In early 2015, an anonymous source leaked 11.5 million confidential documents relating to offshore entities represented by Panamanian corporate services firm, Mossack Fonseca. The International Consortium of Investigative Journalists (ICIJ) has been diligently working their way through this glut of documents for the last year and published their first findings on April 3rd, 2016.
While journalists and law enforcement comb the documents for hints of financial irregularities, we at IP Street can't help be realized that there is an intellectual property angle to this story as well. Assigning patent assets to offshore companies has long been used as an effective tax mitigation strategy and privacy laws in many of the tax haven countries makes it nearly impossible to identify who owns what patents. Patent privateering is a very real thing but analyzing the scope of its use is exceeding difficult. The Panama Papers offer us a unique opportunity to look behind the curtain and see how these shell games really work and who's playing them.
The ICIJ has published the first narrative stories from the leak and the entire archive is said to be 2.6 terabytes so dealing with it is not for the faint of heart. While you wait for more disclosures from the Panama Papers, you can analyze prior ICIJ disclosures such as The Offshore Leaks Database or download The Swiss Leaks by downloading the data from the good people at Archive.org.
Identifying what patents are owned by the entities identified in the Panama Papers would be like walking through deep mud if you used traditional patent search methods. Thankfully, IP Street makes starting this analysis exceeding easy by making direct calls to its Basic Data Feed API endpoint.
Below, we have walked through an example of how one might get started using IP Street to analyze patent ownership networks in the Panama Papers. If you want to follow along on your own, sign up for a free trial. You can also find the full code example in our GitHub Repo.
Step 1: Prepare Your Data
The ICIJ does a great job of making their data freely available and organizing it into a graph database structure. Entities like people and companies are nodes with the relationships between them being called edges.
To start, download the "nodes.csv" file from Archive.org. This file is very large and contains a great deal of data that is not required for our analysis. Clean up the data by removing records that aren't entities. Since we really only need the entity names for our patent ownership query, can exclude all other types of nodes and all other columns.
If you are not interested in dealing with this raw data step, you can download a .csv file we have prepared for you here.
Step 2: Build Your Query
Next, we need to define a POST request to the
/data/patent endpoint which queries for patents owned by each named entity. Python's requests library makes this a fairly straightforward task. You just need to supply the endpoint, an
owner: parameter within the
q: filter object, and your IP Street API key.
Step 3: Send Your Query
The final step is to send a POST request for each name on the list. The response is JSON encoded to make it easy to feed directly into the next module of your data science assembly line.
For simplicity sake, we have made one call per company in this example. Depending on your infrastructure constraints, you may find it more efficient to combine many company names into the same call using an OR operator like so:
"'company abc' OR 'company xyz' OR 'company ijk'".
Below is a truncated example output of the search for "MegaSpirea NV". You can see that we provide you with the current owners, the owner's ultimate parent (if known), the law firm which helped apply for the patent, and the inventor. All of which are great leads for the enterprising investigative journalist.
Many of the entities don't own any patents at all but you will be surprised by the number of strange patent holding companies you will come across in the Panama Papers.
You can find this full program in a ready to run script in our GitHub repo.
The number of companies and patents involved in the Panama Papers is vast and understanding every facet of the shadowy world of offshore companies may be beyond the reach of any individual. We hope that IP Street's suite of APIs can empower the world community to get a better grasp on who owns what in the world of intellectual property. We look forward to showcasing what you discover.
If you are interested in learning how IP Street can help your team better understand patent assets please sign up for a free trial.