Proportionality stricto sensu and the need for an autonomous EU rule-of-law review in direct taxation

Written version of the speech delivered on 21 November 2025 during Assonime’s second International Tax Conference in Rome

Ladies and gentlemen, colleagues, dear friends,

Thanks for the invitation to speak. Thank you for listening.

Article 5 of Protocol No. 2: proportionality stricto sensu

I want to focus on a principle that looks simple but goes to the heart of how the EU should make laws. It is the idea that “the disadvantages caused by EU law must not be disproportionate to the aims to be achieved.”

In the words of the Court of Justice of the European Union (CJEU), an EU measure is not in line with the principle of proportionality stricto sensu:

‘if the resultant disadvantages for certain economic operators were wholly disproportionate to the advantages otherwise offered’1

This is a balancing test: advantages balanced (weight) against the disadvantages.

With this test, the CJEU weighs the importance of the legislative choice made by the Union legislature against the disadvantage it creates for certain market participants. This is known as the proportionality stricto sensu test. The question is whether the seriousness of the interference is proportionate to the weight of the measure’s objective. A measure may be suitable (effective) and necessary (least restrictive), but still excessively burdensome. It is a normative and value-laden assessment.

As Advocate General Saugmandsgaard Øe notes in his Opinion in Tele2 Sverige, proportionality stricto sensu involves “a debate about the values that must prevail in a democratic society and, ultimately, about what kind of society we wish to live in.”2 Although the CJEU generally avoids substituting its own assessment for that of the European Union legislature, the stricto sensu proportionality test nonetheless provides the CJEU with a tool to correct serious and disproportionate interferences on democratic values or on the sovereignty of the Member States. This moral corrective enables the CJEU also to balance individual rights against collective interests.

Proportionality stricto sensu is written in Article 5 of Protocol No. 2 annexed to the EU Treaty and the FEU Treaty3. Even in areas, like in taxation, where the EU legislature must be allowed a broad discretion because it involves political, economic and social choices and in which it is called upon to undertake complex assessments and evaluations4, the EU legislator must take art. 5 of the Protocol into account:

As the Grand Chamber of the CJEU in C-482/17, Czech Republic/Parlement & Council, reminded:

‘Even where it has broad discretion, the EU legislature must base its choice on objective criteria and examine whether the aims pursued by the measure chosen are such as to justify even substantial negative economic consequences for certain operators. Under Article 5 of Protocol (No 2) on the application of the principles of subsidiarity and proportionality, annexed to the EU Treaty and the FEU Treaty, draft legislative acts must take account of the need for any burden falling upon economic operators to be minimised and commensurate with the objective to be achieved’5

But the truth is: we almost never see this principle taken seriously by the EU legislature in EU direct tax lawmaking.

Let me start with a basic question: when the EU proposes new rules, we only see (but not always) an explanation why this exact measure is needed (suitability/effective test), but do we also see an explanation why a lighter alternative would not work? (necessary test) And most importantly: whether the legislation is not disproportionate to the objective to be achieved (stricto sensu)?

Honestly, most of the time, the answer is no.

The rule in Protocol No. 2 says clearly: EU laws must not create unnecessary costs or burdens for citizens. The legislator must show why the chosen solution is not disproportionately burdensome.

But in practice, proportionality is treated as a formal sentence added to the introduction of a proposal, without any real analysis behind it.

The example of the Anti-Tax Avoidance Directive (ATAD)

A good example is the Anti-Tax Avoidance Directive (ATAD)6.

The Commission wrote the following in its explanatory memorandum of the ATAD proposal7:

“The envisaged measures do not go beyond ensuring the minimum necessary level of protection for the internal market. The Directive does not therefore prescribe full harmonisation but only a minimum protection… In this light, the proposal … is therefore compliant with the principle of proportionality.”

This sounds nice. But this is not a proportionality test.

The text basically says:

“We do not fully harmonise; therefore, we are proportionate.”

But proportionality does not ask whether the EU chose full harmonisation or minimum harmonisation.
It asks whether the actual rules of the ATAD creates more burdens than necessary.

That question is not answered.

