TalTech Journal of European Studies recently published our counsel’s PhD. Kaido Künnapas article about the potential use of Regression Tree Model in automation of legal processes related to the general anti-abuse rule in Anti-Tax Avoidance Directive (ATAD).
The Anti-Tax Avoidance Directive and general anti-abuse rule
The European Union has introduced in its Anti-Tax Avoidance Directive a general anti-abuse test which must be transposed into the domestic laws of Member States. Künnapas finds that designing automated application logic of the ATAD-based GAAR (general anti-abuse rule ) may have practical value in application of law and in the article he seeks to answer the following questions: what is the design of the closed-rule decision tree type algorithm on the GAAR and what are the main challenges related to input information to such a decision-making process.
Benefits of the GAAR automation
International tax law in general seems to be an ideal process to be managed by logic-based machine-learning methods—such as decision trees. In very simplistic form, decision trees follow the construction of ‘if… then… else…’, which is familiar to lawyers as being the syllogism in regulative legal norms and the concept of a syllogism is similar to the concept of an algorithm.
Automation of the GAAR would have many benefits, starting from spotlighting abusive tax schemes for third parties, helping taxpayers to understand what is allowed and what is not and demanding more clarity and analysis in legislative processes. The algorithm design would entail decision points which would rely on information either of objective or subjective nature. While development and integration of digital processes and databases would make the objective information sourcing easier over time, the real issue comes from unclear wording of law.
Read more about the regression tree model for automating the GAAR in the TalTech Journal of European Studies article
