Written by Pieter Baert.
Over the last ten years, the EU has taken several key measures to combat aggressive corporate tax planning, aiming to curb the billions in revenue losses suffered by Member States. However, the variety and breadth of the measures introduced have raised concern about their administrative complexity and overall effectiveness. The Subcommittee on Tax Matters will hold a public hearing on this topic on 15 May 2025.
From post-crisis reforms to simplification initiativesFollowing the Great Recession and the subsequent European sovereign debt crisis, and fuelled by publication of various ‘tax leaks’, the EU set out an ambitious range of initiatives to counter corporate tax avoidance. Numerous laws were put into place to close tax loopholes and increase transparency on the tax practices of (large, multinational) companies, including the Anti-Tax Avoidance Directive (ATAD) and the directives on Administrative Cooperation (DAC) and Public Country-by-Country Reporting (Public CbCR). In 2021, a historic global agreement paved the way for the introduction of a minimum effective corporate tax of 15 % for multinationals in the EU, which entered into force in 2024.
Following the 2024 European elections, the European Commission embarked on an overall strategy to cut red tape and simplify EU legislation. In March 2025, the Council called on the Commission to present a road map before the end of Q3 2025 to reduce the reporting burdens for both tax administrations and taxpayers, eliminate outdated and overlapping tax rules, increase the clarity of tax legislation and streamline the application of tax rules, procedures and reporting requirements. This work should cover both direct and indirect taxation, and should ‘preserve the successful achievements’ the EU has made in this area.
Commission evaluations of the ATAD and DAC are already ongoing, and may lead to legislative changes.
Anti-tax Avoidance DirectiveThe Anti-tax Avoidance Directive (ATAD) set out five key provisions – four specific, one general – to close loopholes that were often abused for aggressive tax avoidance purposes. The ATAD rules entered into force between 2019 and 2022.
Table 1 – ATAD overview
Council Directive (EU) 2016/1164 laying down rules against tax avoidance practices that directly affect the functioning of the internal marketArt. 4Interest limitation ruleDiscourages debt arrangements designed to minimise taxation, limiting the deductibility of taxpayers’ excess borrowing costsArt. 5Exit taxation rulePrevents companies from avoiding tax when relocating assetsArt. 6General anti-abuse rule (GAAR)Counters aggressive tax planning when other rules do not applyArt. 7-8Controlled foreign company (CFC) ruleDeters profit shifting to a dependent company in a low-tax country to reduce taxable profitsArt. 9Hybrid mismatch rulePrevents companies from exploiting national mismatches to avoid taxationThe provisions of the ATAD function as ‘minimum standards’, allowing Member States to introduce stricter provisions, if they choose to. This flexibility has led to some divergence in the application of these standards, resulting in legal complexity. Additionally, the directive includes several options for Member States, often to exclude certain entities from the scope of a particular anti-avoidance rule. Some stakeholder feedback on the ATAD encouraged the Commission to seek more harmonisation in this area. A study conducted for the Subcommittee on Tax Matters (FISC) similarly noted that reducing the number of available options ‘should be considered for a more homogenous anti-avoidance landscape’.
Discussions are also ongoing about the continued relevance of certain ATAD provisions – such as the CFC rules – in the context of the broader global minimum corporate tax framework. While ATAD is designed to target specific tax avoidance loopholes, the EU Minimum Tax Directive focuses on a minimum level of effective taxation, irrespective of any tax planning strategies employed.
Directive on Administrative CooperationAs its name implies, the Directive on Administrative Cooperation (DAC) does not govern the imposition or payment of taxes directly. Instead, it facilitates the collection and, increasingly, the automatic exchange of tax-related information between Member States concerning individuals and companies. Benefiting from increasingly efficient digital tools, Member States can track and cross check income streams, swiftly detect evasion or avoidance practices and impose taxes where required according to national legislation. The DAC has undergone eight revisions (DAC1-DAC9) over the past ten years. This has progressively widened the scope of taxpayers and reportable data.
Table 2 – DAC overview
DACInformation being reported and exchangedDACInformation being reported and exchangedDAC1Income from employment, pension, director fees, income and assets from immovable property and life insuranceDAC6Potentially aggressive tax planning schemes of intermediariesDAC2Financial account data (account balances, gross amount of interest and dividends received …)DAC7Income earned by sellers on digital sales platformsDAC3Advance cross-border tax rulings and advance pricing arrangements of companiesDAC8Income earned by crypto-asset tradersDAC4Country-by-country reports on multinationals (data on revenue, profits, tax paid …)DAC9Top-up tax information returns for the purposes of the Minimum Tax DirectiveDAC5Beneficial ownership and due diligence information as collected through the anti-money laundering legal framework.Stakeholder responses to the public consultation on DAC indicated that many would welcome greater transparency on how, and to what extent, tax authorities use the data reported. The European Court of Auditors raised a similar concern, noting Member States ‘generally underused’ the data reported under DAC1 to DAC5, which were subject to limited data quality checks.
