Open data platform reveals complex corporate structures of banks (Wired UK)

"A London-based company called OpenCorporates has developed a platform that collects global corporate datasets -- many of which weren't available as open data -- and presents them as visualisations. One of the most revealing aspects of the platform is a visualisation of the global corporate networks of the six biggest banks in the US -- Goldman Sachs, Bank of America, Morgan Stanley, Wells Fargo, Citigroup and JP Morgan.  The map aims to illustrate how complex multinational companies are. It plots the number of subsidiaries by country, along with the ownership structure.  Taking Goldman Sachs as an example, the platform shows that in Hong Kong there's a company called Goldman Sachs Structured Products (Asia) Limited that's controlled by another company called Goldman Sachs (Asia) Finance, which is registered in Mauritius. That, in turn, is controlled by a company in Hong Kong, which is controlled by a company in New York, which is controlled by a company in Delaware, and that company is controlled by another company in Delaware called GS Holdings (Delaware). This company is a subsidiary of The Goldman Sachs Group in New York City. There are hundreds of such chains, many of which have around 10 layers of control below the HQ. Goldman Sachs has more than 4,000 separate corporate entities, a third of which are registered in countries such as the Cayman Islands and Mauritius, which one might describe as tax havens. Similar patterns can be observed in all of the other banks. The data has all been taken from publicly available data at the Federal Reserve and the Securities and Exchange Commission. Much of the data was available before only in PDF or web page form, so OpenCorporates had to develop scripts that would isolate and scrape the relevant information in an automated way, so that the database is futureproofed. For each data point, users can find out where the data comes from. This means that if any figures are wrong, OpenCorporates can identify the source of the bad data. The plan is to create similar networks for every single corporation. Networks have already been created for large companies such as Starbucks, Barclays and IBM. The dataset is released as open data and third parties can collaborate by contributing to and correcting the data ..."