Image
Top
Navigation
2021 The University of Florida Award for Investigative Data Journalism, Large Newsroom finalist

The FinCEN Files

About the Project

The FinCEN Files began with an unprecedented cache of secret government documents and grew into the largest reporting collaboration in history — and a massive feat of data journalism.

The partnership, which began with BuzzFeed News and the International Consortium of Investigative Journalists, ultimately included more than 400 journalists from 110 news outlets in 88 countries: https://www.icij.org/investigations/fincen-files/fincen-files-media-partners/

Working together, the partnership delivered an unparalleled view of global financial corruption, banks that enable it, and government agencies that watch as it flourishes.

Our investigation revealed that five global banks — JPMorgan, HSBC, Standard Chartered Bank, Deutsche Bank and Bank of New York Mellon — continued to profit from suspect transactions even after they paid fines to U.S. authorities for previous misconduct and in some cases signed deferred prosecution deals.

We showed, transaction by transaction, how easily some of the world’s most notorious criminals used the U.S. banking system to legitimize the profits from deadly drug wars, the cash that funds terror attacks, and the fortunes embezzled from developing countries.

Even amid a global pandemic and a convulsive U.S. presidential election, this project was an international bombshell.

On the first day of publication alone, the partnership put out more than 100 stories. The impact was immediate: Investors drove the stock price of JPMorgan and other big banks down sharply. Governments around the world launched inquiries. Lawmakers in the U.S. and Europe demanded action. New York’s top banking regulator acknowledged dirty money is “wrapped within the guts of financial institutions.”

The FinCEN Files investigation began when BuzzFeed obtained, through a whistleblower, a remarkable collection of highly secret government documents on file with the Treasury Department’s Financial Crimes Enforcement Network (FinCEN). The documents included more than 2,100 “suspicious activity reports,” or SARs, which financial institutions are required to submit to FinCEN when they see the hallmarks of illegal activity.

SARs are dense blocks of information. Turning them into groundbreaking journalism required a massive data-extraction-and-analysis effort.

Their data tables lay out dates, the amount of money under suspicion, and detailed information about the people and organizations involved, such as addresses, bank accounts, and more. These tables can go on for dozens of pages. We wrote custom software to extract those details and put them in a single database for reporters to search and analyze.

But the narrative sections of the documents totaled 3 million words — 14 times the length of Moby-Dick. We initially tried machine learning, but the contents of the narratives proved too complex. That left the old fashioned approach: we read every word. Then we fact-checked it all, three times over. It took 85 journalists in more than two dozen countries more than a year to complete. The result illuminated more than 200,000 suspect transactions totaling more than $2 trillion — all laid out in greater detail than FinCEN itself provides to law enforcement authorities.