Stakeholders have been particularly critical of DAC6, which obliges intermediaries – such as tax advisors and accountants – to report information on potentially aggressive cross-border tax arrangements to the tax authorities. Stakeholders regarded the criteria for identifying such arrangements, known as ‘hallmarks’, as overly broad or difficult to apply in practice. This concern was echoed in a study commissioned by the FISC Subcommittee, which also warned about the additional administrative costs for tax authorities: ‘Tax authorities will also be heavily impacted due to the [… expected] volume of disclosed information they will receive. The danger of over-reporting due to the over-inclusion and multiplicity of hallmarks (…) is a real one, and may cause tax authorities [to miss …] some red flags’.
Read this ‘at a glance’ note on ‘The future of EU anti-tax avoidance rules‘ in the Think Tank pages of the European Parliament.
Written by Andrés García Higuera.
The European Parliament’s Panel for the Future of Science and Technology (STOA) brought together Members of the European Parliament, the European Commission and researchers on 29 April 2025, to discuss the opportunities and challenges of the use of AI in science, at a workshop entitled ‘Generative AI and scientific development’.
STOA Vice‑Chair Lina Gálvez (S&D, Spain) opened the event, and called for a fruitful debate on all angles of the use of AI in scientific development – from the undeniable advantages, to the risks associated with an extensive deployment of AI and ways to counter them.
A first panel focused on the technologies involved in using AI to help foster scientific research in the EU. Serge Belongie (University of Copenhagen) set the scene with a keynote speech about the challenges and opportunities of AI. Maria Cristina Russo, Director of Prosperity at the European Commission’s Directorate-General for Research and Innovation, then commented on the Commission’s efforts to mobilise €200 billion for investment in AI, including €20 million to develop gigafactories devoted to high-performance cloud storage (HPC), as well as the political priorities of the Competitiveness Compass. She also commented on the AI continent action plan, the AI in science initiative, the communication on a European strategy for AI in science, and the creation of the resource for AI science in Europe (RAISE), as well as on the ‘Choose EU‘ initiative to attract researchers. Francesca Campolongo, Director for Digital Transformation and Data at the Joint Research Centre (JRC), highlighted that the EU is at the forefront of AI research and development, but that this does not translate well enough into innovation. She mentioned the GPT@JRC platform as an example of JRC’s commitment to help unlock the full potential of AI in science. Ana García Robles, of the Big Data Value Association, focused on competitiveness and emphasised the need to consider both small and medium-sized enterprises and big companies. Wijnand Ijsselsteijn (Eindhoven University of Technology) highlighted the relation between data science, AI and psychology.
The second panel focused on how AI affects scientific dialogue and gave an overview of the use and the abuse of AI in science. Oxford University Professor Sonia Contera’s keynote speech framed the way AI affects the exchange of information that is essential in science. Lex Bouter of Vrije Universiteit Amsterdam insisted on the need to focus on the concept of research integrity. Anita De Waard, of Elsevier underlined the need for collaboration among academic and industry partners to improve trust and reproducibility, as well as research integrity. Commenting on these issues, Sebastián Ventura Soto from the University of Córdoba noted how new developments in AI can also be used to tackle them. Finally, Elizabeth Gadd of the Coalition for Advancing Research Assessment highlighted the need for responsible research evaluation and for improved metrics for evaluating research. During the Q&A session which followed, Ana Vasconcelos, (EPP, Portugal) discussed the idea of ‘fake’ scientific publications and the reliability of assessing AI manipulation while insisting on the need for trustworthy scientific communication.
STOA Vice-Chair Lina Gálvez closed the event, underlining the need to reap the full benefits of the development and uptake of AI in science, as promoted by the Commission and the JRC, while avoiding drawbacks such as interfering with a necessary and productive scientific dialogue. She ended by announcing that, since the Commission and JRC are already doing excellent work in promoting the deployment of AI with their different programmes, STOA will complement this effort by launching a study analysing the status of open science and the effects of generative AI in scientific exchange.
Prior to the event, the European Science-Media Hub (ESMH) published an article featuring interviews with the keynote speakers. A web-stream recording of the event, a video and photos, are available on our website.
Your opinion matters to us. To let us know what you think, get in touch via stoa@europarl.europa.eu and follow us on X at @EP_ScienceTech.