The interest limitation rule in the ATAD (earning stripping)

Let me give another practical illustration from the ATAD. The directive contains an interest limitation rule, known as the earnings stripping rule (art. 4 ATAD). Under this mechanism, the deductibility of net interest expenses is capped at 30% of the taxpayer’s EBITDA.

In the preamble of the ATAD, the following consideration is set out:

The “key objective of this Directive is to improve the resilience of the internal market as a whole against cross-border tax avoidance practices’’ (BEPS).8

Can the EU legislator clarify why the interest limitation rule also applies in purely domestic situations, where no cross border tax avoidance arises, and why it extends even to interest paid to unrelated third parties such as banks, where there is no BEPS? Neither the proposal nor the preamble offers any justification for this regulatory choice. As an EU citizen, how am I supposed to understand the rationale behind such a broad measure?

No impact assessment for ATAD and pillar 2

There is an additional concern: the ATAD was adopted without an accompanying impact assessment. One of the main reasons was the link with the OECD BEPS project. The logic was: The OECD already did the work, so we don’t need to9.

Exactly the same happened with the Pillar 2 Directive.

Again: no EU impact assessment. Instead, the EU simply relied on OECD estimates. The Commission even wrote that the essential policy decisions were taken at G20/Inclusive Framework level10.

This raises a problem:

If the EU adopts rules negotiated by a non-EU body, without doing its own proportionality test, is the EU legislator really exercising its own judgment, as the Treaties require?

Why is proportionality stricto sensu important?

Let me make three simple points about why proportionality stricto sensu is important:

First, it forces the EU to think before acting.

-       Not only: “Does this help?”

-       But also: “Is this the least burdensome solution?”

-       And: does it justify the burden? (stricto sensu)

This is the heart of the principle. And we almost never see it respected by the EU legislator.

Second, proportionality demands proper preparation.

You cannot claim a measure is the least and not disproportionate burdensome for the taxpayer if you have not compared it with other options. This requires impact assessments. It requires weighing alternatives. It requires studying the real-life consequences. It requires showing your work.

Without this, “proportionality” becomes an empty sentence.

Third, proportionality can bring back a sense of modesty to EU lawmaking.

The EU regulates more and more areas, often at high speed. Proportionality stricto sensu is a reminder: only regulate what is necessary, and do not impose more than needed.

Let me say something about the Court of Justice:

The CJEU usually applies a very light form of review in cases concerning EU legislation. It limits itself to examining whether the Union legislature has committed a “manifest error”, which is an exceptionally high threshold. A measure is not manifestly inappropriate as long as it can contribute effectively to the stated objective11. The instrumental question of whether a less restrictive and equally effective alternative exists is often not asked. This approach means that the CJEU rarely compares different regulatory options itself, but instead verifies whether the EU legislature has remained within reasonable bounds. If a measure can contribute effectively to the aim pursued, the Union legislature is allowed to choose the option it considers necessary.

In more recent cases the CJEU has in fact added a proportionality stricto sensu analysis to its review12. I expect, and welcome, that the CJEU will increasingly focus on the principle of proportionality stricto sensu. In the very recent Familienstiftung case the CJEU has already applied this balancing test directly to national tax law, explicitly referring to the “principle of proportionality stricto sensu” and asking among other things whether the measure “would systematically give rise to a significantly higher tax burden” for the taxpayer13.

It is only a matter of time before the examples I mentioned concerning the interest limitation rule in the ATAD, and other Union law reach the CJEU. Let us see whether the CJEU will take its responsibility and will apply a genuine proportionality stricto sensu test.

But here is more: In case Spain v Council (C-310/04)14, the CJEU ruled that even when the CJEU applies a light test, and the EU legislator has a broad discretion, the EU institutions must still show that they “actually exercised their broad discretion”, meaning that they considered the relevant facts and alternatives.

In the Spain v Council-case, the EU Council could not show this. As a consequence, the CJEU annulled the EU measure.

Imagine the CJEU applying this logic to tax directives15. What would happen if the EU Council had to prove that it truly exercised its broad discretion when adopting ATAD or the Pillar 2 Directive? How can it show that it considered alternatives, when there was no EU impact assessment and when the reasoning mostly relied on OECD assessments?

