US Tax Inspectorate uses data mining and predictive analytics

In Russia and CIS countries, tax crimes are not considered to be something extraordinary for an individual. Underpaid the tax? Well, pay as much as necessary - and free. Quite a different situation in the United States. There, for concealing taxes, you can stay in prison for a long time, this is one of the most serious crimes before the state, along with treason and espionage in favor of another state. So with the IRS (the so-called local tax inspection) it is better not to joke.

But the IRS itself is becoming more intelligent. It starts applying modern technologies to identify those citizens whose costs do not match officially received incomes. This is reported by the Vanderbilt Journal of Entertainment and Technology Law, with reference to the report of the associate professor of business law at the University of Washington, Kimberly A. Houser, and the accounting professor at the University of Washington, Debra Sanders.

It is reported that the IRS is now engaged in data mining of public and commercial data sets (including social networks) in order to create detailed profiles of taxpayers, through which tax specialists perform data analysis.

Authors of the report say that the methods of work of the tax inspection remain largely unknown to the general public. At the same time, there are doubts about their legality. According to some lawyers, collecting data from social networks and profiling users with data-mining of their revenues can itself violate the law.Although users themselves spread the information on public display, but the tax inspectorate begins to collect on them the profile before the audit begins . The existence of data brokers (companies that collect information on users and resell it to interested persons) and the opportunity to buy information on virtually any person via the Internet create a situation when a person loses control over his personal data. This is especially dangerous if the state in the person of the tax service acts as an interested person.

The lack of transparency and accountability in this process raises even more serious questions. The latent actions of the state on such collection to the state are illegal, and besides can lead to discrimination.

The authors of the report list all methods of collection that the IRS uses, as well as articles of laws that may be violated by the tax inspection during the collection and data mining.

So, the sources of tax information:

• Forms of W-2 from the employer (standard practice).
• Records of telephone conversations . According to the American Civil Liberties Union, the IRS was among the customers of equipment for wiretapping (so-called Stingray devices) in 2009-2012.
• Email . Requests for the Freedom of Information Act in 2013 showed that the tax service looked through the personal mail of taxpayers without a warrant .
• Social networks . Upon request under the Freedom of Information Act, the IRS confirmed that it collects information from social networking sites. The IRS training manual seven years ago lists Facebook, MySpace and YouTube as sources of information about taxpayers.
• Data mining involves analyzing large sets of data that were collected for a purpose different from that with which they are analyzed. The goal is to identify previously unknown relationships in the data. The use of data mining is reported in the internal documents of the IRS . Independent sources say that the tax service thus clarifies the profiles of taxpayers.

The following potential violations of the legislation are possible:

• A. Violation of practices of fair treatment of information .

1. Lack of notice.
2. A secret data collection system (violation of the Personal Rights Protection Act of 1974 (Privacy Act of 1974)).
3. Absence of the taxpayer's consent to contact the IRS with a third party (violation of section 7602 of the Internal Revenue Code).
4. The loss of user control over the use of personal information.
   • B. Lack of transparency in the algorithm .
5. Violation of the Administrative Procedure Act.
6. Data-mining mistakes: inaccurate profiles of taxpayers.
7. Potential discrimination. After self-learning, prediction algorithms will be introduced with prejudice to people, given the skin color, residence and other discriminatory characteristics.
8. Arbitrary and whimsical actions of the agency. For all the above reasons, the use of data mining and self-learning algorithms can lead to the fact that an automated state decision-making system will take decisions that are incomprehensible to the human mind.
   • C. Data collection .
9. Violation of the Electronic Communications Privacy Act of 1986 - there are obvious violations in connection with wiretapping phones and unauthorized reading of e-mail.
10. Searches without orders (viewing mail on the server, which was stored longer than 180 days).
11. Inadequate legal procedure.
12. Self-incrimination - people actually give unfavorable testimony, convict themselves, and forcing them to do so is prohibited by the Fifth Amendment to the Constitution.
    • D. Other federal violations .
13. Violation of the Personal Rights Protection Act of 1974.
14. Violation of the 1988 Law on Computer Comparison and Protection of the Right to Privacy (Computer Matching and Privacy Protection Act 1988).
15. Violation of section 6013 of the Internal Revenue Code, which states that information related to taxes should remain confidential and not be disclosed in any way. It is about the appropriate rules for storing and protecting this information from the IRS, penalties are provided for IRS employees for violating the rules.
16. The Data Quality Act requires federal agencies to take action to verify the quality of the data they collect.

In general, the actions of the American IRS can already violate a number of federal laws. It is possible that in the coming years their advanced methods of data mining will be taken over by Russian colleagues, especially in the budget now there is a large deficit - and the state urgently needs to increase tax collections from citizens.

Then it is better to think twice before publishing in the "contact" pictures of the holiday in the Caymans or a snapshot of the new car that you bought, dragging an online poker tournament. Thought that the tax will not find out about it? Now he finds out.