUs explained that they value publicly accessible registers as they are able to see whether a country holds information on an individual before requesting additional information through formal channels. This could be a data-minimised, less intrusive approach to meeting that need, but would come with a number of trade-offs.

Open Ownership’s research on balancing access and privacy, and with users of BO information, both suggest designing access provisions that go beyond the false dichotomy of “public” and “closed” registers can ensure the infringement on privacy is more proportional, and that user needs are better met. For example, a register could allow public access to information about whether or not it holds certain information, but require a user to demonstrate a legitimate interest to access the information itself. This can also help reduce unnecessary applications and the resources required to submit, review, and process those applications. Legitimate-interest access could also be a sufficient safeguard to enable more registers to allow users to search by person name.

Support for data-sharing initiatives

Many countries are exploring solutions to the challenges associated with sharing information across borders at the regional level, and various regional and multilateral bodies are thinking about how to increase data sharing and use between their members. Although there are a number of legal, political, and technical challenges to these initiatives, we are interested in exploring whether a tool like this might help in overcoming some of those. We will be testing the appetite for this sort of product in our discussions with various different actors in the space.

If you are interested in learning more about our experiment or our future plans, please contact Miranda Evans, Open Ownership’s Data Support Manager ([email protected]).

Footnotes

[1] An API is a set of rules which determines how one software program is able to access the data or functionality provided by another software programme.