Bercy manages to get its hands on €640M of tax fraud thanks to a sophisticated algorithm

Bercy manages to get its hands on €640M of tax fraud thanks to a sophisticated algorithm

Bercy is refining its research to fight tax fraud and find taxpayers who fail to pay their taxes. And this stalking pays off. The Ministry of the Economy managed to recover no less than 640 million euros in the first nine months of the year thanks to an algorithm, according to Europe 1. A technique, called “datamining” (data mining), allows computers to cross billions of information from huge databases to spot irregularities.

This tool was strengthened a year ago thanks to an investment of 20 million euros in software to analyze and sort the masses of contextual data. Unveiled two years ago, the CFVR algorithm, for targeting fraud and valuing requests, initially focused on a file comprising 5 million companies in France, before integrating data on 37 million taxpayer households, recalls BFMTV.

Monitored social networks

Banking, tax, real estate data, from the USSSAF, the Security… a whole bunch of information is thus examined and cross-checked by the Bercy algorithm. These controls concern individuals but also companies and result in thousands of reports to local tax services. In 2018, 335 million euros had thus been recovered. While the year is not over yet, we are already approaching the double harvested in 2019.

And the fight against fraudsters should intensify even more. Instagram Facebook, Twitter and Instagram, but also on platforms such as Le Bon Coin, eBay or Amazon in order to spot inconsistencies. For example, a person claiming to live in Switzerland but regularly posting photos of himself in Toulouse or Paris may potentially be detected.

An article of the draft Finance law for 2020 provides for a three-year experiment of the tax and customs administration to collect and exploit this type of data posted on social networks.

Go to our cases Get a free quote