On the one hand the twelve forgive inserted in budget law and it criminal shield for evaders convicted in the first instance, on the other hand the intention – confirmed in the tax delegation – to allow the full use of data of taxpayers for risk analyses evasion. In words the invocation of a new “great and definitive fiscal peace“, in fact the expectation of automating the foreclosure current accounts to facilitate the collection. It is full of ambiguity and contradictions the fiscal policy of the Meloni governmentwho winks at voters that taxes are a kind of abuse but at the same time has his hands tied by the goals of the Pnrr on reducing the tax gap. And so, while the prime minister lashes out against the “protection payment of state“, Revenue is – fortunately – already exploiting the information present in theArchive of financial reports and those that come from electronic invoices to locate thanks a algorithm suspicious locations to check with a good chance of collecting something.
“Here comes the Tax big brother. The Voyeur status will know in real time what every single citizen buys and does”, thundered Meloni in November 2018 when the government Count 1 also allowed the Finance to access the Archive, which contains opening and closing balance, average values And movements of cc and deposit accounts. But last May 19, with the approval of the Mef which supervises it, the Agency led by Enrico Maria Ruffini has published on its website a “informative on the underlying logic of risk analysis models” based precisely on those data. That since last summer, after the long-awaited and tormented go-ahead of the Privacy guarantorcan be crossed with the other databases available to identify risk criteria useful for identifying potential offenders.
Reading the technical attachment, one discovers, among other things, that the offices of the Agency are supported, in the selection of the positions to be subjected to control, by a optimization algorithm which allows for order them based on priority. How does it work? The algorithm was trained starting from the data on already controlled taxpayers, which made it possible to identify the information that “they predict” the probable outcome of the investigation. Using those indicators, the algorithm is able to prepare a list of names to be reported to the provincial directorates responsible for controls. The document prepared by the Risk Analysis sector of the Taxpayers Division also contains a practical example in which to finish are i Merchantsa category that has always been pampered by the majority who last year tried to exempt them from the obligation to accept payments with the pos below a certain figure – except to be forced into a sensational retreat after the intervention of the European Commission.
Retail is an area “typically characterized by high margins,” noted Revenue. This is why, to identify probable tax evaders, taxpayers are selected who have a difference between revenues And costs very contained and declaring a value added very small compared to the number of employees e profits below the tenth percentile, i.e. very low compared to competitors. Within this audience, the individuals who saw theirs in the same year are then identified current account increase by a higher figure of at least 150% to the declared revenues and equal to at least 300 thousand euros, unless the sudden fortune is justified by donations, inheritance or the sale of real estate. The result is a list of taxpayers who have most likely tried to hide something from the tax authorities. The algorithm then applies a model which on the basis of many variables (from age to income passing through the assessments already received) attributes to each position the probability that the taxpayer will take action paying what is due or disputing the assessment and the specular one that does nothing and that the Agency therefore registers a role the sums due.
At the end of the analysis process, the officials are provided with “a single indicator which, by measuring the “value” of each position in terms of risk and profitabilitycan be used to assign a order of priority to the positions themselves e optimize the process of scheduling control activities”. With other models, analyzes can be conducted that start from other information available to the Revenue, such as i costs for raw materials which must be compatible with the revenues declared or anomalies in the number of access to safety deposit boxes, in the frequency of opening and closing of accounts, in the increase of assets in a single year. It will then be up to the administration, obviously, to decide how to use that information: “in no way”, specifies the disclosure, “the issuing of tax measures against taxpayers selected through totally automated procedures” is foreseen. The fact remains that these are measures that the leaders of Brothers of Italyagain in February 2022, described how “harassment of the government against Italian citizens with the excuse of the fight against tax evasion”.