Fighting corporate corruption with open data

on

This blog explains how DataKind and Global Witness partnered to build the world’s first network graph mapping corporate control in the UK.

In June 2016 the UK government launched a new initiative, a world first: it required companies to register who was actually in control.

Recent investigations have added to the case for greater transparency in corporate ownership. Notably, the Panama papers and Paradise papers. This is a good move towards that goal – which in turn means less corruption and all the ills it brings with it.

Global Witness, an international anti-corruption NGO, immediately sought to demonstrate the usefulness of this new data. They partnered with DataKind to explore the data through a weekend-long hackathon back in November 2016, where we dug into the use and flaws of the data. This was followed by a 6-month project, starting in summer 2017. A team of four DataKind volunteers and one Global Witness staff member come together to analyse what corporate control looks like in the UK and demonstrate the value of open ownership data.

And we did! We built the world’s first network graph mapping corporate control in the UK. It has more than 4.5 million companies and 4 million people. Here’s an example of a complex ownership chain:

GW_example

To create this, we used a combo of Python, R and Neo4J (with Cypher), hosted in the cloud. Here are a few titbits from our analyses:

  • 4.1 million companies are now on the beneficial ownership register. 3.7 million of these have disclosed at least one beneficial owner
  • 5 beneficial owners control more than 6,000 companies. This prompts questions about whether some are covers put there by the actual owners
  • 4,000 beneficial owners are listed under the age of 2, including one who has yet to be born. Being ‘in control’ of a company is likely a tough ask of a toddler….
  • There are 7,631 PSCs that are companies registered in secrecy jurisdictions – despite this being a violation of UK laws. Secrecy jurisdictions are places where a lack of transparency and unwillingness of national authorities to share information makes it more attractive for routing illicit money flows from crime and corruption, such as the British Virgin Islands

More on our initial findings in this Global Witness blogpost.

Without the UK’s open data register, Global Witness would never have been able to identify these companies and report them to Companies House, which runs the register of UK companies.

Further analysis using the network graph will yield deeper insight of shady patterns of ownership. This is to be published over the coming month. Watch this space!


 

 

DataKind is a global non-profit dedicated to tackling humanity’s greatest challenges with data science. We do this by bringing together data scientists from the private sector to engage in pro bono data projects with social sector organisations.

Global Witness is a leading anti-corruption NGO. They work to break the links between natural resource exploitation, conflict, poverty, corruption and human rights abuses worldwide.

Leave a Reply

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