Tax, anti-tax evaders: from current accounts to bills, an algorithm for crossing databases


Tax, anti-tax evaders: from current accounts to bills, an algorithm for crossing databases
Written by aquitodovale

On the fight against tax evasion, the government tries to give a decisive turn. The crossing of the databases of the Tax to find the taxpayers at the highest risk of tax evasion is about to leave. And this should be the right time. The final anti-evasion weapon has been ready for two years now, but it cannot be lowered to the ground due to the stakes placed by the Privacy Guarantor on the use of a very high number of databases in the availability of the Revenue Agency and the Guardia di Finanza. The dossier, however, would have been released.

In the last council of ministers, the undersecretary to the presidency of the council, Roberto Garofoli, took stock, ministry by ministry, on the implementation of the NRP. When indicating the objectives of the Ministry of Economy for the end of June, he explained that “the draft ministerial decree containing the procedures for the pseudo-anonymization of data by the Revenue Agency was sent to the Privacy Guarantor, for the purpose of obtaining the relative opinion “. And in fact the Authority, according to what was reconstructed by the Messenger, would have received the new draft of the decree on Thursday 26 May. Text that would have incorporated the observations of the same guarantor sent to the Treasury in February of this year. The exam has just begun, but the government pushes for a go-ahead by June so that checks can be started in July. The opinion of the guarantor is the last step necessary to activate the algorithms capable of extracting the taxpayers to be checked from all the taxpayers databases. The intersection of information is one of the fundamental elements envisaged by the recovery and resilience plan for this year. The government has pledged with Europe to increase the sending of compliance letters to taxpayers by 15 percent, in order to push for spontaneous compliance. In practice, these are missives in which the Tax he complains that the taxpayer has not fully done his duty towards the tax authorities and, therefore, invites him to settle the bill before starting a formal investigation.

The increase

The 15 percent increase in letters must be calculated with respect to 2019, when 2.13 million letters were sent which allowed for 1.2 billion euros to be collected. This year, therefore, 2.5 million letters will have to be sent. For this reason, the Revenue Agency and the Finance Police need to be able to cross-reference the data as soon as possible. In order to hit the targets without errors. The list of databases available to cross-reference data, as mentioned, is long: current accounts, old assessments, credit cards, electricity and gas bills, sports expenses, real estate and securities registers and, even, healthcare expenses . Precisely on the use of the latter the guarantor had posed serious doubts of respect for privacy for which, in the new version of the decree, they could be excluded.

The mechanism

But how will the tax regime work? Thanks to artificial intelligence, two “datasets” will be created, essentially two lists of taxpayers. The first of “analysis” will be used to identify whether there are audiences of taxpayers who present a higher risk of tax evasion than the others. Then there will be a “control” dataset. Taxpayers who present one or more tax risks will be included in this list. To avoid risks related to privacy, this list will not be visible with the real names of the taxpayers, but with pseudonyms. The name of the taxpayer will be revealed only at the time of sending the compliance letter or the assessment.
In addition, a new tax simplification law decree could also arrive in the coming weeks. A provision which could also include a series of reinforcements of the instrument of the compliance letter in order to accelerate the recovery of the sums disputed to taxpayers.



#Tax #antitax #evaders #current #accounts #bills #algorithm #crossing #databases


About the author


Leave a Comment