If proportionality is supposed to ensure decisions are taken “as closely as possible to the citizen,” then copying OECD rules without a proper EU analysis cannot be enough. It reduces the EU legislator to a rubber stamper.

Concluding remarks

Proportionality stricto sensu is not about slowing down the EU. It is about improving the quality of its laws. It is about taking responsibility. It is about doing your own homework rather than repeating what the OECD says. If not, the EU risks being transformed into an implementing agent of OECD rules rather than an independent legislator operating under its own constitutional constraints. The danger is that the Union gradually degrades its role to that of executing externally negotiated standards, while the Treaties require an internally grounded, democratically legitimised review. This is why we need a genuine autonomous EU rule-of-law review in direct taxation.

• 1                 CJEU 17 October 2013, case C-203/12, Billerud Karlsborg,para. 35.
• 2                 See: Opinion in joined Cases C‑203/15 and C‑698/15 Tele2 Sverige, point 248.
• 3                 Protocol (No 2) on the application of the principles of subsidiarity and proportionality.
• 4                 See CJEU 7 March 2017, case C-390/15, RPO, para. 54.
• 5CJEU 3 December 2019, case C-482/17, Czech Republic/Parlement & Council, para. 79.
• 6Council Directive (EU) 2016/1164 of 12 July 2016 laying down rules against tax avoidance practices that directly affect the functioning of the internal market, OJ L 193, 19.7.2016, pp. 1–14.
• 7See: Proposal for a COUNCIL DIRECTIVE laying down rules against tax avoidance practices that directly affect the functioning of the internal market, Brussels, 28.1.2016 COM(2016) 26 final 2016/0011 (CNS), p. 5.
• 8See preamble, para. 16.
• 9See: Proposal for a COUNCIL DIRECTIVE laying down rules against tax avoidance practices that directly affect the functioning of the internal market, Brussels, 28.1.2016 COM(2016) 26 final 2016/0011 (CNS), p. 6.
• 10Proposal for a COUNCIL DIRECTIVE on ensuring a global minimum level of taxation for multinational groups in the Union, Brussels, 22.12.2021 COM(2021) 823 final, p. 4-5: ‘In essence, the important policy decisions have already been taken by the Inclusive Framework and at the highest political level (G20 Finance Ministers and G20 Heads of EN 5 EN States). All EU Member States which are members of the Inclusive Framework have already agreed on the main aspects of Pillar 2 and committed to apply the OECD Model Rules. The EU would have no policy options to choose from as key elements of the framework, such as the scope or tax rates and base, have already been prescribed and agreed on’.
• 11CJEU 10 December 2002, case C-491/01, British America Tobacco, para. 129.
• 12CJEU 3 December 2019, case C-482/17, Czech Republic/Parlement & Council, para. 79.
• 13CJEU 13 November 2025, case C-142/24, Familienstiftung, para. 78-79.Other recent developments in taxation are: the opinion of AG Emiliou in case C-386/24, Centro Petroli Roma, point 109 and AG Rantos, in his Opinion of 13 November 2025 in case C-544/24, Valdymnas, point 49. See for recent literature: Juan Manuel Vázquez, Tax Reporting by Digital Platforms under DAC7: A Proportionality Assessment (Alphen aan den Rijn: Kluwer Law International, 2025) and Vassilis Dafnomilis, Good Intentions, Flawed Design: Balancing Transparency, Fundamental Rights and EU Principles in the EU Public CbC Reporting Directive, World Tax Journal, 2025 (Volume 17), No. 4.
• 14CJEU 7 September 2006, case C-310/04, Spain/EU Council.
• 15  If there is no impact assessment, this does not necessarily mean that the broad discretion has not in fact been exercised. For an example, see: CJEU 3 December 2019, case C-482/17, Czech Republic/Parlement & Council, para. 79. Art 2 of protocol no 2 states: ‘Before proposing legislative acts, the Commission shall consult widely. Such consultations shall, where appropriate, take into account the regional and local dimension of the action envisaged. In cases of exceptional urgency, the Commission shall not conduct such consultations. It shall give reasons for its decision in its proposal’.