Oh wie schön ist (offshore) Panama! (Or what do a fictional bear and tiger have in common with soccer players playing in a defensive position?)

by Luciano Del Corro on in Panama Papers

Panama BeachesA tale of two guys, from different ethnical backgrounds, who live together in the nature and enjoy the small things of life. Adorable characters that came to live in the land of their dreams, right at home. A story rich in metaphors, ideal for progressive parents trying to pass their kids the values that they uphold. However, abusing symbolic language might introduce excessive ambiguity at times. It seems that some of those kids, probably from the practical kind, got a very literal view of the story. (By the way, in case you are interested in the story, have a look here.)

Psychologists use to say that later as adults, we may try to recreate our childhood with elements that bring us back to it. Some may love bears and tigers, some may love the smell of bananas, and some may still regard Panama as their land of dreams. It seems that, with a sense of nostalgia, some kids realized their own version of the Janosch story. Panama is to them not the land of the smell of bananas, but the land of hats, cigars, and, why not, offshore companies.

We have seen in the last couple of weeks how people from all around the globe acquired beloved offshore companies in Panama. You have probably heard enough of local media singling out people, and in turn, these people explaining how convenient is to found a charity in Panama. You heard one name after the other (that you most likely already forgot) but it seems difficult with this information overflow to reach useful conclusions about those findings. The focus so far has been on politicians, and by now you might have gotten the impression that most of the users of offshore companies are indeed politicians. If you examine the Wikipedia category People named in the Panama Papers you will find 5 businessmen, 6 athletes and 35 politicians to date. However, our findings tell a different story: Using our Ambiverse Natural Language Understanding API, some aggregation, and a bit of salt and pepper, we will help you to understand the kind of people that is more inclined to establish offshore companies.

We ran our Natural Language Understanding API on the Panama papers database released by the ICIJ. We linked the names of the officers of the papers to the most likely entity in our knowledge graph. Our knowledge graph contains about 4 million entities like persons, companies, or products which are grouped in more than 600k categories (e.g. singers, politicians, etc). The out-of-the-box knowledge graph contains the most prominent entities in the world (plus the Ambiverse founders. Note: Sadly we also do not run a charity in Panama). Find out more about our knowledge graph here.

Our findings in the Panama Papers

First, what does our knowledge graph know about Panama? Well it knows that it is a country in Central America, that its people speak Spanish, that it is a Republic, a liberal democracy, a member of the United Nations and that it was established in 1903. It also knows that its capital is Panama city but it does not know anything about the smell of bananas (our sources in Panama also confirm that this was probably a poetic license by Janosch).

What about the officers of the offshore companies? Well, you probably have heard a lot about politicians involved in the Offshoregate, however, our finding will surprise you: The most common persons in the Panama papers are not politicians but athletes. We found that more than 20% of the officers in the Panama papers practice some kind of sport professionally and from them, almost half are soccer players. Followed by tennis and basketball players (around 10% each) and hockey and volleyball players (around 5%). In much smaller proportions we also found swimmers, cyclist, cricketers, and gymnasts, among others. Interestingly, if we talk about playing positions, the defenders are the most prominent with more than 10% of all athletes. So if you are thinking of a divorce and your husband is a soccer player playing in defense, you may want to dig a bit deeper.

Now let’s have a look at the second big category (no! they are no politicians. In case you are a bit impatient they will come soon). The second place is occupied by artists with more than 20% (very close to the athletes). It seems that the most popular people in this world (athletes and artists) are more inclined to run offshore companies. Almost the entire artist category is made up by actors, with more actors than actresses (we recently found out that male actors make much more than female actors so it kind of makes sense). Another important subcategory in the artistic domain is musicians (with a prominence of singers and composers). We also found a wide variety of subcategories including presenters, film-makers, dancers and models. It seems that almost every artistic discipline is represented in the Panama papers (maybe it is not a matter of the discipline but the amount of money they make? who knows?).

If you thought that now was the time of politicians you are correct. Around 10% of the persons we found are politicians, including some heads of state, ministers, mayors, governors or legislators; people holding office now or in the past. Also it seems that political ideology has little power to explain if a person may appear in the Panama Papers or not: We found as many conservatives as socialists. There are also judges and executives in companies and in non-governmental organizations like sport bodies. People directly related to money are also very prominent. From the people we found, more than 2% are business persons, 0.5% are economists and also 0.5% are financiers. Interestingly, more than 2% of the people are billionaires.

In total we found more than 19k categories to which the Panama Papers officers belong. Bear in mind that one person can be assigned to more than one category. Among these categories there are some interesting curiosities. We found people that have been convicted as well as benefactors and philanthropist. We found a wide range of professions including medical doctors, lawyers, mathematicians, historians, designers, and many others.

But here comes probably the most important part of the Panama Papers and something that we can very confidently ascertain. As we mentioned before, most of the people in our knowledge graph are the so-called prominent people (i.e. people with a Wikipedia page). But what happens if a person is not in our knowledge graph or even what happens if our technology makes a mistake given the lack of context? (Note that the ICIJ did not release the documents but a database with limited data. Our technology would work much better in plain text making use of more information.) Well, it is highly possible that in case an entity is not in the database, we may have missed it or linked it to the wrong person in our knowledge graph. In fact, we were able to only link around 7.5k persons to our knowledge graph from a total of more than 215k. This means that most of the people in the Panama papers are non-prominent, just plain regular people unlikely to be interesting to the media.

So what can be done in these cases? How can we analyze documents full of entities that are not part of our knowledge graph? Well, if you do already have a database of entities, it can be integrated into our knowledge graph. However, we also developed a tool which allows you to add entities into our knowledge graph so that they can be fully integrated into our system. To learn more about this feature and our technology in general, and how you can do the kind of analysis that we did on the Panama Papers, you can read more details here.

Featured image CC-BY-SA by José Manuel CastrellonFundación Almanaque Azul

Luciano Del Corro

Luciano Del Corro

VP Innovation & Co-Founder at Ambiverse
Luciano has the ability to solve language understanding problems in a principled but tractable manner, making him the key person for bringing research to applications. He completed a PhD in natural language understanding at the Max Planck Institute for Informatics.
Luciano Del Corro