Selectivity and State Aid Law

Selectivity is one of the most misunderstood elements of EU State aid law. Unlike the obvious scenario where the State pays money, selectivity can exist in measures that are presented as neutral, general, or simply “pro-growth”. Yet, in practice, selectivity is often the point where a policy tool crosses into the scope of Article 107(1) TFEU.

Put simply, a measure becomes risky when it does not apply equally to undertakings in a comparable situation, but ends up favouring some over others. The decisive factor is not always the legislator’s intention. It is the effect: how the measure operates in the market and who actually benefits from it.

What selectivity means in plain language

Selectivity is about differential treatment. If two undertakings are in a comparable position in light of the rule’s objective, but one receives a better treatment, selectivity becomes an issue. This arises frequently in taxation, but it also appears in subsidies, regulatory exemptions, licensing schemes, or access conditions to public resources.

The problem starts when a measure is not genuinely horizontal. It may be drafted as “open to all”, but in reality only a limited set of undertakings can access it. Good intentions do not automatically prevent selectivity.

The key tool: the “three-step” logic

Selectivity analysis usually follows a three-step logic. First, you identify the “normal” framework, often called the reference framework. In tax matters, this is the general tax rule: the base, the rate, and the standard calculation method. You cannot assess a derogation unless you know the baseline.

Second, you examine whether the measure derogates from that baseline. A derogation can be an exemption, a reduced rate, a special deduction, or an eligibility condition that narrows access in a way that the normal system would not.

Third, even if there is a derogation, you ask whether it can be justified by the nature or the structure of the system itself. This is where confusion is common. Saying “it promotes development” or “it supports the green transition” describes a policy objective, which may matter for compatibility. It does not, by itself, justify selectivity unless it is rooted in the internal logic of the system.

How “general” measures become selective without anyone noticing

A measure can look general because it does not name beneficiaries. Yet selectivity can result from its criteria. Where criteria are such that only a certain type of undertaking can realistically satisfy them, the measure may be selective in effect.

This is often seen when eligibility is tied to a specific legal form, size threshold, region, or technology. Even if “anyone can apply” in theory, only a few can meet the conditions in practice.

A practical example

Assume a reduced tax rate is introduced for undertakings investing in a specific technology. If that technology is realistically accessible only to a limited number of operators, or requires capital that only a small group can mobilise, the scheme may operate selectively. It may be framed as “general” and “green”, but the real-world impact is concentrated on a narrow set of players.

In such a scenario, selectivity may be established. If the other Article 107(1) criteria are satisfied, the measure may qualify as State aid. Only then does the next question arise: whether the aid can be found compatible under the relevant EU framework. Mixing these two phases is a classic mistake.

What this means for businesses and public authorities

For businesses, selectivity matters because a seemingly safe advantage (a tax break or regulatory carve-out) can later carry recovery risk. For public authorities, poorly designed criteria can lead to legal uncertainty, delays, and eventually the need to unwind a measure.

Good practice is to design criteria with maximum objectivity, document how any differentiation is rooted in the system’s logic, and avoid conditions that create a de facto “closed club” of beneficiaries.

FAQ

Does a green transition objective remove selectivity? No. It may support compatibility, but it does not replace the selectivity analysis.

If the measure is theoretically open to all, is it safe? Not necessarily. Effects matter, and de facto selectivity can still arise.

If it targets a sector, is it automatically selective? Often yes, unless the differentiation is objectively grounded in the system’s structure.

Conclusion

Selectivity is the criterion that often “catches” measures that look well designed on paper. The solution is not to avoid targeted policies, but to build them with clear logic, coherent criteria, and proper documentation from the start.

This is general information and not legal advice.