Legal form: a modelling journey, part I

Abstract vision of a legal form for a company, in a landscape of country and language.

The AnaCredit regulation is an interesting regulation. Having closely worked with the Dutch modeller who drafted the logical model for it at the DNB, it’s one of my favorite models to demonstrate. It compresses literally hundreds of pages of rules and regulations in a single model. To the dismay of most banks when they first saw it, but eventually to the benefit of all of its users.

However, one issue has always been a bit of a pain. And that is the concept of legal form. You are required to report it for the counterparties involved in loans. But at the time the regulation started, some banks had not used foreign legal forms but had mapped them onto their own home countries versions. Others had incomplete registrations. And most of them had issues with languages: in Belgium you can register a company in three different locations, with three different names for the legal form, but they are all the same. Even worse, you can potentially have the situation that you have the same code in different national languages but they mean something different. This means that using the legal form code without any other descriptive attributes is a problem.

Legal form is a bit of a weird duck in a way. Everyone knows it. Everyone uses it. But when asked to describe what it is, you get wildly varying answers. Obviously, the ECB doesn’t know what it is either, or they would not have made the list they use now (you can check out the ECB List of legal forms yourself).


On the internet it’s not that easy to find a definition of legal form. The ones I can find come close to the joke about an elephant, described by a number of blind people: they describe the visible attributes but not the core, the “ding an sich”. See for instance this definition: while it starts good with “the legal form (also known as legal status) is defined according to national legislation” it then goes on and on about its use.

So what is it then? The best I can come up with, is that the legal form under which a company operates, is defined in the national legislation. The name of the legal form and the abbreviation are sometimes given in that legislation, but not always. In the end, any legal form code or name is a shorthand for a referral to a particular paragraph in the national law of any given country or group of countries such as the EU.

For instance, in Dutch law we have a limited liability company. This is called a “besloten vennootschap” and abbreviated as “BV”. This is defined in het Burgerlijk Wetboek (civil law book), book nr. 2, article 175. This says that any BV is a legal person, with named shares, where the shareholders are not liable for any losses over the amount of capital they have put into the company. And that is the basic definition of this particular legal form (although there are more articles describing this legal form in more detail).

Do not make the mistake of assuming that this particular legal form is the same as the limited liability company in other countries: liabilities can be limited in very different ways and can be undone on very different grounds. That is why you really need to include the country when you define legal form.

A legal form is also unique not just by country but by official language in that country. In Belgium, as you can see in the ECB List of legal forms, they have a legal form that has three names (and corresponding acronyms):

  • Unité TVA – UTVA (French)
  • BTW-eenheid – BTWE (Dutch)
  • Mehrwertsteuereinheit – MWSE (German)

This particular legal form has a single surrogate key, meaning that to the ECB, it is just one single legal form. So far, so good though.


Now, so far things are relatively simple. But they’re getting a bit more difficult when you realize that some EU countries have no standardized set of legal forms. Try finding the ones for Portugal, for instance. On the official government website you can find nine legal forms. But the Anacredit list has twenty of them, and one even has no acronym. The thing the website does well, though, is that it refers you to the exact paragraph in the law that describes and regulates the legal form. Portugal is an example, but not an exception.

Things get funnier when you add the ISO standard into the mix. Yes, there is an official ISO standard (ISO 20275:2017 – Financial Services – Entity Legal Form) for legal form, because the mess has not escaped the attention of the standardization committee. The data model is described in the standard, but there is also a registrar for worldwide legal forms, the Global Legal Entity Identifier Foundation (GLEIF). I’ve described this standard in an older post.

Some of the differences between the lists are:

  • The GLEIF list contains legal forms on a country subdivision code level (ISO 3166-2), used in for instance the USA and Canada, where the ECB list does not. This becomes interesting in the case of Madeira, which is subdivided in the GLEIF list but part of Portugal in the ECB list. Where the GLEIF lists two legal forms for Madeira, none are mentioned in the ECB list;
  • The GLEIF list contains 31 legal forms for Portugal, the ECB list only has 20, and the Portuguese government lists 9. Have fun trying to find out which one is valid for your case. Portugal is just an example here, the deviations are similar for many countries;
  • The local name is nice, but do you also want it in the local script? This becomes a relevant question for cyrillic names in Europe, and for other non-Western scripts such as Chinese and Arabic that may not even be read from left to right. The GLEIF lists the local name in the local script and provides a transcription to the Western script as well. But the ECB list does not. If you are lucky you can get them from the GLEIF list, and otherwise, well, there’s always ChatGPT;
  • The GLEIF list sometimes lists an abbreviation, sometimes not. The abbreviations can be in the local script, in that case there are usually, but not always, transliterations into the Western script.
  • The GLEIF list sometimes lists multiple versions of abbreviations for the same legal form for the same language for the same country. The ECB list only does that for different languages in the same country. Good luck matching them up.

So how do we get out of this mess? Well, we don’t 🙂 The mess will remain. But you need to understand for what purpose you need the legal form of a legal entity. The main purpose is to determine the amount of risk and exposure you are taking on if the legal entity in question is your counterparty. If you do business globally, this matter becomes more urgent. A secondary purpose is to report this to the relevant authorities, notably the ECB with the AnaCredit reporting requirement, who does the same but on a higher (aggregate) level, except that exact legal forms are limited to European counterparties, and for global ones you are allowed to approximate them.

We can create a data model that will enable us to fulfill most of these requirements. Given the data issues I doubt it is possible to get full coverage for all countries and all legal forms, but we can certainly do much better than just add the nearest acronym to a legal entity, and hope for the best. How to do this data model is a subject for the next post, however.

Leave a Reply

Your email address will not be published. Required fields are marked *